Datasets:

gollark

/

consciousness

like 0

Consciousness and Cognition 42 (2016) 1–8 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Individual differences in flow proneness are linked to a dopamine D2 receptor gene variant Mate Gyurkovics a,b,⇑, Eszter Kotyuk b,c, Eniko Rozsa Katonai a,b, Erzsebet Zsofia Horvath d, Andrea Vereczkei d, Anna Szekely b a Doctoral School of Psychology, Eötvös Loránd University, Izabella u. 46, H-1064 Budapest, Hungary Institute of Psychology, Eötvös Loránd University, Izabella u. 46, H-1064 Budapest, Hungary c Postdoctoral Research Program, Hungarian Academy of Sciences, Széchenyi István tér 9, H-1051 Budapest, Hungary d } zoltó u. 37-47, H-1094 Budapest, Hungary Department of Medical Chemistry, Molecular Biology and Pathobiochemistry, Semmelweis University, Tu b a r t i c l e i n f o Article history: Received 21 October 2015 Revised 11 January 2016 Accepted 26 February 2016 Keywords: Flow proneness DRD2 C957T Genetic association a b s t r a c t Flow is a special mental state characterized by deep concentration that occurs during the performance of optimally challenging tasks. In prior studies, proneness to experience flow has been found to be moderately heritable. In the present study, we investigated whether individual differences in flow proneness are related to a polymorphism of the dopamine D2 receptor coding gene (DRD2 C957T rs6277). This polymorphism affects striatal D2 receptor availability, a factor that has been shown to be related to flow proneness. To our knowledge, this is the first study to investigate the association between this trait and a specific gene variant. In a sample of 236 healthy Hungarian adults, we found that CC homozygotes report higher flow proneness than do T allele carriers, but only during mandatory activities (i.e., studying and working), not during leisure time. We discuss implications of this result, e.g., the potential mediators of the relationship. Ó 2016 Elsevier Inc. All rights reserved. 1. Introduction Flow is a special mental state that occurs when people devote concentrated attention to their interaction with the environment (Csikszentmihalyi, 1975; Oláh, 2005). For this state to occur, the activity being pursued must have a level of challenge that perfectly matches the skills of the individual. The goals of the activity must be clear and there must be constant feedback on how close the realization of these goals is (Nakamura & Csikszentmihalyi, 2002). Flow experience has many typical attributes, e.g. loss of self-consciousness, or distortion of time experience. Characteristically, actions feel effortless as if individuals were in a ‘‘flow” toward the goal (hence the name of the phenomenon; Csikszentmihalyi & Csikszentmihalyi, 1988). Large individual differences have been found in the frequency and intensity of flow experiences (Asakawa, 2004, 2010; Ullén et al., 2012). Since flow is the product of an interaction between individuals and their environment, these differences are partly due to situational causes (e.g., Kowal & Fortier, 1999; Moreno, Cervelló, & González-Cutre, 2010). However, results also indicate that people participate in these interactions with different levels of proneness to experience flow (De Manzano et al., 2013; Teng, 2011; Ullén et al., 2012). Flow proneness has been found to be linked to certain personality traits. Recent ⇑ Corresponding author at: Institute of Psychology, Eötvös Loránd University, Izabella u. 46, H-1064 Budapest, Hungary. E-mail address: mate.gyurkovics@gmail.com (M. Gyurkovics). http://dx.doi.org/10.1016/j.concog.2016.02.014 1053-8100/Ó 2016 Elsevier Inc. All rights reserved. 2 M. Gyurkovics et al. / Consciousness and Cognition 42 (2016) 1–8 research (Ross & Keiser, 2014; Ullén et al., 2012) suggests that it correlates positively with the Conscientiousness Big Five personality dimension and negatively with the Neuroticism factor. Flow prone individuals were also found to score higher on the Novelty Seeking and Persistence scales of the Temperament and Character Inventory (Teng, 2011). Ullén et al. (2012) found no significant association between flow proneness and intelligence. This indirectly supports the notion that the subjectively effortless attention experienced while in flow stems from different mechanisms than the attentional processes exerted during mental effort (Ullén, de Manzano, Theorell, & Harmat, 2010). According to Dietrich (2004) effortless attention might involve transient hypofrontality on a neurological level. Furthermore, on a neurobiological level high flow proneness has recently been linked to higher striatal dopamine D2 receptor availability in a study using PET scanning (De Manzano et al., 2013). This relationship was most prominent in the dorsal striatum, possibly implying the role of the nigrostriatal dopamine pathway. This is further corroborated by the results of Ulrich, Keller, Hoenig, Waller, and Grön (2014) who found increased activation in the putamen in the dorsal striatum during an experimentally induced flow state. A recent study of 444 adult twin pairs (Mosing, Magnusson et al., 2012) found that flow proneness was moderately heritable (h2 = .29–.35). In most of the studies cited above (De Manzano et al., 2013; Mosing, Magnusson et al., 2012; Ullén et al., 2012) the Swedish Flow Proneness Questionnaire (Ullén et al., 2012) was used to assess flow proneness. This measure comprises three scales, each assessing the frequency of flow experience in a major domain of life: work, maintenance, and leisure. According to the findings of Mosing, Magnusson et al., 2012, the same genetic factors influence flow proneness regardless of the domain assessed. Based on findings above flow proneness is at least partially influenced by genetic factors. However, as of yet there have been no empirical attempts to determine which genes affect this phenotype. According to our best knowledge, this is the first psychogenetic association study in the literature analyzing the genetic background of flow proneness. As flow is a highly rewarding experience, it has been linked theoretically to the dopamine neurotransmitter system (Marr, 2000; Mosing, Pedersen et al., 2012; Peifer, 2012) that plays a key role in the reward circuitry of the brain. Furthermore, dopaminergic pathways are also related to impulse control (e.g., Dalley & Roiser, 2012) which may be important in the sustained task focus necessary to achieve flow (De Manzano et al., 2013). Recently De Manzano et al. (2013) empirically demonstrated a link between flow proneness and dopamine D2 receptor availability. Thus, in the present study we aimed to find associations between individual variance in flow proneness and variations of the gene coding dopamine D2 receptors. The C957T (rs6277) single-nucleotide polymorphism (SNP) in the DRD2 gene which is located on chromosome 11 and codes the dopamine D2 receptor has been shown to influence striatal dopamine receptor availability (see the recent review by Nemoda, Szekely, & Sasvari-Szekely (2011)). With respect to this polymorphism, Hirvonen et al. (2004, 2005, 2009) found that in vivo the T allele had a demonstrable additive effect on dopamine D2 receptor availability. Since striatal dopamine D2 receptor availability is positively linked to flow proneness (see De Manzano et al., 2013 above), in the present study we assumed that the DRD2 C957T SNP may influence flow proneness as well. Specifically, we hypothesized an additive effect of the T allele (thus CC < CT < TT) on flow proneness. 2. Method 2.1. Sample A non-clinical sample of 322 Hungarian subjects (233 females) was used. Mean age in the sample was 21.4 (±3.7 SD) years. Subjects were college students and their genetically non-related acquaintances. Due to the nature of data collection, different subsamples of this main sample were used in the analyses. These three subsamples are further described in the chapter ‘‘Validity and reliability of genotype and phenotype data”. 2.2. Materials To assess flow proneness, the Swedish Flow Proneness Questionnaire (SFPQ; Ullén et al., 2012) was adapted to Hungarian. The original version of this measure has three scales, each concerning one major domain of life (work – FP-Work scale, maintenance – FP-Maintenance, and leisure – FP-Leisure). For each scale, participants indicate the frequency of different elements of the flow experience in the given domain (7 items per domain). In our study, the Maintenance scale was dropped for the second wave of data collection and the instruction of the Work scale was also modified to gauge flow frequency in the context of academic activities (FP-Studies). Instructions were changed to: ‘‘When you are studying or doing academic work, how often does it happen that. . .”. 2.3. Procedure 2.3.1. Questionnaire data Data was collected in two waves. During the first wave, participants (N = 148) completed the Hungarian questionnaire online after a request received via e-mail. As part of previous studies, subjects’ genotypes had already been determined by a team of molecular geneticists collaborating with our psychological laboratory. After analyzing the results of this wave 3 M. Gyurkovics et al. / Consciousness and Cognition 42 (2016) 1–8 of data collection we found that since the sample consisted mostly of college students, many subjects (52.8%) were not actively employed thus had no FP-Work data. For this reason, work-related items were swapped for studying/academic activity items (FP-Studies, see Materials) in the second wave. In this wave, participants (N = 174) received a paper–pencil version of the revised FPQ as part of a questionnaire battery. Initial (post-1st wave) analysis showed no association between FP-Maintenance scores and the examined polymorphism (results are not reported in detail, data available from the authors on request). Therefore this scale was dropped in the paper–pencil version to make the test battery shorter. As a result, all 322 participants had FP-Leisure scores (1st and 2nd waves), 148 participants had FP-Maintenance scores (1st wave only), 65 had FP-Work scores (1st wave only) and 174 had FP-Studies scores (2nd wave only). 2.3.2. DNA isolation and genotyping Buccal DNA samples were collected from the subjects with a non-invasive method. DNA isolation was performed as described elsewhere (Kotyuk et al., 2013). The C957T polymorphism of the dopamine D2 receptor coding gene (rs6277) was determined by real time PCR using about 1 ng purified DNA/sample and a pre-designed TaqManÒ kit (C__11339240_10, Applied BioSystemsÒ, Foster City, US) on a 7300 Real-Time PCR System. The study protocol was designed in accordance with guidelines of the Declaration of Helsinki, and was approved by the appropriate ethics committees. Subjects signed a written informed consent prior to participating. 2.4. Statistical analysis Chi-square analyses were carried out for assessment of genotype frequencies. Independent samples t-tests were used to assess sex differences; relationship with age was tested by Pearson correlation analysis. Association analyses were carried out by multiple linear regression and ANCOVAs. Missing data was handled by removing subjects who had left one or more items unanswered in the given scale. For this reason, 4 participants had no FP-Leisure scores and 2 participants had no FPStudies scores. No outliers with scores outside three standard deviations of the mean were identified. 3. Results 3.1. Validity and reliability of genotype and phenotype data Descriptive statistics for the phenotype scales (FP-Maintenance, FP-Leisure, FP-Work, and FP-Studies) are presented in Table 1. Validity of genotype data was examined in those three subsamples that were later used in association analyses. The measured genotype frequencies (see Table 2 in Section 3.2) corresponded to the Hardy–Weinberg equilibrium in the sample of those who had both FP-Leisure scores and valid DRD2 data (N = 317; v2 = .04, p = .98). Similar results were found for those participants who had both FP-Work scores and valid DRD2 data (N = 65; v2 = .22, p = .90), and in the sample of those who had FP-Studies scores and valid DRD2 data (N = 171; v2 = .02, p = .99). In conclusion, genotype frequencies were in Hardy–Weinberg equilibrium (Hardy, 1908) in all three subsamples. Possible main effect of age and sex was also investigated in the subsamples. Sex had a marginal effect on FP-Studies (t (170) = 1.70; p = .09), scores tended to be lower (3.30 ± .60 SD) for males (N = 55) as compared to females (3.47 ± .58 SD, N = 117). Sex had a significant effect on FP-Leisure (t(316) = 3.23; p = .001): men (N = 87) scored 3.57 (±.56 SD) on average, whereas for women (N = 231) the mean was 3.78 (±.50 SD). Age was not significantly correlated with any of the FP scales. Thus, Sex was used as a covariate in further association analyses. 3.2. Genetic association analysis: DRD2 and flow proneness We investigated associations between the DRD2 C957T polymorphism and the three flow proneness scales (Work, Studies, and Leisure) using ANCOVAs. See Table 2 for descriptive data. Significant results emerged for the Work (F(1, 62) = 5.62; p = .02) and Studies (F(1, 168) = 4.57; p = .03) domains when participants were grouped according to presence of the T allele (see bottom half of Table 2). Table 1 Descriptive statistics for scales of the Flow Proneness Questionnaire. FPQ scale N Mean (±SD) Cronbach’s a FP-Leisure FP-Work FP-Studies FP-Maintenance 318 65 172 148 3.72 (±.53) 3.74 (±.50) 3.41 (±.59) 3.53 (±.54) .68 .61 .71 .60 Note: FP-Leisure = the leisure scale of the Flow Proneness Questionnaire (FPQ); FP-Work = the work scale of the FPQ; FP-Studies = the studies scale of the modified FPQ; FP-Maintenance = the maintenance scale of the FPQ. Standard deviation values in parentheses. 4 M. Gyurkovics et al. / Consciousness and Cognition 42 (2016) 1–8 Table 2 Flow proneness as a function of DRD2 C957T genotype and presence of 957T allele. DRD2 C957T genotype N Mean FP-Work N Mean FP-Studies N Mean FP-Mandatory activities N Mean FP-Leisure CC CT TT 12 34 19 4.04 (±.50) 3.70 (±.54) 3.64 (±.36) 36 86 49 3.58 (±.50) 3.31 (±.57) 3.46 (±.66) 48 120 68 3.70 (±.53) 3.42 (±.58) 3.51 (±.59) 70 156 91 3.65 (±.46) 3.70 (±.52) 3.81 (±.57) T absent T present 12 53 4.04 (±.50) 3.67 (±.48) 36 135 3.58 (±.50) 3.37 (±.60) 48 188 3.70 (±.53) 3.45 (±.59) 70 247 3.65 (±.46) 3.74 (±.54) Note: FP-Work = work scale of Flow Proneness Questionnaire (FPQ); FP-Studies = studies scale of modified FPQ; FP-Mandatory Activities: work and studies scales of FPQ combined (see text). T absent = no T allele present (CC genotype); T present = at least one T allele present (CT and TT genotypes). Standard deviation values in parentheses. Table 3 Multiple linear regression predicting flow proneness during mandatory activities. Variable Standardized coefficient t value T allele present Sex Type of scale .19** .12 .24*** 3.02 1.87 3.83 Note: Predictor variables are coded as follows: T allele present: 0 = ‘‘not present”, 1 = ‘‘present’; Sex: 0 = ‘‘male”, 1 = ‘‘female”; type of scale: 0 = ‘‘Work”, 1 = ‘‘Studies”. Dependent variable was FP-Mandatory Activities scores (flow proneness during working and studying). Adjusted R2 of the model is .10 (F (3, 232) = 9.40; p < .001). ** p < .005. *** p < .001. Because these two domains represent very similar areas of life (mandatory non-leisure activities) and since the pattern of results was similar for both scales, they were combined into a new scale, FP-Mandatory Activities. This also increased sample size (N = 236). We examined the relationship of the DRD2 C957T polymorphism and FP-Mandatory Activities by using multiple linear regression analysis. The main predictor variable was the binary variable ‘‘presence of the T allele” (0 = no T allele; 1 = at least one T allele). Sex was also entered as a predictor of no interest to control for its potential confounding effect. We also entered a binary variable – ‘‘type of scale” – coding which scale (i.e., work or studies) the participant answered. As can be seen in Table 3, presence of the T allele was a significant (p < .005) predictor of flow proneness scores even after controlling for confounding variables. Type of scale also had a significant main effect. In two follow-up ANOVAs we investigated whether presence of the T allele affects FP-Mandatory Activities scores in interaction with sex or type of scale. Neither of the interactions was significant. This suggests that the T allele has a similar association with flow proneness in the life domains of both studying and working, and for women and men (i.e., T allele carriers report lower levels than CC homozygotes). Finally, we examined whether the effect of the polymorphism was significantly different across domains of life (mandatory activities and leisure). Only participants with available information on both scales (FP-Leisure and FP-Mandatory Activities) were included in this analysis (N = 232; 48 CC homozygotes). We conducted a 2 (FP-Leisure, FP-Mandatory Activities)  2 (T allele present, T allele absent) mixed-design ANCOVA with life domain as the withinsubject factor, presence of T allele as the between-subject factor and sex as a covariate of no interest. Life domain had no main effect, but the main effect of presence of T allele was marginally significant (F(1, 229) = 3.11, p = .08). Carriers of the 3.8 Mean FPQ score 3.7 3.6 3.5 3.4 T absent T present 3.3 Mandatory Acvies Leisure Life Domain Fig. 1. The effect of the DRD2 957T allele in two different domains of life. Mandatory Activities refers to FP-Mandatory Activities scores, whereas Leisure refers to FP-Leisure scores. Error bars represent standard errors. M. Gyurkovics et al. / Consciousness and Cognition 42 (2016) 1–8 5 T allele showed lower FP-scores than non-carriers (3.60 ± .43 as opposed to 3.73 ± .44). More importantly, as can be seen in Fig. 1, a significant interaction was found between life domain and presence of T allele (F(1, 229) = 8.025, p = .005). CC homozygotes reported similar levels of flow proneness in both domains (t(47) = .08, p = .94), but T allele carriers reported significantly lower levels during mandatory activities than during leisure (t(183) = 6.44, p < .0001). As shown in our previous analysis, the two groups – carriers and non-carriers – only differed significantly from each other in the mandatory activities domain. 4. Discussion The goal of the present study was to explore the relationship between a dopamine D2 receptor gene polymorphism and proneness to experience flow state in a sample of healthy Hungarian adults. We found a statistically reliable association suggesting that the neurobiology of flow experience does involve dopaminergic pathways, as has been theorized (Marr, 2000) and – partly – demonstrated (De Manzano et al., 2013). Furthermore, our results support the general notion that flow proneness is related to dopamine D2 receptor availability in the striatum. The association we detected was relatively small in magnitude; however, it is important to take into account that this is the effect of a single polymorphism on a highly polygenic and very complex human trait. 4.1. Psychogenetic associations We examined a dopaminergic candidate gene that was found to be linked to flow proneness as measured by the Hungarian adaptation of the Swedish Flow Proneness Questionnaire. The C957T polymorphism of the dopamine D2 receptor coding gene showed an association with proneness to experience flow even after controlling for confounding variables. However, the pattern of results deviated from expectations: the additive effect of the 957T allele was not observed. Quite contrarily, carriers of at least one T allele scored significantly lower in flow proneness in the life domain of mandatory productive activities (working and studying) than did CC homozygotes. To explain these unexpected results, we might look at other correlates of the C957T polymorphism. Colzato and colleagues (Colzato, Van den Wildenberg, & Hommel, 2013; Colzato, Van den Wildenberg, Van der Does, & Hommel, 2010) found that CC homozygotes show higher efficiency in response inhibition than T allele carriers and, in a complementary fashion, TT homozygotes report higher levels of dysfunctional impulsivity in self-report measures than C allele carriers. Thus, CC homozygotes seem to be characterized both by lower impulsivity and higher flow proneness. In turn, flow proneness and impulsivity might be negatively related because highly impulsive individuals may find it difficult to maintain continuous task focus. Thus, they may be less likely to enter and then remain in the flow channel (as proposed in De Manzano et al., 2013; Mosing, Magnusson et al., 2012). As Peifer (2012) suggested, task-irrelevant processes need to be down-regulated during flow enabling the individual to focus attention on channels of importance. In the case of impulsive individuals, this down-regulation may be deficient as impulsivity is characterized by a failure to inhibit prepotent but inappropriate responses (Cross, Copping, & Campbell, 2011) and an inability to adequately focus attention (see attentional impulsivity; Patton, Stanford, & Barratt, 1995). Furthermore, impulsivity and flow proneness also share correlates on the personality level – Conscientiousness and Neuroticism (Ullén et al., 2012), and on the neurobiological level: dopamine D2 receptor availability, specifically in the dorsal striatum, is positively correlated with both flow proneness (De Manzano et al., 2013) and response inhibition efficiency, a behavioral measure of impulsivity (Ghahremani et al., 2012; Robertson et al., 2015). Our research group is currently investigating the relationship between flow proneness and impulsivity (as measured by a selfreport questionnaire), and preliminary results do show a significant medium negative correlation between the two variables. All these considerations might suggest that the polymorphism in question influences flow proneness via impulsivity or that this polymorphism affects core attentional processes that are, in turn, related to both impulsivity and flow proneness. Regarding this potential mediating variable, it is of importance to note that results concerning the relationship between D2 receptor availability and impulsivity are very inconsistent. Some studies (Buckholtz et al., 2010; Lee et al., 2009) suggest a negative correlation in line with the cited results of Ghahremani et al. (2012) and Robertson et al. (2015), whereas some studies report a positive correlation (Kim et al., 2014; Reeves et al., 2012). Furthermore, our results are at odds with the positive correlation between striatal D2 receptor availability and flow proneness. The present findings link higher flow proneness to a genotype associated with lower striatal D2 receptor availability (as demonstrated by Hirvonen et al., 2004, 2005, 2009). In sum, at this point results linking the genetic level, the neurobiological level, and the phenotypic level appear to be inconsistent. All in all, however, it appears clear that flow proneness and striatal D2 receptor availability are associated in some way. There is another interesting possible interpretation of the DRD2 polymorphism-flow proneness link we found, albeit with less firm theoretical basis. As suggested by previous lines of research, there exists a detectable link between the CC genotype and schizophrenia (see Betcheva et al., 2009; Hoenicka et al., 2006; Hänninen et al., 2006; Lawford et al., 2005; Monakhov, Golimbet, Abramova, Kaleda, & Karpov, 2008). Investigating the construct of schizotypy – which the authors related to schizophrenia in a dimensional approach, Nelson and Rawlings (2010) found that in a sample of creative individuals, positive schizotypy was a predictor of absorption, distinct experience, and pleasure during creative activity – all core elements of flow experience. The authors suggest that reduced latent inhibition may be behind the relationship between the constructs of positive schizotypy and creative experience, and it is this reduction in the ‘‘filtering” of outside stimuli that is subjectively experienced as flow state. Our research could suggest that the DRD2 C957T polymorphism may be a common genetic factor 6 M. Gyurkovics et al. / Consciousness and Cognition 42 (2016) 1–8 behind both the schizotypy/schizophrenia dimension and flow proneness/frequency of flow, and that it may also be related to latent inhibition. Future work could further investigate this intriguing possibility. On a molecular level, the T allele of the C957T polymorphism is also linked to decreased mRNA stability and translation, and a decreased response to dopamine-induced up-regulation of DRD2 gene expression (Duan et al., 2003). Based on our results, it is possible that these biological factors are associated with lower flow proneness. It is important to note that in the present study results concerning the DRD2 gene only reached significance in the life domain of working and studying, activities that may be termed both mandatory (in a sense that, although usually chosen at liberty, both activities entail situations that are compulsory and restrictive in nature) and productive (in a sense that usually both have a definable goal toward which the individual works). No effect was found for leisure time activities. According to our analyses, this was due to the fact that T allele carriers reported lower frequency of flow experiences during mandatory activities than during free time activities. CC homozygotes, however, reported similar frequency across domains. This could suggest that T allele carriers are less able to find intrinsic enjoyment in situations that are not entirely free-choice. If T allele carriers are indeed more impulsive, it is probable that these individuals find restrictive situations more frustrating than CC homozygotes. Frustration could then create a mental state that is not conducive to flow. This idea is in accordance with the hypothetical DRD2-impulsivity-flow proneness account suggested above, and could explain why the psychogenetic effect appears domain-specific. In sum, our results could possibly hint at a gene-environment (G  E) interaction where the phenomenologically detrimental effect of the T allele is only apparent in situations that are restrictive in nature. The apparent domain-specificity of the association could also indicate that there might be different neurobiological mechanisms involved in flow in different contexts. Previous results indirectly support this assumption, as the correlations between flow proneness and D2 receptor availability in striatal subregions were stronger in the work domain than in other domains in the study conducted by De Manzano et al. (2013). Nevertheless, they were detectable, albeit smaller, in other contexts too in the Swedish sample. This could suggest that our study was simply underpowered to show the psychogenetic association in the leisure domain as the contribution of striatal dopamine D2 receptor availability is much smaller there. 4.2. Strengths, limitations and future directions To our knowledge, the present study is the first to investigate the genetic background of flow proneness. Our results are generally in line with previous findings of De Manzano et al. (2013) and further support the empirical link between flow theory and dopaminergic functions. We focused on a polymorphism that has been implied clearly by previous research as it is linked to neurobiological factors involved in flow proneness. Although we found a significant association between the DRD2 C957T polymorphism and flow proneness during mandatory activities, the direction of the results is opposite to expectations based on neurobiological data. We have provided alternative explanations for our findings, setting clear paths for future research. Naturally, first and foremost, our results need to be replicated on bigger samples with explicit emphasis on the potential G  E interaction outlined here. Our study is limited in a number of ways. The main limitation is the relatively small sample size in some calculations. Variability in sample size across analyses was a result of the nature of the data collection. A further limitation is that the questionnaire measure that was used may not have been sufficiently sensitive. The FPQ does not differentiate clearly between frequency of flow experience (a function of both situational and personal characteristics) and flow proneness (a personality trait-like concept). Interviews or newly developed questionnaires might mitigate this problem in future studies. In our study we focused solely on the dopamine system and one single genetic variant contributing to this system. A multi-gene approach in future studies could help clarify possible gene  gene interactions affecting this complex trait. 4.3. Conclusion In our study, we demonstrated that the frequency of flow experience during mandatory activities (e.g., working and studying) is related to a dopamine D2 receptor coding gene polymorphism. CC homozygotes of the DRD2 C957T SNP report higher levels of flow proneness than do carriers of the T allele. The association might be mediated by impulsivity (i.e., CC homozygotes might be less impulsive, enabling them to enter – and then remain in – the flow state more easily in certain situations). Research regarding flow proneness is important as it is a trait positively related to many beneficial outcomes, e.g., life satisfaction and self-esteem (Asakawa, 2010). Funding This work was supported by the Hungarian Scientific Research Fund (OTKA K100845). We are also grateful for the support of the Active Psychology Foundation. Declaration of interests The authors have no competing interests to declare. M. Gyurkovics et al. / Consciousness and Cognition 42 (2016) 1–8 7 References Asakawa, K. (2004). Flow experience and autotelic personality in Japanese college students: How do they experience challenges in everyday life? Journal of Happiness Studies, 5(2), 123–154. http://dx.doi.org/10.1023/B:JOHS.0000035915.97836.89. Asakawa, K. (2010). Flow experience, culture, and well-being: How do autotelic Japanese college students feel, behave, and think in their daily lives? Journal of Happiness Studies, 11(2), 205–223. http://dx.doi.org/10.1007/s10902-008-9132-3. Betcheva, E. T., Mushiroda, T., Takahashi, A., Kubo, M., Karachanak, S. K., Zaharieva, I. T., ... Toncheva, D. I. (2009). Case-control association study of 59 candidate genes reveals the DRD2 SNP rs6277 (C957T) as the only susceptibility factor for schizophrenia in the Bulgarian population. Journal of Human Genetics, 54(2), 98–107. http://dx.doi.org/10.1038/jhg.2008.14. Buckholtz, J. W., Treadway, M. T., Cowan, R. L., Woodward, N. D., Li, R., Ansari, M. S., ... Zald, D. H. (2010). Dopaminergic network differences in human impulsivity. Science, 329(5991), 532. http://dx.doi.org/10.1126/science.1185778. Colzato, L. S., Van den Wildenberg, W. P. M., & Hommel, B. (2013). The genetic impact (C957T-DRD2) on inhibitory control is magnified by aging. Neuropsychologia, 51(7), 1377–1381. http://dx.doi.org/10.1016/j.neuropsychologia.2013.01.014. Colzato, L. S., Van den Wildenberg, W. P. M., Van der Does, A. J. W., & Hommel, B. (2010). Genetic markers of striatal dopamine predict individual differences in dysfunctional, but not functional impulsivity. Neuroscience, 170(3), 782–788. http://dx.doi.org/10.1016/j.neuroscience.2010.07.050. Cross, C. P., Copping, L. T., & Campbell, A. (2011). Sex differences in impulsivity: A meta-analysis. Psychological Bulletin, 137(1), 97–130. http://dx.doi.org/ 10.1037/a0021591. Csikszentmihalyi, M. (1975). Beyond boredom and anxiety. Experiencing flow in work and play.San Francisco, CA: Jossey-Bass Publishers. Csikszentmihalyi, M., & Csikszentmihalyi, I. S. (1988). Optimal experience: Psychological studies of flow in consciousness. Cambridge University Press. Dalley, J. W., & Roiser, J. P. (2012). Dopamine, serotonin and impulsivity. Neuroscience, 215, 42–58. http://dx.doi.org/10.1016/j.neuroscience.2012.03.065. De Manzano, Ö., Cervenka, S., Jucaite, A., Hellenäs, O., Farde, L., & Ullén, F. (2013). Individual differences in the proneness to have flow-experiences are linked to dopamine D2-receptor availability in the dorsal striatum. Neuroimage, 67, 1–6. http://dx.doi.org/10.1016/j.neuroimage.2012.10.072. Dietrich, A. (2004). Neurocognitive mechanisms underlying the experience of flow. Consciousness and Cognition, 13(4), 746–761. http://dx.doi.org/10.1016/ j.concog.2004.07.002. Duan, J., Wainwright, M. S., Comeron, J. M., Saitou, N., Sanders, A. R., Gelernter, J., & Gejman, P. V. (2003). Synonymous mutations in the human dopamine receptor D2 (DRD2) affect mRNA stability and synthesis of the receptor. Human Molecular Genetics, 12(3), 205–216. http://dx.doi.org/10.1093/hmg/ ddg055. Ghahremani, D. G., Lee, B., Robertson, C. L., Tabibnia, G., Morgan, A. T., De Shetler, N., ... London, E. D. (2012). Striatal dopamine D2/D3 receptors mediate response inhibition and related activity in frontostriatal neural circuitry in humans. The Journal of Neuroscience, 32(21), 7316–7324. http://dx.doi.org/ 10.1523/JNEUROSCI.4284-11.2012. Hänninen, K., Katila, H., Kampman, O., Anttila, S., Illi, A., Rontu, R., ... Lehtimäki, T. (2006). Association between the C957T polymorphism of the dopamine D2 receptor gene and schizophrenia. Neuroscience Letters, 407(3), 195–198. http://dx.doi.org/10.1016/j.neulet.2006.08.041. Hardy, G. H. (1908). Mendelian proportions in a mixed population. Science, 28(706), 49–50. Hirvonen, M., Laakso, A., Någren, K., Rinne, J. O., Pohjalainen, T., & Hietala, J. (2004). C957T polymorphism of the dopamine D2 receptor (DRD2) gene affects striatal DRD2 availability in vivo. Molecular Psychiatry, 9(12), 1060–1061. http://dx.doi.org/10.1038/sj.mp.4001561. Hirvonen, M., Laakso, A., Någren, K., Rinne, J. O., Pohjalainen, T., & Hietala, J. (2005). C957T polymorphism of the dopamine D2 receptor (DRD2) gene affects striatal DRD2 availability in vivo (Corrigendum). Molecular Psychiatry, 10(9), 889. http://dx.doi.org/10.1038/sj.mp.4001707. Hirvonen, M. M., Laakso, A., Någren, K., Rinne, J. O., Pohjalainen, T., & Hietala, J. (2009). C957T polymorphism of dopamine D2 receptor gene affects striatal DRD2 in vivo availability by changing the receptor affinity. Synapse (New York, N. Y.), 63(10), 907–912. http://dx.doi.org/10.1002/syn.20672. Hoenicka, J., Aragüés, M., Rodríguez-Jiménez, R., Ponce, G., Martinez, I., Rubio, G., ... Palomo, T. (2006). Psychosis Addictions Research Group (PARG) (C957T DRD2 polymorphism is associated with schizophrenia in Spanish patients. Acta Psychiatrica Scandinavica, 114(6), 435–438. http://dx.doi.org/10.1111/ j.1600-0447.2006.00874.x. Kim, J. H., Son, Y. D., Kim, H. K., Lee, S. Y., Kim, Y. B., & Cho, Z. H. (2014). Dopamine D2/3 receptor availability and human cognitive impulsivity: A highresolution positron emission tomography imaging study with [11C] raclopride. Acta Neuropsychiatrica, 26(01), 35–42. http://dx.doi.org/10.1017/ neu.2013.29. Kotyuk, E., Keszler, G., Nemeth, N., Ronai, Z., Sasvari-Szekely, M., & Szekely, A. (2013). Glial cell line-derived neurotrophic factor (GDNF) as a novel candidate gene of anxiety. PLoS ONE, 8(12), e80613. http://dx.doi.org/10.1371/journal.pone.0080613. Kowal, J., & Fortier, M. S. (1999). Motivational determinants of flow: Contributions from self-determination theory. The Journal of Social Psychology, 139(3), 355–368. http://dx.doi.org/10.1080/00224549909598391. Lawford, B. R., Young, R. M., Swagell, C. D., Barnes, M., Burton, S. C., Ward, W. K., ... Morris, C. P. (2005). The C/C genotype of the C957T polymorphism of the dopamine D2 receptor is associated with schizophrenia. Schizophrenia Research, 73(1), 31–37. http://dx.doi.org/10.1016/j.schres.2004.08.020. Lee, B., London, E. D., Poldrack, R. A., Farahi, J., Nacca, A., Monterosso, J. R., ... Mandelkern, M. A. (2009). Striatal dopamine D2/D3 receptor availability is reduced in methamphetamine dependence and is linked to impulsivity. The Journal of Neuroscience, 29(47), 14734–14740. http://dx.doi.org/10.1523/ JNEUROSCI.3765-09.2009. Marr, A. J. (2000). In the zone: A bio-behavioristic analysis of Csikszentmihalyi’s flow experience. <http://www.athleticinsight.com/Vol3Iss1/Commentary. htm#Theory> (Retrieved from July, 2011). Monakhov, M., Golimbet, V., Abramova, L., Kaleda, V., & Karpov, V. (2008). Association study of three polymorphisms in the dopamine D2 receptor gene and schizophrenia in the Russian population. Schizophrenia Research, 100(1), 302–307. http://dx.doi.org/10.1016/j.schres.2008.01.007. Moreno, J. A., Cervelló, E., & González-Cutre, D. (2010). The achievement goal and self-determination theories as predictors of dispositional flow in young athletes. Anales de Psicología, 26(2), 390–399. Mosing, M. A., Magnusson, P. K. E., Pedersen, N. L., Nakamura, J., Madison, G., & Ullén, F. (2012a). Heritability of proneness for psychological flow experiences. Personality and Individual Differences, 53(5), 699–704. http://dx.doi.org/10.1016/j.paid.2012.05.035. Mosing, M. A., Pedersen, N. L., Cesarini, D., Johannesson, M., Magnusson, P. K. E., Nakamura, J., ... Ullén, F. (2012b). Genetic and environmental influences on the relationship between flow proneness, locus of control and behavioral inhibition. PLoS ONE, 7(11), e47958. http://dx.doi.org/10.1371/journal. pone.0047958. Nakamura, J., & Csikszentmihalyi, M. (2002). The concept of flow. In C. R. Snyder & S. J. Lopez (Eds.), Handbook of positive psychology (pp. 89–105). Oxford University Press. Nelson, B., & Rawlings, D. (2010). Relating schizotypy and personality to the phenomenology of creativity. Schizophrenia Bulletin, 36(2), 388–399. http://dx. doi.org/10.1093/schbul/sbn098. Nemoda, Z., Szekely, A., & Sasvari-Szekely, M. (2011). Psychopathological aspects of dopaminergic gene polymorphisms in adolescence and young adulthood. Neuroscience & Biobehavioral Reviews, 35(8), 1665–1686. http://dx.doi.org/10.1016/j.neubiorev.2011.04.002. }ségei: Egy új szituáció-specifikus Flow Kérdo }ív tesztkönyve.Budapest: Hi Press. Oláh, A. (2005). Az optimális élmény mérésének leheto Patton, J. H., Stanford, M. S., & Barratt, E. S. (1995). Factor structure of the Barratt Impulsiveness Scale. Journal of Clinical Psychology, 51(6), 768–774. Peifer, C. (2012). Psychophysiological correlates of flow-experience. In S. Engeser (Ed.), Advances in flow-research (pp. 139–164). New York: Springer. http:// dx.doi.org/10.1007/978-1-4614-2359-1_8. Reeves, S. J., Polling, C., Stokes, P. R. A., Lappin, J. M., Shotbolt, P. P., Mehta, M. A., ... Egerton, A. (2012). Limbic striatal dopamine D2/3 receptor availability is associated with non-planning impulsivity in healthy adults after exclusion of potential dissimulators. Psychiatry Research: Neuroimaging, 202(1), 60–64. http://dx.doi.org/10.1016/j.pscychresns.2011.09.011. 8 M. Gyurkovics et al. / Consciousness and Cognition 42 (2016) 1–8 Robertson, C. L., Ishibashi, K., Mandelkern, M. A., Brown, A. K., Ghahremani, D. G., Sabb, F., ... London, E. D. (2015). Striatal D1-and D2-type dopamine receptors are linked to motor response inhibition in human subjects. The Journal of Neuroscience, 35(15), 5990–5997. http://dx.doi.org/10.1523/ JNEUROSCI.4850-14.2015. Ross, S. R., & Keiser, H. N. (2014). Autotelic personality through a five-factor lens: Individual differences in flow-propensity. Personality and Individual Differences, 59, 3–8. http://dx.doi.org/10.1016/j.paid.2013.09.029. Teng, C.-I. (2011). Who are likely to experience flow? Impact of temperament and character on flow. Personality and Individual Differences, 50(6), 863–868. http://dx.doi.org/10.1016/j.paid.2011.01.012. Ullén, F., de Manzano, Ö., Almeida, R., Magnusson, P. K. E., Pedersen, N. L., Nakamura, J., ... Madison, G. (2012). Proneness for psychological flow in everyday life: Associations with personality and intelligence. Personality and Individual Differences, 52(2), 167–172. http://dx.doi.org/10.1016/j.paid.2011.10.003. Ullén, F., de Manzano, Ö., Theorell, T., & Harmat, L. (2010). The physiology of effortless attention: Correlates of state flow and flow proneness. In B. Bruya (Ed.), Effortless attention: A new perspective in the cognitive science of attention and action (pp. 205–218). Cambridge, MA, USA: The MIT Press. http://dx. doi.org/10.7551/mitpress/9780262013840.003.0011. Ulrich, M., Keller, J., Hoenig, K., Waller, C., & Grön, G. (2014). Neural correlates of experimentally induced flow experiences. Neuroimage, 86, 194–202. http:// dx.doi.org/10.1016/j.neuroimage.2013.08.019.

Consciousness and Cognition 22 (2013) 1061–1073 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Lifespan trends of autobiographical remembering: Episodicity and search for meaning Tilmann Habermas a,⇑, Verena Diel a, Harald Welzer b a b Goethe University, Frankfurt a.M., Germany Kulturwissenschaftliches Institut, Essen, Germany a r t i c l e i n f o Article history: Received 1 March 2013 Available online 13 August 2013 Keywords: Autobiographical memory Autobiographical narrating Lifespan Semantization a b s t r a c t Autobiographical memories of older adults show fewer episodic and more non-episodic elements than those of younger adults. This semantization effect is attributed to a loss of episodic memory ability. However the alternative explanation by an increasing proclivity to search for meaning has not been ruled out to date. To test whether a decrease in episodicity and an increase in meaning-making in autobiographical narratives are related across the lifespan, we used different instructions, one focussing on specific episodes, the other on embedding events in life, in two lifespan samples. A continuous decrease of episodic quality of memory (memory specificity, narrative quality) was confirmed. An increase of search for meaning (interpretation, life story integration) was confirmed only up to middle adulthood. This non-inverse development of episodicity and searching for meaning in older age speaks for an autonomous semantization effect that is not merely due to an increase in interpretative preferences. Ó 2013 Elsevier Inc. All rights reserved. 1. Introduction The ability to remember learned verbal material decreases across adulthood. A similar decrease is apparent in autobiographical event memory. This trend has been described as autobiographical memory becoming less episodic and more semantic with age. In terms of Conway’s model of autobiographical memory (Conway & Pleydell-Pearce, 2000), this would mean that basic event-specific memories become less accessible, especially in terms of temporally sequenced actions and events. This in turn would lead to an increased use of higher level knowledge of repeated or extended events and life phases. There is a reverse age difference between young and old adults in interpretative efforts to render memories more meaningful. This increase might serve as an alternative explanation for the decrease of episodic memory. We present a study that attempts to replicate these two hitherto independent findings and to test their possible inverse interrelatedness in the same data set. Furthermore we compare instructions that maximize specific episodic narrating versus narrating with a focus on search for biographical meaning. Finally we go beyond the simple comparison of young and old adults by using two lifespan samples with four and six age groups respectively to explore the relationship between age, episodicity, and search for meaning in more detail. 1.1. Semantization of autobiographical memory In laboratory studies of verbal learning, working memory and episodic memory decrease in older adulthood. In a large longitudinal study with 35–80-year-olds, Rönnlund, Nyberg, Bäckmann, and Nilsson (2005) found that short-term episodic ⇑ Corresponding author. Address: Department of Psychology, Goethe University, Grüneburgweg 1, D-60323 Frankfurt a.M., Germany. E-mail address: tilmann.habermas@psych.uni-frankfurt.de (T. Habermas). 1053-8100/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.concog.2013.07.010 1062 T. Habermas et al. / Consciousness and Cognition 22 (2013) 1061–1073 memory for verbal material and simple actions began to gradually decrease at age 60, accelerating after age 70. Cultural knowledge (semantic memory) decreased at a more moderate pace in the same age range. Learning studies interpret the concept of episodic memory as originally conceived, i.e. as memory for experimental stimuli (Tulving, 1972). Episodic memory was later reinterpreted to cover autobiographical remembering of personal experiences. This more recent concept refers to the subjective experience of reliving past events in a vivid fashion, including details of the scene (Tulving, 2002). Several studies (reported below) demonstrated that this kind of episodic, i.e. autobiographical episodic memory, also decreases in later adulthood. A further claim is that with age, episodic memories are increasingly reduced to semantic memories, i.e. that remembering how events unfold is transformed into merely knowing that an event happened. To compare remembering how to knowing that something happened, these studies interpret the amount of detail in a memory report as indicating the degree of re-experiencing, or of the episodicity of the memory. All information that does not regard the specific event is termed ‘external’ information and interpreted as resulting from semantic memory. Levine, Svoboda, Hay, Winocur, and Moscovitch (2002) asked 15 young and 15 older adults (24 and 74 years) to freely narrate an event from each of five life periods, followed by a series of probes to elicit more detail. The number of event-related details was counted for each narrative, including happenings, physical, temporal, and locational information as well as cognitive and emotional reactions. In addition, each of these aspects was rated on four-point scales, as were the overall episodic richness and the degree of integration of the episode into the wider context of life. Young adults reported more information internal to, and older adults more information external to the event. Additional probing reduced age differences, especially in the ratings. However older adults still provided more information about things other than the event itself, and still integrated the event more with other parts of life. In related studies Addis, Wong, and Schacter (2008) and Addis, Musicaro, Pan, and Schacter (2010) asked young (early 20s) and older (70s) participants to think of a total of eight recent memories in response to cues and to generate as many details as possible within three minutes. Again, young adults produced more details of the episode itself and less information about anything outside the specific episode. These findings were confirmed by Piolino, Desgranges, Benali, and Eustache (2002) in a sample of 52 40–79-year-olds. The more continuous age distribution age groups allowed a better location of age changes. Participants were asked for personal biographical knowledge and for detailed accounts of four specific episodes for each decade of their lives. Accounts were classified either as specific memories with detail, specific memories without detail, generic memories, or as vague answers. Beginning in the age range of 60–69 years, increasing age lead to a lower percentage of specific memories, both with and without detail. Piolino et al. (2006) replicated this finding with questionnaires in adults aged 25, 62, and 75 years. Again, specific memories, both with and without detail, showed a decrease between all three age groups. Corresponding age decreases in ratings of remembering versus knowing, and of visual field versus observer perspective confirmed a decline in the episodic quality of autobiographical event memories. Age differences were largest in memories from the past 5 years. 1.2. Exceptions to a decrease in episodicity These studies claim that episodic memory for events from the personal past declines between early and late adulthood. Other than two studies which did not find differences in episodic autobiographical memory between middle aged and older adults (Berna, Schönknecht, Seidl, Toro, & Schröder, 2012; Howes & Katz, 1992), there are two major concerns with the scope of the semantization effect of aging. We briefly discuss the first concern and then turn to the second concern which motivated our study. For one, some memories may be spared from the semantization effect. In some flashbulb memory studies that similarly ask for as many details as possible, older adults provided no less detail than younger adults. For example, Bluck, Levine, and Laulhere (1999) found no differences between 20- and 62-year-olds in the amount and accuracy of remembered information about the television news of the O.J. Simpson verdict. Also, older adults remembered historical events as well as younger adults in several studies of events of extraordinary significance (Kvavilashvili, Mirani, Schlagman, Erskine, & Kornbrot, 2010). Memories of highly significant historical events stabilize after about a year (Hirst et al., 2009) and retain a high level of accuracy and detail (Berntsen & Thomsen, 2005). The long-term stability and missing age differences in these memories are probably due to two factors. One factor is that cultural practices of commemorating stabilize memories (Berntsen & Thomsen, 2005; Campbell, Nadel, Duke, & Ryan, 2011; Hirst et al., 2009). The other factor is the personal significance of events in terms of self-reference (Amami, Serbun, & Gutchess, 2011; Howes & Katz, 1992), social identity (Berntsen & Thomsen, 2005), and emotionality (Kensinger, 2009; Kvavilashvili et al., 2010). Every day practices of remembering often involve the repeated sharing of memories which are both personally significant and emotional (e.g., Boden & Bielby, 1983; Kelly, Bohanek, & Fivush, 2008). 1.3. Increase of preference for search for meaning in autobiographical narrating We now turn to the other major concern with the semantization thesis. An alternative explanation for the decrease in the episodic nature of autobiographical memory reports is the increase of the preference for searching for meaning. This shows in an increase in contextualizing and relating memories to other parts of life and oneself, and in a decrease of elaborating the unravelling of the episode itself. T. Habermas et al. / Consciousness and Cognition 22 (2013) 1061–1073 1063 Whereas the semantization explanation of the decrease in episodicity refers to a decrease in mnemonic abilities, the alternative explanation refers to an increase not in ability, but in a preference for remembering in a specific way. The best way to measure mnemonic abilities is to control encoding conditions and compare them to what is recalled. This is not possible for autobiographical memories, as their selection needs to be left to the participants. Therefore studies that attempt to measure autobiographical memory competence give participants ample time to reply (e.g., Addis, Wong, & Schacter, 2008; Piolino et al., 2002), and a first reply is followed up with probes for further detail (Levine et al., 2002; Piolino et al., 2002). With more time and probing, differences should be due more to ability and less to preference, to competence and not mere performance. Consequently, if the decrease in the episodic nature of memory reports is due to a loss of mnemonic abilities, age differences should be greater with probing. Surprisingly, in one study (Levine et al., 2002) probing tended to diminish, rather than increase, age differences. Furthermore, these methods do not access all memories, as evidenced by the phenomenon of hypermnesia (Bluck et al., 1999; Campbell et al., 2011). Therefore, it is unclear whether the findings on which the semantization thesis is based reflect changes of memory abilities or rather of preferences in style of remembering. On the other hand, several studies point to a shift in preferential remembering from a more isolated and event-focussed style to valuing meaning by interpreting the significance of memories and by embedding them in life. For example, in Levine et al. (2002) study, older adults related an event to other parts of life more so than did younger adults. Also, older adults provide more repeated and extended events than specific events (Piolino et al., 2002), possibly to provide memories that are more informative about who they are. The tendency of older adults to focus more on meaning than on detail was also evidenced in the retellings of two tales by 19- and 73-year-olds (Adams, Smith, Nyquist, & Perlmutter, 1997). In another study, comparable age samples described three pictures and talked about their family, their education, and about a holiday. Older adults produced more so-called ‘off-topic speech’ in the personal accounts than young adults (see also Ruffman, Murray, Halberstedt, & Taumoepeau, 2010), but not in the descriptions of pictures. A second group of young and older adults consistently rated the older adults’ personal stories as more informative and interesting, and as better stories (James, Burke, Austin, & Hulme, 1998). This ‘off-topic talk’ may actually have served to contextualize events. James et al. (1998) suggested that older adults communicate more in order to make meaning rather than to report accurately. Trunk and Abrams (2009) could not confirm age differences in communicative intent. However older adults again provided more contextual information. The activity of embedding past events within the context of one’s life and personality development has been termed autobiographical reasoning in developmental literature (Habermas & Bluck, 2000). Autobiographical reasoning uses arguments such as learning lessons and insights (McLean & Thorne, 2003). It develops first between late childhood and early adulthood (Habermas & de Silveira, 2008). Beyond early adulthood, some studies find a further increase in the spontaneous use of autobiographical reasoning up to middle adulthood (Bluck & Glück, 2004). A more simple way of embedding events in a life is by simply referring to other parts of life. For example, self-defining memories of older adults refer more to other parts of life than those of younger adults (Singer, Rexhaj, & Baddeley, 2007; for contradicting evidence see McLean, 2008). It is not clear whether there is any increase in autobiographical reasoning beyond middle adulthood or rather a decrease, as suggested by Pasupathi and Mansour (2006) based on their findings for crisis narratives. Autobiographical reasoning focussed more on personal stability in older and more on change in younger adults (Lilgendahl & McAdams, 2011; McLean, 2008; Rice & Pasupathi, 2010). The only direct comparison between a focus on the remembered event itself and a focus on the meaning of memories found an inverse relationship in young adults (Grysman & Hudson, 2011). To summarize, both a decrease in the ability for episodic remembering as well as an increase in the preference for searching for meaning in life may contribute to decrease of the episodic nature of memory reports with age. 1.4. The study: lifespan age trends in autobiographical narrating 1.4.1. Design To better understand the relationship between episodic autobiographical memory ability and the proclivity to search for autobiographical meaning, we studied both episodicity and searching for meaning in autobiographical memory reports simultaneously. This allowed exploring whether they are in fact inversely related to each other. Secondly, we studied three adult age groups so as to study the constructs’ relationship more continuously across ages. This allowed for curvilinear relationships with age, and offered a simultaneous view on how episodicity and searching for meaning change relative to each other across adulthood. We also added three younger age groups to explore the relation between episodicity and searching for meaning in an age range when the ability for autobiographical reasoning is only emerging (cf. Bohn & Berntsen, 2008; Habermas & de Silveira, 2008). Third, to disentangle ability from preference, we contrasted a sample with instructions that favor narrating the unfolding of an event with a sample with instructions that favor searching for biographical meaning. The former condition is closer to measuring episodic mnemonic abilities and mere preference of searching for meaning, the latter to measuring the abilities to create biographical meaning and mere preference for remembering the unravelling of specific episodes. We present a secondary analysis of autobiographical narratives from two samples. Narratives of Sample A had originally been collected to study neurological correlates of autobiographical remembering (Oddo et al., 2010). Narratives of the four younger age groups of Sample B were originally collected to study the development of global coherence in life narratives (Habermas & de Silveira, 2008). We thus compare effects of age on episodicity and searching for meaning under instructions that differ in their focus on either the unfolding of an episode or searching for meaning. 1064 T. Habermas et al. / Consciousness and Cognition 22 (2013) 1061–1073 1.4.2. Dependent measures The episodicity of memory reports was measured by three indicators. If the definition of episodic remembering stresses subjective reliving (Tulving, 2002), this suggests that the memory report is an ‘‘evocative, and/or vivid description that appears to emerge from a feeling of re-experiencing’’ (Levine et al., 2002, p. 680). If the definition of episodic memory stresses the specificity of the remembered event (Barsalou, 1988; Piolino et al., 2002; Piolino et al., 2006; Williams & Broadbent, 1986), only reports of specific, datable events will count as episodic. If the focus is on the text type of the memory report, then it must contain narrative clauses, which imitate the order of events in the order of sentences, i.e. ‘‘and then . . ., and then . . .’’ (Labov & Waletzky, 1967). Narrative clauses correspond to what St-Laurent, Moscovitch, Tau, and McAndrews (2011) termed ‘temporally precise clustered details’. Any text that contains at least one pair of narrative clauses counts as narrative, as contrasted to chronicles that summarize events, descriptions of states of affairs, and arguments. All three indicators of episodicity were measured. To measure searching for meaning we used two indicators. One was the degree of interpretation involved, the other the degree to which a local event is linked to other elements of life, thereby integrating the event with the life story. While in earlier studies we were concerned with the development of global coherence in life narratives, using the relative frequency of single autobiographical arguments that embed events in a life as indicators of global coherence (e.g., Habermas, 2011), here we are not interested in global text coherence. Rather we adapt to the methodology of memory research by segmenting the text into units, each representing a memory. These segments serve as the unit of measurement. For each segment, we rated episodicity and searching for meaning. 1.4.3. Hypotheses We expected that instructions asking for specific events would produce more episodic reports with less searching for meaning than would instructions asking for autobiographical interpretation and embedding. Age patterns yielded from the specific event instructions could then be interpreted as reflecting the ability to produce episodic memories, and age patterns yielded from the life story instruction as reflecting the ability to embed events meaningfully in a life. To explore whether an increasing preference for searching for meaning may contribute to explaining the decrease of episodicity in memory reports, Hypothesis 1 expected that episodicity and searching for meaning would correlate negatively with each other. Hypothesis 2 expected a decrease of the episodicity of memories across adulthood. In the lower age range, episodic autobiographical remembering should not be a problem and therefore not differ from young adulthood. Hypothesis 3 expected a linear increase in searching for meaning across the lifespan. We expected the increase to start in the lower age range, because children have not yet acquired a concept of life (Habermas & Bluck, 2000). We expected the increase to continue throughout adulthood due to the increasing preference for searching meaning. If the two measures develop inversely with age, this adds to the probability that searching for meaning supports the trend for decreasing episodicity in memory reports in older adults. If, however, episodicity and searching for meaning show different, not strictly inverse developmental patterns, then the development of both can be studied separately, and findings of a decrease of episodicity may not simply reflect an increase in a focus on meaning, but indeed support the semantization effect. 2. Method 2.1. Participants 2.1.1. Sample A A convenience sample of 62 female participants was divided into four age groups (16, 20, 40 and 65 years), with mean ages of 16.53 (SD = .30, N = 16), 20.90 (.67, 15), 40.69 (1.94, 16), and 65.48 years (3.11, 15). Adolescents were recruited from a West German Gymnasium, i.e. a secondary school leading to an A-level exam (‘Abitur’), while young adults were students at a North German University. The two younger groups were interviewed by a female interviewer in her twenties, the two older groups by a female interviewer in her forties. 2.1.2. Sample B A total of 168 females and males from a central German city was distributed across six age groups (8, 12, 16, 20, 40, 65 years) with mean ages of 8.57 (SD = .28, 14 girls, 13 boys), 12.38 (.36, 14 and 15), 16.61 (.43, 14 and 13), 20.52 (.53, 13 and 14), 40.79 (2.87, 14 and 14), and 64.53 years (2.27, 15 and 15). The youngest age group was the higher achieving half of third graders from a grade school, the adolescent and young adult groups were present or former students of a Gymnasium. Participants for the two oldest age groups were recruited with flyers distributed widely in local shops, at sports facilities, doctors’ offices, and among continuing education University students. The younger age groups were heading towards an Abitur, whereas 24 of the middle-aged and 19 of the older adults had Abitur, and four and 9 respectively had finished school after 10 years (‘Mittlere Reife’), while two of the older adults had left school earlier. In the middle aged group 13 and in the older group 17 participants held a College degree. Thus the four younger age groups were well educated, and given the steep historical increase in higher educational, the educational level of the two oldest groups was comparably high for their cohorts. Three female interviewers in their twenties were equally distributed across the four younger age groups, one of T. Habermas et al. / Consciousness and Cognition 22 (2013) 1061–1073 1065 whom (Verena Diel) shared interviewing of the two older age groups with two different female interviewers again in their mid-twenties. 2.2. Procedure 2.2.1. Sample A – memories In a one to two hour long, semi-structured, audio-recorded interview at their homes or in the lab, participants were asked to recount personally significant positive and negative memories. These were then dated by the participants and assigned to a life phase (preschool, grade school, high school, adulthood, and last year). Then the interviewer probed for additional positive and negative memories aimed at eliciting up to nine positive and nine negative memories from each life phase. Frequently interviewers probed for the dates, valence, and other detail of events (Oddo, 2009). 2.2.2. Sample B – life narratives The three younger age groups were interviewed and audio-recorded individually in their schools, the three older age groups in the lab. Participants were asked to write their seven most important memories on cards, date them, and put them in chronological sequence in front of them. Then they were asked to narrate their life in 15 min, integrating the seven memories and recounting how they have become the person they are today (cf. Habermas & de Silveira, 2008, for verbatim instructions). Participants were not interrupted except for a reminder of the time left after 10 min, and otherwise only encouraged to continue with nonverbal cues from the interviewer. The seven memories served to ensure that specific episodes were integrated into the life story. The four younger groups re-narrated the life story again after 2 weeks. Here we analyze only the first life narrative. 2.3. Segmenting Verbatim transcripts from both studies were segmented and coded together by the same researchers. Transcripts were divided into thematic segments containing at least four clauses and extending in exceptional cases up to 3 pages (roughly 250–300 clauses; Diel, Elian, & Weber, 2007). The prototype of a segment is a narrative focussing on a specific, datable event. Ideally segments are explicitly introduced and ended (see Appendix for examples). Two research assistants independently segmented 6 entire narratives of Sample A, reaching an interrater reliability of Cohen’s Kappa = .81 (totalling 421 and 440 segments respectively), and 16 entire life narratives of Sample B, K = .82 (231/241 segments). We counted segments as agreeing when a segment border was marked in the same or immediately neighboring clause. Then each coder segmented half of the remaining narratives, mixed from both samples. In the psychological literature, usually only the initial interrater reliability is calculated and reported. To be exceptionally thorough, we added an extra check of the quality of the ensuing segmenting by calculating an additional interrater reliability (which we term ‘control reliability’) based on all segments of another randomly chosen 6 plus 16 narratives which were not known by the respective main coder, yielding K = .90 and K = .92 for 296/309 and 349/365 segments, respectively. 2.4. Coding Segments served as basic units for coding, each segment receiving a code or rating (see Appendix A for examples). Ratings were made on 4-point-scales (0–3). For each participant, across all segments, mean ratings and relative frequencies of codes were used. First drafts of the coding manuals were written by the first and third authors with Johannes Schröder, while the second author wrote the final manual together with the respective coders (Diel et al., 2009). Single intraclass correlations were used for ratings, Cohen’s Kappa for nominal codes. 2.4.1. Vividness A value of 3 was given if the overall impression was very vivid and the segment included a dramatized narration of a specific event, a 2 if the segment was fairly vivid and contained a narrative of a specific and concrete event, a 1 if the segment was not vivid and resembled more a chronicle than an actual narrative, and a 0 if the segment appeared dry and monotonous. Interrater-reliabilities were ric = .78 (.80) in Sample A and ric = .82 (.81) in Sample B. Reliability calculations were based on 452 (209) and 260 (140) segments respectively. 2.4.2. Episode quality We coded the episodicity of segments in two ways, as memory specificity and as narrative text type. Although the former aims at the content, the latter at the form of text, they should be highly related, because if specific events are talked about at length, they tend to be narrated. Following Barsalou (1988), we coded segments either as containing a specific memory of an event lasting up to a day, as a generalized event, i.e. repeated events or events extending between a day and a year, as a very extended event lasting longer than a year, or as a segment that does not refer to a temporal unit, such as a mere description or comment. Interrater reliabilities were K = .72 (.70) for Sample A and .81 (.71) for Sample B, based on 452 (209) and 261 (140) segments respectively. 1066 T. Habermas et al. / Consciousness and Cognition 22 (2013) 1061–1073 Additionally, segments were coded for text type. A segment was coded as a narrative if it contained at least two consecutive narrative clauses referring to consecutive events (Labov & Waletzky, 1967), as a chronicle if the main bulk of the segment summarized or described events without narrating them (Linde, 1993), or as an argument if it was not coded as a narrative and the main part of the segment contained arguments, comments, evaluations, or interpretations (Rosenthal, 1995). Interrater reliabilities were K = .80 (.68) for Sample A and .83 (.79) for Sample B, based on 435 (209) and 250 (140) segments respectively. For the testing of hypotheses, we only used the percentages of specific memories and narratives among all segments from each participant. 2.4.3. Search for meaning We rated the degree to which each segment was related to and integrated with other parts of the life story, which is a much more basic measure of life story integration than specific autobiographical arguments identified in earlier studies (Habermas & de Silveira, 2008). Life story integration was rated 3 if the segment contained at least three references to other times or to other topics in life, 2 for two, 1 for one, and 0 for no such reference. The degree to which each event was interpreted was also rated. We assigned a value of 3 if the significance of the segment for one’s life, a change in personality, or a profound change of attitude was described, 2 if the narrator described her or his own personality or justified significant emotional reactions to the event, 1 if an event was evaluated or emotions were mentioned, and 0 if events were only factually narrated, but not evaluated. Life story integration achieved an interrater reliability of ric = .77 (.40) in Sample A and ric = .81 (.31) in Sample B, interpretation of ric = .75 (.55) and ric = .82 (.60) respectively, based on 402 (209) and 260 (140) segments respectively. 3. Results 3.1. Data analysis strategy As a manipulation check, effects of the two instructions will initially be compared in the four older age groups for women only. Hypotheses will be tested separately for the two samples, because they differ both in age and gender composition. We explore correlations between indicators of the same construct and test hypothesis 1 that episodicity and searching for meaning are negatively correlated. Hypotheses 2 and 3 will be tested with multivariate analyses of variance (MANOVAs) first with Sample A, then Sample B. Planned contrasts are calculated to test differences between neighboring age groups. Interpretation of results is based both on means and confidence intervals (see figures) as well as multivariate and univariate tests with 5% level of significance and effect sizes. Means and confidence intervals (95%) are provided in graphs, integrating findings from both samples to facilitate visual comparison. Thus graphs show means for Sample A and separate means for females and males of Sample B. Outliers in dependent variables were corrected to the whiskers of respective boxplots for each sample and age group. 3.2. Comparison between samples/instruction Narratives in Sample A were longer than in Sample B (Fig. 1), reflecting the different instructions. Not surprisingly, length also varied more in Sample A, because the duration of the interview was not as standardized as in Sample B. Length did not consistently differ by gender. In Sample A, participants were asked to provide single memories and probed for evaluations and details of time and location. In Sample B the life narratives were instructed to inform the interviewer about how a person had become who she or he is at present, thus encouraging participants to connect the personal past with later life developments and the present. Thus Sample A was expected to provide more concrete and less abstract elements, i.e. more detail, less interpretation, and more specific memories and narrative memory reports. Testing this expectation serves as a manipulation check. A multivariate test with only the female participants in the four older age groups confirmed the overall expectation. All univariate effects were also significant (Table 1; Figs. 2–4). 3.3. Correlations between dependent measures 3.3.1. Zero-order and partial correlations To explore whether the indicators of episodicity and of searching for meaning correlated positively with each other, and to test the first hypothesis that episodicity and searching for meaning correlate negatively, we correlated dependent measures separately for each sample (Table 2). Since age and number of segments correlated highly with dependent measures, and because these correlations differed between samples due to the differing age composition and much larger variation in length in Sample A, we partialled out age and number of segments, and used these partial correlations for data exploration and testing. Otherwise correlations were relatively similar in both samples. T. Habermas et al. / Consciousness and Cognition 22 (2013) 1061–1073 1067 Fig. 1. Length: Mean number of segments and error bars (95%) of memories (Sample A) and life narratives (Sample B) by gender and age group. Table 1 Multivariate and Univariate ANOVAS of mean differences between the two samples (memories versus life narratives) for females in four older age groups (Ages 16, 20, 40, 65). df Multivariate Univariate Vividness Memory specificity Narrativity LS Integration Interpretation (5, 112) (1, 116) (1, 116) (1, 116) (1, 116) (1, 116) g2 F 24.72  .53  .18 .35 .43 .24 .16 25.41 62.45 86.50 36.15 22.02  p < .05. p < .01  p < .001.  3.3.2. Correlations between related dependent measures Memory specificity and narrativity correlated highly and positively in both samples. However, vividness did not correlate with specificity and only weakly with narrativity, thus describing a somewhat different aspect of memory reports. Interpretation and references to other parts of life correlated moderately and positively with each other as expected. 3.3.3. Correlation between episodicity and searching for meaning Of the eight partial correlations between the two indicators of episodicity and the two indicators of searching for meaning, all but one are negative, but only four out of eight are significant, and their size is only moderate (see Table 2). We averaged the two variables each for episodicity and search for meaning, resulting in partial correlations of rp = .14 (ns) in the memories condition and rp = .34 (p < .000) in the life story condition. Thus when instructions demand narrating specific memories, episodicity tends not to correlate negatively with searching for meaning. However when instructions demand making meaning, these two aspects tend to correlate moderately negatively. Thus when testing for memory ability, search for meaning does not substantially reduce episodicity of memory reports, speaking against the alternative explanation of the decrease in episodicity. Only when there is a demand for meaning making, does this reduce to some degree the episodicity of 1068 T. Habermas et al. / Consciousness and Cognition 22 (2013) 1061–1073 2,0 Mean Vividness (0-3) 1,5 1,0 ,5 ,0 8 12 16 20 40 65 Age Fig. 2. Mean differences and error bars (95%) for vividness in memories (Sample A) and life narratives (Sample B) by gender and age group. Fig. 3. Episodicity: Mean percentage of segments and error bars (95%) for specific memories (memory specificity) and narrative segments (text type) in memories (Sample A) and life narratives (Sample B) by gender and age group. memory reports. Vividness showed unexpected values, in this case moderate to large positive partial correlations with indicators of search for meaning. 3.4. Age effects 3.4.1. Vividness In Sample A (memories), age had a significant effect of moderate size (Table 3; Fig. 2). In planned contrasts, there was only a significant decrease between ages 40 and 65 (p = .000). 1069 T. Habermas et al. / Consciousness and Cognition 22 (2013) 1061–1073 Fig. 4. Searching for meaning: Mean ratings and error bars (95%) for life story integration and interpretation in memories (Sample A) and life narratives (Sample B) by gender and age group. Table 2 Zero-order correlations (1st Row) and partial correlations (age, number of segments, 2 row in italics) between dependent measures of episodicity, search for meaning, and number of segments in Sample A (lower left triangle) and Sample B (upper right triangle). 1 2 1. Vividness 2. Specificity 3. Narrativity 4. Lifestory Integration 5. Interpretation 6. N segments 7. Age    3 .09 .05 .22 .01 .38** .21 .03 .42*** .55*** .53*** .38** .35** .80*** .66*** .66*** .27* .16 .22 .21 .73*** 4 .23 .16* .54*** .46*** .42*** .04 .02 .09 .34** .63*** 5 .13 .21** .32*** .24** .35*** .26*** .37*** .45*** .17* .10 .38*** .32*** .34*** .29 .34 .27 .72*** .38** .03 6 7 .12 – .28*** – .16* – .12 – .06 – .21** – .38 – .40*** – .30*** – .23** – .37**** .45 p < .05. p < .01. p < .001. In Sample B (life narratives), age again had a significant moderate effect. Also, females consistently narrated more vividly across age groups. Vividness increased between ages 8 and 12 (planned contrast, p = .01) and, in concordance with findings in Sample A, decreased between ages 40 and 65 (p = .000). Instead of the expected linear decrease across adulthood there was only a decrease in older adulthood. 3.4.2. Episodic quality In Sample A (memories) the episodic quality of memories varied significantly by age, both for the relative frequency of specific memories and of narrative segments. Univariate age effects were even larger (see Table 3). The expected linear decline was monotonous (Fig. 3). Planned contrasts between neighboring age groups were all significant for specificity (p = .016, .001, .049 respectively), but for narrativity only between the two youngest groups (p = .018). Sample B (life narratives) showed a highly similar picture, with age again significantly affecting episodicity. A barely significant interaction with gender resulted from the 16-, 20-, and 40-year-old males recounting more episodic segments than their female counterparts. The expected linear decline was monotonous, but limited to the three younger age groups, while the adult age groups did not differ from each other. Contrasts between neighboring age groups were significant for ages 8 and 12 (p = .000 and p < .05 for specificity and narrativity) and 12 and 16 (p < .05, p = .000). Thus the effect of age on the episodicity of memories varied between instructions. There was a delay in the decrease of episodicity between instructions, which was evident in adolescence in response to the life story instruction, but only in adulthood in response to the memories instruction. 1070 T. Habermas et al. / Consciousness and Cognition 22 (2013) 1061–1073 Table 3 Multivariate and univariate ANOVAs with factor age for vividness, episode quality, and search for meaning in memorie (Sample A, ages 16, 20, 40, 65) and life narratives (Sample B, ages 8, 12, 16, 20, 40, 65). Multivariate univariate Memories (Sample A) Life narratives (Sample B) Age Age df    g F *** 2 df Gender g F *** 2 df AG g F *** 2 df F g2 Vividness (3, 58) 1.77 .28 (5, 156) 2.59 .14 (1, 156) 2.59 .09 (5, 156) .0 .04 Episode quality Specific memory Narrative (6, 116) (3, 58) (3, 58) 8.17*** 25.57*** 14.34*** .30 .57 .43 (10, 312) (5, 156) (5, 156) 10.43*** 18. 66*** 17.55*** .25 .31 .36 (2, 155) (1, 156) (5, 156) .12 . 03 .14 .00 .00 .00 (10, 312) (5, 156) (5, 156) .16* 2.03 2.04 .07 .06 .06 Search for meaning Lifest. Integr. Interpretation (6, 116) (3, 58) (3, 58) 12.03*** 29.89*** 2.44 .38 .61 .11 (10, 312) (5, 156) (5, 156) 8.89*** 11.92*** 12.17*** .22 .28 .28 (2, 155) (5, 156) (5, 156) .19 .31 .03 .00 .00 .00 (10, 312) (5, 156) (5, 156) .67 .89 .36 .02 .02 .01 p < .05. p < .01. p < .001. 3.4.3. Searching for meaning Life story integration and interpretation were used to indicate searching for meaning. In Sample A (memories), age had a significant multivariate effect on searching for meaning, based on a univariate effect for life story integration (Table 2). Contrary to predictions, interpretation did not increase across adulthood, whereas life story integration increased between adolescence and middle adulthood (Fig. 4). Also in Sample B (life narratives) age had a significant effect on searching for meaning, based on univariate effects for both life story integration and interpretation. Life story integration increased cross-sectionally between ages 8 and 12 (p < .05) and 12 and 16 (p = .000) to remain stable across adulthood. Thus for life story integration there was again a delay between instructions, but a reverse one, with an increase across adolescence under life story instructions, and an increase only between young and middle adulthood under the memories instruction. This suggests that the ability to relate specific memories to other parts of life develops during adolescence (Sample B), but that it is spontaneously used increasingly only from mid-adulthood onward (Sample A). Interpretation also increased in Sample B across the entire age range (significant contrast only between ages 12 and 16, p = .001), with the exception of a drop between the two oldest age groups (p = .001). There was no developmental delay between instructions. Thus, overall there were clear developmental increases in searching for meaning, more age-limited in life story integration and more extensive in interpretation, confirming hypothesis 2 for ages up to middle adulthood only. 4. Discussion The present study has the advantages of using multiple age groups between late childhood and older adulthood, multiple instructions, and multiple measures. It is the first study of lifespan trends in autobiographical remembering that includes both episodicity and searching for meaning. 4.1. Searching for meaning cannot explain a decrease of episodicity of remembering The expected adult decrease of episodic memories was confirmed for specificity and narrativity of memory reports in the memories condition, whereas vividness showed a decrease only between middle and older adulthood. Thus the decrease of the episodic nature of autobiographical remembering with age was confirmed when the instructions asked for specific memories, testing for mnemonic ability. This finding confirms the semantization thesis. The expected increase in the ability to search for meaning was confirmed for both indicators in the life story condition for adolescence. Thus the ability to embed events in life develops across adolescence. Results also provide an important qualification of the apparent linear increase in meaning making in studies that only compared young with older adults (Adams et al., 1997; James et al., 1998). In the memories condition, preference for searching for meaning increased only between adolescence and middle adulthood, but not beyond. Thus the potential alternative explanation of the decreasing episodicity of event memories between middle and older adulthood, with an increasing preference for searching for meaning, can be refuted. This conclusion from mean age differences was confirmed by the absence of significant negative correlations between episodicity and searching for meaning in the memories condition. The corresponding negative correlations in the life story condition do not contradict this conclusion, since the semantization thesis concerns not the preference for, but the ability for episodic autobiographical recall as approximated by the memories condition. T. Habermas et al. / Consciousness and Cognition 22 (2013) 1061–1073 1071 4.2. Lifespan abilities and preferences The two different instructions led to an inverse developmental delay in narrativity and specificity on the one hand and life story integration on the other. The adolescent increase in life story integration and the simultaneous decrease of specific memories and narratives in the life story condition are most probably due to the gradual acquisition of autobiographical reasoning during adolescence (Bohn & Berntsen, 2008; Friedman, Reese, & Dai, 2011; Habermas & de Silveira, 2008). The decrease in episodicity in the memories condition between middle and late adulthood can be understood to result from a decrease of episodic autobiographical mnemonic ability, because it is not parallel to an increase in search for meaning. The inclusion of child and adolescent age groups produced a more complete picture of the developmental path of the ability and preference for ways of autobiographical remembering. Older children tell many specific episodes and narratives relative to adults. Finally the cross-sectional increase in vividness between ages 8 and 16 suggest that although older children do possess the ability to narrate events (e.g., Peterson & McCabe, 1983); they continue to refine their narrative abilities up to mid-adolescence. 4.3. Limitations This study is limited to culturally and educationally homogenous samples. Gender influenced only vividness. Participants were relatively free to select memories. Thus the findings do not reflect, like other studies of autobiographical memories, situations of remembering in which the aim is to remember a specified situation, such as when trying to remember where one parked the car the night before. The freedom in the choice of memories tends to result in findings that also reflect preferences and not only abilities. Despite the generally high methodological standards, the reliability of the coding of search for meaning was not optimal. Thus these findings have to be interpreted with caution. On the other hand, reliability scores were calculated on the basis of segments, the ratings of which were averaged for each individual, thereby boosting reliability. The age range was limited to age 65, which may have been too low to catch a similar age decline as in studies of verbal learning. Finally, these cross-sectional data on age trends need to be complemented by a longitudinal follow-up to ensure that group differences are actually due to age and not to other differences between groups. A longitudinal design would also permit studying the actual change or stability of specific memories over time. Thus the criticism of the semantization thesis, that some highly significant memories do not fade or become semantic (I.2), should in future be tested also for non-public, personal autobiographical memories selected by participants themselves. 4.4. Implications Earlier findings of a decrease of episodic autobiographical remembering cannot be explained entirely by an increase in the preference for interpreting and biographically embedding memories. The use of a middle-aged adult group shows that the decrease of episodicity of memory reports does not only begin at age 60, as studies of short-term verbal learning would suggest, but is a continuous process throughout adulthood up to at least age 65. Without additional evidence it is difficult to tell whether this is due to a steady decrease of long-term autobiographical memory, as part of what Baltes, Staudinger, and Lindenberger (1999) termed the mechanics of cognition such as fluid intelligence, or whether it is a combined effect of an increasing preference for searching for meaning in early and middle adulthood and a later decrease of mnemonic abilities. Future studies should simultaneously study verbal learning ability, episodic autobiographical memory ability, and preference for searching for meaning across the adult lifespan to determine both possible differences between verbal learning and autobiographical memory (cf. Berna et al., 2012), and mnemonic ability versus interpretative preferences. The low correlation of vividness of memory reports with their specificity and narrativity shows that also the phenomenological aspects of episodic autobiographical memories merit further study. We rated vividness as perceived by the listener/reader, which should be compared to self-reports of the rememberer’s subjective experience. It could well be that reader/listener-vividness is determined not only by the specificity of the memory, and not necessarily by the amount of situational detail as is sometimes suggested, but also by the form of verbal representation, such as the narrative and dramatic quality of the memory text. This is suggested by the higher correlation of readers’ vividness with narrativity than with memory specificity. This study points to the necessity of studying the relations between different aspects of autobiographical remembering: abilities and preferences, memorizing and recollecting, accuracy and coherence. Also we need studies that include the oldest age without neglecting adolescence, middle and early older adulthood. Furthermore studies need to consider not just single measures of the episodicity of memory narratives, but should encompass multiple measures that cover both traditional memory measures of phenomenal qualities and the more narrative aspects of remembering such as text type and the argumentative aspects of remembering like autobiographical reasoning. Finally, theories of the development of narrative abilities and narrative preferences are rudimentary regarding ages beyond the first decade of life. The present study needs to be complemented by non-autobiographical narratives in varying contexts to arrive at a more general picture of how narrating develops in the adult years. 1072 T. Habermas et al. / Consciousness and Cognition 22 (2013) 1061–1073 Acknowledgments This paper reports a new coding of verbal data from two studies. Data for Study A were collected by Silvia Oddo and Anna Schwab, supported by a grant of the Volkswagenstiftung to H.J. Markowitsch and the third author. Data for the four younger groups of Study B were collected by Cybèle de Silveira, Verena Diel, and Martha N. Havenith, supported by a grant of the German Research Foundation (DFG) to the first author, and data of the two oldest groups of Sample B were collected by the second author with Carolin Elian and Anna Weber, who also segmented all transcripts. The second author coded all transcripts for the purposes of this article together with Julianna Heberer, Anna Kraaz, Anna Monem, and Laura Streck. The secondary data coding was supported by a grant of the Lotte Köhler Stiftung to the first and third authors and Johannes Schröder. We thank Anna Kenney for editorial help. Appendix A. Three examples of typical segments with ratings A.1. Participant from Sample A, 16 years, female. 64 segments Segment 11 ‘‘Class excursion to Rome’’: [Interviewer: Did anything happen at school during the last year? Class excursions, concerts?’’]. Well, we went to Rome. [Interviewer: I thought that something would come to your mind.] Right. Two weeks ago – thank God our class teacher got sick. I think she was looking forward to the trip. But then she could not go. But no one in our class really gets along with her. Then our substitute class teacher, the Math teacher, came along. Well, and with him we had a lot of fun. Went to see the Coliseum. We mostly did things together with the whole class. And the afternoons were mostly free, until the evenings. Well – uhm. Ratings: Vividness 2, extended memory, chronicle, life story integration 0, interpretation 2. Segment 12 ‘‘Wrong class’’: Well, that was kind of embarrassing, on the first or second day. Definitely during the first days. Wanted to meet at nine in the morning, And then us two went downstairs at nine sharp, we were not really late, just very punctual, And then by mistake we lined up with the wrong class. Because we were there with three classes, parallel classes. And we just saw ‘‘Oh, there are some of our people’’, we just stood with them. The others must have been calling us from the other side, but we didn’t hear them. Well, that was a little – unfortunate. [Interviewer: Did your class leave without you?’’] Well, no, they came back to get us. They had already walked like 20 m, well, that leaves a bad impression. Ratings: Vividness 2, specific memory, narrative, life story integration 0, interpretation 1. A.2. Participant from Sample B, 60 years, female, 37 segments Segment 24 ‘‘Father remarries’’: That year my father married for a second time. And that was quite incisive because he totally focussed on his new wife and family, such that I – had done my part, and we have no contact till today. I can’t accept letting myself be hurt anymore. It was like, he was like totally changed, as if what I did, didn’t count for anything anymore, I had passed my youth, my early adulthood, when other women start a family of their own, I spent that time with my father. Therefore for me it was like he gave me a kick. We had been very close, almost like a couple without sex. We had even been addressed as husband and wife. And then he just remarried, very fast after my mother had died, which I could not understand. But well, that’s the way it went. Ratings: Vividness 2, specific memory, argument, life story integration 2, interpretation 2. References Adams, C., Smith, M. C., Nyquist, N., & Perlmutter, M. (1997). Adult age-group differences in recall for the literal and interpretive meanings of narrative text. Journals of Gerontology, Series B—Psychological Sciences and Social Sciences, 52B, 187–195. doi: 10.1093%2Fgeronb%2F52B.4.P187. Addis, D. R., Musicaro, R., Pan, L., & Schacter, D. L. (2010). Episodic simulation of past and future events in older adults: Evidence from an experimental recombination task. Psychology and Aging, 25, 369–376. http://dx.doi.org/10.1037/a0017280. Addis, D. R., Wong, A. T., & Schacter, D. L. (2008). Age-related changes in the episodic stimulation of future events. Psychological Science, 19, 33–41. doi: 10.1111%2Fj.1467-9280.2008.02043.x. Amami, A., Serbun, S. J., & Gutchess, A. H. (2011). Self-referencing enhances memory specificity with age. Psychology and Aging, 26, 636–646. doi: 10.1080%2F09658211.2011.626429. Baltes, P. B., Staudinger, U. M., & Lindenberger, U. (1999). Lifespan psychology: Theory and application to intellectual functioning. Annual Review of Psychology, 50, 471–507. http://dx.doi.org/10.1146/annurev.psych.50.1.471. Barsalou, L. W. (1988). The content and organization of autobiographical memories. In U. Neisser & E. Winograd (Eds.), Remembering reconsidered: Ecological and traditional approaches to the study of memory (pp. 193–243). New York: Cambridge University Press. Berna, F., Schönknecht, P., Seidl, U., Toro, P., & Schröder, J. (2012). Episodic autobiographical memory in normal aging and mild cognitive impairment: A population-based study. Psychiatry Research, 200, 807–812. http://dx.doi.org/10.1016/j.psychres.2012.03.022. Berntsen, D., & Thomsen, D. (2005). Personal memories for remote historical events: Accuracy and clarity of flashbulb memories related to WW II. Journal of Experimental Psychology: General, 134, 242–257. doi: 10.1037%2F0096-3445.134.2.242. Bluck, S., & Glück, J. (2004). Making things better and learning a lesson: Experiencing wisdom across the lifespan. Journal of Personality, 72, 543–572. doi: 10.1111%2Fj.0022-3506.2004.00272.x. T. Habermas et al. / Consciousness and Cognition 22 (2013) 1061–1073 1073 Bluck, S., Levine, L. J., & Laulhere, T. M. (1999). Autobiographical remembering and hyprmnesia: A comparison of older and younger adults. Psychology and Aging, 14, 671–682. doi: 10.1037%2F%2F0882-7974.14.4.671. Boden, D., & Bielby, D. (1983). The past as resource: A conversational analysis of elderly talk. Human Development, 26, 308–319. doi: 10.1159%2F000272892. Bohn, A., & Berntsen, D. (2008). Life story development in childhood: The development of life story abilities and the acquisition of cultural life scripts from late middle childhood to adolescence. Developmental Psychology, 44, 1135–1147. doi: 10.1037/0012-1649.44.4.1135. Campbell, J., Nadel, L., Duke, D., & Ryan, L. (2011). Remembering all that and then some: Recollection of autobiographical memories after a 1-year delay. Memory, 19, 406–415. doi: 10.1080/09658211.2011.578073. Conway, M. A., & Pleydell-Pearce, C. W. (2000). The construction of autobiographical memories in the self-memory system. Psychological Review, 107, 261–288. doi: 10.1037/0033-295X.107.2.261. Diel, V., Elian, C., & Weber, A. (2007). Manual zur thematischen Segmentierung autobiographischer Erzählungen [Manual for segmenting autobiographical narratives]. Unpublished manuscript, Goethe University Frankfurt a.M., Germany. Diel, V., Heberer, J., Kraaz, A., Mahmoudi, A., Streck, S., Habermas, T., Schröder, J., & Welzer, H. (2009). Manual zum Kodieren von Detailliertheit, Deutung, Valenz, Emotionalität, Spezifität und Textsorte in Segmenten autobiographischer Erzählungen [Manual for coding detail, search for meaning, valence, emotionality, memory specificity, and text category in segmented autobiographical narratives]. Unpublished manuscript, Goethe University Frankfurt a.M., Germany. Friedman, W. J., Reese, E., & Dai, X. (2011). Children’s memory for the times of events from the past. Applied Cognitive Psychology, 25, 156–165. doi: 10.1002%2Facp.1656. Grysman, A., & Hudson, J. (2011). The self in autobiographical memory: Effects of self-salience on narrative content and structure. Memory, 19, 501–513. doi: 10.1080/09658211.2011.590502. Habermas, T. (2011). Autobiographical reasoning: Mechanisms and functions. In T. Habermas (Ed.), The development of autobiographical reasoning in adolescence and beyond. New Directions in Child and Adolescent Development, 131, 1–17. San Francisco: Jossey-Bass. doi: 10.1002%2Fcd.285. Habermas, T., & Bluck, S. (2000). Getting a life: The development of the life story in adolescence. Psychological Bulletin, 126, 748–769. doi: 10.1037%2F%2F0033-2909.126.5.748. Habermas, T., & de Silveira, C. (2008). The development of global coherence in life narratives across adolescence: Temporal, causal, and thematic aspects. Developmental Psychology, 44, 707–721. doi: 10.1037%2F0012-1649.44.3.707. Hirst, W., Phelps, E., Buckner, R. L., Budson, A. E., Cuc, A., Gabrieli, J. D. E., et al (2009). Long-term memory for the terrorist attack of September 11: Flashbulb memories, event memories, and the factors that influence their retention. Journal of Experimental Psychology: General, 138, 161–176. doi: 10.1037%2Fa0015527. Howes, J. L., & Katz, A. N. (1992). Remote memory: Recalling autobiographical and public events from across the lifespan. Canadian Journal of Psychology, 46, 92–116. James, L. E., Burke, D. M., Austin, A., & Hulme, E. (1998). Production and perception of ‘‘verbosity’’ in younger and older adults. Psychology and Aging, 13, 355–367. doi: 10.1037%2F%2F0882-7974.13.3.355. Kelly, M., Bohanek, J. G., & Fivush, R. (2008). Positive effects of talking about the negative: Family narratives of negative experiences and preadolescents’ perceived competence. Journal of Research on Adolescence, 18, 573–593. doi: 10.1177%2F0272431607308673. Kensinger, E. A. (2009). How emotion affects older adults’ memories for event details. Memory, 17, 208–219. doi: 10.1080%2F09658210802221425. Kvavilashvili, L., Mirani, J., Schlagman, S., Erskine, J. A. K., & Kornbrot, D. E. (2010). Effects of age on phenomenology and consistency of flashbulb memories of September 11 and a staged control event. Psychology and Aging, 25, 391–404. doi: 10.1037%2Fa0017532. Labov, W., & Waletzky, J. (1967). Narrative analysis: Oral versions of personal experience. In I. Helm (Ed.), Essays on the verbal and visual arts. Proceedings of the 1966 Annual Spring Meeting of the American Ethnological Society (pp. 12–44). Seattle, WA: University of Washington Press. Levine, B., Svoboda, E., Hay, J. F., Winocur, W., & Moscovitch, M. (2002). Aging and autobiographical memory: Dissociating episodic from semantic memory. Psychology and Aging, 17, 677–689. doi: 10.1037%2F%2F0882-7974.17.4.677. Linde, C. (1993). Life stories. Oxford, UK: Oxford University Press Lilgendahl, J. P., & McAdams, D. P. (2011). Constructing stories of self-growth: How individual differences in patters of autobiographical reasoning relate to well-being. Journal of Personality, 779, 391–428. doi: 10.1111%2Fj.1467-6494.2010.00688.x. McLean, K. C. (2008). Stories of the young and the old: Personal continuity and narrative identity. Developmental Psychology, 44, 254–264. doi: 10.1037%2F0012-1649.44.1.254. McLean, K., & Thorne, A. (2003). Late adolescents’ self-defining memories about relationships. Developmental Psychology, 39, 635–645. doi: 10.1037/00121649.39.4.635. Oddo, S. (2009). Zeitabhängige Modulation neuronaler Korrelate beim Abruf episodisch-autobiographischer Erinnerungen [Time-dependent modulation of neural correlates during retrieval of autobiographical memories]. Doctoral Dissertation: University of Bielefeld. Oddo, S., Lux, S., Weiss, P. H., Schwab, A., Welzer, H., Markowitsch, H. J., et al (2010). Specific role of medial prefrontal cortex in retrieving recent autobiographical memories; An fMRI study of young female subjects. Cortex, 46, 29–39. doi: 10.1016%2Fj.cortex.2008.07.003. Pasupathi, M., & Mansour, E. (2006). Adult age differences in autobiographical reasoning in narratives. Developmental Psychology, 42, 798–808. doi: 10.1037%2F0012-1649.42.5.798. Peterson, C., & McCabe, A. (1983). Developmental psycholinguistics: Three ways of looking at a child’s narrative. New York: Plenum. Piolino, P., Desgranges, B., Benali, K., & Eustache, F. (2002). Episodic and semantic remote autobiographical memory in aging. Memory, 10, 239–257. doi: 10.1080%2F09658210143000353. Piolino, P., Desgranges, B., Clarys, D., Guillery-Girard, B., Taconnat, L., Isingrini, M., et al (2006). Autobiographical memory, autonoetic consciousness, and self-perspective in aging. Psychology and Aging, 21, 510–525. doi: 10.1037%2F0882-7974.21.3.510. Rice, C., & Pasupathi, M. (2010). Reflecting on self-relevant experiences: Adult age differences. Developmental Psychology, 46, 479–490. doi: 10.1037%2Fa0018098. Rosenthal, G. (1995). Erzählte und erlebte Lebensgeschichte [The life story as experienced and told]. Frankfurt a,M., GE: Campus. Rönnlund, M., Nyberg, L., Bäckmann, L., & Nilsson, L.-G. (2005). Stability, growth, and decline in adult life span development of declarative memory: Crosssectional and longitudinal data from a population-based study. Psychology and Aging, 20, 3–18. doi: 10.1037%2F0882-7974.20.1.3. Ruffman, T., Murray, J., Halberstedt, J., & Taumoepeau, M. (2010). Verbosity and emotion recognition in older adults. Psychology and Aging, 25, 492–497. doi: 10.1037%2Fa0018247. Singer, J., Rexhaj, B., & Baddeley, J. (2007). Older, wiser, and happier? Comparing older adults’ and College students’ self-defining memories. Memory, 15, 886–898. doi: 10.1080%2F09658210701754351. St-Laurent, M., Moscovitch, M., Tau, M., & McAndrews, M. P. (2011). The temporal unravelling of autobiographical memory narratives in patients with temporal lobe epilepsy or excisions. Hippocampus, 21, 409–421. http://dx.doi.org/10.1002/hipo.20757. Trunk, D. L., & Abrams, L. (2009). Do younger and older adults’ communicative goals influence off-topic speech in autobiographical narratives? Psychology and Aging, 24, 324–337. doi: 10.1080%2F09658210701754351. Tulving, E. (2002). Episodic memory. Annual Review of Psychology, 53, 1–25. doi: 10.1146/annurev.psych.53.100901.135114. Tulving, E. (1972). Episodic and semantic memory. In E. Tulving & W. Donaldson (Eds.), Organization of memory (pp. 381–402). New York: Academic Press. Williams, J. M., & Broadbent, K. (1986). Autobiographical memory in suicide attempters. Journal of Abnormal Psychology, 95, 144–149. doi: 10.1037/0021843X.95.2.144.

Consciousness and Cognition Consciousness and Cognition 14 (2005) 495–519 www.elsevier.com/locate/concog Are strategy shifts caused by data-driven processes or by voluntary processes? Hilde Haidera,*, Peter A. Frenschb, Daniel Joramb a Institut für Psychologie, Universität Köln, Gronewaldstrasse 2, D-50931 Köln, Germany b Humboldt-Universität, Berlin, Germany Received 24 September 2004 Available online 4 February 2005 Abstract The present research investigates the role of voluntary, conscious processing in strategy change. In 2 experiments, we address whether the switch to a new strategy is the result of data-driven, automatic processes or of voluntary processes. Experiment 1 demonstrates that participants performing an alphabet verification task are able to (a) transfer a newly adopted strategy to dissimilar information never encountered before, (b) verbally describe the task regularity that allows for the generation and application of the new strategy immediately after the strategy was adopted. Using the same experimental task, Experiment 2 shows that participants, and (c) decide against adopting a new strategy when the available evidence suggests that the new strategy cannot be used for the entire range of problems encountered. Overall, the obtained results support the view that strategy change is mediated by voluntary controlled processing. They do not support the view that strategy change is an inevitable, automatic consequence of task practice. The present research thus highlights a potential function of conscious human processing.
2004 Elsevier Inc. All rights reserved. Keywords: Skill acquisition; Data-driven learning processes; Voluntary processes * Corresponding author. Fax: +49 221 470 4515. E-mail address: hilde.haider@uni-koeln.de (H. Haider). 1053-8100/$ - see front matter
2004 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2004.12.002 496 H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 1. Are strategy shifts caused by data-driven processes or by voluntary processes? Many reasons can be advanced to explain cognitive scienceÕs recent renewed interest in consciousness. One reason appears to be an attempt to support the treasured belief that our actions are the product of our conscious thoughts and intentions, a belief that stands in stark contrast to the current standard view in cognitive science holding that ‘‘most of what we do goes on unconsciously. It is the exception, not the rule, when thinking is conscious’’ (Lachman, Lachman, & Butterfield, 1979, p. 207). From the latter view it is but a short step to the invention of the zombie within us. This is to assume that all relevant cognitive processes can occur outside of consciousness, and the conscious self is assigned the part of a powerless spectator (Chalmers, 1996). Thus, a central challenge for investigators of consciousness has become to determine the precise functional role of conscious states in the architecture of the mind. The main issue of the research described in this article is to investigate the role of voluntary, conscious processing in strategy change. More specifically, we investigate whether the switch from an old to a new strategy is the result of data-driven, automatic processes or of voluntary processes. For the purpose of this article, strategy shifts are assumed to occur automatically when the change from old to new strategy is an inevitable consequence of task experience; by comparison, strategy shifts are assumed to be voluntary controlled when the change from old to new strategy is based on a deliberate decision to adopt or not to adopt the new strategy. Theoretical accounts and empirical work of how and when strategy shifts occur have tended to focus, almost exclusively, on the change from an algorithm-based task strategy to a memory-retrieval strategy. For example, in LoganÕs (1988, 1992; Logan & Etherton, 1994) Instance Model, a strategy shift occurs when the control of performance is passed from algorithm-based processing to memory-based processing. In the Instance Model, it is assumed that task processing is initially controlled by an algorithm, but that each task processing is stored separately as an instance in memory. Every encounter with a task activates the algorithm that can be used to solve the task, as well as each instance that has been associated with the task in the past. Key to the Instance Model is the assumption of a race between the algorithm and the stored instances in which the fastest process automatically wins the race, that is, determines the response. Due to the obligatory accumulation of memory instances, the probability that the memory retrieval wins the race, increases with practice resulting in a gradual transition from algorithm-based to memory-based processing. By means of Monte Carlo simulations, Logan (1988, 1992) has confirmed that his simple and elegant model produces a gradual transition from algorithm-based to memory-based task processing with task practice. Empirically obtained learning curves for means and standard deviation (that were aggregated over participants) were well fit by power functions with equal exponents. Thus, following the Instance Model, a strategy change from algorithm-based to memory-based processing occurs when the number of memory instances is large enough to control performance; the strategy change is thus an inevitable consequence of task practice (for a similar view see, e.g., Siegler, 1988; Siegler & Stern, 1998). Palmeri (1997, 1999; Nosofsky & Palmeri, 1997) has slightly extended the assumptions of the Instance Model in his Exemplar-Based Random-Walk model (EBRW) by arguing that the degree of similarity between a given task and the stored instances determines the speed of memory retrieval. On the basis of an instance-accumulation process, this model also predicts a gradual transition from algorithm-based processing to memory retrieval, but due to the dynamic H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 497 similarity-based mechanism additionally allows accounting for the transfer to other tasks that is sometimes found in cognitive skill acquisition (e.g., Lassaline & Logan, 1993). The Component Power Laws Model (CMPL) proposed by Rickard (1997, 1999), refers to a strengthening rather than an instance-accumulation process. When a task is performed, two alternative strategies, algorithm-based processing and memory retrieval, are assumed to be strengthened. Due to an inhibitory link between the memory-retrieval strategy and the algorithm, the memory retrieval will win the competition at some point during practice, thus leading to a strategy shift from algorithm to memory retrieval. On the whole, the theoretical views described above all share the assumption that strategy shifts during task practice are driven by automatical learning mechanisms and are inevitable consequences of task practice. Specifically, the CMPL model assumes that performance depends upon the strength of memory representations. If the strength of a memory representation exceeds a threshold, then a strategy shift to memory-retrieval-based processing occurs. Analogously, strategy shifts in the Instance model depend on the number of stored memory instances. If the number is large enough, then the memory instances come to control performance. Put somewhat cautiously, there exist no explicit assumptions in these models on how, when, and why voluntary processes might come to affect performance. Put more strongly, due to the focus on data-driven learning processes, the theories leave no room for the potential influence of voluntary processing. However, Haider and Frensch (2002), using the alphabet verification task (AVT), have recently presented empirical findings that appear to be difficult to reconcile with the view that strategy shifts are an inevitable consequence of task practice. The authors asked participants to evaluate letter strings shown on a computer screen by indicating whether or not the strings followed the alphabet. The strings (e.g., ‘‘C D E F G [4] L’’) consisted of a letter-digit-letter triplet (e.g., ‘‘G [4] L’’) and a varying number (0, 1, 2, 3, or 4 letters) of additional letters (e.g., ‘‘C D E F’’), and contained both task-relevant and task-irrelevant information. The triplet part of the strings was task relevant whereas the additional letters were completely irrelevant for the participantsÕ task because errors never occurred in these letters. Thus, the yes–no decision required of participants could be performed without error based on a consideration of the triplets alone. Importantly, participants were not informed about this characteristic of the task. Instead, they were only instructed to examine whether or not a given letter string followed the alphabet. Results of experiments with the AVT have shown that, with practice, participants become increasingly likely to switch from a processing-all-elements strategy to a new strategy. The new strategy consists of processing the task-relevant letter-digit-letter triplet only (Haider & Frensch, 1996, 1999a, 1999b, in press) and appears to be adopted rather abruptly. Second, the newly adopted strategy is transferred to structure-equivalent but never-before-encountered items (Haider & Frensch, 1996, Experiment 3). Third, speed-accuracy instructions affect how quickly during practice the new strategy is adopted (Haider & Frensch, 1999b). The second and third findings, in particular, suggest that the switch to the new strategy is at least partly based on voluntary processing. Taken together, this pattern of results suggests that participants in the AVT create, with practice, a new and more efficient task strategy that they abruptly adopt. The switch to the new strategy appears to be mediated by a deliberate decision rather than to occur more or less automatically. Although there thus exists some empirical evidence that is consistent with the view that voluntary processes might be involved in a strategy switch, the existing evidence is entirely based on comparisons of experimentally manipulated groups and does not allow for a direct 498 H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 within-subjects comparison of participants who adopted versus participants who did not adopt a new strategy over the course of practice. The main purpose of the present experiments was to test more directly than has been done in the past whether a strategy switch is an inevitable consequence of task processing or is caused by voluntary processing. The logic of the experiments conducted is based on the argument that a strategy switch can be said to be voluntary controlled when participants (a) transfer the new strategy to dissimilar information never encountered before and by doing so accelerate the process of skill acquisition, (b) are able to verbally describe the task regularity that allows for generation and application of the new strategy, and (c) decide against adopting the new strategy when available evidence suggests that the new strategy cannot be used for the entire range of problems encountered. All experiments used the AVT. In Experiments 1a, 1b, and 1c, individual participantsÕ performances were continuously monitored to find out at which point during task practice the new strategy (i.e., processing the response-relevant letter-digit-letter triplet only) was adopted and consistently applied. The performance of participants switching to the new strategy could then be compared to the performance of participants who did not adopt the new strategy. Experiment 1a first established the validity of our assessment of a strategy switch. In Experiments 1b and 1c, we tested whether participants exhibiting a strategy shift are able to (a) transfer the strategy to information never encountered before and (b) verbally describe the task regularity that allows for the construction and use of the new strategy in the first place. In Experiment 2, we tested whether or not the switch to a new strategy is moderated by knowledge about the strategyÕs range of applicability. 2. Experiment 1a The main purpose of Experiment 1a was to introduce an assessment of whether or not an individual participant had adopted a new strategy and to test if the assessment was reliable. The experiment was divided into two phases, a training phase and a transfer block. During the training phase, participants evaluated correct and incorrect alphabetic strings in the AVT; for incorrect strings, the location of errors was fixed such that errors always occurred within the letter-digit-letter triplet component of the strings. ParticipantsÕ performance in the training phase was continuously monitored for the occurrence of an abrupt performance shift. A strategy switch (i.e., from processing all elements of the strings to processing the relevant letter-digit-letter triplet only) was assumed to have occurred when the response times for two consecutive practice blocks differed by more than 1000 ms and when the latter response time was faster than the earlier response time.1 When such a latency discontinuity was observed, participants received one more practice block to ensure that the latency drop was consistent. Then, the transfer block was administered. Participants not exhibiting a latency differ1 A threshold of 1000 ms was selected based on a re-analysis of data from two previously published experiments with the AVT (N = 90, 7 practice blocks, Haider & Frensch, 1996). The re-analysis had shown that almost 85% of participants who produced a latency difference of more than 1000 ms in two consecutive practice blocks also showed a high error rate for incorrect strings in which the error location was outside the triplet component of the string (see Experiment 1a for further details on this relation). H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 499 ence of more than 1000 ms over the entire course of the training phase (i.e., eight practice blocks), received the transfer block after their final practice block. The transfer block served to validate our assessment of a strategy shift. Therefore, some of the incorrect strings in the transfer block contained a correct letter-digit-letter triplet but an error in the formerly redundant string part (i.e., the additional letters). These incorrect strings are called incorrect irregular strings to distinguish them from incorrect regular strings. In the latter strings, errors occur within the letter-digit-letter triplet of a string. If indeed a latency drop of more than 1000 ms signals a switch to the new strategy (i.e., processing the relevant letter-digit-letter triplet only), then participants showing such a latency discontinuity should not notice the errors in incorrect irregular strings; consequently, they should produce high error rates for incorrect irregular strings. Participants who do not switch to the new strategy, on the other hand, and presumably continue to process all elements of the strings, should notice the errors in incorrect irregular strings, and should thus show a much lower error rate for the incorrect irregular strings. 2.1. Method 2.1.1. Research participants Thirty-three male students at the University of the Federal Armed Forces, Hamburg, participated in the experiment. Their age ranged from 21 to 27 years (M = 24.2; SD = 1.8). All participants received course credit in exchange for participation. 2.1.2. Materials 2.1.2.1. Stimulus and apparatus. One hundred and twenty alphabetic strings were used in the experiment. All strings contained a digit in brackets, which was always the digit ‘‘4.’’ The digit indicated that the next 4 consecutive letters of the alphabet needed to be skipped at this string location, and that the string would continue with the fifth letter. Thus, the string ‘‘E [4] J’’, for instance, was to be interpreted as ‘‘E, skip F, G, H, I, continue with J.’’ One hundred of the 120 alphabetic strings were regular; that is, these strings were either correct or contained an error within the letter-digit-letter triplet component of a string. The 100 regular strings were composed of 10 different correct letter-digit-letter triplets (e.g., ‘‘E [4] J’’), and 10 different incorrect letter-digit-letter triplets (e.g., ‘‘E [4] I’’). The triplets could begin with either the letters E, F, G, H, I, J, K, L, M, or N. Each of the correct or incorrect letter-digit-letter triplets could be displayed together with an additional 0, 1, 2, 3, or 4 letters that always preceded the triplet (e.g., ‘‘A B C D E [4] J’’). The remaining 20 alphabetic strings were irregular, as in these strings the triplet was always correct, but an error occurred within the additional letters (e.g., ‘‘ACDEF [4] K’’). The position of this error was systematically varied across all possible string positions and string lengths. The irregular incorrect strings were only used in the transfer block. Strings were presented at the center of a 17-in. (27.8 cm) diagonal video screen controlled by a PC-AT-80486 or by a PC-Pentium. The letters were approximately 1.0 cm · 0.8 cm in size. Consecutive letters appeared approximately 0.8 cm apart on the screen. Participants responded by pressing either the ‘‘<’’ or the ‘‘-’’ key on the second row from the bottom on a PC-AT keyboard. Half of the participants were instructed to use the ‘‘<’’ key to indicate that a string was correct and the ‘‘-’’ key to indicate that the string was incorrect; for the other half, key assignment was reversed. 500 H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 2.1.3. Procedure For all participants, the experiment began with extended computerized instructions. The characteristics of the letter strings were described and participants were shown how to evaluate the strings. The initial instruction was followed by 10 practice trials. Participants were not told that the position of the errors in incorrect strings was fixed. Instead, they were instructed to pay close attention to the entire string because errors could occur anywhere in the string. The training phase followed the initial instruction and the practice trials, and was introduced with an additional short instruction. All participants were again asked to pay close attention to the entire string and to optimize both speed and accuracy of responding. The training phase contained a maximum of 8 practice blocks. Participants received feedback when they responded incorrectly. In addition, participants received feedback concerning their mean latency and their mean percentage of error at the completion of each practice block. The transfer block followed the training phase without announcement in order not to alert participants to the introduction of the irregular incorrect strings. In the transfer block, participants received no feedback when they erroneously accepted a string with an error outside the triplet as correct. Each trial in both the training phase and the transfer block began with the presentation of a fixation cross that appeared for 500 ms at the center of the screen. The screen then turned black, and an alphabetic string was displayed immediately thereafter. The string remained on the screen until a participant made a response. 2.1.4. Diagnosis of strategy switch While participants performed the training phase, it was continuously monitored whether or not a strategy shift had occurred. A strategy shift was indicated by a drop in response time between two consecutive practice blocks that exceeded 1000 ms. When such a drop in response time occurred, participants received one more practice block followed by the transfer block. The additional practice block was given in order to ensure that the latency decline was consistent. When no strategy shift occurred, participants received the transfer block after their final practice block. 2.1.5. Design Dependent variables were the individual median verification times (RTs) and the mean error rates per practice block. Strategy Shift (occurred vs. not occurred) served as the only between-subjects factor (post hoc) in the experiment. Within-subjects factors were Practice Block (1, 2, 3 vs. last practice block) and String Type (correct, incorrect regular, and incorrect irregular). In all analyses reported here and in the remaining experiments, the adopted significance level was a = .05. For significant effects, individual p values are not reported. 2.2. Results For all participants, the mean error rates per practice block were computed first. Participants with error rates higher than 15% in each practice block were excluded from further analyses (N = 2). The mean error rates per practice block for the remaining participants were rather low and varied between 2.65 and 4.33%. No speed-accuracy trade-off was observed, r (223) = .13. H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 501 For each of the remaining participants, median response times were computed separately for each practice block and the transfer block, and for the three types of strings (correct, incorrect regular, incorrect irregular). 2.2.1. Occurrence of strategy shifts Fifteen of the 31 participants in the experiment showed a larger-than-1000-ms latency discontinuity during the training phase. These participants completed an average of 6.4 practice blocks before the transfer block was administered. Thus, almost 50% of the participants showed a strategy shift. 2.2.2. Training phase Fig. 1 compares the mean latencies for correctly evaluated strings in Training Blocks 1, 2, 3, and in the last training block for participants who showed a strategy shift versus participants who did not show a shift. As is evident from the figure, despite a shorter training phase, participants with a shift in strategy were significantly faster by the end of the training phase than participants without such a shift. A 2 (Strategy Shift) · 4 (Practice Block) mixed-block design analysis of variance (ANOVA) yielded a significant main effect of practice block, F (3, 87) = 180.50, MSe = 378734.9, and a significant interaction between strategy shift and practice block, F (3, 87) = 11.85, MSe = 378734.9. As can be seen from the 95% between-subject confidence intervals displayed in Fig. 1, the difference between the two strategy-shift groups was not significant in Training Blocks 1, 2, and 3 but was significant in the last training block. 2.2.3. Transfer block Table 1 displays the mean error rates in the last training block and the transfer block. As can be seen, participants demonstrating a strategy shift do not differ from participants with no shift in their error rates for regular strings, both in the last training block and in the transfer block. However, they visibly differ in their error rates for incorrect irregular strings (transfer block). Fig. 1. Mean RTs in practice blocks 1, 2, 3, and the last practice block of Experiment 1a. Error bars represent the 95% between-subjects confidence intervals (Loftus and Masson, 1994). 502 H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 Table 1 Mean error rates in the last training block and the transfer block (regular and irregular strings) of Experiment 1a Error rate Strategy shift No strategy shift Confidence interval Last training block regular strings Transfer block regular strings Transfer block irregular strings 1.94 2.33 1.60 3.31 70.94 41.67 9.3 8.95 Confidence interval refers to the 95% within-subjects confidence interval (Loftus and Masson, 1994). Table 2 Results of the discriminant function analysis with error rate as independent variable and diagnosed strategy shift as grouping variable for Experiment 1a Error rate Strategy shift No strategy shift Total Total High Low 10 5 15 5 11 16 15 16 31 v2 (df = 1) = 3.89, p < .05. A 2 (Strategy Shift) · 3 (String Type: regular, last training block; regular, transfer block; and irregular, transfer block) mixed-block design ANOVA yielded significant main effects of strategy shift, F (1, 29) = 5.67, MSe = 493.8, and of string type, F (2, 58) = 74.53, MSe = 420, as well as a significant interaction between these two factors, F (2, 58) = 7.75, MSe = 420. To test how well the size of the errors for irregular strings statistically separates participants with and without strategy shift, a discriminant function analysis was computed with error rate for irregular strings as independent variable and strategy shift as grouping variable. Table 2 displays the results. The analysis correctly classified 21 of the 31 participants. Approximately, 67% of the participants who displayed a strategy shift in the training phase showed an error rate that was higher than 50%, whereas 69% of the participants without such a shift made less than 50% errors. This result strongly confirms our assessment of the occurrence of a strategy switch. Participants who are diagnosed as strategy switchers are much more likely to make errors on irregular strings than participants diagnosed as non-strategy switchers. 2.3. Discussion Experiment 1a provided two main results. First, almost 50% of the participants exhibited a strategy shift as indexed by a RT discontinuity of more than 1000 ms in two consecutive practice blocks. Participants who shifted to a new strategy evaluated the alphabetic strings more quickly at the end of the training phase than did participants without a diagnosed strategy shift, although they had performed fewer training blocks. Second, participants who shifted to a new strategy were more likely not to notice errors in the formerly redundant string part in the transfer block. By comparison, only few participants without a diagnosed strategy shift behaved like participants with a shift in the transfer block or the post-experimental questionnaire. Most of the H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 503 participants without a shift detected the errors in the formerly redundant string part. The finding that some participants without a strategy shift behaved as did participants with a diagnosed shift is probably due to the fact that our classification procedure was not perfect. Taken together, the obtained pattern of results strongly supports the assumption that the abrupt RT discontinuities observed during the training with the AVT indicate a qualitative change in task performance due to the adoption of a new strategy. The new strategy adopted by participants who exhibit an abrupt RT discontinuity consists of processing the letter-digit-letter part of the strings exclusively. 3. Experiments 1b and 1c As mentioned before, the logic of Experiments 1b and 1c rests on the argument that a strategy switch is said to be voluntary when participants (a) transfer the new strategy to dissimilar information never encountered before, and when they (b) are able to verbally describe the task regularity that allows for the generation and application of the new strategy. In Experiments 1b and 1c, therefore, we assessed whether participants exhibiting a strategy shift (a) transfer the new strategy to information never encountered before, and (b) are able to verbally describe the task regularity immediately after using the new strategy consistently. The setups of Experiments 1b and 1c were very similar to the setup of Experiment 1a. Again, participants performed the AVT in a training phase, followed by a transfer phase. Again, their performance during the training phase was continuously monitored for the occurrence of an abrupt RT discontinuity signalling the occurrence of a strategy switch (i.e., from processing all elements of the strings to processing the letter-digit-letter triplet only). When a strategy switch was detected, participants received one more practice block and then performed either two transfer blocks with a new structure-equivalent task that consisted of never-before-encountered strings (Experiment 1b), or were asked to immediately answer a post-experimental questionnaire (Experiment 1c). If a strategy switch is mediated by voluntary processing on behalf of the participants (i.e., a deliberate decision to use the new strategy), then participants with a diagnosed strategy switch should be able to transfer their new strategy to novel stimuli, that is, their learning of the novel stimuli should be accelerated relative to the situation where they do not possess the new strategy. Accordingly, they should re-reach their performance level of the last training block within a few practice blocks. On the other hand, if a strategy switch is caused by data-driven, automatic processing, then changing the set of stimuli should bring strategy shiftersÕ speed of processing down to the level of participants who did not switch their strategy (Experiment 1b). In addition, participants with a diagnosed strategy shift should be able to verbally describe the task regularity that allows for the generation and application of the new strategy (Experiment 1c). 3.1. Method 3.1.1. Research participants Thirty male students at the University of the Federal Armed Forces, Hamburg, served as participants in Experiment 1b. Their age ranged from 21 to 30 years (M = 24.9; SD = 1.9). Twenty-five female and seven male students at the University of Cologne served as participants 504 H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 in Experiment 1c. Their age ranged from 19 to 33 years (M = 22.9; SD = 3.2). All participants received course credit in exchange for participation; none of them had participated in Experiment 1a. 3.1.2. Materials 3.1.2.1. Stimulus and apparatus. The materials and apparatus used were the same as those described for Experiment 1a, with the following exceptions: The training phase of Experiment 1b contained only 80 regular strings; eighty different strings were used in the two transfer blocks. Half of the participants received strings that were composed of correct and incorrect letter-digit-letter triplets beginning with the letters E to L as training items and strings composed of correct and incorrect letter-digit-letter triplets beginning with the letters M to T as transfer items. For the other half of participants the assignment of training and transfer items was reversed. Participants were randomly assigned to one of the two assignment groups. In Experiment 1c, the training items were the same as in Experiment 1a. There was, however, no transfer block. Instead, participants were presented with a post-experimental questionnaire asking them to verbally describe the characteristics of the alphabetic strings they had evaluated and the strategies they had used to perform the task. 3.1.3. Procedure The procedure used in Experiments 1b and 1c was the same as that described for Experiment 1a, with the following exceptions: In Experiment 1b, the transfer phase contained two blocks with a structure-equivalent task but new strings. Participants received a short instruction when they had completed their last training block. In Experiment 1c, participants received a post-experimental questionnaire immediately after completing the training phase. They were asked to verbally describe the characteristics of the alphabetic strings they had evaluated and the strategies they had used to process the strings at the beginning and at the end of the training phase. Participants in Experiment 1b also received this questionnaire after completion of the transfer phase. 3.1.4. Diagnosis of strategy switch The assessment of whether or not a strategy change had occurred in the training phase followed the procedure described for Experiment 1a. 3.1.5. Design Dependent variables were again the individual median verification times and the mean error rates per practice block. Strategy Shift (occurred vs. not occurred) served as the only between-subjects factor (post hoc) in both experiments. Within-subjects factors were Practice Block (1, 2, 3 vs. last practice block) and String Type (correct and incorrect, Experiment 1b only). 3.2. Results For all participants, the mean error rates per practice block were again computed. Participants with error rates higher than 15% in each training block were excluded from further analyses H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 505 (N = 0 and N = 3 in Experiments 1b and 1c, respectively). The mean error rates per practice block for the remaining participants were rather low and varied between 1.53 and 6.58% in Experiment 1b and between 2.89 and 6.17% in Experiment 1c. No speed-accuracy trade-offs were observed, r (203) = .21, and r (202) = .01, in Experiments 1b and 1c, respectively. For each of the remaining participants, median response times were computed separately for each training and transfer block, and for the different types of strings (correct and incorrect strings). 3.2.1. Occurrence of strategy shifts Fourteen of the 30 participants in Experiment 1b and 13 of the 29 participants in Experiment 1c showed a larger-than-1000-ms latency discontinuity in the training phase. These participants completed an average of 5.3 training blocks (Experiment 1b) and 5.8 training blocks (Experiment 1c) before the transfer blocks (resp. the questionnaire) was administered. Thus, again, approximately 50% of the participants showed a strategy shift. 3.2.2. Training phase Table 3 (Experiment 1b and Experiment 1c) display the mean latencies for correctly evaluated strings in Training Blocks 1, 2, 3, and in the last Training Block for participants who showed a strategy shift versus participants who did not show a shift. Again, despite shorter training phases, participants with a shift in strategy responded significantly faster by the end of the training phase than participants without a shift. Experiments 1b and 1c yielded significant main effects of strategy switch, F (1, 28) = 5.80, MSe = 2364787.0 and F (1, 27) = 4.70, MSe = 3580425, for Experiments 1b and 1c, respectively, of practice block, F (3, 84) = 115.4, MSe = 398028.3 and F (3, 81) = 89.09, MSe = 300230, for Experiments 1b and 1c, respectively, as well as significant interactions between these two factors, F (3, 84) = 10.74, MSe = 398028.3 and F (3, 81) = 6.91, MSe = 300230, for Experiments 1b and 1c, respectively. As can be seen from the 95% betweensubject confidence intervals given in Table 4, the difference between the two strategy-change conditions was not significant for Training Blocks 1 and 2 but was significant for the last training block. 3.2.3. Transfer phase (Experiment 1b only) Table 4 displays the mean difference between the latency in the first transfer block and the latency in the last training block as well as the mean difference between the latency in the second transfer block and the latency in the last training block, separately for participants with and withTable 3 Mean latencies for correctly evaluated strings in Training Blocks 1, 2, 3, and the last training block of Experiments 1b and 1c Experiment 1b First block Second block Third block Last block Experiment 1c Strategy shift No strategy shift Strategy shift No strategy shift 5088.31 3833.16 2598.37 1190.39 4902.38 (623.99) 4109.47 (479.10) 3831.29 (512.69) 3080.42 (336.73) 3746.55 3129.99 2407.20 965.12 4314.90 (657.23) 3400.33 (535.75) 3107.09 (573.74) 2491.57 (362.12) The values in parentheses represent the 95% between-subjects confidence intervals (Loftus and Masson, 1994). 506 H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 Table 4 Mean differences between the latency in the transfer block and the latency in the last practice block and between the latency in the second transfer block and the last practice block in Experiment 1b First transfer block Second transfer block Strategy shift No strategy shift 1103.82 170.75 1529.27 (425.45) 332.43 (175.27) The values in parentheses represent the 95% between-subjects confidence intervals (Loftus and Masson, 1994). out a detected strategy shift. The mean latencies in the first transfer block increased by around 1000 ms, due to surprise or to item-specific learning during training, or to both. In the second transfer block, however and more importantly, participants with a diagnosed strategy shift reached the latency level of their last training block, whereas participants without a diagnosed shift did not, suggesting that participants with a diagnosed strategy shift indeed were able to accelerate learning with the novel stimuli. In fact, they were even faster verifying the novel strings in the last transfer block than in the last training block (that is after approximately 5 blocks of training). In contrast, participants with no diagnosed strategy shift did not reach their performance level of the last training block within the two transfer blocks. 3.2.4. Post-experimental questionnaire To examine whether the regularity of the task material and the newly adopted strategy was available to verbal report, participants in both experiments were asked, upon completion of the experiment, to verbally describe the characteristics of the alphabetic strings they evaluated and the strategies they had used to perform the task. The results are displayed in Tables 5 and 6. Seventy-nine percent (Experiment 1b) and 92% (Experiment 1c) of the participants with a strategy shift had noticed that errors never occurred Table 5 Verbalization of the task regularity as a function of diagnosed strategy shift in Experiment 1b Regularity in structure Strategy shift No strategy shift Total Total Verbalized Not verbalized 11 4 15 3 12 15 14 16 30 v2 (df = 1) = 8.57, p < .05. Table 6 Verbalization of the task regularity as a function of diagnosed strategy shift in Experiment 1c Regularity in structure Strategy shift No strategy shift Total v2 (df = 1) = 11.02, p < .05. Total Verbalized Not verbalized 12 5 17 1 11 12 13 16 29 H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 507 in the ‘‘irrelevant’’ string part, and had retrieved the answer to the triplet from memory while ignoring the ‘‘irrelevant’’ letters. In contrast, only 25% (Experiment 1b) and 31.3% (Experiment 1c) of the participants without a shift were able to describe the characteristics of the task or an alternative strategy. 3.3. Discussion Experiments 1b and 1c make two important points. First, participants with a diagnosed strategy shift in the training phase are likely to transfer their newly adopted strategy to stimuli they have never encountered before; in other words, the new strategy is item-general (Anderson, 1993; Smith, Langston, & Nisbett, 1992; Whittlesea & Dorken, 1993, 1997). Second, when asked to verbalize the characteristics of the alphabetic strings immediately after the new strategy has been adopted, participants with a diagnosed strategy shift are very likely to be able to do so. On the whole and if one buys into our argument that a strategy switch can be said to be voluntary when participants are able to (a) transfer the new strategy to dissimilar information never encountered before, (b) verbally describe the task characteristics that allow for the generation and application of the new strategy when asked to do so, then the obtained pattern of results is consistent with the idea that a strategy shift is caused by a deliberate decision to adopt a new strategy. 4. Experiment 2 In Experiment 2, we explored the last part of our argument, namely the idea that a strategy switch is voluntary controlled when participants decide against adopting a new strategy when available evidence suggests that the new strategy cannot be used for the entire range of problems encountered. In this experiment, we manipulated the extent to which the alternative strategy used by our strategy-shift participants in Experiments 1a through 1c (i.e., processing the task-relevant letter-digit-letter triplet only) was a valid strategy to solve the problems encountered by participants. In the ‘‘Always-Regular’’ condition, the alternative strategy could be used to correctly solve all strings displayed in the experiment. In other words, all incorrect strings contained errors in the letter-digit-letter triplet; they were all regular, to use the language introduced in Experiments 1a through 1c. In the ‘‘90%-Regular’’ condition, most, but not all, strings could be correctly solved by the alternative strategy. That is, 10% of the incorrect strings contained errors that were not placed in the digit-letter-digit triplet; they were thus irregular. Thus, exclusive use of the alternative strategy would lead to errors. The two groups, Always-Regular and 90%-Regular, were equated in terms of the number of strings evaluated that could be solved by the alternate strategy (i.e., socalled ‘‘regular’’ strings). Therefore, the 90%-Regular group evaluated more strings overall than did the Always-Regular group. Our expectation was that about half of the participants in both conditions (if the results of Experiments 1a to 1c are replicated) would discover the alternative strategy of focusing exclusively on the letter-digit-letter triplet of the strings. If strategy switch is based on a deliberate decision of participants, that is, on voluntary processing, then participants in the Always-Regular condition 508 H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 should be much more likely to adopt the strategy than participants in the 90%-Regular condition on the premise that participants in the latter condition would want to avoid making errors. If, on the other hand, strategy switch is an inevitable consequence of task practice, then participants in the two conditions should be equally likely to show a strategy switch. To assess how many participants in the two conditions became aware of the alternative strategy during the course of practice, 10 correct and 10 incorrect regular speeded trials were added at the end of each block in which participants were asked to evaluate strings but had limited time to do so. While time to generate a response was unlimited in the remainder of the practice block, the time available in the speed trials was continuously adjusted for each participant and each practice block and was set equal to the participantÕs mean latency for evaluating strings of length 7 (i.e., triplet plus 4 letters) in the present practice block. The basic idea was that the reduced time available to participants in the speed block would ‘‘pressure’’ them into applying the alternative strategy even when they had voluntarily decided not to adopt the strategy in general. Ideally, thus, if the assumption of a voluntary controlled strategy shift is correct, then the performance of participants in the Always-Regular and the 90%-Regular conditions should be the same in the speed trials, indicating that an approximately equal number of participants in the two conditions had the new strategy at their disposal. However, the conditions should differ in terms of the number of participants switching to the new strategy for the remainder of the practice blocks. Participants in the Always-Regular condition should be much more likely to change to the new strategy than participants in the 90%-Regular condition. For baseline and comparison purposes, we included an additional 60%-Regular control condition (i.e., 60% of all incorrect trials were regular; 40% were irregular) in which the alternative strategy (i.e., processing letter-digit-letter triplet only) was not a valid strategy to process the strings encountered because errors in incorrect strings were almost as likely to occur outside as inside the triplet. Thus, participants in the control condition should neither discover the alternative strategy nor adopt it. Experiment 2 was again divided into a training phase and a transfer phase. As in Experiment 1a, all participants received a transfer block at the end of the training phase that contained 20% irregular and 80% regular incorrect strings. Experiment 1a has demonstrated that participants showing a strategy shift also exhibit high error rates for strings with errors outside the triplet. Thus, if the consistency of the incorrect strings processed in the training phase affects strategy application, then participants in the Always-Regular condition should show higher error rates for strings with irregular errors than participants in the 90%-Regular and in the 60%-Regular conditions. Finally, in contrast to the previous experiments, individual participantsÕ performances were not monitored for the occurrence of a strategy shift during the training phase in Experiment 2; we also did not administer a post-experimental questionnaire. 4.1. Method 4.1.1. Research participants Forty-six female and 20 male students from introductory courses of psychology at HumboldtUniversity, Berlin, Germany, served as participants. Age of participants ranged from 19 to 38 years (M = 23.3, SD = 3.79). All participants received course credit in exchange for participation. H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 509 4.1.2. Materials 4.1.2.1. Stimuli and apparatus. Overall, 130 alphabetic strings were used. As in Experiment 1a through 1c, the strings were composed of 10 different correct (e.g., ‘‘E [4] J’’) and 10 different incorrect (e.g., ‘‘E [4] I’’) letter-digit-letter triplets beginning with the letters E, F, G, H, I, J, K, L, M, or N. The digit in brackets was again always equal to 4, and each of the correct or incorrect letter-digit-letter triplets could be presented along with either 0, 1, 2, 3, or 4 additional letters. As before, the additional letters always preceded the triplet (e.g., ‘‘A B C D E [4] J’’). As in the transfer block of Experiment 1a, the 30 irregular incorrect alphabetic strings with errors outside the triplet contained a correct triplet. String presentation and response-key assignment followed the format of Experiments 1a through 1c. 4.1.3. Procedure The procedure used in Experiment 2 was the same as that described for Experiment 1a, with the following exceptions: The training phase included 8 practice blocks in the Always-Regular and the 60%-Regular conditions and 9 practice blocks in the 90%-Regular condition. The training of the latter condition was prolonged by one block in order to equalize the number of regular strings for which the additional letters were redundant in the Always-Regular and the 90%-Regular conditions. Except for the first practice block, each of the remaining practice blocks consisted of a training part with 40 correct strings and 40 incorrect strings and a speed part with 10 correct and 10 incorrect letter strings. In the Always-Regular condition, the 40 incorrect alphabetic strings were all regular incorrect strings. In the 90%-Regular condition, participants received 35 regular incorrect strings with the errors located within the triplet and 5 irregular incorrect strings with the errors located outside the triplet. In the 60%-Regular condition, participants received 20 regular incorrect and 20 irregular incorrect strings with errors positioned outside the triplet. The irregular incorrect strings in the 90%-Regular and the 60%-Regular conditions were randomly distributed within the training part of each block. For all participants, the speed part of each practice block contained only regular alphabetic strings. In addition, response time was limited to the individual mean response time needed in the training part of the actual practice block for letter strings of string length 7 (i.e., triplet plus 4 letters). When the time limit was reached, the alphabetic string disappeared from the screen and participants were told to provide an answer. When participants did not respond within the individual determined time limit, an error sign was displayed on the screen for a duration of 1000 ms. Then, the next trial started. When the 8th practice block in the Always-Regular and the 60%-Regular conditions and the 9th practice block in the 90%-Regular condition had been completed, a transfer block followed. The transfer block was identical to that used in Experiment 1a, except that it was divided in a training part with 10 irregular, 30 regular incorrect, and 40 correct strings and a speed part with 10 correct and 10 regular incorrect strings. No further instruction was given before the transfer block was administered. The presentation of the alphabetic strings followed the same format as in Experiments 1a through 1c. However, in contrast to the previous experiments, participants never received feedback. 510 H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 4.1.4. Design Dependent variables were the individual median verification times and the mean error rates per practice block. The only between-subjects factor was Consistency of Errors (Always-Regular, 90%-Regular, and 60%-Regular). Within-subjects factors were Practice Block (1 through 8), String Length (3 through 7), and Error Type (regular vs. irregular). 4.2. Results For all participants, the mean error rates per practice block were computed. As before, participants with error rates higher than 15% in each practice block were excluded from further analyses (N = 3, N = 2, and N = 1, for the Always-Regular, 90%-Regular, and 60%-Regular condition, respectively). Mean error rates for the remaining participants were rather low and varied between 4.10 and 7.10%. No speed-accuracy trade-off was observed (r = .009, r = .117, and r = .068, in the Always-Regular, the 90%-Regular, and the 60%-Regular condition, respectively). For each of the remaining participants, median response times were computed, separately for the training and speed parts of each practice block and separately for the three types of strings (correct, incorrect regular, incorrect irregular) and for each string length, respectively. 4.2.1. Occurrence of a strategy switch As a first assessment of the occurrence of strategy switches in the three experimental conditions, we performed a post hoc analysis of abruptly occurring response-time discontinuities based on the same criterion we had used in the previous experiments (i.e., RT decline between two consecutive blocks of more than 1000 ms). According to this analysis, 63% of the participants in the AlwaysRegular condition, 20% of the participants in the 90%-Regular condition, and 10% of the participants in the 60%-Regular condition showed a strategy shift. Thus, this admittedly crude analysis supports the assumption that a strategy switch is mediated by voluntary processing. 4.2.1.1. Stringlength effect in the training phase. As shown in previous experiments (Haider & Frensch, 1996, 1999a, 1999b), the size of the stringlength effect reflects to what extent participants ignore the letters outside the letter-digit-letter triplet when processing alphabetic strings. The stringlength effect refers to the effect of the length of the string to be evaluated on response time. When a participant processes only the letter-digit-letter triplet of strings, then the processing time for the participant should be independent of the length of the string. Thus, when the results of our crude analysis of the RT discrepancies are any valid, then the stringlength effects should be quite different for the three experimental groups. More precisely, the stringlength effect should disappear, with task practice, in the Always-Regular condition but should not disappear and, in fact, should be comparable for the two remaining conditions. To assess the size of the stringlength effect in the three conditions, we computed the best fitting linear regression lines across the five string lengths separately for each participant and each practice block (for details see Haider & Frensch, 1996). Due to the differing numbers of incorrect irregular strings in the three conditions, we excluded all incorrect strings from this analysis. Fig. 2 displays the means slopes of the individual best-fitting regression lines for correct strings as a function of experimental condition and practice block. As can be seen from Fig. 2, first, the slopes in all three conditions increase from the first to the second practice block. Because we did not H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 511 Fig. 2. Means of best fitting regression slopes for correct strings for Experiment 2. Error bars represent 95% withinparticipant confidence intervals (Loftus and Masson, 1994). obtain this particular result when using alphabetic strings that begin, rather than end, with the letter-digit-letter triplet (Haider & Frensch, 1999a), we suspect that, at the beginning of training, the additional letters might have helped participants to find the first letter of the triplet in the alphabet. However, more importantly, the general pattern depicted in Fig. 2 suggests that the occurrence of irregular strings affected the likelihood of a strategy switch. After Practice Block 2, the sizes of the slopes remained relatively constant in the 60%-Regular and the 90%-Regular conditions, but declined in the Always-Regular condition. A 3 (Consistency of Errors) · 8 (Practice Block) mixed-design ANOVA on participantsÕ individual slope scores confirmed the visual impression. The main effect of consistency of errors was significant, F (2, 57) = 17.55, MSe = 38148.5, as were the main effect of practice block, F (7, 399) = 2.90, MSe = 8289.6, and the two-way interaction between consistency of errors and practice block, F (14, 399) = 3.87, MSe = 8289.6. Results of post hoc comparisons revealed that the 90%-Regular condition did not differ significantly from the 60%-Regular condition; both conditions differed, however, significantly from the Always-Regular condition. Thus, in accordance with the assumption that strategy switch is mediated by voluntary processing, that is, by a deliberate decision to change strategies, the analysis of the stringlength effects revealed that participants in the Always-Regular condition were more likely to change their strategy from processing all elements of the strings to processing the letter-digit-letter triplet only than were participants in the remaining conditions. This finding is consistent with the assumption that the adoption of a new strategy depends on the range of applicability of the strategy. 4.2.1.2. Transfer phase. As in Experiment 1a, participants in the present experiment received a transfer block upon completion of the training phase in which 20% of the incorrect alphabetic strings were irregular, that is, contained an error outside the triplet. In accordance with the results described thus far, we expected that the error rates for irregular incorrect strings should be higher in the Always-Regular condition than in the remaining conditions. To further validate our assumption that diminishing slopes in the training phase indicate a strategy shift, we also computed the correlation between the individual error rates for irregular incorrect strings in the trans- 512 H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 fer block and the slopes of the stringlength effects averaged across the last two training blocks. We expected that participants with flat slopes should also exhibit high error rates for irregular strings in the transfer block. Table 7 displays the mean error rates as a function of error type (last practice block/regular, transfer block/regular vs. transfer block/irregular). As can be seen from Table 7, error rates in the last training block as well as for regular strings in the transfer block did not differ between conditions. By contrast, error rates for irregular incorrect strings was highest for the Always-Regular condition followed by the 90%-Regular and the 60%-Regular conditions. The impression conveyed by Table 7 was confirmed by the result of a 3 (Consistency of Errors) · 3 (Error Type) mixed-design ANOVA (F (2, 57) = 28.75, MSe = 263.5; F (2, 114) = 124.89, MSe = 186.2 and F (4, 114) = 42.91, MSe = 186.2, for the main effects of consistency of errors and of error type as well as for the two-way interaction, respectively). In addition, post hoc comparisons confirmed that participants in the Always-Regular condition produced significantly more errors for incorrect irregular strings than participants in the 90%-Regular condition, F (1, 57) = 31.3, and the latter produced more errors than participants in the 60%-Regular condition, F (1, 57) = 10.2, both MSe = 595.1. If the experimental manipulation of consistency of errors affects the likelihood that the alternative strategy is adopted, then one would expect that participants in the Always-Regular condition should show flat stringlength effect slopes accompanied by high error rates for irregular incorrect strings. In contrast, participants in the 60%-Regular condition should show high slopes accompanied by small error rates. Participants in the 90%-Regular condition should behave similar to participants in the 60%-Regular condition. We therefore computed the correlation for all participants together to examine, at least descriptively, this overall trend. Fig. 3 relates the stringlength effect for correct strings (averaged across the last two training blocks) to participantsÕ mean error rate for irregular errors. As is suggested by Fig. 3, the overall correlation between the size of the stringlength effect and the error rate for irregular strings was highly significant, r (60) = .72. Moreover, a closer inspection of Fig. 3 confirms that, as expected, no participant in the 60%-Regular condition showed error rates higher than 50% and all participants showed relatively high slopes values. In contrast, almost all participants (14 out of 19) in the Always-Regular condition exhibited error rates higher than 60% that were accompanied by flat slopes. In addition, only 4 of the 20 participants in the 90%-Regular condition behaved like participants in the Always-Regular condition, while 16 participants behaved like participants in the Table 7 Mean error rates in the last training block and the transfer block (regular and irregular strings) of Experiment 2 Error Rate Always Regular 90%-Regular 60%-Regular Confidence interval Last training block regular strings Transfer block regular strings Transfer block Irregular strings 2.76 3.78 2.62 1.66 3.03 3.71 4.21 2.21 74.74 31.00 6.67 12.44 H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 513 Fig. 3. Scatter plots: error rate for strings with errors outside the triplet vs. mean slope (averaged over the last two training blocks; Experiment 2). 60%-Regular condition. Thus, in accordance with the results obtained for the training phase, the experimental manipulation clearly affected the likelihood that participants adopted the new strategy. 4.2.2. Knowledge of the alternative strategy Our main interest guiding the analyses of participantsÕ speed-trial performance was to examine whether or not the three conditions differed in their ability to correctly respond within the limited time interval. Our reason for introducing the speed trials was the assumption that the imposed time limit would ‘‘pressure’’ participants who had discovered the alternative strategy to use it when time was limited, although they might not have voluntarily selected the new strategy when time was not limited because they were aware that the strategy was not always applicable. In other words, we hoped that performance on the speed trials would reveal to what extent participants in the three experimental conditions had explicit knowledge about the alternative strategy. First, we determined the proportion of trials in which participants either responded incorrectly or too late (hereafter, individual percentage of errors). Second, in order to examine the stability of our obtained results, we computed individual percentages of errors for four different post hoc-determined time limits (hereafter, time windows), 100% (the originally imposed time limit), 90% (the original limit reduced by 10%), 80% (the original limit reduced by 20%), and 70% (the original limit reduced by 30%). Fig. 4 displays the mean percent error as a function of practice block and condition for the 100% as well as for the 90, 80, and 70% time windows. Two points depicted in Fig. 4 are worthwhile mentioning. First and most importantly, the 90%Regular condition differs from the 60%-Regular condition, but in contrast, does not differ from the Always-Regular condition. A 3 (Consistency of Errors) · 7 (Speed Block: 2–8) mixed-design ANOVA on the individual mean error rates for the 100%-time window confirmed this impression, and yielded a significant main effect of consistency of errors, F (2, 57) = 18.14, MSe = 7268. The post hoc comparisons testing for the differences between the 60%-Regular versus the 90%-Regular and the Always-Regular conditions were all reliable; no other contrast was significant. 514 H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 Fig. 4. Mean percentage errors as a function of practice block and condition for the 100, 90, 80, and the 70% time window in Experiment 2. Second, when comparing the patterns of results for the four time windows, it is obvious that the differences between the three conditions remain remarkably stable. Even at the 70%-time window, the performance of the 90%-Regular condition differed significantly from that of the 60%-Regular condition, F (1, 57) = 6.71, MSe = 2344.2, and did not differ from the Always-Regular condition (p > .7). Thus, manipulating the consistency of errors had very different effects on participantsÕ performance in the speed phase and on their actually changing their strategy. One possible objection to the latter conclusion is that the obtained pattern of results might be due to the fact that in the 60%-Regular condition time pressure in the speed trials was higher than in the remaining conditions because the time limit was based on latencies from 8 correct strings, 2 regular, and 6 irregular incorrect strings. By comparison, the time limit in the 90%-Regular condition was based only on 8 correct, 6 regular, and 2 irregular incorrect strings. Because strings with irregular errors might be evaluated faster than correct strings or regular incorrect strings, the higher frequency of this particular string type in the 60%-Regular condition might have caused higher time pressure and, consequently, might have been responsible for the obtained pattern of results. To address this objection, we conducted an additional analysis in which we used the individual mean latencies for correct strings of stringlength 3 and 4 (triplet only and triplet plus one letter) as time limit, and examined again the percentage of trials falling outside the time window (percentage of errors). The resulting mean percent error for each of the three conditions is displayed in Fig. 5. As can be seen, the picture displayed in Fig. 5 is slightly different from that depicted in Fig. 5. At the beginning of practice (Practice Blocks 2–5), the three conditions do not differ much in their performance suggesting that the ability to respond within the time limit develops during practice. After Practice Block 4, however, the mean percent error increases in the 60%-Regular condition, H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 515 Fig. 5. Mean percentage errors (defined as latencies longer than the individual mean latencies for correct and incorrect strings of stringlength 3 and 4) as a function of practice block and condition in Experiment 2. probably due to the decreasing time needed to evaluate strings of stringlength 3 and 4. In contrast, the error percent remains rather stable in the 90%-Regular condition after Practice Block 3. This impression was supported by the results of a 3 (Consistency of Errors) · 7 (Speed Block) mixed-design ANOVA. This analysis yielded a significant main effect of consistency of error, F (2, 57) = 5.90, MSe = 1129.5. Although the overall contrast between the 90%-Regular and the 60%-Regular conditions failed to reach significance, the post hoc computed difference for Practice Blocks 6 to 8 was significant, F (1, 57) = 5.54, MSe = 723.2. Thus, the results of the additional analysis support the conclusion that participants in the 90%-Regular condition have indeed learned that the letters outside the triplet are redundant. Taken together, the results of the speed-trial analyses show that, in contrast to the results of the analyses concerning actual strategy change, performance in the speed trials was mainly affected by the frequency of the incorrect irregular strings in the training phase. This suggests that manipulating the consistency of errors mainly affects explicit strategy application during training, but that it does not affect knowledge acquisition. 4.3. Discussion Experiment 2 provided two main results. First, the manipulation of the consistency of errors mainly affected the actual, overt strategy application in the training phase. Second, explicit knowledge of structural invariants in the strings processed as well as the availability of the new strategy, in contrast, depended upon the absolute frequency of deviant strings encountered during practice. On the one hand, participants in the 90%-Regular condition behaved very similar to participants in the 60%-Regular condition in the training and transfer phases. On the other hand, their performance did not differ from the performance of the Always-Regular condition in the speed trials. This latter result suggests that participants in the 90%-Regular condition had acquired at least some explicit knowledge about the redundancy of the additional letters in the strings they had evaluated, but did not use this knowledge to consistently change their strategy in the training phase. Taken together, the obtained pattern of results supports the idea that the strategy switch 516 H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 in the AVT is voluntary. Participants in the 90%-Regular condition appear to actively reject the option of adopting the new strategy because the evidence that is available to them suggests that the new strategy cannot be used for the entire range of problems encountered. 5. General discussion The main results of the experiments described in this article can be summarized as follows. First, Experiment 1a demonstrated that our assessment of strategy change in the AVT is valid. Second, Experiments 1b and 1c showed that the new strategy participants in the AVT adopt is item-general, and that participants are able to verbally describe the task characteristics that allow for the generation and application of the new strategy immediately after it has been adopted. Third, Experiment 2 demonstrated that participants are able to decide against adopting a new strategy when the evidence available to them suggests that the new strategy cannot be used for the entire range of problems encountered. On the whole, these findings are more consistent with the view that strategy change, at least in the context of the AVT, is mediated by deliberate, voluntary processing, than with the view that it is an inevitable consequence of task practice. Neither the Instance Model (Logan, 1988), nor the EBRW model (Palmeri, 1997), or RickardÕs CMPL model (1997) appear to be able to explain these findings because these models focus exclusively on data-driven learning processes. 5.1. What leads to verbalization of task constraints experienced during task practice? One specific and noteworthy question that is raised by the findings described above concerns the generation of explicit, verbalizable knowledge about task invariants that can be used to create new strategies. In other words, why is it that participants become aware of the fact that all (or most) strings processed in the AVT obey the constraint that errors always occur within the triplet component of the strings? The question is reminiscent of, and related to, a recent debate in the area of implicit learning, namely why some participants in implicit-learning experiments are able to verbally describe their acquired knowledge whereas other participants are not able to do so (e.g., Frensch et al., 2002; Frensch & Rünger, 2003; Reber, 1989, 1993; Shanks & St. John, 1994). Our current, speculative, answer to this question is as follows. In accordance with other authors (e.g., Bargh, Chen, & Burrows, 1996; Dienes & Perner, 1996, 1999; Farah, 1994; Koriat, 2000; Neumann & Klotz, 1994), we assume that data-driven, implicit learning processes are capable of changing task performance without generating verbally expressible knowledge. Implicit datadriven processes, thus, do not lead to verbally expressible, that is, explicit knowledge. Explicit knowledge, we suspect, is the result of voluntary inferential processes whose goal it is to explain perceived changes in task performance. Note that this position is congruent with KoriatÕs (2000) view distinguishing between two levels of experience, a higher level of controlled, explicit operations and a lower level of implicit operations that is characterized by a low degree of consciousness and is affected by automatic influences. There exist at least two plausible possibilities that would explain how perceived changes in task performance might trigger voluntary processing. First, data-driven implicit learning processes might lead to a continuous increase in speed of processing that might be noticed by participants H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 517 (e.g., Norman, 1968; Rickard, 1997). Second, it is conceivable that data-driven implicit learning processes, such as the accumulation of instances (e.g., Logan, 1988), lead to fast responses that compete with slower responses that are generated by higher cognitive processes (i.e., algorithmbased processing; e.g., Logan, 1988). While in the former case, speed of performance increases continuously, participants in the latter case might produce a mix of short response times that are based on data-driven processes (e.g., instance retrieval) and slow response times that are generated by some higher-order cognitive process. It is at least conceivable that participants might notice the unsystematic changes in their response times because they violate expectations participants might hold regarding their own task processing. The abruptly occurring violations of expectation, we presently assume, might serve as triggers for explicit inferential processes (for details on the Unexpected-Event Hypothesis see Frensch et al., 2002; Haider & Frensch, in press; see also, Whittlesea & Williams, 1998, 2000, 2001a, 2001b). Empirical support for the hypothesis that (a) the detection of a new strategy in the AVT is the result of explicit inferential processing, and that (b) explicit inferential processing, in turn, is initiated by participantsÕ perception of their own behavior, specifically of their own unsystematically occurring, unexpectedly fast responses comes from findings recently reported by Haider and Frensch (in press). The authors demonstrated that it is possible to affect the rate of participants who are able to verbally report an incidentally experienced regularity by manipulating the cause of unexpected events (i.e., premature responses) introduced during training. In their Experiment 1, Haider and Frensch (in press) showed that verbal report is moderated by participantsÕ belief to have produced premature responses. Specifically, when participants were offered a reasonable explanation for their premature responses, then their verbal report was no different from a Control Condition in which no premature responses were produced. When they did not have a reasonable explanation at hand to explain their premature responses, however, then their verbal report was significantly higher. As a final cautionary note, it is important to point out that we do not mean to imply that our findings indicate that strategy change is always mediated by voluntary processes. There exists good empirical evidence that this is not the case. For example, Doane, Sohn, and Schreiber (1999) as well as Woltz, Gardner, and Bell (2000) observed a rather continuous development of item-general learning that was not verbally expressible and, therefore, was probably based on the operation of implicit, data-driven learning mechanisms solely. Thus, a complete understanding of the causes underlying strategy change will necessarily include an explanation of why sometimes strategy switch appears to be mediated by voluntary processes whereas at other times it might well be an inevitable consequence of task practice. References Anderson, J. R. (1993). Rules of the mind. Hillsdale, NJ: Erlbaum.. Bargh, J. A., Chen, M., & Burrows, L. (1996). Automaticity of social behavior: Direct effects of trait construct and stereotype activation on action. Journal of Personality and Social Psychology, 71, 230–244. Chalmers, D. (1996). The conscious mind. In search of a fundamental theory. New York: Oxford University Press.. Dienes, Z., & Perner, J. (1996). Implicit knowledge in people and connectionist networks. In G. Underwood (Ed.), Implicit cognition (pp. 227–256). Oxford UK: Oxford University Press. Dienes, Z., & Perner, J. (1999). A theory of implicit and explicit knowledge. Behavioral and Brain Sciences, 22, 735–808. 518 H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 Doane, S., Sohn, Y. W., & Schreiber, B. (1999). The role of processing strategies in the acquisition and transfer of a cognitive skill. Journal of Experimental Psychology: Human Perception and Performance, 25, 1390–1410. Farah, M. J. (1994). Visual perception and visual awareness after brain damage: A tutorial overview. In C. Umiltá & M. Moscovitch (Eds.), Attention and performance XV: Conscious and nonconscious information processing (pp. 37–76). MA Cambridge: MIT Press. Frensch, P. A., Haider, H., Rünger, D., Neugebauer, U., Voigt, S., & Werg, J. (2002). Verbal report of incidentally experienced environmental regularity: The route from implicit learning to verbal expression of what has been learned. In L. Jimenéz (Ed.), Attention and implicit learning (pp. 335–366). New York: John Benjamins Publishing Company. Frensch, P. A., & Rünger, D. (2003). Implicit learning. Current Directions in Psychological Science, 12, 13–18. Haider, H., & Frensch, P. A. (1996). The role of information reduction in skill acquisition. Cognitive Psychology, 30, 304–337. Haider, H., & Frensch, P. A. (1999a). Eye movement during skill acquisition: More evidence for the information reduction hypothesis. Journal of Experimental Psychology: Learning, Memory, and Cognition, 25, 172–190. Haider, H., & Frensch, P. A. (1999b). Information reduction during skill acquisition: The influence of task instruction. Journal of Experimental Psychology: Applied., 5, 1–23. Haider, H., & Frensch, P. A. (2002). Why aggregated learning follows the power law of practice when individual learning does not: Comment on Rickard (1997, 1999), Delaney et al. (1998), and Palmeri (1999). Journal of Experimental Psychology: Learning, Memory, and Cognition, 28, 391–406. Haider, H., & Frensch, P. A. (in press). The generation of conscious awareness in an incidental learning situation. Psychological Research. Koriat, A. (2000). The feeling of knowing: Some metatheoretical implications for consciousness and control. Consciousness and Cognition, 9, 149–171. Lachman, R., Lachman, J. L., & Butterfield, E. C. (1979). Cognitive psychology and information processing: An introduction. Hillsdale, NJ: Lawrence Erlbaum Associates.. Lassaline, W., & Logan, G. D. (1993). Memory-based automaticity in the discrimination of visual numerosity. Journal of Experimental Psychology: Learning, Memory, and Cognition, 19, 561–581. Loftus, G. R., & Masson, M. E. (1994). Using confidence intervals in within-participant designs. Psychonomic Bulletin Review, 1, 476–490. Logan, G. D. (1988). Toward an instance theory of automatization. Psychological Review, 95, 492–527. Logan, G. D. (1992). Shapes of reaction-time distributions and shapes of learning curves: A test of the instance theory of automatization. Journal of Experimental Psychology: Learning Memory and Cognition, 18, 883–914. Logan, G. D., & Etherton, J. L. (1994). What is learned during automatization. The role of attention in constructing an instance. Journal of Experimental Psychology: Learning, Memory, and Cognition, 20, 1022–1050. Neumann, O., & Klotz, W. (1994). Motor responses to nonreportable, masked stimuli: Where is the limit of direct parameter specification. In C. Umiltá & M. Moscovitch (Eds.), Attention and performance XV: Conscious and nonconscious information processing (pp. 37–76). Cambridge, MA: MIT Press. Norman, D. (1968). Toward a theory of memory and attention. Psychological Review, 75, 522–536. Nosofsky, R. M., & Palmeri, T. J. (1997). An exemplar-based random walk model of speeded classification. Psychological Review, 104, 266–300. Palmeri, T. J. (1997). Exemplar similarity and the development of automaticity. Journal of Experimental Psychology: Learning, Memory, and Cognition, 23, 324–354. Palmeri, T. J. (1999). Theories of automaticity and the power law of practice. Journal of Experimental Psychology: Learning, Memory, and Cognition, 25, 543–550. Reber, A. S. (1989). Implicit learning and tacit knowledge. Journal of Experimental Psychology: General, 118, 219–235. Reber, A. S. (1993). Implicit learning and tacit knowledge: An essay on the unconscious cognition. New York: Oxford University Press.. Rickard, T. C. (1997). Bending the power law: A CMPL theory of strategy shifts and the automatization of cognitive skills. Journal of Experimental Psychology: General, 126, 288–311. Rickard, T. C. (1999). A CMPL alternative account of practice effects in numerosity judgment tasks. Journal of Experimental Psychology: Learning, Memory, and Cognition, 25, 432–442. H. Haider et al. / Consciousness and Cognition 14 (2005) 495–519 519 Shanks, D. R., & St. John, M. F. (1994). Characteristics of dissociable human learning systems. Behavioral and Brain Sciences, 17, 367–447. Siegler, R. S. (1988). Strategy choice procedure and the development of multiplication skills. Journal of Experimental Psychology: General, 117, 258–275. Siegler, R. S., & Stern, E. (1998). Conscious and unconscious strategy discoveries: A microgenetic analysis. Journal of Experimental Psychology: General, 127, 377–397. Smith, E. E., Langston, C., & Nisbett, R. E. (1992). The case of rules in reasoning. Cognitive Science, 16, 1–40. Whittlesea, B. W. A., & Dorken, M. D. (1993). Incidentally, things in general are particularly determined: An episodicprocessing account of implicit learning. Journal of Experimental Psychology: General, 122, 227–248. Whittlesea, B. W. A., & Dorken, M. D. (1997). Implicit learning: Indirect, not unconscious. Psychonomic Bulletin and Review, 4, 63–67. Whittlesea, B. W. A., & Williams, L. D. (1998). Why do strangers feel familiar, but friends donÕt. A discrepancyattribution account of feelings of familiarity. Acta Psychologica, 98, 141–165. Whittlesea, B. W. A., & Williams, L. D. (2000). The source of feelings of familiarity: The discrepancy-attribution hypothesis. Journal of Experimental Psychology: Learning, Memory, and Cognition, 26, 547–565. Whittlesea, B. W. A., & Williams, L. D. (2001a). The discrepancy-attribution hypothesis I The heuristic basis of feelings and familiarity. Journal of Experimental Psychology: Learning, Memory, and Cognition, 27, 3–13. Whittlesea, B. W. A., & Williams, L. D. (2001b). The discrepancy-attribution hypothesis II Expectation, uncertainty, surprise, and feelings of familiarity. Journal of Experimental Psychology: Learning, Memory, and Cognition, 27, 14–33. Woltz, D. J., Gardner, M. K., & Bell, B. G. (2000). Negative transfer errors in sequential cognitive skills: Strong-butwrong sequence application. Journal of Experimental Psychology: Learning, Memory, and Cognition, 26, 601–625.

Consciousness and Cognition 19 (2010) 702–710 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Illusory own body perceptions: Case reports and relevance for bodily self-consciousness q Lukas Heydrich a,b, Sebastian Dieguez a, Thomas Grunwald c, Margitta Seeck b, Olaf Blanke a,b,* a Laboratory of Cognitive Neuroscience, Brain Mind Institute, Ecole Polytechnique Fédérale de Lausanne (EPFL), Switzerland Department of Neurology, University Hospital Geneva, Switzerland c Swiss Epilepsy Center, Zurich, Switzerland b a r t i c l e i n f o Keywords: Epilepsy Self Body representation Depersonalization Multisensory Parietal cortex Medial prefrontal cortex a b s t r a c t Neurological disorders of body representation have for a long time suggested the importance of multisensory processing of bodily signals for self-consciousness. One such group of disorders – illusory own body perceptions affecting the entire body – has been proposed to be especially relevant in this respect, based on neurological data as well as philosophical considerations. This has recently been tested experimentally in healthy subjects showing that integration of multisensory bodily signals from the entire body with respect to the three aspects: self-location, first-person perspective, and self-location, is crucial for bodily self-consciousness. Here we present clinical and neuroanatomical data of two neurological patients with paroxysmal disorders of full body representation in whom only one of these aspects, self-identification, was abnormal. We distinguish such disorders of global body representation from related but distinct disorders and discuss their relevance for the neurobiology of bodily self-consciousness. Ó 2010 Elsevier Inc. All rights reserved. ‘‘I am what I seem to be, yet do not seem to be what I am; even to myself I am an insoluble riddle, for my personality has been torn apart!”1 E.T.A. Hoffmann, ‘‘The Devil’s Elixirs” 1. Introduction Unraveling the neural basis of self-consciousness is a major research topic in the cognitive neurosciences and science at large. One line of research has focused on the contributions of bodily processing and body representation to selfconsciousness (Blanke & Metzinger, 2009; Craig, 2002; Damasio, 1999; Jeannerod, 2007). Behavioral work in healthy subjects has studied multisensory (Botvinick & Cohen, 1998; Ehrsson, 2007) and sensorimotor (Blakemore & Frith, 2003; Blakemore, Frith, & Wolpert, 1999; Fourneret & Jeannerod, 1998) aspects of self-consciousness by revealing some of the mechanisms of how bodily processing influences self-consciousness. Neuroimaging studies in healthy subjects have q This article is part of a special issue of this journal on Self, Other and Memory. * Corresponding author. Address: Laboratory of Cognitive Neuroscience, Brain Mind Institute, Ecole Polytechnique Fédérale de Lausanne (EPFL), Swiss Federal Institute of Technology, Station 19, 1015 Lausanne, Switzerland. Fax: +41 (0) 21 693 96 25. E-mail address: olaf.blanke@epfl.ch (O. Blanke). 1 ‘‘Ich bin das, was ich scheine, und scheine das nicht, was ich bin, mir selbst ein unerklärlich Rätsel, bin ich entzweit mit meinem Ich!!”. E.T.A. Hoffmann, ‘‘Die Elixiere des Teufels” 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.04.010 L. Heydrich et al. / Consciousness and Cognition 19 (2010) 702–710 703 complemented this work on multisensory and sensorimotor mechanisms by measuring some of the associated neural mechanisms – including interoceptive and cognitive aspects of self-consciousness temporo-parietal junction, insula precuneus, and medial prefrontal cortex (for reviews see Blakemore & Frith, 2003; Craig, 2002; Northoff et al., 2006; Vogeley & Fink, 2003). Historically, however, it was the detailed clinical reports of neurological and psychiatric patients suffering from disorders of body representation, and illusory own body perceptions (Critchley, 1953; Hécaen & Ajuriaguerra, 1952; Ionasescu, 1960; Lhermitte, 1939; Sierra, Lopera, Lambert, Phillips, & David, 2002), which have suggested close links between bodily processing and self-consciousness for a long time. This clinical work highlighted the importance of multisensory bodily processing in patients suffering from migraine (Lippman, 1952; Podoll & Robinson, 1999; Todd, 1955), vascular stroke (Critchley, 1953), tumors (Hécaen & Ajuriaguerra, 1952; Sierra et al., 2002), epilepsy (Blanke, Landis, Spinelli, & Seeck, 2004; Blanke, Ortigue, Coeytaux, Martory, & Landis, 2003), and psychiatric disorders (Lukianowicz, 1963). Such disorders of body representation2 have been classified as a variety of alterations in perceptual bodily experience such as the experiences of the absence of a body part (Critchley, 1953; Frederiks, 1969; Hécaen & Ajuriaguerra, 1952), body part transformations (Critchley, 1953; Hécaen & Ajuriaguerra, 1952; Ionasescu, 1960; Lippman, 1952), body part displacement (Lippman, 1952; Nightingale, 1982), disconnection of one body part from the body (Blanke et al., 2003; Lippman, 1952), the delusional misidentification of one’s own body part (i.e. somatoparaphrenia) (Gerstmann, 1942; Vallar & Ronchi, 2009), as well as phantom limbs (Hécaen & Ajuriaguerra, 1952; Lhermitte, 1939), and supernumerary phantom limbs (Khateb et al., 2009; Vuilleumier, Reverdin & Landis, 1997) (for reviews see (Blanke, Arzy, & Landis, 2008; Brugger, 2006; Dieguez, Staub, & Bogousslavsky, 2007; Frederiks, 1969; Hécaen & Ajuriaguerra, 1952; Menninger-Lerchenthal, 1935). The brain damage leading to these disorders of body representation suggested a predominant involvement of the right posterior parietal cortex and ipsilesional subcortical regions, although other areas were also found to be implicated, i.e. the lateral prefrontal and premotor cortex (Arzy, Overney, Landis, & Blanke, 2006; Berti et al., 2005; Critchley, 1953; Dieguez et al., 2007; Hécaen & Ajuriaguerra, 1952) and the right posterior insular cortex (Baier & Karnath, 2008). Extending earlier accounts (Brugger, 2002, 2006; Devinsky, Feldmann, Burrowes, & Bromfield, 1989; Grüsser & Landis, 1991; Hécaen & Ajuriaguerra, 1952; Menninger-Lerchenthal, 1935; Mizumoto & Ishikawa, 2005), we have recently proposed that a well-defined group of disorders of body representation – illusory own body perceptions affecting the entire body (or primarily the head and trunk region) – are especially relevant for the study of bodily self-consciousness (Blanke, 2004; Blanke & Metzinger, 2009). We highlighted this importance by opposing such illusory own body perceptions with those affecting an isolated extremity or body part and have based this on neurological and neurophysiological data as well as philosophical arguments. Based on clinical data, several research groups have recently developed methods to study the mechanisms of full-body processing and its relationship to self-consciousness experimentally in healthy subjects (Altschuler & Ramachandran, 2007; Ehrsson, 2007; Lenggenhager, Tadi, Metzinger, & Blanke, 2007; Mizumoto & Ishikawa, 2005). These experimental and clinical data jointly suggest that the integration of visual and multisensory bodily signals from the entire body is important for three major aspects of bodily self-consciousness: self-location (SL; the volume in space where humans experience the self to be located [‘‘where I experience to be‘‘]), first-person perspective (1PP; the directedness of conscious experience [‘‘where I experience to perceive the world from”]) and self-identification (SI; the degree to which humans identify with their body [‘‘what I experience as my body”]). SL, 1PP, and SI are abnormal in patients with global illusory own body perceptions (Blanke & Metzinger, 2009). Here we present clinical and anatomical data from two epileptic patients suffering from rare illusory own body perceptions associated with abnormal SI and involving predominantly the trunk and head, due to damage to right posterior dorso-medial parietal cortex (patient 1) and the right dorso-medial prefrontal cortex (patient 2). We discuss the phenomenology, etiology, and lesion location in these two cases in regard to the neurological literature and the larger field of the cognitive neuroscience of bodily self-consciousness. 2. Methods Both patients were recruited at the University Hospital of Geneva, where they underwent full diagnostic workup, including a standardized neurological, psychiatric and neuropsychological examination (Pegna, Qayoom, Gericke, Landis, & Seeck, 1998), electroencephalography (EEG, including source imaging of interictal epileptic spikes, Grave de Peralta Menendez, Gonzalez Andino, Lantz, Michel, & Landis, 2001; Lantz, Grave de Peralta Menendez, Gonzalez Andino, & Michel, 2001; Michel et al., 2004), magnetic resonance imaging (MRI) and positron emission tomography (PET). Patient 2 in addition underwent invasive presurgical evaluations at the Swiss Epilepsy Center in Zurich. Epilepsy surgery of patient 2 was performed at the Department of Neurosurgery at the University Hospital of Zurich. In both patients we conducted a semi-structured interview focusing on aspects of bodily self-consciousness, such as 1PP, SL, SI, as well as somatosensory, visual, auditory, and vestibular symptoms and emotions (following Blanke & Mohr, 2005). Additionally patient 2 was asked to complete the Cambridge Depersonalization Scale (CDS) (Sierra & Berrios, 2000). The CDS comprises 29-items inquiring about subjective experiences classically associated with the depersonalization syndrome. 2 Due to the poor definition of the term body schema (as well as the related term body image) and even more importantly its inconsistent use in the neurological literature, we decided to use the more neutral term body representation (see de Vignemont, 2010). 704 L. Heydrich et al. / Consciousness and Cognition 19 (2010) 702–710 Each item is rated on two Likert scales that quantify frequency (range 0–4) and duration (range 1–6) (cutoff score for depersonalization disorder is 70). 3. Results 3.1. Patient 1 Patient 1 is a 55 year old, left-handed male patient suffering from epilepsy since the age of 14 years. His simple partial sensorimotor seizures affected his left hand had been well controlled under anti-epileptic medication until the onset of paroxysmal episodes of vertigo 9 years before the current hospitalization. At that time he additionally started to experience the following, highly stereotypical pattern of symptoms: without any prior warning he would first have the impression of an increasing pressure in the entire left hemi-body. This sensation increased progressively in strength leading eventually to the sensation that he was invaded by a stranger in his left hemi-body. At this time he also sensed that the left half of his head, the upper part of his left trunk, his left arm and his left leg were no longer belonging to him (no misattribution), that these parts were disconnected from the rest of his body, and that his body was divided into two parts (Fig. 1A). Sometimes this was followed by the impression that the left arm was moving unintentionally and would disappear behind the patient’s back. During these episodes he never experienced any deformation or other changes of his body or the environment. Furthermore, no autoscopic hallucinations, no sensation of floating or disembodiment, no change in visuo-spatial or first-person perspective, no disturbance of language or vision and no loss of contact or consciousness were noted. During these sensations the patient localized the self as within the right side of his body (shown in grey in Fig. 1). He managed to remain calm and was able to continue standing, walking, and even give oral presentations while in front of audiences at work (surrounding persons usually did not notice his seizure manifestations). These simple partial seizures occurred on a daily basis and lasted 1 min. Neurological, psychiatric, and neuropsychological examinations were normal. Surface EEG at the time of hospitalization did not reveal any pathological correlates. MRI showed a hypo-dense lesion in the right posterior intraparietal sulcus predominantly in the medial wall in the T1 weighted sequence (Fig. 2), which was confirmed by fluid-attenuated inverse recovery (FLAIR) and PET compatible with arteriovenous malformation. The diagnosis of epileptic seizures was retained, Fig. 1. Phenomenology patient 1. Simple partial seizures were characterized by the impression of increasing pressure in the entire left hemi-body. This sensation increased progressively in strength leading eventually to the sensation that he was invaded by a stranger in his left hemi-body. At this time he also sensed that the left half of his head, the upper part of his left trunk, his left arm and his left leg were no longer belonging to him, that these parts were disconnected from the rest of his body, and that his body was divided into two parts. L. Heydrich et al. / Consciousness and Cognition 19 (2010) 702–710 705 Fig. 2. MRI patient 1. T1-weighted magnetic resonance imaging reveals a hypointense cortical and subcortical lesion in the right posterior intraparietal sulcus (arrows), consistent with an arteriovenous malformation. (A) Axial view, (B) 3-D reconstruction. The hypointense lesion is indicated by an arrow. even though no epileptic discharges were recorded on repeated EEGs. This was based on the history of the patient (left-lateralized seizures as a child), the current symptomatology, and the corresponding neocortical lesion location. 3.2. Patient 2 Patient 2 is a 30 year old, left-handed male patient suffering from pharmaco-resistant epilepsy. He was referred to the presurgical epilepsy unit at the University Hospital of Geneva and the Epilepsy Clinic in Zurich for further evaluation. Fig. 3. Phenomenology patient 2. Complex partial seizures were characterized by altered bodily awareness including a total loss or a strongly diminished awareness of bodily signals, which predominated at the lower trunk and legs and also included – to a lesser degree – his upper trunk and neck (black; the patient could not indicate whether the upper extremities were affected or not, illustrated by dashed black and grey lines). This was accompanied by the impression that everything below his neck was ‘‘numb”, ‘‘useless” and somewhat inaccessible to conscious awareness while the experience of his head region was experienced as light and as detached from the rest of his lower body (light grey). 706 L. Heydrich et al. / Consciousness and Cognition 19 (2010) 702–710 Seizures started at the age of 11 years, with episodes characterized by the following intense feelings of altered bodily awareness. During these episodes he felt a total loss or a strongly diminished awareness of his body (that he compared to numbness). This sensation was localized and predominated at the legs and lower trunk and also included – to a lesser degree – his upper trunk and neck (the patient could not indicate whether the upper extremities were affected or not (Fig. 3). This altered bodily awareness progressively increased in strength and eventually led to the impression that everything below his neck was ‘‘numb”, ‘‘useless” and somewhat inaccessible to conscious awareness whereas the perception of his head remained unaffected and was even experienced as light and as detached from the rest of his lower body. Following this, he described that he felt ‘‘as an observer of his body and as the one being observed” at the same time. However, he never experienced floating, elevation, or vestibular sensations, nor a change in visuo-spatial or first-person perspective, and he never felt physically disembodied (as is typically mentioned in out-of-body experiences; (Blanke et al., 2004; Brugger, Regard, & Landis, 1997; Devinsky et al., 1989) nor saw an image of himself in external space (autoscopy; i.e. (Blanke et al., 2004; Brugger et al., 1997). The patient also noted a feeling of disconnection from his own thoughts and past, and that he was no longer in control of his actions and speech, like being a robot. In addition the patient often felt as if being detached from the environment, as if sounds and voices were being transformed and distant, perceived as if ‘‘through a veil”. These sensations were mostly observed as isolated auras but also could indicate the imminent onset of a complex partial seizure with involuntary movements of the left arm, head deviation to the left, loss of consciousness, and secondary tonic-clonic generalization. The frequency of these auras was variable, ranging from one per day to one per hour and most often occurred during phases of relaxation. The duration was estimated as 30 s. Different anti-epileptic drugs were tried without any success. Family history and personal history were negative for epilepsy or psychiatric disease. Neurological and psychiatric examinations were unremarkable. The interictal neuropsychological evaluation showed executive and attentional deficits: high distractibility and intrusive thoughts, as well as mild impairments in mental flexibility. The CDS yielded a score of 46. Ictal and interictal high resolution EEG recordings (128 channels) revealed right frontal and fronto-temporal spikes, spike-waves, and sharp waves and subsequent source imaging of epileptic spikes (Grave de Peralta Menendez et al., 2001; Lantz et al., 2001; Michel et al., 2004) suggested an epileptogenic focus in the right middle and superior frontal gyrus (Fig. 4A). Single photon emission tomography (SPECT) demonstrated a hyper-perfusion of the right middle frontal gyrus, the right anterior cingulate cortex, and the bilateral operculum. Repeated MRI did not show any lesion, but intracranial EEG recordings and functional mapping using 52 surgically implanted subdural grid electrodes revealed an epileptogenic focus in the right supplementary motor area (SMA) and the right superior frontal gyrus (Fig. 4B). Subsequent complete resection of the right SMA and partial resection of the right superior frontal gyrus resulted in partial seizure control. There were no more seizures during the day and significantly fewer seizures during sleep. No more auras have been described since this first operation (15 months follow-up). Meanwhile, this first topectomy has been extended during a second epilepsy surgical procedure, after which no seizures have recurred yet (4 month follow-up). Fig. 4. A. Interictal Spike Mapping. Interictal Spike Mapping with 256-channel EEG in patient 2 and source imaging of epileptic spikes suggested an epileptogenic focus in the right middle and superior frontal gyrus. The significant voxels at a p < 0.01 threshold (corrected for multiple comparisons, e.g. for the number of electrodes using Bonferroni correction) are indicated in green. The maximum of the estimated source of the averaged interictal spike is indicated in red (Lantz et al., 2001). B. Intracranial EEG and functional mapping. Intracranial EEG recordings and functional mapping using 52 surgically implanted subdural grid electrodes revealed an epileptogenic focus in the right supplementary motor area (SMA) and the right superior frontal gyrus (these electrodes are indicated in red). Subsequent partial resection of the right superior frontal gyrus and the SMA resulted in partial seizure control (postoperative lesion indicated in red). After a second operation extending the topectomy no seizures have recurred yet. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) L. Heydrich et al. / Consciousness and Cognition 19 (2010) 702–710 707 4. Discussion We present two patients suffering from non-visual illusory own body perceptions due to epilepsy that affected the head-trunk region of the body and were accompanied by a disturbance of SI with the body. Patient 1 suffered from abnormal own body perceptions restricted to his left contralesional hemi-body including the head that was experienced as strange, as if invaded by a stranger, and as if disconnected from his right ipsilesional hemi-face and -body. Patient 2 reported abnormal own body perceptions for the trunk, neck, and the legs bilaterally (but not the head), which were experienced as weaker and numb and as if disconnected from the head. In both patients these illusory own body perceptions were caused by neocortical epilepsy involving two dorso-medial brain regions that have been linked to self-related processing by recent cognitive neuroscience work in healthy subjects. Here, we (1) distinguish such rare illusory own body perceptions (those that affect the head-trunk region) from the more commonly reported illusory own body perceptions that affect the contralesional upper extremity and discuss their underlying neural correlates, (2) propose that they share aspects with certain cases of somatoparaphrenia and visual illusory own body perceptions, (3) discuss their relevance for the study of bodily self-consciousness, and (4) highlight contributions of the dorso-medial posterior parietal cortex and prefrontal cortex for the neurobiology of bodily self-consciousness. 4.1. Illusory own body perceptions: head and trunk versus upper extremity Compared to most patients with disorders of body representation, the present two patients differ concerning the affected body part, the size of the affected body surface and their underlying neural correlates. Disorders of body representation are usually characterized by an alteration in bodily experience that mostly concerns the contralesional upper extremity and are mostly caused by damage to the right ventro-lateral prefrontal cortex, right inferior parietal cortex, or right temporo-parietal cortex. Classically, these are the experience of the absence of a body part (Arzy et al., 2006; Critchley, 1953; Frederiks, 1969; Hécaen & Ajuriaguerra, 1952), of body part transformations (Critchley, 1953; Hécaen & Ajuriaguerra, 1952; Ionasescu, 1960; Lippman, 1952), of body part displacement (Lippman, 1952; Nightingale, 1982), of a disconnection of one body part from the body (Blanke et al., 2003; Lippman, 1952), as well as somatoparaphrenia (Gerstmann, 1942; Vallar & Ronchi, 2009) and supernumerary phantom limbs (Khateb et al., 2009). A discussion of the present two patients with respect to somatoparaphrenia seems particularly relevant and we will focus here on this comparison. Somatoparaphrenia is associated with somatosensory and hemiplegic motor deficits of the contralesional (mostly upper) extremity as well as neglect and results mostly from damage to right temporo-parietal cortex and/or the right insula (Baier & Karnath, 2008). Somatoparaphrenic patients typically misidentify the affected hand as belonging to somebody else or misidentify another person’s hand as belonging to themselves (Baier & Karnath, 2008; Gerstmann, 1942; Vallar & Ronchi, 2009). This condition is often considered an important condition to study bodily self-consciousness, especially concerning self-attribution of the affected extremity. Yet, Blanke and Metzinger (2009) have argued that the aspects of bodily self-consciousness that are crucial for the conscious self (SL, 1PP, SI) are rather based on global body representations in the brain. Yet, as no change in 1PP, SL (the self is experienced as localized within the patient’s body, e.g. head-trunk) and SI (e.g. only the contralesional hand is not attributed to the self, whereas the patients still do self-identify with their body) is noted in somatoparaphrenia. Thus, the study of such patients does not allow to study the conscious self as defined by SL, 1PP, SI. This differed in both present patients who suffered from abnormal self-identification. Patient 1 experienced a strange and disconnected contralesional hemi-body including trunk, neck, head and extremities and patient 2 experienced a disconnection from his neck, trunk and both lower extremities, that were felt as less present and numb. In addition, patient 1 reported that the perceptually altered body segments were experienced as invaded by (the body of) another person. Especially the latter aspect, but also the involvement of the midline body region (that affected the head in patient 1) and the larger extension of the affected body surface, suggests the presence of a disorder of self-identification with the body of the patients at a more global level than somatoparaphrenia. These observations of abnormal self-identification with one’s own head and/or trunk are reminiscent of earlier clinical observations in patients with somatoparaphrenia, in whom not only a limb, but also the contralesional hemi-body was affected (Gerstmann, 1942; Glonning, Jellinger, & Tschabiter, 1963; Hoff & Pötzl, 1935/1988; Lhermitte, 1939; MenningerLerchenthal, 1935; Pötzl, 1925, for reviews see Blanke et al., 2008; Brugger, 2006). Thus, Pötzl (1925; case 1) described a patient with left-sided hemiplegia who not only claimed that his left arm belonged to an unknown person, as seen in many patients with somatoparaphrenia, but also that there was another (unseen) person lying in his bed to his left and that this person tried to push him out of the bed. Engerth and Hoff (1929) described a 71 year-old man with left-sided hypoesthesia, hemianopia (with hemianopic hallucinations), and anososognosia who experienced a constant left-sided person who was most often localized next or behind the patient. Biancone described a 72 year-old female patient with left-sided hemiplegia and hemianesthesia that claimed that her left hemi-body belonged to another person that was lying in her bed (quoted after Lhermitte, 1939). In some rare cases, the body of the felt other is even experienced as invading the patient’s body, as noted by patient 1. Thus, a patient with a parasagittal meningioma adjacent to the right posterior parietal cortex reported by Nightingale (1982) described that his left hemi-body felt strange, seemed to have shifted backwards and was invaded (and controlled) by external agents (mostly the patient’s father). Another patient reported by Gloning (1963, case 1) also reported that her left hemi-body felt strange compared to her right hemi-body and was shifted backwards, due to a glioma in right posterior occipito-parietal cortex. These reports point to the importance of the right posterior and medial parietal 708 L. Heydrich et al. / Consciousness and Cognition 19 (2010) 702–710 cortex for illusory own body perceptions that do not only affect the contralesional hand but also the midline trunk/head region making self-identification and self-other distinction more ambiguous. We also note that perceptual aberrations and abnormal self-identification of these patients and patient 1 are also comparable to some patients suffering from heautoscopy and the feeling of a presence (for review, see Blanke et al., 2008) that may be associated with an experienced split of the trunk and/or head (e.g. Blanke et al., 2008; Brugger, Blanke, Regard, Bradford, & Landis, 2006). Finally there might also be a partial overlap with the concept of ego-pathology as described by Scharfetter (1981), which has mainly been applied to psychiatric patients. 4.2. Self-identification and neural full-body representation Here, we argue that disturbances in self-identification and perceptual aberrations with respect to the head and/or trunk region are observed when head and/or trunk representations in medio-dorsal parietal or prefrontal cortex are affected. This contrasts functionally and anatomically with damage to more ventro-lateral regions in parietal and prefrontal cortex that are more typically lesioned in patients with perceptual aberrations affecting a certain body part (Arzy et al., 2006; Berti et al., 2005; Hécaen & Ajuriaguerra, 1952). This dissociation between head and trunk versus arm and hand is compatible with the major functional properties of single neurons in posterior parietal cortex. Neurons in primary somatosensory cortex (for example area 3b) have small tactile receptive fields that respond preferentially or only to stimuli applied to a specific and contralateral body part (such as the finger tip or the toe (Gardner, 1988). Neurons in area 1 and 2 have larger receptive fields and respond to stimuli applied to one hand or foot and neurons in unimodal somatosensory association cortex (area 5; (Taoka, Toda, Iriki, Tanaka, & Iwamura, 2000) have even larger receptive fields responding to cutaneous stimuli applied to proximal parts of an extremity or the trunk. Yet, neurons with somatosensory receptive fields that encode the hemi-body or the entire body (global somatosensory receptive fields) have also been described in the ventral intraparietal region (VIP). These neurons respond to and encode preferentially the head and trunk region, but may encode also the left or right hemi-body, the upper or lower hemi-body, or the entire body surface (Duhamel, Colby, & Goldberg, 1991, 1998). These VIP neurons have not only large somatosensory receptive fields, but also large visual receptive fields, and contain many bimodal neurons integrating visual and somatosensory stimuli (Bremmer, Klam, Duhamel, Ben Hamed, & Graf, 2002; Duhamel et al., 1998; Schlack, Hoffmann, & Bremmer, 2002). VIP in humans has been suggested to be localized in the intraparietal sulcus (Bremmer et al., 2001) and thus overlaps with the regions affected in patient 1. Cells with similar functional properties are likely to exist in premotor cortex and the SMA (Graziano, 1999; Penfield & Jasper, 1954). We therefore speculate that processing of VIP in patient 1 and premotor cortex and SMA (and adjacent regions with similar functional properties) in patient 2 was abnormal and gave rise to disturbances in self-identification and perceptual aberrations affecting head and/or trunk and thus full body representations. 4.3. Cognitive neuroscience of bodily self-consciousness The data in patient 2 may further be of relevance for the neurobiology of depersonalization, as he also reported a non-perceptual detachment of the self and loss of self-relevance. Patients suffering from depersonalization often feel detached and alienated from their body and/or mental processes, feeling as an outside ‘‘observer”, while having no or less control over their actions (DSM-IV., 2000). Although somewhat reminiscent of autoscopic phenomena, such as out-of-body experience or heautoscopy (Brugger, 2002), there is generally no experienced perceptual change in 1PP or in SL in patients with depersonalization. Interestingly, a recent model regarding the neurobiological correlates of depersonalization suggests an involvement of the prefrontal cortex (Sierra & Berrios, 1998) proposing that increased prefrontal activity may lead to limbic inhibition that results in decreased autonomic output and hypo-emotionality towards the self and the world. Together with neuroimaging data showing the medial prefrontal cortex’s role in cognitive aspects of self-related processing (Northoff et al., 2006), its recruitment during behavioral tasks demanding imagined perspective transformations (Vogeley & Fink, 2003), it may be suggested that activity in the medial prefrontal cortex may reflect more cognitive than spatial aspects of the 1PP, including self reference, self concept and a mental representation of oneself as a subject of experience that may be disturbed in patients with depersonalization. Although bodily self-consciousness and its three components, SI, 1PP, and SL, have been linked to multisensory integration at the temporo-parietal junction, careful study of patients with depersonalization as well as patients with disorders of body representation affecting the head and/or trunk region may provide additional insights into the neurobiology of selfhood. Such self-related brain activity is likely to be distributed in a network of brain regions including the temporoparietal junction, the dorso-medial parietal cortex, the prefrontal cortex and the default network (Ehrsson, Spence, & Passingham, 2004; Gusnard, Akbudak, Shulman, & Raichle, 2001; Hanakawa et al., 2003; Kircher et al., 2000; Macrae, Moran, Heatherton, Banfield, & Kelley, 2004; Maguire et al., 1998; Northoff et al., 2006; Ruby & Decety, 2001; Vogeley & Fink, 2003). Acknowledgments LH and MS are supported by the Swiss National Science Foundation (Grants 33CM30-124089, 33CM30-123115, 320030122073, 323530-123718). OB was supported by the Swiss National Science Foundation (Sinergia Grant CRSII1-125135: Balancing Self and Body). L. Heydrich et al. / Consciousness and Cognition 19 (2010) 702–710 709 References Altschuler, E. L., & Ramachandran, V. S. (2007). A simple method to stand outside oneself. Perception, 36(4), 632–634. Arzy, S., Overney, L. S., Landis, T., & Blanke, O. (2006). Neural mechanisms of embodiment: Asomatognosia due to premotor cortex damage. Archieves of Neurology, 63(7), 1022–1025. Baier, B., & Karnath, H. O. (2008). Tight link between our sense of limb ownership and self-awareness of actions. Stroke, 39(2), 486–488. Berti, A., Bottini, G., Gandola, M., Pia, L., Smania, N., Stracciari, A., et al (2005). Shared cortical anatomy for motor awareness and motor control. Science, 309(5733), 488–491. Blakemore, S. J., & Frith, C. (2003). Self-awareness and action. Current Opinion in Neurobiology, 13(2), 219–224. Blakemore, S. J., Frith, C. D., & Wolpert, D. M. (1999). Spatio-temporal prediction modulates the perception of self-produced stimuli. Journal of Cognitive Neuroscience, 11(5), 551–559. Blanke, O. (2004). Out of body experiences and their neural basis. BMJ, 329(7480), 1414–1415. Blanke, O., Arzy, S., & Landis, T. (2008). Chapter 22 Illusory reduplications of the human body and self. Handbook of Clinical Neurology, 88, 429–458. Blanke, O., Landis, T., Spinelli, L., & Seeck, M. (2004). Out-of-body experience and autoscopy of neurological origin. Brain, 127(Pt 2), 243–258. Blanke, O., & Metzinger, T. (2009). Full-body illusions and minimal phenomenal selfhood. Trends in Cognitive Sciences, 13(1), 7–13. Blanke, O., & Mohr, C. (2005). Out-of-body experience, heautoscopy, and autoscopic hallucination of neurological origin implications for neurocognitive mechanisms of corporeal awareness and self-consciousness. Brain Research Brain Research Reviews, 50(1), 184–199. Blanke, O., Ortigue, S., Coeytaux, A., Martory, M. D., & Landis, T. (2003). Hearing of a presence. Neurocase, 9(4), 329–339. Botvinick, M., & Cohen, J. (1998). Rubber hands ’feel’ touch that eyes see. Nature, 391(6669), 756. Bremmer, F., Klam, F., Duhamel, J. R., Ben Hamed, S., & Graf, W. (2002). Visual-vestibular interactive responses in the macaque ventral intraparietal area (VIP). European Journal of Neuroscience, 16(8), 1569–1586. Bremmer, F., Schlack, A., Shah, N. J., Zafiris, O., Kubischik, M., Hoffmann, K., et al (2001). Polymodal motion processing in posterior parietal and premotor cortex: A human fMRI study strongly implies equivalencies between humans and monkeys. Neuron, 29(1), 287–296. Brugger, P. (2002). Reflective mirrors: Perspective-taking in autoscopic phenomena. Cognitive Neuropsychiatry, 7(3), 179–194. Brugger, P. (2006). From phantom limbs to phantom bodies: Varieties of extracorporeal awareness. In G. Knoblich, I. M. Thornton, M. Grosjean, & M. Shiffrar (Eds.), Human body perception from the inside out (pp. 171). Oxford: Oxford University Press. Brugger, P., Blanke, O., Regard, M., Bradford, D. T., & Landis, T. (2006). Polyopic heautoscopy: Case report and review of the literature. Cortex, 42(5), 666–674. Brugger, P., Regard, M., & Landis, T. (1997). Illusory reduplication of own’s own body: Phenomenology and classification of autoscopic phenomena. Cognitive Neuropsychiatry(2), 19–38. Craig, A. D. (2002). How do you feel–now? The anterior insula and human awareness. Nature Reviews Neuroscience, 10(1), 59–70. Critchley, M. (1953). The parietal lobes. New York, London: Hafner Publishing Company. Damasio, A. (1999). The feeling of what happens. London: Random House. de Vignemont, F. (2010). Body schema and body image–pros and cons. Neuropsychologia, 48(3), 669–680. Devinsky, O., Feldmann, E., Burrowes, K., & Bromfield, E. (1989). Autoscopic phenomena with seizures. Archieves of Neurology, 46(10), 1080–1088. Dieguez, S., Staub, F., & Bogousslavsky, J. (2007). Asomatognosia. In O. Godefroy & J. Bogousslavsky (Eds.), The behavioral and cognitive neurology of stroke (pp. 215–253). Cambridge: Cambridge University Press. DSM-IV (2000). Diagnostic and statistical manual of mental disorders–DSM-IV (4th ed.). Washington, DC: American Psychiatric Association. Duhamel, J. R., Colby, C. L., & Goldberg, M. E. (1991). Congruent representations of monkey ventral intra-parietal cortex (area VIP). In J. Paillard (Ed.), Brain and space. Oxford University Press. Duhamel, J. R., Colby, C. L., & Goldberg, M. E. (1998). Ventral intraparietal area of the macaque: Congruent visual and somatic response properties. Journal of Neurophysiology, 79(1), 126–136. Ehrsson, H. H. (2007). The experimental induction of out-of-body experiences. Science, 317(5841), 1048. Ehrsson, H. H., Spence, C., & Passingham, R. E. (2004). That’s my hand! Activity in premotor cortex reflects feeling of ownership of a limb. Science, 305(5685), 875–877. Engerth, G., & Hoff, H. (1929). Ein Fall von Halluzinationen im hemianoptischen Gesichtsfeld. Beitrag zur Genese der optischen Halluzinationen. Monatschr Psychiatr Neurol, 74, 246–256. Fourneret, P., & Jeannerod, M. (1998). Limited conscious monitoring of motor performance in normal subjects. Neuropsychologia, 36(11), 1133–1140. Frederiks, J. A. M. (1969). Disorders of the body schema. In G. W. Bruyn & P. J. Vinken (Eds.). Handbook of clinical neurology (Vol. 4, pp. 207–240). . Gardner, E. P. (1988). Somatosensory cortical mechanisms of feature detection in tactile and kinesthetic discrimination. Canadian Journal of Physiology and Pharmacology, 66(4), 439–454. Gerstmann, J. (1942). Problem of imperception of disease and of impaired body territories with organic lesions. Relation to body schema and its disorders. Archives of Neurology and Psychiatry, 48, 890–913. Glonning, I., Jellinger, K., & Tschabiter, H. (1963). Über einen obduzierten Fall von optischer Körperschemastörung und Heautoskopie. Neuropsychologia, 1, 217–231. Grave de Peralta Menendez, R., Gonzalez Andino, S., Lantz, G., Michel, C. M., & Landis, T. (2001). Noninvasive localization of electromagnetic epileptic activity. I. Method descriptions and simulations. Brain Topography, 14(2), 131–137. Graziano, M. S. (1999). Where is my arm? The relative role of vision and proprioception in the neuronal representation of limb position. Proceedings of the National Academy Sciences of the United States of America, 96(18), 10418–10421. Grüsser, O.-J., & Landis, T. (1991). The splitting of ‘I’ and ‘me’: Heautoscopy and related phenomena. In T. Grüsser & O.-J. Landis (Eds.), Visual agnosias and other disturbances of visual perception and cognition (pp. 297–303). Amsterdam: MacMillan. Gusnard, D. A., Akbudak, E., Shulman, G. L., & Raichle, M. E. (2001). Medial prefrontal cortex and self-referential mental activity: Relation to a default mode of brain function. Proceeding of the National Academy of Sciences of the United States of America, 98(7), 4259–4264. Hanakawa, T., Immisch, I., Toma, K., Dimyan, M. A., Van Gelderen, P., & Hallett, M. (2003). Functional properties of brain areas associated with motor execution and imagery. Journal of Neurophysiology, 89(2), 989–1002. Hécaen, H., & Ajuriaguerra, J. (1952). Méconnaissances et hallucinations corporelles. Paris: Masson. Hoff, H., & Pötzl, O. (1935/1988). Transformation between body image and external worl. In J. Brown (Ed.), Agnosia and apraxia: Selected papers of liepmann, lange, and pötzl (pp. 251–262). Hillsdale, NJ: Lawrence Erlbaum. Ionasescu, V. (1960). Paroxysmal disorders of the body image in temporal lobe epilepsy. Acta Psychiatrica Scandinavica, 35, 171–181. Jeannerod, M. (2007). Being oneself. Journal of Physiology Paris, 101(4–6), 161–168. Khateb, A., Simon, S. R., Dieguez, S., Lazeyras, F., Momjian-Mayor, I., Blanke, O., et al (2009). Seeing the phantom: A functional magnetic resonance imaging study of a supernumerary phantom limb. Annals of Neurology, 65(6), 698–705. Kircher, T. T., Senior, C., Phillips, M. L., Benson, P. J., Bullmore, E. T., Brammer, M., et al (2000). Towards a functional neuroanatomy of self processing: Effects of faces and words. Brain Research Cognitive Brain Research, 10(1–2), 133–144. Lantz, G., Grave de Peralta Menendez, R., Gonzalez Andino, S., & Michel, C. M. (2001). Noninvasive localization of electromagnetic epileptic activity. II. Demonstration of sublobar accuracy in patients with simultaneous surface and depth recordings. Brain Topography, 14(2), 139–147. Lenggenhager, B., Tadi, T., Metzinger, T., & Blanke, O. (2007). Video ergo sum: Manipulating bodily self-consciousness. Science, 317(5841), 1096–1099. Lhermitte, J. (1939). L’image de notre corps (Editions de la Nouvelle Revue Critque ed.). Paris. Lippman, C. W. (1952). Certain hallucinations peculiar to migraine. Journal of Nervous and Mental Disease, 116(4), 346–351. Lukianowicz, N. (1963). ‘‘Body image‘‘ disturbances in psychiatric disorders. British Journal of Psychiatry, 113(494), 31–47. 710 L. Heydrich et al. / Consciousness and Cognition 19 (2010) 702–710 Macrae, C. N., Moran, J. M., Heatherton, T. F., Banfield, J. F., & Kelley, W. M. (2004). Medial prefrontal activity predicts memory for self. Cerebral Cortex, 14(6), 647–654. Maguire, E. A., Burgess, N., Donnett, J. G., Frackowiak, R. S., Frith, C. D., & O’Keefe, J. (1998). Knowing where and getting there: A human navigation network. Science, 280(5365), 921–924. Menninger-Lerchenthal, E. (1935). Das Truggebilde der eigenen Gestalt. Berlin: S. Karger. Michel, C. M., Murray, M. M., Lantz, G., Gonzalez, S., Spinelli, L., & Grave de Peralta, R. (2004). EEG source imaging. Clinical Neurophysiology, 115(10), 2195–2222. Mizumoto, M., & Ishikawa, M. (2005). Immunity to error through misidentification and the bodily illusion experiment. Journal of Consciousness Studies, 12(7), 3–19. Nightingale, S. (1982). Somatoparaphrenia: A case report. Cortex, 18(3), 463–467. Northoff, G., Heinzel, A., de Greck, M., Bermpohl, F., Dobrowolny, H., & Panksepp, J. (2006). Self-referential processing in our brain – A meta-analysis of imaging studies on the self. Neuroimage, 31(1), 440–457. Pegna, A. J., Qayoom, Z., Gericke, C. A., Landis, T., & Seeck, M. (1998). Comprehensive postictal neuropsychology improves focus localization in epilepsy. European Neurology, 40(4), 207–211. Penfield, W., & Jasper, H. (1954). Epilepsy and the functional anatomy of the human brain. Boston: Little, Brown and Company. Podoll, K., & Robinson, D. (1999). Lewis Carroll’s migraine experiences. Lancet, 353(9161), 1366. Pötzl, O. (1925). Über Störungen der Selbstwahrnehmung bei linksseitiger Hemiplegie. Zeitschrift für die gesamte Neurologie und Psychiatrie, 93, 117–168. Ruby, P., & Decety, J. (2001). Effect of subjective perspective taking during simulation of action: A PET investigation of agency. Nature Neuroscience, 4(5), 546–550. Scharfetter, C. (1981). Ego-psychopathology: The concept and its empirical evaluation. Psychological Medicine, 11(2), 273–280. Schlack, A., Hoffmann, K. P., & Bremmer, F. (2002). Interaction of linear vestibular and visual stimulation in the macaque ventral intraparietal area (VIP). European Journal of Neuroscience, 16(10), 1877–1886. Sierra, M., & Berrios, G. E. (1998). Depersonalization: Neurobiological perspectives. Biological Psychiatry, 44(9), 898–908. Sierra, M., & Berrios, G. E. (2000). The Cambridge Depersonalization Scale: A new instrument for the measurement of depersonalization. Psychiatry Research, 93(2), 153–164. Sierra, M., Lopera, F., Lambert, M. V., Phillips, M. L., & David, A. S. (2002). Separating depersonalisation and derealisation: The relevance of the ‘‘lesion method‘‘. Journal of Neurology Neurosurgery and Psychiatry, 72(4), 530–532. Taoka, M., Toda, T., Iriki, A., Tanaka, M., & Iwamura, Y. (2000). Bilateral receptive field neurons in the hindlimb region of the postcentral somatosensory cortex in awake macaque monkeys. Experimental Brain Research, 134(2), 139–146. Todd, J. (1955). The syndrome of Alice in Wonderland. Canadian Medical Association Journal, 73(9), 701–704. Vallar, G., & Ronchi, R. (2009). Somatoparaphrenia: A body delusion. A review of the neuropsychological literature. Experimental Brain Research, 192(3), 533–551. Vogeley, K., & Fink, G. R. (2003). Neural correlates of the first-person-perspective. Trends in Cognitive Sciences, 7(1), 38–42. Vuilleumier, P., Reverdin, A., & Landis, T. (1997). Four legs. Illusory reduplication of the lower limbs after bilateral parietal lobe damage. Archives of Neurology, 54(12), 1543–1547.

Consciousness and Cognition 22 (2013) 677–683 Contents lists available at SciVerse ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Short Communication Episodic future thought: Contributions from working memory Paul F. Hill ⇑, Lisa J. Emery Department of Psychology, Appalachian State University, 222 Joyce Lawrence Ln, Boone, NC 28608, USA a r t i c l e i n f o Article history: Received 18 September 2012 Keywords: Episodic future thought Working memory Episodic buffer Event specificity a b s t r a c t The ability to imagine hypothetical events in one’s personal future is thought to involve a number of constituent cognitive processes. We investigated the extent to which individual differences in working memory capacity contribute to facets of episodic future thought. College students completed simple and complex measures of working memory and were cued to recall autobiographical memories and imagine future autobiographical events consisting of varying levels of specificity (i.e., ranging from generic to increasingly specific and detailed events). Consistent with previous findings, future thought was related to analogous measures of autobiographical memory, likely reflecting overlapping cognitive factors supporting both past and future thought. Additionally, after controlling for autobiographical memory, residual working memory variance independently predicted future episodic specificity. We suggest that when imagining future events, working memory contributes to the construction of a single, coherent, future event depiction, but not to the retrieval or elaboration of event details. Ó 2013 Elsevier Inc. All rights reserved. 1. Introduction Recent years have seen a growing theoretical and empirical interest in the ability to mentally simulate imagined future events, or episodic future thought (for reviews, see Schacter, Addis, Hassabis, Martin, Spreng, & Szpunar, 2012; Szpunar, 2010). Much like episodic memory, episodic future thought is believed to be a constructive process in which disparate pieces of information are retrieved from long-term memory and flexibly recombined into a coherent imagined event (Schacter & Addis, 2007a). Episodic memory and future thought are therefore often thought to serve as complementary functions at opposite ends of one’s temporal narrative and are together viewed as comprising the broader faculty of mental time travel (Suddendorf & Corballis, 2007). Converging evidence from a variety of research domains lends support for this view. For example, a number of clinical populations associated with impaired memory function have been observed to exhibit corresponding deficits in the ability to imagine the future, including amnesic (Hassabis, Kumaran, Vann, & Maguire, 2007; Klein, Loftus, & Kihlstrom, 2002; Tulving, 1985), suicidally depressed (Williams et al., 1996), autistic (Lind & Bowler, 2010), and schizophrenic populations (D’Argembeau, Raffard, & Van der Linden, 2008). Likewise, lifespan developmental patterns of memory functioning parallel the emergence (Busby & Suddendorf, 2005; Suddendorf & Busby, 2005) and decline (Addis, Wong, & Schacter, 2008) of future thought abilities. The phenomenological characteristics of past and future thought have also been reported to be constrained by similar factors, with positive and temporally close events being more richly recalled or imagined (D’Argembeau and Van der Linden, 2004). Furthermore, a number of brain imaging studies have demonstrated striking functional overlap in the neural regions underlying episodic memory and future thought, including prefrontal, medial temporal, and parietal cortices (Addis, Wong, & Schacter, 2007; Okuda et al., 2003; Spreng, Mar, & Kim, 2009; Szpunar, Watson, & McDermott, 2007). ⇑ Corresponding author. Present address: Virginia Tech Carilion Research Institute, 2 Riverside Circle, Roanoke, VA 24016, USA. E-mail address: pfhill@vt.edu (P.F. Hill). 1053-8100/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.concog.2013.04.002 678 P.F. Hill, L.J. Emery / Consciousness and Cognition 22 (2013) 677–683 Although there are considerable similarities between autobiographical recall and future event construction, there are also some important differences. Unlike past episodic recall which requires reconstructing elements of a previously experienced event, future thought depends on a novel recombination of episodic details into a hypothetical event (Addis & Schacter, 2012; Addis et al., 2007). This requirement for novel recombination suggests that future event construction may involve additional cognitive and neural processes that are not as involved in autobiographical memory reconstruction. For example, an early study by Addis et al. (2007) found that some brain regions were more active during future than past event construction. In particular, regions of the prefrontal cortex, including right frontopolar and left ventrolateral areas, and the right hippocampus were more active for future than past events during the event construction phase, when participants were forming a single plausible event (e.g., ‘‘Two weeks from now I will be standing in front of a classroom teaching the first class of the semester’’). During the event elaboration phase, when participants were asked to elaborate on the event by providing as many episodic details as possible, there was considerably more overlap between the regions activated for past and future events. The additional involvement of both prefrontal and hippocampal regions in the construction of future events suggest that a combination of executive and memory binding functions may contribute to this novel constructive process. If this is the case, it points to a potential role for the working memory (WM) system in imagining future episodes, above and beyond the contributions provided by access to the autobiographical memory base. The WM system may be broadly conceptualized as a workspace where information is temporarily held in an active state, such that it can be manipulated, integrated, or accessed in service of higher order cognitive thought. Although there are many theoretical conceptualizations of the WM system (e.g., Baddeley, 2000; Cowan, 1999; Engle, Tuholski, Laughlin, & Conway, 1999; Unsworth & Engle, 2007), each share the notion that maintaining information in an active state requires contributions from both long-term memory and executive processes, the same two processes that have been implicated in constructing future events. Perhaps the most well-known conceptualization of the WM system is Baddeley & Hitch’s (1974) multiple component model and its more recent modifications (Baddeley, 2000; Baddeley, Allen, & Hitch, 2011). This model posits two domain-specific temporary stores, the visuo-spatial sketchpad and the phonological loop, that hold information taken in from the environment or retrieved from long-term memory. These stores have been joined in recent years by the episodic buffer, which is thought to store bound multimodal constructions kept in consciousness through interaction with the central executive. The episodic buffer component of WM has previously been suggested to act as the ‘‘stage’’ on which episodic details are recombined into a hypothetical scenario (Schacter & Addis, 2007b; Suddendorf & Corballis, 2007). This is consistent with the idea that imagining the future involves combining information from multiple modalities, including past experiences, visualization processes (Szpunar, Chan, & McDermott, 2009; Szpunar et al., 2007), and semantic personal information (D’Argembeau & Mathy, 2011; Viard et al., 2011). In addition, the executive function (EF) subcomponent of the WM system has been linked to a number of other future oriented behaviors, such as prospective mind wandering (Baird, Smallwood, & Schooler, 2011; Smallwood, Nind, & O’Connor, 2009), autobiographical planning (Spreng, Stevens, Chamberlain, Gilmore, & Schacter, 2010) and future oriented decision making (Bickel, Yi, Landes, Hill, & Baxter, 2011). It might therefore be hypothesized that the WM system should be particularly involved with the construction of novel future events. This hypothesis, however, has not previously been directly tested. Although not designed to look specifically at the relationship between WM and episodic future thought, two previous behavioral studies provide some support for this hypothesis. In one study (D’Argembeau, Ortoleva, Jumentier, & Van der Linden, 2010), young adults were administered a variety of neuropsychological tests, including one test of WM (Letter-Number Sequencing) and several tests of one subtype of EF, fluency (phonemic, semantic, and figural fluency). Participants’ performance on a composite of the WM and EF measures was then used to predict performance on three different components of future event construction and elaboration: fluency (e.g., how many unique future events can a person generate in 1 min), specificity (e.g., whether a person can construct a unique potential future event, similar to the event construction phase in Addis et al. (2007)), and details (e.g., how many specific episodic details can a person produce about a possible future event, similar to the event elaboration phase in Addis et al. (2007)). The composite WM and EF factor was a consistent predictor of episodic future thought, particularly for specificity and details. Because the composite factor, however, was primarily composed of fluency tasks, it is difficult to determine whether WM capacity itself was involved in episodic future thought, or whether the correlations were driven primarily by performance on the fluency tasks. In a second, smaller study of 17 older adults (Addis et al., 2008), measures of WM (Digit Span Backwards), relational memory (Verbal Paired Associates), phonemic fluency, and EF (Wisconsin Card Sort) were used to predict the number of episodic and semantic details produced during a task combining the event construction and elaboration phases. The number of episodic details produced for both past and future events was significantly correlated with both Verbal Paired Associates and Digit Span Backwards. Because the outcome measure in this study did not distinguish between different phases of future thought, however, it is unclear whether WM was related to the initial construction or subsequent elaboration of an event. Moreover, because of the small sample size and the fact that the WM/Future Thought correlation was only examined in older adults, the generalizability of this result is not clear. Finally, in both the D’Argembeau et al. (2010) and Addis et al. (2008) studies, the measures involving WM appear to be slightly higher for the future measures than for the past, though this was not directly tested in either study. If constructing a novel future event requires greater WM resources than does re-constructing an experienced past event, then WM capacity should be a significant predictor of future event construction even after controlling for past event construction. P.F. Hill, L.J. Emery / Consciousness and Cognition 22 (2013) 677–683 679 The primary aim of the current study, therefore, was to more specifically examine the role of WM capacity in the ability to construct novel representations of possible future autobiographical events. To this end, we administered multiple tests of WM capacity and episodic future thought to a group of young adults, modeling our methods on those of D’Argembeau et al. (2010) in order to distinguish between future event construction and elaboration. Because of the well-established and relatively strong relationship between autobiographical memory and episodic future thought, our analyses were designed to determine whether WM capacity predicted future thought over and above the past-future thought relationship. That is, we directly tested whether WM was predictive of future thought construction more so than past construction, as appeared to be the case in prior studies. 2. Methods 2.1. Participants Participants were 109 college students (71 women, ages 17–30) enrolled in introductory and intermediate psychology classes at Appalachian State University. 2.2. Materials 2.2.1. Working memory Participants completed four automated measures of verbal and visuo-spatial WM. Two of the tasks were measures of ‘‘simple’’ WM, requiring storage and maintenance of verbal and visuo-spatial information (the Letter and Matrix Span tasks, respectively; Kane et al., 2004). The other two ‘‘complex’’ tasks additionally required performance of an unrelated secondary task (the Operation and Symmetry Span tasks; Kane et al., 2004; Unsworth, Heitz, Schrock, & Engle, 2005). These types of tasks are frequently used in studies predicting individual differences in higher level cognitive functioning from WM capacity. 2.2.2. Autobiographical tasks Participants completed a sequence of past and future autobiographical measures taken from D’Argembeau et al. (2010). 2.2.2.1. Fluency task. During the autobiographical fluency task (MacLeod & Byrne, 1996) participants were instructed to think about two time periods from their past (within the past year and within the past 5–10 years) and two time periods in their future (within the next year and within the next 5–10 years). For each prompt, participants were given 60 s to list as many generic events as possible that occurred or may occur within each specified time period. Instructions indicated that the events could be trivial or important. Participants were also instructed that responses were not required to be detailed. The order of cue presentation was counterbalanced across temporal direction (past or future), with the 1-year time period presented first, regardless of temporal direction. Scores were based on the total number of unique items generated for each time period. Following D’Argembeau et al. (2010), the two time periods from each temporal condition were combined to yield a total past and total future fluency score. 2.2.2.2. Specificity task. The episodic specificity task (Williams et al., 1996) assessed the ability to generate specific autobiographical past and future events. Participants were cued to recall specific events occurring in their personal past and specific events that might reasonably occur in their personal future. Participants were instructed to identify unique events occurring at a specific place and time, lasting a few minutes to hours, but not more than 1 day. Two sets of five cue words were used, matched for imageablitity, frequency of use, and word length (Tse & Altarriba, 2007), and counterbalanced across past and future conditions. The order in which the participants completed the two temporal conditions was also counterbalanced. Scoring was based on criteria used in previous studies (e.g., D’Argembeau et al., 2010; Williams et al., 1996). Specific responses were those occurring at a distinct place and time and lasting less than 1 day (e.g., ‘‘My parents will drive to campus this June to attend my graduation ceremony. They will cheer as I walk across the stage and afterwards we will go out to eat’’). The total number of specific events generated for each temporal condition was computed, yielding a past and future specificity score. The highest possible score for either temporal condition was 5. 2.2.2.3. Episodic details task. The episodic details task (D’Argembeau et al., 2010), adapted from the scene construction task used by Hassabis et al. (2007), assessed the number of unique details participants were able to generate when recalling or imagining specific episodic events. Participants were cued to think about a past or future episodic event and were then instructed to elaborate on the details of that event, providing as many sensory and introspective details as possible. Instructions specified that the events should be unique, occurring at a specific place and time, lasting a few minutes or hours, but less than 1 day. Future events were additionally required to be both plausible and novel. The past and future event cues were adopted from D’Argembeau et al. (2010) and the assignment of each cue to a particular temporal direction was counterbalanced across participants (e.g., ‘‘recall the last time you met a friend; imagine something you will do on your next vacation’’ or ‘‘recall something you did on your last vacation; imagine the next time you will meet a friend’’). 680 P.F. Hill, L.J. Emery / Consciousness and Cognition 22 (2013) 677–683 Scoring for the Details task was based on D’Argembeau et al. (2010). Each event description was broken into a set of statements, roughly constituting a single idea unit. These statements were then classified into six categories: spatial reference, entity presence, sensory description, thoughts and emotions, actions, or temporal reference. Spatial references included aspects related to the position or direction of entities relative to the participant’s vantage point (e.g., ‘‘she sat to my left’’), references to explicit measurements (e.g., ‘‘about 50 feet from the water’’), as well as distinct locations (e.g., ‘‘Florida’’). Any reference to a distinct, concrete object was categorized as an entity (objects, people, animals, etc.). The sensory description category consisted of references to the qualitative properties of entities or of the environment (e.g., ‘‘the water is dark blue,’’ ‘‘it is very humid’’). The thoughts and emotions category included introspective thoughts and emotions of the participant or any other entities in the scene. Likewise, actions constituted any action initiated by an entity within the scene. The temporal category was adopted by D’Argembeau et al. (2010) and included references to the temporal context of an event (e.g., ‘‘last winter’’) as well as time measurements (e.g., ‘‘four hours later’’). All unique details that were relevant to an event narrative meeting the specificity requirement were counted towards that event’s total content score. Events not meeting the specificity requirement were not scored. 2.3. Procedure All measures were administered electronically via a desktop computer using E-Prime v.2.0 software (Schneider, Eschman, & Zuccolotto, 2002) and were completed in the following order: simple WM, autobiographical fluency, complex WM, episodic specificity, complex WM, episodic details. The order in which the verbal and visuo-spatial WM tasks were administered was counterbalanced across participants, with both simple WM tasks administered first and the complex WM tasks interposed between the autobiographical measures. Following D’Argembeau et al. (2010), the autobiographical tasks were presented in a fixed order due to the increasingly explicit instructions for generating episodic events, with the order in which the past and future conditions were administered counterbalanced across participants. Participants typed their responses directly into the computer using the keyboard. 3. Results Each of the autobiographical measures was scored by three trained independent raters. Intraclass correlation coefficients (ICC) indicated that there was strong agreement between the raters with respect to the number of autobiographical events generated in the fluency task (past ICC = .98, future ICC = .99), the number of specific events produced in the episodic specificity task (past ICC = .77, future ICC = .79), and the total content scores in the episodic details task (past ICC = .98, future ICC = .98). Descriptive statistics for the WM measures were M = 55.04 (SD = 15.23) for Operation Span, M = 26.39 (SD = 8.00) for Symmetry Span, M = 38.85 (SD = 10.78) for Letter Span, and M = 76.61 (SD = 10.24) for Matrix Span. A principle components analysis to determine whether the WM measures could be reduced into component constructs yielded only one factor with an eigenvalue greater than 1.0 and accounting for 52% of the total variance, which we characterized as a general measure of WM capacity. Varimax rotation produced factor loadings ranging from .52 to .81. The resulting factor score was used as a composite measure of WM capacity in all subsequent analyses.1 Bivariate correlations and descriptive statistics for the past and future autobiographical measures and WM composite are shown in Table 1. Scores for autobiographical fluency and episodic specificity were higher for past events than for future events [tfluency (108) = 2.12, p = .036; tspecificity (108) = 6.40, p < .001] which is consistent with previous research (Addis et al., 2008; D’Argembeau et al., 2010). The total number of episodic details, however, did not significantly differ between past and future events, tdetails (103) = 1.53, p = .129. Each measure of future thought was also correlated with its past counterpart. Among the autobiographical measures, past and future fluency exhibited the most striking relationship, as 64% of the variance in future fluency performance could be accounted for by past fluency. Conversely, a much smaller percentage of shared variance was observed between past and future measures of episodic specificity (18%) and details (25%). Composite WM capacity was significantly correlated with future specificity but none of the other autobiographical tasks. Hierarchical multiple regression analyses were used to test the independent contributions of WM to future thought, over and above autobiographical memory. For each future measure, the corresponding autobiographical memory task was entered as a predictor of future thought in Step 1; composite WM was entered in Step 2 to determine any unique shared variance not accounted for by the corresponding autobiographical memory task. As can be seen in Table 2, each of the past autobiographical tasks was a significant predictor of its corresponding measure of future thought. Furthermore, when controlling for past specificity, residual WM capacity independently predicted an additional 5% of the variance in performance on the future specificity task. WM did not significantly predict future fluency or future details. Moreover, when reversing 1 We had initially chosen to measure both simple and complex, verbal and visuospatial WM in order to test hypotheses about the potential relative contribution of verbal vs. visuospatial processes, and controlled attention vs. storage processes, to different components of future thought. Because, however, a principle components analysis of the WM measures failed to yield evidence of these factors, we chose to combine the WM measures into one composite. This finding is supportive of conceptualizations of the WM system emphasizing the unitary nature of storage (e.g., Cowan, 1999) and contributions of long-term memory and controlled attention processes to both simple and complex WM tasks (e.g., Unsworth & Engle, 2007). 681 P.F. Hill, L.J. Emery / Consciousness and Cognition 22 (2013) 677–683 Table 1 Bivariate correlations and descriptive statistics for autobiographical measures and composite working memory scores. Measure 1 1. Past fluency 2. Future fluency 3. Past specificity 4. Future specificity 5. Past details 6. Future details 7. WM composite – .80** .21* .18 .36** .31** .06 2 – .11 .16 .26** .24* .15 3 – .42** .25* .21* .01 4 5 – .18 .24* .22* 6 – .50** .19 – .09 7 M SD – 16.61 15.79 2.22 1.20 17.17 15.61 – 5.73 6.76 1.57 1.54 11.83 11.62 – Note: WM = Working Memory. p < .05. p < .01. * ** Table 2 R2 statistics from multiple regression analyses. Dependent variable Future fluency Future fluency Future specificity Future specificity Future details Future details Predictor(s) Past fluency Past fluency, WM Past specificity Past specificity, WM Past details Past details, WM R2 DR2 F(DR2) df .64 .65 .18 .23 .25 .25 ** 176.30 2.83 21.18 6.00 31.66 .01 1, 98 1, 97 1, 98 1, 97 1, 97 1, 96 .64 .01 .18** .05* .25** .00 Note: WM = Working Memory Composite. p < .05. ** p < .01. * these analyses, composite WM did not significantly predict performance on any of the past measures independently of future thought performance (p P .10). 4. Discussion We took an individual differences approach to investigate the role WM plays in the ability to construct future episodic events. Participants completed measures of episodic memory and future thought as well as verbal and visuo-spatial WM tasks. Using a composite score of WM capacity, we were able to examine the extent to which residual WM variance contributed to future thought while controlling for autobiographical memory. Our results indicated that: (a) the ability to imagine personally relevant events in the future was strongly related to autobiographical memory; and (b) even when controlling for autobiographical memory, WM capacity independently predicted the ability to construct specific future events. Consistent with previous research indicating functional overlap between episodic memory and future thought (Schacter et al., 2012; Suddendorf & Corballis, 2007), each of the three measures of autobiographical memory was significantly correlated with its corresponding measure of future thought. The fluency tasks exhibited the most striking overlap, with more variance shared between the past and future tasks (64%) than remains unexplained (36%). Unlike the specificity and details tasks, the fluency tasks typically did not require retrieval and/or binding of specific sensory and contextual details. Instead, successful performance on the fluency tasks required quick access to a large number of personally relevant events. For example, an acceptable response on the fluency test might be simply ‘‘Graduation,’’ or ‘‘Vacation at the Beach.’’ These represent broad semantic event categories, but not necessarily highly integrated event representations. The strong correlation between past and future autobiographical fluency therefore suggests that accessing general autobiographical event information – regardless of temporal direction – appears to load on largely overlapping retrieval processes. This is consistent with recent research suggesting that both re-construction of past events and construction of future ones start with access to personally relevant semantic information and general event categories (Conway & Pleydell-Pearce, 2000; D’Argembeau & Mathy, 2011), which is then used as a scaffold to construct an integrated and detailed representation of a single event. Notably, the overlap between the past and future measures was not as strong for the specificity and details tests as for the fluency tests, suggesting that binding and elaboration of event details for future events rely more heavily on additional cognitive functions. WM appears to be one of those functions, but only for the ability to construct a future event occurring at a specific place and time. That is, composite WM capacity was significantly correlated with future episodic specificity, over and above the relationship between future and past specificity, but did not significantly contribute to the other future thought measures. This finding is largely consistent with neuroimaging research that has identified preferential engagement in portions of right frontopolar and right hippocampal cortices during future, relative to past, event construction, but not necessarily during event elaboration (Addis et al., 2007). Notably, our observed specificity scores were lower than those previously reported by D’Argembeau et al. (2010). This is likely the result of a computerized method of data collection, which did not include 682 P.F. Hill, L.J. Emery / Consciousness and Cognition 22 (2013) 677–683 experimenter prompting. It should be noted, however, that our findings were strikingly similar to the findings of D’Argembeau et al., 2010, in which the combined WM/EF composite was uncorrelated with the past specificity measure (r = .05), but modestly and significantly correlated with future specificity (r = .23). The fact that WM was not significantly related to the number of episodic details reported for future events may seem surprising given that past behavioral research has found significant relationships between the production of episodic details and both EF (D’Argembeau et al., 2010) and WM (Addis et al., 2008). As described in Section 1, these discrepancies are likely a result of differences in the methodology and populations used across studies. For example, the EF factor in the D’Argembeau study consisted primarily of fluency tasks, which may be particularly important for the production of a large number of details of a single event. Although we did not include an independent assessment of fluency in the current study, it is relevant to note that our autobiographical fluency measures were consistently correlated with the details task(s), but not the specificity task(s), supporting this explanation. The details task also used more specific cues (e.g., ‘‘Imagine something you will do on your next vacation’’) than the specificity task, which used single open-ended cue words, such that the construction of a single event may have played less of a role in the details task than in the specificity task. It is possible that providing supplementary details in the former task simply involved accessing and listing information not directly tied to any previous experiences (e.g., imagining a blue sky and crashing waves in response to a generic beach scene; Addis & Schacter, 2012), thus minimizing demands on the recombination of episodic details into a coherent, unitary future event (Summerfield, Hassabis, & Maguire, 2010). In conclusion, the current study suggests that WM contributes to the construction of a single, coherent, future event depiction, but not to the retrieval or elaboration of event details. This finding is broadly consistent with the previous neuroimaging study by Addis et al. (2007), in which the unique frontal and right hippocampal activation for future events was found when constructing (analogous to our specificity task), but not elaborating on (analogous to our details task), episodic events. Though future experimental research using dual-task interference methodology (e.g., Baddeley et al., 2011) will help to clarify the relative role of binding vs. EF in future event construction, the present results provide an important theoretical and empirical link between WM capacity and episodic future thought. This has important implications for why some populations (e.g., older adults) might have difficulty imagining future events. Acknowledgments This study was conducted in fulfillment of Paul Hill’s master’s thesis under the direction of Lisa Emery. We thank Douglas Waring and Mark Zrull for serving on the thesis committee, Rose Mary Webb for providing valuable guidance, and Elizabeth Gabel, Rachel Kimel, Michele DiCio, and Alexander Stanley for their assistance with data collection. References Addis, D. R., & Schacter, D. L. (2012). The hippocampus and imagining the future: Where do we stand? Frontiers in Human Neuroscience, 5, 1–15. http:// dx.doi.org/10.3389/fnhum.2011.00173. Addis, D., Wong, A., & Schacter, D. (2007). Remembering the past and imagining the future: Common and distinct neural substrates during event construction and elaboration. Neuropsychologia, 45, 1363–1377. http://dx.doi.org/10.1016/j.neuropsychologia.2006.10.016. Addis, D. R., Wong, A. T., & Schacter, D. L. (2008). Age-related changes in the episodic simulation of future events. Psychological Science, 19, 33–41. http:// dx.doi.org/10.1111/j.1467-9280.2008.02043.x. Baddeley, A. (2000). The episodic buffer: A new component of working memory? Trends in Cognitive Sciences, 4, 417–423. http://dx.doi.org/10.1016/S13646613(00)01538-2. Baddeley, A. D., Allen, R. J., & Hitch, G. J. (2011). Binding in visual working memory: The role of the episodic buffer. Neuropsychologia, 49, 1393–1400. http:// dx.doi.org/10.1016/j.neuropsychologia.2010.12.042. Baddeley, A. D., & Hitch, G. J. (1974). Working memory. In G. H. Bower (Ed.). The psychology of learning and motivation: Advances in research and theory (Vol. 8, pp. 47–89). New York, NY: Academic Press. Baird, B., Smallwood, J., & Schooler, J. W. (2011). Back to the future: Autobiographical planning and the functionality of mind-wandering. Consciousness and Cognition, 20, 1604–1611. http://dx.doi.org/10.1016/j.concog.2011.08.007. Bickel, W. K., Yi, R., Landes, R. D., Hill, P. F., & Baxter, C. (2011). Remember the future: Working memory training decreases delay discounting among stimulant addicts. Biological Psychiatry, 69, 260–265. http://dx.doi.org/10.1016/j.biopsych.2010.08.017. Busby, J., & Suddendorf, T. (2005). Recalling yesterday and predicting tomorrow. Cognitive Development, 20, 362–372. http://dx.doi.org/10.1016/ j.cogdev.2005.05.002. Conway, M., & Pleydell-Pearce, C. (2000). The construction of autobiographical memories in the self-memory system. Psychological Review, 107, 261–288. http://dx.doi.org/10.1037/0033-295X.107.2.261. Cowan, N. (1999). An embedded-processes model of working memory. In A. Miyake & P. Shah (Eds.), Models of working memory: Mechanisms of active maintenance and executive control (pp. 62–101). New York, NY: Cambridge University Press. D’Argembeau, A., & Mathy, A. (2011). Tracking the construction of episodic future thoughts. Journal of Experimental Psychology: General, 140, 258–271. http:// dx.doi.org/10.1037/a0022581. D’Argembeau, A., Ortoleva, C., Jumentier, S., & Van der Linden, M. (2010). Component processes underlying future thinking. Memory & Cognition, 38, 809–819. http://dx.doi.org/10.3758/MC.38.6.809. D’Argembeau, A., Raffard, S., & Van der Linden, M. (2008). Remembering the past and imagining the future in schizophrenia. Journal of Abnormal Psychology, 117, 247–251. http://dx.doi.org/10.1037/0021-843X.117.1.247. D’Argembeau, A., & Van der Linden, M. (2004). Phenomenal characteristics associated with projecting oneself back into the past and forward into the future: Influence of valence and temporal distance. Consciousness and Cognition: An International Journal, 13, 844–858. http://dx.doi.org/10.1016/ j.concog.2004.07.007. Engle, R. W., Tuholski, J. E., Laughlin, J. E., & Conway, A. R. A. (1999). Working memory, short-term memory, and general fluid intelligence: A latent-variable approach. Journal of Experimental Psychology: General, 128, 309–331. Hassabis, D., Kumaran, D., Vann, S., & Maguire, E. (2007). Patients with hippocampal amnesia cannot imagine new experiences. Proceedings of the National Academy of Sciences of the United States of America, 104, 1726–1731. http://dx.doi.org/10.1073/pnas.0610561104. P.F. Hill, L.J. Emery / Consciousness and Cognition 22 (2013) 677–683 683 Kane, M. J., Hambrick, D. Z., Tuholski, S. W., Wilhelm, O., Payne, T. W., & Engle, R. W. (2004). The generality of working memory capacity: A latent-variable approach to verbal and visuospatial memory span and reasoning. Journal of Experimental Psychology: General, 133, 189–217. http://dx.doi.org/10.1037/ 0096-3445.133.2.189. Klein, S., Loftus, J., & Kihlstrom, J. (2002). Memory and temporal experience: The effects of episodic memory loss on an amnesic patient’s ability to remember the past and imagine the future. Social Cognition, 20, 353–379. http://dx.doi.org/10.1521/soco.20.5.353.21125. Lind, S. E., & Bowler, D. M. (2010). Episodic memory and episodic future thinking in adults with autism. Journal of Abnormal Psychology, 119, 896–905. http:// dx.doi.org/10.1037/a0020631. MacLeod, A. K., & Byrne, A. (1996). Anxiety, depression, and the anticipation of future positive and negative experiences. Journal of Abnormal Psychology, 105, 286–289. Okuda, J., Fujii, T., Ohtake, H., Tsukiura, T., Tanji, K., Suzuki, K., et al (2003). Thinking of the future and past: The roles of the frontal pole and the medial temporal lobes. NeuroImage, 19, 1369–1380. http://dx.doi.org/10.1016/S1053-8119(03)00179-4. Schacter, D. L., & Addis, D. R. (2007a). The cognitive neuroscience of constructive memory: Remembering the past and imagining the future. Philosophical Transactionsof the Royal Society of London Series B: Biological Sciences, 362, 773–786. http://dx.doi.org/10.1098/rstb.2007.2087. Schacter, D. L., & Addis, D. R. (2007b). On the constructive episodic simulation of past and future events. Behavioral and Brain Sciences, 30, 331–332. Schacter, D. L., Addis, D. R., Hassabis, D., Martin, V. C., Spreng, R. N., & Szpunar, K. K. (2012). The future of memory: Remembering, imagining, and the brain. Neuron, 76, 677–694. http://dx.doi.org/10.1016/j.neuron.2012.11.001. Schneider, W., Eschman, A., & Zuccolotto, A. (2002). E-Prime reference guide. Pittsburgh, PA: Psychology Software Tools, Inc.. Smallwood, J., Nind, L., & O’Connor, R. C. (2009). When is your head at? An exploration of the factors associated with the temporal focus of the wandering mind. Consciousness and Cognition, 18, 118–125. http://dx.doi.org/10.1016/j.concog.2008.11.004. Spreng, R., Mar, R., & Kim, A. (2009). The common neural basis of autobiographical memory, prospection, navigation, theory of mind, and the default mode: A quantitative meta-analysis. Journal of Cognitive Neuroscience, 21, 489–510. http://dx.doi.org/10.1162/jocn.2008.21029. Spreng, R. N., Stevens, W. D., Chamberlain, J. P., Gilmore, A. W., & Schacter, D. L. (2010). Default network activity, coupled with the frontoparietal control network, supports goal-directed cognition. NeuroImage, 53, 303–317. http://dx.doi.org/10.1016/j.neuroimage.2010.06.016. Suddendorf, T., & Busby, J. (2005). Making decisions with the future in mind: Developmental and comparative identification of mental time travel. Learning and Motivation, 36, 110–125. http://dx.doi.org/10.1016/j.lmot.2005.02.010. Suddendorf, T., & Corballis, M. (2007). The evolution of foresight: What is mental time travel, and is it unique to humans? Behavioral and Brain Sciences, 30, 299–313. http://dx.doi.org/10.1017/S0140525X07001975. Summerfield, J. J., Hassabis, D., & Maguire, E. A. (2010). Differential engagement of brain regions within a ‘core’ network during scene construction. Neuropsychologia, 48, 1501–1509. http://dx.doi.org/10.1016/j.neuropsychologia.2010.01.022. Szpunar, K. (2010). Episodic future thought: An emerging concept. Perspectives on Psychological Science, 5, 142–162. http://dx.doi.org/10.1177/ 1745691610362350. Szpunar, K. K., Chan, J. K., & McDermott, K. B. (2009). Contextual processing in episodic future thought. Cerebral Cortex, 19, 1539–1548. http://dx.doi.org/ 10.1093/cercor/bhn191. Szpunar, K., Watson, J., & McDermott, K. (2007). Neural substrates of envisioning the future. Proceedings of the National Academy of Sciences of the United States of America, 104, 642–647. http://dx.doi.org/10.1073/pnas.0610082104. Tse, C., & Altarriba, J. (2007). Testing the associative-link hypothesis in immediate serial recall: Evidence from word frequency and word imageability effects. Memory, 15, 675–690. http://dx.doi.org/10.1080/09658210701467186. Tulving, E. (1985). Memory and consciousness. Canadian Psychologist, 26, 1–12. Unsworth, N., & Engle, R. W. (2007). The nature of individual differences in working memory capacity: Active maintenance in primary memory and controlled search from secondary memory. Psychological Review, 114, 104–132. http://dx.doi.org/10.1037/0033-295X.114.1.104. Unsworth, N., Heitz, R. P., Schrock, J. C., & Engle, R. W. (2005). An automated version of the operation span task. Behavior Research Methods, 37, 498–505. Viard, A., Chételat, G., Lebreton, K., Desgranges, B., Landeau, B., Sayette, V. D. L., et al (2011). Mental time travel into the past and the future in healthy aged adults: An fMRI study. Brain and Cognition, 75, 1–9. Williams, J. M. G., Ellis, N. C., Tyers, C., Healy, H., Rose, G., & MacLeod, A. K. (1996). The specificity of autobiographical memory and imageability of the future. Memory & Cognition, 24, 116–125. http://dx.doi.org/10.3758/BF03197278.

Consciousness and Cognition 19 (2010) 1138–1139 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Reply A Reply to Uttl and Morin’s (2010) Commentary of Hughes and Nicholson (2010) q Susan M. Hughes ⇑, Julia Heberle Albright College, Reading, PA, United States a r t i c l e i n f o Article history: Available online 16 September 2010 Keywords: Self-recognition Face Voice Hemispheric specialization Laterality a b s t r a c t In response to a commentary provided by Uttl and Morin (2010) regarding the recent study by Hughes and Nicholson (2010), we evaluate their suggestion to modify our study’s design to reduce ceiling effects (similar to that of Rosa, Lassonde, Pinard, Keenan, & Belin, 2008). Also, the commentators failed to take into account our data on reaction times (a measure that is arguably less affected by ceiling effects), which help substantiate our conclusions regarding self-face and self-voice recognition. This rejoinder encourages readers to consider the relevance of the ecological validity of Hughes and Nicholson’s findings. Ó 2010 Elsevier Inc. All rights reserved. We thank Uttl and Morin (2010) for their scholarly effort in examining our article, Hughes and Nicholson (2010). First, we would like to correct Uttl and Morin’s report of our finding; they stated that we found ‘‘that the left hemisphere is superior in self-recognition to the right hemisphere,” (p. 1) when in fact we found a left-hand advantage, thereby implicating a right hemisphere advantage, on our vocal and combined voice/face self-recognition tasks (listed in Abstract, pp. 5, 7, 9, 10). Uttl and Morin also claim that we ‘‘did not consider much of the evidence showing no such right hemisphere dominance in self-recognition (see Morin, in press) as well as in other forms of self-awareness such as autobiographical memory and emotion awareness” (p. 1). Our review of the scientific literature on laterality of recognition of self-stimuli clearly highlights right-hemisphere involvement, left-hemisphere involvement, and dual hemisphere involvement for self-recognition tasks (p. 3), as well as our inclusion of studies examining lateralized hemispheric dominance not just for self-face and self-voice processing but also several other types of self-related information processing (p. 3). While the majority of scientific studies implicate the right hemisphere in this arena, we are not denying evidence for the role of the left hemisphere on these tasks, as well. We also pointed readers to a more detailed review of the literature on visual self-recognition by Uddin, Kaplan, Molnar-Szakacs, Zaidel, and Iacobonib (2005) who concluded that the extent to which laterality of self-recognition remains open to question, but acknowledged that the majority of studies report a more prominent role of the right hemisphere. As there is an appreciable amount of literature on this topic, not all of the studies and commentaries on this subject can be reviewed within one article given the limited space available in most journals. As our study had additional aims, we were also responsible for presenting literature beyond lateralization. Although we thank the authors for pointing out that they have recently produced a valuable work reviewing relevant evidence of laterality of self-information (Morin, in press), it proves rather difficult to consult a work that is in press and not in print, as our work was submitted well before the suggested article was publicly available. q Reply to Commentary on Hughes, S. M., Nicholson, S. E. (2010). The processing of auditory and visual recognition of self-stimuli. Consciousness and Cognition, 19, 1124–1134. ⇑ Corresponding author. Address: Albright College, Psychology Department, 13th and Bern Streets, Reading, PA 19612, United States. E-mail address: shughes@alb.edu (S.M. Hughes). 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.06.013 S.M. Hughes, J. Heberle / Consciousness and Cognition 19 (2010) 1138–1139 1139 Given the ample literature showing that human voices are generally more difficult to recognize than are human faces (e.g., Ellis, Jones, & Mosdell, 1997; Hanley, Smith, & Hadfield, 1998; Joassin, Maurage, Bruyer, Crommelinck, & Campanella, 2004, 2008), it is not implausible to predict that the same also hold true for the recognition of self-face versus self-voice stimuli. It is also not surprising that simple tasks of identifying one’s own face and voice would produce nearly perfect accuracy and exhibit fast reaction times. Thus, one aim of our study (given the simplicity of the task of recognizing self-voice and selfface) was to determine if there were any differences in this ability across these sensory modalities. Uttl and Morin claim that due to the simplicity of our self-recognition tasks, high accuracy scores for some of our comparison groups created a situation where our data was confounded by ceiling effects, thereby rendering our results inconclusive. Uttl and Morin suggested that to overcome purported ceiling effects, we should have considered manipulating the difficulty of our self-recognition task as did Rosa et al. (2008) who had also examined vocal self-recognition. When Rosa et al. created a more difficult task to help control for certain confounds such as task difficulty, their findings were comparable to ours – with there being a lefthand (right hemisphere) advantage for vocal self-recognition. So whereas Rosa et al.’s original vocal self-recognition task produced null findings, our tasks had not, and yielded the same results as did Rosa et al.’s subsequent tasks of greater difficulty. Furthermore, our experiment included an additional element of difficulty in considering both vocal and visual stimuli together, making it a unique contribution to the literature. That is, unlike previous studies, we examined how simple tasks of self-recognition across two sensory modalities compare to and impact one another when presented simultaneously. Uttl and Morin’s suggestion of manipulating the difficulty of the self-recognition tasks, however, is duly noted and could be employed in future incarnations of this line of research. What makes Uttl and Morin’s critique even less compelling is that they are ignoring half of the data presented in our research report. We found similar results throughout our analyses with respect to reaction time measures (i.e., a measure that is arguably less affected by ceiling or floor effects given that this measure was recorded in milliseconds), providing converging evidence with our findings regarding task accuracy, and thus justifying our overall conclusion that self-face recognition was superior to self-voice recognition. We appreciate Uttl and Morin’s attention to our study, their suggestions for plausible alternative explanations to our data, and their indications for future avenues for this research. While an intellectual discourse of the interpretation of findings is always encouraged and vital to science, this can only be accomplished when the commentators provide an accurate interpretation and account of the target article. Uttl and Morin’s commentary presents a misconception of our aims, findings, and interpretations. We (Hughes & Nicholson, 2010) addressed a heretofore unexamined phenomenon empirically, and substantiated our predictions. We feel that the ecological validity of our findings adds new breadth and depth to the cumulative body of knowledge on this topic and encourage others to replicate and extend our findings. References Ellis, H., Jones, D., & Mosdell, N. (1997). Intra- and inter-modal repetition priming of familiar faces and voices. British Journal of Psychology, 88, 143–156. Hanley, J., Smith, S., & Hadfield, J. (1998). I recognize you but I can’t place you: An investigation of familiar-only experiences during tests of voice and face recognition. The Quarterly Journal of Experimental Psychology A: Human Experimental Psychology, 51A, 179–195. Hughes, S. M., & Nicholson, S. E. (2010). The processing of auditory and visual recognition of self-stimuli. Consciousness and Cognition, 19, 1124–1134. Joassin, F., Maurage, P., Bruyer, R., Crommelinck, M., & Campanella, S. (2004). When audition alters vision: An event-related potential study of the crossmodal interactions between faces and voices. Neuroscience Letters, 369, 132–137. Joassin, F., Maurage, P., & Campanella, S. (2008). Perceptual complexity of faces and voices modulates cross-model behavioral facilitation effects. Neuropsychological Trends, 3, 29–44. Morin, A. (in press). Self-recognition, theory-of-mind, and self-awareness: What side are you on? Laterality. Rosa, C., Lassonde, M., Pinard, C., Keenan, J., & Belin, P. (2008). Investigations of hemispheric specialization of self-voice recognition. Brain and Cognition, 68, 204–214. Uddin, L. Q., Kaplan, J. T., Molnar-Szakacs, I., Zaidel, E., & Iacobonib, M. (2005). Self-face recognition activates a frontoparietal ‘‘mirror” network in the right hemisphere: An event-related fMRI study. Neuroimage, 25, 926–935.

Consciousness and Cognition 19 (2010) 1124–1134 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog The processing of auditory and visual recognition of self-stimuli Susan M. Hughes *, Shevon E. Nicholson Department of Psychology, Albright College, Reading, PA 19612, United States a r t i c l e i n f o Article history: Received 1 August 2008 Available online 26 March 2010 Keywords: Auditory self-recognition Visual self-recognition Lateralization Voice recognition a b s t r a c t This study examined self-recognition processing in both the auditory and visual modalities by determining how comparable hearing a recording of one’s own voice was to seeing photograph of one’s own face. We also investigated whether the simultaneous presentation of auditory and visual self-stimuli would either facilitate or inhibit self-identification. Ninetyone participants completed reaction-time tasks of self-recognition when presented with their own faces, own voices, and combinations of the two. Reaction time and errors made when responding with both the right and left hand were recorded to determine if there were lateralization effects on these tasks. Our findings showed that visual self-recognition for facial photographs appears to be superior to auditory self-recognition for voice recordings. Furthermore, a combined presentation of one’s own face and voice appeared to inhibit rather than facilitate self-recognition and there was a left-hand advantage for reaction time on the combined-presentation tasks. Ó 2010 Elsevier Inc. All rights reserved. 1. Introduction The majority of self-recognition studies in humans have focused on self-face recognition (see Uddin, Kaplan, MolnarSzakacs, Zaidel, and Iacobonib (2005) for review), whereas studies examining self-voice recognition are not as abundant (e.g., Holzman, Rousey, & Snyder, 1966; Nakamura et al., 2001; Olivos, 1967; Rosa, Lassonde, Pinard, Keenan, & Belin, 2008; Rousey & Holzman, 1967). To our knowledge, there are even fewer investigations specifically examining the interaction of the visual and auditory senses when processing self-information simultaneously. Therefore, the present study examined reactions to hearing a recording of one’s own voice in comparison to seeing a photograph of one’s own face, and to explore whether a simultaneous, cross-modal presentation of self-stimuli would facilitate or inhibit self-identification. Many investigations have shown that voices are generally more difficult to recognize than faces (Ellis, Jones, & Mosdell, 1997; Hanley, Smith, & Hadfield, 1998; Joassin, Maurage, Bruyer, Crommelinck, & Campanella, 2004; Joassin, Maurage, & Campenella, 2008). Furthermore, people are more capable of retrieving biographical information about a person from seeing their face than from hearing their voice, regardless of how familiar the voice is to the listener (Damjanovic & Hanley, 2007; Hanley et al., 1998). When there is simultaneous, multimodal presentation of voices and faces, two behavioral effects have been observed – facilitation effects and interference effects. Joassin, Maurage, Bruyer, Crommelinck, and Campanella (2004) found that the recognition of face-voice associations led to an interference effect, whereby performance had decreased in the bimodal condition relative to the recognition of faces presented in isolation. They found that the addition of auditory information interfered with the processing of visual information, and the addition of visual information did not facilitate the processing of auditory information. Cook and Wilding (1997) found that the presence of a face along with a voice was, in fact, detrimental to voice-recognition memory, as well. * Corresponding author. Fax: +1 610 921 7883. E-mail address: shughes@alb.edu (S.M. Hughes). 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.03.001 S.M. Hughes, S.E. Nicholson / Consciousness and Cognition 19 (2010) 1124–1134 1125 Other studies have reported that simultaneous presentations of auditory and visual information produce a facilitation effect than when presented alone (Molholm et al., 2002). Specifically, when faces and voices are presented concurrently, facilitation effects are seen when facial stimuli are blurred to create a situation where they were just as equally difficult to recognize as voices (Joassin, Maurage, & Campanella, 2008). Joassin et al. used morphing software so that recognition performances for faces were equivalent as those with voices, and found that the simultaneous presentation of these equally difficult to recognize stimuli facilitated rather than hampered recognition. Walker, Bruce and O’Malley (1995) examined how familiarity played a role in facial identity and facial speech recognition when these two tasks are carried out simultaneously. Their results indicated that those who were familiar with the faces shown had fewer correct expected-combination responses than those unfamiliar with the seen faces, regardless of whether the face/voice combinations were congruent (i.e. face paired with voice of same person) or incongruent (i.e. face paired with the voice of another that was either the same or opposite gender). Although Walker et al.’s findings may shed further light as to the interaction between how voice and face information is processed together (especially when considering familiar stimuli), the study was directly testing the McGurk effect (i.e. when discrepant auditory and visual information is presented while seeing a speaker talk) and was dependent upon examining a speaker’s lip movement while speaking to gather both the auditory and visual information at the same time. While the present study does not test for the McGurk effect, it addresses whether the processing of facial and vocal information involving self-stimuli extend these previous findings on face and voice recognition. When experimentally comparing visual and auditory self-stimuli, there is the inherent problem that the two modalities may not be exactly comparable and the self-stimuli obtained for testing may present some perceptual distortions for the participant. For instance, the sound of hearing one’s own voice internally while speaking is perceived through both air and bone conduction while hearing a recording of one’s own voice, as others hear it, occurs through air conduction alone (Holzman, Berger, & Rousey, 1967; Maurer & Landis, 1990). Therefore, individuals hear their own voices internally in a distorted manner and different from that of hearing a recording of their own voice. The extent to which bone- to air-conduction ratio is implicated when reacting to hearing one’s own voice is an issue confronted by several previous investigations that have examined vocal self-recognition (Holzman & Rousey, 1966; Holzman et al., 1966, 1967). There is difficulty in experimentally trying to approximate how playback usually sounds to a person as it does internally, and unfortunately, there has not yet been a way to experimentally control and account for this distortion. Despite the concerns for using these self-stimuli, it is likely the case that individuals have been exposed to both seeing their own photos and hearing samples of their own voice recordings numerous times and both represent somewhat familiar self-stimuli, albeit individuals are probably less familiar with hearing their own voice than seeing their own face. When people are confronted with hearing a recording of their own voice, they are often vexed by the experience. Holzman and Rousey (1966) showed that individuals tend to first respond with a negative affective reaction after hearing their own voice, note discrepancies between what they expected to hear and what they heard, and then react with a defensive negation. In contrast, these sorts of reactions do not occur when hearing voice recordings of others. Moreover, people tend to focus on the grammar, syntax, and psychological characteristics of other speakers, while they focus on the tonal qualities when hearing their own voice (Holzman & Rousey, 1966). Just as with voices, reactions towards seeing one’s own photograph can create perturbing reactions; individuals may experience negative reactions immediately following evaluation of their own facial photo if that image deviates from what their ideas of self-image are (Morita et al., 2008). Hearing a recording of one’s own voice may not just be a reaction of the unfamiliar, but is also a complex reaction of familiar characteristics (i.e., awareness of aspects of one’s own personality mirrored in their voice), (Holzman & Rousey, 1966), and could still conjure a sense of self as does viewing a photo of one’s own face. Furthermore, Holzman et al. (1967) showed that reactions towards hearing a recording of one’s own voice cannot be solely attributed to the physical differences between one’s speaking voice and their recorded voice. The authors suggested that the self-voice confrontation experience is converting the voice from the usual position of mediator of expression into a percept, permitting one to hear the qualities of paralanguage of communication that they may not otherwise realize while speaking (Holzman & Rousey, 1966; Holzman et al., 1967). It also appears that vocal self-recognition is dependent upon the length of the vocal segment heard and prior exposure to hearing recordings of one’s own voice. Rousey and Holzman (1967) found that only 38% of participants could accurately identify their own voices immediately using voice segment lengths of either 1 or 5 s, while Olivos (1967) demonstrated that 55% of participants could discriminate their own voices if allowed 15 s to respond. The low rates of vocal self-recognition were attributed to participants experiencing the perceptual distortion occurring when hearing one’s own voice through both air and bone conduction as opposed to hearing a recording of one’s own voice through only air conduction (Holzman et al., 1967; Maurer & Landis, 1990). However, Rousey and Holzman (1967) demonstrated that mere exposure/familiarity to hearing recordings of one’s own voice plays a role in successfully recognizing one’s own voice and facilitates self-recognition. Participants who heard their recorded voices on a frequent basis of more than once a week showed an 83% accuracy rate for recognizing their own voice, while radio announcers (who frequently hear their own recorded voices) showed a 100% accuracy rate. Self-recognition of facial identity has been extensively investigated in humans using lesion and brain imaging studies. Previous research has found evidence of right hemisphere dominance for the recognition of a photograph of one’s own face (Keenan, Nelson, O’Connor, & Pascual-Leone, 2001; Keenan, Wheeler, Platek, Landi, & Lassonde, 2003; Keenan et al., 1999; Platek, Myers, Critton, & Gallup, 2003; Platek, Thomson, & Gallup, 2004; Platek et al., 2006). Using fMRI techniques, Morita et al. (2008) found evidence self-face recognition of a shown photograph in the right prefrontal cortex, specifically the right 1126 S.M. Hughes, S.E. Nicholson / Consciousness and Cognition 19 (2010) 1124–1134 precentral gyrus. Devue et al. (2007) also showed greater activation of the right frontal cortex in response to seeing one’s own self-photo as compared with visual processing of other highly familiar persons. Keenan, Wheeler, Platek et al. (2003) and Platek et al. (2003) examined a patient with a complete callosotomy who showed right hemisphere processing advantage (as indicated with left hand) for self-morphed faces. Claims for right hemisphere specialization for self-recognition are not universal, and evidence for left hemisphere or bilateral involvement exists. For instance, Kircher et al. (2001) found evidence of left-hemisphere activation when viewing images of one’s own face. Similarly, Brady, Campbell, and Flaherty (2004) used mirror-symmetric images made from both the right half and the left half of participants’ faces to examine which image participants would more closely self-identify. Participants consistently reported that the mirror-symmetric image of the left half of the face (seen in the right visual field when looking in a mirror) more closely resembled their own image. These findings suggest that because the right visual field projects to the left hemisphere, there may be a left hemisphere bias for facial self-recognition. Turk et al. (2002) examined self-recognition in a split-brain patient and claimed that while both hemispheres are capable of self-recognition, cortical networks in the left hemisphere played the more important role in the execution of this process for this patient. Uddin, Rayman, and Zaidel (2005) also studied self-recognition in a split-brain patient by presenting photographs of the patient, a familiar acquaintance, or an unknown face to independent hemispheres. Their results revealed an equal ability in both hemispheres for self-recognition, but right hemisphere dominance for the recognition of familiar acquaintances. Sugiura et al. (2005) showed activations of the right occipito–temporo-parietal junction and frontal operculum, as well as in the left fusiform gyrus for self-face recognition as opposed to the temporoparietal junction in both hemispheres and the left anterior temporal cortex which was activated when viewing familiar faces. Uddin et al. (2005) provided a detailed review of the current literature on visual self-recognition and concluded that the extent to which self-recognition is lateralized is still an open question, although the majority of studies report the role of the right hemisphere. Morin (2002) provides a commentary of how the ‘‘full sense of self” cannot solely be attributed to the role of the right hemisphere. Others have concluded that different lateral activations are dependent upon the self-face task employed (Sugiura et al., 2000). While researchers have extensively studied visual, facial processing of self-recognition in the brain, neuroimaging research concerning auditory self-recognition is more limited. Nakamura et al. (2001) showed greater activation of several brain areas (left frontal pole, right temporal pole, right entorhinal cortex, and left precuneus) during familiar voice discrimination tasks compared to controls, whereas discrimination of one’s own voice activated bilateral regions of the frontal cortex and were different from those activated by discrimination of familiar voices. Among the regions, the right inferior frontal sulcus and parainsular cortex were activated to a greater extent during the self-task and were implicated as being part of the neural network involved in vocal self-recognition. Nakamura et al. also found that adjusted regional cerebral blood flow in both the left frontal and right temporal poles is correlated with the number of participants’ correct identification of familiar voices, but had not correlated with the number of correct identification of own voice in all regions activated. Kaplan, AzizZadeh, Uddin and Iacoboni (2008) also showed that similar to seeing one’s own face in comparison to a friend’s face activated regions in the right interior frontal gyrus (IFG), listening to one’s own voice as compared to a friend’s voice, also showed increased activity in the right IFG. Moreover, Rosa et al. (2008) attempted to assess functional asymmetries of self-voice recognition and found a left hand/right hemisphere advantage for self-voice compared to other-voice recognition. Several studies have suggested that there may be lateralized hemispheric dominance for not just self-face and self-voice processing, but also several other types of self-related information processing. For instance, Molnar-Szakacs, Uddin and Iacabonib (2005) investigated the representation of self through the linguistic modality by measuring excitability of the hand representation of the motor cortex while participants were reading self-descriptive personality trait words. They found that both self-related and unrelated personality trait words yielded similar significant right hemisphere facilitation, suggesting that the right hemisphere plays a role in the processing of both self-relevant and affective stimuli. Platek et al. (2003) also found right hemisphere dominance for non-facial discrimination tasks, such as the recognition of self-descriptive adjectives. In another study, Platek et al. (2004) found that additional visual self-information (i.e., seeing one’s own name), auditory selfinformation (i.e., hearing one’s own name) and olfactory self-information (i.e., exposure to one’s self-odor) had all facilitated self-face recognition. Craik et al. (1999) demonstrated that self-related encoding of memory involves both specific activations of the left frontal and right frontal lobes as measured from PET scans. The aim of this study is to examine if hearing a recording of one’s own voice, as others would hear it, will elicit comparable reaction times and performance accuracy for self-recognition tasks as would seeing a photograph of one’s own face. In line with previous investigations showing that voices are generally recognized with greater difficulty than faces (Ellis et al., 1997; Hanley et al., 1998; Joassin et al., 2004, 2008), we hypothesized that participants would also display higher accuracy rates and faster reaction times in visual self-recognition tasks in comparison to auditory self-recognition tasks. Furthermore, we expected that cross-modal, simultaneous presentation of visual and auditory self-stimuli would inhibit self-identification. Joassin et al. (2004) observed that the bimodal presentation of faces paired with voices was significantly slower than the recognition of faces presented alone, and other studies have shown that even a well-practiced, ‘‘simple” task can be slowed due to dual–task interference across sensory modalities (Levy, Pasheler, & Boer, 2006). Furthermore, there is evidence that processing of familiar faces with voices results in poorer performance on speech perception tasks than unfamiliar stimuli (Walker, Bruce, & O’Malley, 1995). Lastly, we expected that there may be greater accuracy and faster response times for these self-recognition tasks when using the left versus the right hand since the majority of studies report the role of the right hemisphere/left-hand advantage for both the processing of one’s own face (Keenan, Wheeler, & Ewers, 2003; Keenan et al., S.M. Hughes, S.E. Nicholson / Consciousness and Cognition 19 (2010) 1124–1134 1127 1999; Platek et al., 2003) as well as one’s own voice (Rosa et al., 2008). It has been previously demonstrated that hemispheric cortical localization can translate into quicker reaction times for the contralateral hand (Hodges, Lyons, Cockell, & Reed, 1997) and the use of hand response to test for hemispheric specialization have been employed by several previous investigations (e.g., Keenan et al., 1999, 2003; Platek et al., 2003, 2004; Rosa et al., 2008). Therefore, following the procedures of these previous investigations, hand use was included as a measure to test for possible laterality on these auditory and visual self-recognition tasks. 2. Methods 2.1. Participants Ninety-one participants (males = 50, females = 41) were recruited from summer classes and other summer programs at Albright College. The mean age of the participants was 24.4 years (SD = 10.9; range = 17–60). The majority of the participants (91.2%) indicated they were right handed, with the remaining 8.8% being left-handed. Of the participants, 76.9% were Caucasian, 14.3% were African American, 4.4% were Hispanic, 2.2% were Asian, and 2.2% indicated being another ethnicity. Participants who were chronic smokers (smoked more than a pack of cigarettes a week), had throat, larynx, or auditory surgery, had a cold or illness that affected their speech, had a hearing impairment, or had a severe accent were eliminated so as to only include those who could provide clear voice samples. Two participants were excluded based on these grounds, thereby totaling our sample to 91 participants. 2.2. Materials Black-and-white photographs of male and female faces used as stimuli in this study were obtained from the internet from websites containing information about self-reported age and ethnicity (i.e. MySpace.com, etc.). Each photograph was standardized in color, size, and shape by Microsoft Photo Editor so that it could be embedded into a presentation of pictures of oneself and others. Pictures were standardized so that the overall view of a picture’s color, size and shape would not be the cues the participants used when making quick assessments of pictures in the timed reaction tasks. Voice recordings were obtained from a previous study of individuals reciting a count from one to ten and were recorded using Microsoft Sound Recorder software and the same microphone (Andrea NC-8) as the voices obtained from participants in this study so that both sets of comparable recordings could also be embedded within a presentation to identify self from other. All stimuli were presented using SuperLab 2.0.4 presentation software. 2.3. Procedure After obtaining informed consent, participants completed a brief demographic questionnaire regarding such information as their sex, age, ethnicity, handedness, etc. Then, we took three different black-and-white headshot photographs of each participant using a Kodak V803 digital camera with different backgrounds in each picture so that it could be used as a stimulus in the self-recognition presentation tasks. The photos were standardized in size and shapes identical to the photos obtained from the internet of others and were embedded within one of several different SuperLab presentations that closely matched the age, ethnicity, and sex based of the participants’ own demographic information provided. In other words, participants viewed presentations of stimuli that most closely matched themselves on these self-reported demographic dimensions. Participants then provided three different voice samples counting from one to ten at a pace of approximately one numeral per second that were used as stimuli for the self-recognition tasks. Because of the possibility that the content of what was said, or prosody of normal sentence-structured speech could influence the perception of the voice recordings, a number recitation was used. Therefore, this procedure was utilized in an attempt to obtain vocal samples that were both neutral and of comparable content and followed the methods of previous investigations of subjective vocal perceptions (Hughes, Dispenza, & Gallup, 2004; Hughes, Harrison, & Gallup, 2002; Pipitone & Gallup, 2008). A microphone (Andrea NC-8) positioned approximately 2.5 cm from the participant’s mouth was used to record voices with Microsoft Sound Recorder software. Three separate voice recordings were obtained so that participants would not be cued by particular features of a voice recording (e.g., longer pause made between numbers while speaking, etc.) when trying to self-identify. The participants completed a reaction time task which evaluated self-recognition under three conditions that presented stimuli containing: (1) the participant’s own photograph (visual self-recognition), (2) the participant’s own voice (auditory self-recognition), and (3) a pairing of the participant’s own photograph and voice. In the first two conditions, participants were presented with 30 faces or voices, respectively, six of which were their own face or voice, and 24 were of others. They were asked to indicate whether the stimulus was themselves or someone else by pressing one of two keys (1 or 2) on the number pad of the keyboard. The next stimulus was presented only after a response was made on the keyboard. Participants were told that the accuracy of their response as well as their reaction time was being recorded so that they were aware that they needed to complete the task as quickly as possible but with as few errors as possible. For the third condition, participants were presented with a total of 24 pictures paired with voices (six slides consisted of both their own face and voice paired together, six slides of their own voice paired with someone else’s face, six slides with their own face paired with 1128 S.M. Hughes, S.E. Nicholson / Consciousness and Cognition 19 (2010) 1124–1134 someone else’s voice, and six slides were of both face and voice of others paired together). Participants were instructed that only if their own face and voice were correctly paired together, was it considered a correct choice for the identification of self-recognition. Pilot data revealed that the combined condition took an overall slightly longer response time than the individual conditions. Therefore, the reasoning behind using 30 presentations for the independent visual and vocal conditions versus 24 in the combined condition was an attempt to avoid any fatigue effects for the combined condition and to make the tasks similar in overall response time. Furthermore, we obtained three different voice samples and pictures of the participants so as to reduce the possibility that repeated exposure of the exact same self-stimuli would be an advantage for selfidentification. Participants were tested for reaction time and errors made when responding with both the right and left hand to determine if there were lateralization effects for facial and vocal self-recognition in each of these conditions. It has been previously demonstrated that hemispheric cortical localization can translate into quicker reaction times for the contralateral hand (Hodges et al., 1997). We counterbalanced the order of the three condition sets (i.e., only visual, only vocal, and combined visual and vocal stimuli) and the order of the stimuli presented within each condition. Furthermore, since participants were asked to complete each of the three conditions using their right hand and left hand, thereby completing a total of six tasks, we had counterbalanced each of these six tasks. 3. Results 3.1. Accuracy for vocal and facial stimuli presented separately First, we compared the two conditions where visual and vocal stimuli were presented separately (i.e., participants had to identify whether photographs shown were of their own face or that of another and whether a count recitation from one to ten was their own voice or that of another person). We calculated the number of times a participant accurately identified a stimulus when presented with the self-stimulus and when presented with other-stimulus in each condition. We then converted these accuracy scores into percentages to analyze comparisons between groups and to consider the independent measures of gender, handedness, and hand used to complete the task. This follows the type of analysis used in similar previous investigations (Rosa et al., 2008; Weems & Zaidel, 2005). Accuracy of response was analyzed with a 2 (hand used)  2 (face/voice stimuli)  2 (self/other stimuli) repeated measures ANOVA. There was no main effect for hand used, F(1, 90) = 0.16, p = .691, but there was a main effect for face/voice stimuli, F(1, 90) = 30.05, p < .001, g2 = .250. Participants responded to the face condition with a significantly higher accuracy rate (M = 99.2%, SE = 0.2) than to the voice condition (M = 94.5%, SE = 0.8). There was also a main effect for self/other stimuli, F(1, 90) = 29.95, p < .001, g2 = .250. Participants responded to the self condition with a 94.8% (SE = 0.7) accuracy rate, whereas they responded to the other condition with a significantly higher 98.9% (SE = 0.3) accuracy rate. There was a significant twoway interaction between hand used to respond to stimuli and face/voice stimuli, F(1, 90) = 6.21, p = .015, g2 = .065, whereby the face recognition using the right hand (M = 99.7%, SE = .13) was greater than when using the left hand (M = 98.6%, SE = .38), t(90) = 2.82, p = .006, but was the opposite for voice stimuli, with recognition using the left hand (M = 95.3%, SE = .72) was greater than using the right hand (M = 93.8%, SE = 1.13), t(90) = 1.98, p < .05. There was also a significant two-way interaction between face/voice stimuli and self/other stimuli, F(1, 90) = 14.29, p < .001, g2 = .137. Although statistically significant, there was less of a difference between recognizing faces of self (M = 98.5%, SE = .38) versus other (M = 99.8%, SE = .08), t(90) = 3.30, p = .001, as there was between recognizing one’s own voice (M = 91.1%, SE = 1.45) and recognizing the voices of others (M = 97.9%, SE = .50), t(90) = 4.78, p < .001. Furthermore, a significant three-way interaction was found between hand used to respond to stimuli, face/voice stimuli, and self/other stimuli on accuracy scores, F(1, 90) = 6.10, p = .015, g2 = .063. An illustration of this three-way interaction is presented in Fig. 1, and post hoc pair-wise comparisons of all groups indicted that there was no significant difference between right and left-handed tasks when presented with stimuli of faces and voices of others, but there was a hand effect when it came to responding to self stimuli. Participants were more accurate at responding to their own voice with their left hand (M = 92.7%, SD = 12.9) than right hand (M = 89.6%, SD = 9.2), t(90) = 1.99, p = .05, whereas they were more slightly more accurate at responding to their own face with their right hand (M = 99.6%, SD = 2.5) than with their left hand (M = 97.5%, SD = 6.9), t(90) = 2.77, p = .007. When considering handedness and gender of the participants as variables, there was no main effect for handedness, F(1, 87) = 0.13, p = .718, nor for gender F(1, 87) = 2.58, p = .112, and no significant interactions were found for these two variables and any other measure. 3.2. Reaction time for vocal and facial stimuli presented separately A 2 (hand used)  2 (face/voice stimuli)  2 (self/other stimuli) repeated measures ANOVA was also used to examine reaction time for these tasks. There was no main effect for hand used during task, F(1, 90) = 0.37, p = .546. However, there was a main effect for face/voice stimuli, showing a faster response time to faces (M = 547.49 ms, SE = 14.03) versus voices (M = 1570.91 ms, SE = 72.73), F(1, 90) = 217.88, p < .001, g2 = .250. Similarly, another main effect revealed that participants responded faster to other stimuli (M = 1009.64 ms, SE = 28.33) than to self-stimuli (M = 1108.05 ms, SE = 55.71), F(1, 90) = 5.91, p = .017, g2 = .062. When gender and handedness were also considered as variables, there were no main effects for handedness, F(1, 87) = 0.53, p = .469, or for gender, F(1, 87) = 0.71, p = .401, and neither showed any significant interactions with other variables. Percent Correct S.M. Hughes, S.E. Nicholson / Consciousness and Cognition 19 (2010) 1124–1134 1129 Left Hand Right Hand Self Face Self Voice Other Face Other Voice Type of Stimuli Fig. 1. When faces and voices were presented separately on the self-recognition tasks, a significant three-way interaction was found between hand used to respond to stimuli, face/voice stimuli, and self/other stimuli on accuracy of response. 3.3. Accuracy for vocal and facial task presented concurrently Percent Correct For the third condition, participants were shown stimuli that were (1) both their own voice and face paired together, (2) their own face paired with someone else’s voice (3) their own voice paired with someone else’s face or (4) neither their own face nor own voice paired together. Similarly to the previous analysis, we calculated the number of times a participant accurately identified a stimulus for each of the four combinations stated above and converted the accuracy scores into percentages to analyze comparisons between groups. A 2 (hand used for response)  2 (self-face/other face)  2 (self-voice/other voice)  2 (sex of participant) factorial mixed model design was used to analyze accuracy of responses made during the third condition self-recognition task. There was no main effect for either gender, F(1, 89) = 0.20, p = .660, or for hand used to respond, F(1, 89) = 0.15, p = .701. However, there was a main effect difference in accuracy when recognizing one’s own face (M = 95.38, SE = 0.80) versus others’ faces (M = 99.75, SE = 0.13), F(1, 89) = 28.84, p < .001, g2 = .245. Similarly, there was a main effect difference in accuracy when recognizing one’s own voice (M = 96.36, SE = 0.75) versus others’ voices (M = 98.77, SE = 0.23), F(1, 89) = 9.65, p = .003, g2 = .098. There was also a significant interaction between recognition for self versus other for faces and self versus other for voices, F(1, 89) = 7.24, p = .009, g2 = .075. As shown in Fig. 2, post hoc analysis revealed that respondents showed a significantly lower level of accuracy responding to stimuli that included both their own face and voice paired together (93.1%, SE = 1.5) as compared to a 97.6% (SE = 0.46) accuracy rate when responding to their face paired with another’s voice, t(90) = 2.90, p = .005, a 99.6% (SE = 0.26) accuracy rate when responding to their own voice paired with another’s face, t(90) = 4.23, p < .001, and a 99.9% (SE = .06) accuracy rate when identifying stimuli that was neither their face nor voice, t(90) = 4.53, p < .001. Additionally, accuracy rate when responding to their face paired with another’s voice (M = 97.6%, SE = 0.46), was significantly lower than a when responding to their own voice paired with another’s face (M = 99.6%, SE = 0.26), t(90) = 3.65, p < .001, and to identifying stimuli that was neither their face nor voice Face and Voice Face Only Voice Only Neither Face nor Voice Type of Self Stimuli Fig. 2. Accuracy in response rates to self-identification of face and voice stimuli that were presented simultaneously. 1130 S.M. Hughes, S.E. Nicholson / Consciousness and Cognition 19 (2010) 1124–1134 (M = 99.9%, SE = .06), t(90) = 4.91, p < .001. When handedness of the participants was included as a variable, no main effect was found, F(1, 89) = 0.44, p = .508, nor any significant interactions. 3.4. Reaction time for vocal and facial task presented concurrently A 2 (hand used for response)  2 (self-face/other face)  2 (self-voice/other voice)  2 (sex of participant) factorial mixed model design was also used to analyze reaction time during this self-recognition task. There was a main effect for hand used for responding, F(1, 89) = 7.89, p = .006, g2 = .081, whereby it took respondents less time to respond with their left hand (M = 1144.1 ms, SE = 47.9) than with their right hand (M = 1251.6 ms, SE = 53.6). There was no main effect found for sex of participant, F(1, 89) = 0.89, p = .347. However, there was a main effect for self-face recognition (M = 1497.2 ms, SE = 67.6) compared to other-face recognition (M = 898.4 ms, SE = 46.1), F(1, 89) = 79.55, p < .001, g2 = .472. Similarly, there was a main effect for self-voice recognition (M = 1305.6 ms, SE = 61.6) compared to other-voice recognition (M = 1090.0 ms, SE = 38.5), F(1, 89) = 27.91, p < .001, g2 = .239. There was a significant two-way interaction between recognition for self versus other stimuli when voices and faces were presented simultaneously, F(1, 89) = 15.19, p < .001, g2 = .146. Post hoc analysis revealed that there were significant differences between each pairwise comparison, whereby participants took the least amount of time responding to stimuli that were neither their own voice and face (M = 872.9 ms, SE = 44.5), followed by stimuli that was their own voice, but not face (M = 923.9 ms, SE = 49.2), their own face but not voice (M = 1307.2 ms, SE = 41.9), and took the longest amount of time responding to their own face paired with their own voice (M = 1687.3 ms, SE = 103.3). We also found a significant twoway interaction between hand used for responding by recognition for self versus other voices on reaction time, F(1, 89) = 16.12, p < .001, g2 = .153. Response to the voices of others was similar when using the right hand (M = 1108.1 ms, SE = 44.1) versus left hand (M = 1072.0 ms, SE = 40.0), t(90) = 0.93, p = .355, whereas there was much more of a delay in response to one’s own voice when using the right hand (M = 1393.0 ms, SE = 70.6) versus the left hand (M = 1216.1 ms, SE = 61.5), t(90) = 3.76, p < .001. Furthermore, as shown in Fig. 3, a significant four-way interaction was found between sex of participant, hand used for responding, and self/other stimuli for faces/voices stimuli combined, F(1, 89) = 5.36, p = .023, g2 = .057. With the exception of the face only condition, males showed an overall significantly faster response using their left hand (M = 1100.7, SE = 69. 5) than their right hand (M = 1205.8, SE = 71.5) , F(1, 49) = 7.94, p = .007, whereas there was not a significant difference between the overall left hand responses (M = 1187.4 , SE = 63.3) and right hand responses (M = 1297.3 , SE = 80.1) for females, F(1, 40) = 2.38, p = .131. There were no main effect for handedness, F(1, 89) = 0.01, p = .921, nor any significant interactions with this variable. 4. Discussion 4.1. Visual versus auditory tasks When voice and face stimuli were presented separately in the self-recognition tasks, we found that participants showed faster response times and greater accuracy when completing the facial recognition tasks compared to the vocal recognition tasks. This finding supports previous studies that demonstrate an overall greater difficulty for recognizing voices than faces 2200 2000 Mean Reaction Time (ms) . 1800 1600 1400 1200 Face and Voice Face Only Voice Only Neither Face nor Voice 1000 800 600 400 200 0 Left Hand Right Hand Male Participants Left Hand Right Hand Female Participants Fig. 3. When faces and voices were presented together, a significant interaction was found between sex of participant, hand used for responding, and the combination of voice and face self-stimuli presented together for reaction time. S.M. Hughes, S.E. Nicholson / Consciousness and Cognition 19 (2010) 1124–1134 1131 (Ellis et al., 1997; Hanley et al., 1998; Joassin et al., 2004, 2008) and is also consistent with studies demonstrating the tendency to perform visual name identification tasks faster than auditory name identification tasks (Adams & Janata, 2002). Our hypothesis that individuals would perform better on the visual task may have also been likely confirmed because of the fact that visual information predominates human processing of information (Wilentz, 1971), and because hearing one’s own voice recording from an outside source may sound dissimilar than hearing one’s voice internally (i.e., there is distortion that occurs from bone conduction when perceiving one’s own voice as it is produced during normal speech, Holzman et al., 1967; Maurer & Landis, 1990). Furthermore, participants were overall much more accurate at identifying stimuli presented that was not themselves versus when it was, suggesting that it may take more effort to process self-information. Similarly, Kircher et al. (2001) demonstrated a greater delayed response time for self-recognition of more extensively self-faced morphed stimuli. Holzman and Rousey (1966) showed that listening to one’s own voice can represent a disrupting experience which mobilizes a defensive behavior, and their participants showed a shift toward the negative pole of the evaluative scale immediately following hearing their voices. Likewise for self-face recognition, Morita et al. (2008) notes that negative reactions can occur immediately following viewing an image of oneself if that image contradicts their expectation of what their self-image should appear as being. Perhaps these negative reactions that occur immediately following exposure to self-stimuli might be one reason why there was a slight delay in response and decrease in accuracy for recognition of self-information versus others. It may also be possible that participants were responding to physiological changes that occurred when hearing their own voices and seeing photos of their faces. Both Holzman et al. (1966) and Olivos (1967) demonstrated that psychophysiological responses to hearing one’s own voice is greater than when hearing others’ voices, whether or not the participant was consciously aware they were listening to his or her own voice. However, it was argued that either the somatic response could occur as a result of the conscious recognition of hearing one’s own voice, or it may be that the somatic response occurs prior to the recognition of one’s own voice, thereby increasing the probability of conscious self-recognition. According to our findings, since participants showed a greater delay and lower accuracy scores when identifying their own voice recording relative to those of others’, we can speculate that any physiological response occurring when hearing their own voice did not seem to enhance or facilitate self-recognition. In fact, it may have inhibited/delayed a response to self. Similarly, psychophysiological changes also occur when viewing one’s own face in self-recognition tasks; Sugiura et al. (2000) showed an increased skin conductance response during recognition of one’s own face compared to others’ faces in both explicit and non-explicit discrimination tasks. But just as with the voices, any physiological response occurring when seeing one’s own face did not seem to enhance self-recognition as compared to the recognition of others since participants were generally better at identifying others’ faces than their own. Perhaps the perceptual experience of identifying self-stimuli involves the complicated interaction of reciprocal feedback of the autonomic response and cognitive appraisal (Olivos, 1967), thereby delaying response to self-information. Nonetheless, the accuracy of recognizing one’s own voice was considerably higher (89–93%) in our study than the overall estimates presented in previous studies (38–55%), (Holzman et al., 1966; Rousey & Holzman, 1967), but were consistent with more recent studies that showed 94–96% accuracy rates for vocal self-recognition (Rosa et al., 2008). Familiarity and exposure to hearing recordings one’s own voice impact the success rate of vocal self-identification; Rousey and Holzman (1967) found that for participants who heard their recorded voices on a basis more than once a week showed a 83% accuracy rate for recognition of own voice, and radio announcers, who frequently hear their own voices, showed a 100% accuracy rate. Given that today’s technology may allow for greater exposure to hearing one’s own recorded voice (i.e., via phone messaging, internet phone chatting, etc.), this may explain the higher rates of accurate voice recognition of recordings in our study and that of Rosa et al. (2008) as compared to studies conducted several decades ago. However, this conclusion is only a conjecture, as we did not manipulate frequency of exposure to their own recordings in the study, nor take a record of the participants’ prior exposure to hearing recordings of their voice. 4.2. Combined vocal and facial stimuli Consistent with studies demonstrating that simultaneous presentation of facial and vocal stimuli interferes with identification tasks (Cook & Wilding, 1997; Joassin et al., 2004), we also found that this was the case when processing self-stimuli, whereby auditory self-stimuli presented concomitantly with visual self-stimuli inhibited performance and reaction times on the self-recognition tasks. When voice and faces were presented together, participants were comparably the least accurate and took the longest amount of time responding to their own voice and face correctly paired together for all conditions. On the other hand, participants responded the fastest at identifying if both a face and voice was not their own. The participants also performed better and took less time at identifying a combination of their own voice paired with another’s face than their own face paired with another’s voice. It makes sense that participants were using the visual information (the face) as an initial screening for trying to identify the correct combination of their own face and voice since visual information is presented immediately. In other words, as soon as they saw that it is not their own face, they could continue without paying attention to the voice. However, if their own face is presented, it takes additional time to process whether or not it is their own voice. Therefore, it appears that participants had to process the two forms of self-stimuli almost independently. Furthermore, when people are confronted with hearing a recording of their voice, they usually first perceive the recording as being discrepant from what they first intended to hear followed by an affective reaction, commonly associated with a disavowal of the voice heard (Holzman & Rousey, 1966). Therefore, the cognitive and affective appraisal of this response may slightly delay both 1132 S.M. Hughes, S.E. Nicholson / Consciousness and Cognition 19 (2010) 1124–1134 accuracy and response time for vocal self-recognition tasks regardless of its pairing with visual information and regardless of repeated exposure to the self-stimuli. 4.3. Laterality On the tasks that presented vocal and visual stimuli separately, there was no difference in task performance accuracy when using the right or left hand to respond to faces and voices of others, but there was an effect found for hand used when responding to self-stimuli. Participants were more accurate at responding to their own voice using their left hand, suggesting right hemisphere activation. This supports similar findings for a left hand/right hemisphere advantage for self-voice recognition (Rosa et al., 2008) and with those of Nakamura et al. (2001) and Kaplan et al. (2008) who found a significant right hemisphere activation bias through neuroimaging testing for self-voices. It is also consistent with findings that show that impairment of familiar voice recognition is correlated with damage to the right and not left hemisphere (Van Lancker & Canter, 1982). On the other hand, our participants showed greater accuracy using the right hand for self-face recognition, a finding which fails to replicate and contradicts previous investigations showing a left-hand advantage on such tasks (Keenan et al., 1999, 2003; Platek et al., 2003, 2004). However, a right hand advantage for face recognition was not found when considering reaction time (as was the measure used in these previous studies). Platek et al. (2004) also could not find a left-hand advantage for self-face recognition when primes from other senses were presented. However, in our study, when it came to the identification facial and auditory self-stimuli presented together, reaction time was again significantly faster using the left hand. Therefore, right hemisphere dominance seems to prevail when processing both visual and auditory self-stimuli simultaneously. While several previous investigations have used reaction times for contralateral hand use as a measure of hemispheric cortical localization (Hodges et al., 1997; Keenan et al., 1999, 2003; Platek et al., 2003, 2004; Rosa et al., 2008), other studies have questioned whether hand response manipulations really reveal any strong laterality effects (Weems & Zaidel, 2005). The use of dichotic listening may have be a more effective method to assess hemispheric lateralization in the processing of auditory cues, just as presentation to the right and left visual fields can do the same for visual processing. Therefore, our data involving laterality effects from hand use should be viewed with caution and future investigations should consider presenting the stimuli contra-laterally rather than utilize hand use as an index for lateralization effects. Moreover, in addition to self-report, perhaps use a more objective assessment technique, such as the Edinburgh Handedness Inventory could have been used to assess handedness. Nonetheless, given the fact that hand response is only a gross measure of hemispheric specialization (for a discussion, see Weems & Zaidel, 2005), the difference in response to self-stimuli found between left and right-hands for visual and vocal tasks can be seen as impressive. Furthermore, since we had included self-identified lefthanded individuals in this study, we were able to show that handedness of the participants seemed to have little impact on any of our findings, whereby the majority of the sample being right handed, still showed a left-hand advantage for the auditory and the combined visual and auditory tasks. 4.4. Sex differences In addition, a sex difference was found when considering the interaction for response hand used across each of the four types of combined facial and vocal stimuli presented (i.e. both self-face and voice, self-voice only, self-face only, or neither self-face nor voice), as shown in Fig. 3. There appears to be more of a left-hand advantage when identifying both one’s own voice and face paired correctly together for males than there was for females. This may be due to the fact that female brains function more bilaterally, as opposed to male brains which function more unilaterally for face processing recognition tasks (Proverbio, Brignone, Matarazzo, Del Zotto, & Zani, 2006). In other words, females have a reduced need for asymmetric transfer due to their greater connectivity between hemispheres, whereas males tend to show greater laterality for information processing (Zaidel, Aboitiz, Clarke, Kaiser, & Matteson, 1995). Proverbio et al. (2006) showed men had strong right hemispheric dominance for identifying both neutral and affective faces, whereas there was a lack of asymmetry for face processing in the amplitude of the occipito-temporal N1 response to these stimuli for women. Kurosaki, Shirao, Yamashita, Okamoto, and Yamawaki (2006) found sex differences when discriminating altered from intact versions of one’s own body; women confronted with an altered configuration of their body elicited activity in the prefrontal and limbic areas, whereas men confronted with altered configurations of their body showed activation in the right occipital cortex. However, this study did not report data regarding intact self-stimuli. Nonetheless, few self-recognition studies have either reported or found sex differences on self-recognition tasks, and some even control for the sex of the participants in their study so as to maintain uniformity in their sample since the existence of sex differences in self-processing has not been experimentally characterized (Devue et al., 2007; Platek et al., 2006). Future investigations should further consider the impact of sex on self-recognition processing. 4.5. Conclusion Altogether, these findings are among the first to empirically examine the self-assessment of vocal recognition in direct comparison to facial recognition. We have been able to demonstrate: (1) that visual self-recognition processing is superior to auditory self-recognition processing for similar tasks, (2) that a cross-modal exposure to self-stimuli does not facilitate, S.M. Hughes, S.E. Nicholson / Consciousness and Cognition 19 (2010) 1124–1134 1133 but rather slightly inhibits self-recognition, (3) that there may be lateralization effects for vocal self-recognition alone and when combined with visual information (i.e. with there being a left hand/right hemisphere dominance when voices and faces are presented concurrently), and (4) that self-information requires more effort to process than information about others across both the auditory and visual modalities. Future investigations should consider the use of stimuli of those who are familiar as a comparison group to see if that impacts discrimination for self versus other on these auditory and visual tasks. Acknowledgments We wish to thank David Osgood and The Summer Albright College Research Experience Program (ACRE) for their support of this research, Rodney Warfield for his assistance in obtaining participants for this study, Dinna Pich and Bradley C. Rhodes for their assistance in data and stimulus preparation, and also Marissa A. Harrison and Gordon G. Gallup, Jr. for their helpful comments on the manuscript. References Adams, R. B., & Janata, P. (2002). A comparison of neural circuits underlying auditory and visual object categorization. Neuroimage, 16, 361–377. Brady, N., Campbell, M., & Flaherty, M. (2004). My left brain and me: A dissociation in the perception of self and others. Neuropsychologia, 42, 1156–1161. Cook, S., & Wilding, J. (1997). Earwitness testimony 2: Voices, faces and context. Applied Cognitive Psychology, 11, 527–541. Craik, F., Moroz, T., Moscovitch, M., Stuss, D., Winocur, G., Tulving, E., et al (1999). In search of self: A positron emission tomography study. Psychological Science, 10, 26–34. Damjanovic, L., & Hanley, R. J. (2007). Recalling eposodic and semantic information about famous faces and voices. Memory and Cognition, 35, 1205–1210. Devue, C., Collette, F., Balteau, E., Degueldre, C., Luxen, A., Maquet, P., et al (2007). Here I am: The visual correlates of self-recognition. Brain Research, 1143, 169–182. Ellis, H., Jones, D., & Mosdell, N. (1997). Intra- and inter-modal repetition priming of familiar faces and voices. British Journal of Psychology, 88, 143–156. Hanley, J., Smith, S., & Hadfield, J. (1998). I recognize you but I can’t place you: An investigation of familiar-only experiences during tests of voice and face recognition. The Quarterly Journal of Experimental Psychology A: Human Experimental Psychology, 51A, 179–195. Hodges, N., Lyons, J., Cockell, D., & Reed, A. (1997). Hand, space and attentional asymmetries in goal directed manual aiming. Cortex, 33, 251–269. Holzman, P. S., Berger, A., & Rousey, C. (1967). Voice confrontation: A bilingual study. Journal of Personality and Social Psychology, 7, 423–428. Holzman, P. S., & Rousey, C. (1966). The voice as a percept. Journal of Personality and Social Psychology, 4, 79–86. Holzman, P. S., Rousey, C., & Snyder, C. (1966). On listening to one’s own voice. Effects on psychophysiological responses and free associations. Journal of Personality and Social Psychology, 4, 432–441. Hughes, S. M., Dispenza, F., & Gallup, G. G. Jr., (2004). Ratings of voice attractiveness predict sexual behavior and body configuration. Evolution and Human Behavior, 25, 295–304. Hughes, S. M., Harrison, M. A., & Gallup, G. G. Jr., (2002). The sound of symmetry: Voice as a marker of developmental instability. Evolution and Human Behavior, 23, 173–180. Joassin, F., Maurage, P., Bruyer, R., Crommelinck, M., & Campanella, S. (2004). When audition alters vision: An event-related potential study of the crossmodal interactions between faces and voices. Neuroscience Letters, 369, 132–137. Joassin, F., Maurage, P., & Campanella, S. (2008). Perceptual complexity of faces and voices modulates cross-model behavioral facilitation effects. Neuropsychological Trends, 3, 29–44. Kaplan, J. T., Aziz-Zadeh, L., Uddin, L. Q., & Iacoboni, M. (2008). The self across the senses: An fMRI study of self-face and self-voice recognition. Social Cognitive and Affective Neuroscience, 3, 218–223. Keenan, J., McCutcheon, B., Freund, S., Gallup, G. G., Jr., Sanders, G., & Pascual-Leone, A. (1999). Left hand advantage in self-recognition task. Neuropsychologia, 37, 1421–1425. Keenan, J., Nelson, A., O’Connor, M., & Pascual-Leone, A. (2001). Self-recognition and the right hemisphere. Nature, 409, 305. Keenan, J., Wheeler, M., & Ewers, M. (2003). The neural correlates of self-awareness and self-recognition. In T. Kircher & D. Anthony (Eds.), The self in neuroscience and psychiatry (pp. 166–179). New York: Cambridge University Press. Keenan, J., Wheeler, M., Platek, S. M., Landi, G., & Lassonde, M. (2003). Self-face processing in a callosotomy patient. European Journal of Neuroscience, 18, 2391–2395. Kircher, T., Senior, C., Phillips, M. L., Rabe-Hesketh, S., Benson, P., Bullmore, E. T., et al (2001). Recognizing one’s own face. Cognition, 78, 1–15. Kurosaki, M., Shirao, N., Yamashita, H., Okamoto, Y., & Yamawaki, S. (2006). Distorted images of one’s own body activate the prefrontal cortex and limbic/ paralimbic system in young women: A functional magnetic resonance imaging study. Biological Psychiatry, 59, 380–386. Levy, J., Pasheler, H., & Boer, E. (2006). Central interference in driving: Is there any stopping the psychological refractory period? Psychological Science, 17, 228–235. Maurer, D., & Landis, T. (1990). Role of bone conduction in the self-perception of speech. Folia Phoniatrica, 42, 226–229. Molholm, S., Ritter, W., Murray, M. M., Javitt, D. C., Schroedera, C. E., & Foxe, J. J. (2002). Multisensory auditory–visual interactions during early sensory processing in humans: A high-density electrical mapping study. Cognitive Brain Research, 14, 115–128. Molnar-Szakacs, I., Uddin, L. Q., & Iacobonib, M. (2005). Right-hemisphere motor facilitation by self-descriptive personality-trait words. European Journal of Neuroscience, 21, 2000–2006. Morin, A. (2002). Right hemispheric self-awareness: A critical assessment. Consciousness and Cognition, 11, 396–401. Morita, T., Itakura, S., Saito, D., Nakashita, S., Harada, T., Kochiyama, T., et al (2008). The role of the right prefrontal cortex in self-evaluation of the face. A functional magnetic resonance imaging study. Journal of Cognitive Neuroscience, 20, 342–355. Nakamura, K., Kawashima, R., Sugiura, M., Kato, T., Nakamura, A., Hatano, K., et al (2001). Neural substrates for recognition of familiar voices: A PET study. Neuropsychologia, 39, 1047–1054. Olivos, G. (1967). Response delay, psychophysiologic activation, and recognition of one’s own voice. Psychosomatic Medicine, 29, 433–440. Pipitone, R. N., & Gallup, G. G. Jr., (2008). Women’s voice attractiveness varies across the menstrual cycle. Evolution and Behavior, 29, 268–274. Platek, S. M., Loughead, J. W., Gur, R. C., Busch, S., Ruparel, K., Phend, N., et al (2006). Neural substrates for discriminating self-face from personally familiar faces. Human Brain Mapping, 27, 91–98. Platek, S. M., Myers, T. E., Critton, S. R., & Gallup, G. G. Jr., (2003). A left-hand advantage for self-description: The impact of schizotypal personality traits. Schizophrenia Research, 65, 147–151. Platek, S. M., Thomson, J. W., & Gallup, G. G. Jr., (2004). Cross-modal self-recognition: The role of visual, auditory, and olfactory primes. Consciousness and Cognition, 13, 197–210. Proverbio, A. M., Brignone, V., Matarazzo, S., Del Zotto, M., & Zani, A. (2006). Gender differences in hemispheric asymmetry for face processing. BMC Neuroscience, 7, 44–54. Rosa, C., Lassonde, M., Pinard, C., Keenan, J., & Belin, P. (2008). Investigations of hemispheric specialization of self-voice recognition. Brain and Cognition, 68, 204–214. 1134 S.M. Hughes, S.E. Nicholson / Consciousness and Cognition 19 (2010) 1124–1134 Rousey, C., & Holzman, P. S. (1967). Recognition of one’s own voice. Journal of Personality and Social Psychology, 6, 464–466. Sugiura, M., Kawashima, R., Nakamura, K., Okada, K., Kato, T., Nakamura, A., et al (2000). Passive and active recognition of one’s own face. NeuroImage, 11, 36–48. Sugiura, M., Watanabe, J., Maeda, Y., Matsue, Y., Fukuda, H., & Kawashimaa, R. (2005). Cortical mechanisms of visual self-recognition. Neuroimage, 24, 143–149. Turk, D., Heatherton, T., Kelley, W., Funnell, M., Gassaniga, M., & Macrae, C. (2002). Mike or me? Self-recognition in a split-brain patient. Nature Neuroscience, 5, 841–842. Uddin, L. Q., Kaplan, J. T., Molnar-Szakacs, I., Zaidel, E., & Iacobonib, M. (2005). Self-face recognition activates a frontoparietal ‘‘mirror” network in the right hemisphere: An event-related fMRI study. Neuroimage, 25, 926–935. Uddin, L. Q., Rayman, J., & Zaidel, E. (2005). Split-brain reveals separate but equal self-recognition in the two cerebral hemispheres. Consciousness and Cognition, 14, 633–640. Van Lancker, D. R., & Canter, G. J. (1982). Impairment of voice and face recognition in patients with hemispheric damage. Brain and Cognition, 1, 185–195. Walker, S., Bruce, V., & O’Malley, C. (1995). Facial identity and facial speech processing: Familiar faces and voices in the McGurk effect. Perception and Psychophysics, 57, 1124–1133. Weems, S. A., & Zaidel, E. (2005). The effect of response mode on lateralized lexical decision performance. Neuropsychologia, 43, 386–395. Wilentz, J. (1971). The senses of man. New York: Crowell. Zaidel, E., Aboitiz, F., Clarke, J., Kaiser, D., & Matteson, R. (1995). Sex differences in interhemispheric relations for language. In F. L. Kitterle (Ed.), Hemispheric communication: Mechanisms and models (pp. 85–175). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.

Consciousness and Cognition Consciousness and Cognition 14 (2005) 377–389 www.elsevier.com/locate/concog Procedural memory in dissociative identity disorder: When can inter-identity amnesia be truly established?q Rafaële J.C. Huntjensa,*, Albert Postmaa, Liesbeth Woertmana, Onno van der Harta, Madelon L. Petersb a Department of Clinical Psychology, Research Institute for Psychology and Health, Faculty of Social Sciences, Utrecht University, P.O. Box 80.140, 3508 TC, Utrecht, The Netherlands b Department of Clinical, Medical, and Experimental Psychology, Maastricht University, Maastricht, The Netherlands Received 30 January 2003 Available online 25 November 2004 Abstract In a serial reaction time task, procedural memory was examined in Dissociative Identity Disorder (DID). Thirty-one DID patients were tested for inter-identity transfer of procedural learning and their memory performance was compared with 25 normal controls and 25 controls instructed to simulate DID. Results of patients seemed to indicate a pattern of inter-identity amnesia. Simulators, however, were able to mimic a pattern of inter-identity amnesia, rendering the results of patients impossible to interpret as either a pattern of amnesia or a pattern of simulation. It is argued that studies not including DID-simulators or simulation-free memory tasks, should not be taken as evidence for (or against) amnesia in DID.
2004 Elsevier Inc. All rights reserved. Keywords: Procedural memory; DID; Dissociation; Inter-identity amnesia q Albert Postma was supported by a grant from the Netherlands Organization for Fundamental Research (NWO, No. 440-20-000). We thank Paul Knuijt and Willem Verwey for their help in designing the study. We especially thank the patients who participated in the study and the clinicians for their help in gathering this large patient sample and for their assistance in testing. * Corresponding author. Fax: +30 253 4718. E-mail address: R.Huntjens@fss.uu.nl (R.J.C. Huntjens). 1053-8100/$ - see front matter
2004 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2004.10.001 378 R.J.C. Huntjens et al. / Consciousness and Cognition 14 (2005) 377–389 1. Introduction Overactive, underactive, obsessive, or avoidant utilizations of memory characterize numerous psychopathologies (Spiegel, Frischholz, & Spira, 1988). A disorder in which a functional failure of memory is considered to be a core phenomenon is Dissociative Identity Disorder (DID), previously referred to as Multiple Personality Disorder (MPD). In the Diagnostic and Statistical Manual of Mental Disorders, 4th ed. (American Psychiatric Association, 1994), DID is characterized by the presence of two or more distinct identities or personality states, who recurrently take control of the personÕs behavior and who each have their own relatively enduring pattern of perceiving, relating to, and thinking about the environment and self. DID patients very frequently report episodes of inter-identity amnesia, in which an identity claims amnesia for events experienced by other identities (Boon & Draijer, 1993; Coons, Bowman, & Milstein, 1988; Putnam, Guroff, Silberman, Barban, & Post, 1986; Ross et al., 1990; for a review see Gleaves, May, & Cardeña, 2001). However, this does not mean that patients report a dense amnesia between all identities. Different degrees of amnesia may exist between various identities and reported amnesia may either be mutually or one-way, that is, identity A reports awareness of the experiences of identity B, while B reports no knowledge of the experiences of identity A (Ellenberger, 1970; Janet, 1907; Peters, Uyterlinde, Consemulder, & Van der Hart, 1998). Whereas most clinical DID experts agree that DID is accompanied by a disturbance in episodic memory, they seem to disagree as to whether identities share implicit memory, such as priming and procedural memory (cf. Merckelbach, Devilly, & Rassin, 2001), that is, the expression of information without conscious recollection (Schacter, 1987). Putnam (1997), for example, stated that ‘‘fluctuations in the level of basic skills, in habits, and in recall of knowledge are classic forms of memory dysfunction in dissociative patients’’ (p. 82) and ‘‘paradoxically, it seems as if overlearned information and skills are especially susceptible to intermittent failures of memory retrieval’’ (p. 83). On the other hand, Cardeña (2000) stated ‘‘in dissociative amnesia, the individual loses explicit memory for personal experience, while implicit memory for general knowledge, skills, habits, and conditioned responses is unimpaired’’ (p. 57). Six experimental studies have examined implicit memory transfer between identities, most of them focusing on inter-identity priming (Dick-Barnes, Nelson, & Aine, 1987; Eich, Macaulay, Loewenstein, & Dihle, 1997a, Eich, Macaulay, Loewenstein, & Dihle, 1997b; Huntjens et al., 2002; Nissen, Ross, Willingham, Mackenzie, & Schacter, 1988; Peters et al., 1998; for a review see Dorahy, 2001). Priming studies have yielded mixed results, which Eich et al. (1997a) and Nissen et al. (1988) ascribed to the influence of what they called identity-specific factors at the time of encoding and retrieval. In terms of encoding, evidence of amnesia in DID was obtained on conceptually driven tasks that make use of semantically rich materials that they argued was interpreted in different ways by different identities. In contrast, evidence of transfer between identities was obtained on data-driven tasks, in which, according to their reasoning, encoding leaves little room for identity-specific interpretation. In terms of retrieval, transfer of information was obtained on tasks allowing for only a single response on each trial and evidence of amnesia was obtained on tasks allowing a wide range of responses. However, in the most recent study on inter-identity priming in DID, which was performed by our group, we found no objective evidence for inter-identity amnesia on a variety of priming tasks including both conceptually driven and perceptually driven tasks, and both tasks with single and multiple responses (Huntjens et al., 2002). R.J.C. Huntjens et al. / Consciousness and Cognition 14 (2005) 377–389 379 Of the above mentioned, only two studies have included tasks that pertain to the procedural memory system, that is, the memory system that is involved is learning skills and ‘‘knowing how’’ to do things: riding a bicycle, typing words on a keyboard, or solving a jigsaw puzzle (Schacter, 1996). The first study on procedural memory in DID was performed by Dick-Barnes et al. (1987), who used a pursuit-rotor task designed to assess the transfer of perceptual-motor training. Results indicated a practice effect, which is transfer of procedural knowledge learning across the three identities tested. In this study, however, no information was given about the a priori reported amnesia between the participating identities, making the results inapt as a case against inter-identity amnesia. Nissen et al. (1988) performed the second study on procedural memory in DID. Two identities were tested, both reporting amnesia for experiences of the other identity. The authors made use of the serial reaction time (SRT) task introduced by Nissen and Bullemer (1987) that has become a standard task to assess the acquisition and retention of new procedural associations. We will discuss this task in more detail because in the present study we also used a SRT task. Participants are asked to respond as quickly as possible to a stimulus (e.g., a light, an asterisk) that is presented at one of four horizontally aligned locations on a computer screen. Four keys are spatially mapped to the four locations, and participants are asked to press the key in response to the stimulus as fast as possible without making errors. Each response triggers the presentation of the next stimulus, which in turn requires a new response, etc. The critical experimental variation lies in the sequence of stimuli. Subjects respond either to a cyclically repeating sequence (resulting also in a cyclically repeating sequence of responses) or to a random sequence, the constraint being that the same position cannot be used on successive trials. In the Nissen et al. (1988) study, first one identity was given three blocks of trials in a randomsequence condition. Then, the other identity was given four blocks of trials in a 10-trial repeating sequence and a fifth block consisting of a random sequence instead of the repeating sequence. Response time (RT) decreases more when a repeating sequence is presented than when a random sequence is presented, and RT increases when the stimulus presentation switches from a repeating to a random sequence. These sequence-specific RT effects indicate sequential learning. This identity showed some learning of the sequence. Finally, the first identity performed three blocks of the repeating sequence blocks and then one random block. Results indicated this identityÕs performance was facilitated by the other identityÕs acquisition of the sequence. The Nissen et al. (1988) study has some limitations. Similar to the Dick-Barnes et al. (1987) study, only 1 patient was tested. Furthermore, no statistical tests were applied, which makes the interpretation of the data somewhat difficult. The assessment of the degree of the patientÕs learning was also complicated by the omission of a normal control group. Finally, no measures to prevent or detect simulation were included, which seems important given that the so-called ‘‘sociocognitive’’ model considers DID to be a syndrome of social creation or iatrogenesis in the treatment of suggestible individuals (Allen & Movius, 2000; Lilienfeld et al., 1999; Spanos, 1996). Procedural memory is relevant to our everyday functioning because it connects traces of previous experiences to direct motor actions. It is therefore important to establish if and to what extent DID patients suffer limitations (viz., amnesia) in their procedural skill learning ability. The purpose of the present study thus was to systematically investigate procedural memory in DID. 380 R.J.C. Huntjens et al. / Consciousness and Cognition 14 (2005) 377–389 Specifically, we designed an experiment to overcome some of the methodological shortcomings of previous studies, by including a relatively large sample of female DID patients (n = 31) as well as a normal control group comparable on sex, mean age, and education-level (n = 25). Subjects were presented with eight blocks of trials, with the first and the last block containing repetitions of a random sequence and the other blocks containing the same repeating sequence. To diminish the possibility of simulation of inter-identity amnesia by conscious influencing of task performance, we took several measures to discourage explicit memory processing and encourage implicit memory processing. First, following Pascual-Leone, Wasserman, Grafman, and Hallett (1996), we told participants that the location of the stimulus on each successive trial was random and we used a 12-trial instead of a 10-trial sequence to prevent recognition of the repeating sequence of stimuli. For the same reason, we instructed participants to react as accurately, but above all, to react as fast as possible, and we repeated this instruction several times to ensure high-speed performance. Finally, to prevent recognition of the sequence, we used a sequence of stimuli with less statistical structure than the sequence used by Nissen et al. (1988). As statistical structure increases, there are fewer unique runs of trials of a given size, and specific runs are repeated more often. An example of a low structure sequence is BDBCABADAC, in which no run of two or more trials is repeated (Stadler, 1992). Finally, to detect if simulation of inter-identity amnesia indeed was not possible on the task employed, we included a second control group instructed to simulate DID (n = 25). The DID simulators were asked to make up an imaginary, ‘‘amnesic’’ identity and to ‘‘switch’’ upon request to this amnesic identity during the experiment. Controls were expected to show evidence of sequence learning, which would be evident in a decrease in response times in the blocks containing a repeating sequence (blocks 2–7) and an increase in response times when the stimulus presentation switches from a repeating to a random sequence (block 7 vs. 8). Patients as well as simulators were asked to switch to their amnesic identity after the fourth block. In case of inter-identity amnesia, patients were believed not to show evidence of previous exposure to the task, which is learning of the repeating sequence. They were thus expected to show an increase in response times after the switch to their amnesic identity, indicative of ‘‘starting all over again.’’ Because of the measures we took to prevent simulation, simulators were not expected to be able to simulate inter-identity amnesia in their imagined identity. Their scores were thus hypothesized to equal the control scores. 2. Method 2.1. Participants Thirty-one female DID patients participated in the study. These are the same patients who participated in the Huntjens et al. (2002) and Huntjens, Postma, Peters, Woertman, and Van der Hart (2003) papers. Patients were recruited with the help of clinicians in the Netherlands and Belgium. To be eligible for participation, patients had to meet the DSM-IV (1994) criteria and the criteria of the Structured Clinical Interview for DSM-IV Dissociative Disorders (SCID-D), a semi-structured interview used to diagnose the DSM-IV dissociative disorders (Boon & Draijer, 1994; Steinberg, 1993). The mean number of years since diagnosis of DID for patients was 4.42 years (range 3 months–11 years), and DID was always the main reason for patients to be in treatment. R.J.C. Huntjens et al. / Consciousness and Cognition 14 (2005) 377–389 381 Participants were informed that the aim of the study was to understand more about the memory problems often reported by DID patients. Patients self-selected two identities that would participate in the experiment. Borrowing terms prevalent in DID clinical practice, conditions for participation were described as follows: (1) at least one of the identities is completely amnesic for the events experienced by the other participating identity during the experiment; (2) the two identities are able to perform the tasks without interference from other identities; (3) the two identities are able to perform the tasks without spontaneous switches to other identities; (4) the patient is able to switch on request between the two identities. The selected identities could be either of the female or of the male perceived gender type. The switching process was assisted either by the patientsÕ own clinician or by one of the authors (RH or OvdH). The transition was initiated by asking the patient to let an identity ‘‘come forward’’ and take control over the patientÕs consciousness and behavior. Also, the patient was asked to let the other participating identity ‘‘step back,’’ and move out of consciousness. In addition, 50 female control participants participated. Groups were comparable on age and education (see Table 1). Control participants did not report any relevant memory, visual, or attentional problems, or psychiatric disorders. Control participants were divided into two groups, called the ‘‘controls’’ and the ‘‘simulators.’’ Simulators were instructed to imitate DID. They were shown a documentary about a DID-patient and were given additional written information about DID. They were subsequently asked to make up an imaginary, amnesic identity and come up with detailed characteristics of this identity. Following Silberman, Putnam, Weingartner, Braun, and Post (1985), they were given a 17-item data sheet for the identity on which they were asked to assign name, age, sex, physical description, personal history, and personality style. Examination of the completed data sheets confirmed that participants had invested considerable effort in inventing an identity. Finally, they were asked to practice during the week preceding the experiment switching to their new identity and taking on its state of mind. Both the controls and the simulators completed the Dissociative Experiences Scale (DES; Carlson & Putnam, 1993) and the Creative Experiences Questionnaire (CEQ; Merckelbach, Muris, Schmidt, Rassin, & Horselenberg, 1998) (see Table 1). The DES is a 28-item self-report questionnaire with scores ranging from 0 to 100. Scores above 20 or, more conservatively, above 30, are thought to be indicative of Table 1 Participant characteristics for the three groups: DID patients, controls, and simulators DID patients (n = 31) M SD Controls (n = 25) M SD Simulators (n = 25) M SD Age (years) Education DES CEQ 38.48 8.68 5.39 1.20 — — — — 37.72 11.29 5.88 1.13 6.31 4.10 5.48 3.24 32.48 10.31 5.84 1.14 6.54 3.93 4.20 2.58 Note. Education is assessed in categories from 1 (low) to 7 (high) (Verhage, 1964); The DES is the Dissociative Experiences Scale with score range from 0 to 100, and the CEQ is the Creative Experiences Questionnaire with score range from 0 to 25. 382 R.J.C. Huntjens et al. / Consciousness and Cognition 14 (2005) 377–389 pathological dissociation. The CEQ is a 25-item self-report questionnaire with scores ranging from 0 to 25. Scores are thought to be indicative of fantasy proneness, that is, the inclination to be immersed in daydreams and fantasies. The controls and the simulators did not differ significantly on DES-scores and CEQ scores. Neither controls nor simulators showed pathological levels of dissociation as measured by the DES. Written informed consent was obtained from all participants prior to participation. 2.2. Stimuli and apparatus Participants performed a Serial Reaction Time (SRT) task. On each trial, four locations arranged horizontally on a computer monitor were underscored, and a small rectangle appeared above one of them. The stimulus was a yellow character on a black background and 0.5 cm wide · 1 cm high. All four locations were easily discriminable and 5 cm from the bottom of the monitor screen and separated horizontally by 7 cm. Participants responded by pressing the z, x, n, and m keys on the computer keyboard, which was positioned below and in front of the monitor such that the four keys were approximately aligned with the four stimulus locations. The four keys were marked and the z key was the correct key for the leftmost position, the x key for the position second from left, and so on. The stimulus remained on the screen until the participant pressed the correct key, upon which the next stimulus appeared without an inter-stimulus delay. If the subject pressed the incorrect key, the stimulus changed color to gray and the correct key had to be pressed before the next trial was presented. No feedback was given regarding response latency. Each block consisted of 120 trials, which was followed by a short break of 30 s, after which subjects initiated the next block by pressing a key when they were ready. The blocks consisted either of a random sequence, the only constraint being that the same event could not occur on two successive trials, or of an ordered sequence, in which the location of the stimulus followed a particular 12-trials sequence. Designating the four locations A, B, C, and D from left to right, the sequence was as follows: B-D-B-C-A-B-A-D-A-C-D-C. Each block comprised 10 repetitions of this 12-trial sequence, but the end of one 12-trials sequence and the beginning of the next was not marked in any way. Thus, in the absence of knowledge of the sequence itself, each block would seem to be a continuous series of 120 trials. 2.3. Procedure The task was presented in 8 blocks of 120 trials each and two practice blocks of 12 trials, one preceding block 1 and one preceding block 5. Participants were instructed to respond by pressing the key that corresponded to the location in which the stimulus appeared. They responded to locations A, B, C, and D with their left middle, left index, right index, and right middle fingers, respectively, and were asked to rest their fingers lightly on the keys as they performed the task. Subjects were told to respond as accurately and as fast as possible and the instruction to respond as fast as possible was repeated at the beginning of each block. Participants were told that the location of the stimulus on each successive trial was random. However, for all participants, blocks 2–7 followed a repeating sequence, while only blocks 1 and 8 followed a random sequence. Block 1 functioned as a baseline measure of performance. R.J.C. Huntjens et al. / Consciousness and Cognition 14 (2005) 377–389 383 Patients performed a practice block and blocks 1–4 in one identity. After this, they were requested to switch to the identity claiming amnesia for experiences in the present of the identity performing the first series of blocks. The switching process was always accomplished in less than 2 min. When the patient confirmed the presence of the second identity, this identity was directly asked if and what she knew of the learning phase and the material the other identity had seen. Patients answered with either ‘‘Yes’’ or ‘‘No.’’ The identity subsequently performed a practice block and blocks 5–8. So although at this stage, the procedure allows for the acquisition of new associations by identity 2, what is critical is the activation (or not) of existing procedural memory structures learned by identity 1 in the performance of identity 2. Simulating controls performed blocks 1–4 without simulating, after which they received the following instruction: ‘‘You have now performed a task as yourself. We are now asking you to switch to your imagined identity, which will perform the same task you did just now. However, your identity doesnÕt know you have performed the same task so he or she doesnÕt know you saw small blocks on the screen and pressed corresponding keys. Your identity thus has no practice in performing this task. So try to start all over again, at the same speed and with the proportion of errors you responded when you started this task as yourself. Your identity has no other difficulties in performing the task. He or she remembers what he/she does and learns and performs as well as any other person. Your identity just doesnÕt profit from the practice you have had as yourself. Now take a few minutes to let your imagined identity come forward. We will then explain the task to him/her.’’ Subjects then performed blocks 5–8. Normal controls performed all blocks 1–8 including the practice blocks in the same order with a 2-min break after the 4th block to keep the procedure equal. At the end, we questioned participants about the sequence. We asked them whether they had noted a repeating sequence at any point. If they responded positively, we asked them to type the sequence on the keyboard. 3. Results Of the 31 DID patients tested, the three patients that reported some explicit knowledge of the study phase in the test phase, either of the material used or of the instructions given to the other participating identity, were left out of the analyses. Two control participants and one patient were left out of the analyses because of extreme high error scores (mean percentage correct responses lower than 80%). The results described therefore pertain to 27 DID patients, 23 control participants, and 25 simulators. The subjectsÕ mean percentage of correct responses and mean RT were calculated for each block, including only those trials in each block on which the subject responded correctly in the RT measure. Results are presented in Fig. 1 and Table 2. In control subjects, the gradual decrease in mean RT over blocks 2–6 and the increase in RT from blocks 7 to 8 indicated learning of the sequence. Mean RT decreased from 572 ms in block 2 to 453 ms in block 6. Unexpectedly, response times then increased with 9 ms in block 7, possibly reflecting a fatigue effect. As expected, mean response times increased with 52 ms to block 8, when the random sequence was introduced. The mean percentage of correct responses in controls gradually decreased from blocks 2 to 7 (except from blocks 4 to 5, see Table 2) and also decreased from blocks 7 to 8. The decrease in response times compared with the increase in percentage of correct 384 R.J.C. Huntjens et al. / Consciousness and Cognition 14 (2005) 377–389 Fig. 1. Mean response times (in milliseconds) in each block for DID patients (n = 27), controls (n = 23), and simulators (n = 25). Table 2 Mean percentage correct responses in each block for DID patients (n = 27), controls (n = 23), and simulators (n = 25) Blocks 1 2 3 4 5 6 7 8 Group [M (SD)] DID patients Controls Simulators 97.75 (2.93) 95.59 (4.11) 94.57 (3.85) 94.23 (5.66) 96.67 (4.63) 96.48 (3.21) 95.71 (4.88) 93.83 (6.50) 98.37 (1.32) 95.98 (3.58) 95.00 (2.85) 93.44 (3.94) 93.48 (4.91) 92.43 (4.40) 91.70 (5.19) 88.99 (6.15) 97.67 (2.38) 94.03 (4.30) 91.80 (4.33) 89.67 (5.79) 98.23 (1.58) 95.37 (4.00) 93.50 (4.29) 86.83 (8.93) responses in blocks 2–6 is indicative of an accuracy-speed trade-off, that is, participants respond faster to stimuli but trade this increase in speed for a decrease in accuracy. In patients, response times decrease from blocks 2 to 4 with 53 ms. Then, after having made the switch to their imagined amnesic identity, their response times increased with 201 ms, after which they again decreased with 137–668 ms in block 7. Finally, response times again increased with 31 ms from blocks 7 to 8 indicating a learning effect. Mean percentages of correct responses decreased from blocks 2 to 4, then increased after the switch, and again decreased from block 5 onwards. SimulatorsÕ RTs and percentages of correct responses showed a pattern comparable to patients. Their response pattern shows a decrease in response times in blocks 2–4, then an increase from R.J.C. Huntjens et al. / Consciousness and Cognition 14 (2005) 377–389 385 blocks 4 to 5 with 168 ms and again a decrease from blocks 5 to 7. Finally, they also showed an increase from blocks 7 to 8 that is indicative of sequence learning. A 8 Block · 3 Diagnosis group [patients-controls-simulators] MANOVA on the mean response times revealed that there was a significant Block effect F (7, 66) = 32.15, p < .001. Within-subjects contrasts, which compare the mean response times in each block except the first block to the mean response times in the preceding block, revealed that mean response times decreased significantly over blocks (all pÕs < .001). The MANOVA also revealed a significant Block · Diagnosis Group interaction F (14, 134) = 3.97, p < .001. The interaction proved significant only in block 4 vs. block 5 (p < .001), block 5 vs. block 6 (p < .001), and block 6 vs. block 7 (p = .001), the blocks containing a repeating sequence after the switch. While controls thus gave evidence of continuous learning over blocks, patients, and simulators started all over again after their switch to the amnesic identity. The Diagnosis Group main effect was also significant, F (2, 72) = 13.60, p < .001, indicating that diagnosis groups differed significantly in overall mean response times. TukeyÕs Honestly Significant Difference (HSD) pairwise comparison procedures indicated that patients differed significantly from control participants (p < .001), and from simulators (p < .001) with slower responses overall. Controls participants did not differ from simulators (p = .961). A corresponding MANOVA on the mean percentages of correct responses revealed that there was a significant Block effect F (7, 66) = 21.11, p < .001. Within-subjects contrasts revealed that the mean percentage of correct responses significantly decreased over blocks (p 6 .002 for all comparisons). The analysis also revealed a significant Block · Diagnosis Group interaction F (14, 134) = 4.78, p < .001. The Block · Diagnosis Group interaction proved significant only for block 4 vs. block 5 (p < .001), block 5 vs. block 6 (p = .011), and block 7 vs. block 8 (p = .001), the blocks after the ‘‘switch,’’ indicating the difference between the continuous decrease in correct responses of control subjects and the sudden increase in correct responses after the switch for patients and simulators. The Diagnosis Group main effect did not reach significance, F (2, 72) = 3.11, p = .051. To control for a possible accuracy-speed trade-off, we calculated the correlation between the mean response time in each block and the percentage of correct responses in each block for each subject. We then reanalyzed the data only including subjects with correlations 6.50, as indicative for learning free of accuracy-speed tradeoff. Sixteen patients, 13 control subjects, and 13 simulators fulfilled the criterion of a relatively small correlation between response times and percentages of correct responses. Results are presented in Fig. 2 and Table 3. This analysis, however, showed the same pattern of response times, with a significant Block main effect, F (7, 33) = 19.68, p < .001, and a significant Block · Diagnosis Group interaction F (14, 68) = 2.12, p = .021. The interaction proved significant only in block 5 vs. block 6 (p = .011), and block 6 vs. block 7 (p = .015). The interaction from block 4 vs. block 5 did not reach significance (p = .076). Again, the pattern we found was that while controls gave evidence of continuous learning over blocks, patients and simulators did not seem to profit from their previous learning experience with the task in their ÔamnesicÕ identity. 3.1. Awareness of the sequence To the question whether they had noted a repeating sequence at any point, 17 out of 23 controls, 10 out of 25 simulators, and 10 out of 27 patients responded with yes. However, participants 386 R.J.C. Huntjens et al. / Consciousness and Cognition 14 (2005) 377–389 Fig. 2. Mean response times (in milliseconds) in each block for DID patients (n = 16), controls (n = 13), and simulators (n = 13). Table 3 Mean percentage correct responses in each block for DID patients (n = 16), controls (n = 13), and simulators (n = 13) Blocks 1 2 3 4 5 6 7 8 Group [M (SD)] DID patients Controls Simulators 97.66 (1.53) 95.10 (3.73) 93.65 (4.36) 93.02 (6.73) 95.47 (5.52) 95.42 (3.56) 94.69 (5.86) 92.34 (7.83) 98.40 (1.38) 97.12 (2.27) 96.28 (2.09) 94.87 (2.80) 95.71 (1.66) 94.68 (2.84) 94.04 (4.59) 89.87 (5.95) 97.37 (2.86) 94.42 (4.30) 92.82 (3.66) 91.22 (4.31) 97.69 (1.87) 95.13 (4.22) 94.94 (3.70) 89.23 (6.50) were not able to describe the procedure used. They differed very much in the number and designation of blocks they thought consisted of sequences. For example, one participant said she thought every block contained a different sequence and another participant thought the first block contained a sequence, while actually this block consisted of a random sequence. Also, several participants thought the sequence only consisted of 2 or 3 trials that were repeated amongst random trials. Two control participants were able to type in a maximum substring of 6 trials in a row out of the 12-trials sequence in among other incorrect trials. Four controls, 5 simulators, and 3 patients were able to type in a maximum substring of four correct trials in a row; 7 controls, 5 simulators, and 2 patients were able to type in 3 trials in a row; and 4 controls and 5 patients were only able to type in 2 trials. R.J.C. Huntjens et al. / Consciousness and Cognition 14 (2005) 377–389 387 4. Discussion The purpose of this study was to objectively test procedural memory functioning in DID. Results of control subjects in this study showed the expected decrease in response times over blocks containing a repeating sequence and the expected increase in response times when the stimulus presentation switched from a repeating to a random sequence. It is somewhat difficult to establish what exactly was learned in this task due a possible accuracy-speed tradeoff, even in the low correlation group. Rather than revealing the learning of better predictions of the expected stimulus and response in a repeating sequence trial, a distinctive feature of procedural learning, the pattern may reflect the learning of a faster motor response to the stimulus. The results of patients showed they responded slower overall as is evident from their increased response times when compared to normal controls and simulators. Secondly, the results of patients seemed to indicate a pattern of inter-identity amnesia, that is, a decrease in response times after their ‘‘switch’’ to their amnesic identity. However, the most important finding in this study is that despite of their lack of explicit processing of the sequence learned in the SRT task, simulators were able to mimic the patient pattern. The measures we took to promote implicit memory processing, that is, the speeded performance instruction, telling the participants the sequence of the trials was random, the 12-trial sequence instead of a the more usual 10-trial sequence, and the increased statistical structure of the sequence, did result in making most of the participants unaware of the nature of the repeating sequence. And those participants who did report noticing a sequence, did not even come close to typing in the correct sequence. Explicit knowledge of the nature of the repeating sequence was thus often completely absent. Importantly, even without this explicit knowledge, simulators were able to slow down their responses comparable to the pattern of inter-identity amnesia that was explained to them was to be expected in DID. Because of the ability of simulators to mimic inter-identity amnesia, the results of patients cannot be interpreted unambiguously. Their pattern of performance can both indicate inter-identity amnesia or simulation of inter-identity amnesia. In our previous study on implicit memory functioning in DID (Huntjens et al., 2002), we used memory tasks on which simulation by instructed simulators proved impossible. On these implicit memory tasks, no objective evidence of inter-identity amnesia in DID was found. The results of this previous study concur with the two previous studies on procedural memory in DID performed by Dick-Barnes et al. (1987) and Nissen et al. (1988). It would thus be unlikely to expect amnesia on the SRT task employed in this study, also because the SRT task is data-driven and therefore, given the reasoning of Eich et al. (1997a, 1997b) and Nissen et al. (1988), the least expected memory system for amnesia in DID. Speaking against the possibility of amnesia-simulation by patients is a study performed by Eich et al. (1997b), in which simulation of inter-identity amnesia was possible on a picture fragment completion task. On this task, results indicated that patients did not try to simulate inter-identity amnesia. In sum, this study shows that even if measures are taken to reduce or exclude explicit stimulus knowledge, simulation on implicit memory tasks is possible. This conclusion is very important in interpreting results of previous studies and for designing new studies on the subject. Results of all studies on memory in DID not including tasks which are known to be simulation-resistant or not including a control group of DID simulators, cannot be taken as evidence for or against interidentity amnesia in DID. Simply providing statements that simulation is unlikely on the tasks used certainly does not constitute convincing evidence. 388 R.J.C. Huntjens et al. / Consciousness and Cognition 14 (2005) 377–389 Future studies should thus include memory tasks which are simulation-resistant in order to be able to make definite claims about inter-identity amnesia in DID. Furthermore, tasks on which simulation is easy so that a clear simulation profile can be established should be used in future studies to shed light on the question as to whether patients with DID are simulating their reported memory phenomena. The present results indicate that even without awareness of exactly what is learned procedurally, simulation is possible if subjects possess an advanced enough simulation strategy, which is detailed knowledge about the amnesia profile that is expected of patients. References Allen, J. J. B., & Movius, H. L. (2000). The objective assessment of amnesia in dissociative identity disorder using eventrelated potentials. International Journal of Psychophysiology, 38, 21–41. American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: American Psychiatric Association. Boon, S., & Draijer, N. (1993). Multiple personality disorder in the Netherlands: A clinical investigation of 71 patients. American Journal of Psychiatry, 150, 489–494. Boon, S., & Draijer, N. (1994). Gestructureerd Klinisch Interview voor de vaststelling van DSM-IV Dissociatieve Stoornissen (SCID-D) [Structured Clinical Interview for the Diagnosis of DSM-IV Dissociative Disorders (SCIDD)]. Lisse: Swets & Zeitlinger. Cardeña, E. (2000). Dissociative disorders. In A. E. Kazdin (Ed.). Encyclopedia of psychology (Vol. 3, pp. 55–59). Oxford: Oxford University Press. Carlson, E. B., & Putnam, F. W. (1993). An update on the Dissociative Experiences Scale. Dissociation, 6, 16–27. Coons, P. M., Bowman, E. S., & Milstein, V. (1988). Multiple personality disorder: A clinical investigation of 50 cases. Journal of Nervous and Mental Disease, 176, 519–527. Dick-Barnes, M., Nelson, R. O., & Aine, C. J. (1987). Behavioral measures of multiple personality: The case of Margaret. Journal of Behavioral Therapy and Experimental Psychiatry, 18, 229–239. Dorahy, M. J. (2001). Dissociative identity disorder and memory dysfunction: The current state of experimental research, and its future directions. Clinical Psychology Review, 21, 771–795. Eich, E., Macaulay, D., Loewenstein, R. J., & Dihle, P. H. (1997a). Memory, amnesia, and dissociative identity disorder. Psychological Science, 8, 417–422. Eich, E., Macaulay, D., Loewenstein, R. J., & Dihle, P. H. (1997b). Implicit memory, interpersonality amnesia, and dissociative identity disorder: Comparing patients with simulators. In J. D. Read & D. S. Lindsay (Eds.), Recollections of trauma: Scientific research and clinical practice (pp. 469–474). New York: Plenum Press. Ellenberger, H. F. (1970). The discovery of the unconscious. The history and evolution of dynamic psychiatry. New York: Basic Books. Gleaves, D. H., May, M. C., & Cardeña, E. (2001). An examination of the diagnostic validity of dissociative identity disorder. Clinical Psychology Review, 21, 577–608. Huntjens, R. J. C., Postma, A., Hamaker, E. L., Woertman, L., Van der Hart, O., & Peters, M. (2002). Perceptual and conceptual priming in patients with dissociative identity disorder. Memory & Cognition, 30, 1033–1043. Huntjens, R. J. C., Postma, A., Peters, M. L., Woertman, L., & Van der Hart, O. (2003). Interidentity amnesia for neutral, episodic information in dissociative identity disorder. Journal of Abnormal Psychology, 112, 290–297. Janet, P. (1907). The major symptoms of hysteria. London: Macmillan. Lilienfeld, S. O., Lynn, S. J., Kirsch, I., Chaves, J. F., Sarbin, T. R., Ganaway, G. K., & Powell, R. A. (1999). Dissociative identity disorder and the sociocognitive model: Recalling the lessons of the past. Psychological Bulletin, 125, 507–523. Merckelbach, H., Devilly, G. J., & Rassin, E. (2001). Alters in dissociative identity disorder: Metaphors or genuine entities?. Clinical Psychology Review, 22, 1–17. R.J.C. Huntjens et al. / Consciousness and Cognition 14 (2005) 377–389 389 Merckelbach, H., Muris, P., Schmidt, H., Rassin, E., & Horselenberg, R. (1998). De Creatieve Ervaringen Vragenlijst als maat voor ‘‘fantasy proneness’’ [The Creative Experiences Questionnaire (CEQ) as a measure of fantasy proneness]. De Psycholoog, 33, 204–208. Nissen, M. J., & Bullemer, P. (1987). Attentional requirements of learning: Evidence from performance measures. Cognitive Psychology, 19, 1–32. Nissen, M. J., Ross, J. L., Willingham, D. B., Mackenzie, T. B., & Schacter, D. L. (1988). Memory and awareness in a patient with multiple personality disorder. Brain and Cognition, 8, 117–134. Pascual-Leone, A., Wasserman, E. M., Grafman, J., & Hallett, M. (1996). The role of the dorsolateral prefrontal cortex in implicit procedural learning. Experimental Brain Research, 107, 479–485. Peters, M. L., Uyterlinde, S. A., Consemulder, J., & Van der Hart, O. (1998). Apparent amnesia on experimental memory tests in dissociative identity order: An exploratory study. Consciousness and Cognition, 7, 27–41. Putnam, F. W. (1997). Dissociation in children and adolescents. New York: The Guilford Press. Putnam, F. W., Guroff, J. J., Silberman, E. K., Barban, E. K., & Post, R. M. (1986). The clinical phenomenology of multiple personality disorder: 100 recent cases. Journal of Clinical Psychiatry, 47, 285–293. Ross, C. A., Miller, S. D., Reagor, P., Bjornson, L., Fraser, G. A., & Anderson, G. (1990). Structured interview data on 102 cases of multiple personality disorder from four centers. American Journal of Psychiatry, 147, 596–601. Schacter, D. L. (1987). Implicit memory: History and current status. Journal of Experimental Psychology: Learning, Memory and Cognition, 13, 501–518. Schacter, D. L. (1996). Searching for memory: The brain, the mind, and the past. New York: Basic Books. Silberman, E. K., Putnam, F. W., Weingartner, H., Braun, B. G., & Post, R. M. (1985). Dissociative states in multiple personality disorder. Psychiatry Research, 15, 153–160. Spanos, N. P. (1996). Multiple identities and false memories: A sociocognitive perspective. Washington, DC: American Psychological Association. Spiegel, D., Frischholz, E. J., & Spira, J. (1988). Functional disorders of memory. American Psychiatric Press Review of Psychiatry, 12, 747–782. Stadler, M. A. (1992). Statistical structure and implicit learning. Journal of Experimental Psychology: Learning, Memory, and Cognition, 18, 318–327. Steinberg, M. M. D. (1993). Structured Clinical Interview for DSM-IV Dissociative Disorders (SCID-D). Washington, DC: American Psychiatric Press. Verhage, F. (1964). Intelligentie en leeftijd: onderzoek bij Nederlanders van twaalf tot zevenenzeventig jaar [Intelligence and age: Study with Dutch people from age 12 to 77]. Assen: Van Gorcum.

Consciousness and Cognition xxx (2013) xxx–xxx Contents lists available at SciVerse ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Action understanding: How low can you go? Daniel D. Hutto School of Humanities, University of Hertfordshire, de Havilland Campus, Hatfield, Hertfordshire AL10 9AB, United Kingdom a r t i c l e i n f o Article history: Available online xxxx Keywords: Mirror neurons Action understanding Folk psychology Theory of mind Mindreading Enactivism a b s t r a c t This paper begins by reminding the reader of the standard arguments that sceptics offer for doubting that mirror neurons could constitute any kind of action understanding (Section 2). It then outlines the usual response to these sceptical worries made by believers (Section 3). An attempt to put flesh on this idea in terms of what brains understand is critically examined and found wanting (Section 4). The ensuing analysis shows that it is prima facie possible to develop a more tenable account of enactive understanding that would fit the bill (Section 5). However, a second look raises further questions about (A) what mirror neurons target and (B) what such targeting involves (Section 6). Finally it is concluded that while mirror neurons may play a central role in enabling non-mentalistic forms of intersubjective engagement this falls short of action understanding (Section 7). Ó 2013 Elsevier Inc. All rights reserved. 1. Introduction First discovered exactly two decades ago, mirror neurons continue to inspire debate about what function, if any, they might play in social cognition. Literally thousands of papers have devoted speculation to this and related topics in the wave of research in social neuroscience and philosophy of mind that followed in the wake of Gallese and Goldman’s (1998) seminal opinion piece. That paper asked, for the first time, if a possible function of mirror neurons might be to ‘‘detect certain mental states of observed conspecifics . . . as part of, or a precursor to, a more general mind reading ability’’ (p. 494). Years later the focus has shifted. Hence, the question this paper investigates is slightly updated. It asks: Might mirror neuron activity – in its own right – suffice for any kind of action understanding? There are two clearly divided camps of opinion about how this question is to be answered on the current scene. For convenience, let us call these two camps the sceptics – who say ‘No’, and the believers, who say ‘Yes’. Importantly, the difference of opinion separating sceptics and believers over how to answer the question is not rooted in disagreements about the neuroscientific or empirical details about what mirror neurons do or how they do it. Mirror neurons have a unique kind of dual functionality; they fire both when certain end-directed doings of conspecifics are perceived (visually or audibly) and when a perceived observer engages in the same type of activity (Gallese, Fadiga, et al., 1996). Mirror neurons are sensitive to more than the mere kinematic features of simple movements, if movement is narrowly defined as mere changes in a body part’s position. Mirror neurons are selectively sensitive to something more than this – they target doings that are embedded in wider chains of coordinated activity (e.g., reaching, grasping, bringing to the mouth) and that are in the service of particular ends (e.g., eating vs. placing) (Fogassi, Ferrari, et al., 2005). Mirror neurons are ‘set up to be set off’ by such activity and are so in ways that abstract from the specific modality or particular means of achieving a given end. Thus they fire whether the relevant doings are seen (Umilta, Kohler, et al., 2001) or heard (Kohler, Keysers, et al., 2002) or, indeed, even if the motor activity required for achieving a particular end E-mail address: d.d.hutto@herts.ac.uk 1053-8100/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.concog.2013.01.002 Please cite this article in press as: Hutto, D. D. Action understanding: How low can you go? Consciousness and Cognition (2013), http:// dx.doi.org/10.1016/j.concog.2013.01.002 2 D.D. Hutto / Consciousness and Cognition xxx (2013) xxx–xxx involves non-standard sequences of movements (Rochat, Caruana, et al., 2010). This establishes that whatever it is that mirror neurons do they are not mere movement detectors. Both sides in this debate concede this much. Inspired by these findings the believers advance the view that the activity of mirror neurons, by itself, enables or suffices for an understanding of (1) motor acts; (2) why a motor act is done (the goal of the action of which the motor act is part); (3) the intention behind observed motor acts (Fabbri-Destro & Rizzolatti, 2008, see also Becchio, Manera, Sartori, Cavallo, & Castiello, 2012).1 Some believers have even gone so far as to claim that ‘‘the primary function of [mirror neuron] matching is to enable us to immediately understand the meaning of the actions performed by others’’ (Sinigaglia, 2008, pp. 19–20, emphasis added. See also Rizzolatti & Sinigaglia, 2006, p. 124). Sceptics deny all such claims. In what follows, I begin by briefly reminding the reader of the standard arguments that the sceptics offer for doubting that mirror neurons could suffice for or constitute any kind of action understanding (Section 2). I then outline the usual response to these sceptical worries made by the believers, a response which stresses that sceptics have overlooked the possibility of a non-folk psychological kind of action understanding (Section 3). Rizzolatti and Sinigaglia’s (2006) attempt to put flesh on this idea in terms of what brains understand is then critically examined and found wanting (Section 4). The ensuing analysis shows that, with some important adjustments, it is prima facie possible to develop a more tenable account of enactive understanding that would fit the bill and serve the believers needs (Section 5). However, a second look raises further questions about (A) what mirror neurons target and (B) what doing so involves (Section 6). And, finally, worries about what embodied understanding might mean completely scupper the credibility of the believers’ reply. It is concluded that while mirror neurons may play a central role in enabling non-mentalistic forms of intersubjective engagement that this falls short of action understanding, strictly understood (Section 7). 2. The sceptics Sceptics doubt that – whatever else they may do for social cognition, whatever part they might play in enabling it – the kinds of responses to others that mirror neuron activity engenders does not by itself suffice for any kind of action understanding. A number of recent papers conclude this (Borg, 2007; Goldman, 2009; Jacob, 2008). The general consensus amongst sceptics is that mirror neuron activity might play a restricted role in enabling or prompting action understanding but that mirror neuron activity, by itself, cannot suffice for any kind of action understanding. For example, sceptics allow that it is possible that mirror neuron activity might play a part in detecting target behaviours for the attention of mindreading devices, while insisting that action understanding only occurs with the latter capacities are brought to bear. If so mirroring processes might, at most, play a causal role in helping to prompt action understanding by initiating or triggering bona fide mentalizing activity but without qualifying as sufficing for such understanding. The arguments for this sceptical verdict share a common form. They start by assuming that action understanding necessarily involves making mentalistic attributions of contentful attitudes of some kind – i.e. beliefs, desires or their analogues. It is then argued that mirror neuron activity does not – by itself – suffice for any kind of action understanding because such activity lacks some or other aspect that is required for attributing contentful mental states. Thus it is concluded that mirror neuron activity – in lacking the relevant features –necessarily falls short of what is required for action understanding proper. Borg (2007) highlights that even if, thanks to mirror neuron activity, one’s own motor system is naturally sensitive to another’s dispositions to behave in specific ways (for example firing reliably when another picks up a cup and moves it to their lips to drink) that this falls manifestly short of intentional attribution. And this is primarily because, as we learned long ago, there is no way credible way to ‘‘move between descriptions of behaviour and claims about mental states’’ (p. 17). This is the gist of her complaint against the believers. In a similar vein Jacob (2008) argues that neural resonance is not sufficient for divining intentions understood as propositional attitudes or combinations thereof.2 He holds that while mirroring based on neural resonance might succeed in replicating the content of another’s motor intention, in order to do their work the content of motor intentions is necessarily too-fine grained (indeed nonconceptual) to qualify as action-guiding intentional content. This is established by the fact that motor intentions (should there be such things) stand in a one-to-many relation even to the most basic goals and intentions of the sort associated with action-sponsoring propositional attitudes. Jacob (2008) illustrates this indeterminacy worry nicely with the following example: In grasping the red apple with her right hand, was the agent’s goal to eat it? To give it to her little daughter? To throw it away? Or to display it in order to draw it? I can mentally rehearse her reach-to-grasp movement: this might enable me to know what it is like for the agent to grasp and feel the apple within the palm of her right hand. As this example illustrates, the notion of a goal is more abstract than the notion of a target (Jacob, 2008, p. 205). 1 Fleshed out with an example, the claim is that ‘‘There are two distinct series of information that one can get observing an action done by another individual. One is ‘‘what’’ the actor is doing; the other is ‘‘why’’ the actor is doing it. If we see, for example, a girl grasping an apple, we understand that she is grasping an object. Often, we can also understand, in addition, why she is doing it, that is we can understand her intention. We can infer if she is grasping the apple for eating it, or for putting it into a basket’’ (Fabbri-Destro & Rizzolatti, 2008, p. 2055). 2 In an earlier publication Jeannerod and Jacob (2005) argued that mirror neuron activity is not necessary for action understanding, based on the work Heider and Simmel (1944). Please cite this article in press as: Hutto, D. D. Action understanding: How low can you go? Consciousness and Cognition (2013), http:// dx.doi.org/10.1016/j.concog.2013.01.002 D.D. Hutto / Consciousness and Cognition xxx (2013) xxx–xxx 3 With this in mind he identifies the root problem with the idea that mirror neurons might constitute a kind of action understanding as follows: ‘‘there is no unique mapping from the agent’s motor intention to her prior intention, the observer would not thereby be enabled to represent the content of the agent’s prior intention’’ (Jacob, 2008, p. 207). Other related worries have been identified too. According to orthodox thinking, understanding actions and intentions necessarily requires being able to ascribe relevant mental state concepts and contents. Thus ‘‘mindreading consists of attributing (ascribing, imputing) a mental state to someone . . . [where] to attribute a mental state to an individual is to represent that individual as being in that state’’ (Goldman, 2009, p. 235). Thus even if it is imagined that by instantiating the same neural state as another it is possible to replicate another’s motor intention, there would have to be additional grounds for thinking that in doing so contents were being ascribed or attributed to the other on this basis. An immediate problem, just discussed in reviewing Jacob’s worry, is that even if we are prepared to imagine the existence of ‘motor intention’ contents to be ascribed they would be of the wrong sort of grain to enable action understanding. But still more worrying for the believers’ proposal is the fact that the neuroscientific evidence suggests that direct matching by means of mirroring does not involve the further steps that would be required to make an attributional story work. Apparently there are no grounds to suppose that mirror neuron activity itself involves anything like (1) a simulation (no pretence or ‘as if’ modelling) or inference phase, nor (2) any subsequent attribution phase (Gallagher, 2007a, p. 358; Gallagher, 2007b, p. 64; Gallagher & Zahavi, 2008, pp. 178–179). Goldman sums up these combined reasons for scepticism about the believers’ proposal in a single sentence: [I]t looks implausible that motor mirroring events are states of attribution [i.e. representations] (of beliefs) containing mentalistic contents. This explains some of my reasons for doubting the correctness, or fruitfulness, of what I take to be the Parma approach to the issue (Goldman, 2009, p. 238). 3. The believers’ reply Believers are unmoved by these observations. This is because, they insist, in speaking of a pre-reflexive understanding of biological action they are offering ‘‘a radical alternative to the mainstream view endorsed by classical cognitive science and many who work on the philosophy of mind’’ (Gallese, Rochat, Cossu, & Sinigaglia, 2009, p. 104). They identify the ‘main problem’ with the sceptical arguments stemming from mainstream analytic philosophers is that they all mistakenly assume that mindreading is ‘‘the sole and primary way to achieve intentional understanding’’ (Gallese et al., 2009, p. 104). We are assured that believers are completely averse to ‘‘intellectualist and hypermentalist accounts of our primary contact with others . . . [and that] the mirror-based responsiveness to others’ motor goals and intentions can be construed in terms of action understanding without thereby falling into the morass of an unwarranted cognitivism and abstract mentalism . . . [such that] this kind of understanding has nothing to do with the classical picture of mindreading’’ (Sinigaglia, 2009, p. 324, emphasis added). Over and against mainstream sceptics, the believers’ stress that the ‘immediate understanding of the acts of others’ is neither based on, nor is (or involves) any form of mentalizing (Rizzolatti & Sinigaglia, 2006, p. 131). By their lights basic action understanding involves no attribution of mental states or contents. The line of reasoning is clear: if a quite distinctive, nonfolk psychological kind of action understanding exists then it is at least possible that mirror neurons enable it. Very well, but if mirror neuron-based action understanding is completely unlike that of the folk psychological variety then what kind of understanding is it exactly? Believers tell us that this kind of action understanding ‘‘is a pragmatic, pre-conceptual, and pre-linguistic form of understanding’’ (Rizzolatti & Sinigaglia, 2006, p. xi). That’s a start. But the credibility of this initial answer depends on developing a more detailed account of the nature and basis of this distinctive, and most fundamental, kind of non-folk psychological action understanding – and explaining how it can be constituted by the activity of mirror neurons alone. Here the devil is surely in the details. On close scrutiny is there a tenable account of the nature and basis of this special kind of action understanding? 4. The understanding brain? Rizzolatti and Sinigaglia (2006) attempt to develop this proposal by making good on the claim that ‘‘the acting brain is also and above all a brain that understands’’ (p. xi). These authors stress that: Certainly, we can use our higher cognitive faculties to reflect on what we have perceived and infer – the intentions, expectations, or motivations of others that would provide us with a reason for their acts . . . but [1] our brain is able to understand these latter immediately [2] on the basis of our motor competence alone, without the need for any kind of reasoning (Rizzolatti & Sinigaglia, 2006, p. xi). We might wonder: how, exactly, are brains able to understand action? Or more precisely: How is it conceivable that neural activity (even of a very special kind) could, by itself, constitute any kind of action understanding? We are told that: This understanding is completely devoid of any reflexive, conceptual, and/or linguistic mediation as it is based exclusively on the vocabulary of acts and the motor knowledge on which our capacity to act depends (Rizzolatti & Sinigaglia, 2006, p. 125, see also pp. 97–98 for the most minimal reading). Please cite this article in press as: Hutto, D. D. Action understanding: How low can you go? Consciousness and Cognition (2013), http:// dx.doi.org/10.1016/j.concog.2013.01.002 4 D.D. Hutto / Consciousness and Cognition xxx (2013) xxx–xxx Apparently, the answer is to be found in the idea that the brain contains meaningful content concerning possible acts and that it’s the practical, embodied understanding of such acts that is generated by the exploitation of this motor knowledge. It is claimed that ‘‘motor knowledge of our own [possible] acts is a necessary and sufficient condition for an immediate understanding of the meaning of the actions of others’’ (Rizzolatti & Sinigaglia, 2006, p. 106, emphasis added. See also pp. 131, 153, 154). To assess this proposal properly it is necessary to articulate the nature of the core claim in greater detail, in so far as this is possible. Critically important is the idea that the motor knowledge exploited for the purpose of action understanding equates to ‘‘knowledge of motor principles that regulate [possible] actions [i.e. those within an individual’s motor repertoire]’’ (Rizzolatti & Sinigaglia, 2006, p. 100). The vocabulary of motor acts informs the content of these principles (Rizzolatti & Sinigaglia, 2006, p. 46). With this idea in place these authors attempt to make sense of the special properties of F5 visuo-motor or ‘mirror’ neurons by appeal to the idea that the brain operates with its own meaningful language of motor acts. Thus specific assemblies of neurons (or patterns of neural activity) are taken to indicate or stand for ‘goals’ (holding, grasping, etc.) while yet others indicate or stand for the ‘manner’ of the act (precision grip, power grip, etc.) or its ‘temporal segmentation’ (opening and closing, etc.). Assuming all of this, it is then proposed that the responses that mirror neurons sponsor (i.e. firing when an agent performs a certain motor act or observes the same type of motor act performed by a relevant other) have the ‘‘same functional meaning’’ (Rizzolatti & Sinigaglia, 2006, p. 47). This result is secured by the fact that the ‘messages’ sent by such neurons are the ‘same’ whether the subject acts or whether the subject merely watches another acting in the same way.3 So, on this story, it appears that in ‘grasping the meaning’ of an act (when performed by another) the brain calls upon its unique knowledge of motor principles and thereby is able to immediately recognize that the other’s act has the same functional meaning as an act the observer might have performed. And it turns out that this is possible only if the act is one that the observer could perform. Note, crucially, that for this account to be tenable we must be prepared to assume that ‘‘a basic motor knowledge of the meaning of acts [is] coded by the various neurons’’ (Rizzolatti & Sinigaglia, 2006, p. 100, emphasis added). But this assumption is problematic. The idea that neurons literally encode the content of the elements of possible motor acts – i.e. that there is a meaningful vocabulary of such acts – is licensed by the familiar idea that perceptual mechanisms are crypto-analysts: As many contemporary philosophers see it, sensory systems are like decoding machines. An encoded signal is the tangled result of two different variables: the code used by the sender, and the message she wishes to send (Matthen, 2005, pp. 5– 6). Use of the ‘code’ metaphor is rife in neuroscience; indeed in many ways it finds a natural home there. It is therefore not surprising that philosophers often turn to neuroscience for parade cases of its invocation. For example, Thompson writes: A neurobiological example can help illustrate these ideas. Certain kinds of cortical neurons are often described as feature detectors because they respond preferentially (fire above their base rate) to various types of stimuli, such as edges, lines and moving spots . . . Such neurons are said to ‘represent’ features of objects and to make that information available for further processing by various systems in the brain (Thompson, 2007, p. 52).4 Yet despite its widespread popularity the theoretical implications and costs of talk of neural ‘codes’ and ‘representations’ are rarely fully articulated or evaluated. Goldman offers a timely health warning: There is no generally accepted treatment of what it is to be such a mental code, and little if anything has been written about the criteria of sameness or difference for such codes. Nonetheless, it’s a very appealing idea, to which many cognitive scientists subscribe (Goldman, 2012, p. 73). This may not be terribly important for the casual uses to which ‘code’ talk is often put – i.e. where it serves to do no more than highlight that the neural states in question are just reliably correlated with specific worldly features. But when such talk is meant to support serious claims and explanations about the existence of a meaningful neural vocabulary, which in turn is meant to ground a special kind of action understanding, it requires careful scrutiny. A detailed argument against taking such talk seriously is provided in Hutto and Myin (2013). For our purposes it will suffice to consider Ramsey’s critique of the popular tendency to describe states (or ensembles of states) that are reliably caused by, or nomically depend upon, the occurrence of certain external features under specific conditions as ‘representations’. He regards any account of representation that involves no more than this commitment as subscribing to the ‘receptor notion’ of 3 Rizzolatti and Sinigaglia go so far as to claim that when it comes to understanding the behaviour of mirror neurons this is ‘‘the only possible interpretation’’ (2006, p. 47, see also pp. 49–50). 4 Thompson notes that ‘‘According to the received view in cognitive science, in order to explain cognitive abilities we need to appeal to information-bearing states inside the system. Such states, by virtue of the semantic information they carry about the world, qualify as representations . . . representations are internal structures that encode context-independent information about the world, and cognition is the processing of such information’’ (Thompson, 2007, p. 52). He blames the unwarranted popularity of this view on ‘‘a confused metaphorical transference from culture to individual psychology lies at the very origin of the cognitivist view . . . The original model of a computational system was a person – a mathematician or logician manipulating symbols with hands and eyes, and pen and paper’’ (Thompson, 2007, p. 7). Please cite this article in press as: Hutto, D. D. Action understanding: How low can you go? Consciousness and Cognition (2013), http:// dx.doi.org/10.1016/j.concog.2013.01.002 D.D. Hutto / Consciousness and Cognition xxx (2013) xxx–xxx 5 representation. This notion comes into play when scientists speak of states or processes that serve as ‘detectors’ or ‘indicators’ of some external feature or other in precisely the way that Thompson does in the above quotation. Ramsey (2007) argues that the receptor notion fails to meet what he calls the job description challenge, which he takes as criterial for deciding if something qualifies as a representation. Satisfying the job description challenge requires saying not only (1) what determines the content of a state or structure but also, critically, saying (2) how the state or structure serves or functions as a representation in or for a larger system. That is to say to qualify as a representation by these lights ‘‘there needs to be some account of how the structure’s possession of intentional content is (in some way) relevant to what it does in the cognitive system’’ (Ramsey, 2007, p. 27, emphasis added. See also p. xv). This verdict has important and wide-reaching implications, for if it is correct then a great deal of cognitive science research, despite familiar advertisements, really only posits and investigates reliable, on-line causal mediators or causal relays and not structures that can be thought of as representations proper. Certainly with respect to the kind of work mirror neurons might be doing the right conclusion appears to be that: Their actual stated functional role – reliably responding to some external condition – is not, by itself, a role sufficiently representational to justify treating some state or structure as a representation. There are several non-representational internal states that must, in their proper functioning, reliably respond to various states of the world. Our immune system, to take one example, functions in part by consistently reacting to infections and insults to our bodies. Yet no one suggests that any given immune system response (such as the production of antibodies) has the functional role of representing these infections (Ramsey, 2007, p. 125). Such states, at most, ‘‘have the function of causing something to happen in certain conditions’’ (Ramsey, 2007, p. 126). As a result ‘‘one of the most popular ways of thinking about representations in cognitive science is confused and ought to be discontinued’’ (Ramsey, 2007, pp. 148–149). It is important to be clear about what Ramsey is saying here. He is not saying that neuroscience does not tell us something important about the basis of cognition. It is rather that if what it tells us is correct then a great deal of cognitive work in fact gets done without the need for the manipulation of anything that plays the explanatory role of contentful representations. Or to put the point another way: no one denies the existence of reliable correlations between neural states and certain worldly features. Yet the existence of such correlations alone does not provide adequate grounds for thinking that neural states or processes represent those states of affairs, or that they are genuine ‘bearers of content’, or that they otherwise contain or communicate ‘meanings’. This would not follow even if the correlations in question were to take the form of strict nomic dependencies such that the neural states would qualify as properly informational bearing states under the most stringent definition of what ‘carrying information’ requires. Nor would it matter if such neural patterns were the outcome of active learning, such that the neurons or populations of neurons had developed a habit of selectively responding to specific stimuli. These supplements would do nothing to change the status of a neural state (or assembly of such states) that fire reliably into properly representational states with meaningful content. Even augmented in the suggested ways such neural states would still fail to qualify as contentful representations. Certainly we are not justified in simply assuming that there is a pre-existing encrypted message or content to be received and passed along in the brain. Nothing in the details of the standard neuroscientific story, if taken seriously, supports this picture. But if so then the ‘vocabulary’ metaphor, which Rizzolatti and Sinigaglia call on in order to make sense of the claim that mirror neuron firings ‘samesay’ in brainese, is just that – a metaphor. And if so, it cannot do the required work. Or it is meant seriously, but breaks down at the crucial point. Either way this account will not fly. 5. Enactivism to the rescue There might be a way forward. If talk of an ‘understanding brain’ is edited out, a softer reading of the claim that another’s action exhibits ‘motor significance’ for us becomes available. For we could say, with an important adjustment, that our ‘‘motor system goes into ‘resonance mode’ by which we recognize [i.e. the whole organism undergoes a certain kind of response] the intentional aspect of the movements’’ (Rizzolatti & Sinigaglia, 2006, p. 100, see also pp. 97–98). Rather than focusing solely on the brain’s (imagined) capacity to understand action believers could focus on the role mirror neuron activity plays in enabling the whole organism to do certain things. This raises the important question: Perhaps the understanding in question occurs at the organismic level in a more properly enactive way? This offers a very different way of unpacking the idea that mirror neuron activity constitutes a distinctive kind of action understanding that comes before and below capacities for mentalistic attribution and the emergence of folk psychological competence proper. This seems to be something like what Gallese (2001, 2003, 2005, 2007) has in mind when he defends the idea that our primary engagements with others involves a kind of ‘implicit’, embodied, non-mindreading, form of embodied simulation. He maintains that such embodied simulation does not involve: attributing mental states or propositional attitudes; a grasp of the self/other distinction; or inference or interpretation of any sort. Indeed, he tells us that it is ‘‘immediate, automatic, and almost reflexlike’’ (Gallese, 2005, p. 101). An obvious attraction of this sort of approach is that it has the potential to deal with another serious worry concerning the original version of the claim under scrutiny – i.e. that action understanding might somehow reside wholly within the activity of mirror neurons. For once we let go of the idea that the brain might understand the relevant motor acts (using its own Please cite this article in press as: Hutto, D. D. Action understanding: How low can you go? Consciousness and Cognition (2013), http:// dx.doi.org/10.1016/j.concog.2013.01.002 6 D.D. Hutto / Consciousness and Cognition xxx (2013) xxx–xxx special vocabulary) it is hard to make sense of the claim that mirror neuron activity could – in isolation – constitute embodied action ‘understanding’. In contrast, if we opt to focus on what such activity enables organisms to do when interacting with one another it is easier to see the importance of appropriate connections with other parts of the brain and body. If we take the embodiment thesis seriously then the firing of single cells, or even longer chains of mirror neuron inspired activity, would require that yet wider links are forged with an organism’s affective reactions and/or tools for effective coordination (see Dennett (2006) on sequelae, p. 162). This observation becomes particularly pertinent if embodied simulation is to be understood as a hot, emotionally laden form of cognition. Fleshing out this proposal requires going well beyond the strict letter of Rizzolatti and Sinigaglia’s (2006) original formulation of the believers’ credo. For that formulation only speaks of motor competences alone yielding the requisite understanding. Accordingly, the original claim ‘‘doth give . . . no jot of blood; the words expressly are, a pound of [mirror neuron activity]’’ (Shakespeare, The Merchant of Venice, Scene 1, Act IV). Thus the original claim needs to be modified along the following lines: this is a pragmatic, pre-conceptual, and pre-linguistic form of understanding . . .[that] lies at the base of many of our celebrated cognitive abilities . . . This type of understanding is also reflected in [in part enabled by] the activation of the mirror neurons, [inter alia]’’ (Rizzolatti & Sinigaglia, 2006, p. xi, with adjustments). This yields what I take to be at least a prima facie credible way of making good on the idea that ‘‘mentalizing is not our sole or primary way of intentional understanding . . . [i.e. that there is] an ‘enactive’ understanding below and before any mindreading ability’’ (Sinigaglia, 2008, p. 18). And it surely fits with the believers’ insistence that ‘‘enactive understanding is . . . different in nature and content from the modalities of mindreading’’ (Sinigaglia, 2008, p. 30). 6. Further adjustments But even this does not go far enough. Further adjustments are needed. As discussed in Section 3, sceptics give strong reasons for thinking the function of mirror neurons in social cognition cannot be to target folk psychological attitudes such as beliefs and desires. This pushes them to accept a weaker interpretation of what they may target. For those wedded to cognitivism this translates in a deflationary interpretation of the representational properties – or the content – of mirror neuron activity. In this respect Saxe’s (2009) conclusion is fairly typical: ‘‘there is no evidence that mirror neurons represent the internal states of the target (e.g. his intention) rather than external properties of the action’’ (p. 450). There are two questionable assumptions at play here. The first is (A) that if mirror neurons do not function to represent folk psychological attitudes then they must function to represent purely mindless properties that are at best associated with actions. The second is (B) that mirror neurons represent. Lets us consider each of these assumptions in turn. It is widely taken for granted that a denial that mirror neurons target the familiar propositional attitudes – beliefs, desires, intentions and their ilk – entails that they must target something purely mindless and mechanical – e.g. mere or pure physical movement. Saxe, for one, brings this hidden assumption into the light when she writes: Mirror neurones [sic] may contain representations of action sequences that make fine-grained predictions about an unfolding action (Csibra, 2007), but only in terms of the physical movement, not the internal states (or, a fortiori, the propositional attitudes) (Saxe, 2009, p. 450, first emphasis added). Yet elsewhere – over and against this – she also notes that ‘‘[there is] elegant evidence that mirror neurones [sic] represent relatively abstract properties of observed actions, going beyond the currently visible movements’’ (Saxe, 2009, p. 450).5 Since no one seriously denies the elegant evidence to which she directs our attention, this suggests that whatever we decide that mirror neuron activity targets we already have excellent reason to exclude two possibilities – that it targets either (1) mere movements or (2) the familiar propositional attitudes. However a different rendering of what mirror neuron activity reliably targets becomes visible if we rethink basic mentality in terms of purposeful, end-directed engagements of organisms with their environments. Let me reiterate the problem. On the one hand, the sceptics provide good reason to suppose that mirror neurons are not sensitive to what lies behind actions proper, if the notion of action is strictly defined such that it entails a necessary link with propositional attitudes. At best mirror neurons are incidentally tracking contentful attitudes and intentions of the sort associated with folk psychological explanations. Yet, on the other hand, mirror neurons exhibit special sensitivities to end-directed activity of the sort that, if the believers are to be believed, is not reducible to a series of purely mechanical movements. Diagnosis: Perhaps the philosophical debate between the sceptics and the believers is fostered by a failure to recognize that there is another, quite distinct explanandum that might be at play here. It is almost a mantra in analytic philosophical circles that the minimum requirement for something to be an action is that it is connected with (and likely produced by) a contentful mental attitude of some sort or other. Fodor (2008) highlights the standard reason for holding this view. He insists that we have no choice but to accept that ‘‘the ability to think 5 The discharge of mirror neurons is critically dependent on the observation of actions performed in specific contexts. This has been shown in an important series of widely discussed experiments (see Caggiano, Fogassi, Rizzolatti, Thier, & Casile, 2009; Iacoboni et al., 2005). There also is strong evidence which shows that ‘‘the richer our motor repertoire is the sharper our sensitivity to other’s motor goals and motor intentions – our ability to act shapes our experiencing and making sense of others’ behaviour’’ (Sinigaglia, 2009, p. 320). Please cite this article in press as: Hutto, D. D. Action understanding: How low can you go? Consciousness and Cognition (2013), http:// dx.doi.org/10.1016/j.concog.2013.01.002 D.D. Hutto / Consciousness and Cognition xxx (2013) xxx–xxx 7 the kind of thoughts that have truth-values is, in the nature of the case, prior to the ability to plan a course of action. The reason is perfectly transparent: Acting on plans (as opposed to, say, merely behaving reflexively or just thrashing about) requires being able to think about the world’’ (p. 13). In short, if all bona fide acting involves planning, and all bona fide planning involves computing and contentful representation then acting always entails the existence of contentful attitudes. Without a doubt some problems, indeed, perhaps whole classes of problems, are best addressed through planning of the sort that requires the rule-governed manipulation of contentful representations. As linguistic beings, humans are representation mongers and thus regularly adopt this basic strategy to plan action and solve problems. Our cultural heritage provides us with a store of represented knowledge – in many and various formats – that enables us to do so. But it hardly follows that this type of cognitive engagement is the basis of, required for, or indeed suitable for, all sorts of enddirected activity, always and everywhere. Noë (2009) hits the nail on the head, noting that, ‘‘the real problem with the intellectualist picture is that it takes rational deliberation to be the most basic kind of cognitive operation’’ (p. 99, emphasis added). Intellectualism of a wholly unqualified kind – of the sort that assumes the existence of contentful representations wherever there is end-directed activity, across the board – has fallen on hard times of late. It finds only a few stalwart adherents in today’s cognitive sciences. Indeed, if anything has promoted the fortunes of more radical enactive and embodied approaches to cognition, it has been the dismal failure of the standard ‘rules and representation’ approach when it comes to dealing with the most basic forms of intelligent activity. This is the headline-grabbing lesson of recent efforts in robotics and artificial intelligence, which have provided – what appear to be – a series of existence proofs against the idea that basic cognition is always content involving. Pioneering work by Brooks (1991a, 1991b) and Beer (1996), for example, reveals that standard versions of intellectualism are a bad starting point when thinking about how to build robots that actually work. Important lessons have been learnt by paying attention to architectonic requirements of robots that are able to complete quite basic sorts of tasks, such as navigating rooms while avoiding objects or recognizing simple geometrical forms and shapes. Inverting standard intellectualist thinking, Brooks famously built robots that dynamically and frequently sample features of their local environments in order to directly guide their responses. They do so without going through the extra steps of generating and working with contentful descriptions of those environments. Breaking faith with these problematic assumptions of classical cognitive science, Gallese and Sinigaglia (2011) have joined a new wave of theorists in positing action-oriented representations. This is meant to provide a more modest way of making sense of the hypothesized motor commands and plans that actively inform goal-directed activity. Action oriented representations are unlike detached, amodal representations in that they are specially geared – both in their content and formats – to drive specific sorts of actions. They are assumed to be states or processes, local or distributed, whose functional role is to indicate the presence of, and sometimes to ‘stand in’ for, states of affairs specifically in order to guide certain kinds of action. In line with this general idea, Gallese and Sinigaglia (2011) posit bodily-formatted representations encoded in a representational format that is ‘‘associated with characteristic processing profiles. These profiles (motor, viscero-motor and somatosensory) characterize a bodily formatted representation, distinguishing it from, say, a propositional representation, even in the presence of (partially) overlapping content’’ (Gallese & Sinigaglia, 2011, p. 100). This is a decisive step away from the idea that basic actions require the existence of propositional attitudes of the familiar folk psychological sort. Still positing bodilyformatted representations allows one to hold onto the idea that contentful attitudes of another sort is required to make an action an action (as opposed to, say, a mere bodily movement or reflex). Making this sort of adjustment to standard cognitive science and positing action-oriented representations is not the only option. Hutto and Myin (2013) and Hutto (in press) have argued that it is possible that manual activity – including acts of reaching and grasping – can be explained without appeal to the existence of representations of any kind at all. They identify problems for any theory that makes serious appeal to the existence of contentful attitudes that exist prior to and independently of linguistic practices. If those arguments hold up then not only do they motivate more radically enactive and embodied approaches to cognition, they show that such approaches have the capacity to advance well beyond dealing only with the antics of behaviour-based robots and insects and instead enable us to rethink what lies at the basis of much distinctively human and animal cognition. For the purposes of this inquiry, radical enactivism raises the possibility that the cases of manual activity of interest to the believers in this debate are not – if we go by the strict concept of action or even a weakened variant of it – actions proper. Recall that the strict concept of action assumes a constitutive connection between actions and contentful attitudes of some sort. Proponents of the strict concept of action usually conceived of the relevant contentful attitudes in terms of propositional attitudes of the folk psychological sort. But a variation on this idea could appeal to the existence of contentful attitudes understood in terms of action-oriented or body-formatted representations. But note that if radical enactivism is correct then we should not understand basic activity such as reaching and grasping as involving contentful attitudes of any sort at all. If we accept this and the strict concept of action, in either variant, if follows, as Rowlands (2006) observes, that ‘‘most of what we do does not count as action’’ (p. 97). In response to the stipulated criterion that the existence of contentful attitudes of some sort is required for the existence of actions, many philosophers are now prepared to acknowledge the existence of non-intentional doings, motivated activities and/or deeds. For example, Velleman (2000) recognizes the need to: Please cite this article in press as: Hutto, D. D. Action understanding: How low can you go? Consciousness and Cognition (2013), http:// dx.doi.org/10.1016/j.concog.2013.01.002 8 D.D. Hutto / Consciousness and Cognition xxx (2013) xxx–xxx define a category of ungoverned activities, distinct from mere happenings, on the one hand, and from autonomous actions, on the other. This category contains the things one does rather than merely undergoes, but that one somehow fails to regulate in the manner that separates autonomous human action from merely motivated activity (p. 4). On the face of it, the great bulk of animal doings take the form of sophisticated forms of highly coordinated, motivated or end-directed activity that falls short of action, if acting entails the existence of some kind of contentful, even if non-conscious and subpersonal, attitudes. To accept this is to allow there exists unplanned intelligent activity of a sort that is not sponsored by contentful attitudes at all. If mirror neurons are sensitive to activity of that sort then they are sensitive to familiar doings which while they do not qualify as actions per se (according to strict concept of action, in either its original or weakened variant) they are also a far cry from being mere ‘thrashings about’ or ‘reflexive behaviours’. Having clarified what mirror neuron activity targets we can turn to the second point about how it targets what it targets – i.e. whether in doing so it represents its target. Assuming standard cognitivism, Goldman asks: what type of mental state or event is instantiated in the activity of primary interest in the premotor region of an agent’s brain? Presumably, it is motor plans or intentions, i.e. propositional attitudes with contents of the form ‘let [my] effector E perform motor act M with respect to goal-object G’. We can notate the content of such an event as follows <effector E, motor act M, goal-object G>. If something like this construal is right then premotor activation in the execution mode has representational content. But does its representational content include mentalistic content? Apparently not . . . Intends, for example, is not part of its content (Goldman, 2009, p. 238). Goldman supposes that mirror neuron activity lacks mentalistic content (and hence is unsuitable for understanding actions) but he does not assume that it lacks representational content full stop. But adopting the latter stance is surely possible: one might reject the idea that mirror neuron activity has any kind of representational content at all. Indeed, doing so can be motivated by the arguments discussed in Section 4. In today’s climate many still fear that a denial of representationalism entails endorsement of some kind of behaviourism. But cognitivism and behaviourism are, of course, not the only options. On a radical enactivist approach there is absolutely no difficulty in supposing that what is reliably targeted by mirror neuron activity are organismic doings – patterns of activity that do not reduce to mere behaviours or sequences of movements. But such an account differs from cognitivism in denying that the reliable response patterns that mirror neurons have to such doings are contentful representations of those doings. If we adjust our thinking about what mirror neurons target and what such targeting involves along non-representationalist, radically enactive lines it is possible to characterize what it is that mirror neurons do in basic social cognition by understanding their job description as responding to end-directed doings. This revised proposal neatly taps into the insight that embodied engagements with others’ doings has special advantages on the motivational and emotional front. With substantial edits it can be agreed that: primary forms of social cognition . . . [heavily rely] . . . on our first-person abilities for action and emotion (Sinigaglia & Sparaci, 2008, p. 329) the observed actions entail[s] a first-person involvement ‘on the spectator’s part’ (Rizzolatti & Sinigaglia, 2006, p. 138, emphasis added). Such remarks highlight that the ‘observer’ gets in on the act because he or she is already naturally attuned to and viscerally moved by what the other does. For what the other does makes a meaningful or significant difference for the observer’s own emotional reactions and action possibilities. We might say that what the other does transforms or shapes the observer’s possibility space in meaningful ways. In saying this, pace Slors (2010) we need not downplay or ignore any of the special properties of mirror neurons. Rather we should rethink how exactly and why those properties might allow for unique forms of intersubjective engagement.6 7. Without understanding Can’t one embrace enactivism of this sort and support the claim that in playing their part in basic social cognition mirror neurons yield an immediate, embodied understanding of the doings of others? Why shouldn’t we say this? Barresi and Moore (2008) ask: To what extent do we need to understand the mental processes governing our own and other’s actions or can we function socially based on simple mechanisms by which we come to share psychological states, without understanding them? (p. 39). I suspect the answer to this is – not at all. We should be very cautious in characterizing basic forms of intersubjective engagement in terms of understanding. But why say this? After all, at least one believer remarks: 6 Looking at primary forms of social cognition through an enactivist lens of this specific sort has the potential to address a complaint that has been raised against the ‘direct social perception’ interpretation – i.e. that it risks underplaying the ‘remarkable fact’ that these engagements are special precisely because they tap, in a surprising way, into what is common or overlapping between intersubjective partners (Slors, 2010). Please cite this article in press as: Hutto, D. D. Action understanding: How low can you go? Consciousness and Cognition (2013), http:// dx.doi.org/10.1016/j.concog.2013.01.002 D.D. Hutto / Consciousness and Cognition xxx (2013) xxx–xxx 9 I agree with Hutto’s radical enactive approach to intentionality, in particular with his distinction between intentional and propositional attitudes . . . however I do not believe that these arguments can be fruitfully used against the construal of mirror-based making sense of others as action understanding (Sinigaglia, 2009, p. 323). Sinigaglia’s reason for persisting with ‘understanding’ talk is that it helps to differentiate motor contagion and mimicry from the peculiar kind of sensitivity to motor goals provided through direct matching (see Sinigaglia, 2009, p. 323). But this concern can be dealt with – not by fighting for retention of a word7 – but by providing an appropriate and detailed taxonomy of the different kinds of responsiveness in question along with giving details about the special aspects of the kinds of intersubjective engagements that mirror neuron activity supports. Such a strategy can do justice to both phenomenology and neuroscience without assuming that the special kind of responsiveness that mirror neurons engender takes the form of any kind of distinctive, mirror-based embodied understanding. Thus if there is no other compelling reason on the table than this, then there seems no good reason to think that mirror neuron activity constitutes a distinctive form of embodied understanding as opposed to a very special variety of embodied responsiveness to the doing of others.8 References Barresi, J., & Moore, C. (2008). The neuroscience of social understanding. In J. Zlatev, T. Racine, C. Sinha, & E. Itkonen (Eds.), The shared mind: Perspectives on intersubjectivity (pp. 39–66). Amsterdam/Philadelphia: John Benjamins. Becchio, C., Manera, V., Sartori, L., Cavallo, A., & Castiello, U. (2012). Grasping intentions: From thought experiments to empirical evidence. Frontiers in Human Neuroscience, 6, 117. Beer, R. D. (1996). Toward the evolution of dynamical neural networks for minimally cognitive behavior. In P. Maes, M. Mataric, J. Meyer, J. Pollack, & S. Wilson (Eds.), From animals to animats 4: Proceedings of the fourth international conference on simulation of adaptive behavior (pp. 421–429). Cambridge, MA: MIT Press. Borg, E. (2007). If mirror neurons are the answer, what was the question? Journal of Consciousness Studies, 14, 5–19. Brooks, R. (1991a). New approaches to robotics. Science, 253, 1227–1232. Brooks, R. (1991b). Intelligence without representation. Artificial Intelligence, 47, 139–159. Caggiano, V., Fogassi, L., Rizzolatti, G., Thier, P., & Casile, A. (2009). Mirror neurons differentially encode the peripersonal and extrapersonal space of monkeys. Science, 324, 403–406. Csibra, G. (2007). Action mirroring and action interpretation: An alternative account. In P. Haggard, Y. Rosetti, & M. Kawato (Eds.), Sensorimotor Foundations of Higher Cognition, Attention and Performance XXII (pp. 435–459). Oxford: Oxford University Press. Dennett, D. C. (2006). Sweet dreams: Philosophical obstacles to a science of consciousness. Cambridge, MA: MIT Press. Fabbri-Destro, M., & Rizzolatti, G. (2008). Mirror neurons and mirror systems in monkeys and humans. Physiology (Bethesda), 23(17), 1–9. Fodor, J. A. (2008). LOT 2: The language of thought revisited. Oxford: Oxford University Press. Fogassi, L., Ferrari, P., Gesierich, B., Rozzi, S., Chersi, F., & Rizzolatti, G. (2005). Parietal lobe: From action organization to intention understanding. Science, 308(5722), 662–667. Gallagher, S. (2007a). Simulation trouble. Social Neuroscience, 2, 353–365. Gallagher, S. (2007b). Logical and phenomenological arguments against simulation theory. In D. D. Hutto & M. Ratcliffe (Eds.), Folk psychology re-assessed (pp. 63–78). Dordrecht: Springer. Gallagher, S., & Zahavi, D. (2008). The phenomenological mind: An introduction to philosophy of mind and cognitive science. London: Routledge. Gallese, V. (2001). The ‘shared manifold’ hypothesis: From mirror neurons to empathy. Journal of Consciousness Studies, 8, 33–50. Gallese, V. (2003). The manifold nature of interpersonal relations: The quest for a common mechanism. Philosophical Transactions of the Royal Society of London, 358, 517–528. Gallese, V. (2005). ‘Being like me’: Self-other identity, mirror neurons and empathy. In S. Hurley & N. Chater (Eds.). Perspectives on imitation: From neuroscience to social science. Mechanism of imitation and imitation in animals (Vol. 1). Cambridge, MA: MIT Press. Gallese, V. (2007). Before and below ‘theory of mind’: Embodied simulation and the neural correlates of social cognition. Philosophical Transactions of the Royal Society Biological Sciences, 362, 659–669. Gallese, V., Fadiga, L., Fogassi, L., & Rizzolatti, G. (1996). Action recognition in the premotor cortex. Brain, 119, 593–609. Gallese, V., & Goldman, A. (1998). Mirror neurons and the simulation theory of mind-reading. Trends in Cognitive Sciences, 2(12), 493–501. Gallese, V., Rochat, M., Cossu, G., & Sinigaglia, C. (2009). Motor cognition and its role in the phylogeny and ontogeny of action understanding. Developmental Psychology, 45, 103–113. Gallese, V., & Sinigaglia, C. (2011). What is so special about embodied simulation? Trends in Cognitive Sciences, 15(11), 512–519. Goldman, A. I. (2009). Mirroring, simulating and mindreading. Mind and Language, 24, 235–252. Goldman, A. I. (2012). A moderate approach to embodied cognitive science. Review of Philosophy and Psychology, 3(1), 71–88. Heider, F., & Simmel, M. (1944). An experimental study of apparent behaviour. American Journal of Psychology, 57(2), 243–259. Hutto, D. D., (2013). Radically enactive cognition in our grasp. In Z. Radman (Ed.), The hand, an organ of the mind: What the manual tells the mental. Cambridge, MA: MIT Press. 7 Besides, just which notion of understanding is at stake? At one point Sinigaglia lists – with rhetorical flair – the relevant questions needed to address this, but he does not provide any answers. ‘‘But what do we really mean when we talk about the ‘understanding’ and the ‘meaning’ of an action? What kind of understanding can be associated with the direct matching mechanism of the mirror neurons? What does such a mechanism tell us about the meaning – i.e. the ‘intentional content’ of an action? Is it not misleading to resort to an intentional language by using terms such as ‘understanding’, ‘meaning’ and ‘content’?’’ (Sinigaglia, 2008, p. 20) Until believers’ clarify what they mean by claiming that mirror neuron activity constitutes a form of ‘understanding’ there is little point in arguing the toss. 8 An anonymous reviewer of this paper suggests that the understanding in question might be a kind of pragmatic (and not an intellectually grounded) understanding of an action that should be explicated in terms of knowing how to performed it. Accordingly, knowing how to ride a bike affords an understanding of someone’s else riding a bike – i.e. one knows how it is possible for them to do it. The idea is that just as in one’s own case such knowledge does not rest on having a theory of bike riding or even a way to explain how to ride a bike. The know how in question is meant to take the form or a pragmatic, nonlinguistic grasp of what certain actions require and the suggestion is that that this sort of knowledge can apply, vicariously, the actions of others too. While the invocation of know how in this context is intuitively attracted if this proposal is to be taken seriously it needs more development. There is a need to work out precisely what the nature of such knowledge amounts to and what kind of understanding it supplies. For example, does knowing how to X really afford an understanding of how it is possible to X? Are these notions equivalent? Moreover, what evidence suggests that mirror neuron activity affords knowledge of ‘how it is possible’ to grasp items or to reach for them as opposed to merely being specially sensitive and responsive to others’ reaching and grasping of items? Please cite this article in press as: Hutto, D. D. Action understanding: How low can you go? Consciousness and Cognition (2013), http:// dx.doi.org/10.1016/j.concog.2013.01.002 10 D.D. Hutto / Consciousness and Cognition xxx (2013) xxx–xxx Hutto, D. D., & Myin, E. (2013). Radicalizing enactivism: Basic minds without content. Cambridge, MA: MIT Press. Iacoboni, M., Molnar-Szakacs, I., Gallese, V., Buccino, G., Mazziotta, J. C., & Rizzolatti, G. (2005). Grasping the intentions of others with one’s own mirror neuron systems. PloS Biology, 3, 529–535. Jacob, P. (2008). What do mirror neurons contribute to human social cognition? Mind and Language, 23, 190–223. Jeannerod, M., & Jacob, P. (2005). The motor theory of social cognition: A critique. Trends in Cognitive Sciences, 9, 21–25. Kohler, E., Keysers, C., Umilta, M. A., Fogassi, L., & Gallese, V. (2002). Hearing sounds, understanding actions: Action representation in mirror neurons. Science, 297(5582), 846–848. Matthen, M. (2005). Seeing, doing and knowing: A philosophical theory of sense perception. Oxford: Oxford University Press. Noë, A. (2009). Out of our heads. New York: Hill and Wang. Ramsey, W. M. (2007). Representation reconsidered. Cambridge: Cambridge University Press. Rizzolatti, G., & Sinigaglia, C. (2006). Mirrors in the brain: How our minds share actions and emotions. Oxford: Oxford University Press. Rochat, M. J., Caruana, F., Jezzini, A., Escola, L., Intskirveli, I., Grammont, F., et al (2010). Responses of mirror neurons in area F5 to hand and tool grasping observation. Experimental Brain Research, 204(4), 605–616. Rowlands, M. (2006). Body language. Cambridge, MA: MIT Press. Saxe, R. (2009). The neural evidence for simulation is weaker than I think you think it is. Philosophical Studies, 144, 447–456. Sinigaglia, C. (2008). Mirror neurons: This is the question. Journal of Consciousness Studies, 15, 70–92. Sinigaglia, C. (2009). Mirror in action. Journal of Consciousness Studies, 16, 309–334. Sinigaglia, C., & Sparaci, L. (2008). The mirror roots of social cognition. Acta Philosophica, 2, 307–330. Slors, M. (2010). Neural resonance: Between implicit simulation and social perception. Phenomenology and the Cognitive Sciences, 9(3), 437–458. Thompson, E. (2007). Mind in life: Biology, phenomenology, and the sciences of the mind. Cambridge, MA: Harvard University Press. Umilta, M. A., Kohler, E., Gallese, V., Fogassi, L., Fadiga, L., Keysers, C., et al (2001). I know what you are doing: A neurophysiological study. Neuron, 31, 155–165. Velleman, J. D. (2000). The possibility of practical reason. Oxford: Oxford University Press. Please cite this article in press as: Hutto, D. D. Action understanding: How low can you go? Consciousness and Cognition (2013), http:// dx.doi.org/10.1016/j.concog.2013.01.002

Consciousness and Cognition Consciousness and Cognition 14 (2005) 22–29 www.elsevier.com/locate/concog Guest Editorial Toward a science of ultimate concern The existence of consciousness in other animals has remained a contentious issue in science and medicine ever since Descartes brought us the ontological mischief of mind–body dualism. An extreme, albeit implicit, variant of this mode of thought exists in animal behavior research to this day. Animals continue to be granted bodily reflexes, a complex instinctual apparatus and fundamental learning abilities, but rarely experiential states (Rollin, 1998). Although such views have traditionally brought peace of mind to those who pursue a livelihood through the use of animals (Thomas, 1996), they should have provided little solace to those interested in the nature of the brain or the mind. Although we all appreciate that mental processes remain relatively invisible to the tools of science, they are no more so than electrons and gravity to the tools of physics. Behavioral scientists are simply more timid in accepting the weight of evidence and theoretical inference as a basis for scientific reasoning. Also, the neglect of consciousness is partly due to the fact that the emphasis in the field is so heavily weighted toward research on cognitive topics, especially learning and memory, where information-processing metaphors seem to suffice. Those perspectives begin to falter when it comes to the epigenetically integrated intrinsic tools for living that evolution provided organisms, such as basic emotional and other affective abilities. Evolution surely provided more than learning capacities to organisms so they could begin to coherently navigate and respond to the world before they had an appreciation of the contingencies of the environment. Primary-process consciousness may be closely linked to some of these unconditioned responses, some of which, such as vocalizations can be used as self-reports of affect in animals (Knutson, Burgdorf, & Panksepp, 2002; Panksepp, Knutson, & Buradorf, 2002). On the basis of first principles, it is quite unlikely that consciousness is a learned feature of brain organization. It could well be that the cerebral sources of conscious experience are more critically linked to the evolved aspects of brain dynamics—intrinsic value guides for existence— rather than to the abundant learning that proceeds rather automatically and unconsciously once organisms begin to confront real-life contingencies. After all, human minds dwell readily on, and seek out, aspects of the world that are critically linked to survival issues. How do animals ‘‘know,’’ in highly variable environments, that certain aspects of the world sustain life, while others detract therefrom? Might they not be ‘‘agents’’ of actions, as opposed ‘‘scarecrows’’ responding to world events? The simple answer has always been that human animals are utilitarian, because 1053-8100/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2004.10.002 Guest Editorial / Consciousness and Cognition 14 (2005) 22–29 23 life-affirming events feel good, and life-detracting events, feel bad. As utilitarian Jeremy Bentham said in 1789, ‘‘utility’’ reflects ‘‘that property in any object, whereby it tends to produce benefit, advantage, pleasure, good, or happiness . . . or . . . to prevent the happening of mischief, pain, evil, or unhappiness’’ (Introduction to the Principles of Morals and Legislation). Of course, now we know that these feelings are not properties of external objects, but the evolved nature of our brains. Thus, a more sophisticated answer is that we and our fellow animals are inheritors of various tools for living that neurobehaviorism, the now prevailing school of thought in functional neuroscience, is still too timid to conceptualize (Panksepp, 1990). In any event, a key issue about animal consciousness is whether other mammals experience emotional and motivational values in ways not dissimilar our own diversity of basic affective feelings. Except for studies of fear conditioning, there has emerged little institutional devotion to the study of brain mechanisms relevant for understanding affective processes in non-human animals. However, even in well-funded areas such as fear conditioning, there is scarcely an investigator who dares explicitly address the ever present worry—do animals experience fear? In the current intellectual climate (hopefully the tail-end of Cartesian dualism), it still does not pay to consider such issues (for critique, see Panksepp, 2002). However, if such brain processes do exist, we may never understand the sources of behavior until we begin explicitly to consider such options and then to study the relevant brain processes with more devotion than ever emerged in the past century. It would be a lonely and peculiar world if we humans were the only conscious species on the face of the earth. Despite recent efforts to recognize cognitive variants of consciousness in animals (Griffin, 2001), all too many behavioral experts are still unwilling to explicitly acknowledge the high probability that many other animals are, in fact, conscious beings and thereby to promote experimental work on such topics. This is understandable from an epistemological perspective—taking sides on questions that cannot be easily resolved empirically is not an attractive option for scientists to consider. It is generally deemed wiser to wrap oneself in a cloak of agnosticism—an ontologically superficial stance that often makes us behavioral scientists seem more foolish than we are, especially in the eyes of many intelligent animal lovers whose opinions may better reflect natureÕs ways—individuals who are not about to back down on such issues of ultimate concern (Bekoff, 2000). Of course, as our body of substantive knowledge increases, all ontological positions need to be re-molded by accruing empirical advances. At this point, I adhere to a dual-aspect monism perspective (certain aspects of mind and instinctual behavior are opposite sides of the same neural coin), for that is the only way not to be immoblized in the animal consciousness laboratory. Here, I will argue that we may already have reached a point in our intellectual history where the denial of consciousness in animals is as improbable as the pre-scientific anthropocentric view that the sun revolves around the earth or that the ‘‘soul’’ is something other than a neurobiological process. However, too many scientists across too many generations have too willingly assumed a mantle of disbelief about such matters (without contemplating all the available evidence), so we may now have as much of a social as a scientific problem on our hands. It would be good if there were a solid basis for a new and shared consensus on this contentious topic, for that could facilitate many new and important research initiatives. My goal here is to coax the field at large to consider emerging scientific ways to conceptualize the core emotions and affective experiences of other animals, and thereby humans as well. There is now enough sound argumentation and abundant data to allow all reasonable behaviorists to remove their shrouds of agnosticism, and to provisionally accept the default position that has long seemed 24 Guest Editorial / Consciousness and Cognition 14 (2005) 22–29 evident to many thoughtful observers who are not as constrained by the rigors of experimental evidence as scientists need to be: namely, it is evolutionarily more coherent to entertain the working hypothesis that all other mammals (and probably many other creatures as well) do have experiential states that help guide their behavior than to work from the premise that they do not. This could open up the field to new types of empirical inquiries, some of them unimagined through most of the 20th century (e.g., Panksepp & Burgdorf, 1999, 2003). Classical behaviorism, still very much alive in behavioral neuroscience, chose to assert that there was nothing to weigh and measure in the realm of mental constructs, and hence chose to actively disregard such spooky, and supposedly superfluous, matters (Panksepp, 1990, 2000). Some impaled themselves on the horns of the dilemma of how immaterial mind processes might ever control physical processes without contemplating and affirming that mind is a complex physiochemical process of wide-scale brain–body neurodynamics. This is a pity, for a detailed understanding of primary process consciousness in humans may only be obtained by studying the relevant brain mechanisms in other animals where the necessary neuroscientific work can be conducted in sufficient detail. Contrary to what used to be asserted rather too boldly during the classic behaviorist era, there is now an emerging consensus that the ‘‘black box’’ must be probed to understand behavior. However, there is still widespread denial that a study of psychological processes in animals is an aspect of nature that needs to be opened up for scientific inquiry and discussion. In the arena of drug addiction, where self-report measures in animals are emerging (Panksepp et al., 2002) we might well ask ‘‘Would individuals exhibit addictive behaviors if there were no affective payoffs?’’ (Panksepp, Nocjar, Burgdorf, Panksepp, & Huber, 2004, p. 93). My own answer to that question is ‘‘Clearly the answer is no—and not just for our own species.’’ Considering that flies react to such drugs, and even planaria and crayfish prefer places where they received cocaine and amphetamine (see Panksepp & Huber, 2004), we must now contemplate how widely affective experience may be distributed in animal life. To not actively consider such possibilities, only sustains a neo-dualistic chasm between our increasingly rich neuro-behavioral work on the brain substrates of animal responses and/or actions and the mental products of such neurodynamics within human lives. Credible bridging principles are possible, and hopefully they will become ever more welcome, as we emerge, ever so slowly, from the century-long denial of neuro-mental faculties in 20th century animal brain research. There are, of course, potential scientific as well as social costs and benefits involved with retaining or discarding the cloak of agnosticism. With sustained agnosticism we can avoid difficult ambiguities concerning the nature of mental realities, perhaps ultimately unknowable to us in any fine detail. However, by being too religiously agnostic, we may also seal doors to the discovery of relevant new knowledge about how the brain/mind is truly organized. On the other hand, by shelving traditional agnosticism and provisionally accepting common wisdom—that animals do have experiential feelings of their own—we may open up a PandoraÕs box of confusions that could be hard to close, even as we are again accepted by the broader society as defenders of naturalistic wisdom as opposed to being regarded, too commonly, as purveyors of pretentious academic positions. One obvious advantage of agnosticism is that the stance can be efficiently deployed to keep animal rights advocates at bay (and this may currently be the biggest implicit reason for equivocating about animal emotions . . . especially in laboratories that stress their animals in ways that would be deemed morally reprehensible in humans). Guest Editorial / Consciousness and Cognition 14 (2005) 22–29 25 However, the biggest problem with agnosticism is that if the brains of other animals are built around survival concerns that are encoded in affects, as our own mental apparatus seems to be, then we may never really understand how our own brains operate without a forthright confrontation with the neuro-evolutionary mystery of affective consciousness in other animals. It should be obvious that the details of the relevant biological mechanisms cannot be worked out in humans. Surely it would be an intellectual tragedy, indeed perhaps a continuance thereof, if neurobehavioral scientists were the last among concerned observers—among a highly diverse public interested in such topics—to recognize emotional experiences in the lives of other animals . . . if, in fact, they do exist. In my estimation, the weight has long favored the conclusion that many other animals do have various affective experiences that resemble our own, rather than the conclusion that they do not. The reason such naturalistic categories and dimensions of neuro-mental existence have never been adequately discussed or experimentally considered in neuroscience, is a fear of being tarred with the brush of anthropomorphism. But evolutionary theory suggests that our own unique mental faculties were surely built upon ancestral ones that existed before human walked the face of the earth (Panksepp & Panksepp, 2000), and scientific anthropomorphism should become increasingly useful for guiding work as we identify homologous structure-function relationships in mammalian brains (Panksepp, 2003a). Indeed, far-fetched as it may seem, it is certainly possible that our first-order sensory-perceptual consciousness was built upon the core values that were encoded as basic emotions and motivations at some earlier point in neural evolution. As highlighted by this special issue of Consciousness and Cognition, we have again reached a time when it may be wise to re-weigh all relevant arguments and empirical approaches. For my own part, I will extend my discussion of affective consciousness in animals, which is, in my estimation, easier to understand than more cognitive variants of consciousness (Panksepp, 2003b)—easier because such state processes of the brain have massive circuitries and neurochemical codes that can be increasingly evaluated for affective properties in human beings (Panksepp & Harro, 2004). I also believe that cross-species behavioral brain research of this type is the optimal way to decode the neuro-evolutionary foundations of human consciousness, especially its basic affective and motivational underpinnings (Damasio, 1999; Panksepp, 1982; Panksepp, 1998). Unfortunately, the re-emerging tendency to make hard discriminations between emotional behaviors (which animals certainly have) and emotional feelings (which, according to some, only humans may have), is a strategy advanced by investigators of human emotions who may not wish to get bogged down on the topic of animal consciousness (e.g., Damasio, 1999, 2003; Dolan, 2002). Those who pursue animal neurobehavioral research to gain insight into the underpinnings of human mind cannot avoid such issues as readily (Panksepp, 2003b, 2003c). For substantive future progress, we must attempt to deconstruct the conceptual prison–house that ‘‘never-mind’’ behaviorism created for the field (e.g., Panksepp, 1990, 2000). In my estimation, a scientifically coherent liberalization of our strictures against psychological analyzes of animal brain functions, at an institutional level (e.g., NSF and NIH), is long overdue. This could be achieved by permitting—even encouraging—investigators to discuss their results not only in traditional positivistic behavioral but also basic psychobiological ways (a proposal that is elaborated in my main contribution to the present issue–see Panksepp, 2005). Indeed, better answers to such 26 Guest Editorial / Consciousness and Cognition 14 (2005) 22–29 questions, in animal models, are essential ingredients for a scientifically coherent biological psychiatry (Panksepp, 2004). Animal brain research, properly conducted, has the best chance to inform us of the deep neural under pinnings of our own affective nature—those all important ‘‘ancestral voices of our genes’’ that may emerge from our genetically guided brain organizations interacting with the diverse life-sustaining affordances and life-detracting dangers of the world. Because of advances in neuroscience, we can finally effectively triangulate among cross-species studies of brain, behavior, and mind, and make substantive progress on such topics. Since there is remarkably little ongoing work using such strategies, my aim is to promote enthusiasm for an interdisciplinary approach whose power remains underutilized and potential underestimated. My hope is that future generations of mind scientists develop the conceptual flexibility to discuss emotional and affective issues more openly than previous ones and to promote research on questions that have commonly been scorned in the behavioral community—the nature of those core biological values that emerge from ancient subcortical neural systems of mammalian brains. This obviously cannot happen until the extended community of mind scientists accepts, as axiomatic, a new, but not so radical, view of neuro-mental nature—that our affective experiences are deeply grounded in various ancient state-control process of the brain that we share with other animals, because of the evolutionary journey we shared in deep-time. Contrary to what generations of behaviorists have insisted, the topic of consciousness may not be as irrelevant for understanding animal behavior as is commonly still assumed. Since proof-positive is impossible to obtain in this as in all such incredibly difficult areas of scientific knowledge, it is critical to weigh the evidence for various reasonable theoretical possibilities that are linked to workable epistemological strategies. My own goal has been to reveal the neural infrastructure of basic affective-emotional experiences in humans and other animals, a topic that has traditionally been derided in behavioral science and deemed too complex to be tackled by cognitive science (Norman, 1980; Panksepp, 1988). Although modern brain imaging has been rapidly changing that bias (e.g., Murphy, Nimmo-Smith, & Lawrence, 2003; Phan, Wager, Taylor, & Liberzon, 2002), such visually entrancing correlational approaches are only a preliminary step toward a more causal analysis, much of which is still best effected by behavioral brain research in other animals. Although there are robust strategies for pursuing the causal correlates of cognitive consciousness (Baars, Ramsoy, & Laureys, 2003), other strategies need to be taken to decipher affective consciousness (perhaps as described in my other contribution to this issue). Cross-species, experimental analysis, where key brain variables are evaluated, provides an approach that is not as full of false negatives and misleading ‘‘neuro-echos’’ as is modern fMRI. The interweaving of animal neurobehavioral and human neuropsychological research permits substantive dynamic analyses of the core emotions and their associated feelings. It is the animal research that is revealing many experimentally manipulable contributory variables (i.e., neurochemistries) that control basic emotional and motivational processes. There are now many pharmacological agents waiting in the wings to be evaluated for their affect modulating properties (Roques, 2000). All need to be studied experientially in humans using methodologies that take first-person subjective experience seriously (Panksepp, 1999). Without coherent cross-species conceptual bridges between work on brain, mind, and behavioral functions, such practical ‘‘discovery’’ work will continue to be delayed. Psychologically informed behavioral brain research in animals can guide hypothesis-driven introspective studies of Guest Editorial / Consciousness and Cognition 14 (2005) 22–29 27 human affective experiences (Panksepp & Harro, 2004). Thereby we may also get glimmers of the even more well-hidden subjective experiences of animals, especially in the realm of basic emotions and motivations where subcortical neurochemical homologies abound both at neuroanatomical and psychological levels. Only the perennial neo-dualistic possibility that all aspects of consciousness require higher human-specific neocortical circuits not possessed by other animals—potentially a ‘‘read-out fallacy’’—prevents us from openly considering that a study of certain ‘‘instinctual’’ animal emotions may help us bridge toward an understanding of corresponding human psychological issues. Such an evolutionary strategy neither denies the many important differences that exist among species, nor the unique self-awareness permitted by our expansive cortico-cognitive thinking-cap. Other animals probably do not worry about their own mortality as we humans are prone to do, but they surely detect that they are hungry and thirsty and may even recognize having missed a scheduled meal. Animals clearly exhibit robust social preferences and aversions. Even laboratory rats enjoy people who treat them especially well, and dislike those that do not (as suggested by our preliminary evaluation of their ultrasonic self-reports—Knutson et al., 2002; Panksepp et al., 2002). If it were the case that animals had no internal experiences—that they were without the varieties of emotional and motivational feelings so evident in their actions and expressions—there might be little reason, other than aesthetic ones, for us to be concerned about how we treat them. However, if their experiences of the world even remotely resemble our own, then we have profound reasons to reflect upon and to feel sympathy and responsibility for their life qualities—to respect them and to honor them for the many ways they contribute to our own quality of life (McMillan, 2005). Scientifically we may gain much and lose nothing, by accepting other animals into the circle of sentient life. Relevant hypothesis-driven questions can be pursued with rigorous scientific methodologies (Broom, 2001). Our traditional laboratory procedures and rules of inference would not change substantially, even though we may become more enthralled by psychologically relevant ‘‘network doctrines’’ of brain function rather than a mere ‘‘neuron doctrine’’ which has too fine a resolution to ‘‘see’’ affects in action. In the midst of such a metamorphosis, we would still have to recognize that much of mental life transpires at unconscious levels of neural processing. More than anything, we just need to abandon various form of neo-dualism that still prevent us from trying to shed light on the deepest mysteries of existence in humans and other animals. As soon as we accept the possibility that dual-aspect monism may be a more reasonable, a more lifeaffirming and scientifically productive world-view than any form of ‘‘never-mind’’ neo-dualism, many of our conceptual errors, and resulting societal problems, may fade into the historical past. Most of our neuroscientific questions will not require consideration of such psychological subtleties, but discussions of behavioral output in mammals certainly will. References Baars, B. J., Ramsoy, T. Z., & Laureys, S. (2003). Brain, conscious experience and the observing self. Trends in Neurosciences, 26, 671–675. Bekoff, M. (2000). Animal emotions: Exploring passionate natures. BioScience, 50, 861–870. Broom, D. M. (2001). Coping with challenge: Welfare in animals including humans. Berlin: Dahlem University Press. 28 Guest Editorial / Consciousness and Cognition 14 (2005) 22–29 Damasio, A. R. (1999). The feeling of what happens. New York: Harcourt Brace. Damasio, A. R. (2003). Looking for Spinoza. Orlando, FL: Harcourt. Dolan, R. J. (2002). Emotion, cognition, and behavior. Science, 298, 1191–1194. Griffin, D. R. (2001). Animal minds: Beyond cognition to consciousness. Chicago: University of Chicago Press. Knutson, B., Burgdorf, J., & Panksepp, J. (2002). Ultrasonic vocalizations as indices of affective states in rats. Psychological Bulletin, 128, 961–977. McMillan, F. D. (2005). Mental health and well-being in animals. Ames, IO: Iowa State University Press, in press. Murphy, F. C., Nimmo-Smith, I., & Lawrence, A. D. (2003). Functional neuroanatomy of emotions: A meta-analysis. Cognitive, Affective, and Behavioral Neuroscience, 3, 207–233. Norman, D. A. (1980). Twelve issue for cognitive science. Cognitive Science, 4, 1–32. Panksepp, J. (1982). Toward a general psychobiological theory of emotions. The Behavioral and Brain Sciences, 5, 407–467. Panksepp, J. (1988). Brain emotional circuits and psychopathologies. In M. Clynes & J. Panksepp (Eds.), Emotions and psychopathology (pp. 37–76). New York: Plenum Press. Panksepp, J. (1990). Can ‘‘mind’’ and behavior be understood without understanding the brain?. New Ideas in Psychology, 8, 139–149. Panksepp, J. (1998). Affective neuroscience: The foundations of human and animal emotions. London: Oxford University Press. Panksepp, J. (1999). Emotions as viewed by psychoanalysis and neuroscience, An exercise in consilience. NeuroPsychoanalysis, 1, 15–38. Panksepp, J. (2000). On preventing another century of misunderstanding: Toward a psychoethology of human experience and a psychoneurology of affect. Neuro-Psychoanalysis, 2, 240–255. Panksepp, J. (2002). The MacLean legacy and some modern trends in emotion research. In G. A. Cory, Jr. & R. Gardner, Jr. (Eds.), The evolutionary neuroethology of Paul MacLean (pp. ix–xxvii). Westport, CT: Praeger. Panksepp, J. (2003a). Can anthropomorphic analyses of separation cries in other animals inform us about the emotional nature of social loss in humans? Comment on Blumberg and Sokoloff 2001. Psychological Review, 110, 376–388. Panksepp, J. (2003b). At the interface of affective, behavioral and cognitive neurosciences. Decoding the emotional feelings of the brain. Brain and Cognition, 52, 4–14. Panksepp, J. (2003c). DamasioÕs Error?. Consciousness and Emotion, 4, 111–134. Panksepp, J. (Ed.). (2004). Textbook of biological psychiatry. New York: Wiley. Panksepp, J. (2005). Affective consciousness: Core emotional feelings in animals and humans. Consciousness and Cognition, 14, in press. Panksepp, J., & Burgdorf, J. (1999). Laughing rats? Playful tickling arouses high frequency ultrasonic chirping in young rodents. In S. Hameroff, D. Chalmers, & A. Kazniak (Eds.), Toward a science of consciousness III (pp. 124–136). Cambridge, MA: MIT Press. Panksepp, J., & Burgdorf, J. (2003). ‘‘Laughing’’ rats and the evolutionary antecedents of human joy?. Physiology and Behavior, 79, 533–547. Panksepp, J., & Harro, J. (2004). The future of neuropeptides in biological psychiatry and emotional psychopharmacology: Goals and strategies. In J. Panksepp (Ed.), Textbook of biological psychiatry (pp. 627–660). New York: Wiley. Panksepp, J. B., Huber, R. (2004). Ethological analyses of crayfish behavior: A new invertebrate system for measuring the rewarding properties of psychostimulants. Behavioural Brain Research, 153, 171–180. Panksepp, J., Knutson, B., & Burgdorf, J. (2002). The role of brain emotional systems in addictions, A new neuroevolutionary and new self-perspective report animal model. Addiction, 97, 459–469. Panksepp, J., Nocjar, C., Burgdorf, J., Panksepp, J. B., & Huber, R. (2004). The role of emotional systems in addiction: A neuroethological perspective. In R. A. Bevins (Ed.), 50th Nebraska symposium on motivation, motivational factors in the etiology of drug abuse (pp. 85–126). Lincoln: Nebraska, in press. Panksepp, J., & Panksepp, J. B. (2000). The seven sins of evolutionary psychology. Evolution and Cognition, 6, 108–131. Phan, K. L., Wager, T., Taylor, S. F., & Liberzon, I. (2002). Functional neuroanatomy of emotion: A meta-analysis of emotion activation studies in PET and fMRI. Neuroimage, 16, 331–348. Guest Editorial / Consciousness and Cognition 14 (2005) 22–29 29 Rollin, B. E. (1998). The unheeded cry: Animal consciousness, animal pain and science. Ames, Iowa: Iowa University Press. Roques, B. P. (2000). Novel approaches to targeting neuropeptide systems. Trends in Pharmacological Sciences, 21, 475–483. Thomas, K. (1996). Man and the natural World. New York: Oxford University Press. Jaak Panksepp Department of Psychology Bowling Green State University Bowling Green, OH 43403, USA E-mail address: jpankse@bgnet.bgsu.edu Available online 10 December 2004

Consciousness and Cognition Consciousness and Cognition 14 (2005) 4–6 www.elsevier.com/locate/concog In Memoriam Donald R. Griffin A pioneer in the study of consciousness and cognition died this past November. Donald Redfield Griffin shocked the scientific community in 1976 with his book The Question of Animal Awareness, Evolutionary Continuity of Mental Experience. Although the grip of behaviorism had begun to weaken, talking about awareness in relation to nonhuman animals was just not done. But Griffin was a fearless, open-minded observer and thinker, and after his earlier imaginative and rigorous investigation of bats (with Robert Galambos, while they were still undergraduates), leading to the discovery he named ‘‘echolocation,’’ it was clear that he could not be ignored. His explanation of bat behavior had at first drawn vehement attacks because of the complex cognition he was claiming for a mere animal, and now he was attacked again. The fact that a highly respected scientific investigator with such intellectual weight would suggest that animals might be aware shook many to the core. This attitude Griffin labeled ‘‘mentophobia’’; he saw it as a deep-seated fear of the loss of human uniqueness. Griffin’s articles and books (e.g., Animal Thinking, 1984; Animal Minds, 1992) that followed The Question of Animal Awareness drew together examples of complex behavior from a wide variety of doi:10.1016/j.concog.2004.05.005 In Memoriam / Consciousness and Cognition 14 (2005) 4–6 5 animals, giving researchers the incentive and direction to carry out programs of their own in the new field that he called ‘‘cognitive ethology.’’ As a result, animal behavior that had previously been overlooked or disregarded now became the focus of observations and experiments by ethologists and comparative psychologists. Although interest in animal cognition grew steadily, the study of animal consciousness lagged behind, and Griffin was impatient. He was fond of reminding his colleagues that, even though much of our cognition is unconscious, our conscious experience is extremely important to us, and if animals share anything at all similar, it is undoubtedly important to them and warrants our attention. He deeply regretted the dearth of academic opportunities for scientists with an interest in animal consciousness, and he was at a loss to understand why there was so little response to his suggestion that animal communication could serve as a ‘‘window’’ on an animal’s thoughts and feelings, since it was clearly used, albeit imperfectly, for that purpose in humans. Griffin was educated at Harvard, receiving his Ph.D. in zoology in 1942. After teaching at Cornell and then at Harvard (he chaired the Biology Department from 1961 to 1965), he spent the majority of his academic career at Rockefeller University, where he directed the Institute of Animal Behavior. Among his many interests, which he pursued with great imagination, never daunted by practical difficulties, were bird navigation, honeybee communication, and marine mammal acoustic orientation. He developed new techniques for studying animals in their natural environment, using the latest technology when it was available or else tinkering with whatever was at hand. He was a member of the National Academy of Sciences, American Academy of Arts and Sciences, Animal Behavior Society, American Philosophical Society, American Physiological Society, and Corporation of Woods Hole Oceanographic Institution. He served as president of the Harry Frank Guggenheim Foundation. He was married to Ruth Castle, with whom he had four children. He later married Jocelyn Crane. Upon his retirement from Rockefeller as an emeritus professor, Griffin returned to Harvard and continued his investigations of beaver, honeybees, and bats. He welcomed everyone who sought him out, taught a seminar and guest-lectured, met with colleagues at numerous universities, spoke at conferences around the world, published articles, and updated Animal Minds (Animal Minds: Beyond Cognition to Consciousness, 2001). Griffin was gratified that advances in cognitive neuroscience were contributing to the investigation of animal consciousness. He continually attempted to broaden and clarify his own thinking, working with me, up until his last days, on an article (‘‘New Evidence of Animal Consciousness’’) for the January 2004 issue of Animal Cognition. Challenging the common wisdom to the end, he proposed that the adaptive value of consciousness is so powerful that it is possible it accounts for a larger part of the activity in an animal’s small brain than it does in our large human brains. Griffin was 88 years old at his death. His mind was still in its prime. I had the privilege of working with Don for his final eight years. He was a mentor and teacher, and then a colleague and friend. He was the epitome of a person who is passionate and painstaking about searching for the truth, placing no limits on what can be questioned and goodhumoredly urging others to do the same. He approached each day with hopeful enthusiasm. His 6 In Memoriam / Consciousness and Cognition 14 (2005) 4–6 days were long and full, energetic, and creative. His successes are memorable. He will live on as an inspiration to all, a superlative example of a life well spent. Gayle B. Speck Vision Sciences Laboratory Psychology Department Harvard University 33 Kirkland Street Cambridge, MA 02138, USA E-mail address: gayle@wjh.harvard.edu Available online 7 June 2004

Consciousness and Cognition 22 (2013) 1132–1141 Contents lists available at SciVerse ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog How from action-mirroring to intention-ascription? q Pierre Jacob ⇑ Institut Jean Nicod, UMR 8129, ENS/EHESS/CNRS, Ecole Normale Supérieure, Pavillon Jardin, 29, rue d’Ulm, 75005 Paris, France a r t i c l e i n f o Article history: Available online 13 March 2013 Keywords: Mirror neurons Mindreading Prior intention Motor intention Embodied simulation a b s t r a c t This paper is devoted to an assessment of the three-step model offered by Gallese and colleagues in support of the thesis that the function of the mirror mechanism is to mindread an agent’s intention. The first step of the model is the acceptance of the direct-matching model of action understanding. The second step is the endorsement of a different model of mirror neuron activity, i.e. the model of chains of logically related mirror neurons (or motor chains) whose application to action-mirroring is supposed to show that the mirror mechanism enables an observer to predict the goal of the agent’s forthcoming action. The third step is the endorsement of the ‘deflationary’ account of intention-ascription according to which to ascribe an intention to an agent is to predict the goal of the agent’s forthcoming action. I argue that each step of the model faces insuperable objections. Ó 2013 Elsevier Inc. All rights reserved. 1. Introduction The discovery of mirror neurons by Giacomo Rizzolatti and his group in the early 1990’s was a great discovery and a fascinating one. Mirror neurons were first discovered in the ventral premotor cortex and subsequently in the inferior parietal lobule of non-human primates by single-cell recording. They were reported to fire both when an animal performs a goaldirected action onto some target and also when the same animal observes another perform the same kind of transitive action. Evidence for the existence of mirror neuron activity in humans – i.e. a human ‘mirror mechanism’, as Gallese and Sinigaglia (2011) have recently called it –, has been reported on the basis of a variety of experimental techniques. Many studies have also reported that the human mirror mechanism can be elicited by a wider variety of actions than in non-human primates, including intransitive actions, not directed towards a specific physical target. While the mirror mechanism was first documented in the brains of both non-human and human primates during the execution and the perception of actions, further evidence has been taken to show that the human mirror mechanism also underlies both the first-personal experiences of sensations and/or emotions and the third-personal recognition of (and perhaps also the second-personal response to) another’s sensations and/or emotions. One plausible way to interpret the evidence for the activity of the mirror mechanism in the domains of sensations and/or emotions is simply to take notice of the fact that human actions encompass not only goal-directed but also expressive actions. While an agent’s overt goal-directed action is a cue to the agent’s goal, an agent’s overt expressive action is a cue to the agent’s sensation and/or emotion. In what follows I will focus on the activity of the mirror mechanism in the execution and the perception of goal-directed actions.1 q This article is part of a special issue of this journal on Mirror Neurons and Intentionality. I am grateful to Dr Antonella Corradini for both inviting me to the Workshop on Mirror Neurons and Intentionality in Milan on September 4, 2011, and for editing the special issue of Consciousness and Cognition, and to two anonymous referees for their comments on a previous version of this paper. This work was supported by a grant from the French Ministry of Research (ANR-BLAN SOCODEV). ⇑ Fax: +33 146333933. E-mail address: Jacob@ehess.fr 1 Goldman (2008, 2009) has argued that the definition of mirroring should not be restricted to action-mirroring on the grounds that it must also apply to sensations and emotions. But if one draws the distinction between goal-directed and expressive actions, then one may economically restrict mirroring to action. 1053-8100/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.concog.2013.02.005 P. Jacob / Consciousness and Cognition 22 (2013) 1132–1141 1133 Ever since the publication of Gallese and Goldman’s (1998) seminal paper, the cognitive neuroscientists who discovered mirror neurons have interpreted their discovery in terms of the simulation approach to mindreading: the mirror mechanism is taken to enable an observer to ascribe an intention to the agent of an overt action via a process of mental simulation. As Goldman (2008, 2009) has pointed out, the mirror mechanism might contribute to mindreading another’s intention in at least two ways: the mirror mechanism might either causally contribute to, or be constitutive of, mindreading. If the relation is causal, then the mirroring event and the mindreading event are distinct. If the former constitutes the latter, then there is a single event which is both a mirroring and a mindreading event. While Goldman (2008, 2009) has explicitly endorsed the former weaker view, Gallese, Rochat, Cossu, and Sinigaglia (2009, p. 108) have explicitly endorsed the stronger identity view: ‘‘The activation of the mirror neuron system is intrinsically constitutive of action and intention understanding, at least at the level of basic actions’’. On the face of it, Gallese and colleagues’ strong identity claim faces the following challenge. The mirror mechanism might perhaps enable an observer to share an agent’s intention. But to mindread another’s intention is to ascribe an intention to another. So the challenge for Gallese’s strong identity claim is to close the gap between sharing an agent’s intention and ascribing the intention to the agent. In recent years, Gallese has offered a three-step model of intention-ascription based on action mirroring that provides a tentative answer to this challenge.  The first step is the acceptance of the direct-matching model of action understanding, whose application to actionmirroring is supposed to enable an observer to understand an agent’s goal and to ascribe it to the agent.  The second step is the endorsement of a different model of the mirror mechanism, i.e. the model of chains of logically related mirror neurons (or model of motor chains) whose application to action-mirroring is taken to show that the mirror mechanism enables an observer to predict the goal of the agent’s forthcoming action.  The third and final step is the endorsement of the so-called ‘deflationary’ account of intention-ascription according to which to ascribe an intention to an agent is to predict the goal of the agent’s forthcoming action. In addition to endorsing this three-step model of intention-ascription based on action-mirroring, Gallese also embraces what he calls an embodied approach to mental simulation. My twofold goal in this paper is to argue not only that the three-step model fails to establish Gallese’s strong identity claim between action-mirroring and mindreading but also that there is a tension between features of the three-step model and the embodied approach to mental simulation. The paper is in six sections. In the second section, I argue that the uncritical acceptance of the mindreading interpretation of the mirror mechanism may have to some extent prematurely discarded a different interpretation of mirror neuron activity. In the third section, I examine the dilemma Csibra (2007) raised for the direct-matching model of action understanding and the solution favored by advocates of the three-step model of intentionascription. In the fourth section, I argue that there is a tension between the solution to Csibra’s dilemma accepted by advocates of the mindreading interpretation and the embodied approach to mental simulation. In the penultimate section, I turn to the model of chains of logically related mirror neurons and ask the question whether sharing an agent’s chain of logically related mirror neurons could be sufficient for predicting the goal of the agent’s next act. In the final section, I assess the deflationary model of intention-ascription. 2. Two basic options As Rizzolatti, Fogassi, and Gallese (2004, p. 431) have characterized their initial discovery, ‘‘mirror neurons are a specific class of neurons that discharge both when the monkey performs an action and when it observes a similar action done by another monkey or the experimenter’’.2 By advancing the cautious hypothesis that the mirror mechanism might be ‘‘a primitive version, or possibly a precursor in phylogeny, of a simulation heuristic that might underlie mindreading’’, Gallese and Goldman (1998, p. 498) have stressed the synchronous interpersonal neural similarity between the agent’s and the observer’s brains, which results from mirror neuron activity, thereby linking the mirror mechanism to the simulation approach to mindreading. Thus, the mindreading interpretation of the mirror mechanism is entirely consistent with the above definition of the mirror mechanism: it depends on the simulation approach to mindreading, which in turn is rooted in the claim that action-mirroring generates a synchronous interpersonal neural similarity between the agent’s and the observer’s brains. However, Rizzolatti et al. (2004) do not define mirror neuron activity directly in terms of a synchronous interpersonal neural similarity across two distinct brains, but rather in terms of a non-synchronous intrapersonal neural similarity within single brains, at different times, in different tasks, i.e. the execution and the perception of action. In fact, synchronous interpersonal neural similarity between the agent’s and the observer’s brains could hardly be achieved unless there were a mechanism active in single brains, at different times, in different tasks (i.e. the execution and the perception of action). Two distinct brains could not stand in some appropriate similarity relation at a single time unless both brains were independently endowed with a mechanism (the mirror mechanism) active in two different tasks at different times. In other words, 2 Because he thinks that the definition of mirroring should not be restricted to action but should also apply to emotions and sensations, Goldman (2008) has offered an original definition of a mirroring process that both considerably internalizes the definition of mirror neuron activity in an agent’s brain and considerably liberalizes the definition of mirror neuron activity in an observer’s brain. As I explained above, I think that the distinction between goal-directed and expressive actions makes this costly maneuver unnecessary, but I will not examine it any further in this paper. 1134 P. Jacob / Consciousness and Cognition 22 (2013) 1132–1141 synchronous interpersonal neural similarity across two distinct brains at a single time presupposes asynchronous intrapersonal neural similarity at different times in two different tasks. Presumably, natural selection operates on individuals’ brains, not on pairs of individuals’ brains. Only individual members of a species whose brains contained a mechanism active at different times in the execution and the perception of action could have been selected by evolution – not sets of pairs of individual members of a species, whose brains stood in some suitable similarity relation when one executed a grasping action and the other watched the former. Admittedly then non-synchronous intrapersonal neural similarity is a more basic property of the mirror mechanism than synchronous interpersonal neural similarity. If so, then the latter must be a by-product of the former. This point was made by Hurley (2008, p. 758), as part of her criticism of Goldman’s (2006) mental-simulation approach to mindreading, when she argued that ‘‘re-use’’, not similarity, is ‘‘the core generic sense of process-driven simulation’’. However, while Hurley (2008) emphasized the primary role of intrapersonal re-use, and de-emphasized the role of interpersonal similarity, she took it for granted, as Goldman (2006) did, that the primary function of action-mirroring is mindreading and that the correct approach to third-person mindreading must be mental simulation.3 But I want to consider an entirely different interpretation of action-mirroring based on the principle that there could not be synchronous interpersonal neural similarity between two different individuals unless there was a single mechanism in each individual’s brain active at different times in different tasks. It looks as if a neural mechanism (the mirror mechanism) that is active in both the execution and the perception of instances of grasping is a mechanism whose function seems to be to abstract away from the many differences between executing and observing an act of grasping. For example, only an agent executing a task of grasping a target, not an observer perceiving the action, will both have efference copy information about his motor instruction and also haptic information about the target. Only the agent, not an observer, will be in a position to predict the sensory consequences of her action before executing it. Furthermore, among several studies, Kohler et al. (2002) and Keysers et al. (2003) have reported that single neurons in the monkey ventral premotor cortex selectively fire when the animal both executes and also sees, hears and both hears and sees such actions as peanut breaking, ring grasping or paper ripping. So not only does the mirror mechanism seem able to discard the many differences between executing and perceiving an action, but in perceptual tasks, it also seems able to abstract away from the further differences between vision and audition. In a nutshell, the evidence shows that the mirror mechanism is active when an animal executes and perceives one and the same action (e.g. grasping). It also shows that it is able to achieve cross-modal integration in perceptual tasks. Now it seems as if the function of a mechanism able to deliver a representation of an action whose content brackets the differences between the motor, visual and auditory representations of one and the same action is to deliver a representation of this action with conceptual content. The view that the function of the mirror mechanism is to provide conceptual representations of actions makes sense of the variations in the statistical congruency between the motor and the perceptual properties of mirror neurons noted by Csibra (2005, 2007).4 This conceptualist interpretation of the mirror mechanism is in agreement with the view that mirror neuron activity underlies action recognition.5 It is also in agreement with the view expressed by Rizzolatti et al. (2000, p. 542) and Craighero, Metta, Sandini, and Fadiga (2007) according to which area F5 of the monkey ventral premotor cortex (where mirror neurons were first discovered) is ‘‘a store of motor schemas or a vocabulary of actions’’, i.e. a motor vocabulary: ‘‘Neurons forming these vocabularies store both knowledge about an action and the description. . . of how this knowledge should be used’’. Thus, the motor vocabulary stored in F5 consists of mental schemas or symbols whose semantic role is to denote actions and such that the meanings of complex symbols depend systematically on the meanings of their constituents. If so, then mirror neuron activity does not directly underlie the attribution of a psychological state (e.g. an intention) to an agent. But it may contribute the abstract (amodal) content of the concept of the action (e.g. grasping) to determining the content of the agent’s intention (to e.g. grasp a target). Finally, Gallese (2003, p. 1238) himself endorsed the conceptualist interpretation when he wrote: ‘‘If the different modes of presentation of events as intrinsically different as sounds, images, or willed effortful acts of the body are nevertheless bound together within a circumscribed, informational lighter level of semantic reference, what we have here is a mechanism instantiating conceptualization’’. Now, the cognitive neuroscientists who discovered mirror neurons have often construed the output of action-mirroring from the standpoint of what they call motor intentionality, by which they mean to refer to the motor representation of an agent’s action (or goal of her action), which they take to be shared by the agent and the observer (for example, cf. Rizzolatti & Sinigaglia, 2007; Rizzolatti & Sinigaglia, 2010). Building on Jeannerod’s (1994) seminal investigation of motor imagery in humans, several authors (including Jeannerod, 2006 and Pacherie, 2000) have argued that while the content of an agent’s prior intention to perform a goal-directed action is conceptual (or propositional), the non-conceptual content and format of the agent’s motor intention may suitably be linked to the content of motor imagery, i.e. the content of the output of 3 She also wrongly, I think, inverted the respective functions of simulation and mindreading, when she wrote that ‘‘according to simulation theory, mindreading aims at simulation of, and hence, matching the target’s mental states’’ (Hurley, 2008, p. 757). Matching the target’s mental state is the function of mental simulation, which is a step in the process of third-person mindreading, whose function is to ascribe the mental state to another. Ascription requires exiting the simulation stage. By contrast, Goldman’s (2006) own account is a simulation-plus-projection account. 4 Gallese and Sinigaglia (2011, p. 513) recognize that ‘‘interpersonal similarity between a simulator’s and a target’s mental state or process does not qualify as mental simulation unless it arises from intrapersonal reuse of the simulator’s own mental state or process’’. But they do not further take intrapersonal neural similarity as a step towards a conceptualist interpretation of the mirror mechanism (cf. Section 3). 5 Following Sperber (2005), I have argued for such an interpretation of mirror neuron activity in Jacob (2009). Meini and Paternoster (2012) also argue for a similar interpretation, but they hold a different view of concept-possession. P. Jacob / Consciousness and Cognition 22 (2013) 1132–1141 1135 the process whereby the agent imagines performing the action without executing it. The reason why the content of the agent’s prior intention (to e.g. turn the light on) must be conceptual is that it must be able to interact with the conceptual content of the agent’s beliefs and desires. Recently, Butterfill and Sinigaglia (2012) have further argued that a proper account of an agent’s goal-directed action requires that the agent be able to coordinate two distinct representations of the outcome of her action: the outcome of the agent’s action must be represented by both her intention and the motor representation of her own action. They take the outcome of the agent’s action to be conceptually represented by the agent’s intention, on the grounds that the content of the agent’s intention must interact appropriately with the conceptual contents of her beliefs and desires. But they argue that the outcome of the agent’s action must be non-conceptually represented by the agent’s motor representation of her action. On their account, the content and format of the agent’s motor representation of the outcome of her action is on a par with the non-conceptual content and format of motor imagery. Since the outcome of the agent’s action is both represented by the agent’s intention and her motor representation, the two kinds of representation must differ in format. They further offer a solution to the ‘‘interface’’ problem of how the two distinct representations of the outcome of the agent’s action can match each other. They argue that the agent’s intention conceptually represents the outcome by means of a demonstrative action concept that ‘‘defers’’ to the agent’s motor representation of the outcome. I would like to make three short points about Butterfill and Sinigaglia’s (2012) interesting line of thinking about the distinction between the contents and formats of respectively an agent’s intention and the motor representation of her action. First, mirror neuron activity was discovered in the brains of non-human primates and what is known about the content and format of motor imagery is based on the investigation of humans, not non-human primates. As a result, it is not entirely clear to what extent human motor imagery can shed light on the content of the output of action-mirroring. Secondly, it is not entirely clear whether Butterfill and Sinigaglia’s distinction between the agent’s intention and the agent’s motor representation of her action leaves any room for the agent’s motor intention, which shares the world-to-mind direction of fit of the agent’s prior intention and the non-conceptual content of the agent’s motor representation. Thirdly, the distinctive ways the outcome of an agent’s action is being represented by respectively the agent’s intention and her motor representation can be captured along several different lines. For example, the agent’s intention (whether her prior intention or her motor intention) has a world-to-mind direction of fit, not a mind-to-world direction of fit. So it represents the outcome in a prescriptive, not a descriptive, way. Furthermore, the intention is a cause of the achievement of the outcome. By contrast, the motor representation of the agent’s action has a mind-to-world direction of fit and so it represents the outcome in a descriptive, not a prescriptive, way. This distinction is consistent with the hypothesis that the content of the output of action-mirroring must be sufficiently abstract to be commonly entertained by an agent who performs an act and by an individual who sees, hears and both sees and hears the act being performed by another. 3. Answering Csibra’s dilemma Not only is the motor vocabulary or conceptualist interpretation of the mirror mechanism an alternative to the mindreading interpretation, but the latter asymmetrically depends on the former: there could not be synchronous interpersonal neural similarity between two distinct brains unless there were non-synchronous intrapersonal neural similarity within a single brain in different tasks (but of course the converse does not hold). Nonetheless, the simulation-based approach to mindreading has become the orthodox interpretation of the mirror mechanism among the cognitive neuroscientists who discovered mirror neurons. On Gallese’s three-step model, the main function of the mirror mechanism is to mindread another’s intention. The first step of Gallese’s model is the acceptance of the direct-matching model of action understanding, whose application to action-mirroring is supposed to enable an observer to understand an agent’s goal and to ascribe it to the agent. The direct-matching model of action understanding itself involves the three following steps. First, perceiving an agent’s goal-directed action causes the observer to covertly rehearse (or imitate) the agent’s movements by mapping them onto the observer’s motor repertoire (without executing them). Secondly, by covertly rehearsing the agent’s movements, the observer comes to share the agent’s goal. Thirdly, by sharing the agent’s goal, the observer comes to understand it and somehow to ascribe it to the agent.6 Csibra (2007) has raised the following dilemma for the direct-matching model of action-understanding, the first horn of which is that action-mirroring consists in mapping the agent’s executed movements onto the observer’s motor repertoire in accordance with the first step of the direct-matching model. The second horn is that action-mirroring enables the observer to share and understand the agent’s goal in accordance with the second and third steps of the direct-matching model. The dilemma between the two horns arises from the fact that an agent’s movements do not stand in a one-to-one relation to her goal: not only can an agent recruit different movements in the service of a single goal, but she can also recruit one and the same movement in the service of different goals. So, if what is being mapped onto the observer’s motor repertoire by action-mirroring is the agent’s overt motor act (or bodily movements), then it is unlikely that mirroring could deliver a representation (or understanding) of the agent’s goal. Conversely, if the output of mirroring is a representation (or understanding) of the agent’s goal, then it is unlikely to be generated by the mapping of the agent’s observed movements onto the observer’s motor repertoire. 6 While ‘intention’ refers to an agent’s psychological state, ‘goal’ may refer either to a possible non-mental state of affairs in the environment, i.e. a goal-state, which the agent aims to achieve, or to a mental representation of such a goal-state (cf. Gergely & Csibra, 2003). 1136 P. Jacob / Consciousness and Cognition 22 (2013) 1132–1141 In response, Rizzolatti and Sinigaglia (2010) have opted for the second horn of Csibra’s dilemma: on the basis of experimental results reported by Umiltà et al. (2008), they have argued that in the monkey brain, mirror neuron activity encodes goals. Umiltà et al. (2008) trained monkeys to grasp objects using both normal pliers and so-called ‘reverse’ pliers: grasping an object with normal pliers involves closing the fingers, but grasping with reverse pliers involves opening the fingers. Single cell recordings in area F5 in executive and perceptual tasks by Umiltà et al. (2008) show that F5 neurons discharged during the same phase of grasping ‘regardless of whether this involved opening or closing of the hand’ (Rizzolatti & Sinigaglia, 2010, p. 266). In other words, what seems to matter to the firing of F5 neurons is the agent’s goal (grasping) irrespective of the difference between the agent’s closing and opening his fingers. 4. What is so special about embodied simulation?7 Since they published their joint (1998) paper together in support of the mindreading interpretation of the mirror mechanism, Gallese’s and Goldman’s respective views have diverged to some extent. While Goldman (2006, 2008, 2009) has argued for a merely causal link between the mirror mechanism and mindreading another’s intention, Gallese defends a stronger constitutive link: ‘‘The activation of the mirror neuron system is intrinsically constitutive of action and intention understanding, at least at the level of basic actions’’ (Gallese et al., 2009, p. 108). Gallese further advocates an embodied approach to mental simulation. I will examine Gallese’s strong claim in favor of the constitutive link between mirroring and mindreading another’s intention in the next section, when I turn to his deflationary view of intention-ascription. In the present section, I want to suggest that there is a deep tension (or a dilemma) between Rizzolatti and colleagues’ solution to Csibra’s dilemma and the endorsement of an embodied approach to mental simulation. This tension shows up quite clearly in a recent paper by Gallese and Sinigaglia (2011, p. 516), in which they construe embodied simulation as ‘‘the reuse of mental representations that are bodily in format’’ and they further argue that ‘‘there is substantial evidence that the mirror mechanism is selective for motor goals and motor intentions, regardless of the body effectors and kinematic features enabling their accomplishments’’. The dilemma for this view is the following: on the one hand, in response to Csibra’s dilemma, Gallese and Sinigaglia (2011) clearly take it that the mirror mechanism codes an agent’s goals (and intentions), not the agent’s movements. But on the other hand, they also characterize the mirror mechanism as a process of mental simulation whereby the goal is coded in a bodily format. So the question is: how could the mirror mechanism code an agent’s goal in a bodily format ‘‘regardless of the body effectors and kinematic features enabling its accomplishments’’? So the next question arises: What makes an approach to mental simulation embodied? What makes the question fairly complicated is that many, but not all, advocates of the embodied cognition program mean to reject what Wheeler (2005, p. 81) calls the neurocentric assumption ‘‘at work within orthodox cognitive science’’, according to which ‘‘the causal factors that explain the adaptive richness and flexibility of naturally occurring intelligent behavior are located neither in the agent’s non-neural body nor in her environment, but pretty much exclusively in her brain’’. As Clark (2008a, 2008b) and Jacob (2012b) have further argued, there are currently two distinct ways of rejecting the neurocentric assumption: one is Shapiro’s (2011) body-centric view according to which an individual’s mind (or cognition) is constituted by both the individual’s brain and the brain’s non-neural bodily environment, which makes an individual’s mind a two-place relation between the individual’s brain and its non-neural bodily environment. The other one is Clark and Chalmers’ (1998) and Clark’s (2008b) extended mind thesis according to which an individual’s mind (or cognition) is constituted by the individual’s brain, the brain’s nonneural bodily environment and the individual’s non-bodily environment, which makes an individual’s mind a three-place relation between an individual’s brain, its non-neural bodily environment and its non-bodily environment. Not all advocates of embodied cognition, however, are willing to reject the neurocentric assumption. Instead, some advocates of embodied cognition (including Gallese) seem happy to recognize that an individual’s brain is part of her body and want to emphasize instead the contribution of the human motor and/or perceptual systems to higher human cognitive capacities (e.g. mindreading). Gallese himself has offered three answers to the question what makes mental simulation embodied, all of which seem consistent with neurocentrism. First, Gallese (2009a, p. 493) argues that mental simulation is embodied if it is ‘‘a mandatory, pre-rational, non-introspectionist functional mechanism’’. Secondly, Gallese (2009b, p. 520) argues that what makes mental simulation embodied is that it is ‘‘a crucial functional mechanism of intersubjectivity by means of which the actions, emotions, and sensations of others are mapped by the same neural mechanisms that are normally activated when we act or experience similar emotions and sensations’’. While Gallese’s second condition amounts to defining the process of mental simulation as interpersonal neural similarity, both answers are clearly consistent with a neurocentric approach to mental simulation. Finally, Gallese and Sinigaglia (2011, p. 513) have recently adopted Goldman and de Vignemont’s (2009) proposal that what makes a representation embodied is its bodily format: ‘‘the idea is that just as a map and a series of sentences might represent the same route with a different format, so mental representations might have partly overlapping contents (e.g. a motor goal, an emotion or sensation) while differing from one another in their format (e.g. bodily instead of propositional)’’. Building on Jeannerod’s (1994, 2006) insight (echoed by Pacherie, 2000), Butterfill and Sinigaglia have argued that, unlike thinking about an action of grasping a cup, executing the action, perceiving the action executed by another and imagining 7 This is the title of a Gallese and Sinigaglia’s (2011) paper. P. Jacob / Consciousness and Cognition 22 (2013) 1132–1141 1137 performing the action without executing it, all three may be said to recruit motor representations with the same ‘‘performance profile’’, to the extent that these representations may interfere with one another. In line with Gallese and Sinigaglia (2011), they conclude that what makes embodied simulation embodied is not the content, but the bodily format, of the motor representation of the (goal of the) action. I claimed above that there is a tension between two strands in Gallese and Sinigaglia’s (2011) approach to embodied simulation: on the one hand, the solution to Csibra’s dilemma requires that detailed information about bodily effectors and the kinematics of the agent’s movements be disregarded from the content of the representation of the agent’s goal generated by action-mirroring. On the other hand, the stress on embodiment pulls towards including detailed information about the agent’s bodily effectors and the kinematics of her movements as part of the content of the representation of the agent’s goal generated by action-mirroring. Could the tension be resolved by appeal to the distinction between the content and the format of a representation of the agent’s goal?8 Given Goldman and de Vignemont’s (2009) approach to embodied cognition further elaborated by Goldman’s (2012), I don’t think that appeal to the content/format distinction can resolve the tension. Their approach to embodied cognition is two-tiered. On the one hand, as Goldman (2012, pp. 73–74) makes clear, a representation derives its bodily format from its bodily content: a representation qualifies as embodied if it is generated by a mechanism (e.g. the somatosensory system or the motor system) whose primary function is to represent one’s own bodily parts and states (e.g. pain, tickle, temperature, itch, muscular and visceral sensations, sensual touch, and other feelings from, and about, the body). For example, on Goldman’s (2012, p. 79) account, action-mirroring counts as an embodied prescriptive representation of how the agent is to move her own effectors (in e.g. a task of executing an act of grasping). On the other hand, Goldman (2012) further assumes that an agent’s representation of her own body with a bodily format can be re-deployed (or re-used) in a novel cognitive task, for representing other things than her own body parts. According to Goldman, the neuroscientific findings show that the embodied prescriptive representation of how to move the agent’s effectors generated in an agent’s brain by mirror neuron activity is re-deployed in action observation. Thus, acceptance of Goldman’s two-tiered conception of embodied cognition seems to mandate a solution to Csibra’s dilemma which runs against the option favored by Rizzolatti, Gallese and Sinigaglia. On Goldman’s conception, the output of action-mirroring represents an agent’s motor command to one of the agent’s effectors, not a representation of the agent’s goal irrespective of particular effectors and the kinematics of the effectors’ movements. In a nutshell, it is really unclear how to reconcile an embodied version of the simulation approach to mindreading with a solution to Csibra’s dilemma. 5. The motor chain model of mirror neuron activity So far, I have only discussed the first assumption of the three-step model of intention-ascription based on action-mirroring, i.e. the direct matching model of action-understanding. The three-step model of intention-ascription also involves a different model of action-mirroring (or mirror neuron activity): the model of ‘‘chains of logically related mirror neurons’’. The second assumption of the three-step model is that by sharing chains of logically related mirror neurons with the agent of a goaldirected action, an observer can predict the goal of the agent’s forthcoming action. In the present section, I turn to this second assumption. The model of chains of logically related mirror neurons, which was first considered in an early paper by Di Pellegrino, Fadiga, Fogassi, Gallese, and Rizzolatti (1992), was re-introduced in order to account for intriguing findings based on single cell-recordings in the monkey by Fogassi et al. (2005) and on brain-imaging in humans by Iacoboni et al. (2005). Fogassi et al. (2005) recorded single mirror neurons in the monkey inferior parietal lobule (IPL) while the animal executed an act of grasping a target which was embedded into two distinct more complex actions, one of which was eating (by bringing the target to the mouth) and the other of which was placing the target into a container. They also recorded single mirror neurons while the animal was watching another agent perform the same action of grasping embedded into one or the other of the two more complex actions. They found that different mirror neurons fire when the animal either performed or observed the same action of grasping, according to whether the following act was bringing to the mouth or placing into a container. These findings are currently interpreted as evidence that mirror neurons in the monkey IPL form motor chains or are organized along chains of intentional motor acts. Iacoboni et al. (2005) used functional magnetic resonance imaging to study the brain activations of participants while they watched video-clips displaying three possible conditions. In the Context condition, they saw a teapot, a cup and a plate with cookies either orderly as if before tea or disorderly as if after tea. In the Action condition, they saw a human hand grasp a cup either by full prehension or precision grip with no contextual elements. In the Intention condition, they saw a human hand grasp a cup by either full prehension or precision grip in one of the two contexts described above. Iacoboni et al. (2005) found a significantly stronger signal in the posterior part of the inferior frontal gyrus and the adjacent sector of the ventral premotor cortex where hand actions are represented for the observation of the Intention condition than for the other two conditions. Both Fogassi et al. (2005) and Iacoboni et al. (2005) take their findings to support the model of chains of logically related mirror neurons (or motor chains). On this model, mirror neurons are taken to form chains: one chain might link cells firing 8 I am grateful to an anonymous reviewer for forcing me to be more explicit on this point. 1138 P. Jacob / Consciousness and Cognition 22 (2013) 1132–1141 for grasping to cells firing for bringing to the mouth. Another chain might link cells firing for grasping to cells firing for placing into a container. In accordance with the second assumption of the three-step model, the authors assume that sharing an agent’s chain of logically related mirror neurons enables an observer to predict the goal of the agent’s next act.9 Now, the question is whether sharing the agent’s motor chain of logically related mirror neurons could be a sufficient condition for predicting the goal of the agent’s next act. In the experiments by either Fogassi et al. (2005) or Iacoboni et al. (2005), the inference of the goal of the agent’s next act depends on two dissociable parameters: the observation of the agent’s executed act of grasping and contextual cues. What the data show is that the activation of chain CH1 of logically related mirror neurons in an observer’s brain is triggered by the perception of the act of grasping in context C1, while the activation of chain CH2 of logically related mirror neurons in the brain of the same observer is triggered by the perception of the act of grasping in context C2. Given that the perceived act of grasping is the same in both cases, the data do not rule out the possibility that what enables the observer to infer the two distinct goals of the agent’s two distinct next acts is the difference between context C1 and context C2. As the findings by Iacoboni et al. (2005) in the Context condition clearly show, the perception of the context alone fails to trigger mirror neuron activity. If so, then sharing chains of logically related mirror neurons with the agent will not be sufficient for predicting the goal of the agent’s next act because knowledge of the context (which is not achieved by mirror neuron activity at all) is also necessary. According to the model of motor chains, mirror neurons are organized into chains of motor acts. But clearly if mirror neurons are organized into motor chains, mirror neuron activity itself cannot select the right motor chain in the observers’ brain: the right motor chain cannot be self-selected in the observer’s brain; only the processing of contextual cues, which cannot be achieved by mirror neuron activity, may enable the observer to select the right motor chain. 6. The deflationary account of intention-ascription The third assumption of the model of intention-ascription is what Gallese (2007, p. 662) has labeled the deflationary account of intention-ascription, according to which determining an agent’s intention (or ascribing it to the agent) amounts to determining why a given act (e.g. grasping a cup) was performed, which in turn is equivalent to detecting the goal of the agent’s ‘‘still not executed and impending subsequent act (e.g. bringing the cup to the mouth’’).10 As a result of accepting both the view that sharing another’s chain of logically related mirror neurons is a sufficient basis for predicting the goal of her forthcoming act and the deflationary account of intention-ascription, one can conclude that sharing an agent’s chain of logically related mirror neurons enables an observer to ascribe an intention to the agent. In accordance with the deflationary account of intention-ascription, Iacoboni et al. (2005, p. 0533) have argued that their own findings ‘‘strongly suggest that coding the intention associated with the actions of others is based on the activation of a neuronal chain formed by mirror neurons coding the observed motor act and by ‘logically related’ mirror neurons coding the motor acts that are most likely to follow the observed one, in a given context’’. The application of the deflationary account of intention-ascription here is open to the two following critical comments. First, when applying the deflationary account of intention ascription, Iacoboni et al. (2005, p. 0530) entertain the hypothesis that the mirror neuron system might code ‘‘the global intention associated with the observed action’’. What is meant by ‘‘the agent’s global intention’’? In the context of their study, Iacoboni et al. (2005) must mean that the activity of a given chain of logically related mirror neurons codes the agent’s prior intention to e.g. drink and that by sharing the agent’s chain of logically related mirror neurons, an observer can ascribe to the agent the prior intention to e.g. drink. If so, then it is surprising that Rizzolatti and Sinigaglia (2010, p. 271) backtrack from this strong claim and instead endorse the weaker claim that ‘‘the studies reviewed above [by which they refer to Iacoboni et al. (2005) study] indicate that the parieto-frontal mirror network may subserve the understanding of the motor intention underlying the actions of others’’ [my emphasis]. In the Action and Intention conditions of Iacoboni et al.’s (2005) study, the agent has the prior intention to either drink or clean and the motor intention to grasp the cup using either full prehension or precision grip. One must choose between two strategies: one can argue that mirror neuron activity in participants’ brains codes the agent’s motor intention to grasp the cup with either full prehension or precision grip. Alternatively, one can argue that mirror neuron activity in participants’ brains codes the agent’s prior intention to either drink or clean. But one cannot argue at once that mirror neuron activity in the observer’s brain codes both the agent’s prior intention and her motor intention. In particular, one cannot do so in the context of Iacoboni et al.’s (2005) study because what constitutes decisive evidence in favor of ascribing to the agent one prior intention (drinking) over the other (cleaning) are the contextual elements (before tea vs. after tea), not the distinctive ways in which the agent grasps the cup (by full prehension vs. by precision grip). As it happens, in the Intention condition, as in the Action condition, participants see alternations between instances of grasping the cup with full prehension and with precision grip. However, only by arguing that mirror neuron activity in the observer’s brains codes the agent’s motor intention (not her prior intention) could one further claim that the content of mirror neuron activity is coded in a specific bodily format and therefore that the process of mental simulation involved in mirror neuron activity is embodied. 9 I accept Barlassina’s (2011) well-taken reply to my earlier criticisms of the model of chains of logically related mirror neurons in Jacob (2008). As Iacoboni et al. (2005, p. 0533) put it, ‘‘to ascribe an intention is to infer a forthcoming new goal, and this is an operation that the motor system does automatically’’. For a very interesting assessment of the deflationary account of intention-ascription, cf. Spaulding (in press). 10 P. Jacob / Consciousness and Cognition 22 (2013) 1132–1141 1139 Secondly, referring to mirror neuron activity, as mirror neuron theorists do, as ‘‘coding’’ an agent’s intention is ambiguous between two distinct interpretations: ‘‘coding’’ an agent’s intention might either mean sharing the agent’s intention or ascribing the intention to the agent. Sharing the agent’s intention is to have the same intention as the agent, namely to intend to do something or other (cf. Jacob, 2012a). But ascribing an intention to another is not to intend to do something or other: to ascribe an intention to an agent is to judge or believe that another intends to do so and so. While having the concept INTENTION is not necessary to intend to do something and therefore to share another’s intention, it is necessary for believing that another intends to so something, and therefore to ascribe an intention to another. Nor (I believe) is intending to do something a necessary condition for believing that another intends to do something and therefore for ascribing the intention to do something to another. I am inclined to say that by sharing the agent’s chain of logically related mirror neurons, an observer may share the agent’s ability to predict the goal of the agent’s next act. But it would, I think, be a mistake to identify the agent’s prior intention to drink with her ability to predict the goal of her next act (bringing to the mouth) while she is performing the act of grasping the cup. Similarly, it would, I think, be a mistake to identify the observer’s ability to ascribe to the agent the prior intention to drink with her ability to predict the goal of the agent’s next act (bringing to the mouth) upon observing the agent grasp the cup. 7. Conclusions In this paper, I have assessed each step of the impressive three-step model offered by Vittorio Gallese and his colleagues, who discovered mirror neurons, in support of the thesis that the function of mirror neuron activity is to enable an observer to ascribe an intention to an agent who performs a goal-directed action. Distancing myself from the detailed criticisms of each step of the three-step model of intention-ascription, I would now like to make explicit one of the fundamental bases of my disagreement with Gallese and colleagues’ mindreading interpretation of the mirror mechanism in how one construes what it takes to achieve a task of mindreading. What I think Premack and Woodruff’s (1978) seminal paper and the subsequent discussion of this paper by the three philosophers Bennett (1978), Dennett (1978) and Harman (1978) showed is that one can predict an agent’s likely behavior by sharing her motivations (goals and desires) and epistemic states (beliefs about the way the world is). But predicting another’s likely behavior by sharing her epistemic and motivational states falls short of providing evidence that one can read another’s mind for the simple reason that to share another’s belief and desire about some states of affairs in the world just is to represent the world in the same way as the other individual does. What would decisively show that one can read another’s mind (or possesses a ‘theory-of-mind’) is that one could ascribe to an agent motivations and/or epistemic states that one does not share. For example, it is one thing to share an agent’s intention; it is another thing to ascribe an intention to the agent. To share an agent’s intention is to have an intention. To ascribe an intention to an agent is to believe or judge that another intends to do so and so. (Furthermore, one could not come to judge or believe that another intends to do so and so unless one possessed the concept INTENTION.) To judge or believe that an agent intends to do so and so, not to share her intention, is to mindread her intention. The significance of the famous standard elicited false belief task (designed by Wimmer and Perner, 1983) lies in the fact that while the participant has a true belief about some object’s location, her task is to predict where an agent with a false belief will look for the object. Only if she can ascribe to the agent a belief whose content the participant does not share can she pass the standard elicited false belief task. While there are reasons to believe that there is more to passing the standard elicited false belief task than ascribing false beliefs to another, passing the standard false belief task is clearly a sufficient condition for displaying the ability to ascribe to another a psychological state distinct from one’s own (cf. Baillargeon, Scott, & He, 2010; Bloom & German, 2000). Thus it is that the ability to ascribe to others psychological states that one does not share has been taken to be a hallmark of the ability to read others’ minds. My claim that sharing another’s intention falls short of mindreading her intention for the reasons offered in the above paragraph seems open to two criticisms. On the one hand, while the evidence (reviewed by Baillargeon et al., 2010) based on spontaneous non-elicited tasks is taken by many (but not all) developmental psychologists to suggest that preverbal human infants can make sense of an agent’s instrumental action by taking into account the contents of both her motivations and epistemic states (including her false belief), it sounds unlikely (if not outrageous) to credit infants with the ability to ascribe to an agent a belief that the infants themselves take to be false.11 On the assumption that preverbal human infants achieve tasks of mindreading, it would seem that ascribing an intention to another (as opposed to sharing another’s intention) might be too strong a requirement on mindreading. On the other hand, the findings reported by Kovacs, Téglas, and Endress (2010) show that human adults come to share automatically another’s perspective. On the assumption that this counts as mindreading another’s perspective, these findings too show that ascribing a mental state to another would seem unnecessary as a condition on mindreading.12 I will consider the first objection first. In most (if not all) so-called spontaneous non-elicited ‘‘false belief tasks’’, what is being tested is participants’ ability to make sense of an agent’s false belief by measuring participants’ surprise when the agent with a false belief reaches for the object at its actual location or by measuring participants’ ability to look towards 11 12 See e.g. Perner and Ruffman (2005) for skepticism. I am grateful to an anonymous referee for pressing me to be more explicit on these points. 1140 P. Jacob / Consciousness and Cognition 22 (2013) 1132–1141 the empty location where an agent with a false belief is going to reach for the object in anticipation of the agent’s action. If and when an infant displays surprise upon observing an agent with a false belief reach for an object at its actual location or looks at the empty location in anticipation of the action of an agent with a false belief, it makes sense to assume that the infant is able to take the agent’s epistemic perspective, represent the content of the agent’s false belief and, therefore, to ascribe to the agent a belief that happens to be false. What the infants’ surprise or anticipatory looking behavior does not require is that the infant draws an explicit comparison between the content of her own true belief and the content of the agent’s false belief. It is not required that the infant recognizes that, while her own belief is true, the agent’s belief is false. In other words, it is not required that she assesses the agent’s belief as false or that she predicates the property of falsehood of the agent’s belief. I now turn to the findings by Kovacs et al. (2010): in their experiment, adult participants see a video displaying a blue smurf who places a ball on a table on which there is also an occluder. After the ball has been placed on the table, its own self-propelled motion takes it behind the occluder and it either remains there or not. The participants’ psychophysical task is to press a button as fast as possible upon detecting the ball after the removal of the occluder. While Kovacs et al. (2010) report that adults are faster to press the button when they expect the ball to be behind the occluder than when they do not, they also found that adults are faster when they do not expect the ball to be there, but the smurf falsely expects it to be there. This shows that adults automatically take into account the smurf’s false expectation, in spite of the irrelevance of the smurf’s expectation to the psychophysical task. In Kovacs and colleagues’ experiment, participants are not presented with an instrumental action performed by an agent as a means towards satisfying her desire or intention in light of her belief. Their task is to press a button as fast as possible, not to make sense of an agent’s belief and desire. Although they see the blue smurf bring the ball and place it on the table, it is questionable whether participants are representing the content of the smurf’s motivation at all. Thus, Kovacs et al.’s findings are surprising for two related reasons: one is that in spite of its complete irrelevance to the task, the smurf’s false expectation influences participants’ performance. The other is that participants seem to automatically take the smurf’s epistemic perspective without ascribing any motivation to the smurf. While this is evidence that participants take the smurf’s epistemic perspective into account, it is still an open question whether participants specifically tie the content of the smurf’s false expectation to the smurf or whether they use this content as a way of mitigating their confidence in the truth of their own conflicting expectation. In some of their writings, Gallese and his colleagues seem aware of the fact that they must face the challenge of offering an alternative conception of mindreading. Thus, Gallese, Keysers, and Rizzolatti (2004, p. 396) have described the mirror mechanism as ‘‘the fundamental mechanism that allows us a direct experiential grasp of the mind of others is not conceptual reasoning but direct simulation of the observed events’’. Presumably, by ‘‘a direct experiential grasp of the mind of others’’ what Gallese et al. (2004) mean is sharing another’s mental state (e.g. an intention). Gallese and Sinigaglia (2011, p. 518) explicitly recognize that the mirror mechanism can only play a causal, not a constitutive, role in mindreading, ‘‘when mindreading needs a representational attribution, that is, when attribution concerns propositional attitudes such as beliefs and desires that can be construed as putative reasons for action’’. But they maintain that the mirror mechanism constitutes mindreading ‘‘when mindreading involves functional attribution of mental states or processes (such as a motor goal or a motor intention. . .) having a bodily format’’. So now the question arises what are the grounds for the distinction between functional and representational attributions of psychological states to others? While the purpose of representational attribution is to construe others’ mental states as putative reasons for their actions, what else could be accomplished by functional attributions? As Gallese and Sinigaglia (2011, p. 517) put it, ‘‘functionally, an attribution is a representation of a goal, intention or belief which plays some role in enabling one to deal with an agent by virtue of its being appropriately related to that agent’s goal, intention or belief’’. As I understand it, a an instance of a functional (non-representational) attribution of a goal and/or intention to another would consist in forming a joint goal (i.e. moving a piece of furniture together) and taking the steps towards fulfilling it by engaging in joint action with another. Thus, sharing a joint goal in this sense does seem to ‘‘enable one to deal with an agent by virtue of being appropriately related to that agent’s goal’’. Now the ability to share a joint goal and engage in joint action with another has been highlighted by Tomasello, Carpenter, Call, Behne, and Moll (2005) as a uniquely human cognitive and motivational capacity, not shared by non-human primates. If so, then mirror neuron activity would not seem sufficient to support sharing a joint goal and engaging in joint action. In fact, it seems to me that establishing a joint goal and fulfilling it by successfully executing a joint action require for each agent not only to track and mindread the other’s motivational and epistemic states and to construe them as putative reasons for her action, but also to engage in communicative action. If this is correct, then far from being achieved by mirror neuron activity, functional attribution (exemplified by sharing a joint goal) seems to rely on, and require, representational attributions of mental states to others, i.e. construing others’ mental states as reasons for their actions. References Baillargeon, R., Scott, R. M., & He, Z. (2010). False-belief understanding in infants. Trends in Cognitive Sciences, 14(3), 110–118. Barlassina, L. (2011). After all, it’s still replication: A reply to Jacob on simulation and mirror neurons. Res Cogitans, 8(1), 92–111. Bennett, J. (1978). Some remarks about concepts. Behavioral and Brain Sciences, 4, 557–560. Bloom & German, T. (2000). Two reasons to abandon the false belief task as a test of theory of mind. Cognition, 77, B25–B31. Butterfill, S. & Sinigaglia, C. (2012). Intention and motor representation in purposive action. Philosophy and Phenomenological Research. Clark, A. (2008a). Pressing the flesh: a tension in the study of the embodied, embedded mind? Philosophy and Phenomenological Research, 76(1), 37–59. P. Jacob / Consciousness and Cognition 22 (2013) 1132–1141 1141 Clark, A. (2008b). Supersizing the mind, embodiment, action, and cognitive extension. Oxford: Oxford University Press. Clark and Chalmers (1998). The extended mind. Analysis, 58(1), 7–19. Craighero, L., Metta, G, Sandini, G. & Fadiga, L. (2007). The mirror-neurons system: data and models. In von Hofsten & Rosander (Eds.), Progress in Brain Research, Vol. 164. Csibra, G. (2005). Mirror neurons or emulator neurons? <http://mirrorneurons.free.fr>. Csibra, G. (2007). Action mirroring and action understanding. In P. Haggard, Y. Rossetti & M. Kawato (Eds.), Sensorimotor foundations of higher cognition. Attention and Performance XXII. Oxford: Oxford University Press. Dennett, D. C. (1978). Beliefs about beliefs. Behavioral and Brain Sciences, 4, 568–570. Di Pellegrino, G., Fadiga, L., Fogassi, L., Gallese, V., & Rizzolatti, G. (1992). Understanding motor events: A neurophysiological study. Experimental Brain Research, 91, 176–180. Fogassi, L., Ferrari, P. F., Gesierich, B., Rozzi, S., Chersi, F., & Rizzolatti, G. (2005). Parietal lobe: From action organisation to intention understanding. Science, 308, 662–667. Gallese, V. (2003). A neuroscientific grasp of concepts: from control to representation. Philosophical Transactions of the Royal Society, London B, 358, 1231–1240. Gallese, V. (2007). Before and below theory of mind: Embodied simulation and the neural correlates of social cognition. Philosophical Transactions of the Royal Society, 362(1480), 659–669. Gallese, V. (2009b). Mirror neurons, embodied simulation, and the neural basis of social identification. Psychoanalytic Dialogues, 19(5), 519–536. Gallese, V., & Goldman, A. I. (1998). Mirror neurons and the simulation theory of mindreading. Trends in Cognitive Sciences, 2, 493–501. Gallese, V., Keysers, C., & Rizzolatti, G. (2004). A unifying view of social cognition. Trends in Cognitive Sciences, 8, 396–403. Gallese, V. (2009a). Mirror neurons and the neural exploitation hypothesis: From embodied simulation to social cognition. In J. A. Pineda (Ed.), Mirror neuron systems (pp. 163–190). New York: Humana. Gallese, V., Rochat, M., Cossu, G., & Sinigaglia, C. (2009). Motor cognition and its role in the phylogeny and ontogeny of action understanding. Developmental Psychology, 45(1), 103–113. Gallese, V., & Sinigaglia, C. (2011). What is so special about embodied simulation? Trends in Cognitive Sciences, 11, 512–519. Gergely, G., & Csibra (2003). Teleological reasoning about actions: The naïve theory of rational actions. Trends in Cognitive Sciences, 7, 287–292. Goldman, A. I. (2006). Simulating Minds the Philosophy, Psychology, and Neuroscience of Mindreading. Oxford: Oxford University Press. Goldman, A. I. (2008). Mirroring, mindreading, and simulation. In J. A. Pineda (Ed.), Mirror neuron systems (pp. 311–330). New York: Humana. Goldman, A. I. (2009). Mirroring, simulating and mindreading. Mind and Language, 24(2), 235–252. Goldman, A. I. (2012). A moderate approach to embodied cognitive science. Review of Philosophy and Psychology, 3, 71–88. Goldman, A. I., & de Vignemont, F. (2009). Is social cognition embodied? Trends in Cognitive Sciences, 13(4), 154–159. Harman, G. (1978). Studying the chimpanzee’s theory of mind. Behavioral and Brain Sciences, 4, 576–577. Hurley, S. (2008). Understanding simulation. Philosophy and Phenomenological Research, 77(3), 755–774. Iacoboni, M., Molnar-Szakacs, I., Gallese, V., Buccino, G., Mazziotta, J. C., & Rizzolatti, G. (2005). Grasping the intentions of others with One’s own mirror neuron system. Plos Biology, 3, 529–535. Jacob, P. (2008). What do mirror neurons contribute to human social cognition? Mind and Language, 23(2), 190–223. Jacob, P. (2009). A philosopher’s reflections on the discovery of mirror neurons. Topics in Cognitive Science, 1(3), 570–595. Jacob, P. (2012a). Sharing and ascribing goals. Mind and Language, 27(2), 202–229. Jacob, P. (2012b). Embodying the mind by extending it. Review of Philosophy and Psychology, 3(1), 33–51. Jeannerod, M. (1994). The representing brain: neural correlates of motor intention and imagery. Behavioral and Brain Sciences, 17(2), 187–245. Jeannerod, M. (2006). Motor cognition. Oxford: Oxford University Press. Keysers, C., Kohler, E., Umiltà, M. A., Fogassi, L., Rizzolatti, G., & Gallese, V. (2003). Audiovisual mirror neurons and action recognition. Experimental Brain Research, 153, 628–636. Kohler, E., Keysers, C., Umiltà, M. A., Fogassi, L., Gallese, V., & Rizzolatti, G. (2002). Hearing sounds, understanding actions; action representation in mirror neurons. Science, 297, 846–848. Kovacs, A. M., Téglas, E., & Endress, A. D. (2010). The social sense: Susceptibility to others’ beliefs in human infants and adults. Science, 330, 1830–1834. Meini, C., & Paternoster, A. (2012). Mirror neurons as a conceptual mechanism? Mind and Society, 11(2), 183–201. Pacherie, E. (2000). The content of intentions. Mind and Language, 15(4), 400–432. Perner, J., & Ruffman, T. (2005). Infants’ insight into the mind: How deep? Science, 308, 214–216. Premack, D., & Woodruff, G. (1978). Does the chimpanzee have a theory of mind? Behavioral and Brain Sciences, 4, 515–526. Rizzolatti, G., Fogassi, L., & Gallese, V. (2000). Cortical mechanisms subserving object grasping and action recognition: A new view on the cortical motor functions. In M. S. Gazzaniga (Ed.), The new cognitive neurosciences. Cambridge, Mass: MIT Press. Rizzolatti, G., Fogassi, L., & Gallese, V. (2004). Cortical mechanisms subserving object grasping, action understanding, and imitation. In M. Gazzaniga (Ed.), The cognitive neurosciences III (pp. 427–440). Cambridge, MA: MIT Press. Rizzolatti, G., & Sinigaglia, C. (2007). Mirror neurons and motor intentionality. Functional Neurology, 22(4), 205–210. Rizzolatti, G., & Sinigaglia, C. (2010). The functional role of the parieto-frontal mirror circuit: interpretations and misinterpretations. Nature Reviews. Neuroscience, 11(4), 264–274. Shapiro, L. (2011). Embodied cognition. London: Routledge. Spaulding, S. (in press). Mirror Neurons and Social Cognition. Mind and Language. Sperber, D. (2005). Mirror neurons or concept neurons? <http://www.interdisciplines.org/mirror/papers/1>. Tomasello, M., Carpenter, M., Call, J., Behne, T., & Moll, H. (2005). Understanding and sharing intentions: The origins of cultural cognition. Behavioral and Brain Sciences, 28, 675–691. Umiltà, M. A., Escola, L., Intskirveli, I., Grammont, F., Rochat, M., Caruana, F., et al (2008). When pliers become fingers in the monkey motor system. Proceedings of the National Academy of Sciences, 105(6), 2209–2213. Wheeler, M. (2005). Reconstructing the cognitive world: The next step. Cambridge MA: MIT Press.

Consciousness and Cognition 43 (2016) 27–37 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Visual discrimination of delayed self-generated movement reveals the temporal limit of proprioceptive–visual intermodal integration Mark Jaime ⇑, Kelly O’Driscoll, Chris Moore Department of Psychology and Neuroscience, Dalhousie University, 1355 Oxford Street, Halifax, Nova Scotia B3H 4R2, Canada a r t i c l e i n f o Article history: Received 26 November 2015 Revised 29 March 2016 Accepted 11 May 2016 Keywords: Proprioceptive–visual integration Intermodal perception Visual proprioception a b s t r a c t This study examined the intermodal integration of visual–proprioceptive feedback via a novel visual discrimination task of delayed self-generated movement. Participants performed a goal-oriented task in which visual feedback was available only via delayed videos displayed on two monitors—each with different delay durations. During task performance, delay duration was varied for one of the videos in the pair relative to a standard delay, which was held constant. Participants were required to identify and use the video with the lesser delay to perform the task. Visual discrimination of the lesser-delayed video was examined under four conditions in which the standard delay was increased for each condition. A temporal limit for proprioceptive–visual intermodal integration of 3–5 s was revealed by subjects’ inability to reliably discriminate video pairs. Ó 2016 Elsevier Inc. All rights reserved. 1. Introduction The spatiotemporal integration of proprioception and vision specifies body self-awareness and its movements (e.g., Lewis & Brooks-Gunn, 1979; Mitchell, 1993; Rochat, 2012). For example, in the Rubber Hand Illusion (see Botvinick & Cohen, 1998) a rubber hand can be made to feel as if it belongs to one’s body by providing synchronous visual and proprioceptive feedback to the perceiver (e.g., Dummer, Picot-Annand, Neal, & Moore, 2009). However, asynchrony between proprioceptive, tactile, and visual feedback has been shown to disrupt the sense of body ownership (e.g., Lloyd, 2007; Palluel, Aspell, & Blanke, 2011; Pavani, Spence, & Driver, 2000; Petkova & Ehrsson, 2009). One example of this notion involves the perception of selfgenerated movements presented via a video monitor and delayed by some interval of time. In such cases the perceiver may report a feeling of disembodiment—as if his or her movements on the video do not feel self-initiated. Delayed video of self-generated movement typically has this effect because it disrupts the temporal synchrony between proprioceptive and visual feedback that is characteristic of the experience of self-generated movement. Thus, integrated visual–proprioception serves as an intersensory basis for experiential self-awareness—the phenomenal experience of being and acting in the world (Rochat, 2012). Despite a number of studies examining the integration of asynchronous audio–visual stimulation (e.g., Arrighi, Alais, & Burr, 2006; Conrey & Pisoni, 2006; Dixon & Spitz, 1980; Grant, Wassenhove, & Van Poeppel, 2004; Hillock, Powers, & Wallace, 2011; Hillock-Dunn & Wallace, 2012; Lewkowicz, 1996, 2010; Maier, Di Luca, & Noppeney, 2011; Powers, ⇑ Corresponding author at: Division of Science, Indiana University-Purdue University – Columbus, 4601 Central Avenue, Columbus, IN 47203, United States. E-mail address: mjaime@iupuc.edu (M. Jaime). http://dx.doi.org/10.1016/j.concog.2016.05.002 1053-8100/Ó 2016 Elsevier Inc. All rights reserved. 28 M. Jaime et al. / Consciousness and Cognition 43 (2016) 27–37 Hillock, & Wallace, 2009; Spence & Squire, 2003; van Wassenhove, Grant, & Poeppel, 2007; Vatakis, Navarra, Soto-Faraco, & Spence, 2007; Zampini, Shore, & Spence, 2003), relatively few studies have examined the integration of proprioceptive–visual asynchrony. There is however evidence suggesting that, like audio–visual asynchrony, perceptual systems can bind small proprioceptive–visual asynchronies into a unified percept (e.g., Jaime, Longard, & Moore, 2014; Leube et al., 2003; Spence & Squire, 2003). For example, in an fMRI study that used video delays of self-generated movement to probe the neural substrates of experiential self-awareness, Leube et al. (2003) demonstrated that delays less than 80 ms were predominantly reported incorrectly by adults as being live (synchronous). More recently, Jaime et al. (2014) examined proprioceptive–visual asynchrony detection in children using a similar task (but outside the MRI scanner) and demonstrated that the threshold for the integration of asynchronous proprioceptive–visual feedback appears to change with age. For example, children’s threshold (6- to 8-year-olds) ranged within 100–200 ms whereas the threshold for adults was within 52–100 ms. The kind of threshold discussed above represents the lower temporal threshold for proprioceptive–visual integration, which may serve to bind proprioceptive and visual information presented to the perceptual systems at slightly different times. Theoretically, there is also an upper temporal threshold, one that represents the maximum amount of temporal lag between proprioceptive and visual feedback that can be tolerated during intermodal matching of their spatiotemporal properties (see Fig. 1). This threshold is best exemplified by child studies that have modified the classic self-recognition paradigm (Amsterdam, 1972) with delayed video feedback, thereby creating a temporal asynchrony between what is felt (proprioception) and seen (vision) during self-generated movement. For example, in one study Miyazaki and Hiraki (2006) surreptitiously placed a sticker on a child’s forehead while the child observed a live video stream of her own moving self that was delayed by 1 or 2 s. In order to pass the delayed self-recognition task the child had to effectively match her proprioceptive feedback with her subsequent visible movements in the delayed video. The child was considered to have self-recognized if she made attempts to touch or remove the sticker from her forehead—an indication that she realized that her delayed visible movements corresponded with her previously registered proprioceptive feedback. Results revealed that 4-year-olds, but not 3-year-olds, were able to self-recognize when the video delay (i.e., proprioceptive–visual asynchrony) was 2 s in duration. However when the video delay was decreased to 1 s, 3-year-olds were also able to self-recognize. These findings suggest that children’s ability to self-recognize may initially hinge on their ability to perceive the spatiotemporal correspondence between proprioception and vision. However, whereas some degree of asynchrony between proprioception and vision can be integrated to specify self-related information, there is a temporal threshold at which point the asynchrony between proprioception and vision may be too long for the child to detect spatiotemporal correspondence—an upper temporal threshold for proprioceptive–visual intermodal integration (Fig. 1). Moreover, the fact that 3-year-olds were able to self-recognize when the video delay was decreased from 2 s to 1 s may also suggest that the upper temporal threshold for proprioceptive–visual intermodal integration increases with age. We speculate that a potential mechanism underlying proprioceptive–visual intermodal integration during Miyazaki and Hiraki’s (2006) delayed self-recognition task may involve working memory for proprioceptive information. Consider that in order to solve the delayed self-recognition task one has to retain proprioceptive information for some interval of time and then compare that information to subsequent visual feedback. For example, if a real time video stream of one’s movements is delayed by 2 s, one would presumably have to store in memory proprioceptive information of those movement dynamics for at least 2 s in order to determine if the visible movements correspond to the previously felt movements. It is therefore plausible that the 4-year-olds in Miyazaki and Hiraki’s (2006) study were able to retain proprioceptive information for the full 2 s interval, but the 2- and 3-year-olds ‘‘forgot” their felt movements within 2 s. In other words, greater proprioceptive working memory capacity could therefore have mediated older children’s ability to self-recognize. Fig. 1. Conceptual diagram of the lower and upper proprioceptive–visual integration thresholds. The arrow represents an increase of the proprioceptivevisual asynchrony interval from left to right. M. Jaime et al. / Consciousness and Cognition 43 (2016) 27–37 29 To date, the temporal limits for the retention of proprioceptive information remains poorly understood. What we do know from early work that has assessed memory for proprioception via the recall of simple motor tasks in adults is that proprioceptive memory shows a rapid decay over retention intervals ranging between 5 and 120 s (Adams & Dijkstra, 1966). Other early studies have also demonstrated that the encoding process for kinesthetic information via proprioception is independent from the encoding process for kinesthetic information via vision. For example, whereas visual kinesthetic information shows little decay over a 20 s retention interval, its rate of decay is significantly increased by interpolation (e.g., the classification of digit pairs into high or low and odd or even). By contrast, the relatively rapid rate of decay for proprioceptive information does not appear to be affected by interpolation and shows rapid decay whether or not the retention interval is unfilled (Posner, 1967; Posner & Konick, 1966). However, the retention of proprioceptive information does appear to be subject to interference when the interpolated task involves the processing of proprioceptive information (Williams, Beaver, Spence, & Rundell, 1969). It has therefore been posited that visual kinesthetic information shows little decay over unfilled intervals because the information lends itself to rehearsal (Posner, 1967) whereas kinesthetic information specified via proprioception is under less conscious awareness (Broadbent, 1958). Thus, young children’s inability to recognize themselves in videos delayed by 2 s (Miyazaki & Hiraki, 2006) may reflect both the fleeting and/or preconscious nature of proprioceptive memory and the fact that, during a bout of self-generated movement, the more recently perceived proprioceptive signals retroactively interfere with the retention of earlier proprioceptive signals. There are also two prior studies that have explored working memory for kinesthetic information with adults. One study adapted a memory span test (typically used for words) to assess the recall of discrete simple body movements. The authors reported that adults have a memory span limit of approximately 4 discrete movements (Smyth, Pearson, & Lindsay, 1988). In a later study, Miall, Haggard, and Cole (1995) examined the ability for a deafferented participant, and three neurologically normal participants, to make wrist movements between two visual targets. They were then were asked to pause for an interval of time ranging between 0 and 24 s and then resume their movements without visual feedback. The authors reported that the degree to which movement accuracy was affected when the visual feedback was removed increased as a function of the amount of time the participants maintained the paused position. There was also a degradation of the retained information within a period of 6 s. However, it is important to note that the above studies examined the recall of discrete body movement information when the information was encoded via the visual modality (Smyth et al., 1988) or via both proprioceptive and visual modalities (Miall et al., 1995). To date, no study has quantified the duration of proprioceptive representations of unrehearsed dynamic movements that are encoded only via proprioception. In this study, we were interested in quantifying the temporal limits of proprioceptive–visual integration. We therefore developed a visual discrimination paradigm for delayed self-generated movement that served as a proxy measure of proprioceptive–visual intermodal integration. Participants were asked to perform a goal-oriented motor task, best performed through visual guidance. However, online visual feedback was available only via video that was played on two side-byside monitors and was delayed by several seconds. Participants were asked to use the video with the lesser of the two delays to guide their movements. As in the self-recognition work using delayed videos described above (Miyazaki & Hiraki, 2006), this task presumably required intermodal integration between proprioceptive feedback and the subsequent corresponding visible movements. We developed this visual discrimination paradigm instead of a self-recognition task because participants might be more likely to use explicit knowledge to solve a delayed self-recognition task. Since participants in this task did not have to self-recognize, there was a lesser likelihood that they would use explicit knowledge and a greater likelihood that they would rely on proprioceptive–visual intermodal integration to perform the task. In addition, we chose to test adults instead of young children because of the present task’s complexity relative to the delayed self-recognition task. That is, young children might have difficulty comprehending the task’s instructions. They might not understand how to choose the lesser of the two delays or be able to refrain from looking down at their hands. Furthermore, 2- to 4-year olds might also not have the fine motor ability or attentional capacity required to effectively perform the task. Nonetheless, quantifying the temporal limits of proprioceptive–visual integration is important for a better understanding of the intersensory basis of selfawareness—and given that no other studies have quantified the temporal limits of proprioceptive-visual intermodal integration—testing adults provides a good starting point by which future developmental studies can expand on. Given the novelty of this paradigm, we had no specific a priori hypotheses for what the exact temporal limit of proprioceptive–visual integration might be. However, based on pilot testing we predicted that visual discrimination performance would show the following patterns: (1) Greater differences in delay between videos results in better visual discrimination performance. (2) Large asynchrony intervals between proprioceptive and visual feedback results in a decline in visual discrimination performance above and beyond the amount of difference in delay between videos because the amount of asynchrony between proprioceptive and visual feedback is too long for effective intermodal matching. 2. Materials and methods 2.1. Participants A total of 96 adults (71 females, 25 males) were tested and recruited through the undergraduate subject pool at a Canadian university. Participants received either course credit or a small monetary incentive for their participation. All participants had normal or corrected vision. Study procedures were conducted in compliance with the university’s 30 M. Jaime et al. / Consciousness and Cognition 43 (2016) 27–37 Table 1 Mean age (in years), age range, and gender distribution of participants per condition. Condition N Mean age Age range Females Males 1 2 3 4 21 24 25 26 20.7 20.6 23.6 21.5 18–27 18–37 18–38 18–36 16 22 17 16 5 2 8 10 Totals 96 71 25 institutional review board. Each participant was assigned to one of four conditions. Table 1 provides a summary of the participants’ sex and age distributions. 2.2. Apparatus Participants were seated at a small square table (52 cm in length, 52 cm in width, and 61 cm in height). Two metal nuts measuring 1.5 cm in diameter, 2 metal nuts measuring 2 cm in diameter, four small Styrofoam cups, and a pair of metal tongs (measuring 20 cm in length) with rubber ends were placed on top of the table. Participants’ direct view of the table was obstructed by a clothed partition (35 cm in height) that was mounted on the side of the table facing the adult. The height of the chair in which participants were seated was adjusted such that the table could not be viewed even if participants gazed down at the table. The clothed partition had two small slits that allowed participants to insert either their left of right arm depending on their handedness. Two 48 cm LCD computer monitors were positioned side-by-side approximately 30 cm in front of the table, directly facing the participant. The height of the monitors was positioned such that the participants seated behind the clothed partition had an unobstructed view of the monitors even while their view of the table was occluded. Each monitor displayed a top view of the table, which was captured by a video camera positioned above the table. The video camera output was split and processed through two video delay processors (D1 Pipeline, Prime Image, Inc.) before being displayed on the monitors. The video delay processors allowed for delays in the order of standard video frames or seconds. It is important to note that the D1 Pipeline produced a small propagation delay that was less than one frame (or 0.03 s). Therefore, delays are reported according to the D1 Pipeline’s settings and may contain additional delay in the signal. A second video camera was positioned in between and slightly behind the two monitors. This second camera was used to record the participants’ looks to either the left or right monitor. An image of the experimental setup is provided in Fig. 2. 2.3. Video stimuli The stimuli used in this study consisted of two side-by-side delayed videos of the participant’s hand and arm during task performance. One of the two delayed videos was termed the standard delay because the delay duration never changed across trials. The other delayed video was termed the varying delay because its delay duration changed across trials. Thus, for every presentation of the two side-by-side delayed videos the delay in one of the videos was held constant and the delay in the other video was varied. The position (left or right) of the standard or varying delayed videos was randomized across trials. 2.4. Task Prior to performing the task, informed consent was obtained from the participants. Then the participants were asked to complete the Edinburgh Handedness inventory in order to identify each participant’s preferred hand (Oldfield, 1971). The task involved using the tongs with the dominant hand to grasp and place one metal nut into each of the four different Styrofoam cups. Because participants’ direct view of their movements was occluded by the cloth partition, visual feedback was only available in the two side-by-side monitors (Fig. 2). To ensure that the participants did not become aware of the purpose of the experiment, they were told that the objective of the study was to assess hand-eye coordination and that they should perform the task as quickly and accurately as possible. Based on pilot testing, we determined that shorter delay durations decreased task difficulty, whereas longer delay duration increased task difficulty because the longer the delay in visual feedback the slower one had to move in order to perform the task as accurately as possible. Therefore, participants were informed that the two video monitors would each show differing delay durations over a series of trials and they should use the video with the lesser delay to make the task easier. Participants were also instructed not to attempt to stand up and lean over the partition while performing the task and there were no participants who did this. An entire session consisted of 24 individual trials, each lasting 30 s in duration. The onset and offset of each trial was signalled by the sound of a bell. Prior to the onset of each trial, subjects were instructed to keep their hand outside the partition until signalled to begin. At the end of each trial, subjects were instructed to put down the tongs regardless of whether they had completed the task or not (i.e., placed one metal nut in each of the four cups). After each trial there was approximately a 10 s pause while an experimenter arranged the materials on the table to their starting position and reset the video delay processor settings before commencing the subsequent trial. If the task was completed in less than 30 s, the trial was 31 M. Jaime et al. / Consciousness and Cognition 43 (2016) 27–37 MONITORS TABLE PARTITION Fig. 2. Experimental setup. Arial view of the table with the clothed partition and the two side-by-side monitors displaying participant’s visual feedback. terminated early. If a metal nut fell on the floor, an experimenter would immediately place another metal nut near the edge of the table surface. 2.5. Study design This study consisted of four conditions carried out in consecutive order in which both the standard and varying delays differed across conditions. For example, in Condition 1 the standard delay was held at 1 s, in the Condition 2 the standard was held at 2 s, in Condition 3 the standard was held at 3 s, and in Condition 4 the standard was held at 6 s. In each condition, there were also six different varying delays presented. The varying delays were calculated by subtracting the following amounts of delay from the standard delay: 1 s, 0.67 s, 0.50 s, 0.33 s, 0.04 s, and 0 s—from here on we refer to these delay amounts as contrasts. A contrast simply equates to the amount of delay difference between the standard and varying delay. For example, in Condition 1 there were six different standard vs. varying delayed video pairings or contrasts: (1) 1 s vs. 0 s (or no delay) pairing, (2) 1 s vs. 0.33 s pairing, (3) 1 s vs. 0.50 s pairing, (4) 1 s vs. 0.66 pairing, (5) 1 s vs. 0.96 s pairing, and (6) 1 s vs. 1 s pairing (or no contrast). Thus, the varying delay was typically less than the standard delay except for when both video delays were the same. It is also important to note that contrast was consistent across conditions. That is, although the standard and varying delay durations were different in each condition the contrast remained the same across all conditions. There were 4 trials for each of the 6 contrasts thereby yielding 24 individual trials per participant. There order of contrasts were randomized across trials. A summary of the study design is provided in Table 2. 2.6. Dependent measures Our primary dependent variable was visual discrimination. This variable was used as a proxy measure of proprioceptive– visual intermodal integration. Visual discrimination of self-generated movement was operationalized as the mean proportion of looking time (MPL) to the monitor showing the varying delayed video. To obtain the MPL measure, individual looks were coded via frame-by-frame analysis of the video recordings for each of the 24 trials and per each participant. Given that the first few initial looks to either monitor may just reflect chance looking in either direction, we excluded the first half of the number of looks from the analysis in order to reduce the amount of random variation in the scores. For example, if a particular 30 s trial contained a total number of 6 looks, then we excluded the first 3 looks from the analysis. If the number Table 2 Delayed video pairings per condition (in seconds). Standard = video monitor with unchanging delay duration. Varying = video monitor with changing delay duration. Contrast = standard minus varying. Contrasts 1 0.67 0.50 0.33 0.04 0 Condition 1 Condition 2 Condition 3 Condition 4 Standard Varying Standard Varying Standard Varying Standard Varying 1 1 1 1 1 1 0 0.33 0.50 0.66 0.96 1 2 2 2 2 2 2 1 1.33 1.50 1.66 1.96 2 3 3 3 3 3 3 2 2.33 2.50 2.66 2.96 3 6 6 6 6 6 6 5 5.33 5.50 5.66 5.96 6 32 M. Jaime et al. / Consciousness and Cognition 43 (2016) 27–37 of looks happened to be an odd number (say 3 looks) then we excluded the lesser first half of the number of looks (e.g., 2) for that trial. We chose to remove the first half of the number in order increase the likelihood that all subjects had sufficient practice in each trial. Accordingly, the MPL scores were computed by dividing the total number of frames in which the subject was looking at the monitor displaying the varying delay by the total number of frames in which the subject was looking at both monitors (varying + standard). Thus, MPL scores that are closer to 1.0 represent better visual discrimination. MPL scores were computed across each of the four trials in each contrast. There were six MPL scores for each participant—one for each contrast level. A task performance response variable was also obtained via a review of the participants’ recorded videos. Task performance was operationalized as the overall total number of metal nuts placed in the cups for the entire testing session (across the 24 trials). Thus, greater numbers represent better performance on the task. This measure was obtained in order to examine the extent to which task performance was being affected by the delay manipulations in each condition and whether it was associated with visual discrimination. 3. Results We used SAS version 9.4 with PROC MIXED to perform a linear mixed effects analysis of the effects of condition and contrast on MPL. We used a random intercepts model with a covariance structure of variance components. Condition and contrast (and their interaction effects) were entered as fixed effects and a random effect of subjects was entered into the model due to the fact that multiple measurements were taken within-subjects. Statistical significance (p < 0.05) was obtained by restricted maximum likelihood (REML) tests. A total of 96 subjects each performed 24 trials, yielding a total of 2304 observations. However, 5 trials were not coded and included in the analysis due to lost video footage (e.g., camera batteries died). Therefore, a total of 2299 observations were included in the final analysis. Results of the mixed effects analysis revealed a significant effect of contrast, F(5, 2183) = 26.61, p < 0.0001, condition, F(15, 2183) = 3.64, p < 0.0001, and a contrast by condition interaction, F(3, 92) = 23.78, p < 0.0001, on the MPL scores. The main effects of contrast and condition were then examined with a series of multiple comparisons (Tukey-corrected). With respect to the effect of contrast, results showed that the MPL scores for the 0.33 s, 0.5 s, 0.67 s, and 1 s contrasts were significantly greater than the MPL scores for the 0 s contrast, t(2183) = 4.36, p = 0.0003, t(2183) = 6.03, p < 0.0001, t (2183) = 6.81, p < 0.0001, and t(2183) = 6.64, p < 0.0001, respectively. MPL scores for the 0.33 s, 0.5 s, 0.67 s, and 1 s contrasts were also significantly greater than the MPL scores for the 0.04 s contrast, t(2183) = 6.02, p < 0.0001, t(2183) = 7.69, p < 0.0001, t(2183) = 8.47, p < 0.0001, and t(2183) = 8.30, p < 0.0001, respectively. Together, these results indicate that visual discrimination improved for contrast amounts that were greater than 0.04 s. This pattern of results can be visually inspected in Fig. 3. Fig. 3. Mean proportion of looks (MPL) to the varying delay monitor plotted as a function of contrast amount. Error bars represent standard error of the mean. M. Jaime et al. / Consciousness and Cognition 43 (2016) 27–37 33 Condition affected visual discrimination such that there was a decline in the MPL scores for every subsequent condition. As also indicated by Fig. 3, one can see that the MPL scores for the 1 s standard condition were significantly greater than MPL scores for the 2 s standard condition, t(92) = 4.60, p < 0.0001, the 3 s standard condition, t(92) = 4.16, p = 0.0002, and the 6 s standard condition, t(92) = 8.43, p < 0.0001. In addition, the MPL scores in the 2 s and 3 s conditions were significantly greater than the MPL scores in the 6 s condition, t(92) = 3.88, p = 0.0007 and t(92) = 4.42, p < 0.0001, respectively. However, the MPL scores did not differ between the 2 s and 3 s conditions, t(92) = 0.623, p = 0.962. To unpack the contrast by condition interaction, we first examined the effect of condition on MPL at each of the six levels of contrast. Mean proportion of looks significantly differed between conditions at contrast levels of 0.33 s, 0.50 s, 0.67 s, and 1 s, F(3, 2183) = 7.66, p < 0.0001, F(3, 2183) = 12.36, p < 0.0001, F(3, 2183) = 14.7, p < 0.0001, and F(3, 2183) = 10.4, p < 0.0001, respectively. However, MPL did not differ between conditions when there was no contrast between delayed videos (0 s contrast), F(3, 2183) = 0.70, p = 0.555, or when the contrast between delayed videos was very small (0.04 s contrast), F(3, 2183) = 0.39, p = 0.758. Second, we examined the effect of each condition on the relationship between MPL and contrast. Thus, four separate simple linear regressions—one for each condition—were performed. As shown in Fig. 4, contrast was a significant predictor of MPL in Conditions 1, 2, and 3, F(1, 502) = 112, p < 0.0001, R2 = 0.128, F(1, 574) = 18.6, p < 0.0001, R2 = 0.031, and F(1, 594) = 25.5, p < 0.0001, R2 = 0.041, respectively. However, in Condition 4 contrast was not a predictor of MPL, F (1, 622) = 2.06, p = 0.151, R2 = 0.003. To evaluate under which level of contrast and condition visual discrimination performance was occurring beyond chance level (>0.50), a series of one-sample t-tests were applied to the MPL scores at each contrast level and per each condition Fig. 4. Scatterplots (with regression lines) of mean proportion of looks (MPL) as a function of contrast for the 1 s standard (upper left), 2 s standard (upper right), 3 s standard (lower left), and 6 s standard (lower right) conditions. 34 M. Jaime et al. / Consciousness and Cognition 43 (2016) 27–37 Table 3 Mean proportion of looks (MPL), standard deviations (SD), and parameter estimates for each condition and at each contrast level. * Condition Contrast MPL SD Estimate 1s None 0.04 s 0.33 s 0.50 s 0.67 s 1.00 s 0.53 0.46 0.77* 0.83* 0.90* 0.85* 0.36 0.38 0.32 0.25 0.17 0.26 0.24 0.34 0.05 0.02 0.05 0.55 2s None 0.04 s 0.33 s 0.50 s 0.67 s 1.00 s 0.55 0.50 0.59* 0.64* 0.71* 0.67* 0.37 0.36 0.36 0.35 0.32 0.34 0.05 0.12 0.05 0.01 0.04 0.55 3s None 0.04 s 0.33 s 0.50 s 0.67 s 1.00 s 0.53 0.46 0.66* 0.73* 0.65* 0.71* 0.40 0.38 0.37 0.34 0.38 0.34 0.10 0.20 0.01 0.06 0.05 0.55 6s None 0.04 s 0.33 s 0.50 s 0.67 s 1.00 s 0.48 0.50 0.52 0.52 0.54 0.55 0.40 0.41 0.41 0.41 0.41 0.41 0.55 0.55 0.55 0.55 0.55 0.55 MPL is significantly greater than chance level (0.50) at p < 0.05 (one sample t-test, two-tailed). (with a Bonferroni correction). Results revealed that in Condition 1, MPL scores were significantly greater than chance level for the following contrasts: 0.33 s, t(83) = 7.79, p < 0.0001, 0.5 s, t(83) = 12.3, p < 0.0001, 0.67 s, t(83) = 21.9, p < 0.0001, and 1 s, t(83) = 12.7, p < 0.0001. In Condition 2, MPL scores were significantly greater than chance level for the following contrasts: 0.33 s, t(95) = 2.50, p = 0.014, 0.5 s, t(95) = 4.01, p < 0.0001, 0.67 s, t(95) = 6.26, p < 0.0001, and 1 s, t(95) = 4.93, p < 0.0001. In Condition 3, MPL scores were significantly greater than chance level in the following contrasts: 0.33 s, t(98) = 4.39, p < 0.0001, 0.5 s, t(98) = 6.67, p < 0.0001, 0.67 s, t(99) = 3.94, p < 0.0001, and 1 s, t(98) = 6.05, p < 0.0001. In Condition 4, MPL scores were not significantly greater than chance level in any of the contrasts. Thus, although there was a decline across conditions, visual discrimination was above chance level for all conditions except in the 6 s standard condition (Condition 4). A summary of these results are provided in Table 3. To examine the effect of condition on task performance, a one-way ANOVA was performed on the total number of metal nuts placed in a cup with condition as the between-subjects factor. The results of this analysis revealed that condition had a significant effect on task performance, F(3, 92) = 10.4, p < 0.0001. Post hoc multiple comparisons (Tukey HSD) showed that the number of metal nuts placed in a cup was significantly greater in Condition 1 (M = 63) than in Condition 3 (M = 43) and Condition 4 (M = 44). The task response measure was also significantly greater in Condition 2 (M = 55) than in Conditions 3 and 4. These results suggest that the conditions with the longer delay durations increased task difficulty. Finally, the influence of task performance on visual discrimination was examined with a series of bivariate Pearson correlations (two-tailed) on the number of metal nuts placed in a cup and MPL (averaged across all trials). The results indicate that task performance was not associated with visual discrimination in any of the four conditions, r = 0.35, p = 0.132 (Condition 1), r = 0.14, p = 0.509 (Condition 2), r = 0.01, p = 0.975 (Condition 3), r = 0.22, p = 0.313 (Condition 4), and likely did not influence visual discrimination. 4. Discussion The objective of this study was to quantify the temporal limit of proprioceptive–visual intermodal integration. The temporal limit of proprioceptive–visual intermodal integration represents the maximum amount of temporal asynchrony between proprioception and vision possible for effective intermodal integration. Despite the paucity of research on the temporal limits of proprioceptive–visual integration, one prior study using delayed videos for self-recognition suggests that children may be able to integrate asynchronies between proprioceptive and visual feedback of up to two seconds (Miyazaki & Hiraki, 2006). However, it is not known whether they can integrate greater asynchrony intervals. It is also not known whether children used explicit knowledge or intermodal information to self-recognize. Thus, we designed a novel visual discrimination paradigm of delayed self-generated movement as a proxy measure of proprioceptive–visual intermodal integration because effective visual discrimination of delayed self-generated movement presumably requires intermodal integration between proprioceptive and visual feedback. To our knowledge, no other study has used this particular visual M. Jaime et al. / Consciousness and Cognition 43 (2016) 27–37 35 discrimination paradigm to examine proprioceptive–visual integration. Given that this paradigm did not involve selfrecognition, the use of explicit knowledge was unlikely. In this study four conditions were carried out in consecutive order. The purpose of Condition 1 was to establish a baseline pattern of visual discrimination. Throughout each participant’s testing session in Condition 1, the amount of difference in delay between the two videos (or contrast) was gradually increased from having no contrast, where the delay in each of the two videos was exactly the same, to a 1 s contrast between delays, where the amount of delay in one video differed by one second from the other (Table 2). We expected that contrast amount would affect visual discrimination such that greater contrast between delayed videos facilitates better visual discrimination. In line with this prediction, the results of Condition 1 demonstrate that visual discrimination performance was at chance level for delay pairings with no contrast (0 s) or at very low contrast amounts (0.04 s), but statistically above chance level at greater contrast levels. Given the pattern of results in Condition 1, Conditions 2, 3, and 4 were subsequently conducted to examine the effect of the size of the proprioceptive–visual asynchrony interval on visual discrimination. We increased the amount of overall delay in both videos with each subsequent condition but maintained the contrasts between the delayed video pairings consistent. This allowed us to unpack the effect of contrast from proprioceptive–visual asynchrony. For example, if proprioceptive– visual integration is involved in the visual discrimination of the delayed videos then an increase in the asynchrony interval should affect visual discrimination performance. Visual discrimination performance was indeed best in Condition 1—which had the shortest proprioceptive–visual asynchrony interval—but then showed a progressive decline with each subsequent condition (i.e., increasing proprioceptive–visual asynchrony). Additionally, there was a gradual decrease in the slope of the regression line in each subsequent condition suggesting that visual discrimination performance was affected by the size of the proprioceptive–visual asynchrony interval (Fig. 4). We propose that the temporal limit of proprioceptive–visual intermodal integration was revealed by the inability for participants to discriminate between the two delayed videos of self-generated feedback, even when the contrast amounts were high, because the proprioceptive–visual asynchrony interval (delay) was too long for effective intermodal integration. This was evidenced by the chance-level visual discrimination performance in the final condition at all levels of contrast. Thus, the results of Condition 4 indicate that the proprioceptive–visual asynchrony interval had exceeded the temporal limit for intermodal integration. Taken together, this study’s results suggest that the temporal limit for proprioceptive–visual intermodal integration may be between 3 and 5 s. One possible explanation for the existence of a temporal limit for proprioceptive–visual intermodal integration may be that delays between 3 and 5 s in duration are beyond the retention capacity of proprioceptive memory. That is, if proprioceptive–visual intermodal integration requires the ability to hold proprioceptive feedback in working memory then a video with 6 s of delay, for example, likely requires participants to hold their movements in proprioceptive working memory for at least 6 s—a duration that is too long to retain in memory. Although the temporal limits of iconic and echoic sensory memory are well known (see Coltheart, 1980; Cowan, 1984 for reviews), the exact retention limit of proprioceptive working memory has not been quantified. However, prior work that has examined memory for body movements has reported that 4 discrete body movements performed by a model every 1.5 s is the retention limit of kinesthetic information (Smyth et al., 1988). Interestingly, if this span of body movements is converted into a retention interval (6 s), its value exceeds the temporal limit for proprioceptive-visual integration revealed in this study. Miall et al. (1995) have also demonstrated that self-generated visuo-motor information previously encoded via vision and/or proprioception degrades within 6 s of the removal of visual feedback. Nonetheless, caution must be taken in relating the above studies with the current findings, specifically because the above studies examined the capacity to reproduce movements that were encoded via vision and/or proprioception; the temporal duration of kinesthetic information encoded only via proprioception was not examined. Thus, the current study is unique in that participants did not have to reproduce modeled movements. Instead, the paradigm elicited encoding of their movement dynamics via proprioception so as to match those movements to the delayed visual feedback—what we have defined here as proprioceptive–visual intermodal integration. Given that integrated visual–proprioceptive feedback is a characteristic of the first-person experience, the ability to detect spatiotemporal correspondence between asynchronous proprioceptive and visual feedback reflects and experiential awareness of self-generated movement. Our results suggest that human perceptual systems can detect spatiotemporal correspondence between proprioception and vision despite relatively long proprioceptive–visual asynchrony intervals. However, there are limits to the amount of asynchrony tolerated by the perceptual system, which may serve to decrease the salience of ambiguous multisensory signals in the sensory array and increase the salience of self-generated multisensory feedback (Ernst & Bülthoff, 2004; Lalanne & Lorenceau, 2004). The temporal limits of proprioceptive–visual integration revealed in this study may have implications for early selfawareness. Prior research suggests that young children may initially rely on the tight temporal coupling between proprioceptive and visual feedback for self-recognition (Povinelli & Simon, 1998). For example, Miyazaki and Hiraki (2006) have demonstrated that 2- and 3-year-olds, but not 4-year-olds, have difficulty associating visual feedback that is delayed by 2 s to the corresponding movements of their current self. Thus, it appears that younger children may not be able to detect the spatiotemporal correspondence between asynchronous proprioceptive and visual feedback whereas the older children can. One potential explanation for the age-related differences in the Miyazaki and Hiraki (2006) study is that the temporal limit of proprioceptive-visual integration may vary between 3 and 4 years of age. In other words, 2 s of asynchrony between proprioceptive and visual feedback may be beyond the temporal limit for intermodal integration in 2- to 3-year olds but still within the temporal limit of intermodal integration in 4-year-olds. Interestingly, the temporal limit for proprioceptive–visual 36 M. Jaime et al. / Consciousness and Cognition 43 (2016) 27–37 intermodal integration of 3–5 s (revealed in this study) may suggest that the temporal limit for proprioceptive-visual intermodal integration may be greater for adults than for young children. Although the notion of a developmental change in the upper temporal limit for proprioceptive-visual intermodal integration has yet to be examined empirically, prior research that has examined the lower temporal threshold for proprioceptive–visual integration has demonstrated an age-related change (Jaime et al., 2014). Therefore it is plausible that the upper temporal limits for proprioceptive–visual integration may also show age-related changes. There are limitations to the current study. First, male participants were underrepresented in our sample. However, given that no other studies have examined sex differences in proprioceptive–visual intermodal integration it is difficult to speculate on how the underrepresentation of males relative to females might have affected this study’s results. It is also important to note that because this paradigm did not have a recognition or recall response measure the use of proprioceptive memory in this paradigm is implied and thus it is possible that a different strategy might have been at play during visual discrimination. For example, participants could have chosen the lesser of the two delays based simply on which monitor displayed movement information first rather than using proprioceptive memory. Thus, future studies that aim to examine the retention limits of proprioceptive working memory should consider using a recall or recognition response variable. 5. Conclusions This study adds to a very scarce literature on the integration of vision and proprioception during the perception of selfgenerated movement (e.g., Leube et al., 2003; Jaime et al., 2014). Here we demonstrate that visual discrimination of delayed self-generated movement can serve to probe the temporal limits of proprioceptive–visual intermodal integration. Our findings suggest that human perceptual systems can integrate corresponding seen and felt movements with relatively large asynchrony intervals. However, the extent of intermodal integration is not without limits and depends on whether the size of the asynchrony interval is within the upper proprioceptive–visual integration threshold. That is, asynchrony intervals that are subthreshold are perceived as sharing the same spatiotemporal dynamics, whereas asynchrony intervals beyond the threshold may be too difficult to perceive spatiotemporal correspondence (Fig. 1). Given that the integration of vision with proprioception is a characteristic of the first-person experience of self-generated movement, the ability to integrate asynchronous visual and proprioceptive feedback reflects an experiential awareness of the moving and acting self. We posit that proprioceptive–visual intermodal integration plays an important role in body self-awareness because it directs attention to spatiotemporal properties that are invariant across vision and proprioception such as changes in directionality, impact, spatial co-location, acceleration, and relative position—properties that provide real time feedback about the physical status of the self. Acknowledgment This study was supported by a grant from the National Sciences and Engineering Research Council of Canada (NSERC). References Adams, J. A., & Dijkstra, S. (1966). Short-term memory for motor responses. Journal of Experimental Psychology, 71, 314–318. Amsterdam, B. (1972). Mirror self-image reactions before age two. Developmental Psychobiology, 5, 297–305. Arrighi, R., Alais, D., & Burr, D. (2006). Perceptual synchrony of audiovisual streams for natural and artificial motion sequences. Journal of Vision, 6, 260–268. Botvinick, M., & Cohen, J. (1998). Rubber hands ‘feel’ touch that eyes see. Nature, 391, 756. Broadbent, D. E. (1958). Perception and communication. London: Pergamon. Coltheart, M. (1980). Iconic memory and visible persistence. Perception & Psychophysics, 27, 183–228. Conrey, B., & Pisoni, D. B. (2006). Auditory-visual speech perception and synchrony detection for speech and nonspeech signals. The Journal of the Acoustical Society of America, 119, 4064–4073. Cowan, N. (1984). On short and long auditory stores. Psychological Bulletin, 96, 341–370. Dixon, N. F., & Spitz, L. (1980). The detection of auditory visual desynchrony. Perception, 9, 719–721. Dummer, T., Picot-Annand, A., Neal, T., & Moore, C. (2009). Movement and the rubber hand illusion. Perception, 38, 271–280. Ernst, M. O., & Bülthoff, H. H. (2004). Merging the senses into a robust percept. Trends in Cognitive Science, 8, 162–169. Grant, K. W., Wassenhove, V., & Van Poeppel, D. (2004). Detection of auditory (cross-spectral) and auditory–visual (cross-modal) synchrony. Speech Communication, 44, 43–53. Hillock, A. R., Powers, A. R., & Wallace, M. T. (2011). Binding of sights and sounds: Age-related changes in multisensory temporal processing. Neuropsychologia, 49, 461–467. Hillock-Dunn, A., & Wallace, M. T. (2012). Developmental changes in the multisensory temporal binding window persist into adolescence. Developmental Science, 15, 688–696. Jaime, M., Longard, J., & Moore, C. (2014). Developmental changes in visual-proprioceptive integration in children. Journal of Experimental Child Psychology, 125, 1–12. Lalanne, C., & Lorenceau, J. (2004). Crossmodal integration for perception and action. Journal of Physiology, 98, 265–279. Leube, D. T., Knoblich, G., Erb, M., Grodd, W., Bartels, M., & Kricher, T. T. (2003). The neural correlates of perceiving one’s own movements. NeuroImage, 20, 2084–2090. Lewis, M., & Brooks-Gunn, J. (1979). Social cognition and the acquisition of self. New York: Plenum Press. Lewkowicz, D. J. (1996). Perception of auditory-visual temporal synchrony in human infants. Journal of Experimental Psychology, 22, 1094–1106. Lewkowicz, D. J. (2010). Infant perception of audio-visual speech synchrony. Developmental Psychology, 46, 66–77. Lloyd, D. M. (2007). Spatial limits on referred touch to an alien limb may reflect boundaries of visuo-tactile peripersonal space surrounding the hand. Brain & Cognition, 64, 104–109. Maier, J. X., Di Luca, M., & Noppeney, U. (2011). Journal of Experimental Psychology: Human Perception and Performance, 37, 245–256. M. Jaime et al. / Consciousness and Cognition 43 (2016) 27–37 37 Miall, R. C., Haggard, P. N., & Cole, J. D. (1995). Evidence of a limited visuo-motor memory used in programming wrist movements. Experimental Brain Research, 107, 267–280. Mitchell, R. W. (1993). Mental models of mirror-self-recognition: Two theories. New Ideas in Psychology, 11, 295–326. Miyazaki, M., & Hiraki, K. (2006). Delayed intermodal contingency affects young children’s recognition of their current self. Child Development, 77, 736–750. Oldfield, R. C. (1971). The assessment and analysis of handedness: The Edinburgh inventory. Neuropsychologia, 9, 97–113. Palluel, E., Aspell, J., & Blanke, O. (2011). Leg muscle vibration modulates bodily self-consciousness: Integration of proprioceptive, visual, and tactile signals. Journal of Neurophysiology, 105, 2239–2247. Pavani, F., Spence, C., & Driver, J. (2000). Visual capture of touch: Out-of-body experiences with rubber gloves. Psychological Science, 11, 353–359. Petkova, V. I., & Ehrsson, H. H. (2009). When right feels left: Referral of touch and ownership between the hands. PLoS One, 4(9), e6933. http://dx.doi.org/ 10.1371/journal.pone.0006933. Posner, M. I. (1967). Characteristics of visual and kinesthetic memory codes. Journal of Experimental Psychology, 75, 103–107. Posner, M. I., & Konick, A. F. (1966). Short-term retention of visual and kinesthetic information. Journal of Organizational Behavior and Human Performance, 1, 71–88. Povinelli, D. J., & Simon, B. B. (1998). Young children’s understanding of briefly versus extremely delayed images of self: Emergence of the autobiographical stance. Developmental Psychology, 34, 188–194. Powers, A., Hillock, A., & Wallace, M. T. (2009). Perceptual training narrows the temporal window of multisensory binding. Journal of Neuroscience, 29, 12265–12274. Rochat, P. (2012). Primordial sense of embodied self-unity. In V. Slaughter & C. Brownell (Eds.), Early development of body representations (pp. 1–18). Cambridge: Cambridge University Press. Smyth, M., Pearson, N., & Lindsay, R. (1988). Movement and working memory: Patterns and positions in space. The Quarterly Journal of Experimental Psychology, 40, 497–514. Spence, C., & Squire, S. (2003). Multisensory integration: Maintaining the perception of synchrony. Current Biology, 13, R519–R521. van Wassenhove, V., Grant, K. W., & Poeppel, D. (2007). Temporal window of integration in auditory-visual speech perception. Neuropsychologia, 45, 598–607. Vatakis, A., Navarra, J., Soto-Faraco, S., & Spence, C. (2007). Temporal recalibration during asynchronous audiovisual speech perception. Experimental Brain Research, 181, 173–181. Williams, H. L., Beaver, W. S., Spence, M. T., & Rundell, O. H. (1969). Digital and kinesthetic memory with interpolated information processing. Journal of Experimental Psychology, 80, 530–536. Zampini, M., Shore, D. I., & Spence, C. (2003). Audiovisual temporal order judgments. Experimental Brain Research, 152, 198–210.

Consciousness and Cognition 42 (2016) 93–100 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Only pre-cueing but no retro-cueing effects emerge with masked arrow cues Markus Janczyk a,⇑, Heiko Reuss b a b Eberhard Karls University of Tübingen, Department of Psychology, Germany Julius Maximilians University of Würzburg, Department of Psychology III, Germany a r t i c l e i n f o Article history: Received 4 November 2015 Revised 5 February 2016 Accepted 6 February 2016 Keywords: Retro-cue Pre-cue Attention Masked cueing a b s t r a c t The impact of masked stimulation on cognitive control processes is investigated with much interest. In many cases, masked stimulation suffices to initiate and employ control processes. Shifts of attention either happen in the external environment or internally, for example, in working memory. In the former, even masked cues (i.e., cues that are presented for a period too short to allow strategic use) were shown efficient for shifting attention to particular locations in pre-cue paradigms. Internal attention shifting can be investigated using retro-cues: long after encoding, a valid cue indicates the location to-be-tested via change detection, and this improves performance (retro-cue effect). In the present experiment, participants performed in both a pre- and a retro-cue task with masked and normally presented cues. While the masked cues benefitted performance in the pre-cue task, they did not in the retro-cue task. These results inform about limits of masked stimulation. Ó 2016 Elsevier Inc. All rights reserved. 1. Introduction The limits and possibilities of masked stimulation have been heavily researched and debated in recent years and are considered as ‘‘one of the most controversial issues in psychology” (Kouider & Dehaene, 2007, p. 857). Especially intriguing is the finding that processes other than simple motor activations, such as many typical executive functions, have been found to be enabled by and susceptible to influences of masked stimulation (Kunde, Reuss, & Kiesel, 2012). Contradicting classical conceptions of consciousness and control (e.g., Jack & Shallice, 2001), information that one is not aware of is able to induce the initiation of processes like task set activation (Lau & Passingham, 2007; Mattler, 2006; Reuss, Kiesel, Kunde, & Hommel, 2011), criterion setting (Reuss, Kiesel, & Kunde, 2015), conflict adaptation (Desender, Van Opstal, & Van den Bussche, 2014; Reuss, Desender, Kiesel, & Kunde, 2014), and shifts of external attention (Palmer & Mattler, 2013). In the present study, we investigated the effect of masking retro-cues on shifts of internal attention within working memory (WM). We will continue with giving a brief overview over external attention shifting, with a special focus on masked cues, and then will introduce the particular retro-cue paradigm employed in our experiment. ⇑ Corresponding author at: Eberhard Karls University of Tübingen, Department of Psychology, Cognition and Action, Schleichstraße 4, 72076 Tübingen, Germany. E-mail address: markus.janczyk@uni-tuebingen.de (M. Janczyk). http://dx.doi.org/10.1016/j.concog.2016.02.003 1053-8100/Ó 2016 Elsevier Inc. All rights reserved. 94 M. Janczyk, H. Reuss / Consciousness and Cognition 42 (2016) 93–100 1.1. External attention shifting and masked cues External shifts of attention are investigated by spatially cueing target locations before target onset, either by presenting a (so-called exogenous) cue directly at the potential target location, or by presenting (so-called endogenous) central cues that are mapped to particular locations (Folk, Remington, & Johnston, 1992; Jonides, 1981; Müller & Rabbitt, 1989; Posner, 1980; Posner & Cohen, 1984; Theeuwes, 1991; Yantis & Jonides, 1990). Comparing performance after invalid cues (i.e., the cued location was not the target location) and after valid cues allows to indirectly measure whether attention was impacted by the cues. Faster response times with valid cues compared to invalid cues are thought to reflect shifts of attention to the cued location. Valid cues allow for fast processing of the target as attention already resides at the target location, whereas invalid cues demand a reorienting of attention before the target can be processed. Exogenous cues lead to shifts of attention even when they are (1) not predictive of the target location (e.g., Folk & Remington, 1998, 1999), that is, when valid cues occur equally often as invalid cues, and (2) when they are presented in a way that renders them invisible to the participant (e.g., Ansorge & Heumann, 2006; Ansorge & Neumann, 2005; Eimer, 1997; Mulckhuyse, Talsma, & Theeuwes, 2007). In contrast, endogenous cues (1) usually need to be predictive (i.e., more often valid than invalid) for participants to orient their attention according to the cues, and (2) their effectiveness under masked conditions is very limited (Reuss, Kiesel, Kunde, & Wühr, 2012). This reflects a necessary strategical, conscious effort, which is prevented by masking (cf. Daza, Ortells, & Fox, 2002; Dehaene & Naccache, 2001; Merikle & Joordens, 1997; Merikle, Joordens, & Stolz, 1995). Yet, centrally presented cues that possess an inherent or overlearned spatial meaning, like gaze cues or arrow cues, have been found to orient attention under masked conditions as well (Gayet, Van der Stigchel, & Paffen, 2014; Reuss, Pohl, Kiesel, & Kunde, 2011). For example, in the study by Reuss and colleagues, arrow cues were randomly presented either visible or masked. In a first experiment, the overall validity of the arrow cues was 50%, so that they were in fact not predictive of the target location. While visible cues still led to shifts of attention (thus akin to the automatic effect of exogenous cues), masked cues did not impact on attention. However, when the overall cue validity was raised so that the cues were predictive in a second experiment, both visible and masked cues oriented attention. The authors theorized that the effectiveness of the masked cues depends on the intent to utilize the arrow cues, which is only provided when the cues are predictive (see also Gayet et al., 2014). From this point of view, the effect of masked arrow cues can be characterized as automatic in contrast to a truly endogenously driven shift of attention, but this kind of automaticity is constrained by the top–down contingency of the effect. There are also findings that even unpredictive unconscious arrow cues lead to validity effects (Gabay, Avni, & Henik, 2012). Notably, this study did not use masked cues, but employed the inattentional blindness paradigm (Mack & Rock, 1998), in which the stimulus is in principle clearly visible, but remains unconscious because it is not attended to. Overall, one can conclude that arrow cues indeed orient spatial attention even when presented in a way that diminishes awareness of the cue’s identity and prevents its strategic use. 1.2. Internal attention shifting in working memory As we have seen so far, directing attention to particular locations in the external environment enhances performance if subsequent targets appear at these locations. However, attention can also act on internally represented information, for example, in memory (see Chun, Golomb, & Turk-Browne, 2011, for a review). Here we use the fact that attention can be shifted internally within visual WM, even long after objects have been encoded into visual WM and are not present in the sensory environment any longer (Griffin & Nobre, 2003; Landman, Spekreijse, & Lamme, 2003). This is typically being shown by using the retro-cue paradigm: Participants encode objects into visual WM and following a retention interval (>500 ms to exclude effects of iconic memory; Sperling, 1960), a cue indicates one spatial location that is subsequently probed via change detection. The robust finding is a performance benefit following valid than following neutral (noninformative) cues; the retro-cue effect (RCE; e.g., Astle, Summerfield, Griffin, & Nobre, 2012; Berryhill, Richmond, Shay, & Olsen, 2011; Makovski, Sussman, & Jiang, 2008; Matsukura, Luck, & Vecera, 2007). An ongoing debate revolves about whether pre- and retro-cueing involve similar or distinct mechanisms (e.g., Astle, Nobre, & Scerif, 2012; Griffin & Nobre, 2003; Li & Saiki, 2014; Makovski & Jiang, 2007; Nobre et al., 2004; Shimi, Nobre, Astle, & Scerif, 2014; Tanoue & Berryhill, 2012) and commonalities but also differences have been reported. The potential effects of masked retro-cues are thus of great theoretical interest for multiple reasons: To investigate the limits of masked stimulation, and to gain insight into the underlying mechanisms of retro-cuing and its relation to pre-cueing. 1.3. The present experiment As described earlier, arrow pre-cues impact on attention in a non-endogenous way, which is indicated by the observation that (1) visible arrow cues shift attention automatically, even when they are not predictive (e.g., Pratt, Radulescu, Guo, & Hommel, 2010), and (2) the effectiveness of masked arrow cues, although predictiveness seems to play a role here (Gayet M. Janczyk, H. Reuss / Consciousness and Cognition 42 (2016) 93–100 95 et al., 2014; Reuss et al., 2011). Because this appears to be a property of arrows, the same effect would be expected for shifting internal attention in WM with arrow retro-cues.1 Yet, whether this were the case is not entirely clear in light of recent results. For example, several studies investigated whether the RCE suffers from dual-task interference. Some studies applied a secondary task between cue offset and the change detection screen and reported no changes in the magnitude of the RCE (Hollingworth & Maxcey-Richard, 2013; Rerko, Souza, & Oberauer, 2014). In contrast, when the secondary task appeared in close temporal proximity to or before the cue, the RCE became reliably smaller (Janczyk & Berryhill, 2014). Thus, retro-cue identification (rather than the processes initiated by the retro-cue) might be critical, suggesting that it is necessary to perceive the retro-cue in a conscious, supraliminal way. While the role of cue awareness with shifting of external attention has already been investigated (Palmer & Mattler, 2013; Reuss et al., 2011, 2012), there exists no such research with retro-cues. In our experiment, we used a retro-cue task with normal and masked cues. Given the general power of masked stimuli to influence behavior, an RCE for masked arrows is possible in principle, especially considering that we used predictive cues of 100% validity (randomly intermixed with neutral cues). There are, however, also reasons to question this possibility. If indeed no RCE emerges with masked retro-cues, one of course has to demonstrate that this absent effect of masked cues is not simply attributable to another reason, like the specific masks or presentation times. Therefore, participants also performed in a pre-cue task, for which we expected effects of masked cues (Gayet et al., 2014; Reuss et al., 2011). Results of our study therefore provide valuable information to further define which processes can or cannot be triggered by masked cues. 2. Method 2.1. Participants Twenty-four undergraduates from University of Würzburg participated for course credit (mean age 21.5 years; 22 female).2 Participants reported correct or corrected-to-normal vision and were naïve regarding the hypotheses of the experiment. Written informed consent was obtained prior to the experiment. 2.2. Apparatus and stimuli Stimulus presentation and response collection were done by a standard PC connected to a 17-in. CRT monitor (refresh rate: 100 Hz). Participants’ viewing distance was approximately 60 cm. Visual WM and change detection stimuli were circles (radius: 1.7 cm; 1.6° visual angle) of nine different colors (blue, brown, yellow, gray, green, purple, orange, red, and pink). Cues were white arrows (2.1 cm long; 2.0° visual angle) pointing up-left, up-right, down-left, or down-right as valid cues, and a white X (0.8 cm high; 0.8° visual angle) as the neutral cue. Filler stimuli on the change detection screen were white annuli. All visual stimuli were presented against a black background. Pre- and post-masks were pictures of random white dots that covered the area where the cues were presented. Such pattern masks are one of the standard ways to prevent conscious awareness of stimuli (e.g., Boy, Husain, & Sumner, 2010; Reuss et al., 2014). Responses were given on the left and right CTRL-key of a standard computer keyboard. 2.3. Tasks and procedure All participants were tested individually in a single session of about 90 min, and all participants performed in two tasks, a pre-cue and a retro-cue task (see Fig. 1). In the pre-cue task, a trial started with a fixation cross (200 ms) followed by three pre-masks (each 30 ms), and either the valid (arrow) or the neutral cue (X). With a normal, non-masked presentation, the cue was visible 120 ms followed by one post-mask (30 ms). With a masked presentation, the cue was visible 30 ms followed by four post-masks (each 30 ms). Following a blank screen (100 ms), the learn screen set on (300 ms) with four circles each filled with one of four randomlydrawn colors. No color repeated within this screen. After another blank screen (1000 ms), the change detection screen with one colored circle and three white annuli set on (visible until response). The task was to indicate whether the colored circle had the same color as the circle at this position in the learn screen (left CTRL-key) or not (right CTRL-key). The next trial started following an inter-trial-interval (ITI) of 1000 ms. In the retro-cue task, a trial started with a fixation cross (200 ms) followed by the learn screen (300 ms). After a blank screen (1000 ms), three pre-masks (each 30 ms) and the valid or neutral cue were presented for 30 or 120 ms. The cue was then followed by one or four post-masks (each 30 ms), depending on presentation condition. Following a blank screen (400 ms), the change detection screen set on (visible until response). The task was to indicate whether the colored circle had 1 It was discussed whether there is something special about arrow retro-cues. Berryhill et al. (2011) compared centrally presented arrows, centrally presented symbolic retro-cues (numbers), and peripheral retro-cues and only observed an RCE with the arrows. However, other authors reported RCEs with peripheral retro-cues as well (e.g., Matsukura, Cosman, Roper, Vatterott, & Vecera, 2014). 2 In previous studies, 20 (Reuss et al., 2011) and 19 (Gayet et al., 2014) participants were sufficient to detect an impact of masked arrows in pre-cueing. We used these numbers as an orientation and will come back to issues of statistical power in Section 4. 96 M. Janczyk, H. Reuss / Consciousness and Cognition 42 (2016) 93–100 Fig. 1. Examples for trial sequences in the pre-cue task and the retro-cue task. (Note: The figure is not drawn to scale and, for example, the masks did only cover the area where the cues were presented in the actual experiment.) the same color as the circle at this position in the learn screen (left CTRL-key) or not (right CTRL-key). The next trial started after an ITI of 1000 ms. All participants performed eight blocks, four with the pre-cue and four with the retro-cue task, in alternating order. The task administered in the first block was counterbalanced across participants. Each block comprised 64 trials resulting from two repetitions of 2 (cuetype: valid vs. neutral)
2 (cue-presentation: normal vs. masked)
2 (required response: yes vs. no)
4 (test positions) trial types. Two preceding blocks were unanalyzed practice blocks of 32 randomly drawn trials. After the main task, a cue identification task was applied. Participants were presented only masked and non-masked arrows and were to indicate their direction with a manual keypress. This is a standard procedure for assessing the efficiency of masks. Instructions were provided on-screen before the first block of each task variant. These instructions mentioned the 100% validity of the (informative) arrow cues. 2.4. Design and analyses ‘Task’ (pre-cue vs. retro-cue), ‘cuetype’ (valid vs. neutral), and ‘cue-presentation’ (normal vs. masked) were repeatedmeasures. Analyses focused on d0 (a measure representing the ability to correctly determine the match between learn and change detection display), which was calculated for each participant and design cell. The resulting d0 -values were then analyzed by means of analyses of variance (ANOVAs) and paired t-tests. A significance level of a = .05 was adopted. 3. Results Results are visualized in Fig. 2. We initially submitted the d0 values to a 2
2
2 ANOVA (see Table 1). Because the second-order interaction was significant, we followed-up this analysis by analyzing data from the two tasks separately with 2
2 ANOVAs. For the pre-cue task, better performance was observed with a valid compared with a neutral cue, F(1,23) = 126.83, p < .001, gp2 = .85, and with a normal compared with a masked presentation, F(1,23) = 39.65, p < .001, gp2 = .63. The pre-cue effect was smaller for the masked presentation resulting in a significant interaction, F(1,23) = 54.58, p < .001, gp2 = .70. Paired t-tests confirmed a pre-cue effect for both the normal, t(23) = 13.70, p < .001, d = 2.80, and the masked presentation, t(23) = 4.55, p < .001, d = 0.93. For the retro-cue task, better performance was observed with a valid compared with a neutral cue, F(1,23) = 18.56, p < .001, gp2 = .45, and with a normal compared with a masked presentation, F(1,23) = 7.45, p = .012, gp2 = .25. The RCE was smaller for the masked presentation resulting in a significant interaction, F(1,23) = 7.41, p = .012, gp2 = .24. Importantly, while an RCE was observed for the normal presentation, t(23) = 4.38, p < .001, d = 0.89, it was absent with the masked presentation, t(23) = 0.70, p = .491, d = 0.14. An ANOVA with task and cuetype as repeated-measures was run whilst including only trials with masked cuepresentation. It revealed a significant interaction of task and cuetype, F(1,23) = 9.15, p = .006, gp2 = .29, thus confirming a larger cueing-effect for the pre- compared with the retro-cue task for masked cues. An ANOVA with task and cue-presentation as repeated-measures with only neutral cue trials yielded no significant effects, all Fs 6 2.32, all ps P .142. 97 M. Janczyk, H. Reuss / Consciousness and Cognition 42 (2016) 93–100 Fig. 2. Mean d0 values as a function of task, cue-presentation, and cuetype. Error bars are 95% within-subject confidence intervals calculated separately for each comparison of neutral and valid cues (see Pfister & Janczyk, 2013). Table 1 Test statistics for the full 2
2
2 ANOVA. Effect F(1,23) p gp 2 Task Cuetype Cue presentation Task
cuetype Task
cue-presentation Cuetype
cue-presentation Cuetype
cue-presentation
task 26.56 135.80 39.27 35.47 3.96 40.82 5.51 <.001 <.001 <.001 <.001 .059 <.001 .028 .54 .86 .63 .61 .15 .64 .19 4. Discussion To extend our knowledge about which processes can be triggered by stimuli presented too briefly to allow their strategic use, we here tested whether masked retro-cues are able to orient attention within visual WM in a similar way as masked pre-cues do in the external environment. The latter has been shown in previous research (Gayet et al., 2014; Reuss et al., 2011) and similarly we observed enhanced performance for the cued location, when a masked pre-cue indicated the tobe-memorized display location. A standard RCE was also present, but in contrast, when the retro-cues were masked in the same way as the pre-cues were, the RCE was completely eliminated.3 4.1. External and internal attention shifting An impact of masked arrow pre-cues on external spatial attention is expected when the masked cues are contingent with the top–down settings of the participant (Reuss et al., 2011), for example, when the participant intends to utilize the cues due to their predictiveness. This precondition was met in our study, because half of the cues were predictive with 100% validity, and the other half was neutral. In other words: predictive cues were always valid and thus provided an overall incentive to be used. A theoretically important aspect of this part of our results is that the pre-cueing effect cannot be attributed to response priming. It was recently debated whether previous findings do not demonstrate an impact of masked cues on attention, but merely an instance of response priming (Al-Janabi & Finkbeiner, 2014). However, such a mechanism is only possible with a localization task (i.e., when the response corresponds with the target location). As this was not the case here, our results show that masked arrow pre-cues are indeed able to orient attention. 3 One might contest the validity of this comparison because the time from the learn screen to the change detection screen was not identical in the pre- and the retro-cue condition. However, in the retro-cue literature the time from learn screen to cue onset is sometimes considered the retention interval and retrocue conditions are compared to no-cue conditions where the change detection screen sets on after the same time where otherwise the cue sets on. Also note that the critical statistical comparison does not include the task variable but was done within the retro-cue task itself. Additionally, the d0 -values in the pre- and retro-cue task were equal for neutral cues. Thus the confound of the different delay between learn and change detection screens and task type should not be a principle problem. 98 M. Janczyk, H. Reuss / Consciousness and Cognition 42 (2016) 93–100 The absence of an effect with masked retro-cues presents a dissociation of internal and external attention shifting: retrocues apparently must be presented in a way that allows their conscious perception and not in a way preventing this, such as with subsequent pattern masks. This result corroborates earlier reports that the RCE becomes smaller when a dual-task (attracting attention away from the memory task) is applied close to retro-cue presentation (Janczyk & Berryhill, 2014), but not when applied a considerable time after retro-cue presentation (Hollingworth & Maxcey-Richard, 2013; Rerko et al., 2014). Caution is necessary, however, regarding further implications for explanations of the RCE and this experiment was not designed to distinguish them (see, e.g., Astle et al., 2012; Gözenman, Tanoue, Metoyer, & Berryhill, 2014; Matsukura et al., 2007; Pertzov, Bays, Joseph, & Husain, 2013; Souza, Rerko, & Oberauer, 2014), because none of these make assumptions whether the processes initiated as a result of the retro-cue are automatic or act in a controlled, strategic way. A difference between the pre- and retro-cue conditions (that applies to other studies as well; see Lepsien, Griffin, Devlin, & Nobre, 2005, for a similar point) is cognitive or memory load. While in the retro-cue task, a certain load was present at the time of cue presentation due to the previous memorization of the visual WM items, no such load was present at cue presentation in the pre-cue task. Two points are noteworthy in this context, however. First, it has been shown that the effectiveness of masked stimuli (that do not rely on semantic analysis) is not impaired under conditions of high cognitive load (Fischer, Kiesel, Kunde, & Schubert, 2011). Second, results concerning an interaction of (memory) load and the RCE for normally presented retro-cues are not consistent at present. Although some studies found an interaction of this kind (e.g., Lepsien et al., 2005; Souza et al., 2014) others did not (e.g., Gressmann & Janczyk, 2016; Makovski et al., 2008). Based on the present data, however, we cannot exclude with certainty that an impact of masked retro-cue re-emerges with less (memory) load, that is, less items to be encoded into visual WM at the outset of a trial. That the retro-cue must be represented in a way allowing its strategic use, also relates to a study by Shimi et al. (2014) who compared both central and peripheral pre- and retro-cues across different ages. In Experiment 1, cues were 100% valid and pre- and retro-cue benefits were reported overall. To test whether these benefits reflect automatic orienting or hinge on ‘‘controlled voluntary orienting” (p. 586), in Experiment 2 only retro-cues were employed and their validity was reduced to 50% to eliminate all incentives to use them in a voluntary way. Thus, half of the retro-cues were now invalid, and no impact was observed. The authors concluded that adults as well as children used cues in a voluntary way, but neither cue directed attention automatically. Concerning retro-cues, this fits well with our conclusion based on a different manipulation. In fact, we used 100% valid cues to provide the necessary incentive for using the cues. Further, our experiment provides a direct comparison with a pre-cueing condition. This is lacking in the Shimi et al. study, leaving open whether the particular experimental setup would have yielded an effect of pre-cues at least (as one can certainly expect, e.g., Hommel, Pratt, Colzato, & Godijn, 2001). In line with our conclusion, Shimi et al. (2014) interpret their results as a developmental dissociation between external and internal attention orienting: The former appeared to be efficient from young age on, while the latter continues to improve with ontogenetic development. Regarding cue visibility, it should be noted that the cue identification task showed that participants correctly identified masked arrows in about 68%. A comparison to the identification rate in previous studies with masked arrow cues indicates that the cues were only moderately masked. Thus, the absence of the RCE with masked cues cannot be attributed to their stronger masking in the present study. If anything, the cues were more easily identifiable than in previous studies that reported a cueing effect of masked (pre-)cues. Notably, however, even the moderate decrease in cue visibility according to our cue identification procedure was fully sufficient to eliminate the RCE. Note that this does not undermine our main conclusions or hamper interpretation of the results: Certainly, a more efficient masking of the cues would not yield other results for the retro-cue condition as those we observed. However, because the crucial finding hinges on retaining the null-hypothesis of one particular t-test, the question of sufficient power arises. It was difficult to anticipate a precise effect size for the RCE with masked cues. However, all reported ds in the other cases were >0.8, that is, at least large effects. Assuming d = 0.8 the achieved power was 1  b = .96 (a = .05, two-tailed, n = 24). We are thus confident that the study had sufficient power for detecting an RCE with masked cues if existing. 4.2. Limits of masked stimuli Limits (and possibilities) of masked stimuli were the topic of considerable debate, as they address basic psychological questions like the function of our consciousness. In the last few years, this debate was rekindled regarding the relation of conscious stimulus representations and cognitive control processes like task-set activation (Lau & Passingham, 2007; Reuss et al., 2011), response inhibition (van Gaal, Ridderinkhof, Fahrenfort, Scholte, & Lamme, 2008), conflict adaptation (Desender et al., 2014; Reuss et al., 2014), criterion setting (Reuss et al., 2015), or orienting of attention (Palmer & Mattler, 2013; Reuss et al., 2011). Strikingly, many of these processes are available even with unconscious stimulation (contrary to traditional concepts of consciousness and control). For example, participants were able to adapt to contexts of high conflict even though they were neither aware of the context information nor of the conflict information (Reuss et al., 2014). However, it is important to note that these studies also demonstrate that there are limitations that come with non-conscious stimulation, for example, stronger constraints on timing than with conscious stimulation. Also, some studies failed to find an effect of masked stimuli on cognitive control, but observed that some processes were only triggered by consciously perceived events (e.g., Ansorge, Fuchs, Khalid, & Kunde, 2011; Kunde, 2003). The present study speaks to this debate by showing that masked stimulation impacts on orienting of external attention, but not of internal attention within visual WM. Overall, however, it remains to be a very complex challenge to further define the limits of unconscious processing by identifying which M. Janczyk, H. Reuss / Consciousness and Cognition 42 (2016) 93–100 99 processes are susceptible to unconscious stimulation and which are not. The present study adds a piece to the complex puzzle, and is thus important for the enterprise of elucidating which factors lead to a (non-)susceptibility to unconscious stimuli. In the present experiment, we have purposefully chosen the first delay interval as approximately 1000 ms to tap into visual WM. Whether the lack of a masked cueing effect in this particular retro-cue condition is a general feature of retrocues or specific to visual WM remains unclear though. Future research should address this, for example, by including a shorter first delay (e.g., 200 ms) and thus tap into iconic memory. 4.3. Conclusion While masked arrows orient attention to locations in the external environment, they do not shift attention within visual WM in a similar way. This indicates a dissociation between external and internal attention shifting, strengthens the conceptual distinction between both kinds of attention, and indicates an example where masked stimulation seems not to be sufficient. Acknowledgments Work of MJ is supported by the Institutional Strategy of the University of Tübingen (Deutsche Forschungsgemeinschaft [German Research Foundation], ZUK 63). References Al-Janabi, S., & Finkbeiner, M. (2014). Responding to the direction of the eyes: In search of the masked gaze-cueing effect. Attention, Perception and Psychophysics, 76, 148–161. Ansorge, U., Fuchs, I., Khalid, S., & Kunde, W. (2011). No conflict control in the absence of awareness. Psychological Research Psychologische Forschung, 75, 351–365. Ansorge, U., & Heumann, M. (2006). Shifts of visuospatial attention to invisible (metacontrast-masked) singletons: Clues from reaction times and eventrelated potentials. Advances in Cognitive Psychology, 2, 61–76. Ansorge, U., & Neumann, O. (2005). Intentions determine the effects of invisible metacontrast-masked primes: Evidence for top–down contingencies in a peripheral cueing task. Journal of Experimental Psychology: Human Perception and Performance, 31, 762–777. Astle, D. E., Nobre, A. C., & Scerif, G. (2012). Attentional control constrains visual short-term memory: Insights from developmental and individual differences. The Quarterly Journal of Experimental Psychology, 65(2), 277–294. Astle, D. E., Summerfield, J., Griffin, I., & Nobre, A. C. (2012). Orienting attention to locations in mental representations. Attention, Perception & Psychophysics, 74, 146–162. Berryhill, M. E., Richmond, L. L., Shay, C. S., & Olsen, I. R. (2011). Shifting attention among working memory representations: Testing cue types, awareness, and strategic control. The Quarterly Journal of Experimental Psychology, 65, 426–438. Boy, F., Husain, M., & Sumner, P. (2010). Unconscious inhibition separates two forms of cognitive control. Proceedings of the National Academy of Sciences, 107 (24), 11134–11139. Chun, M. M., Golomb, J. D., & Turk-Browne, N. B. (2011). A taxonomy of external and internal attention. Annual Review of Psychology, 62, 73–101. Daza, M. T., Ortells, J. J., & Fox, E. (2002). Perception without awareness: Further evidence from a Stroop priming task. Perception & Psychophysics, 64, 1316–1324. Dehaene, S., & Naccache, L. (2001). Towards a cognitive neuroscience of consciousness: Basic evidence and a workspace framework. Cognition, 79, 1–37. Desender, K., Van Opstal, F., & Van den Bussche, E. (2014). Feeling the conflict. The crucial role of conflict experience in adaptation. Psychological Science, 25, 675–683. Eimer, M. (1997). Uninformative symbolic cues may bias visual-spatial attention: Behavioral and electrophysiological evidence. Biological Psychology, 46, 67–71. Fischer, R., Kiesel, A., Kunde, W., & Schubert, T. (2011). Selective impairment of masked priming in dual-task performance. Quarterly Journal of Experimental Psychology, 64, 572–595. Folk, C. L., & Remington, R. W. (1998). Selectivity in distraction by irrelevant featural singletons: Evidence for two forms of attentional capture. Journal of Experimental Psychology: Human Perception and Performance, 24(3), 847. Folk, C. L., & Remington, R. W. (1999). Can new objects overrride attentional control settings? Perception & Psychophysics, 61, 727–739. Folk, C. L., Remington, R. W., & Johnston, J. C. (1992). Involuntary covert orienting is contingent on attentional control settings. Journal of Experimental Psychology: Human Perception and Performance, 18, 1030–1044. Gabay, S., Avni, D., & Henik, A. (2012). Reflexive orienting by central arrows: Evidence from the inattentional blindness task. Psychonomic Bulletin & Review, 19, 625–630. Gayet, S., Van der Stigchel, S., & Paffen, C. L. E. (2014). Seeing is believing: Utilization of subliminal symbols requires a visible relevant context. Attention, Perception, & Psychophysics, 76, 489–507. Gözenman, F., Tanoue, R. T., Metoyer, T., & Berryhill, M. E. (2014). Invalid retro-cues can eliminate the retro-cue benefit: Evidence for a hybridized account. Journal of Experimental Psychology: Human Performance and Perception, 40, 1748–1754. Gressmann, M., & Janczyk, M. (2016). The (un)clear effects of invalid retro-cues. Frontiers in Psychology, 7, 244. Griffin, I. C., & Nobre, A. C. (2003). Orienting attention to locations in internal representations. Journal of Cognitive Neuroscience, 15, 1176–1194. Hollingworth, A., & Maxcey-Richard, A. M. (2013). Selective maintenance in visual working memory does not require sustained visual attention. Journal of Experimental Psychology: Human Perception and Performance, 39, 1047–1058. Hommel, B., Pratt, J., Colzato, L., & Godijn, R. (2001). Symbolic control of visual attention. Psychological Science, 12, 360–365. Jack, A. I., & Shallice, T. (2001). Introspective physicalism as an approach to the science of consciousness. Cognition, 79, 161–196. Janczyk, M., & Berryhill, M. E. (2014). Orienting attention in visual working memory requires central capacity: Decreased retro-cue effects under dual-task conditions. Attention, Perception, & Psychophysics, 76, 715–724. Jonides, J. (1981). Voluntary versus automatic control over the mind’s eye’s movement. In J. B. Long & A. D. Baddeley (Eds.), Attention & performance IX (pp. 187–203). Hillsdale, NJ: Erlbaum. Kouider, S., & Dehaene, S. (2007). Levels of processing during non-conscious perception: A critical review. Philosophical Transactions of the Royal Society of London B, 362, 857–875. Kunde, W. (2003). Sequential modulations of stimulus-response correspondence effects depend on awareness of response conflict. Psychonomic Bulletin & Review, 10, 198–205. Kunde, W., Reuss, H., & Kiesel, A. (2012). Consciousness and cognitive control. Advances in Cognitive Psychology, 8, 9–18. 100 M. Janczyk, H. Reuss / Consciousness and Cognition 42 (2016) 93–100 Landman, R., Spekreijse, H., & Lamme, V. A. (2003). Large capacity storage of integrated objects before change blindness. Vision Research, 43, 149–164. Lau, H. C., & Passingham, R. E. (2007). Unconscious activation of the cognitive control system in the human prefrontal cortex. Journal of Neuroscience, 27, 5805–5811. Lepsien, J., Griffin, I. C., Devlin, J. T., & Nobre, A. C. (2005). Directing spatial attention in mental representations: Interactions between attentional orienting and working-memory load. NeuroImage, 26, 733–743. Li, Q., & Saiki, J. (2014). The effects of sequential attention shifts within visual working memory. Frontiers in Psychology, 5, 965. Mack, A., & Rock, I. (1998). Inattentional blindness.Cambridge, MA: MIT Press. Makovski, T., & Jiang, Y. V. (2007). Distributing versus focusing attention in visual short-term memory. Psychonomic Bulletin & Review, 14, 1072–1078. Makovski, T., Sussman, R., & Jiang, Y. V. (2008). Orienting attention in visual working memory reduces interference from memory probes. Journal of Experimental Psychology. Learning, Memory, and Cognition, 34, 369–380. Matsukura, M., Cosman, J. D., Roper, Z. J. J., Vatterott, D. B., & Vecera, S. P. (2014). Location specific effects of attention during visual short-term memory maintenance. Journal of Experimental Psychology: Human Perception and Performance, 40, 1103–1116. Matsukura, M., Luck, S. J., & Vecera, S. P. (2007). Attention effects during visual short-term memory maintenance: Protection or prioritization? Perception & Psychophysics, 69, 1422–1434. Mattler, U. (2006). On the locus of priming and inverse priming effects. Perception & Psychophysics, 68, 975–991. Merikle, P. M., & Joordens, S. (1997). Parallels between perception without attention and perception without awareness. Consciousness and Cognition, 6, 219–236. Merikle, P. M., Joordens, S., & Stolz, J. A. (1995). Measuring the relative magnitude of unconscious influences. Consciousness and Cognition, 4, 422–439. Mulckhuyse, M., Talsma, D., & Theeuwes, J. (2007). Grabbing attention without knowing: Automatic capture of attention by subliminal spatial cues. Visual Cognition, 15, 779–788. Müller, H. J., & Rabbitt, P. M. A. (1989). Reflexive and voluntary orienting of visual attention: Time course of activation and resistance to interruption. Journal of Experimental Psychology: Human Perception and Performance, 15, 315–330. Nobre, A. C., Coull, J. T., Maquet, P., Frith, C. D., Vandenberghe, R., & Mesulam, M. M. (2004). Orienting attention to locations in perceptual versus mental representations. Journal of Cognitive Neuroscience, 16, 363–373. Palmer, S., & Mattler, U. (2013). Masked stimuli modulate endogenous shifts of spatial attention. Consciousness and Cognition, 22, 486–503. Pertzov, Y., Bays, P. A., Joseph, S., & Husain, M. (2013). Rapid forgetting prevented by retrospective attention cues. Journal of Experimental Psychology: Human Perception and Performance, 39, 1224–1231. Pfister, R., & Janczyk, M. (2013). Confidence intervals for two sample means: Calculation, interpretation, and a few simple rules. Advances in Cognitive Psychology, 9, 74–80. Posner, M. I. (1980). Orienting of attention. Quarterly Journal of Experimental Psychology, 32, 3–25. Posner, M. I., & Cohen, Y. (1984). Components of visual orienting. In H. Bouma & D. G. Bowhui (Eds.), Attention and performance X (pp. 531–556). Hillsdale, NJ: Erlbaum. Pratt, J., Radulescu, P., Guo, R. M., & Hommel, B. (2010). Visuospatial attention is guided by both the symbolic value and the spatial proximity of selected arrows. Journal of Experimental Psychology: Human Perception and Performance, 36, 1321–1324. Rerko, L., Souza, A. S., & Oberauer, K. (2014). Retro-cue benefits in working memory without sustained focal attention. Memory & Cognition, 42, 712–728. Reuss, H., Desender, K., Kiesel, A., & Kunde, W. (2014). Unconscious conflicts in unconscious contexts: The role of awareness and timing in flexible conflict adaptation. Journal of Experimental Psychology: General, 143, 1701–1718. Reuss, H., Kiesel, A., & Kunde, W. (2015). Adjustments of response speed and accuracy to unconscious cues. Cognition, 134, 57–62. Reuss, H., Kiesel, A., Kunde, W., & Hommel, B. (2011). Unconscious activation of task sets. Consciousness and Cognition, 20, 556–567. Reuss, H., Kiesel, A., Kunde, W., & Wühr, P. (2012). A cue from the unconscious – masked symbols prompt spatial anticipation. Frontiers in Psychology, 3, 397. Reuss, H., Pohl, C., Kiesel, A., & Kunde, W. (2011). Follow the sign! Top–down contingent attentional capture of masked arrow cues. Advances in Cognitive Psychology, 7, 82–91. Shimi, A., Nobre, A. C., Astle, D., & Scerif, G. (2014). Orienting attention within visual short-term memory: Development and mechanisms. Child Development, 85, 578–592. Souza, A. S., Rerko, L., & Oberauer, K. (2014). Unloading and reloading working memory: Attending to one item frees capacity. Journal of Experimental Psychology: Human Perception and Performance, 40, 1237–1256. Sperling, G. (1960). The information available in brief visual presentations. Psychological Monographs, 74, 1–29. Tanoue, R. T., & Berryhill, M. E. (2012). The mental wormhole: Internal attention shifts without regard for distance. Attention, Perception & Psychophysics, 74, 1199–1215. Theeuwes, J. (1991). Exogenous and endogenous control of attention: The effect of visual onsets and offsets. Perception & Psychophysics, 49, 83–90. van Gaal, S., Ridderinkhof, K. R., Fahrenfort, J. J., Scholte, H. S., & Lamme, V. A. F. (2008). Frontal cortex mediates unconsciously triggered inhibitory control. Journal of Neuroscience, 28, 8053–8062. Yantis, S., & Jonides, J. (1990). Abrupt visual onsets and selective attention: Voluntary versus automatic allocation. Journal of Experimental Psychology: Human Perception and Performance, 16, 121–134.

Consciousness and Cognition 19 (2010) 711–720 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Grasping language – A short story on embodiment q Doreen Jirak a,e, Mareike M. Menz a, Giovanni Buccino b, Anna M. Borghi b,c,1, Ferdinand Binkofski a,d,*,1 a Department of Systems Neuroscience and Neuroimage Nord, University Medical Center Hamburg Eppendorf, Hamburg, Germany Magna Graecia, University of Catanzaro, Catanzaro, Italy c Institute of Cognitive Science and Technology, CNR, Rome, Italy d Section for Neurological Cognitive Research, RWTH Aachen University, Aachen, Germany e MIN-Faculty, Department of Informatics, University of Hamburg, Germany b a r t i c l e i n f o Article history Available online 23 August 2010 Keywords: Embodiment Language Perception Action fMRI Meta-analysis ALE a b s t r a c t The new concept of embodied cognition theories has been enthusiastically studied by the cognitive sciences, by as well as such disparate disciplines as philosophy, anthropology, neuroscience, and robotics. Embodiment theory provides the framework for ongoing discussions on the linkage between ‘‘low” cognitive processes as perception and ‘‘high” cognition as language processing and comprehension, respectively. This review gives an overview along the lines of argumentation in the ongoing debate on the embodiment of language and employs an ALE meta-analysis to illustrate and weigh previous findings. The collected evidence on the somatotopic activation of motor areas, abstract and concrete word processing, as well as from reported patient and timing studies emphasizes the important role of sensorimotor areas in language processing and supports the hypothesis that the motor system is activated during language comprehension. Ó 2010 Elsevier Inc. All rights reserved. 1. The idea of embodiment ‘‘Grasp the subject, the words will follow” was advised by Cato the Elder. Grasping an explanation, giving an example, posing a threat – language is full of actions and objects, and the ties between language and motion are under continuous investigation. Embodied cognition theories are becoming more and more popular in cognitive (neuro)science, as well as in philosophy, anthropology, cognitive psychology, and robotics (e.g. Nolfi & Floreano, 2000; Ziemke, 2002). According to the embodied view, there is no separation between the so-called ‘‘low” cognitive processes, such as perception and action, and ‘‘high” cognitive processes, such as language and thought. Generally, embodiment links the individual sensorimotor experiences with higher cognitive functions such as language processing and comprehension. Connecting motor abilities with cognitive capacities contradicts with the classic amodal view, which assumes a clear-cut separation between low and high level processes and which states that cognition derives from computational processes in separate domains. When applied to language, embodied cognition views claim that when we understand words, the same sensorimotor areas are recruited as for interacting with the objects and entities the words refer to. Similarly, when we comprehend sentences, we internally simulate the state of the world the sentences describe (Zwaan, 2004). In the past years much behavioral q This article is part of a speical issue of this journal on Self, Other and Memory. * Corresponding author. Address: Section for Neurological Cognitive Research, RWTH Aachen University, Pauwelstrasse 30, D-52074 Aachen, Germany. Fax: +49 241 80 82139. E-mail address: fbinkofski@ukaachen.de (F. Binkofski). 1 These authors contributed equally. 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.06.020 712 D. Jirak et al. / Consciousness and Cognition 19 (2010) 711–720 and neural evidence has been collected, showing that the process of language comprehension activates a motor simulation (Gallese, 2008) and involves the motor system (see Barsalou, 2008; Fischer & Zwaan, 2008; Pulvermüller, 2005). In a neuroscientific context, this perspective also implies that brain areas related to action and language can no be longer be seen as independent, but rather working in concert. Areas traditionally regarded as pure motor areas as e.g. the primary motor or the premotor cortex, as well as areas that have traditionally been assigned to the processing of language, e.g. Broca’s or Wernicke’s region, are not modularized, but rather provide the linkage of action and language (Pulvermüller, 2005). Participation of Broca’s region has already been revealed in different motor experiments, such as grasping experiments (Grafton, Arbib, Fadiga, & Rizzolatti, 1996), object manipulation (Binkofski et al., 1999) and action imitation (Rizzolatti & Arbib, 1998). Based on numerous results, it turns out, that Broca’s region is more than just representing a language processing area (Menz & Binkofski, 2008). In this short story on embodied language we want to review recent literature on language and embodiment – first along the main lines of argumentation in the ongoing debate, and second in an activation likelihood estimation (ALE) meta-analysis on reported neural activity patterns. 1.1. Somatotopic activation during processing of action words and sentences One of the most important issues in the ongoing debate about embodied cognition is the somatotopy of activation. If the hypothesis held true that language recruited the same sensorimotor areas as for action and interaction, motor areas should display the same somatotopy for processing language as for processing actions. Although the general involvement of premotor and motor cortices has been demonstrated repeatedly, the issue of somatotopy still needs to be clarified. There is strong evidence for a somatotopic activation of premotor cortices from studies with different techniques (fMRI, MEG, etc.). Tettamanti et al. (2005) investigated the underlying neural processes while presenting sentences expressing actions performed with the mouth, the hand or the foot. Specifically, hand action and related words were found activated in the left precentral gyrus, the posterior intraparietal sulcus and the left posterior inferior temporal area. In contrast, leg activity has been identified in the left dorsal premotor and left intraparietal sulcus, but located more dorsally and rostrally in relation to the parietal hand activities. In addition, detection of a bilateral pattern in the posterior cingulate shows clear distinction of activation in processing abstractness. Activity in Broca’s region has been detected detached from any effector-specificity, thus implying a special role in language processing. Summarized, the fMRI results presented in that study display activity in a frontal-parietal circuit with temporal participation in the left hemisphere. Furthermore, Pulvermüller (2005) recorded neurophysiological and behavioral responses to verbs referring to actions performed with the face, the arms and the legs. Using a lexical decision task, they found topographical differences in the brain activity patterns generated by the different verbs, starting 250 ms after word onset. Consequently, the English verbs ‘lick’, ‘pick’, and ‘kick’ engage different neural sites in a topographical pattern. Moreover, a near-simultaneous activity pattern in the inferior frontal gyrus and the superior temporal gyrus could be identified, which supports both speech production as well as word comprehension. Another study on somatotopic organization of the motor cortex is presented in Hauk, Johnsrude, and Pulvermüller (2004). The differentiation of arm-related and leg-related action, respectively action words revealed distinct patterns in the middle frontal and the precentral gyrus for arm actions and, on the contrary, activations in dorsal areas in left and midline of the pre-and postcentral gyrus and dorsal premotor cortex. Induced by a dissociation on word category processing, the cortical activity found in this study displays effector-dependent processing along the motor strip. In addition, EEG-recordings showed an activation of the effector-specific motor regions occurring quite early, less than 200 ms after word onset (Pulvermüller, Lutzenberger, & Preissl, 1999). A combined behavioral and TMS study by Buccino et al. (2005) strengthens this suggestion. A decrease in amplitude of MEPs was recorded from hand muscles when listening to hand-action related sentences, and from foot muscles when listening to foot related sentences. In line with this evidence, further results obtained with behavioral tasks (Borghi & Scorolli, 2009; Scorolli & Borghi, 2007) suggest that the simulation activated during combinations of nouns and verbs is sensitive to the congruency between the effector implied by the sentence (e.g. mouth vs. foot) and the effector used to produce the motor response. Also other studies are able to identify body-part specificity in premotor cortices, but not in other motor areas. In an fMRI study with a lexical decision task Willems, Hagoort, and Casasanto (2010) found a preferential activation of the left premotor cortex for right handers, and of the right premotor cortex for left handers, while responding to manual-action verbs (compared to nonmanual action verbs). This suggests that the simulation evoked during language processing is body-specific. However, whereas imagery activated both motor and premotor cortices in a differential way, language comprehension activated only the premotor cortex. In a similar vein, Tomasino, Werner, Weiss, and Fink (2007, 2008) found activation of the primary motor cortex during explicit mental motor imagery, whereas no activation was found in a letterdetection task. Although several studies showed evidence towards action word comprehension in connection with somatotopy, a more critical view concerning this topic is discussed in Postle, McMahon, Ashton, Meredith, and de Zubicaray (2008). D. Jirak et al. / Consciousness and Cognition 19 (2010) 711–720 713 Focussing on different motor areas Postle et al. (2008) used fMRI in cytoarchitectonically regions of interest (primary and premotor cortices) to compare action verbs related to different effectors (hand, foot and mouth) with other concrete nouns unrelated to body parts and actions, as well as to non-words and to strings of hashes. Action and non-action words were matched for imageability. Whereas an expected somatotopic organization for observation of simple movements could be identified, i.e. activations of motor areas BA 4 and BA 6 in a ventral-to-dorsal fashion according to the succession mouth, hand and foot. The same applies to a posterior-to-anterior pattern across the lateral surfaces of the mentioned motor areas. Although there was no evidence of a somatotopic organization for action words, the pre-supplementary motor area (preSMA) displayed a different activation for foot-related action compared to non-action words, thereby possibly playing a rather cognitive-motor role instead of a pure motor one. A main difference of Postle et al. in comparison to the other studies cited above is the use of cytoarchitectonically defined probability maps. This suggests that studies on somatotopy connected to word meaning extraction should be also related to cytoarchitectural information and functional criteria, in order to correctly interpret activation distribution as somatotopy. Summarized, several language comprehension studies show evidence on an at least effector-specific activation during language comprehension, though not primarily investigated in all studies introduced above. In general, but also considering some critical aspects by Postle et al. (2008), it can be stated, that somatotopy can be identified in premotor regions, but not consistently in primary motor cortices. This might be due to the complexity of tasks, but also to the different roles the motor cortices play in conceptualizing and execution of actions. 1.2. Embodiment and the mirror neuron system Along with a somatotopy in processing language, also the involvement of the mirror neuron system in processing language is based on the theoretical principle, that the processing of language is grounded in the same neural units as the actions the words refer to. According to embodied theories, canonical and mirror neurons represent the neural basis of the simulation activated during language comprehension (Gallese, 2008). Mirror neurons, which were originally identified in the ventral premotor cortex of monkeys, are not only firing during active motion, but they are also triggered by observing a conspecific performing actions with objects (Rizzolatti & Craighero, 2004). Similar investigations in humans lack the single-neuron resolution, thus it is rather referred to as the mirror neuron system (MNS; Buccino et al., 2001, 2005). The linking element between monkey and man is the area F5 of the monkey ventral premotor cortex which is regarded as a homolog to Broca’s region, the inferior frontal region in the human cortex, which is primarily known as a speech processing area (Rizzolatti & Craighero, 2004). This leads to the assumption, that homolog to F5, also Broca’s region contains mirror neurons (Buccino et al., 2005) and Broca’s region is no longer regarded as a pure language area, but as also as a region linking action and language (Binkofski & Buccino, 2004; Menz & Binkofski, 2008). As Broca’s region is the core region of the MNS and implies an additional link between motor processing and speech, the whole MNS possibly also has an impact on language processing and comprehension. Glenberg et al. (2008) show mirror neuron activation in an experimental setting by either presenting a typical action sound or a verbal action description. In differential contrasts of typically MNS-activating tasks, as e.g. object observation compared to language processing, the activation of Broca’s region to either seems to differ. Aziz-Zadeh, Koski, Zaidel, Mazziotta, and Iacoboni (2006), Aziz-Zadeh, Wilson, Rizzolatti, and Iacoboni (2006) and Aziz-Zadeh and Damasio (2008) describe close but not completely overlapping patterns for action observation and reading phrases. They draw the conclusion that mirror neurons are not directly mediating the understanding of language, but possibly play an important role as a precursor in the development of language (Rizzolatti & Arbib, 1998). However, this does not contradict an integrating role of Broca’s region in processing both sounds and actions. In monkey, Kohler et al. (2002) detected firing of mirror neurons in the presence of action specific noise. Mirror neurons showed a specialization for an action and the sound it produces, e.g. they fire when breaking a peanut and also when only the sound is played. In the human domain, D’Ausilio et al. (2009) reported a facilitation of the perception of a given speech sound when the motor articulator responsible for that sound was stimulated with TMS for motor cortex controlling. Apart from Broca’s region, the left inferior parietal lobe (IPL), also part of the MNS (Buccino et al., 2005) was reported to have an important role in the integration of sounds and actions. McNamara et al. (2008) asked subjects to learn associations between previously unrelated novel sounds and meaningless gestures. Both IPL and Broca’s region showed a strong, bilateral, negative correlation of BOLD response with learning of sound–action associations during data acquisition. Together with decrease due to the sharpening of the network, connectivity between the areas increased and the strongest learning related connectivity between regions was found in Broca’s region and left IPL. This leads to the conclusion that the involvement of motor regions in language processing is closely linked to regions of the mirror neuron system, thus possibly relying on using mirror neurons to integrate sounds and actions or even to simulate in order to understand action words. However, this claim is strong and will need further evidence. 1.3. Abstract and concrete word processing A second strong claim of embodied theories relates to the grounding of abstract language. Embodied theories assume that abstract concepts, just like concrete ones, are grounded in the sensorimotor system. Within this general framework, at least three different explanations of abstraction have been proposed (see Glenberg et al., 2008). A key issue in the literature is to what extent concrete and abstract words (e.g., ‘‘bottle” vs. ‘‘truth”) are represented differently. However, starting from this 714 D. Jirak et al. / Consciousness and Cognition 19 (2010) 711–720 general assumption different explanations have been proposed, the most influential one is based on metaphors. Lakoff and Johnson (1980) proposed up to 50 metaphor schemes in everyday language. ‘‘Cancer finally caught him up” serves as example for personification, whereas ‘‘Your claims are indefensible” has yet another categorization. In line with the work, Casasanto and Boroditsky (2008) show connection of metaphorical language processing and experiences built on perception and action. The argumentation of metaphors contributing to access meaning and also appealing to the conceptual system, which is also involved by action experiences and thinking, speaks in favor of the embodiment perspective of language comprehension. According to very recent proposals, abstract words involve more emotional aspects than concrete ones (Vigliocco, 2009); in addition, abstract words rely more on linguistic and social information as their modality of acquisition employs more linguistic information as compared to concrete words (Borghi & Cimatti, 2009). According to the strong version of the embodied framework, both the literal and the more abstract meaning of language (e.g., ‘‘grasp” in the context of ‘‘grasping an apple” and in the context of ‘‘grasping a notion”) are processed in the same neural units. Thus, action words should be represented in the same sensory-motor areas as their simple motor analog. As an example, Aziz-Zadeh and Damasio (2008) proposed that the verb ‘to kick’ (literal) and ‘kick of the year’ (abstract) imply the same ‘kick’ motor representation. There is TMS evidence supporting this view. Glenberg and collaborators (2008) have shown that abstract transfer sentences (e.g., give some news) activate motor information exactly as concrete transfer sentences (e.g., give a pizza). A more specific description of the processing of abstract and concrete words is given by a study of Rüschemeyer, Brass, and Friederici (2007). They performed an fMRI study comparing verbs with motor meaning, such as ‘‘to beat”, and verbs with abstract meanings, such as ‘‘to guess”. Participants had to respond by pressing a key to pseudo-words, while no response for words was required. The authors found enhanced activation of motor verbs compared to abstract verbs in the posterior premotor, primary motor, primary and secondary somatosensory cortex; this activation was bilateral but was higher in the left hemisphere. However, they did not find any difference in the activation of frontal mirror neurons areas, the ventral premotor cortex, and in the inferior parietal lobule for processing simple motor verbs and abstract verbs. In addition, no difference was found while comparing German verbs with motor stems and verbs with abstract stems. Tettamanti et al. (2005), however, describe a premotor activation during the processing of action-related sentences as compared to their control condition (e.g. ‘‘now I appreciate loyalty”), thus reporting a unique activation of a motor area to concrete sentences containing a manipulable object as opposed to sentences containing abstract objects. Another recent fMRI study, in turn suggests that the abstract words and sentences activate motor representations (Raposo, Moss, Stamatakis, & Tyler, 2009). The authors compared single verbs and literal and idiomatic sentences in order to verify whether involvement of motor regions is automatic and invariable or whether it is modulated by the sentential context. In the task they used, particularly adequate in case of ambiguous sentences, participants listened to sentences; on half of them they were presented with a visual probe and had to determine by pressing a key whether the visual probe was related or not to the sentence meaning. Listening to leg-related and arm-related action verbs (e.g., grab, kick) activated a fronto-parietal system typically involved in action execution. This leads to two assumptions. First, verbs and nouns are possibly processed differently regarding their abstractness, and second, that different levels of derivation from a word’s literal meaning might lead to different activations. Hence, abstract (‘‘to kick around an idea”), metaphorical (‘‘to kick in the dugout”), idiomatic (‘‘to kick the bucket”), and morphological (rare in English but in German ‘‘treten/to kick” and ‘‘eintreten”/to occur”) should be investigated separately and not be subsumed under the term ‘‘abstract” or ‘‘non-motor” or even be regarded as homogeneous control conditions. Altogether we can only conclude that demonstrations of the sensorimotor grounding of abstract words have so far been confined to rather specific domains and further evidence for grounding of abstract language contents is needed (for further discussion of this issue, see Borghi & Cimatti, 2009). 1.4. Are sensorimotor areas essential for language comprehension? One core issue in discussing embodied language lies in the question whether the involvement of sensorimotor areas is auxiliary, concomitant, or necessary for language processing and comprehension. Even though the majority of studies demonstrate that the motor system is activated during words and sentences processing, there is some controversial evidence, and some issues remain open (see also Willems & Hagoort, 2007). Partly this might be related to the afore mentioned disagreement as to the definition of abstractness and the usage of control conditions. However there are two points of view contributing to an essential role of sensorimotor areas to language processing: evidence from patients and evidence on timing. The first promising evidence is given by studies on patients with impairments of the motor system. Boulenger et al. (2008) found no priming effect for action verbs for patients affected by Parkinson disease off dopaminergic treatment, i.e. when there is no normal activation level in premotor and motor areas. The priming effect, instead, was present in both controls and Parkinson patients after dopaminergic intake. This study provides strong evidence that the integrity of the motor system is necessary for verb processing. Along with this evidence, Bak et al. (2006) found selective deficits in verb processing in two patients, father and son, with a familial motor disorder; in addition, Bak and Hodges (2004) found a selective difficulty for verb processing in motor neuron disease, a neurodegenerative disease of the motor system. Even if this evidence is not conclusive, a number of studies report action comprehension deficits in patients with premotor and parietal lesions (for a review and discussion, see Aziz-Zadeh & Damasio, 2008). Altogether, it can be claimed that lesions of the motor system selectively impair language processing, and particularly verb comprehension. This evidence can support the argument that an integer motor system might be part of the comprehension process. D. Jirak et al. / Consciousness and Cognition 19 (2010) 711–720 715 The second good argument in favor of the involvement of the motor system as an essential part of language comprehension would be an early activation of the motor system. Papeo, Vallesi, Isaja, and Rumiati (2009) recorded TMS-induced motor-evoked potentials from right hand muscles in order to measure M1 activity during comprehension of action verbs. They found an increase of M1-activity only at 500 ms, while no increase was present as they delivered single pulse TMS at 170 and 350 ms after action words appearance. This suggests that M1 is involved during post-conceptual processing of action words, and it is not necessarily implied and does not contribute to words comprehension. However, other studies report an early involvement of motor areas (for a review and a model reproducing results on both early and late activation of motor areas, see Chersi, Thill, Ziemke, & Borghi, 2010). In a lexical decision task, presenting action verbs and nouns, Pulvermüller et al. (1999) detected neural activity recorded from EEG less than 200 ms after word onset and in an automatic way. Even in this short time window, the signals highlighted electrocortical differences. As discussed by the authors, the study revealed nonclassical language areas involved in language processing, and proposed additional cell assemblies in the motor cortex for action verbs and neural signals from the visual cortex for nouns, respectively. Thus, we think that evidence collected so far is promising and that in the near future the issue of the necessity of activation of the motor system concerning semantic processing will be definitively solved. 2. Embodiment – a meta-analysis Opposed to narrative reviews or label-based anatomical approaches, the coordinate-based meta-analysis methods statistically aggregate activation foci (‘peaks’) derived from neuroimaging data and emphasize specific neuronal patterns across multiple studies following a common paradigm or hypothesis. Recent experimental series can be evaluated to a meta-statement. For the meta-analysis included in this review we used the ALE approach as implemented in the Ginger ALE software provided by BrainMap (Laird et al., 2005). 2.1. Literature search and criteria An exhaustive literature search was conducted on fMRI studies indexed in the Medline database. We focused on studies, which contained the pre-specified keywords ‘‘embodiment”, ‘‘language”, ‘‘motor”, ‘‘action”, and ‘‘perception”. The filter criteria do not distinguish between concrete and abstract word processing, as the objective of the analysis is to cover the whole range of embodied language. In addition, the selection of experiments took into account both extremities (hand/arm, foot/ leg) to increase the variance of particular action-related patterns in literal and abstract meanings. Furthermore, we included studies on words (verbs, nouns) and on sentences or both. 2.2. Statistical procedure A statistical map was generated by using a collection of 468 foci from the 21 studies reported in Table 1 after transferring them into Talairach space (Talairach & Tournoux, 1988). In order to account for the uncertainty, that is technically inherent to the actual location of the peaks, each coordinate was modeled not as a single point, but by a three-dimensional (3D) Gaussian function with 12 mm FWHM. Thus, the localization probability distributions describe the probability that a given focus actually lay within a particular voxel. Statistical significance is gained via a permutation test of randomly generated foci using the same FWHM and number of foci. The voxel-wise comparison is tested against the null-hypothesis of uniformly distributed peaks, giving a set of ALE-values necessary for thresholding the probability map. Using the False Discovery Rate (FDR) with q = .01, the test was corrected for multiple comparisons. 2.3. Results of the meta-analysis The activation clusters of the meta-analysis are summarized in Table 2 and Fig. 1. All coordinates are in Talairach space and anatomical labels as well as Brodmann areas were obtained with the Talairach Daemon (Lancaster et al., 1997, 2000). Major activity sites are displayed in the left hemisphere, predominantly in the frontal lobe, comprising the inferior frontal gyrus (cluster 1, BA 44, BA 46) and the precentral gyrus (cluster 1, cluster 3, cluster 9, BA 4, BA 6). In the parietal lobe, distinct clusters could be detected in the left supramarginal gyrus (cluster 1, BA 40), as well as in the right superior parietal lobulus (cluster 5, BA 7) and in the left precuneus (cluster 1, BA 19) area. In the left temporal lobe, activations could be found in the middle temporal gyrus (cluster 2, BA 22, BA 39) and more prominently in the fusiform area (cluster 8, BA 37). Further findings include the insula (cluster 2, BA 13) in the left hemisphere as well as the bilateral cerebellum. Additionally, the analysis reveals two clusters comprising the posterior cingulate (cluster 6, cluster 11, BA 30) in both hemispheres. 2.4. Integrating the meta-analysis into previous findings The main advantage of the ALE meta-analysis is to give an overview on previously reported findings, to un-weight interpretations, and to re-weight results within, but even more outside of regions of interest. Due to the method, there 716 D. Jirak et al. / Consciousness and Cognition 19 (2010) 711–720 Table 1 Publications included in the meta-analysis, task they employed, number of subjects that were investigated and number of selected foci for the ALE metaanalysis. Study Task N Foci Bedny et al. (2008) Boulenger et al. (2008) Canessa et al. (2008) Action verb comprehension Arm- and leg-related action verbs Semantic decisions on picture pairs of (non-)manipulable objects (e.g. tools) Nouns (objects), verbs (action) Classification of real (e.g. tools) and nonsense objects (e.g. artificial shapes) Arm- and leg-relates action words Hand- and leg-related action verbs Action-/object-naming, verbs or nouns Responses on hand manipulation and phonological word decisions Hand- and foot-related action verbs Arm- and leg-related action verbs Lexical decision between functional and volumetric manipulability (by hand) Go/NoGo tasks to words stimuli Motor- and abstract verbs/wordstems Action- and object naming (verbs, nouns),hand- and leg-related Action- and object naming (verbs, nouns) Word reading and generation from semantic categories Hand- and foot related sentences Action-related and abstract sentences, supplemental data to Tettamanti (2005) Motor and non-motor verbs, hand related Word decision on concrete hand actions 12 18 15 6 43 35 17 20 5 48 14 16 15 14 18 22 15 15 45 62 25 6 5 5 19 20 13 12 12 17 18 7 19 25 6 18 14 10 15 20 29 40 Gennari, Mac Donald, Postle, and Seidenberg (2007) Gronau et al. (2008) Hauk et al. (2004) Kemmerer et al. (2008) Liljeström et al. (2008) Meister and Iacoboni (2007) Postle et al. (2008) Raposo et al. (2008) Rüschemeyer, von Rooij, Lindemann, Willems, and Bekkering (2009) Rüschemeyer, Pfeiffer, and Bekkering (2010) Rüschemeyer et al. (2007) Saccuman et al. (2006) Siri et al. (2007) Tremblay and Gracco (2006) Tettamanti et al. (2005) Tettamanti et al. (2008) Tomasino et al. (2007) Willems, Ozyürek, and Hagoort (2009) Table 2 Results from the ALE meta-analysis. Clusters of activation connected above threshold, activation sites, Talairach-Coordinates (x, y, z) of maximum ALE-value, and maximum ALE-value of this cluster. Cluster Area 1 L. Inferior frontal gyrus (BA 44) L. Inferior frontal gyrus (BA 45) L. Precentral gyrus (BA 6) L. Supramarginal gyrus (BA 40) L. Postcentral (BA 3) L. Precuneus (BA 19) L. Precentral gyrus (BA 4) L. Insula (BA13) L. Middle frontal gyrus (BA 6) L. Middle temporal gyrus (BA 22) L. Middle temporal gyrus (BA 39) L. Insula (BA 13) L Medial frontal gyrus (BA 6) L. Cerebellum anterior lobe R. Superior parietal lobe (BA 7) L. Posterior cingulate (BA 30) L. Middle temporal gyrus (BA 30) L. Fusiform gyrus (BA 37) R. Precentral gyrus (BA 4) R. Fusiform gyrus (BA 37) R. Posterior cingulate (BA 30) 2 3 4 5 6 7 8 9 10 11 x y 46 46 40 44 46 28 52 32 26 52 50 54 4 36 26 20 52 46 40 38 16 z 12 26 6 38 20 66 10 22 6 40 58 32 0 42 64 62 4 56 16 48 64 20 8 48 36 40 42 24 4 58 2 6 18 54 22 42 4 10 14 36 18 10 Cluster size ALEmax 32,144 0.0310 0.0239 0.0199 0.0187 0.0186 0.0168 0.0165 0.0148 0.0127 0.0268 0.0204 0.0120 0.0341 0.0515 0.0165 0.0145 0.0162 0.0143 0.0140 0.0112 0.0123 8856 5760 1224 1072 848 816 736 312 136 136 will be no new evidence, however, the focus on activated areas might change and a more general pattern can be identified. As described in the first section of this short review, embodied cognition theories propose several ideas. The first important one is the assumption that there is no separation between low and high cognitive processes. This assumption is tightly linked to the second claim, that sensorimotor systems are recruited, when verbal material is processed. Indeed, the main finding from this meta-analysis shows the clear involvement of a variety of regions, including mainly temporal (cluster 1, 2, 8) and frontal (cluster 1, 3), but also cerebellar activity (cluster 4, 10). Moreover, there is a clear predominance of activations in the (language and motor areas of the) left hemisphere. This could be due to a variety of reasons. First, language processing naturally occurs in the left hemisphere. In addition, the majority of participants in the present studies had right dominant effectors, hand and foot, which are processed contra-laterally. In D. Jirak et al. / Consciousness and Cognition 19 (2010) 711–720 717 Fig. 1. ALE results of the meta-analysis. Images display maximum ALE-values thresholded at p < .01 (FDR corrected). Left column displays 3D surface renderings from posterior-left, left, frontal, and right viewpoints. Lightbox images illustrate sections at x, y ,z = 38, 42, 20, x, y ,z = 38, 10, 26, x, y ,z = 4, 5, 56, and x, y ,z = 42, 16, 36 (from top to bottom). an fMRI study with a lexical decision task Willems et al. (2010) found a preferential activation of the left premotor cortex for right handers, and of the right premotor cortex for left handers, while responding to manual-action verbs (compared to nonmanual action verbs). However, it should be noted that during the experiments, participants were not performing actions. Rather, they had to read, listen and name actions without concomitant motor activity. Hence, it is important to emphasize that the motor areas such as SMA, the precentral gyrus or the premotor cortex were active when language only was processed. The meta-analysis also detects neural activations in the right hemisphere, especially in frontal (cluster 9) and parietal (cluster 5) regions. Indeed, there is evidence that the right hemisphere contributes to the processing of semantic components (Canessa et al., 2008; Gronau, Neta, & Bar, 2008; Kemmerer, Castillo, Talavage, Patterson, & Wiley, 2008). Another interesting result speaking in favor of motor activation irrespective of a distinction between low and high cognitive processes is the here identified participation of the cerebellum in language processing and, consequently, in language comprehension. This region has not yet been extensively focused on in the context of language, although it is well known that is plays a crucial role in motor learning. However, our results show a significant contribution of the left cerebellar hemisphere to language processing. The neural connections in participation with the cerebellum could be another interesting aspect pointing to the neural interoperability of different brain areas in embodied language. By relocating this result to the studies in the literature collection (Table 2) it becomes evident, that the cerebellum activity could be recorded within experiments which investigate action comprehension (Boulenger, Hauk, & Pulvermüller, 2009), action naming (Liljeström et al., 2008) or semantic processing (Saccuman et al., 2006). Therefore, the activation of the cerebellum suggests that language is embodied not only because sensorimotor areas are active, but also because words are processed along a frontal-parietal-temporal network including (clusters 1, 2) subcortical activity. An explanation for the distributed character of activity might lie in the distinction between the concepts retrieved from semantic knowledge and the perceptual component of words as hypothesized in Bedny, Caramazza, Grossman, Pascual-Leone, and Saxe (2008). The bilateral activity in the temporal lobe, namely the fusiform gyri, suggest that the posterior part of the temporal lobe is an area organizing concepts, rather than visual properties of words, which is in line with findings from Hauk and Pulvermüller (2004). Also Rüschemeyer et al. (2007) highlight the role of the right temporal area processing rather complex verbs with abstract meanings. Although they propose further work on that topic, our work provides supporting results for the implication of temporal areas and conceptualization or categorization. 718 D. Jirak et al. / Consciousness and Cognition 19 (2010) 711–720 Only partly reflected in this meta-analysis are the two proposals of embodied cognition theories, namely the somatotopic activation during the processing of action words and sentences, as well as the grounding of abstract concepts in the sensorimotor system. Neither has been investigated per se, as the number of included studies did not allow for a differential contrast between different effectors. 3. Concluding remarks This short review was conceptualized to give an overview along the lines of argumentation in the ongoing debate on the embodiment of language. We further employed an ALE meta-analysis to illustrate and relate previous findings. Both the narrative review, and the meta-analysis confirm the connection of language and motor areas. Primary motor, supplementary motor and premotor cortices are repeatedly reported to be active during language processing, and they are significantly present in the ALE results. If the hypothesis held true that the same neural units are responsible for action and simulation of action during language processing, classical characteristics of motor activation should be shared by language comprehension. Namely, language should produce the same somatotopy in activation as real action does. There is growing evidence supporting this view, especially for premotor areas. However, there is not enough data to validate this claim in a meta-analysis, which could theoretically be of immense help clarifying this issue. The idea, that the mirror neuron system plays an important role in the embodiment of language, is backed by a variety of findings in recent publications. Additionally, a set of peak points in the meta-analysis are very close to the locations reported for the mirror neuron system (Buccino et al., 2001, 2005). However, the strong claim of embodied theories, that mirror neurons represent the neural basis of the simulation activated during language comprehension, needs further thorough investigation. A region of interest in this can be Broca’s region, as this seems to have a core role in the integration of motion and sound. The claim of embodied theories is the grounding of abstract language in sensorimotor areas. Currently the findings on abstract words and sentence processing provide inconsistent results which are possibly due to the varieties in the definition of abstractness, but also to the variety of stimuli used in control conditions. Although our meta-analysis is not able to further enlighten this debate, the regions identified would be well suitable as regions of interest for further analyses on the processing of abstract language. The strongest results in all lines of argumentation are provided by the findings from patients and timing, both of which argue in favor of a necessary instead of an auxiliary role of sensorimotor areas in language processing. This very fast activation, its automaticity, taken together with the likely somatotopic organization render the hypothesis advanced among others by Mahon and Caramazza (2008), that information is first transduced in an abstract format and then influences the motor system, rather unlikely. The hypothesis that the motor system is activated in a direct and straightforward way is much more plausible and economical, even if evidence on timing and somatotopy still leave some unsolved issues. Finally, it should be noted that the discussion on embodiment should take into account in a sufficient way the strong plasticity and distributed character of the human brain. Consider some of the results we discussed. Even if abstract words are not represented in the same motor areas as concrete words, this would not necessarily render a problem. For example, it is possible that abstract words, due to the fact that they do not have a concrete referent, activate more language related areas (for a discussion, see Borghi & Cimatti, 2009; submitted). Even if the activation during language processing pertains the premotor cortices and not the primary motor ones, this would not undermine the embodied hypotheses. If we found that the activation of explicit imagery differs from that elicited by language comprehension, this result would even strengthen the embodied hypothesis. If some patients preserve their ability to comprehend language despite their motor disabilities, as it might happen with apractic patients, this might suggest that the brain is distributed and plastic enough. Acknowledgments This work was supported by the FP7 project ROSSI, Emergence of communication in Robots through Sensorimotor and Social Interaction, Grant Agreement No: 216125 and by German Ministry of Research BMBF (01GW0571 and 01GW0752). We thank the anonymous reviewers for helpful comments on an earlier version of the manuscript. References Aziz-Zadeh, L., & Damasio, A. (2008). Embodied semantics for actions: Findings from functional brain imaging. Journal de Physiologie – Paris, 102(1–3), 35–39. Aziz-Zadeh, L., Koski, L., Zaidel, E., Mazziotta, J., & Iacoboni, M. (2006). Lateralization of the human mirror neuron system. Journal of Neuroscience, 26(11), 2964–2970. Aziz-Zadeh, L., Wilson, S. M., Rizzolatti, G., & Iacoboni, M. (2006). Congruent embodied representations for visually presented actions and linguistic phrases describing actions. Current Biology, 16(18), 1818–1823. Bak, T. H., & Hodges, J. R. (2004). The effects of motor neurone disease on language: Further evidence. Brain and Language, 89(2), 354–361. Bak, T. H., Yancopoulou, D., Nestor, P. J., Xuereb, J. H., Spillantini, M. G., & Pulvermüller, F. (2006). Clinical, imaging and pathological correlates of a hereditary deficit in verb and action processing. Brain, 129, 321–332. Barsalou, L. W. (2008). Grounded cognition. Annual Review of Psychology, 59, 617–645. D. Jirak et al. / Consciousness and Cognition 19 (2010) 711–720 719 Bedny, M., Caramazza, A., Grossman, E., Pascual-Leone, A., & Saxe, R. (2008). Concepts are more than percepts: The case of action verbs. Journal of Neuroscience, 28(44), 11347–11353. Binkofski, F., & Buccino, G. (2004). Motor functions of the Broca’s region. Brain and Language, 89, 362–369. Binkofski, F., Buccino, G., Posse, S., Seitz, R. J., Rizzolatti, G., & Freund, H. (1999). A fronto-parietal circuit for object manipulation in man: Evidence from an fMRI-study. European Journal of Neuroscience, 11(9), 3276–3286. Borghi, A. M., & Cimatti, F. (2009). Words as tools and the problem of abstract words meanings. In N. Taatgen & H. van Rijn (Eds.), Proceedings of the 31st annual conference of the cognitive science society (pp. 2304–2309). Amsterdam: Cognitive Science Society. Borghi, A. M., & Scorolli, C. (2009). Language comprehension and hand motion simulation. Human Movement Science, 28, 12–27. Boulenger, V., Hauk, O., & Pulvermüller, F. (2009). Grasping ideas with the motor system: Semantic somatotopy in idiom comprehension. Cerebral Cortex, 19(8), 1905–1914. Boulenger, V., Mechtouff, L., Thobois, S., Broussolle, E., Jeannerod, M., & Nazir, T. A. (2008). Word processing in Parkinson’s disease is impaired for action verbs but not for concrete nouns. Neuropsychologia, 46(2), 743–756. Buccino, G., Binkofski, F., Fink, G. R., Fadiga, L., Gogassi, L., Gallese, V., et al (2001). Action observation activates premotor and parietal areas in a somatotopic manner: An fMRI study. European Journal of Neuroscience, 13, 400–404. Buccino, G., Riggio, L., Melli, G., Binkofski, F., Gallese, V., & Rizzolatti, G. (2005). Listening to action-related sentences modulates the activity of the motor system: A combined TMS and behavioural study. Brain Research, Cognitive Brain Research, 24(3), 355–363. Canessa, N., Borgo, F., Cappa, S. F., Perani, D., Falini, A., Buccino, G., et al (2008). The different neural correlates of action and functional knowledge in semantic memory: An FMRI study. Cerebral Cortex, 18(4), 740–751. Casasanto, D., & Boroditsky, L. (2008). Time in the mind: Using space to think about time. Cognition, 106, 579–593. Chersi, F., Thill, S., Ziemke, T., & Borghi, A. M. (2010). Sentence processing: Linking language to motor chains. Frontiers in Neurorobotics. D’Ausilio, A., Pulvermüller, F., Salmas, P., Bufalari, I., Begliomini, C., & Fadiga, L. (2009). The motor somatotopy of speech perception. Current Biology, 19(5), 381–385. Fischer, M., & Zwaan, R. (2008). Embodied language: A review of the role of the motor system in language comprehension. The Quarterly Journal of Experimental Psychology, 61, 825–850. Gallese, V. (2008). Mirror neurons and the social nature of language: The neural exploitation hypothesis. Social Neuroscience, 3, 317–333. Gennari, S. P., Mac Donald, M. C., Postle, B. R., & Seidenberg, M. S. (2007). Context-dependent interpretation of words: Evidence for interactive neural processes. Neuroimage, 35, 1278–1286. Glenberg, A. M., Sato, M., Cattaneo, L., Riggio, L., Palumbo, D., & Buccino, G. (2008). Processing abstract language modulates motor system activity. The Quarterly Journal of Experimental Psychology, 61, 905–919. Grafton, S. T., Arbib, M. A., Fadiga, L., & Rizzolatti, G. (1996). Localization of grasp representations in humans by positron emission tomography. 2. Observation compared with imagination. Experimental Brain Research, 112(1), 103–111. Gronau, N., Neta, M., & Bar, M. (2008). Integrated contextual representation for objects’ identities and their locations. Journal of Cognitive Neuroscience, 20(3), 371–388. Hauk, O., Johnsrude, I., & Pulvermüller, F. (2004). Somatotopic representation of action words in human motor and premotor cortex. Neuron, 41(2), 301–307. Hauk, O., & Pulvermüller, F. (2004). Neurophysiological distinction of action words in the fronto-central cortex. Human Brain Mapping, 2, 1191–1201. Kemmerer, D., Castillo, J. G., Talavage, T., Patterson, S., & Wiley, C. (2008). Neuroanatomical distribution of five semantic components of verbs: Evidence from fMRI. Brain and Language, 107(1), 16–43. Kohler, E., Keysers, C., Umilta, M. A., Fogassi, L., Gallese, V., & Rizzolatti, G. (2002). Hearing sounds, understanding actions: Action representation in mirror neurons. Science, 297(5582), 846–848. Laird, A. R., Fox, P. M., Price, C. J., Glahn, D. C., Uecker, A. M., Lancaster, J. L., et al (2005). ALE meta-analysis: Controlling the false discovery rate and performing statistical contrasts. Human Brain Mapping, 25(1), 155–164. Lakoff, G., & Johnson, M. (1980). Metaphors we live by. Chicago: University of Chicago Press. Lancaster, J. L., Rainey, L. H., Summerlin, J. L., Freitas, C. S., Fox, P. T., Evans, A. C., et al (1997). Automated labeling of the human brain: A preliminary report on the development and evaluation of a forward-transform method. Human Brain Mapping, 5, 238–242. Lancaster, J. L., Woldorff, M. G., Parsons, L. M., Liotti, M., Freitas, C. S., Rainey, L., et al (2000). Automated Talairach Atlas labels for functional brain mapping. Human Brain Mapping, 10, 120–131. Liljeström, M., Tarkiainen, A., Parviainen, T., Kujala, J., Numminen, J., Hiltunen, J., et al (2008). Perceiving and naming actions and objects. Neuroimage, 41(3), 1132–1141. Mahon, B. Z., & Caramazza, A. (2008). A critical look at the embodied cognition hypothesis and a new proposal for grounding conceptual content. Journal de Physiologie – Paris, 102(1–3), 59–70. McNamara, A., Buccino, G., Menz, M. M., Gläscher, J., Wolbers, T., Baumgärtner, A., et al (2008). Neural dynamics of learning sound–action associations. PLoS ONE, 3(12), e3845. Meister, I. G., & Iacoboni, M. (2007). No language-specific activation during linguistic processing of observed actions. PLoS ONE, 9, e891. Menz, M. M., & Binkofski, F. (2008). Broca’s region beyond language. The omnifariousness of Broca’s region. Rivista Medica, 14, 51–55. Nolfi, S., & Floreano, D. (2000). Evolutionary robotics. MIT Press. Papeo, L., Vallesi, Antonino, Isaja, Alessio, & Rumiati, Raffaella Ida (2009). Effects of TMS on different stages of motor and non-motor verb processing in the primary motor cortex. PLoS ONE, 4(2), e4508. Postle, N., McMahon, K. L., Ashton, R., Meredith, M., & de Zubicaray, G. I. (2008). Action word meaning representations in cytoarchitectonically defined primary and premotor cortices. Neuroimage, 43, 634–644. Pulvermüller, F. (2005). Brain mechanisms linking language and action. Nature Reviews Neuroscience, 6(7), 576–582. Pulvermüller, F., Lutzenberger, W., & Preissl, H. (1999). Nouns and verbs in the intact brain: Evidence from event-related potentials and high-frequency cortical responses. Cerebral Cortex, 9, 498–508. Raposo, A., Moss, H. E., Stamatakis, E. A., & Tyler, L. K. (2009). Modulation of motor and premotor cortices by actions, action words and action sentences. Neuropsychologia, 47, 388–396. Rizzolatti, G., & Arbib, M. A. (1998). Language within our grasp. Trends in Neurosciences, 21(5), 188–194. Rizzolatti, G., & Craighero, L. (2004). The mirror–neuron system. Annual Review of Neuroscience, 27, 169–192. Rüschemeyer, S.-A., Brass, M., & Friederici, A. D. (2007). Comprehending prehending: Neural correlates of processing verbs with motor stems. Journal of Cognitive Neuroscience, 19, 855–865. Rüschemeyer, S.-A., Pfeiffer, C., & Bekkering, H. (2010). Body schematics: On the role of the body schema in embodied lexical-semantic representations. Neuropsychologia, 48, 774–781. Rüschemeyer, S.-A., von Rooij, D., Lindemann, O., Willems, R., & Bekkering, H. (2009). The function of words: Distinct neural correlates for words denoting differently manipulable objects. Journal of Cognitive Neuroscience, 1–8. Saccuman, C. M., Cappa, S. F., Bates, E. A., Arevalo, A., Della Rosa, P., Danna, M., et al (2006). The impact of semantic reference on word class: An fMRI stud of action and object naming. Neuroimage, 32, 1865–1878. Scorolli, C., & Borghi, A. M. (2007). Sentence comprehension and action: Effector specific modulation of the motor system. Brain Research, 1130, 119–124. Siri, S., Tettamanti, M., Cappa, S. F., Della Rosa, P., Saccuman, C., Scifo, P., et al (2007). The neural substrate of naming events: Effects of processing demands but not of grammatical class. Cerebral Cortex, 171–177. Talairach, J., & Tournoux, P. (1988). A stereotactic coplanar atlas of the human brain. New York: Thieme. 720 D. Jirak et al. / Consciousness and Cognition 19 (2010) 711–720 Tettamanti, M., Buccino, G., Saccuman, M. C., Gallese, V., Danna, M., Scifo, P., et al (2005). Listening to action-related sentences activates fronto-parietal motor circuits. Journal of Cognitive Neuroscience, 17(2), 273–281. Tettamanti, M., Manenti, R., Della Rosa, P. A., Falini, A., Perani, D., Cappa, S. F., et al (2008). Negation in the brain: Modulating action representations. Neuroimage, 43, 358–367. Tomasino, B., Fink, G. R., Sparing, B., Dafotakis, M., & Weiss, P. H. (2008). Action verbs and the primary motor cortex: A comparative TMS study on silent reading, frequency judgment, and motor imagery. Neuropsychologia, 48(7), 1915–1928. Tomasino, B., Werner, C. J., Weiss, P. H., & Fink, G. R. (2007). Stimulus properties matter more than perspective: An fMRI study of mental imagery and silent reading of action phrases. Neuroimage, 36, T128–T141. Tremblay, P., & Gracco, V. L. (2006). Contribution of the frontal lobe to externally and internally specified verbal responses: fMRI evidence. Neuroimage, 33, 947–957. Vigliocco, G. (2009). Happiness is an abstract word. In N. Taatgen & H. van Rijn (Eds.), Proceedings of the 31st annual conference of the cognitive science society (pp. 1115–1120). Amsterdam: Cognitive Science Society. Willems, R. M., & Hagoort, P. (2007). Neural evidence for the interplay between language, gesture, and action: A review. Brain and Language, 101, 278–289. Willems, R. M., Hagoort, P., & Casasanto, D. (2010). Body-specific representations of action verbs: Neural evidence from right- and left-handers. Psychological Science, 21, 67–74. Willems, R. M., Ozyürek, A., & Hagoort, P. (2009). Differential roles for left inferior frontal and superior temporal cortex in multimodal integration of action and language. Neuroimage 1.47(4), 1992–2004. Ziemke, T. (Ed.). (2002). Special issue on Situated and embodied cognition. Cognitive Systems Research, 3, 3. Zwaan, R. (2004). The immersed experiencer: Toward an embodied theory of language comprehension. In B. H. Ross (Ed.). Psychology of learning and motivation (Vol. 44, pp. 35–62). New York: Academic.

Consciousness and Cognition 19 (2010) 1119–1121 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Commentary Duality and nonduality in meditation research q Zoran Josipovic Psychology Dept. and Center for Neural Science, New York University, 6 Washington Pl, Rm 957, New York, NY 10003, United States a r t i c l e i n f o Article history: Received 16 February 2010 Available online 10 April 2010 Keywords: Meditation Focused attention Open monitoring a b s t r a c t The great variety of meditation techniques found in different contemplative traditions presents a challenge when attempting to create taxonomies based on the constructs of contemporary cognitive sciences. In the current issue of Consciousness and Cognition, Travis and Shear add ‘automatic self-transcending’ to the previously proposed categories of ‘focused attention’ and ‘open monitoring’, and suggest characteristic EEG bands as the defining criteria for each of the three categories. Accuracy of current taxonomies and potential limitations of EEG measurements as classifying criteria are discussed. Ó 2010 Elsevier Inc. All rights reserved. Cognitive and affective neuroscience studies of meditation have a potential to make important contributions to the understanding of the brain’s functioning and neural plasticity (Lutz, Dunne, & Davidson, 2007; Lutz, Slagter, Dunne, & Davidson, 2008; Lazar et al., 2005). Furthermore, extensive first-person reports of changes brought on by these practices, may, once verified, enhance our views of the nature and functions of consciousness. The enormous variety of meditation techniques that have been developed over the centuries in the world’s contemplative traditions have presented an ongoing challenge to finding consistent and encompassing taxonomies. Only recently has this challenge emerged into full focus in the field of contemporary meditation research (Cahn & Polich, 2006; Lutz, Brefczynski-Lewis, et al., 2008; Lutz, Slagter, et al., 2008). The current paper by Travis and Shear is an important contribution in this direction. A recent effort by Lutz, Brefczynski-Lewis, et al. (2008) and Lutz, Slagter, et al. (2008) to operationalize meditation techniques in terms of the deployment of attentional strategies has resulted in a categorization of meditation techniques as belonging to either a ‘focused attention’ or ‘open monitoring’ style. This appears to be accurate, as most meditation techniques rely, at least at some stage of practice, on either endogenous or exogenous attentional systems, respectively (Austin, 2009). While this generalization covers a fairly large number of meditation techniques, especially within the Buddhist tradition, it surprisingly leaves out a very important, if not critical meditation style. Travis and Shear correctly point to this flaw and argue that Transcendental Meditation actually belongs to a new category, which they label ‘automatic self-transcending’. The term ‘transcendence’ has its inherent problems, and placing it in the context of cognitive psychology and neuroscience brings with it some unavoidable awkwardness. What the authors imply here is that something about the meditation techniques belonging to this category makes them auto-transcending, and leads to establishing of automaticity due to implicit learning. Transcending the technique, in this sense, is common among experienced practitioners of other meditation styles as well, as the authors note in reference to QiGong (Qin, Jin, Lin, & Hermanowicz, 2009), and the focused attention meditation in Tibetan Buddhist practitioners (Brefczynski-Lewis, Lutz, Schaefer, Levinson, & Davidson, 2007). Their claim in relation to TM is that developing a certain level of automaticity and ‘effortlessness’, happens relatively quickly because of the way in which q Commentary on Travis, F. & Shear, J. (2010). Focused attention, open monitoring and automatic self-transcending: Categories to organize meditations from Vedic, Buddhist and Chinese traditions. Consciousness and Cognition, 19, 1110–1118. E-mail address: zoran@cns.nyu.edu 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.03.016 1120 Z. Josipovic / Consciousness and Cognition 19 (2010) 1119–1121 the technique is set up, and that this is one of its chief differentiating features. Thus, in addition to differentiating meditations based on attentional strategy, this new category involves differences in working memory load. However, what is at stake here is a more profound difference, one that cannot be adequately captured within a singledimension characterization of attentional strategy. Both focused attention and open monitoring styles of meditation contain an essentially dualistic orientation of ‘subject-observing-object’. Yet, there is another group of meditations that do not employ this strategy, but instead rely on accessing a level of awareness that is inherently free from this dualistic subject–object construct. This non-conceptual awareness has sometimes been termed nondual awareness, open awareness or open presence (Kozhevnikov, Louchakova, Josipovic, & Motes, 2009; Lutz et al., 2007). Recognizing it within one’s waking experience, and learning to abide in it, is known in the Tibetan Buddhist tradition of Dzogchen as ‘‘taking the goal of meditation as the path”. In other traditions, such as various branches of Yoga and Vedanta, and in some other sects of Tibetan Buddhism, this non-conceptual awareness is first isolated from experience, and later, with practice, established in daily life. With this approach, there may be, initially, various degrees of focused attention deployment, until one can access this nondual awareness. This initial stage is perhaps what has led some to classify TM as a focused attention style of meditation. Thus, in terms of the actual goal of meditation practice, the fundamental differentiating feature of a meditation technique is whether it remains within the dualistic subject–object cognitive structure, or whether it transcends this structure to reveal the underlying nondual awareness. Some current research in cognitive neuroscience has attempted to differentiate attention from awareness (Brascamp, van Boxtel, Knapen, & Blake, 2009; Koch & Tsuchiya, 2007). The problem is that nondual awareness is not a level of consciousness that is known about or has been operationalized in cognitive neuroscience. So, while operationalizing different meditation techniques in terms of already established constructs of cognitive science allows for experimental tractability, it also creates a problem of accuracy as to the intended goal of meditation. It also leads to confounding the neural correlates of the meditation techniques that are used to get to particular ‘states’ of consciousness, with the correlates of the ‘states’ themselves. In order to solve this problem, both the taxonomy and the research of meditation need to be approached in a multidimensional fashion, where some of the dimensions could be: targeted states of consciousness; duality to nonduality scale; (which may or may not overlap with) stages of expertise; cognitive processes such as attentional strategies and working memory load; and objects of meditation. Travis and Shear suggest that the differences between meditation styles should be evident as fairly simple distinctions in EEG signatures. There is elegance and parsimoniousness to this idea, but the reality may be more complicated. The EEG signatures of meditation tend to be fairly complex across all bands and differ, as well, with the degree of the subject’s proficiency (Cahn & Polich, 2006). Changes in the gamma band, which the authors use as one of the indicators of focused attention style, have been found in other styles of meditation as well (Cahn, Delorme, & Polich, 2010). It is also questionable whether compassion meditation (Lutz, Greischar, Rawlings, Ricard, & Davidson, 2004) which produced some of the largest changes in the gamma band found in meditation to date, belongs to the focused attention style, as the authors suggest. The non-referentiality of compassion makes it more akin to meditations in the nondual or ‘automatic self-transcending’ category. Most importantly, certain aspects of synchrony in the gamma range await further clarification due to artifacts from scalp muscles and eye movement (Yuval-Greenberg, Tomer, Keren, Nelken, & Deouell, 2008). The two EEG signatures of meditation that at present appear to be most consistent are the increase in frontal midline theta, and the forward spread and increase in alpha. Whether they can be used as reliable indices for meditation categories requires further research. Finally, the lingering question is whether the changes in the EEG signal accurately reflect the subtle meditative states of consciousness or whether they reflect the overall levels of arousal in the brain and the specifics of various cognitive processes associated with the techniques of meditation. Expanding the current taxonomy of meditation and defining the characteristic neurophysiological signatures of various meditation categories are important issues in meditation research. Travis and Shear’s paper makes a significant contribution to their clarification. Advancing the research of meditation will add to the scientific understanding of brain functioning, and may help answer the larger social and psychological question about what it is to be an authentic, integrated and realized human being. References Austin, J. H. (2009). Selfless insight. Cambridge, MA: MIT Press. Brascamp, J. W., van Boxtel, J. J., Knapen, T., & Blake, R., (2009). A dissociation of attention and awareness in phase-sensitive but not phase-insensitive visual channels. Journal of Cognitive Neuroscience, 1–19 [Epub ahead of print]. Brefczynski-Lewis, J. A., Lutz, A., Schaefer, H. S., Levinson, D. B., & Davidson, R. J. (2007). Neural correlates of attentional expertise in long-term meditation practitioners. PNAS 104, 11483–11488. Cahn, B. R., Delorme, A., & Polich, J. (2010). Occipital gamma activation during Vipassana meditation. Cognitive Processing, 11, 39–56. Cahn, B. R., & Polich, J. (2006). Meditation states and traits: EEG, ERP, and neuroimaging studies. Psychological Bulletin, 132(2), 180–211. Koch, C., & Tsuchiya, N. (2007). Attention and consciousness: Two distinct brain processes. Trends in Cognitive Sciences, 11(1), 16–22. Kozhevnikov, M., Louchakova, O., Josipovic, Z., & Motes, M. A. (2009). The enhancement of visuospatial processing efficiency through Buddhist deity meditation. Psychological Science, 20(5), 645–653. Lazar, S., Kerr, C. E., Wasserman, R. H., Grey, J. R., Greve, D. N., Treadway, M. T., et al (2005). Meditation experience is associated with increased cortical thickness. NeuroReport, 16, 1893–1897. Lutz, A., Brefczynski-Lewis, J., Johnstone, T., & Davidson, R. J. (2008). Regulation of the neural circuitry of emotion by compassion meditation: Effects of meditative expertise. PLoS One, 26, e1897. Lutz, A., Dunne, J., & Davidson, R. (2007). Meditation and the neuroscience of consciousness. In P. D. Zelazo, M. Moscovitch, & E. Thompson (Eds.), The Cambridge handbook of consciousness (pp. 499–551). Cambridge, England: Cambridge University Press. Z. Josipovic / Consciousness and Cognition 19 (2010) 1119–1121 1121 Lutz, A., Greischar, L. L., Rawlings, N. B., Ricard, M., & Davidson, R. J. (2004). Long-term meditators self-induce high-amplitude gamma synchrony during mental practice. Proceedings of the National Academy of Science, 101(46), 16369–16373. Lutz, A., Slagter, H. A., Dunne, J. D., & Davidson, R. J. (2008). Attention regulation and monitoring in meditation. Trends in Cognitive Sciences, 12(4), 163–169. Qin, Z., Jin, Y., Lin, S., & Hermanowicz, N. S. (2009). A forty-five year follow-up EEG study of qigong practice. International Journal of Neuroscience, 119(4), 538–552. Travis, F., & Shear, J. (2010). Focused attention, open monitoring and automatic self-transcending: Categories to organize meditations from Vedic, Buddhist and Chinese traditions. Consciousness and Cognition, 19, 1110–1118. Yuval-Greenberg, S., Tomer, O., Keren, A. S., Nelken, I., & Deouell, L. Y. (2008). Transient induced gamma-band response in EEG as a manifestation of miniature saccades. Neuron, 58(3), 429–441.

Consciousness and Cognition 22 (2013) 955–964 Contents lists available at SciVerse ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Is ‘‘R’’ purple or green? Bistable grapheme-color synesthesia induced by ambiguous characters Suhkyung Kim a, Randolph Blake b,c, Chai-Youn Kim a,⇑ a Department of Psychology, Korea University, Seongbuk-Gu, Anam-Dong, Seoul 136-701, Republic of Korea Department of Psychology/Vanderbilt Vision Research Center, Vanderbilt University, 512 Wilson Hall, 111 21st Avenue South, Nashville, TN 37240, United States c Brain and Cognitive Sciences, Seoul National University, Daehak-dong, Gwanak-gu, Seoul 151-742, Republic of Korea b a r t i c l e i n f o Article history: Received 28 April 2011 Available online 17 July 2013 Keywords: Grapheme-color synesthesia Bistability Temporal relation Grapheme identification Color naming a b s t r a c t People with grapheme-color synesthesia perceive specific colors when viewing different letters or numbers. Previous studies have suggested that synesthetic color experience can be bistable when induced by an ambiguous character. However, the exact relationship between processes underlying the identity of an alphanumeric character and the experience of the induced synesthetic color has not been examined. In the present study, we explored this by focusing on the temporal relation of inducer identification and color emergence using inducers whose identity could be rendered ambiguous upon rotation of the characters. Specifically, achromatic alphabetic letters (W/M) and digits (6/9) were presented at varying angles to 9 grapheme-color synesthetes. Results showed that grapheme identification and synesthetically perceived grapheme color covary with the orientation of the test stimulus and that synesthetes were slower naming the experienced color than identifying the character, particularly at intermediate angles where ambiguity was greatest. Ó 2013 Elsevier Inc. All rights reserved. 1. Introduction Synesthesia is the condition where a stimulus in one sensory modality induces perceptual experience not only in the relevant sensory modality but also in another, usually irrelevant modality (Cytowic & Eagleman, 2009). Synesthesia can involve cross-modal associations such as vision-induced auditory experiences and taste-induced tactile experiences (Cytowic, 1989). Other cases of synesthesia involve intra-modal associations such as grapheme-color synesthesia, one of the most prevalent forms of synesthesia (Rich, Bradshaw, & Mattingley, 2005; Simner et al., 2006). It is this type of synesthesia that we focus on in this paper. People with grapheme-color synesthesia describe seeing achromatic letters and/or numerals as colored. Some people (dubbed associators) report that these illusory colors appear in their mind’s eye, whereas others (dubbed projectors) actually see the colors on the inducing characters themselves (Dixon, Smilek, & Merikle, 2004). Regardless of which type of grapheme-color synesthesia we’re dealing with, the associations of character and color are highly specific and enduring. Much remains to be learned about this fascinating propensity, but one thing can be said with certainty: a given color cannot be evoked without the identity of the character being established. The direction of the causal arrow is unequivocal. Note, by the way, that this conclusion does not necessarily mean that a character must be consciously perceived before color can emerge. Indeed, there is conflicting evidence concerning the role of awareness in grapheme-color synesthesia, with some results pointing to color induction without conscious character identification (e.g., Ramachandran & Hubbard, 2001; Smilek, ⇑ Corresponding author. E-mail address: chaikim@korea.ac.kr (C.-Y. Kim). 1053-8100/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.concog.2013.05.012 956 S. Kim et al. / Consciousness and Cognition 22 (2013) 955–964 Dixon, Cudahy, & Merikle, 2001; Smilek, Dixon, & Merikle, 2005; Wagar, Dixon, Smilek, & Cudahy, 2002; but see also Hubbard, Arman, Ramachandran, & Boynton, 2005) and other results implying that synesthetic color is available after the inducing stimulus is consciously identified (Mattingley, Rich, Yelland, & Bradshaw, 2001; Rich & Mattingley, 2010). In the study described in this paper we have not attempted to resolve this issue but, instead, have sought to answer a related question about the emergence of synesthetic color under conditions where the identity of the inducing character is ambiguous and, therefore, requires additional time to be identified. Stimulus ambiguity, in general, provides a useful tool for studying fluctuations in perception despite unchanging physical stimulation (Kim & Blake, 2005). There are situations where a given inducing character can take on one of several different identities. Consider, for example, the two character sequences in Fig. 1A, where the middle character in both sequences is physically identical but semantically different owing to the context in which it appears. When synesthetes view character sequences like this, the reported color of the ambiguous character corresponds to the type of character – letter vs numeral – implied by the sequence (Dixon, Smilek, Duffy, Zanna, & Merikle, 2006). Likewise, when a large character is produced by appropriate arrangement of small characters different from the global shape they form (Fig. 1B), synesthetes experience either of two colors dependent on whether their attention is focused at the local or the global level of representation (Palmeri, Blake, Marois, Flanery, & Whetsell, 2002). Recently, Bridgeman, Winter, and Tseng (2010) described another instance where synesthetic color experience evoked by ambiguous inducing characters is ambiguous. They had synesthetic observers view alphabetic letters that continuously rotated in a clockwise direction, and the observers were instructed to verbalize the color they were experiencing throughout the presentation, being sure to report if one color replaced another. At the end of each rotation presentation, the synesthetes were required to remember and report the orientation of the letter at the time the color switched, if it did. Bridgeman et al. found that synesthetic colors tended to change. The results of Bridgeman et al. fit nicely with the earlier studies using ambiguous inducing characters, but they do not tell us anything about the time course of this change in character and color identification with orientation, information that could be potentially revealing with respect to processes underlying grapheme-color synesthesia. Accordingly, we performed a different version of the letter rotation experiment that allowed us to assess psychometric performance together with reaction time performance. In our study, we took advantage of particular pairs of synesthesia-inducing characters that could be Fig. 1. Examples of ambiguous characters inducing bistable synesthetic colors. (A) The physically identical middle character can be identified either as the digit ‘5’ (top) or the letter ‘S’ (bottom) depending on the context in which it appears; the associated synesthetic color varies depending on the letter perceived (Myles et al., 2003). (B) A global–local figure, in this example a large ‘5’ composed of a number of small ‘2’s. Synesthete WO experiences orangish red (his ‘‘color’’ for ‘2’) when he attends to the local feature and he experiences green (his ‘‘color’’ for ‘5’) when he attends to the global shape. S. Kim et al. / Consciousness and Cognition 22 (2013) 955–964 957 recognized as either of two letters or two numbers when rotated through 180 degrees. For example, the Roman alphabetic letter ‘‘M’’ becomes ‘‘W’’ when turned upside down and, similarly, the digit 9 becomes 6. When those particular characters assume intermediate degrees of rotation, their identities become ambiguous; befitting their ambiguous status, people report sometimes seeing a given, rotated character as M (or 6) and other times as W (or 9). Using such character pairs in a speeded judgment task, we measured response times (RTs) for identifying a character and for identifying its (synesthetic) color. We tested a group of four grapheme-color synesthetes whose native language is English (the ‘‘US group’’) and a group of five bilingual grapheme-color synesthetes who learned English as young children (the ‘‘Korean group’’). We reasoned that both groups would have had equivalent years of experience with digits (the Korean number system being identical to the US) but somewhat different total experience with Roman alphabetic characters (since English was their second language, and the vast majority of their reading throughout their lives was in Korean). The Korean synesthetes, not surprisingly, also experience grapheme-color synesthesia when viewing Korean characters. But we wish to stress that they characterize their synesthetic photisms for Roman alphabetic characters as being as vivid and unchanging as the photisms experienced when viewing Korean characters. 2. Material and methods 2.1. Participants Nine grapheme-color synesthetes participated in the Experiment. Four of them (2 females; mean age = 32.75 years, range: 20–65) were native English speakers and they were recruited and tested at Vanderbilt University. All four were ‘projectors’, meaning that they ‘‘see’’ their synesthetic colors located in visual space on the alphanumeric characters themselves (Dixon et al., 2004). The other five (all female; mean age = 23.2 years, range: 20–28) were native Korean speakers having about 10 years of experience learning English and they were recruited and tested at Korea University. All five of them reported their synesthetic colors to appear in their mind’s eye and, therefore, were classified as associators (Dixon et al., 2004). All nine participants had normal or corrected-to-normal visual acuity. Prior to the experiment, they gave written informed consent approved by Vanderbilt University Institutional Review Board (4 US synesthetes) or by Korea University Institutional Review Board (5 Korean synesthetes). 2.2. Apparatus For both the Korea University and Vanderbilt University experiments, stimuli were presented on a 17-inch CRT monitor (1024  768 resolution, 60-Hz frame rate) under the control of an Intel PC using Matlab 7.0.4 (MathWorks, Co.) and Psychophysics Toolbox 2.54 (Brainard, 1997; Pelli, 1997). Testing was done in a quiet, dark room in which the video monitor provided the only source of illumination. 2.3. Stimulus and task conditions Two pairs of graphemes – W/M and 6/9 – were used as stimuli. Printed in Skia font, the upside down and right-side up versions of the members of a pair were identical, and when shown at intermediate orientations at and around 90 deg the character was ambiguous (Fig. 2A and B). Participants confirmed that both members of both stimulus pairs induced comparably vivid synesthetic colors (Fig. 2C). For all observers except two, the synesthetic colors of the two members of a pair were distinctly different; L.R. (a member of the US group) experiences 6 and 9 as subtly different shades of purple and S.Y. (a member of the Korean group) experiences W and M as subtly different shades of blue. On each trial, a member of a given stimulus pair was presented in one of seven orientations (0, 30, 60, 90, 120, 150 or 180 angular degrees, where 0 refers to the upright position for a given character); when appearing in its non-canonical orientation, the direction of its rotated orientation could be in either the clockwise (CW) or the counterclockwise (CCW) direction. Each stimulus subtended 5.13  5.35 degree of visual angle and was shown in white against a dark background, yielding a contrast close to 100%. The two stimulus conditions – alphabetic letters (W/M) and digits (6/9) – were presented in separate blocks. For both the digit condition and the letter condition, the two task conditions were administered in separate blocks of trials, one devoted to character identification and the other to perceived synethetic color. In the character identification task, participants indicated the perceived identity of the stimulus (e.g., ‘‘W or M’’) by pressing one of two pre-designated keyboard keys, and in the synesthetic color task they indicated the perceived ‘‘color’’ of the stimulus (e.g., ‘‘purple or green’’) again by pressing one of two keys. The experiment included a total of 6 blocks (3 stimulus conditions  2 task conditions), and each block contained 280 trials (7 orientation values  2 directions  20 repetitions). Four US synesthetes were tested with the alphabetic letters, but only three were tested on the digit task and one of those participants (L.R.) had to be dropped from that condition because she perceived the 6 and the 9 as highly similar in color. All five Korean synesthetes were tested with the digits, but only four were tested on the letter task and one of those participants (S.Y.) had to be dropped from that condition because she 958 S. Kim et al. / Consciousness and Cognition 22 (2013) 955–964 Fig. 2. Examples from the two stimulus conditions and the color matching results from the synesthetes tested. (A) An alphabet letter can be recognized either as ‘‘W’’ or as ‘‘M’’ when shown at intermediate orientations at and around 90 deg. Synesthetic color experience will change accordingly; the letter can be experienced either as ‘‘purple’’ or as ‘‘green’’ (Based on HP’s color matching results, see figure C). (B) A digit can be recognized either as ‘‘6’’ or as ‘‘9’’ when shown at intermediate orientations at and around 90 deg. Synesthetic color experience will change accordingly; the digit can be experienced either as ‘‘red’’ or as ‘‘pink’’. (Based on HP’s color matching results, see figure C). (C) Color matching results for W, M, 6, and 9 from eight of nine synesthetes in our two groups; one synesthete - WO - was not tested with this color-matching procedure. Matched colors from three repeated trials for each of the four chracters are shown. The variability of those three color matches for each character is expressed numerically by the VS index values and by the small, orange horizontal bars appearing at the right-hand side of each row of character color matches; those values were obtained using a standardized, web-based synesthesia test (Eagleman, Kagan, Nelson, Sagaram, & Sarma, 2007). perceived the W and the M as highly similar in color. The order of the blocks was counterbalanced among participants, and the order of the trials within a block was unpredictable within the limits of the MATLAB randomization routine. 2.4. Procedure Fig. 3 shows the sequence of an example trial. A trial began with a white stimulus presented on the exact center of the black background for 100 msec. A white square-shaped border (6.89  6.89 degrees in visual angle) was presented to indicate the location of the stimulus. In each trial, the target was presented at one of the seven orientations in either of the two possible directions. A pattern mask appeared immediately following the stimulus and remained present until the observer responded. Observers were urged to respond as quickly as possible. S. Kim et al. / Consciousness and Cognition 22 (2013) 955–964 959 Fig. 3. Experimental Procedure. Stimuli were presented for 100 msec and then masked until a response was made by the observer. The angular orientation of the inducer characters varied randomly over trials, with orientation ranging from 0 to 180 degrees in 30 degree steps. In separate blocks of trials observers performed a character identification task and synesthetic color identification task. 2.5. Data analysis To examine the difference between task (character and synesthetic color) conditions, we performed psychometric curve fitting and RT analyses. Accuracy could only be assessed on trials where the characters were at or very near their cardinal orientations; when characters appeared at intermediate orientations, there was no explicitly correct answer because the character’s identity and, by inference, its perceived ‘‘color’’ were ambiguous. Both character and synesthetic color psychometric curves in the digit and letter conditions were generated by psignifit toolbox version 2.5.6 with Matlab (http://bootstrap-software.org/psignifit/), implementing the maximum-likelihood method (Wichmann & Hill, 2001a). Each curve was fitted by a general psychometric function: W (x;a,b,c,k) = c + (1 - c- k) F (x; a,b) where F was Cumulative Gaussian function. Four parameters (a, b, c,k) were estimated from the simulations. Confidence intervals spanned .05 to .95 based on 4999 simulations (Wichmann & Hill, 2001b). Mean RTs were also computed to compare the perceptual latency between the character and color task conditions. Outliers were eliminated based on ±3 standard deviation cutoff. 3. Results Fig. 4 shows the proportion of trials on which given character identification and synesthetic color responses were made to digits (Fig. 4A) and to letters (Fig. 4B) whose orientations varied over trials. Combined results (the panel labeled Total) are shown together with results categorized by the observers’ nationality. All synesthetes in both groups tested on these conditions showed the expected dependence of character identification on the orientation of the stimulus: at and very near the cardinal orientations, character identity was corresponded to the expected character (e.g., ‘‘9’’ was always identified as ‘‘9’’), but at intermediate orientations centered around 90 deg identification of character and of color varied over trials (e.g., ‘‘R’’ was sometimes reported as W and other times as M). We found no evidence for a shift in the ‘‘color’’ psychometric curve relative to the character psychometric curve. Just to reiterate a point made earlier, the character identification judgments and the ‘‘color’’ judgments were made on separate trials administered in separate blocks; the design of our task did not permit collecting ‘‘color’’ and character identification RTs on the same trial. This equivalence in the patterns of results for character identification and synesthetic color is gratifying but not surprising, and it dovetails with earlier findings that the particular color experienced during grapheme-color synesthesia depends on how a given stimulus is perceptually organized (Bridgeman et al., 2010; Myles, Dixon, Smilek, & Merikle, 2003; Palmeri et al., 2002). Of more significance are the RT results obtained on these tasks. Considering first the results from the digit condition, Fig. 5 plots the RTs for the digit identification task and the ‘‘color’’ judgment task as a function of stimulus orientation (with angular deviation expressed relative to ‘‘6’’). The blue lines plot RTs for the character task and the red lines plot RTs for the ‘‘color’’ task; the shaded regions demarcate ±1 standard error of the mean. The seven small panels with initials above them show results for each individual, the two intermediate sized panels show results for the US participants and the Korean participants, and the large panel shows the average across all participants. One clearly sees that RTs for the digit condition tend to be longer at and near 90 deg orientation, which makes sense because the stimuli were more ambiguous and, presumably, the task more difficult at these intermediate orientations. One could construe these curves as evidence for the involvement of mental rotation of the sort documented by Shepard and Metzler (1971), although there was nothing in the task instructions in our study implying that this was the strategy to be adopted. Of particular relevance to the question motivating our study are the consistently faster RTs for character identification compared to ‘‘color’’ identification. A 2-way repeated measures ANOVA was performed across all seven participants (2 US 960 S. Kim et al. / Consciousness and Cognition 22 (2013) 955–964 Fig. 4. Proportion of trials on which given character identification and color responses were made. (A) Fitted psychometric functions show the averaged total result (upper), averaged results from the two sub-groups of synesthetes (the US (lower left) and the Korean (lower right)). The proportion of perceiving ‘‘9’’ (blue; character task) and the idiosyncratic ‘‘color’’ of ‘‘9’’ (red; synesthetic color task) are plotted against degree of angular deviation from ‘‘6’’ (and the ‘‘color’’ of 6). (B) Fitted psychometric functions show the averaged total result (upper), averaged results from the two sub-groups of synesthetes (the US (lower left) and the Korean (lower right)). The proportion of perceiving ‘‘M’’ (blue; character task) and the idiosyncratic ‘‘color’’ of ‘‘M’’ (red; synesthetic color task) are plotted against degree of angular deviation from ‘‘W’’. S. Kim et al. / Consciousness and Cognition 22 (2013) 955–964 961 Fig. 5. RT results from the digit (6/9) condition; the averaged total (upper center), averages of the two sub-groups of synesthetes (lower center; the US (left) and the Korean (right)), and individual results. The RTs in the character task (blue line) and in the synesthetic color task (red line) are plotted against degree of angular deviation from ‘‘6’’. Shaded areas denote ± 1 standard error of the mean (blue: character task, red: synesthetic color task). and 5 Korean) in which task and orientation were within-factor variables. The main effects of task and orientation were both significant statistically (F(1, 6) = 6.879, p < 0.05; F(6, 36) = 11.097, p < 0.0001). There was no statistically significant interaction between task and orientation (F(6, 36) = 1.345, p = 0.2331). These findings clearly imply that digits are being identified more quickly than their synesthetic colors are being identified. Results from the letter condition are shown in Fig. 6, using the same format as panels in Fig. 5. Again we find the tendency for RTs to vary with character orientation, being slower at orientations where the letter’s identity is most ambiguous. Moreover, we again see a consistent trend for the character identity RTs to be faster than the ‘‘color’’ identity RTs. These impressions are borne out by the repeated measures 2-way ANOVA performed on these data. As in the digit RT analysis, task and orientation were within-factor variables. The main effects of task and orientation were both significant statistically (F(1, 7) = 6.591, p < 0.05; F(6, 42) = 10.918, p < 0.0001), and in this condition the interaction between task and orientation was also marginally significant (F(6, 42) = 2.018, p = 0.0599). So, again, we are finding that an inducer (a letter in this case) is identified faster than its associated color. One intriguing observation that deserves comment is revealed when we compare the average RTs for the Korean synesthetes on the two conditions: they are slower in their responses on the letter condition (W vs M) compared to the digit condition (6 vs 9). No such tendency is seen in the RT results for the US participants. Why might this be? Perhaps it is related to the Koreans’ differential exposure to digits and Roman letters; they encountered digits very early in and very often throughout their lives, but their exposure to Roman characters did not start until just before adolescence and continues to be less frequent than their exposure to Korean characters. Nonetheless, it is noteworthy that their RTs to the synesthetic colors experienced when viewing these Roman characters are still slower, on average, than their RTs to identifying those characters. 962 S. Kim et al. / Consciousness and Cognition 22 (2013) 955–964 Fig. 6. RT results from the letter (W/M) condition; the averaged total (upper center), averages of the two sub-groups of synesthetes (lower center; from the US (left) and the Korean (right)), and individual results. The RTs in the character task (blue line) and in the synesthetic color task (red line) are plotted against degree of angular deviation from ‘‘W’’. Shaded areas denote ±1 standard error of the mean (blue: character task, red: synesthetic color task). 4. Discussion Using inducing characters whose identity varies with orientation, we have found that grapheme identification and synesthetically perceived grapheme color covary with the orientation of the test stimulus. In our study, we did not collect identification and ‘‘color’’ responses on the same trial. Consequently, we cannot state for certain whether the covariation in those two performance measures seen in the psychometric curves occurs on an individual, trial by trial basis. We are gratified, however, by the nearly complete overlap of the two classes of psychometric curves. Using rather different techniques, Bridgeman et al. (2010) also found that synethetic color depended on character orientation. On each trial they presented a single alphabetic letter that rotated in clockwise direction and instructed their synesthetic participants ‘‘to name the color they saw, and to name a new color when they saw a change. After each rotation trial, they also reported the remembered letter orientations at the times of color changes, if any (p. 672)’’. We, on the other hand, used a speeded forced-choice judgment task to assess character identification and perception of synesthetic color. Results from this objective task confirmed the phenomenological reports obtained by Bridgeman et al. and, furthermore, disclosed that ‘‘color’’ judgments take longer than do character identification judgments. In the following paragraphs we consider several possible implications of these RT differences. To what can we attribute the difference in processing time between character and synesthetic color identification? One popular account of grapheme-color synesthesia posits that synesthetes have unusually dense fiber connections between the visual word-form area and color selective area within the temporal lobe/ventral stream pathway (e.g., Baron-Cohen, Harrison, Goldstein, & Wyke, 1993), a prediction that was subsequently verified using brain imaging techniques (Rouw & Scholte, 2007). There are other variants of this kind of model that emphasize long-range disinhibition (Grossenbacher & Lovelace, 2001) or aberrant re-entrant processing (Smilek et al., 2001) that amplifies feedback signals from high-tiered cortical areas to color processing areas. On all versions of this model, the signals innervating color-selective visual areas arise in those areas later in time relative to the signals originating within visual areas registering information about the inducers. It stands to reason, then, that perceptual judgments about synesthetic color could lag judgments about inducer identity. It is arguable however, whether this model would predict a difference in processing time of almost 1/10th a second for character vs ‘‘color’’ identification, yet that is the RT difference value we calculated from the RT data pooled across digit and letter conditions. S. Kim et al. / Consciousness and Cognition 22 (2013) 955–964 963 Another, altogether different account of synesthesia questions whether grapheme-color synesthesia is perceptual in nature in the first place (Gheri, Chopping, & Morgan, 2008). Perhaps, the argument goes, people describing synesthetic experiences are basing those descriptions on overlearned associations, meaning that synesthesia is cognitive, not perceptual, in origin. Our RT results with ambiguous characters could be construed as consistent with this view, with the additional time required to name a characters’ synesthetic color arising from the involvement of lexical processing required to pair the appropriate color name with the character currently being viewed. We have no rejoinder to that interpretation of our results, other than an appeal to the subjective reports of our participants who denied employing such a strategy. But psychologists have long known that verbal reports are unreliable reflections of underlying mental processes, even when the person reporting has high confidence in his/her introspective awareness (e.g., Nisbett & Wilson, 1977). So for now, we shall avoid overinterpreting our results in terms of the origin(s) of the RT differences between character and ‘‘color’’ identification, and let the findings speak for themselves. We want to reiterate two notable aspects of the present results for the two subject groups. As pointed out in the Methods section, all four US synesthetes were classified as projectors and all five Korean synesthetes were classified as associators. However, all members of these two groups performed comparably on both of the conditions reported in this paper. Evidently, then, the co-dependence of synesthetic color and character identification on orientation does not seem to differ between these putative categories of synesthesia. The second notable finding was the overall slower RTs in the Korean group when performing the tasks involving letters compared to performance of the digit task. This task-related difference in overall RT was not observed in the US group, leading us to speculate that this difference in the Koreans may have something to do with their more enduring, intense exposure to digits compared to Roman letters. In light of this possibility, it will be informative to utilize ambiguous Korean figures that evoke colors in our Korean subjects. One such figure that accomplishes this is the Korean word pair / (‘‘ ’’ means ‘‘door and ‘‘ ’’ means ‘‘bear’’). Note, these constitute words, not single letters, but still the logic of the task is the same as the characters we used in the present study. Our US subjects, of course, would be baffled by the task. Finally, we want to mention some additional observations collected from some of our participants at the end of the study. For one set of observations, we had each of three participants view one of the characters on the video monitor with their heads maintained in the upright position. All three readily named the character and the color that it reliably evokes. Next we had them turn around and view the same display while bending over and looking between their legs. This maneuver, of course, inverts the image of the character on the retina, although participants readily understand that nothing has changed on the screen. All observers promptly and with confidence asserted that the color of the character had changed, befitting the new orientation of the image on their retina. This simple test suggests that the character orientations that we are talking about in this paper are registered in head-based, retinal coordinates and not environmental, gravitational-based coordinates. But confounding this conclusion is a second set of observations made by the four Korean participants. Specifically, we asked them to look at an inducer (e.g., upright ‘W’) on the screen, to imagine it being upside down, and to described what it’s color appearance. All now reported a color associated with the inverted version of the character (e.g., ‘M’) they were looking at. In other words, mental rotation of an inducer that remains physically the same in the world can change the synesthetic color associated with that inducer. Interestingly, two synesthetes (S.K. and Y.K.) voluntarily reported that it takes long time with conscious mental efforts before they experience their (synesthetic) color of ‘M’ while viewing ‘W’ on the monitor screen. By showing this interplay between retinal image orientation and mental image orientation, these results pose an interesting challenge for any theory of synesthesia claiming that synesthesic colors are exclusively associated the meanings of the letters as they exist in the environment. Acknowledgments This work was supported by Basic Science Research Program through the National Research Foundation of Korea funded by the Ministry of Education, Science and Technology (NRF-2009-0089090) to C.-Y.K. and by the World Class University program through the Korea National Research Foundation funded by the Ministry of Education, Science and Technology (R3210142) to R.B. We thank Sang Wook Hong for his help in collecting data and for useful discussion throughout the project. References Baron-Cohen, S., Harrison, J., Goldstein, L. H., & Wyke, M. (1993). Coloured speech perception: Is synesthesia what happens when modularity breaks down? Perception, 22, 419–426. Brainard, D. H. (1997). The psychophysics toolbox. Spatial Vision, 10, 433–436. Bridgeman, B., Winter, D., & Tseng, P. (2010). Dynamic phenomenology of grapheme-color synesthesia. Perception, 39(5), 671–676. Cytowic, R. E. (1989). Synesthesia and mapping of subjective sensory dimensions. Neurology, 39(6), 849–850. Cytowic, R. E., & Eagleman, D. M. (2009). Wednesday is indigo blue – Discovering the brain of synesthesia. Cambridge, MA: MIT Press. Dixon, M. J., Smilek, D., Duffy, P. L., Zanna, M. P., & Merikle, P. M. (2006). The role of meaning in grapheme-colour synaesthesia. Cortex, 42(2), 243–252. Dixon, M. J., Smilek, D., & Merikle, M. P. (2004). Not all synaesthetes are created equal: Projector versus associator synaesthetes. Cognitive, Affective and Behavioral Neuroscience, 4(3), 335–343. Eagleman, D. M., Kagan, A. D., Nelson, S. S., Sagaram, D., & Sarma, A. K. (2007). A standardized test battery for the study of synesthesia. Journal of Neuroscience Methods, 159, 139–145. Gheri, C., Chopping, S., & Morgan, M. J. (2008). Synaesthetic colours do not camouflage form in visual search. Proceeding of Royal Society B, 275, 841–846. Grossenbacher, P. G., & Lovelace, C. T. (2001). Mechanisms of synesthesia: Cognitive and physiological constraints. Trends in Cognitive Sciences, 5(1), 36–41. 964 S. Kim et al. / Consciousness and Cognition 22 (2013) 955–964 Hubbard, E. M., Arman, A. C., Ramachandran, V. S., & Boynton, G. M. (2005). Individual differences among grapheme-color synesthetes: Brain-behavior correlations. Neuron, 45, 975–985. Kim, C.-Y., & Blake, R. (2005). Psychophysical magic: Rendering the normally visible ‘‘invisible’’. Trends in Cognitive Sciences, 9(8), 381–388. Mattingley, J. B., Rich, A. N., Yelland, G., & Bradshaw, J. L. (2001). Unconscious priming eliminates automatic binding of colour and alphanumeric form in synesthesia. Nature, 410, 580–582. Myles, K. M., Dixon, M. J., Smilek, D., & Merikle, P. M. (2003). Seeing double: The role of meaning in alphanumeric-colour synaesthesia. Brain and Cognition, 53, 342–345. Nisbett, R. E., & Wilson, T. D. (1977). Telling more than we can know – Verbal reports on mental processes. Psychological Review, 84(3), 231–259. Palmeri, T., Blake, R., Marois, R., Flanery, M., & Whetsell, W. (2002). The perceptual reality of synesthetic colors. Proceedings of the National Academy of Sciences of the United States of America, 99(6), 4127–4131. Pelli, D. G. (1997). The Video Toolbox software for visual psychophysics: Transforming numbers into movies. Spatial Vision, 10, 437–442. Ramachandran, V. S., & Hubbard, E. M. (2001). Synesthesia: A window into perception, thought and language. Journal of Consciousness Studies, 8, 3–34. Rich, A. N., Bradshaw, J. L., & Mattingley, J. B. (2005). A systematic, large-scale study of synaesthesia: Implications for the role of early experience in lexical colour associations. Cognition, 98(1), 53–84. Rich, A. N., & Mattingley, J. B. (2010). Out of sight, out of mind: The attentional blink can eliminate synaesthetic colours. Cognition, 114(3), 320–328. Rouw, R., & Scholte, H. S. (2007). Increased structural connectivity in grapheme-color synesthesia. Nature Neuroscience, 10(6), 792–797. Shepard, R. N., & Metzler, J. (1971). Mental rotation of three dimensional objects. Science, 171, 701–703. Simner, J., Mulvenna, C., Sagiv, N., Tsakanikos, E., Witherby, S. A., Fraser, C., et al (2006). Synaesthesia: The prevalence of atypical cross-modal experiences. Perception, 35(8), 1024–1033. Smilek, D., Dixon, M. J., Cudahy, C., & Merikle, P. M. (2001). Synesthetic photisms influence visual perception. Journal of Cognitive Neuroscience, 13, 930–936. Smilek, D., Dixon, M. J., & Merikle, P. M. (2005). Binding of graphemes and synesthetic colors in color-graphemic synesthesia. In L. C. Robertson & N. Sagiv (Eds.), Synesthesia: Perspectives from cognitive neuroscience (pp. 74–89). Oxford University Press. Wagar, B. M., Dixon, M. J., Smilek, D., & Cudahy, C. (2002). Coloured photisms prevent substitution masking in digit colour synaesthesia. Brain and Cognition, 48, 606–611. Wichmann, F. A., & Hill, N. J. (2001a). The psychometric function: I. Fitting, sampling, and goodness of fit. Perception and Psychophysics, 63(8), 1293–1313. Wichmann, F. A., & Hill, N. J. (2001b). The psychometric function: II. Bootstrap-based confidence intervals and sampling. Perception and Psychophysics, 63(8), 1314–1329.

Consciousness and Cognition 19 (2010) 690–701 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Visuospatial perspective taking in a dynamic environment: Perceiving moving objects from a first-person-perspective induces a disposition to act q H. Kockler a,*, L. Scheef b, R. Tepest a, N. David c, B.H. Bewernick d, A. Newen e, H.H. Schild b, M. May f, K. Vogeley a a Department of Psychiatry, University of Cologne, Germany Department of Radiology, University of Bonn, Germany c Department of Neurophysiology and Pathophysiology, University Medical Center Hamburg-Eppendorf, Germany d Department of Psychiatry, University of Bonn, Germany e Department of Philosophy, Ruhr University Bochum, Germany f Faculty of Mind and Social Sciences, Helmut-Schmidt-University, Hamburg, Germany b a r t i c l e i n f o Article history: Received 21 December 2009 Available online 2 April 2010 Keywords: Perspective taking First-person-perspective Virtual reality Functional magnetic resonance imaging (fMRI) Time-to-collision (TTC) Embodiment Enactment Intraparietal sulcus (IPS) Motion Perception Action a b s t r a c t Spatial perspective taking is an everyday cognitive process that is involved in predicting the outcome of goal directed behavior. We used dynamic virtual stimuli and fMRI to investigate at the neural level whether motion perception interacts with spatial perspective taking in a life-like design. Subjects were asked to perform right-left-decisions about the position of either a motionless, hovering (STATic) or a flying ball (DYNamic), either from their own (1PP) or from the perspective of a virtual character (avatar, 3PP). Our results showed a significant interaction of STIMULUS TYPE and PERSPECTIVE with significantly increased activation in right posterior intraparietal sulcus (IPS) for 1PPDYN condition. As the IPS is critically involved in the computation of object-directed action preparation, we suppose that the simple perception of potentially action-relevant dynamic objects induces a ‘readiness for (re)action’, restricted to the 1PP. Results are discussed against the background of current theories on embodiment and enactive perception. Ó 2010 Elsevier Inc. All rights reserved. 1. Introduction Integrating internal models of the outer world and of one’s own organism for the purpose of orientation and survival in the world is the central function of human self-consciousness and self-understanding. In the framework of Damasio this corresponds to the ‘core consciousness’ the scope of which is the ‘here and now’ (Damasio, 1999, p. 16) and the so-called ‘core self’ as a ‘transient entity, ceaselessly re-created for each and every object with which the brain interacts’ (Damasio, 1999, p. 17). One essential component of self-consciousness is a first-person-perspective, which refers to the experience of one’s own subjective multimodal experiential space centered around one’s own body. The ability of perspective taking enables us not only to live in a coherent self-centered world, but also to simulate how other people experience their environment and to predict their behavior by taking their perspective (Vogeley & Fink, 2003). Neurological and psychiatric syndromes demonstrate that perspective taking as body-centered representation of the self and the world is indispensable. Neglect as disturbance of the self-centered spatial representation of the world and/or the own person leads to a distortion of the q This article is part of a special issue of this journal on Self, Other and Memory. * Corresponding author at: Department of Psychiatry, University of Cologne, Kerpener St. 62, 50924 Cologne, Germany. E-mail address: hanna.kockler@uk-koeln.de (H. Kockler). 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.03.003 H. Kockler et al. / Consciousness and Cognition 19 (2010) 690–701 691 egocentric reference frame so that – in cases of right-sided lesions – the left side of the body and/or the surrounding world is ignored (Fink et al., 2003; Halligan, Fink, Marshall, & Vallar, 2003). Deficits in taking a mental third-person-perspective (theory of mind) are assumed to be associated with autism (Baron-Cohen, 1995; Frith, 2001) and schizophrenia (Doody, Gotz, Johnstone, Frith, & Owens, 1998; Frith & Corcoran, 1996; Vogeley, Kurthen, Falkai, & Maier, 1999). However, to what extent spatial and mental perspective taking rely on similar or different neural mechanisms is currently still a matter of debate (see e.g. Aichhorn, Perner, Kronbichler, Staffen, & Ladurner, 2006; Uddin, Iacoboni, Lange, & Keenan, 2007). Several functional neuroimaging studies have focused on perspective taking in space and were recently reviewed by Zacks and Michelon (2005). In one of their own studies, Zacks and colleagues asked healthy subjects to make right-leftjudgements about line drawings of human bodies (front- or back-facing, upright or inverted) either from their own, or from the figure’s perspective. They found differential activation of the medial prefrontal cortex and cingulate gyrus during 1PP and differential activation of the left parietal–temporal–occipital junction during 3PP (Zacks, Rypma, Gabrieli, Tversky, & Glover, 1999). Vogeley et al. (2004) investigated the neural correlates of spatial perspective taking in a 3D-visuospatial task in which a virtual scene consisting of a virtual character surrounded by red balls was presented. Subjects were asked to assess the number of objects as seen either from their own or from the avatar’s perspective. Results showed a predominantly right sided superior parietal activation during 3PP as opposed to 1PP, while bilateral anterior medial prefrontal, posterior cingulate and superior temporal cortices were associated with 1PP as compared to 3PP. With the exception of a recent study of our own group (David et al., 2006), all fMRI-studies with animated stimulus material examining spatial perspective taking so far employed static stimuli. For the purpose of a systematic study of the influence of real life-like dynamic stimuli on spatial perspective taking, we conducted a study with a two-factorial design in which we systematically varied the factors STIMULUS TYPE (DYN (dynamic stimuli) versus STAT (static stimuli)) and PERSPECTIVE (1PP (first-person-perspective) versus 3PP (third-person-perspective)). Stimulus material showed a virtual character (avatar) in a virtual 3D-scenery. Subjects were asked to decide whether a static, hovering ball was located to the left or right side of their own (1PPSTAT) or the avatar’s visual field (3PPSTAT) or – in the case of dynamic stimuli – whether a flying ball passed onés own (1PPDYN) or the avataŕs left or right side (3PPDYN). This design allowed us to investigate the influence of a dynamic environment on perspective taking as interaction of the factors STIMULUS TYPE and PERSPECTIVE. 2. Methods 2.1. Subjects Eighteen right-handed, healthy male volunteers (age 24.4 ± 2.1 years) without neurological or psychiatric illness, difficulties in handedness-judgements or disturbed vision participated in the study. Informed consent was obtained before participation. Participants were naïve with respect to the experimental task as well as the purpose of the study. Before taking part in the experiment, subjects were instructed and trained via a standardized slide show. The study was approved by the ethics committee of the Medical Faculty of the University of Bonn. 2.2. Stimulus material Fully controlled three-dimensional virtual scenes were generated using the software packages 3D Max (Version 4.0, Discreet, Division of Autodesk, Montreal, Canada) and Poser (Poser5, Curious Labs Inc., Santa Cruz, CA). The basic scene consisted of a quadrangular room with a male avatar positioned in the center of the room. A blue ball was either hovering motionless in the room or flying through the scenery. There were no additional visual cues in the scene like windows, doors, or furniture. The camera view was directed from the frontal aspect of the room onto the opposite wall. This basic scenery was systematically varied with respect to: (1) the orientation of the avatar, (2) the position of the ball in STAT and (3) the ball’s trajectory in DYN. (1) The orientation of the avatar was varied in a clockwise fashion around its vertical axis at angular distances of 30° so that 12 different positions were created. Avatar positions were counterbalanced throughout the whole experiment. Head and body orientation and gaze direction were always congruent. (2) In static stimuli (pictures) the ball was posted in the room at 24 equidistant positions on a circle around the avatar’s position. Only those ball positions inside the avatar’s field of view (covering 150°) were allowed. To avoid ambiguous scenes with respect to 3PP conditions, midline positions and positions at 15° besides midline were excluded. These rules gave rise to 96 different static stimuli. (3) In dynamic stimuli (video animations), ball trajectories were created by defining 12 different points of departure and arrival with a spacing of 30° equidistantly on a circle around the avatar’s position. The ball always flew directly from one of these points to another with a spacing of at least 120° between the point of departure and the point of arrival. A spacing of 180° was excluded to avoid collisions of ball and avatar. From all possible combinations of avatar orientations and ball trajectories, we chose 96, which allowed unambiguous solutions of the left–right task from both perspectives as pilot data showed. Each trajectory was used in both directions with respect to the subject’s position (towards and away from the observer). Both static and dynamic stimuli systematically varied presentation in both hemifields. The height of the flying ball was on the level of the avatar’s head, like in static stimuli, and its velocity was approximately constant within a given video sequence. In order to standardize the length of video sequences, the velocity varied between ±30% between videos due to length differences of the trajectories. To minimize interstimulus differences in velocity the longest and the shortest possible trajectories were not included. Stimuli 692 H. Kockler et al. / Consciousness and Cognition 19 (2010) 690–701 with trajectories that passed the camera too close were excluded as well to avoid a startle response or emotional arousal. In total, 192 stimuli (half of them static, half dynamic) were created. Each stimulus was presented twice per subject. All stimuli had a spatial resolution of 200  400 pixels. The video sequences were shown with a rate of 50 frames per second. Stimulus examples are shown in Fig. 1. 2.3. Tasks and study design We constructed the study in a two-factorial design with the factors PERSPECTIVE and STIMULUS TYPE with four different conditions, namely 1PPSTAT, 1PPDYN, 3PPSTAT, 3PPDYN. Stimuli were presented by Presentation (Version 0.76, Neurobehavioral Systems, Albany, CA) in a block design fashion. For this purpose, a series of 12 stimuli per block was shown on a black background and projected into the scanner environment via a mirror. Before each block of pictures subjects either read the instruction ‘Is the ball on the right or on the left of the AVATAR?’ (3PPSTAT) or ‘Is the ball on the right or on the left side of YOU?’ (1PPSTAT). Before each block of videos either ‘Does the ball pass the AVATAR on his right or on his left side?’ (3PPDYN) or ‘Does the ball pass YOU on your right or on your left side?’ (1PPDYN) was shown. Four experimental runs were presented consisting of 8 blocks with 12 stimuli each in randomized order, systematically varying the four conditions. Each block corresponded to approximately 11 scans per block at a TR of 3.0 s. In advance of each block, a black screen (18 s) and the instruction for the following block (6 s) was shown as a low-level-baseline condition (rest). The first seven scans were discarded later to allow for T1 saturation effects. Each stimulus was shown for approximately 2.7 s. The duration of the stimulus pictures was timed exactly, while the video sequences were presented with some duration uncertainty because of technical limits of the Presentation software. Because of these video timing uncertainties, which accumulated over time, runs differed in their length and lasted between 155 and 159 scans. To obtain an equal length for analyses, the last baseline of each run was cut off after TR 154. The overall duration of the scanning session was approximately 34 min. As dependent variables, reaction times and errors (correctness scores) were recorded using an fMRI compatible response device (Lumitouch, Lightwave Medical Industries, CST Coldswitch Technologies, Richmond, CA). Answers were given by index and ring fingers of the right or left hand in alternating order to avoid differential hemispheric activation due to finger movement. Independent of the answering hand, the left of both fingers indicated the answer ‘left’ and the right finger indicated the answer ‘right’ for the convenience of the test person. 2.4. Statistical analyses of behavioral data Statistical analyses of the behavioral data were performed using SPSS for Windows (Version 11.0). Dependent variables were reaction times (RT) and correctness scores (CS) as percentage of correct answers given within a maximum timeframe (stimulus length). Median values were calculated over each block per subject and then subsequently mean-averaged for each condition across subjects. A repeated-measures ANOVA with the main factors PERSPECTIVE and STIMULUS TYPE as withinsubject factors was performed. The significance level for all analyses was set at p < .05 (two-tailed). 2.5. Functional magnetic resonance imaging Functional MR imaging was performed employing a 3.0 T whole-body imager (Intera; Philips Medical Systems, Best, Netherlands), equipped with a quadrature head coil. Functional images were obtained using a single-shot gradient echo, Fig. 1. Experimental design. Two-factorial design with PERSPECTIVE (third-person-perspective, 3PP, versus first-person-perspective, 1PP) and STIMULUS TYPE (static stimuli, STAT, versus dynamic stimuli, DYN) as main factors. Instructions and stimuli examples are shown. The static stimulus example is shown as used in the experiment, the dynamic stimulus example is a photomontage of single frames extracted from one video sequence used in the study. Stimuli originally had been colored. H. Kockler et al. / Consciousness and Cognition 19 (2010) 690–701 693 echoplanar imaging (EPI) sequence using the following imaging parameters: TR = 3000 ms, TE = 35 ms, flip angle = 90°, FOV 230  230 mm2, voxel size 3.6  3.6  3.6 mm3. The optimal echo time was chosen to match the T2* of gray matter to ensure optimal echo time, compensating for the inherently shorter T2* relaxation times at 3.0 T (Clare, Francis, Morris, & Bowtell, 2001). Each session contained 170 functional images. To aid in localization of activation, a high-resolution T1-weighted magnetization-prepared rapid gradient-echo imaging (MP-RAGE) as 3D MRI image was acquired (TR: 7.7 ms, TE: 3.93 ms, flip angle = 15°, TFE-shot interval 2200 ms, FOV 256  256 mm2, matrix 256  256, voxel size 1  1  1 mm3). 2.6. Image processing and analysis Functional images were preprocessed and statistically analyzed using SPM2 (Statistical Parametric Mapping Software, Wellcome Department of Imaging Neuroscience, London, UK) implemented in MATLAB 6.5 (Mathworks Inc., Sherborn, MA, USA). The preprocessing comprised eye masking in order to avoid artefacts, reorientation according to AC–PC-line, realignment and unwarping to correct for head movement, co-registration of functional and anatomical images, normalization to the MNI stereotactic space, and smoothing with a 10  10  10 mm3 Gaussian kernel. For statistical analyses, the general linear model and a boxcar waveform convolved with the hemodynamic response function were used. Subject-specific, low frequency drifts in signal changes were removed by a high pass filter. Global signal changes were treated as a covariate of no interest. Specific effects were tested by applying appropriate linear contrasts to the parameter estimates for each condition resulting in a t-statistic for each voxel. These t-statistics constitute a statistical parametric map (SPM). The threshold was set at p < .05 throughout the whole study (family-wise error (FWE) corrected) and an extent threshold of 20 contiguous voxels. The following contrast images were calculated for every subject: main effect of 1PP [(1PPDYN plus 1PPSTAT) relative to (3PPDYN plus 3PPSTAT)], main effect of 3PP [(3PPDYN plus 3PPSTAT) relative to (1PPDYN plus 1PPSTAT)], main effect of DYN [(1PPDYN plus 3PPDYN) relative to (1PPSTAT plus 3PPSTAT)], main effect of STAT [(1PPSTAT plus 3PPSTAT) relative to (1PPDYN plus 3PPDYN)]. Additionally, data were analyzed for interactions by calculating the contrasts for the interactions of PERSPECTIVE and STIMULUS TYPE as interaction 1 (IA1) employing the formula [(1PPDYN relative to 1PPSTAT) relative to (3PPDYN relative to 3PPSTAT)] and interaction 2 (IA2) employing the formula [(3PPDYN relative to 3PPSTAT) relative to (1PPDYN relative to 1PPSTAT)]. To determine the main effects of the factors PERSPECTIVE and STIMULUS TYPE and their interactions on the group level, corresponding contrast images were calculated and analyzed as described above applying a random effects model (Friston et al., 1995). Stereotactic MNI coordinates of the voxels of local maximal activation were determined within regions of significant relative signal change associated with the different conditions. The anatomic localization of local maxima and other activated voxels was transformed to Talairach coordinates and assessed by reference to the standard stereotactic atlas (Talairach & Tournoux, 1988). Furthermore, approximations of the corresponding Brodmann areas (BA) are provided on the basis of this atlas. However, it should be kept in mind that Talairach’s and Tournoux’ version of the Brodmann map does not reflect the complete and up-to-date cytoarchitectonic organization of the human brain. Nevertheless Brodmann areas are given for the reader’s convenience and for comparability to other studies. Figs. 2 and 3 provide the SPM {t}-maps of the significant main effects on the group level as overlay images onto one normalized 3D data set (p < .05, FWE corrected, extent threshold = 20 voxels). Fig. 2. Neural correlate of IA1. Significantly activated clusters are superimposed on a single, normalized T1 data (p < .05, FWE corrected, extent threshold = 20 voxels). The statistical interaction IA1 of the main factors PERSPECTIVE and STIMULUS TYPE, mainly reflecting the differential effect of 1PPDYN > 1PPSTAT, is associated with increased neural activity in the right posterior intraparietal sulcus (IPS). Data in Figs. 2 and 3 correspond to those in Table 2. Fig. 2 is shown in neurological orientation (right = right, left = left in coronal sections). 694 H. Kockler et al. / Consciousness and Cognition 19 (2010) 690–701 Fig. 3. Neural correlates of 3PP > 1PP (a), and DYN > STAT (b). Significantly activated clusters are superimposed on a single, normalized T1 data (p < .05, FWE corrected, extent threshold = 20 voxels). Data in Figs. 2 and 3 correspond to those in Table 2. (a) Taking a third-person-perspective (3PP > 1PP) is associated with increased neural activity in the pre-SMA bilaterally, the posterior parietal cortex bilaterally with predominance of the right hemisphere, inferior occipital cortex (EBA), cerebellum, subregions of the dorsolateral prefrontal cortex and the anterior insula bilaterally. (b) During the dynamic condition (DYN > STAT), differential activations in occipitotemporal (V5/MT+) cortex, superior frontal areas and superior and medial parietal (precuneus) cortex bilaterally can be observed. In order to characterize the differential influences of the different conditions in IA1 more precisely, the areas activated in IA1 were defined as regions-of-interest (ROI) and additional analyses were performed in these ROIs: 1PPDYN > 1PPSTAT; 1PPSTAT > 1PPDYN; 3PPDYN > 3PPSTAT; 3PPSTAT > 3PPDYN. For ROI-analyses, the WKU PickAtlas (www.fmri.wfubmc.edu/download.htm) and Marsbar (www.marsbar.sourceforge.net) were used. Furthermore, the BOLD signal changes at the principally activated voxels of the significant key contrasts were taken from raw data of every participant and their means were calculated for each condition as well as their standard errors (Fig. 4). 3. Results 3.1. Behavioral data Overall means and standard deviations (SD) of the dependent variables are reported in Table 1. Both the ANOVA on reaction times (RT) and the ANOVA on correctness scores (CS) revealed a significant effect of PERSPECTIVE (for RT: Fig. 4. Mean percent BOLD signal changes compared to baseline for IA1 (a), 3PP (b) and DYN (c) at the respective principally activated voxel. The signal changes at these voxels were taken from raw data and plotted over time. Means were calculated for each condition and show the differential increase in signal over the four conditions. The mean standard errors are given as error bars. Voxel coordinates are provided according to Talairach and Tournoux (1988). (a) Principally activated voxel in IA1: x = 14, y = 74, z = 56 (MNI coordinates). As the figure shows, 3PPDYN (entering negatively into the contrast) and 3PPSTAT (entering positively into IA1) approximately neutralize their influence in contrast IA1 ((3PPSTAT – 3PPDYN) plus (1PPDYN – 1PPSTAT)). Therefore, IA1 is mainly due to the difference of 1PPDYN versus 1PPSTAT. Please note that the relatively high activation of the IPS in 3PP conditions reflects the mental rotation operation that is necessary to solve the task (see Section 4 of main effects). (b) Principally activated voxel in 3PP: x = 4, y = 20, z = 46. (c) Principally activated voxel in DYN: x = 46, y = 64, z = 4. Table 1 Behavioral data. Mean values and standard deviations (SD) of the dependent variables (reaction times and correctness scores) are summarized for the main conditions. 3PPDYN Reaction time (msec) Correctness score (%) 3PPSTAT 1PPDYN 1PPSTAT Mean SD Mean SD Mean SD Mean SD 1565.3 95.0 141.3 5.6 989.2 87.9 173.3 5.3 1230.2 98.3 128.5 1.7 552.7 99.7 118.6 0.6 H. Kockler et al. / Consciousness and Cognition 19 (2010) 690–701 695 F(1,17) = 205.19; p < .001; for CS: F(1,17) = 50.73; p < .001) and STIMULUS TYPE (for RT: F(1,17) = 828.04; p < .001; for CS: F(1,17) = 19.28; p < .001) as well as a significant interaction PERSPECTIVE  STIMULUS TYPE (for RT: F(1,17) = 18.36; p 6 .001; for CS: F(1,17) = 32.61; p < .001). Whereas the difference of RT between DYN and STAT was due to the latency with which objects in DYN had to pass until an adequate decision could be made, the RT difference in PERSPECTIVE most likely reflected task difficulty. Correctness scores higher than 75% assured correct task solving and allowed adequate interpretation of fMRI data. In the 3PPDYN-condition, reaction times (means of individual medians ± standard deviation) show a dependence on the degree of rotation of the avatar (Fig. 5). 3.2. Neural correlates of perspective, stimulus type and their interaction Differential neural activity as correlate of IA1 [(1PPDYN relative to 1PPSTAT) relative to (3PPDYN relative to 3PPSTAT)] was located in the right posterior intraparietal sulcus (IPS) and related superior parietal areas (Table 2a and Fig. 2). The neural correlate of the main effect of 3PP was associated with increased neural activity in the pre-supplementary motor area (pre-SMA) bilaterally, the posterior parietal cortex bilaterally predominantly on the right side, cerebellum, subregions of the dorsolateral prefrontal cortex bilaterally and the anterior insula bilaterally (Table 2b and Fig. 3a). Areas of significantly increased neural activity associated with the main effect of DYN were located in the occipitotemporal (approximately V5/ MT+), superior frontal and superior/medial parietal cortex (precuneus) bilaterally (Table 2c and Fig. 3b). With regard to the second interaction IA2 as well as to the main effects for 1PP and STAT, no differential neural activity was found. These findings are further illustrated by plots of the mean BOLD signal changes (Fig. 4). It can be demonstrated that for the main effects 3PP and DYN, the mean BOLD signal changes at the principally activated voxel (3PP: 4, 20, 46; DYN: 46, 64, 4) correspond to the expected values, as each of the relevant conditions that contributes to a particular main effect shows an adequate and comparable increase in signal change (Fig. 4b and c). The significant BOLD increase in contrast IA1 [(1PPDYN relative to 1PPSTAT) relative to (3PPDYN relative 3PPSTAT)] at the principally activated voxel (14, 74, 56) cannot be related to any relevant difference between 3PPSTAT and 3PPDYN and is thus due to the difference between 1PPDYN and 1PPSTAT (Fig. 4a). This effect of the interaction IA1 was further corroborated by subsequent region-of-interest (ROI) analyses as the ROI corresponding to the clusters activated during IA1 were also significantly activated in 1PPDYN > 1PPSTAT, but neither in 3PPDYN > 3PPSTAT nor in 3PPSTAT > 3PPDYN (p < .05; FWE corrected; not illustrated). These confirmatory analyses allow to conclude that the neural correlate of IA1 is essentially constituted by the differential effect of 1PPDYN relative to 1PPSTAT although the overall activation of the IPS during 3PP-conditions compared to baseline is at least as high as during 1PPDYN. 4. Discussion The main experimental question of this study was to examine the influence of life-like dynamic stimuli on spatial perspective taking as a necessary prerequisite of human self-consciousness and the ‘core self’ (Damasio, 1999). Behavioral data Fig. 5. Reaction times of 3PPDYN-condition, dependent on the position of the avatar. Reaction times show the typical pattern of a mental rotation operation. 696 H. Kockler et al. / Consciousness and Cognition 19 (2010) 690–701 Table 2 Significant clusters and their peak coordinates for the neural correlates of IA1, 3PP, and DYN. Volume statistics of the main effects and the interaction between the experimental factors (p < .05, FWE corrected, extent threshold = 20 voxels). Voxel coordinates are provided according to Talairach and Tournoux (1988). BA = Brodmann’s area, L = left hemisphere, R = right hemisphere. Region (a) IA1 Superior parietal R IPS R a b c Cluster size (voxels) MNI-coordinates (x, y, z) t-value BA 28 39 14 34 74 82 56 30 8.53 8.20 7 (b) 3PP (3PP > 1PP) Medial frontal R/La Superior parietal R Cerebellum L Inferior occipital Rb Inferior/middle front R Superior frontal L IPS/superior parietal L Cerebellum L Midbrain Anterior insula R Superior frontal R Anterior insula L Inferior/middle front L Middle occipital L 271 1596 416 891 543 238 1080 587 452 150 312 97 94 57 4 40 36 44 50 26 36 12 2 34 28 32 42 46 20 52 64 66 12 4 38 78 34 28 2 22 4 68 46 62 28 10 30 58 40 28 4 6 48 2 26 4 10.44 10.43 10.16 9.72 9.38 9.28 9.27 8.76 8.41 8.29 8.15 8.09 7.64 7.37 8 40 (c) DYN (DYN > STAT) Middle temporal Lc Middle temporal Rc/superior occipital Rb Superior/middle front L Superior parietal L/R/precuneus L/R Middle frontal R Postcentral L 646 883 1033 2259 272 34 46 48 26 34 28 60 64 66 4 42 0 18 4 4 56 60 62 36 10.11 9.91 9.12 9.05 7.13 7.04 37/19 44 6 40 47 6 47 9 37 37 37 6 6 43 Corresponding to pre-SMA. Corresponding to extrastriate body area (EBA). Corresponding to V5/MT+. as well as extensive statistical analyses of fMRI data revealed a significant interaction between the experimental factors PERSPECTIVE and STIMULUS TYPE with differential activation of the right posterior intraparietal sulcus (IPS) due to an increase in neural activation during 1PPDYN relative to 1PPSTAT. 4.1. The interaction effect While it is not surprising that the right IPS is highly activated in 3PP-states due to its well documented involvement in mental rotation tasks (see Section 4.2), the additional IPS-activation in 1PPDYN compared to 1PPSTAT is a new and unexpected finding. The IPS until now has neither been reported as a primary motion-processing area nor as a neural correlate of a first-person-perspective as such (Vogeley et al., 2004). Instead, the majority of studies on the IPS demonstrated a role in visually guided actions that are directed to objects (Goodale & Milner, 1992; Rizzolatti & Matelli, 2003). Anatomy and function of the IPS has been extensively studied in macaques (for reviews see Colby and Goldberg (1999) and Grefkes and Fink (2005)), these studies provide evidence for a central role in action organization, such as reaching and grasping and eye-armcoordination (Battaglia-Mayer, Caminiti, Lacquaniti, & Zago, 2003; Burnod et al., 1999; Gregoriou & Savaki, 2001). During the last years, a number of studies in humans demonstrated that the IPS of humans and macaques share a similar functional organization (Astafiev et al., 2003; Grefkes & Fink, 2005; Orban et al., 2006; Shikata et al., 2007). Furthermore, there is now converging evidence on the IPS’ function that can be inferred from clinical data, virtual lesion studies (transcranial magnetic stimulation, TMS) and functional neuroimaging. Lesions of the right IPS and adjacent superior parietal cortex result in a neurological syndrome called ‘optic ataxia’: patients show an impairment in the visual control of target-directed arm movements (Battaglia-Mayer & Caminiti, 2002; Perenin & Vighetto, 1988; Weiss et al., 2006). Functional MRI studies in humans underlined this clinical finding: The IPS shows increased BOLD signal during visually guided, objectdirected action, especially during reaching (Kertzman, Schwarz, Zeffiro, & Hallett, 1997; Levy, Schluppeck, Heeger, & Glimcher, 2007), grasping (Begliomini, Wall, Smith, & Castiello, 2007; Frey, Vinton, Norlund, & Grafton, 2005), attending and pointing to a (peripheral) visual location (Astafiev et al., 2003). Strongest evidence is provided by studies using transcranial magnetic stimulation (TMS). Transient lesions of the left posterior IPS impaired the adjustment of reaching movements to a moving environment (Della-Maggiore, Malfait, Ostry, & Paus, 2004), whereas TMS-induced transient lesions to the anterior part of the IPS caused a disturbance of the on-line control of reach-to-grasp movements (Tunik, Frey, & Grafton, 2005). Anatomically, extensive input from visual areas such as the temporo-parietal motion-responsive brain region V5/MT+ and projections from the IPS to ventral prefrontal and primary motor areas (Grefkes & Fink, 2005; Sakata, Taira, Kusunoki, H. Kockler et al. / Consciousness and Cognition 19 (2010) 690–701 697 Murata, & Tanaka, 1997) further support that the IPS is crucial for the integration and coordination of external peripersonal visual information on the one hand and internal representations of action schemata and reference frames on the other hand which are both essential components of object-directed action. Andersen and Buneo (Andersen, Snyder, Bradley, & Xing, 1997; Buneo & Andersen, 2006) as well as Rizzolatti and Matelli (2003) proposed that one of the major functions of the IPS and IPS-frontal circuits is the representation of object positions in an effector-centered, thus egocentric reference frame in order to allow for motor descriptions of stimulus positions. In humans, it could be shown that the visuomotor coordinate transformation subserving goal-directed hand movements activates the superior parietal cortex with a maximum in the medial wall of the IPS (Andersen et al., 1997; Grefkes, Ritzl, Zilles, & Fink, 2004). In short, the function of the IPS (as part of the dorsal stream of vision) is concisely described by the term ‘vision for action’ coined by Goodale (1998). The present finding that the IPS is specifically involved during 1PPDYN importantly extends the previous evidence of its involvement in a variety of object-directed action-processing operations and enriches the concepts on its functions by suggesting that already the mere visual perception of dynamic objects from a first-person-perspective leads to its recruitment. We propose that this reflects an automatic, prereflexive disposition to act or – in our context – an early´ readiness to react́ that is independent of any conscious intention to act. This ‘readiness for (re)action’ is activated even if an action is neither explicitly required nor actively performed (as in our study) and best understood as an unconscious state of ‘being prepared’ to react to the object if it should become necessary. This is plausible from an evolutionary point of view: Moving objects are more action-relevant than static ones due to their potential live-threatening character and thus can require very fast modifications of behavior. Our own finding is underlined by direct evidence from monkey studies that showed that an IPS subregion, the ventral intraparietal area (VIP), contributes to collision avoidance and defensive behavior (Graziano & Cooke, 2006). In a similar vein, Cooke, Taylor, Moore, and Graziano (2003) evoked complex defensive facial expressions by microstimulation of the VIP in the macaque. In direct relevance to the present study, Field and Wann (2005) recently investigated the neural correlates of a so-called ‘time-to-collision (TTC)’ mechanism in humans. TTC refers to the estimation of the remaining time until an approaching object reaches a target (here: the participating subjects of the study). During a collisional condition compared with control conditions of similar visual quality that did not induce the impression of an object moving in depth, increased neural activity in sensorimotor areas were observed that are target of the dorsal visual system. Similar to our study, this activation could not be attributed to any actual movement or any intention to act, supporting the idea of an automatically activated disposition to react induced by the mere visual perception of a potentially threatening dynamic object. Thus, subjects are prepared for fast modifications of behavior that are required in dynamic situations, e.g. defensive responses in order to avoid collisions with objects. Interestingly, recent studies provide evidence that the ability to detect and predict a collision seems to be a function of perceptuomotor brain development and to depend on perceptuomotor experiences over lifetime (Lobjois, Benguigui, Bertsch, & Broderick, 2008; Van der Weel & van der Meer, 2009). Extending these previous results, our findings implicate that the ‘readiness for (re)action’ is stronger in 1PP-states than in 3PP-states: If the ‘readiness for reaction’ would be independent of perspective, we should expect a higher IPS-activation in 3PPDYN than in 3PPSTAT corresponding to the higher activation in 1PPDYN than in 1PPSTAT. This is not the case. We can therefore presume that the disposition to act is restricted to 1PP-states which is intuitively plausible because one’s own disposition to act can have no functional role for someone else. An anticipatory coding of objects in terms of motor descriptions is only ecologically useful if it implies a behavioral advantage. This advantage of a readiness for actions does not exist in a third-person-perspective because we simply cannot act for another person. So, the restriction to one’s own perspective as well shows that the disposition to act depends on an action-relevant context – an inference that is further corroborated by studies of the neural computation of near versus far space. Processing of objects that are located in ‘near space’ defined as the peripersonal space within arm’s reach differentially activate the dorsal stream of vision including the IPS while processing of objects that are presented in far space show a stronger activation of the ventral stream of vision (Weiss, Marshall, Zilles, & Fink, 2003). Correspondingly, in monkeys, extrapersonal space within reaching distance is represented by two subpopulations of neurons within the IPS, the medial intraparietal area (MIP, ‘near space’) and the ventral intraparietal area (VIP, ‘ultra-near space’), respectively (Colby & Duhamel, 1991). These findings as well as our results may lead to the hypothesis that the visual perception of objects in events in general recruits action-relevant areas as soon as the perception of the object becomes action-relevant, in the sense of enabling a disposition to act in the immediate body-related environment. This is an observation of theoretical importance for current concepts on ‘embodiment’ and ‘enactive perception’: first of all, the IPSactivation in our experiment is indicative of what is meant by the ‘embodiment’ of the mind (Gallagher, 2005): in the perception of events certain features of objects can be strongly associated with motor activations in the brain reflecting a readiness to act on a behavioral level. Secondly, however, from these ‘enactive’ perceptual states, we can distinguish other perceptual states that are not – or at least significantly less – associated with this disposition to act. In our experiment, the question if the perception of an event leads to this neural response in cortical areas associated with action preparation or not critically depends on the subject’s perspective while the visual input is the same from both perspectives. Our finding thus leads us to the final assumption that the degree of enactment in the perception of events depends on the context, more precisely on the question if an event is action-relevant for a subject of not. Thus, the general claim that the perception of objects is equivalent to the preparation of object-directed actions (O’Regan & Noë, 2001; Noë, 2004) must be rejected. Reflecting critically our study, there are also alternative options to interpret our findings. Unfortunately, we could not obtain measurements of eye movements. As the posterior IPS is known to be involved in the generation of saccades we cannot completely rule out that the effect could be confounded by task dependent eye movements that could have been processed in the IPS. However, this appears unplausible to us based on the following considerations. First, eye movement related 698 H. Kockler et al. / Consciousness and Cognition 19 (2010) 690–701 parietal areas of each hemisphere topographically respond to targets for pro- and antisaccades in the contralateral hemifield (Medendorp, Goltz, & Vilis, 2005; Sereno, Pitzalis, & Martinez, 2001). In our study, each stimulus, both dynamic and static, was shown twice, the second instance right–left-mirrored compared to the first instance to avoid laterality effects. If the observed activity was due to saccadic eye movements, one had to assume a bilateral activation and the asymmetric, isolated activity in right parietal regions during 1PPDYN could not be explained. Second, if different eye movements during conditions would cause the intraparietal activation in our interaction, we would at least expect an additional activation of the frontal eye fields (FEF) because (at least in monkeys) the frontal eye fields as motor areas are even more strongly associated with the execution of eye movements than the parietal eye regions (Kusunoki et al., 2000). Third, in the experiment that served as basis for the current study and that employed similar stimulus material, Vogeley et al. (2004) showed that there were no differences in eye movements between 3PP and 1PP. Attentional effects can also be discussed as potential confounding mechanisms that recruit the IPS. Culham et al. (1998) showed in an fMRI study that attentive tracking of moving objects (versus passive viewing) without eye movement leads to increased activity in different regions including the IPS. In our opinion, such an effect of covert attention would not reject our interpretation of a ‘readiness for (re)action’ as attentive tracking of moving objects could also be understood as the constitutive cognitive mechanism of this disposition to act that is induced by moving objects: The higher the relevance of these objects is, the more attention will be allocated towards them. In this sense, recent electrophysiological studies in monkeys raised the idea that the role of the IPS in spatial attention consists in providing spatial ‘salience maps’ of objects by both topdown and bottom-up modulated coding of behaviorally relevant objects in terms of spatial coordinates of attention targets (see Goldberg, Bisley, Powell, and Gottlieb (2006), for a review). Recapitulating, Rizzolatti and Matelli (2003) precisely described the function that we suppose to be reflected by the neural correlate of the interaction in our experiment: ‘‘When a visual stimulus is presented in the peripersonal space, it evokes automatically a ‘potential motor action’, which, regardless of whether the action is subsequently executed or not, maps the spatial stimulus position in motor terms” (Rizzolatti & Matelli, 2003; similar: Culham & Valyear, 2006). For the fist time, our results provide evidence that potential motor actions are triggered by dynamic stimuli more than by static stimuli and probably are restricted to 1PP-states. These findings possibly reflect a more general mechanism that induces an early and prereflective disposition to act – neurally represented by the recruitment of action-relevant areas – in the presence of objects with an action-relevant character. 4.2. Main effect of perspective: 3PP The differential activation during the main effect of 3PP was mainly located in the predominantly right-sided posterior parietal cortex (PPC), the pre-supplementary motor area (pre-SMA) bilaterally and dorsolateral prefrontal areas. These findings in essence replicate the results of Vogeley and colleagues (2004) and show exceeding overlaps with the pattern that was found in a recent meta-analysis of 32 mental rotation studies (Zacks, 2008), suggesting that the process of taking the spatial perspective of another person is more closely related to mental rotation operations than to mental perspective taking tasks (theory of mind) on a neural and cognitive level. This conclusion is also supported by reaction time patterns that show a dependence of 3PP-reaction times on the degree of rotation of the avatar that is typical for mental rotation studies (Fig. 5). Accordingly, Harris et al. (2000) found that the right superior parietal cortex centered on the IPS and extending down into the superior posterior occipital cortex shows a correlation between increase in activation and increase in the amount of mental rotation during parametric modulation of the degree of rotation. The high percent signal change in the right IPS during 3PP-conditions thus reflects the mental rotation computations that are necessary to solve the task. Interestingly – although not based on mental rotation operations – various studies also found the right (superior, inferior or medial) parietal cortex being involved in distinguishing self from other (e.g. Decety & Sommerville, 2003; Feinberg & Keenan, 2005; Feinberg & Roane, 2005; Ruby & Decety, 2004; Seger, Stone, & Keenan, 2004; Stuss & Anderson, 2004; Uddin, Molnar-Szakacs, Zaidel, & Iacoboni, 2006). Additionally to the regions involved in mental rotation operations reported by Zacks, we found one region in the right inferior occipital region. Although we did not use an appropriate localizer, we assume that this region corresponds to the extrastriate body are (EBA) which is involved in the perception of human bodies and body parts (David et al., 2007; Downing, Jiang, Shuman, & Kanwisher, 2001). Right EBA especially computes bodies that are shown from an allocentric perspective – a finding that has been interpreted as the neural correlate of perceiving others (Saxe, Jamal, & Powell, 2005). 4.3. Main effect of stimulus type: DYN Activation patterns during the main effect of DYN were also in accordance with previous evidence. For example, the principally activated region during DYN was the visual motion area V5/MT+ (Born & Bradley, 2005; Watson et al., 1993; Wilms et al., 2005). The precuneus, also found here, has been associated with visual imagery, supporting our capacity to – metaphorically speaking – ‘inspection of internal images’ (Cavanna & Trimble, 2006; Ogiso, Kobayashi, & Sugishita, 2000), or ‘the mind’s eye’ (Fletcher et al., 1995), and the spatial localization of objects (Rao, Zhou, Zhuo, Fan, & Chen, 2003). According to these results, one could speculate that the precuneus possibly performs an imagined visual anticipation of the ball’s motion by generating an estimated ‘internal image’ of the ball’s trajectory. A similar function has been ascribed to the lateral premotor cortex that is also activated during the dynamic condition. Mental anticipation of the location of moving objects H. Kockler et al. / Consciousness and Cognition 19 (2010) 690–701 699 increased activation of the dorsolateral premotor cortex as one part of a so-called ‘prehension network’ associated with object-related attention (Schubotz & von Cramon, 2001). 4.4. Other main effects: 1PP and STAT With respect to the other main effects, neither 1PP nor STAT reached the level of statistical significance applied in the present study for all other conditions. While we did not expect a certain differential activation for STAT, we had strong hypotheses for 1PP. Similar experiments of our group contrasting 1PP versus 3PP consistently showed a characteristic pattern of increased activation comprising medial prefrontal, medial parietal, inferior parietal and temporoparietal cortex bilaterally and for the present study we expected a similar activation pattern, accordingly. However, all these studies used exclusively static stimuli. As our results show, there is a major difference in 1PP if stimuli are dynamic. We therefore assume that his dissociation of 1PP explains the missing statistical power of a common 1PP [(1PPDYN plus 1PPSTAT) relative to (3PPDYN plus 3PPSTAT)]. 5. Conclusion Studying the influence of dynamic stimuli on spatial perspective taking revealed a significant increase in neural activation during motion perception from a first-person-perspective in the right posterior intraparietal sulcus (IPS). Our results indicate that the neural key process in the perceptual situation examined is related to the computation of object-directed action preparation. They suggest that potentially threatening and thus action-relevant, moving objects induce a state of ‘oneself in readiness for (re)action’. Acknowledgment We thank Jeffrey Zacks for very helpful comments on a previous version of the manuscript. M. M., A. N. and K. V. were supported by the German VolkswagenFoundation. References Aichhorn, M., Perner, J., Kronbichler, M., Staffen, W., & Ladurner, G. (2006). Do visual perspective tasks need theory of mind? NeuroImage, 30(3), 1059–1068. Andersen, R. A., Snyder, L. H., Bradley, D. C., & Xing, J. (1997). Multimodal representation of space in the posterior parietal cortex and its use in planning movements. Annual Review of Neuroscience, 20, 303–330. Astafiev, S. V., Shulman, G. L., Stanley, C. M., Snyder, A. Z., Van Essen, D. C., & Corbetta, M. (2003). Functional organization of human intraparietal and frontal cortex for attending, looking, and pointing. Journal of Neuroscience, 23(11), 4689–4699. Baron-Cohen, S. (1995). Mindblindness: An essay on autism and theory of mind. Boston: MIT Press/Bradford Books. Battaglia-Mayer, A., & Caminiti, R. (2002). Optic ataxia as a result of the breakdown of the global tuning fields of parietal neurones. Brain, 125(2), 225–237. Battaglia-Mayer, A., Caminiti, R., Lacquaniti, F., & Zago, M. (2003). Multiple levels of representation of reaching in the parieto-frontal network. Cerebral Cortex, 13(10), 1009–1022. Begliomini, C., Wall, M. B., Smith, A. T., & Castiello, U. (2007). Differential cortical activity for precision and whole-hand visually guided grasping in humans. European Journal of Neuroscience, 25(4), 1245–1252. Born, R. T., & Bradley, D. C. (2005). Structure and function of visual area MT. Annual Review of Neuroscience, 28, 157–189. Buneo, C. A., & Andersen, R. A. (2006). The posterior parietal cortex: Sensimotor interface for the planning and online control of visually guided movements. Neuropsychologia, 44(13), 2594–2606. Burnod, Y., Baraduc, P., Battaglia-Mayer, A., Guigon, E., Koechlin, E., Ferraina, S., et al (1999). Parieto-frontal coding of reaching: An integrated framework. Experimental Brain Research, 129(3), 325–346. Cavanna, A. E., & Trimble, M. R. (2006). The precuneus: A review of its functional anatomy and behavioral correlates. Brain, 129(3), 564–583. Clare, S., Francis, S., Morris, P. G., & Bowtell, R. (2001). Single-shot T2* measurement to establish optimum echo time for fMRI: Studies of the visual, motor, and auditory cortices at 3.0 T. Magnetic Resonance in Medicine, 45, 930–933. Colby, C. L., & Duhamel, J. R. (1991). Heterogeneity of extrastriate visual areas and multiple parietal areas in the macaque monkey. Neuropsychologia, 29(6), 517–537. Colby, C. L., & Goldberg, M. E. (1999). Space and attention in parietal cortex. Annual Review of Neuroscience, 22, 319–349. Cooke, D. F., Taylor, C. S., Moore, T., & Graziano, M. S. (2003). Complex movements evoked by microstimulation of the ventral intraparietal area. Proceedings of the National Academy of Sciences, USA, 100(10), 6163–6168. Culham, J. C., Brandt, S. A., Cavanagh, P., Kanwisher, N. G., Dale, A. M., & Tootell, R. B. (1998). Cortical fMRI activation produced by attentive tracking of moving targets. Journal of Neurophysiologie, 80(5), 2657–2670. Culham, J. C., & Valyear, K. F. (2006). Human parietal cortex in action. Current Opinion in Neurobiologie, 16(2), 205–212. David, N., Bewernick, B. H., Cohen, M. X., Newen, A., Lux, S., Fink, G. R., et al (2006). Neural representations of self versus other: Visual-spatial perspective taking and agency in a virtual ball-tossing game. Journal of Cognitive Neuroscience, 18(6), 898–910. David, N., Cohen, M. X., Newen, A., Bewernick, B. H., Shah, N. J., Fink, G. R., et al (2007). The extrastriate cortex distinguishes between the consequences of one’s own and others’ behavior. NeuroImage, 36(3), 1004–1014. Damasio, A. (1999). The feeling of what happens. Body and emotion in the making of consciousness. Harcourt, Inc. Decety, J., & Sommerville, J. A. (2003). Shared representations between self and other: A social cognitive neuroscience view. Trends in Cognitive Sciences, 7(12), 527–533. Della-Maggiore, V., Malfait, N., Ostry, D. J., & Paus, T. (2004). Stimulation of the posterior parietal cortex interferes with arm trajectory adjustments during the learning of new dynamics. Journal of Neuroscience, 24(44), 9971–9976. Doody, G. A., Gotz, M., Johnstone, E. C., Frith, C. D., & Owens, D. G. (1998). Theory of mind and psychoses. Psychological Medicine, 28(2), 397–405. Downing, P. E., Jiang, Y., Shuman, M., & Kanwisher, N. (2001). A cortical area selective for visual processing of the human body. Science, 293(5539), 2470–2473. Feinberg, T. E., & Keenan, J. P. (2005). Where in the brain is the self? Consciousness and Cognition, 14(4), 661–678. Feinberg, T. E., & Roane, D. M. (2005). Delusional misidentification. The Psychiatric Clinics of North America, 28(3), 665–683 [678–679]. 700 H. Kockler et al. / Consciousness and Cognition 19 (2010) 690–701 Field, D. T., & Wann, J. P. (2005). Perceiving time to collision activates the sensorimotor cortex. Current Biology, 15(5), 453–458. Fink, G. R., Marshall, J. C., Weiss, P. H., Stephan, T., Grefkes, C., Shah, N. J., et al (2003). Performing allocentric visuospatial judgments with induced distortion of the egocentric reference frame: An fMRI study with clinical implications. NeuroImage, 20, 1505–1517. Fletcher, P. C., Frith, C. D., Baker, S. C., Shallice, T., Frackowiak, R. S., & Dolan, R. J. (1995). The mind’s eye – Precuneus activation in memory-related imagery. NeuroImage, 2(3), 195–200. Frey, S. H., Vinton, D., Norlund, R., & Grafton, S. T. (2005). Cortical topography of human anterior intraparietal cortex active during visually guided grasping. Brain Research. Cognitive Brain Research, 23(2–3), 397–405. Friston, K. J., Homes, A. P., Worsley, K. J., Poline, J. P., Frith, C. D., & Frackowiak, R. S. (1995). Statistical parametric maps in functional imaging: A general linear approach. Human Brain Mapping, 2, 189–210. Frith, C. D., & Corcoran, R. (1996). Exploring ‘theory of mind’ in people with schizophrenia. Psychological Medicine, 26(3), 521–530. Frith, U. (2001). Mind blindness and the brain in autism. Neuron, 32(6), 969–979. Gallagher, S. (2005). How the body shapes the mind. Oxford University Press. Goldberg, M. E., Bisley, J. W., Powell, K. D., & Gottlieb, J. (2006). Saccades, salience and attention: The role of the lateral intraparietal area in visual behavior. Progress in Brain Research, 155, 157–175. Goodale, M. A., & Milner, A. D. (1992). Separate visual pathways for perception and action. Trends in Neurosciences, 15(1), 20–25. Goodale, M. A. (1998). Vision for perception and vision for action in the primate brain. Novartis Foundation Symposium, 218, 21–34. Graziano, M. S., & Cooke, D. F. (2006). Parieto-frontal interactions, personal space, and defensive behavior. Neuropsychologia, 44(6), 845–859. Grefkes, C., Ritzl, A., Zilles, K., & Fink, G. R. (2004). Human medial intraparietal cortex subserves visuomotor coordinate transformation. NeuroImage, 23(4), 1494–1506. Grefkes, C., & Fink, G. R. (2005). The functional organization of the intraparietal sulcus in humans and monkeys. Journal of Anatomy, 207(1), 3–17. Gregoriou, G. G., & Savaki, H. E. (2001). The intraparietal cortex: Subregions involved in fixation, saccades, and in the visual and somatosensory guidance of reaching. Journal of Cerebral Blood Flow and Metabolism, 21(6), 671–682. Halligan, P. W., Fink, G. R., Marshall, J. C., & Vallar, G. (2003). Spatial cognition: Evidence from visual neglect. Trends in Cognitive Sciences, 7(3), 125–133. Harris, I. M., Egan, G. F., Sonkkila, C., Tochon-Danguy, H. J., Paxinos, G., & Watson, J. D. (2000). Selective right parietal lobe activation during mental rotation: A parametric PET study. Brain, 123(1), 65–73. Kertzman, C., Schwarz, U., Zeffiro, T. A., & Hallett, M. (1997). The role of posterior parietal cortex in visually guided reaching movements in humans. Experimental Brain Research, 114(1), 170–183. Kusunoki, M., Gottlieb, J., & Goldberg, M. E. (2000). The lateral intraparietal area as a salience map: The representation of abrupt onset, stimulus motion, and task relevance. Vision Research, 40(10–12), 1459–1468. Levy, I., Schluppeck, D., Heeger, D. J., & Glimcher, P. W. (2007). Specificity of human cortical areas for reaches and saccades. Journal of Neuroscience, 27(17), 4687–4696. Lobjois, R., Benguigui, N., Bertsch, J., & Broderick, M. P. (2008). Collision avoidance behavior as a function of aging and tennis playing. Experimental Brain Research, 184, 457–468. Medendorp, W. P., Goltz, H. C., & Vilis, T. (2005). Remapping the remembered target location for anti-saccades in human posterior parietal cortex. Journal of Neurophysiology, 94(1), 734–740. Noë, A. (2004). Action in perception. Cambridge, MA: The MIT Press. Ogiso, T., Kobayashi, K., & Sugishita, M. (2000). The precuneus in motor imagery: A magnetoencephalographic study. NeuroReport, 11(6), 1345–1349. Orban, G. A., Claeys, K., Nelissen, K., Smans, R., Sunaert, S., Todd, J. T., et al (2006). Mapping the parietal cortex of human and non-human primates. Neuropsychologia, 44(13), 2647–2667. O’Regan, J. K., & Noë, A. (2001). A sensorimotor account of vision and visual consciousness. Behavioral and Brain Sciences, 24, 939–1031. Perenin, M. T., & Vighetto, A. (1988). Optic ataxia: A specific disruption in visuomotor mechanisms. I. Different aspects of the deficit in reaching for objects. Brain, 111(3), 643–674. Rao, H., Zhou, T., Zhuo, Y., Fan, S., & Chen, L. (2003). Spatiotemporal activation of the two visual pathways in form discrimination and spatial location: A brain mapping study. Human Brain Mapping, 18(2), 79–89. Rizzolatti, G., & Matelli, M. (2003). Two different streams form the dorsal visual system: Anatomy and functions. Experimental Brain Research, 153(2), 146–157. Ruby, P., & Decety, J. (2004). How would you feel versus how do you think she would feel? A neuroimaging study of perspective-taking with social emotions. Journal of Cognitive Neuroscience, 16(6), 988–999. Sakata, H., Taira, M., Kusunoki, M., Murata, A., & Tanaka, Y. (1997). The TINS lecture. The parietal association cortex in depth perception and visual control of hand action. Trends in Neuroscience, 20(8), 350–357. Saxe, R. I., Jamal, N., & Powell, L. (2005). My body or yours? The effect of visual perspective on cortical body representations. Cerebral Cortex, 16(2), 178–182. Schubotz, R. I., & von Cramon, D. Y. (2001). Functional organization of the lateral premotor cortex: fMRI reveals different regions activated by anticipation of object properties, location and speed. Brain Research. Cognitive Brain Research, 11(1), 97–112. Seger, C. A., Stone, M., & Keenan, J. P. (2004). Cortical activations during judgements about the self and an other person. Neuropsychologia, 42, 1168–1177. Sereno, M. I., Pitzalis, S., & Martinez, A. (2001). Mapping of contralateral space in retinotopic coordinates by a parietal cortical area in humans. Science, 294(5545), 1350–1354. Shikata, E., McNamara, A., Sprenger, A., Hamzei, F., Glauche, V., Büchel, C., et al (2007). Localization of human intraparietal areas AIP, CIP, and LIP using surface orientation and saccadic eye movement tasks. Human Brain Mapping, 29(4), 411–421. Stuss, D. T., & Anderson, V. (2004). The frontal lobes and theory of mind: Developmental concepts from adult focal lesion research. Brain and Cognition, 55(1), 69–83. Talairach, J., & Tournoux, P. (1988). Co-planar stereotaxic atlas of the human brain. Stuttgart: Georg Thieme Verlag. Tunik, E., Frey, S. H., & Grafton, S. T. (2005). Virtual lesions of the anterior intraparietal area disrupt goal-dependent on-line adjustments of grasp. Nature Neuroscience, 8(4), 505–511. Uddin, L. Q., Molnar-Szakacs, I., Zaidel, E., & Iacoboni, M. (2006). RTMS to the right inferior parietal lobule disrupts self-other discrimination. Social Cognitive and Affective Neuroscience, 1(1), 65–71. Uddin, L. Q., Iacoboni, M., Lange, C., & Keenan, J. P. (2007). The self and social cognition: The role of cortical midline structures and mirror neurons. Trends in Cognitive Neuroscience, 11(4), 153–157. Van der Weel, F. R., & van der Meer, A. L. H. (2009). Seeing it coming: Infants’ brain responses to looming danger. Naturwissenschaften, 96, 1385–1391. Vogeley, K., Kurthen, M., Falkai, P., & Maier, W. (1999). Essential functions of the human self model are implemented in the prefrontal cortex. Conscious and Cognition, 8(3), 343–363. Vogeley, K., & Fink, G. R. (2003). Neural correlates of the first-person-perspective. Trends in Cognitive Sciences, 7(1), 38–42. Vogeley, K., May, M., Ritzl, A., Falkai, P., Zilles, K., & Fink, G. R. (2004). Neural correlates of first-person perspective as one constituent of human selfconsciousness. Journal of Cognitive Neurosciences, 16(5), 817–827. Watson, J. D., Myers, R., Frackowiak, R. S., Hajnal, J. V., Woods, R. P., Mazziotta, J. C., et al (1993). Area V5 of the human brain: Evidence from a combined study using positron emission tomography and magnetic resonance imaging. Cerebral Cortex, 3(2), 79–94. Weiss, P. H., Marshall, J. C., Zilles, K., & Fink, G. R. (2003). Are action and perception in near and far space additive or interactive factors? NeuroImage, 18(4), 837–846. Weiss, P. H., Rahbari, N. N., Lux, S., Pietrzyk, U., Noth, J., & Fink, G. R. (2006). Processing the spatial configuration of complex actions involves right posterior parietal cortex: An fMRI study with clinical implications. Human Brain Mapping, 27(12), 1004–1014. H. Kockler et al. / Consciousness and Cognition 19 (2010) 690–701 701 Wilms, M., Eickhoff, S. B., Specht, K., Amunts, K., Shah, N. J., Malikovic, A., et al (2005). Anatomy and Embryology, 210(5–6), 485–495. Zacks, J. M., Rypma, B., Gabrieli, J. D., Tversky, B., & Glover, G. H. (1999). Imagined transformations of bodies: An fMRI investigation. Neuropsychologia, 37(9), 1029–1040. Zacks, J. M., & Michelon, P. (2005). Transformations of visuospatial images. Behavioral and Cognitive Neuroscience Reviews, 4(2), 96–118. Zacks, J. M. (2008). Neuroimaging studies of mental rotation: A meta-analysis and review. Journal of Cognitive Neuroscience, 20(1), 1–19.

Consciousness and Cognition 19 (2010) 1151–1153 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Commentary Reflections on the varieties of hypnotizables: A commentary on Terhune and Cardeña q V.K. Kumar Department of Psychology, West Chester University of Pennsylvania, West Chester, PA 19383, United States a r t i c l e i n f o Article history: Received 29 March 2010 Available online 28 April 2010 Keyword: Hypnotic types a b s t r a c t This commentary reflects on the varieties of high hypnotizable subjects suggested in the works by Barber, Barrett, Pekala and colleagues, and Terhune and Cardeña (2010). These different studies point to the existence of different types of low, medium, and high hypnotizable subjects. However, types of high hypnotizables have received the most attention. Two main concerns are raised in this commentary: (a) drawing parallels between the suggested typologies is not without problems given methodological differences among different studies, and (b) the low base rates of these special types is likely not to appeal to a typical clinician, already resistant to testing for hypnotizability, to conduct initial assessments so as to tailor suggestion to fit specific typologies. Ó 2010 Elsevier Inc. All rights reserved. 1. Introduction Terhune and Cardeña study contributes to the idea there are indeed different types of hypnotizable subjects. Barber’s (see 2000 for a good review) differentiated between three types of good hypnotic subjects: Amnesia-prone, fantasy-prone, and positively-set. Barber’s discovery of the fantasy-prone was based on extensive interviews of good and poor hypnotic subjects focused on their childhood memories, fantasies, sensory experiences, psychosomatic illnesses, etc. (p. 211). Barber’s (2000) positively-set subjects were based on the observation that the ‘‘best subjects... had very favorable attitude toward hypnosis, wanted to be a hypnotized, and expected to experience the suggested effects” (p. 224). Barber’s amnesia-prone comes from the work of Barrett (1990, 1996), who found, using similar methodology to Barber, support for two types of high hypnotizables: fantasizers and dissociaters. Barrett (1996) observed that the fantasizers, were more likely, than dissociaters, to (a) score higher on the Tellegen Absorption Scale (Tellegen & Atkinson, 1974), (b) lower on the Field’s (1965) Inventory Scale of Hypnotic Depth (ISHD), (c) need less time to reach deep trance, (d) remember trance details, and (e) experience hypnosis as similar to their other everyday experiences. Barber (2000), however, renamed Barrett’s ‘‘dissociaters” as ‘‘amnesia-prone” given their many reported amnesia experiences, and noted his preference for the term amnesia-prone to dissociaters since it ‘‘is ambiguous and confusing because it is loaded with surplus meaning and is used in different ways by different investigators (Cardeña, 1994)” (p. 218). In contrast to Barber and Barrett’s methodology, Pekala and colleagues used the Phenomenology of Consciousness Inventory (PCI; Pekala, 1991a), a state measure of subjective experiences, as the basis of their suggested typologies. They had subjects complete the PCI retrospectively in reference to a brief sitting-quietly interval embedded in the Harvard Group Scale of Hypnotic Susceptibility, Form A (HGSHS: A; Shor & Orne, 1962). Using cluster analysis on the 12 major PCI dimensions, q Commentary on Terhune, D. B., & Cardeña., E. (2010). Differential patterns of spontaneous experiential response to a hypnotic induction: A latent profile analysis. Consciousness and Cognition, 19, 1140–1150. Fax: +1 610 436 2845. E-mail address: Kkumar@wcupa.edu 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.04.006 1152 V.K. Kumar / Consciousness and Cognition 19 (2010) 1151–1153 they reported finding a variety of lows (e.g., classic lows, relaxed lows), mediums (e.g., nondialoging mediums, dialoging mediums), and highs (e.g., classic highs, fantasy highs, and compliant highs) in four studies (Forbes & Pekala, 1996; Pekala, 1991b; Pekala & Forbes, 1997; Pekala, Kumar, & Marcano, 1995). Terhune and Cardeña’s study (2010) had participants complete the PCI in reference to a resting epoch embedded in the Waterloo-Stanford Group Scale of Hypnotic Susceptibility, Form C (WSGC Bower, 1993). Additionally, they administered the ISHD in reference to the entire session. Their study differed in important ways from the studies of Pekala and colleagues inasmuch as they used: (a) the WSGC (and not HGSHS: A); (b) five-factor based scores (Dissociated Control, Positive Affect, Negative Affect, Visual Imagery, and Attention to Internal Processes) computed on the PCI (Kumar, Pekala, & Cummings, 1996), instead of the 12 major dimensions; and (c) Latent Profile Analysis (LPA) to identify profiles of hypnotizability, instead of cluster analysis. Their LPA revealed four profiles—the first two containing high hypnotizables and the last two containing medium and low hypnotizables. Based on further analyses, the first two profiles were named as ‘‘inward attention” and ‘‘dissociative” subtypes, respectively. Consistent with Barrett, they found the dissociative profile to be associated with higher involuntariness, but lower vivid imagery, compared with the inward attention profile. Contrary to expectations, the two profiles did not differ on memory and the dissociative profile was associated with greater vividness of imagery than profiles 3 and 4 consisting of the mediums and lows. Can we reconcile Barber’s three types and Barrett’s two types with those of Pekala and colleagues’ three (high) types and Terhune and Cardeña’s two types? Any efforts toward reconciliation of these respective outcomes need to consider methodological differences among them. While Barber and Barrett primarily used a trait approach, Pekala and colleagues, as is also true of Terhune and Cardeña (2010), used a state approach (i.e., analyzing ratings of subjective experiences during a resting epoch in hypnosis) to identify various subtypes. Barber (2000) saw Pekala and his colleagues’ finding of fantasy highs as a ‘‘statistical confirmation” (p. 231) of his fantasyprone, compliant highs of his positively-set, and classic highs of the amnesia-prone. However, we can ask: Can we fully equate Barber’s and Barrett’s fantasy-prone with Pekala and colleagues’ fantasy highs, despite methodological differences in their studies? As noted earlier, while Barber’s and Barrett’s characterizations were based on lengthy interviews concerning particular experiences, Pekala and colleagues’ characterizations were based on the subjects’ pattern of responding to particular PCI items in reference to a sitting-quietly interval during hypnosis. For the same reasons, we can ask: Can we equate Terhune and Cardeña’s dissociative type with Barrett’s dissociaters? It appears that the Terhune and Cardeña’s dissociative type resembles the classic highs of Pekala and colleagues inasmuch as they were characterized to ‘‘exhibit pronounced distortions in awareness, affect, and volitional control and reduced attention and imagery.” Furthermore, the inward attention type resembles the fantasy highs showing greater inward attention, visual imagery, and absorption than the dissociative types. But again, procedural differences in these studies raise caution in concluding that the typologies are indeed parallel. 2. Concluding remarks Despite different methodologies, taken together, these different studies suggest that high hypnotizables are not all alike, neither are all mediums and lows alike, in terms of the ways they attempt to achieve or experience the stated goals of hypnotic suggestions. However, examining for convergence among these putative types is not without problems given methodological differences among the studies. No doubt, there is theoretical and practical significance to identifying types of hynotizables. For example, Barber (2000) noted that while the fantasy-prone may not need a hypnotic induction due to their high responsiveness, the amnesia-prone may benefit from a hypnotic induction to achieve higher levels of responsiveness. For the positively-set, Barber‘s task- motivational instruction (e.g., think-with) should work well. However, Barber (2000) observed that the positively-set are not profoundly responsive to the ‘‘cognitive test suggestions such as amnesia and visual hallucination” (p. 230) and lamented that ‘‘far too little attention is given to the large number of useful suggestions that fit the special attributes of positively-set subjects...” (p. 230). A point worth considering is that only a small proportion of individuals seeking therapy will fit into these special types given their low base rates in the population. Any attempts to tailor suggestions to such small numbers will require additional time and resources to conduct initial assessments on the part of the clinician. However, knowing a typical clinician’s resistance to testing for hypnotizability, it remains to be seen as to how the notion of types will alter everyday practice. References Barber, T. X. (2000). A deeper understanding of hypnosis: Its secrets, its nature, its essence. American Journal of Clinical Hypnosis, 42, 208–272. Barrett, D. (1990). Deep trance subjects: A schema of two distinct subgroups. In R. G. Kunzendorf (Ed.), Mental imagery (pp. 101–112). NY: Plenum Press. Barrett, D. (1996). Fantasizers and dissociaters: Two types of high hypnotizables, two different imagery styles. In R. G. Kunzendorf, N. P. Spanos, & B. Wallace (Eds.), Hypnosis and imagination. Amityville, NY: Baywood. Bower, K. S. (1993). The Waterloo-Stanford Group C (WSGC) scale of hypnotic susceptibility: Normative and comparative data. International Journal of Clinical and Experimental Hypnosis, 42, 35–46. Cardeña, E. (1994). The domain of dissociation. In S. J. Lynn & J. W. Rhue (Eds.), Dissociation: Clinical and theoretical perspectives (pp. 15–31). NY: Guilford Press. Forbes, E. J., & Pekala, R. J. (1996). Types of hypnotically (un)susceptible individuals as a function of phenomenological experience: A partial replication. Australian Journal of Clinical and Experimental Hypnosis, 24, 92–109. V.K. Kumar / Consciousness and Cognition 19 (2010) 1151–1153 1153 Kumar, V. K., Pekala, R. J., & Cummings, J. (1996). Trait factors, state factors, and hypnotizability. International Journal of Clinical and Experimental Hypnosis, 44, 232–249. Pekala, R. J. (1991a). The phenomenology of consciousness inventory. West Chester, PA: Mid-Atlantic Educational Institute. Pekala, R. J. (1991b). Hypnotic types: Evidence from a cluster analysis of the phenomenal experience. Contemporary Hypnosis, 8, 95–104. Pekala, R. J., & Forbes, E. J. (1997). Types of hypnotically (un)susceptible individuals as a function of phenomenological experience: Toward a typology of hypnotic types. American Journal of Clinical Hypnosis, 39, 212–224. Pekala, R. J., Kumar, V. K., & Marcano, G. (1995). Hypnotic types: A partial replication concerning phenomenological experience. Contemporary Hypnosis, 12, 194–200. Shor, R. E., & Orne, E. C. (1962). The Harvard group scale of hypnotic susceptibility, form A. Palo Alto, CA: Consulting Psyhcologists Press. Tellegen, A., & Atkinson, G. (1974). Openness to absorbing and self-altering experiences (‘‘absorption”), a trait related to hypnotic susceptibility. Journal of Abnormal Psychology, 83, 268–277. Terhune, D. B., & Cardeña, E. (2010). Differential patterns of spontaneous experiential response to a hypnotic induction: A latent profile analysis. Consciousness and Cognition, 19, 1140–1150.

Consciousness and Cognition 43 (2016) 38–47 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Theory of mind in non-suicidal self-injury (NSSI) adolescents Fiorenzo Laghi a,⇑, Arianna Terrinoni b, Rita Cerutti c, Fiorella Fantini c, Serena Galosi b, Mauro Ferrara b, Francesca Marina Bosco d a Department of Developmental and Social and Psychology, Faculty of Medicine and Psychology, Sapienza, University of Rome, Via dei Marsi, 78, 00185 Roma, Italy Department of Pediatrics and Child and Adolescent Neurology and Psychiatry, Sapienza, University of Rome, Via dei Sabelli 108, 00185 Rome, Italy c Department of Dynamic and Clinical Psychology, Faculty of Medicine and Psychology, Sapienza, University of Rome, Via dei Marsi, 78, 00185 Roma, Italy d Department of Psychology and Center for Cognitive Science, University of Torino, Via Po 14, 10123 Torino, Italy b a r t i c l e i n f o Article history: Received 26 February 2016 Revised 10 May 2016 Accepted 13 May 2016 Available online 26 May 2016 Keywords: Non-suicidal self-injury Theory of mind Self-injuring behavior Adolescent inpatients Suicidal ideation a b s t r a c t The aim of the present study is to investigate different facets of the theory of mind (ToM), i.e. first vs. third-person, first vs. second-order ToM, egocentric vs. allocentric perspective, in a clinical sample of 20 non-suicidal self-injury (NSSI) adolescent inpatients and 20 healthy controls. Methods: We investigated whether performance in ToM tasks was related to both the type and frequency of self-injuring behavior and attitude toward life and death, using a semi-structured interview and different self-report questionnaires. Results: NSSI participants performed less well than the control group in all the ToM dimensions investigated. Furthermore, ToM performance was negatively related to Attraction to Death, in terms of both the type and frequency of self-injuring behavior, and it was positively related to Attraction to Life. Conclusions: These preliminary findings have interesting implications for future clinical investigations, in that they provide previously unavailable information regarding the association between ToM and NSSI behavior. Ó 2016 Elsevier Inc. All rights reserved. 1. Introduction Non-suicidal self-injury (NSSI) is broadly defined as a direct, not socially sanctioned behavior that causes physical injury which results in the destruction of one’s own body tissue in the absence of any observable intent to die (Muehlenkamp, 2005; Nock, 2010). In recent years the increase in all Western countries of self-injurious behaviors, especially during adolescence and young adulthood, has made NSSI a major public health issue (Klonsky, 2011; Klonsky, May, & Glenn, 2013). Studies suggest that NSSI is prevalent among preadolescents (approximately 7.5%; Hilt, Nock, Lloyd-Richardson, & Prinstein, 2008), and it peaks in middle and late adolescence, with lifetime prevalence rates ranging from 13% to 41.5% within community adolescent samples (Bifulco et al., 2014; Cerutti, Manca, Presaghi, & Gratz, 2011; Giletta, Scholte, Engels, Ciairano, & Prinstein, 2012; Jacobson & Gould, 2007; Laye-Gindhu & Schonert-Reichl, 2005; Zetterqvist, Lundh, Dahlström, & Svedin, 2013). Further studies involving clinical samples of adolescent inpatients have indicated notably higher lifetime rates of NSSI (ranging from 37% to 80%; Darche, 1990; Di Clemente, Ponton & Hartley,1991; Ferrara, Terrinoni, & Williams, 2012; Jacobson, Muehlenkamp, Miller, & Turner, 2008; Nock & Prinstein, 2004) suggesting that self-injurers may manifest more or different ⇑ Corresponding author. E-mail addresses: fiorenzo.laghi@uniroma1.it (F. Laghi), Arianna.terrinoni00@gmail.com (A. Terrinoni), rita.cerutti@uniroma1.it (R. Cerutti), fiorellafantini@hotmail.com (F. Fantini), mauro.ferrara@uniroma1.it (M. Ferrara), francesca.bosco@unito.it (F.M. Bosco). http://dx.doi.org/10.1016/j.concog.2016.05.004 1053-8100/Ó 2016 Elsevier Inc. All rights reserved. F. Laghi et al. / Consciousness and Cognition 43 (2016) 38–47 39 psychiatric problems than others (Klonsky & Olino, 2008). However, it should be noted that the variability in the prevalence estimates of NSSI among different studies may be related to the various ways in which NSSI has been defined and assessed including frequency of each self-injury behavior (e.g., one or more behaviors during the lifetime versus one or more behaviors occurring in the last 12 months) (Laye-Gindhu & Schonert-Reichl, 2005; Muehlenkamp, Claes, Havertape, & Plener, 2012). Generally, self-injurious behavior has been considered as relating specifically to severe mental health conditions, such as personality disorders (e.g., borderline personality disorder) and in the context of a wide range of different Axis I psychiatric disorders (Nock, Joiner, Gordon, Lloyd-Richardson, & Prinstein, 2006) but self-injury in the absence of a conscious suicidal intent may also be present without any psychiatric comorbidities (Wilkinson, 2013) supporting the hypothesis of NSSI as a potentially separate diagnostic entity (Selby, Bender, Gordon, Nock, & Joiner, 2012). In this perspective, NSSI is currently taken into consideration in DSM-5 as a potential distinct clinical entity with respect to other disorders including suicidal behavior (American Psychiatric Association, 2013). Although there is not yet an official diagnosis, its delimitation is helpful for converging on a clear definition. It is important in research and clinical settings to test suggested criteria sets for its diagnostic validity (In-Albon, Ruf, & Schmid, 2013), since NSSI takes a variety of forms that differ in terms of severity and frequency (Nock, 2010). Recent research has highlighted that individuals who engage in self-injury may deliberately use more than one method, either at the same time or on different occasions (Ferrara et al., 2012; Klonsky, 2011). NSSI among those who perform it repetitively can be mild, moderate, or severe depending on the lethality of the injury (Klonsky & Olino, 2008; Nock, 2010). The most common form of self-injury in clinical samples appears to be skin-cutting, utilized by 70% (Nock & Prinstein, 2004) or more (Klonsky et al., 2013) of those who self-injure, with most cases among females (Claes, Vandereycken, & Vertommen, 2006). Previous studies have described the typical self-cutter as female, adolescent or young adult, single, usually from a middle-to-upper class background (Darche, 1990; Favazza & Conterio, 1989). As noted by Whitlock and Selekman (2014), the chosen method of NSSI often communicates a lot about the intent of the behavior and cutting is frequently observed in association with suicidal thoughts and behaviors. Although the distinction between NSSI and suicidal behaviors has been made (Nock, 2010) the question is still controversial (Joiner, Di Ribeiro, & Silva, 2012; Klonsky et al., 2013). The DSM-5 (American Psychiatric Association, 2013), refers to suicidal behavior disorder (SBD) and NSSI as different conditions, placing them within Section III. 1.1. NSSI and theory of mind For most individuals who engage in NSSI, its purpose appears to be to reduce the intensity of painful emotions (e.g., anxiety, tension) and previous studies have highlighted that the reduction of emotion dysregulation will decrease the need for maladaptive behaviors that are employed to regulate emotions, such as self-injury (Gratz, 2007). Despite research documenting a strong association between emotion dysregulation and non-suicidal self-injury (Gratz & Chapman, 2007), there is a notable lack of research on the relationship between NSSI and Theory of Mind (ToM). Theory of mind is the specific ability of people to attribute mental states, such as intentions, emotions, desires and beliefs, to themselves and others in order to explain and predict behavior (Premack & Woodruff, 1978). The understanding of mind is one of the most important attainments in childhood which allows children to function socially and to distinguish accidental and intended behavior, wishes and reality, truth and deception (Bellagamba, Laghi, Lonigro, & Pace, 2012; Bellagamba, Laghi, Lonigro, Pace, & Longobardi, 2014). Thus, theory of mind is fundamental for the understanding of the social world and engaging in human interactions. Difficulties with mentalizing and emotional awareness, that are part of ToM, have implications for several severe personality disorders (e.g., borderline personality disorder) as well as psychological problems deriving from the low capacity to think about mental states (Allen, Fonagy, & Bateman, 2008). In a recent study, Sharp et al. (2011) investigated the mentalizing capacity of adolescents with borderline traits. This was the first study to evaluate the impairment of mentalizing in BPD considering, specifically, its potential dysfunctions such as ‘‘hypermentalizing” (excessive, inaccurate mentalizing), which result in an incorrect and reduced attribution of the mental state rather than in a complete absence of ToM. The results of this study revealed that neither undermentalizing nor its complete absence are linked to borderline traits. Hypermentalizing (i.e. excessive interpretation of mental states) is, instead, strongly associated with the traits characterizing borderline adolescents. An assessment of ToM in NSSI may be important, since recent studies have produced evidence to suggest that individuals who engage in NSSI have significant difficulties with managing specific mental states, i.e. emotions. It is generally assumed that NSSI often represents a dysfunctional form of emotional regulation (Klonsky, 2007; Klonsky et al., 2013; Nock & Prinstein, 2004); Fonagy and Bateman (2008) suggested that mentalization may be temporarily inhibited by intense emotional arousal. NSSI can be considered as a maladaptive coping strategy to regulate and control emotion (Di Pierro, Sarno, Gallucci, & Madeddu, 2014; Gratz, 2007), and it is performed with the intent of alleviating negative affect (Klonsky, 2007); similarly, increased mentalization might reduce NSSI behaviors. Since ToM allows psychic and symbolic representation of one’s own internal state and at the same time it serves to regulate and control one’s emotions, increased mindfulness, defined as attentiveness to present experience, may serve to reduce the symptoms of NSSI (Lundh, Karim, & Quilisch, 2007). Lundh et al. (2007) demonstrated the association between NSSI and mindfulness, using The Mindful Attention Awareness Scale (MAAS) that was constructed by Brown and Ryan (2003) to assess individual differences in the frequency of mindful 40 F. Laghi et al. / Consciousness and Cognition 43 (2016) 38–47 states over time, defined as attention to and awareness of what is occurring in the present. This study found self-injurers to have lower mindfulness than non-self-injurers. In another recent study, Rossouw and Fonagy (2012) examined whether mentalization-based treatment for adolescents (MBT-A) was more effective than treatment as usual (TAU) for adolescents who engaged in self-harm behaviors. The authors determined that MBT-A may be more effective than routine care in reducing the recurrence of self-harm behavior as well as depression. However, this study considered self-harm acts from a broader perspective without distinguishing between suicidal and non-suicidal intentions. Despite a growing body of research on the clinical correlates of NSSI in adolescents, to date there has been limited research to examine the relationship between ToM and NSSI and mentalizing capacity comparing adolescent inpatients with healthy adolescents. In particular, to the best of our knowledge, no previous studies have provided an articulated assessment of ToM abilities in NSSI patients. A large amount of research has highlighted the complex nature of ToM (Bosco, Colle, & Tirassa, 2009; Laghi et al., 2014; Lonigro, Laghi, Baiocco, & Baumgartner, 2014; Tirassa, Bosco, & Colle, 2006a; Tirassa, Bosco, & Colle, 2006b) and hinted at the possibility of breaking it down into different aspects or components (Abu-Akel, 2003; Vogeley & Fink, 2003). Nevertheless, the classical ToM tasks (see for example Baron-Cohen, O’Riordan, Stone, Jones, & Plaisted, 1999; Happé, 1994; Wimmer & Perner, 1983; but also see more recent tools such as Hutchins, Prelock, & Bonazinga, 2012; Sivaratnam, Cornish, Gray, Howlin, & Rinehart, 2012) focus on a specific component of ToM, called third-person ToM, i.e. the ability to reason about another person’s mental states, overlooking first-person ToM, i.e. the ability to reason about one’s own mental states (see Nichols & Stich, 2003). Another important distinction made in the literature is between first and second-order ToM. First-order ToM engages the ability to understand a person’s belief about a state of the world, while second-order ToM involves the ability to ascribe nested mental states, that is to understand a person’s beliefs about the beliefs of another person. Empirical data have shown that in clinical populations second-order ToM tasks are more difficult than first-order ones (Chiavarino et al., 2015; Laghi et al., 2014). Finally, a distinction has been made between egocentrism and allocentrism (Frith & De Vignemont, 2005). In the egocentric perspective, others are represented in relation to the self, while in the allocentric perspective others’ mental states are represented independently from the self. In order to investigate all these ToM aspects, i.e. first vs. third person, first vs. second order, egocentric vs. allocentric, with a single tool, we used the Theory of Mind Assessment Scale (Th.o.m.a.s.; Bosco et al., 2009). This is a semi-structured interview based on open questions (see Appendix A). Th.o.m.a.s. has proved to be useful for investigating healthy adolescents (Bosco, Gabbatore, & Tirassa, 2014), young adults with bulimia (Laghi et al., 2014) and various clinical and non-clinical populations such as: individuals with schizophrenia, (Bosco, Colle, De Fazio, et al., 2009), sex offenders (Castellino, Bosco, Marshall, Marshall, & Veglia, 2011), people with alcohol use disorder (Bosco, Capozzi, Colle, Marostica, & Tirassa, 2013) and those with congenital heart disease (Chiavarino et al., 2015). In all the above-mentioned studies Th.o.m.a.s. has been found to be a useful clinical instrument, able to discriminate between clinical and healthy participants; it has offered the opportunity to directly compare and observe various patterns of performance in relation to several aspects of ToM and has also shown good inter-reliability. Globally, these studies testify to the appropriateness of Th.o.m.a.s as a tool for assessing and comparing different ToM components in clinical populations. To summarize, the present study was designed to address the existing gap in the NSSI literature and provide additional information about NSSI in the clinical population of adolescent inpatients. The key issue is whether a focus on theory of mind abilities is useful, i.e. provides an appropriate domain for therapeutic intervention (Fonagy, Bateman, & Bateman, 2011). 1.2. Study aims The purpose of our study was to describe the phenomenological aspects of non-suicidal self-injury in a clinical sample of NSSI adolescents. The comparison between a clinical sample of adolescent psychiatric inpatients and a control group may add information for theoretical and clinical models of NSSI. More specifically, this study aimed to provide an articulated assessment of these patients’ theory of mind abilities. Since there is little empirical evidence regarding the ability of NSSI subjects to perform theory of mind (ToM) tasks, we wanted to explore different facets of theory of mind abilities, i.e. first vs. third person ToM, first vs. second order, egocentric vs. allocentric perspective, in order to gain a better understanding of the complex nature of this behavior. Additionally, we wanted to explore whether performance in ToM tasks: (i) might be significantly related to both the type and frequency of self-injuring behaviors; and (ii) to attitude toward life and death. 2. Methods 2.1. Participants The participants for this study were adolescents consecutively admitted to a psychiatric inpatient unit ((Department of Pediatrics and Child and Adolescent Neurology and Psychiatry, Sapienza), for blinded review) that specializes in treating severe behavioral problems and psychotic episodes. All youths who presented NSSI over a 12-month period preceding admission were included. Information collected during the clinical assessment included duration of illness, past medical history, F. Laghi et al. / Consciousness and Cognition 43 (2016) 38–47 41 and demographics. Each patient was evaluated by one of two senior clinicians who established whether they met the inclusion criteria listed below and whether they had engaged in repetitive NSSI according to the proposed DSM-5 criteria for a NSSI condition. Diagnosis of NSSI was made according to the criteria proposed by DSM 5 (Conditions for Further Studies Section III; APA, 2013), screened using DSHI (Deliberate Self Harm Inventory), and R-NSSI-Q (Repetitive-Not Suicidal Self Injury-Questionnaire): on 5 or more days, in the last year, the individual engaged in intentional self-inflicted damage to the surface of his or her body, without suicidal intent (criterion A); the individual engages in the self-injurious behavior with the expectation to obtain relief from a negative feeling or cognitive state, to resolve an interpersonal difficulty, or to induce a positive feeling state (criterion B); the intentional self-injury is associated with interpersonal difficulties or negative feelings or thoughts, an antecedent period of preoccupation with the intended behavior that is difficult to control, or thinking about self-injury that occurs frequently, even when it is not acted upon (criterion C); the behavior is not socially sanctioned (criterion D); it causes clinically significant distress or interference in interpersonal, academic, or other important areas of functioning (criterion E); and it does not occur exclusively during psychotic episodes, delirium, substance intoxication, or substance withdrawal (criterion F). The diagnosis was corroborated using the Italian version of the Schedule for Affective Disorders and Schizophrenia for School-Age Children/Present and Lifetime Version (K–SADS–PL; Kaufman et al., 1997). At the K–SADS–PL, 11 patients presented Mood disorders, 7, Impulse control disorders, and 2 presented Eating disorders. All adolescents with an admission diagnosis of attempted suicide were triaged to a separate unit and not included. Patients diagnosed with intellectual disabilities, pervasive developmental disorders, schizophrenia spectrum disorders or associated neurological conditions were excluded, given the relevance of such conditions for stereotypic self-injury. The selected clinical sample consisted of 20 female adolescents (age 12–17: mean = 14.74; SD = 1.64). Healthy controls (HCs; n = 20 females) were students, matched to the study group for age and years of formal education, who were evaluated at a research laboratory at the Faculty of Medicine and Psychology. Exclusion criteria for the HC group were Axis I mental disorder, neurological disease, history of head trauma and current use of psychotropic medication. All participants were native Italian speakers. 2.2. Materials 2.2.1. The Deliberate Self-harm Inventory (DSHI) The DSHI (Gratz, 2001) is a 17-item self-report measure that assesses lifetime history of various aspects of NSSI (defined as the deliberate, direct destruction of body tissue without suicidal intent), including frequency, duration, and type of NSSI behavior. Specifically, the DSHI asks participants whether and how often they have engaged in a variety of behaviors ‘‘intentionally (i.e., on purpose)”. Further, for the one behavior that could also be used to end one’s life (cutting), participants are asked whether they have cut themselves ‘‘without intending to kill yourself”. The DSHI has been found to demonstrate high internal consistency, adequate test–retest reliability, and adequate construct, discriminant, and convergent validity (Fliege et al., 2006; Gratz, 2001). The DSHI was recently translated into Italian by Cerutti et al. (2011) using the translation–backtran slation method, and was found to have adequate psychometric properties within an Italian adult sample, including adequate internal consistency, and good convergent and discriminant validity, as evidenced by significant correlations with both clinical and non-clinical dimensions of personality and body perception. For descriptive purposes we used the indexes of frequency = number of episodes per-month (seldom = episodic self-injury; sometimes to always = repetitive self-injury), types of self-injurious behaviors (e.g., self-cutting, self-burning, etc.), diversification = occurrence of multiple types of selfinjurious behaviors measured on a two-level scale (2–4: moderate diversification; and 5–11: high diversification). 2.2.2. The repetitive non-suicidal self-injury questionnaire The Repetitive Non-Suicidal Self-Injury Questionnaire (R-NSSI-Q; Manca, Presaghi, & Cerutti, 2014) is composed of 15 items, and in conjunction with the DSHI, is a valid and reliable measure that may also be helpful for identifying an area of NSSI risk. The items fall within one of the following criteria of the current DSM-5: failing to resist the impulse to selfinjure or being frightened of being unable to resist the impulse to self-injure (criterion B2); increased levels of tension just before committing the self-injurious act (criterion B1); a pervasive sensation of gratification and relief immediately after having committed the self-injurious behavior (criterion B4); feelings of estrangement and/or guilt and difficulty in sharing the self-injurious experience with others (criterion C); check for the repetitiveness of the self-injuring behavior. This last item only partially satisfies criterion B3 (‘‘The urge to engage in self-injury occurs frequently, although it might not be acted upon”), because it did not ask for an intention but rather asked for endorsed self-injurious behaviors. The R-NSSI-Q has been used in a number of studies, and its reliability have been shown to be satisfactory (Cerutti, Presaghi, Manca, & Gratz, 2012; Manca et al., 2014). A score of 21 of the R-NSSI-Q resulted in the ‘‘optimal” cut-off score that maximally discriminated between occasional and repetitive NSSI adolescents. Internal reliability of the R-NSSI-Q in the present study was 0.83. 2.2.3. The Multi-Attitude Suicide Tendency (MAST) The Multi-Attitude Suicide Tendency (MAST; Osman et al., 2000) scale is a 30-item self-report measure of adolescent attitude toward life and death. The four types of conflicting attitudes identified are Attraction to Life (AL, 7 items), Repulsion by Life (RL, 7 items), Attraction to Death (AD, 7 items) and Repulsion by Death (RD, 9 items). Aspects concerning robustness and reliability of the Scale are included in Orbach et al. (1991). All dimensions exhibited acceptable levels of internal consistency (Cronbach’s a in the present study ranged from 0.83 to 0.94). 42 F. Laghi et al. / Consciousness and Cognition 43 (2016) 38–47 2.2.4. Theory of Mind The Theory of Mind Assessment Scale (Th.o.m.a.s.) is a semi-structured interview aimed at assessing a subject’s theory of mind (Bosco, Colle, De Fazio, et al., 2009; see the Appendix A for a more detailed description). Recently, the psychometric proprieties of ToM has been provided in a group of healthy adolescents and adults (Bosco, Gabbatore, Tirassa, & Testa, 2016) showing good internal consistency and inter-rater agreement. Furthermore factor analysis pointed out a good fit of the 4 scales composing it that also showed an high correlation. All the interviewees gave their written consent to the tape-recording of their interviews, that were then transcribed. The transcripts were rated by two independent judges, who were blind to the goal of the research and to which participants belonged to the experimental or control group. Each judge assigned each answer a score from 0 to 4 (see the Appendix B for a more detailed description). Following previous studies (Bosco, Colle, De Fazio, et al., 2009; Bosco et al., 2014), coscore reliability on the theory of mind assessment was established at 90% before data collection began. Reliability for the two coders was calculated by correlating their scores on four scales for each interview. Range reliability was 0.83–0.87, and was calculated by using the Spearman Brown correction formula. Disagreements were resolved by conferencing. 2.3. Procedure All participants were tested individually in a quiet room. They gave their written informed consent before completing the questionnaires and within one week of the first session where they completed the questionnaires, they were administered the Th.o.m.a.s. interview. This was done during a two-to-three-hour session by two interviewers – both of whom were trained experienced clinicians – under the supervision of the first author. Participants in the control group completed the questionnaires in a laboratory at the Department of Developmental and Social and Psychology, Faculty of Medicine and Psychology, Sapienza, and were interviewed within one week of the first session. This survey was reviewed and approved by the Ethics Committee of the Department of Developmental and Social and Psychology, Faculty of Medicine and Psychology, Sapienza. 2.4. Data analysis The Statistical Package for the Social Sciences (SPSS 18.0) was used to conduct bivariate and multivariate analyses relating to independent variables. We conducted ANOVAs to investigate differences between the two groups regarding demographic characteristics and the Th.o.m.a.s scale and dimensions. Within-subjects ANOVAs with four levels on within-subjects factors (both scale type and subscale dimensions) were also carried out. We used ANOVA to verify significant differences between the groups. Mann–Whitney U tests were conducted to examine differences between the two clinical groups for the occurrence of multiple types of self-harming behaviors (1–4 vs. 5–11 types). Additionally, for the clinical sample, we performed correlations to investigate the relationship between the Total Th.o.m.a.s score and Attraction to Life, in terms of both the type and frequency of self-harming behaviors. 3. Results 3.1. Demographic and clinical characteristics All individuals in the clinical sample had performed more than 5 NSSI in the previous 12 months and obtained a score above the cut-off (mean = 31; SD = 4.24; range: 22–39) for The Repetitive Non-Suicidal Self-Injury Questionnaire. Consistent with past research on NSSI within clinical samples of adolescents, most of them (N = 18) reported using at least two different methods to injure themselves, with the most commonly endorsed methods (see Table 1) being cutting, head-banging, and interference with wound healing. Table 1 provides information both on the frequency of each NSSI behavior among this clinical sample and on the reported age of onset for each of the NSSI behaviors. 3.2. ToM performance in NSSI subjects and controls The ANOVA showed a significant group effect on all theory of mind dimensions and on the subscales, where the NSSI group achieved significantly lower scores than the controls, as reported in Table 2. Focusing on the NSSI group’s performance on the Th.o.m.a.s., we conducted a within-subjects ANOVA with four levels on within-subjects factors (scale type: A, I–Me, first-person ToM in an egocentric perspective; B, Other–Self, third-person ToM in an allocentric perspective; C, Me–Other, third-person ToM in an egocentric perspective; D, Other–Me; second-order ToM task). We found no significant differences in the NSSI subjects’ mean scores on the four individual scales, F(3, 76) = 1.09, p = 0.36. We did not find any significant differences between the NSSI subjects’ mean scores on the three individual Th.o. m.a.s. subscales (Awareness, Relation, and Realization), F(2, 57) = 0.30, p = 0.74. Focusing on the control group’s performance on the Th.o.m.a.s., we found significant differences in subjects’ mean scores on the four individual scales, F(3, 76) = 18.93, p < 0.001. In particular, post hoc pairwise comparison (Bonferroni corrected; 43 F. Laghi et al. / Consciousness and Cognition 43 (2016) 38–47 Table 1 Clinical characteristics of NSSI group. NSSI behavior Age of onset Types f Cutting 19 Banging head 13 Interference with wound healing 13 Severe scratching 11 Punching self 11 Biting 10 Carving words or pictures into skin 8 Sticking pins, needles, staples into skin 6 Burning 3 Other forms of self-injury 1 Dripping acid on skin Occurrence of multiple types of self-injurious behaviors 1–4 9 5–11 11 Illness duration (months) f% Min-max M SD Min-max M SD 95 65 65 55 55 50 40 30 15 5 – 11–16 8–17 6–16 12–17 6–14 6–14 12–16 8–17 13–15 6 12.74 13.00 10.77 13.73 11.91 9.80 13.50 12.50 13.67 6.00 1.48 2.20 3.09 1.90 2.66 3.33 1.77 2.95 1.15 – 1–54 1–60 1–108 2–60 1–132 4–132 2–40 2–96 6–54 132 19.53 20.69 43.92 15.36 26.36 48.90 15.13 31.33 26.00 132 16.30 19.24 38.96 17.37 38.56 45.30 15.94 36.28 24.98 – 45 55 Table 2 ToM performance in NSSI subjects and controls: descriptive statistics and significant group effects. HC (N = 20) NSSI (N = 20) F(1,38) Partial g2 M SD M SD Th.om.as scales A I-Me B Other-Self C Me-Other D Other-Me 3.26 3.15 3.09 2.55 0.34 0.27 0.35 0.34 2.39 2.04 2.20 2.01 0.59 0.81 0.79 0.75 32.38** 33.76** 21.30** 8.66* 0.46 0.47 0.36 0.19 Th.om.as subscales Awareness Relation Realization 3.15 3.09 2.80 0.22 0.26 0.31 2.21 2.09 2.05 0.65 0.64 0.72 37.87** 41.80** 18.08* 0.49 0.52 0.32 Th.om.as Total score 3.01 0.20 2.16 0.62 34.44** 0.47 Note: NSSI = non-suicidal self-injurious group; HC = healthy controls. A I-Me = first person, egocentric; B Other-Self = third person, egocentric; C MeOther = first person, egocentric; D Other-Me = second order ToM task. * p < 0.01. ** p < 0.001. p < 0.01) revealed that the subjects scored lower on scale D (I–Me), which investigates second-order ToM, than on scales A, B, and C. There were no significant differences between the latter three scales. We found significant differences between the control subjects’ mean scores on the three individual Th.o.m.a.s. subscales (Awareness, Relation, and Realization), F(2, 57) = 10.06, p < 0.001. In particular, the post hoc pairwise comparison (Bonferroni corrected; p < 0.01) revealed that the subjects scored lower on the Realization scale than on Awareness and Relation, which did not differ significantly. Table 3 Relationship between Theory of Mind, self-harming behaviors, and attitude toward life and death. 1 1. Th.om.as total score 2. AL (MAST) 3. RL (MAST) 4. AD (MAST) 5. RD (MAST) 6. Types (DSHI) 7. Frequency (DSHI) 2 3 4 5 6 7 – 0.87*** – – 0.45** 0.31 0.44* 0.10 0.54** 0.44* – 0.59*** 0.64*** 0.27 0.35 0.45** – 0.72*** 0.09 0.50** 0.45** – 0.08 0.46** 0.41 – 0.14 0.33 Note: Th.om.as Total Score = Theory of Mind Assessment Scale; AL = Attraction to Life; RL = Repulsion by Life; AD = Attraction to Death; RD: Repulsion by Death; Types (DSHI) = Types of self-injury behavior-The Deliberate Self-harm Inventory; Frequency (DSHI) = Frequency of self-injury behavior-The Deliberate Self-harm Inventory. * p = 0.05. ** p < 0.05. *** p < 0.01. 44 F. Laghi et al. / Consciousness and Cognition 43 (2016) 38–47 Additionally, Mann–Whitney U tests were conducted to examine differences between the two clinical groups for the occurrence of multiple types of self-injuring behaviors (1–4 vs. 5–11 types). The results revealed that adolescents with high diversification scored lower (M = 1.70; SD = 0.53) on scale D (I–Me) than subjects with low diversification (M = 2.40; SD = 0.81; Z = 2.01, p < 0.05). 3.3. Theory of Mind, self-harming behaviors, and attitude toward life and death The total ToM score (overall Th.o.m.a.s, score) was negatively related to Attraction to Death, in terms of both the types and frequency of self-injuring behaviors, and it was positively related to Attraction to Life, as reported in Table 3. 4. Discussion and conclusions To our knowledge this is the first study to investigate the relationship between mild or severe NSSI disorder diagnosed according to the DSM-5 proposed criteria and Theory of Mind comparing a clinical group with a non-clinical control group of adolescents. We examined DSM-5 criteria for an NSSI disorder in female inpatient adolescents. As we expected, our findings indicate that in the clinical group, adolescent inpatients showed more than one type of NSSI and displayed a repetitive modality of these behaviors. The most reported method was cutting (i.e., cutting wrists, arms, or other parts of the body); this datum is consistent with the majority of studies that have shown this modality to be more frequent among girls who have engaged in NSSI both in clinical and non-clinical populations (Claes et al., 2006; Laye-Gindhu & Schonert-Reichl, 2005). Interestingly, we found an impaired ToM to be associated with NSSI. Adolescent inpatients, who reported having engaged in at least one kind of self-injury for more than five days, exhibited poorer performance on the overall Th.o.m.a.s. scale than HCs. Specifically, the NSSI group scored lower on scales A (I-me: first-person ToM in an egocentric perspective), B (Otherself: third person and allocentric perspective), C (I-other: third-person ToM in an egocentric perspective), D (Other–Me: second-order ToM task) with respect to the control group. Unlike in the NSSI group, where no difference emerged among the four scales, the control group obtained lower scores on scale D, showing no significant differences among the other three scales. Additionally, adolescent inpatients who engaged in NSSI also performed worse than healthy adolescents on the Awareness, Relation, and Realization subscales. Furthermore, while the control subjects scored lower on the Realization scale than on Awareness and Relation, no difference emerged among the three subscales in adolescents with NSSI. These findings suggest that adolescent inpatients with NSSI are more likely to have a pervasive and comprehensive ToM impairment than adolescents in the control group. Self-injury inpatients find it more difficult to be aware of their mental states, i.e., emotions, desire and beliefs and physical sensations, to adapt to these and adopt effective strategies to achieve a desired state, that is, to reduce their psychological suffering. Conversely, in the control group the differences that emerged belong to a profile of the typical adolescent phase of development for which the more complex mentalizing abilities, such as second-order ToM (Scherzer, Leveillé, Achim, Boisseau, & Stip, 2012) are not yet completely acquired (Bosco, Colle, De Fazio, et al., 2009; for a review see Brizio, Gabbatore, Tirassa, & Bosco, 2015). However, considering that no other studies have exhaustively examined ToM in patients with NSSI disorder, it is not yet possible to make a comparison between our results and those of other investigations. Nonetheless, we observed that some characteristics of ToM abilities in NSSI patients are similar to those of patients with other disorders, e.g. eating disorders, which frequently co-occur with NSSI (Claes, Klonsky, Muehlenkamp, Kuppens, & Vandereycken, 2010; Muehlenkamp, Claes, Peat, Smits, & Vandereycken, 2011). Specifically, our findings are similar to those of a recent study (Laghi et al., 2014) exploring the different dimensions of ToM in patients with eating disorders (i.e., bulimia nervosa, eating disorder not otherwise specified in the form of sub-threshold bulimia nervosa). The results of this latter study highlighted an overall impairment of ToM abilities, with patients with eating disorders performing less well than control subjects on the Th.o.m.a.s. scale, in particular on scales B and D as well as on the Awareness and Realization subscales. Nevertheless, patients with eating disorders showed a less severe impairment of ToM in comparison to NSSI patients. It has been argued that NSSI may contribute to later suicidal thoughts and behaviors (Nock et al., 2006) and that it plays a role in the development of the acquired capability for suicide because pain tolerance, combined with a desire to die, might increase the risk of lethal self-injury (Joiner et al., 2012). Nevertheless, an important distinction should be made between NSSI as an act to preserve life as a way of terminating an intolerable emotional condition (Klonsky, 2007) or alleviating negative affect (Gratz & Chapman, 2007), and suicidal behavior disorder that is considered a pathological condition characterized by the presence of one or more suicide attempts (Klonsky et al., 2013). In line with such perspective, the present findings show that the clinical sub-group reported a severe self-injury, - i.e., at least 1 NSSI act that required surgical treatment or NSSI acts over 12 or more days using a single method or NSSI acts over 8 or more days using more than one method (APA, 2013) - with sub-threshold depressive symptoms, suicidal thoughts and anhedonic traits. Our results are in line with a previous study (Ferrara et al., 2012) in which the patients were found to have an impaired ToM profile similar to that of the participants in the present study. Also in line with the study by Ferrara et al. (2012), our results showed that performance by our participants on the MAST scale of Attitude toward life (AL) was positively related to their total ToM score and that performance on the Attitude toward death (AD) scale was related to a higher frequency and diversification of NSSI. F. Laghi et al. / Consciousness and Cognition 43 (2016) 38–47 45 Some limitations of the study must be pointed out. Firstly, the relationship identified in the study is correlational and not causal. Therefore, this study only represents a first step in understanding the key components of mindreading abilities as a protective factor in NSSI adolescents. Secondly, the results of this study could thus be considered as a pilot study. Further research is required to replicate and confirm these findings considering a larger sample of NSSI adolescent inpatients. Thirdly, because of the small sample, it has not been possible to consider potential covariates, such as sociodemographic variables, in the research design. It would be useful if future research were to provide an assessment using different tools, for instance including the Happe’s Strange Stories Test (Happé, 1994) or other instruments that can evaluate competences in reading mental states in more typical everyday life situations. Despite this limit, the results of our research seem promising, especially in the light of a recent study by Scott, Pilkonis, Hipwell, Keenan, and Stepp (2015) showing that adolescent girls who have engaged in NSSI also report suicide ideation. NSSI thus represents a particularly high-risk group in need of prevention and intervention efforts. Specialized treatment programs designed to enhance mentalizing, as well as perceived meaning in life and developing positive relationships with others, could therefore be particularly helpful in reducing the risk of suicide among adolescents with NSSI (Whitlock et al., 2012). It has been demonstrated that Mentalization-based Therapy (MBT) is more effective than routine care in reducing suicidal and non-suicidal self-injury as well as depression in female adolescents (Bateman & Fonagy, 2010). To conclude, our findings seem to indicate that ToM is an interesting focus for both future research and forthcoming work on the prevention and treatment of mild and severe types of NSSI that might plausibly benefit from targeted intervention. Improved mentalizing may prevent the repetition and the dangerousness of NSSI behaviors as well as the risk for suicidal ideation. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.concog. 2016.05.004. References Abu-Akel, A. (2003). A neurobiological mapping of theory of mind. Brain Research Reviews, 43, 29–40. http://dx.doi.org/10.1016/S0165-0173(03)00190-5. Allen, J. G., Fonagy, P., & Bateman, A. (2008). Mentalizing in clinical practice. Washington, DC: American Psychiatric Publishing. American Psychiatric Association (2013). Diagnostic and statistical manual of mental disorders (5th ed.) Arlington, VA: American Psychiatric Publishing. Baron-Cohen, S., O’Riordan, M., Stone, V., Jones, R., & Plaisted, K. (1999). Recognition of faux pas by normally developing children and children with Asperger syndrome or high-functioning autism. Journal of Autism and Developmental Disorders, 29(5), 407–418. Bateman, A., & Fonagy, P. (2010). Mentalization based treatment for borderline personality disorder. World Psychiatry, 9(1), 11–15. Bellagamba, F., Laghi, F., Lonigro, A., & Pace, C. S. (2012). Re-enactment of intended acts from a video presentation by 18- and 24-month-old children. Cognitive Processing, 13(4), 381–386. Bellagamba, F., Laghi, F., Lonigro, A., Pace, C. S., & Longobardi, E. (2014). Concurrent relations between inhibitory control, vocabulary and internal state language in 18- and 24-month-old children Italian-speaking infants. European Journal of Developmental Psychology, 11(4), 420–432. Bifulco, A., Schimmenti, A., Moran, P., Jacobs, C., Bunn, A., & Rusu, A. C. (2014). Problem parental care and teenage deliberate self-harm in young community adults. Bulletin of the Menninger, 78(2), 95–114. Bosco, F., Capozzi, F., Colle, L., Marostica, P., & Tirassa, M. (2013). Theory of Mind deficit in subjects with alcohol use disorder: An analysis of mindreading processes. Alcohol Alcoholism, 49, 299–307. Bosco, F. M., Colle, L., De Fazio, S., Bono, A., Ruberti, S., & Tirassa, M. (2009). Th.o.m.a.s.: An exploratory assessment of theory of mind in schizophrenic subjects. Consciousness and Cognition, 18, 306–319. Bosco, F. M., Colle, L., & Tirassa, M. (2009b). The complexity of theory of mind. Consciousness and Cognition, 18, 323–324. Bosco, F. M., Gabbatore, I., & Tirassa, M. (2014). A broad assessment of theory of mind in adolescence: The complexity of mindreading. Consciousness and Cognition, 24, 84–97. Bosco, F. M., Gabbatore, I., Tirassa, M., & Testa, S. (2016). Psychometric properties of the Theory of Mind Assessment Scale in a sample of adolescents and adults. Frontiers in Psychology, 7, 566. http://dx.doi.org/10.3389/fpsyg.2016.00566. Brizio, A., Gabbatore, I., Tirassa, M., & Bosco, F. M. (2015). No more a child, not yet an adult”: Studying social cognition in adolescence. Frontiers in Psychology, 6, 1011. http://dx.doi.org/10.3389/fpsyg.2015.01011. Brown, K. W., & Ryan, R. M. (2003). The benefits of being present: Mindfulness and its role in psychological well-being. Journal of Personality and Social Psychology, 84, 822–848. Castellino, N., Bosco, F. M., Marshall, W. L., Marshall, L. E., & Veglia, F. (2011). Mindreading abilities in sexual offenders: An analysis of theory of mind processes. Consciousness and Cognition, 20, 1612–1662. http://dx.doi.org/10.1016/j.concog.2011.08.011. Cerutti, R., Manca, M., Presaghi, F., & Gratz, K. L. (2011). Prevalence and clinical correlates of deliberate self-harm among a community sample of Italian adolescents. Journal of Adolescence, 34, 337–347. Cerutti, R., Presaghi, F., Manca, M., & Gratz, K. L. (2012). Deliberate self-harm behavior among Italian young adults: Correlations with clinical and nonclinical dimensions of personality. American Journal of Orthopsychiatry, 82, 298–308. http://dx.doi.org/10.1111/j.1939-0025.2012.01169.x. Chiavarino, C., Bianchino, C., Brach-Prever, S., Riggi, C., Palumbo, L., Bara, B. G., & Bosco, F. M. (2015). Theory of mind deficit in adult patients with congenital heart disease. Journal of Health Psychology, 20(10), 1253–1262. http://dx.doi.org/10.1177/1359105313510337. Claes, L., Klonsky, E. D., Muehlenkamp, J., Kuppens, P., & Vandereycken, W. (2010). The affect-regulation function of nonsuicidal self-injury in eatingdisordered patients: Which affect states are regulated? Comprehensive Psychiatry, 51(4), 386–392. Claes, L., Vandereycken, W., & Vertommen, H. (2006). Self-injury in female versus male psychiatric patients: A comparison of characteristics, psychopathology and aggression regulation. Personality and Individual Differences, 42, 611–621. Darche, M. A. (1990). Psychological factors differentiating self-mutilating and non-self-mutilating adolescent inpatient females. Psychiatric Hospital, 21(1), 31–35. Di Clemente, R. J., Ponton, L. E., & Hartley, D. (1991). Prevalence and correlates of cutting behavior: Risk for HIV transmission. Journal of the American Academy of Child & Adolescent Psychiatry, 30, 735–740. Di Pierro, R., Sarno, I., Gallucci, M., & Madeddu, F. (2014). Nonsuicidal self-injury as an affect-regulation strategy and the moderating role of impulsivity. Child and Adolescent Mental Health, 19(4), 259–264. http://dx.doi.org/10.1111/camh.12063. 46 F. Laghi et al. / Consciousness and Cognition 43 (2016) 38–47 Favazza, A. R., & Conterio, K. (1989). Female habitual self-mutilators. Acta Psychiatrica Scandinavica, 79, 238–239. Ferrara, M., Terrinoni, A., & Williams, R. (2012). Non-suicidal self-injury (Nssi) in adolescent inpatients: Assessing personality features and attitude toward death. Child and Adolescent Psychiatry and Mental Health, 6, 12. Fliege, H., Kocalevent, R. D., Walter, O. B., Beck, S., Gratz, K. L., Gutierrez, P. M., & Klapp, B. F. (2006). Three assessment tools for deliberate self-harm and suicide behavior: Evaluation and psychopathological correlates. Journal of Psychosoatic Research, 61, 113–121. Fonagy, P., & Bateman, A. (2008). Attachment, mentalization and borderline personality disorder. European Psychotherapy, 8(1), 35–47. Fonagy, P., Bateman, A., & Bateman, A. (2011). The widening scope of mentalizing: A discussion. Psychology and Psychoterapy: Theory and Research and Practice, 84, 98–110. Frith, U., & De Vignemont, F. (2005). Egocentrism, allocentrism and Asperger syndrome. Consciousness and Cognition, 14, 719–738. http://dx.doi.org/10.1016/ j.concog.2005.04.006. Giletta, M., Scholte, R. H., Engels, R. C., Ciairano, S., & Prinstein, M. J. (2012). Adolescent non-suicidal self-injury: A cross-national study of community samples from Italy, the Netherlands and the United States. Psychiatry Research, 197(1–2), 66–72. Gratz, K. L. (2001). Measurement of deliberate self-harm: Preliminary data on the deliberate self-harm inventory. Journal of Psychopathology and Behavioral Assessment, 23, 253–263. Gratz, K. L. (2007). Targeting emotion dysregulation in the treatment of self-injury. Journal of Clinical Psychology, 63(11), 1091–1103 (in Session). Gratz, K. L., & Chapman, A. L. (2007). The role of emotional responding and childhood maltreatment in the development and maintenance of deliberate selfharm among male undergraduates. Psychology of Men and Masculinity, 8, 1–14. Happé, F. (1994). An advanced test of theory of mind: Understanding of story characters’ thoughts and feelings by able autistic, mentally handicapped, and normal children and adults. Journal of Autism and Developmental Disorders, 24, 129–154. Hilt, L. M., Nock, M. K., Lloyd-Richardson, E. E., & Prinstein, M. J. (2008). Longitudinal study of nonsuicidal self-injury among young adolescents: Rates, correlates, and preliminary test of an interpersonal model. Journal of Early Adolescence, 28, 455–469. Hutchins, T. L., Prelock, P. A., & Bonazinga, L. (2012). Psychometric evaluation of the theory of mind inventory (ToMI): A study of typically developing children and children with autism spectrum disorder. Journal of Autism and Developmental Disorders, 42(3), 327–341. In-Albon, T., Ruf, C., & Schmid, M. (2013). Proposed diagnostic criteria for the DSM-5 of nonsuicidal self-injury in female adolescents: diagnostic and clinical correlates. Psychiatry Journal. http://dx.doi.org/10.1155/2013/159208. Jacobson, C. M., & Gould, M. (2007). The epidemiology and phenomenology of non-suicidal self-injurious behavior among adolescents: A critical review of the literature. Archives of Suicide Research, 11, 129–147. Jacobson, C. M., Muehlenkamp, J. J., Miller, A. L., & Turner, J. B. (2008). Psychiatric impairment among adolescents engaging in different types of deliberate self-harm. Journal of Clinical Child and Adolescent Psychology, 37, 363–375. Joiner, T. E., Di Ribeiro, J. D., & Silva, C. (2012). Non suicidal self-injury, suicidal behavior and their co-occurrence as viewed through the lens of the interpersonal theory of suicide. Current Directions in Psychological Science, 21, 342–347. Kaufman, J., Birmaher, B., Brent, D., Rao, U. M. A., Flynn, C., Moreci, P., ... Ryan, N. (1997). Schedule for affective disorders and schizophrenia for school-age children-present and lifetime version (K-SADS-PL): Initial reliability and validity data. Journal of the American Academy of Child & Adolescent Psychiatry, 36(7), 980–988. http://dx.doi.org/10.1097/00004583-199707000-00021. Klonsky, E. D. (2007). The functions of deliberate self-injury: A review of the evidence. Clinical Psychology Review, 27, 226–239. Klonsky, E. D. (2011). Non-suicidal self-injury in United States adults: Prevalence, sociodemographics, topography, and functions. Psychological Medicine, 41, 1981–1986. Klonsky, E. D., May, A. M., & Glenn, C. R. (2013). The relationship between non suicidal self-injury and attempted suicide: Converging evidence from four samples. Journal of Abnormal Psychology, 122(1), 231–237. Klonsky, E. D., & Olino, T. M. (2008). Identifying clinically distinct subgroups of self-injurers among young adults: A latent class analysis. Journal of Consulting and Clinical Psychology, 76, 22–27. Laghi, F., Cotugno, A., Cecere, F., Sirolli, A., Palazzoni, D., & Bosco, F. M. (2014). An exploratory assessment of theory of mind and psychological impairment in patients with bulimia nervosa. British Journal of Psychology, 105, 509–523. Laye-Gindhu, A., & Schonert-Reichl, K. A. (2005). Nonsuicidal self-harm among community adolescents: Understanding the ‘‘whats” and ‘‘whys” of selfharm. Journal of Youth and Adolscence, 34, 447–457. Lonigro, A., Laghi, F., Baiocco, R., & Baumgartner, E. (2014). Mind reading and empathy: Evidence for nice and nasty ToM behaviours in school-aged children. Journal of Child and Family Studies, 23, 581–590. Lundh, L. G., Karim, J., & Quilisch, E. (2007). Deliberate self-harm in 15-year-old adolescents: A pilot study with a modified version of the Deliberate SelfHarm Inventory. Scandinavian Journal of Psychology, 48, 33–41. Manca, M., Presaghi, F., & Cerutti, R. (2014). Clinical specificity of acute versus chronic self-injury: Measurement and evaluation of Repetitive Non Suicidal Self-Injury. Psychiatry Research, 215, 111–119. http://dx.doi.org/10.1016/j.psychres.2013.10.010. Muehlenkamp, J. J. (2005). Self-injurious behavior as a separate clinical syndrome. American Journal of Orthopsychiatry, 75, 324–333. Muehlenkamp, J. J., Claes, L., Havertape, L., & Plener, P. L. (2012). International pre- valence of adolescents non-suicidal self-injury and deliberate self-harm. Child Adolescent Psychiatry Mental Health, 6(10), 2000. http://dx.doi.org/10.1186/1753–6-10. Muehlenkamp, J. J., Claes, L., Peat, C., Smits, D., & Vandereycken, W. (2011). Non-suicidal self-injury in eating disordered patients: A test of a conceptual model. Psychiatry Research, 188, 102–108. Nichols, S., & Stich, S. P. (2003). Mindreading. An integrated account of pretence, self-awareness, and understanding other minds.Oxford: Clarendon Press. Nock, M. K. (2010). Self-Injury. Annual Review of Clinical Psychology, 6, 339–363. Nock, M. K., Joiner, T. E., Gordon, K. H., Lloyd-Richardson, E., & Prinstein, M. J. (2006). Non-suicidal self-injury among adolescents: Diagnostic correlates and relation to suicide attempts. Psychiatry Research, 144, 65–72. Nock, M. K., & Prinstein, M. J. (2004). A functional approach to the assessment of self-mutilative behavior. Journal of Consulting and Clinical Psychology, 72, 885–890. Orbach, I., Milstein, I., Elizur, A., Apter, A., Tiano, S., & Elizur, A. A. (1991). Multi-attitude suicide tendency scale for adolescents: Psychological assessment. Journal of Consulting and Clinical Psychology, 3, 398–404. Osman, A., Gilpin, A. R., Panak, W. F., Kopper, B. A., Barrios, F. X., Gutierrez, P. M., & Chiros, C. E. (2000). The multi attitude suicide tendency scale: further validation with adolescent psychiatric inpatients. Suicide Life Threat Behavior, 30(4), 377–385. Premack, D., & Woodruff, G. (1978). Does the chimpanzee have a theory of mind? Behavioral and Brain Sciences, 1, 512–526. Rossouw, T. I., & Fonagy, P. (2012). Mentalization-based treatment for self-harm in adolescents: A randomized controlled trial. Journal of the American Academy of Child & Adolescent Psychiatry, 12, 1304–1313. Scherzer, P., Leveillé, E., Achim, A., Boisseau, E., & Stip, E. (2012). A study of theory of mind in paranoid schizophrenia: A theory or many theories? Frontiers in Psychology, 14(3), 432. http://dx.doi.org/10.3389/fpsyg.2012.00432. Scott, L. N., Pilkonis, P. A., Hipwell, A. E., Keenan, K., & Stepp, S. D. (2015). Non-suicidal self-injury and suicidal ideation as predictors of suicide attempts in adolescent girls: A multi-wave prospective study. Comprehensive Psychiatry, 58, 1–10. Selby, E. A., Bender, T. W., Gordon, K. H., Nock, M. K., & Joiner, T. E. Jr., (2012). Non-suicidal self-injury (NSSI) disorder: A preliminary study. Personality Disorders: Theory, Research, and Treatment, 3, 167–175. Sharp, C., Pane, H., Ha, C., Venta, A., Patel, A. B., Sturek, J., & Fonagy, P. (2011). Theory of mind and emotion regulation difficulties in adolescents with borderline traits. Journal of the American Academy of Child & Adolescent Psychiatry, 50(6), 563–573. Sivaratnam, C. S., Cornish, K., Gray, K. M., Howlin, P., & Rinehart, N. J. (2012). Brief report: Assessment of the social-emotional profile in children with autism spectrum disorders using a novel comic strip task. Journal of Autism and Developmental Disorders, 42(11), 2505–2512. F. Laghi et al. / Consciousness and Cognition 43 (2016) 38–47 47 Tirassa, M., Bosco, F. M., & Colle, L. (2006a). Rethinking the ontogeny of mindreading. Consciousness and Cognition, 15, 197–217. Tirassa, M., Bosco, F. M., & Colle, L. (2006b). Sharedness and privateness in human early social life. Cognitive Systems Research, 7, 128–139. Vogeley, K., & Fink, G. R. (2003). Neural correlates of the first-person perspective. Trends in Cognitive Sciences, 7, 38–42. Whitlock, J., Muehlenkamp, J., Eckenrode, J., Purington, A., Abrams, G. B., Barreira, P., & Kress, V. (2012). Nonsuicidal self-injury as a gateway to suicide in young adults. Journal of Adolescent Health, 52(4), 486–492. Whitlock, J. L., & Selekman, M. (2014). Non-suicidal self-injury (NSSI) across the lifespan. In M. Nock (Ed.). Oxford handbook of suicide and self-injury. Oxford Library of Psychology. Oxford University Press. Wilkinson, P. (2013). Non-suicidal self-injury. European Child and Adolescent Psychiatry, 22(1), 75–79. Wimmer, H., & Perner, J. (1983). Beliefs about beliefs: Representation and constraining function of wrong beliefs in young children’s understanding of deception. Cognition, 13, 103–128. http://dx.doi.org/10.1016/0010-0277(83)90004-5. Zetterqvist, M., Lundh, L. G., Dahlström, O., & Svedin, C. G. (2013). Prevalence and function of non-suicidal self-injury (NSSI) in a community sample of adolescents, using suggested DSM-5 criteria for a potential NSSI disorder. Journal of Abnormal Child Psychology, 41(5), 759–773.

Consciousness and Cognition 19 (2010) 1095–1096 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Reply The threshold of wakefulness, the experience of control, and theory development q Timothy Lane a, Chien-Ming Yang a,b,* a b The Research Center for Mind, Brain, and Learning, National Chengchi University, Taipei, Taiwan Department of Psychology, National Chengchi University, Taipei, Taiwan a r t i c l e i n f o a b s t r a c t Keywords: Sleep onset Control Subjective experience Perception of sleep Ó 2010 Elsevier Inc. All rights reserved. We are very grateful to Professor Wackermann for his constructive and insightful comments. We take it that one among his main concerns is the possibility of variation that might go undetected due to our choice of methodology. For example, when inquiring as to the logicality or coherence of thoughts, we seem to be presupposing a shared internal norm that can be accurately reflected, despite multiple, mediating steps of memory and evaluation. Professor Wackermann’s concern is by no means an idle one. But (1) we believe that queries of the sort employed here are necessary if we are to begin making progress in eliciting the structure of conscious experience. Were we to limit ourselves to questions that concern just raw, sensory experience, we would risk arbitrarily ignoring significant aspects of the subjects’ phenomenology. Further, (2) our statistical analysis is indeed intended to balance out individual differences. Although this strategy does risk obscuring important variation, it can still be helpful in identifying significant, albeit not universal, indicators. Finally, (3) while it is possible that the transition from wakefulness to sleep follows different paths, the issue is an empirical one. Just as it would be unwise to arbitrarily ignore individual variation, so too would it be unwise to arbitrarily inflate individual variation. We realize that Professor Wackermann’s concerns though are not mere methodological quibbles as regards how best to address a single psychological phenomenon. As Wackermann (2006) lucidly expresses elsewhere, he seeks to develop a strategy for discovering universal laws that is compatible with the study of entities that exhibit great variation, human beings. Indeed, we are in sympathy with his view that more attention should be given to what he terms the ‘‘idiomatic” regularities. On this view, research should proceed in a two-step fashion: first, one should attend to intra-individual regularities and render these in logical or mathematical form. Only after this step has been completed should one seek inter-individual comparisons. As regards the research that actuated Professor Wackermann’s critique, we are not able to present results in such a way that they would satisfy strict standards for ‘‘distributed nomothesis.” But, motivated by this strategy, we have re-evaluated the data, attending more carefully to individual variation. In so doing we found that, for most subjects, ratings on more than one item were associated with the perception of falling asleep. Moreover, for ten of the twenty subjects, ‘‘control over think- q Reply to Commentary by Wackermann on Yang, C.-M., Han, H.Y., Yang, M.H., Su, W.C., & Lane, T. (2010). What subjective experiences determine the perception of falling asleep during sleep onset period? Consciousness and Cognition 19, 1084–1092. * Corresponding author. Address: Department of Psychology, National Chengchi University, 64, Sec. 2, Chih-Nan Rd., Taipei 116, Taiwan. E-mail address: dryangcm@gmail.com (C.-M. Yang). 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.05.015 1096 T. Lane, C.-M. Yang / Consciousness and Cognition 19 (2010) 1095–1096 ing process” was associated with the perception of falling asleep; for eight subjects, ‘‘control over perception”; and, for seven, ‘‘thinking experience,” ‘‘logic of thinking process,” and ‘‘orientation.” This reanalysis suggests that the experience of control might be a key factor in the subjective experience of sleep onset. Not only is it cited explicitly with reference to thought process and perception, it seems to be implied by those who indicate that their thoughts were not logical. Speculating, perhaps further investigations of the relevant cognitive processes would reveal that one among the significant idiomatic regularities related to this transitional state involves diminution of the sense of control. Presumably the relevant meaning of control here does not concern the obsessive or intrusive thoughts that are commonly associated with insomnia—after all, the reported experiences are regarded by the subjects as indications of sleep onset. It would seem to be far more likely that the relevant sense of control bears greater similarity to the thought insertions experienced by schizophrenics, what Frith (1992) refers to as ‘‘passivity experiences.” Frith’s account might also help to explain the association with ‘‘control over perception,” as his model emphasizes our capacity to distinguish between changes in our perception of the external world that result from our own actions and changes that result from alterations in the external environment itself. In schizophrenics this ability is impaired, an impairment that Frith attributes to a failure to monitor intentions. Inability to monitor intentions might well be experienced as an inability to control perception. Naturally we do not intend to suggest that sleep onset and schizophrenia are one and the same. Clearly the two differ in many respects. But exploration of the nature and degree of difference might well lead to significant insights. Such exploratory work would, we believe, be consistent with Wackermann’s (2006) view that science should be dedicated to the search for a ‘‘beautiful linking of facts.” He believes that too much experimental work is nothing more than ‘‘a game played by its own rules on an isolated playground.” He advocates regarding experimental work as ‘‘materialized reasoning”: that is, experiments should be motivated by careful theory development that is relatively independent of particular databases. One goal of such development should be a ‘‘beautiful linking of facts,” where previously there had only been a disconnected jumble. As a very preliminary step in the direction of finding pattern amid jumble, recent research into control of action and goal maintenance suggests a separate, but arguably relevant domain. For example, Suhler and Churchland (2009) have proposed a neurobiological model of control that is applicable both to quotidian states wherein control is exercised (e.g. getting out of a warm bed on a cold morning) and to prototypical cases wherein persons feel ‘‘out of control” (e.g. addiction). They propose a model of multiple parameters that includes neurochemicals, connectivity among brain structures, and so forth. The intent is to identify an ‘‘in control” region within multi-dimensional space, a space that reflects the likelihood that there are many different ways of being in, or out, of control. Were we to further develop such a model, because the prefrontal cortex is implicated in self-monitoring and in goaldirected thought and because it is relatively inactive during NREM sleep (Muzur, Pace-Schott, & Hobson, 2002), we would likely include it as one among several parameters that needs to be highlighted. Nevertheless, we are keenly aware that our brief discussion here merely gestures in the direction of one possible line of inquiry. But an especially attractive feature of such theorizing is that it allows for the possibility of mathematical modeling in a way that can accommodate ‘‘idiomatic” regularities: that is, it can account for different ways of being in and out of control. Of course whether or not thinking along the lines adumbrated here will yield fruitful results, we cannot say. But we are grateful for Professor Wackermann’s gentle encouragement to search for the idiomatic and to take seriously the role of theorizing. References Frith, C. (1992). The cognitive neurobiology of schizophrenia. New Jersey: Lawrence Erlbaum Associates. Muzur, A., Pace-Schott, E. F., & Hobson, J. A. (2002). The prefrontal cortex in sleep. Trends in Cognitive Sciences, 6, 475–481. Suhler, C., & Churchland, P. S. (2009). Control: Conscious and otherwise. Trends in Cognitive Sciences, 13, 341–347. Wackermann, J. (2006). Rationality, universality, and individuality in a functional conception of theory. International Journal of Psychophysiology, 62, 411–426. Yang, C.-M., Han, H. Y., Yang, M. H., Su, W. C., & Lane, T. (2010). What subjective experiences determine the perception of falling asleep during sleep onset period? Consciousness and Cognition, 19, 1084–1092.

Consciousness and Cognition 42 (2016) 9–14 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Minding gaps on the skin: Opposite bisection biases on forehead and back of one’s head Bigna Lenggenhager a,b,⇑, Christine Busch a, Peter Brugger a,b a b Neuropsychology Unit, Department of Neurology, University Hospital Zurich, Zurich, Switzerland Center for Integrative Human Physiology (ZIHP), University of Zurich, Zurich, Switzerland a r t i c l e i n f o Article history: Received 11 January 2016 Revised 1 March 2016 Accepted 2 March 2016 Keywords: Perspective Embodiment Line bisection Cutaneous perception Body space Pseudoneglect Visual–somatosensory interactions a b s t r a c t Humans perceive the world from an egocentric perspective, while being able to mentally take a third person’s perspective. Graphesthesia tasks revealed that letters written on the back of one’s own head are consistently perceived from an embodied perspective, while the perspective on one’s front is less consistent and often disembodied. We developed a cutaneous gap bisection task as a more discrete measure of the perspective on the body. In analogy to a visual pseudoneglect, we expected bisections to deviate toward the left ear when perceived from an embodied perspective. While this hypothesis was confirmed for gap bisections on the back, the results on the front suggest overall a disembodied perspective. Contrary to our expectation, this pattern was not predicted by the spontaneous perspective participants took in a graphesthesia task, indicating different cognitive mechanisms. We discuss these findings in the frame of the current literature on spatial attention and perspective taking. Ó 2016 Published by Elsevier Inc. 1. Introduction A central aspect of human self-consciousness is our stable embodied first person perspective on the world (e.g. Blanke, 2012). In contrast, a central aspect of human social life is the ability to mentally simulate another person’s perspective (e.g. Costantini, Committeri, & Sinigaglia, 2011) albeit it is hotly debated to what extent such process is automatic or rather effortful (e.g. Arnold, Spence, & Auvray, 2016; Cole, Atkinson, Le, & Smith, 2016). Such dyadic perspective exists not only for the space around us, but also for our own body and self, linked to the perception of the self as a subject (‘‘I”) or as an object (‘‘me”). One task that has widely been used to investigate the spontaneous perspective on one’s own body surface is a graphesthesia task (Natsoulas & Dubanoski, 1964), in which letters or numbers are drawn on body parts of a blindfolded participant, who is required to identify the orientation of the stimulus. During this task the letter ‘‘b” drawn by an experimenter on the front, for example, can either be perceived as a ‘‘d” (embodied perspective) or as a ‘‘b” (disembodied perspective). An early, but thorough investigation by Stracke (1947) suggested that signs drawn on the back are perceived as they are drawn by the experimenter (i.e. from an embodied perspective), while those on the front are rather perceived as mirror-inverted, i.e. as if the participant ‘‘looked” through the head on the number (”frontal plane hypothesis” (Duke, 1966)), suggesting an embodied perspective as well. Yet, already Stracke (1947) noticed that participants were slower and less consistent in their responses ⇑ Corresponding author at: Neuropsychology Unit, Department of Neurology, University Hospital Zurich, CH-8091 Zurich, Switzerland. E-mail address: bigna.lenggenhager@gmail.com (B. Lenggenhager). http://dx.doi.org/10.1016/j.concog.2016.03.001 1053-8100/Ó 2016 Published by Elsevier Inc. 10 B. Lenggenhager et al. / Consciousness and Cognition 42 (2016) 9–14 for drawings on the front than an on the back. And while the results on the back of the head are consistent across subjects and studies (Duke, 1966; Parsons, 1987; Stracke, 1947), the results on front stimulation are much less consistent and have been shown to depend for example on the participant’s gender (Duke, 1966), body posture (Stracke, 1947), on the currently available vestibular input (Ferrè, Lopez, & Haggard, 2014) and, in a slight task modification, on the participant’s momentary self-focus (Hass, 1984). This suggests that in graphesthetic perception on body parts with which we typically face other people in social interactions (i.e. front face), we more readily take the observer’s perspective, while both situational and inter-personal aspects might play a role to what extent we spontaneously do so. There are, however, two important limitations to this oftenused task: first, the participant’s responses allow only a dichotomous classification into perception as mirror-inversed (embodied perspective) or not (disembodied perspective), not allowing for any intermediate state. Second, the task involves writing and reading, which is a special situation that strongly involves communicative and social aspects, which are mostly not controlled for (see e.g. Arnold et al., 2016; Hass, 1984 for exceptions). Participants might thus more or less explicitly take into account the experimenter’s perspective to solve the task. Here, we used an alternative, purely spatial task with a continuous measure to investigate the perspective on the own body. We used a cutaneous gap bisection task on healthy participants’ front and back, which required them to bisect the empty space between two endpoints of a line (Bradshaw, Bradshaw, Nathan, Nettleton, & Wilson, 1986). In the visual domain, it has been shown that in healthy, right-handed participants, bisections of a visible line from a first person perspective are slightly, but consistently shifted toward the left of the objective midpoint, a phenomenon termed ‘‘pseudoneglect” (Bowers & Heilman, 1980). Bisecting the empty space between two visually presented endpoints of a line (‘‘gap bisection”) leads to similar, albeit smaller (left-sided) lateral displacements (Bradshaw et al., 1986; McIntosh, McClements, Dijkerman, & Milner, 2004). Whether a comparable pseudoneglect exists in the tactile domain as well (bisection of felt distances on the skin) has been studied much less systematically. Some previous investigations described a leftward shift of the subjective meridian when participants were asked to point to the body midline (Richard, Honoré, & Rousseaux, 2000; Spidalieri & Sgolastra, 1997). Spidalieri and Sgolastra (1999) found a leftward displacement of the anterior head midline, yet they did not test midline pointing on the back of the head. A proper characterization of pseudoneglect in the somatosensory modality has, however, never been provided, and was one of the aims of the present study. We investigated healthy subjects’ pointing biases to the spot laying halfway between two touch stimuli applied simultaneously to either the forehead or, in separate trials, to the back of the head. We generally assumed an existence of a pseudoneglect in the somatosensory modality. Thus, we predicted a leftward deviation relative to the actual midpoint on the back (i.e. a pseudoneglect from an embodied perspective, see Fig. 1). For the front we predicted that participants using spontaneously a disembodied perspective in a graphesthesia task, which requires writing letters on one’s forehead (Shimojo et al., 1989; Hass, 1984), would show a rightward bias, while those using an embodied perspective should also show a leftwards bias. 2. Methods 2.1. Participants The 40 subjects (20 women) had a mean age of 37.5 years (SD = 6.2 years) and an average educational training of 14.8 years (SD = 1.4 years). All subjects were right-handed (Chapman and Chapman, 1987), and none had a history of Fig. 1. Schematic illustration of the different perspectives and the predicted deviation. Predicted deviation from the midline if a somatosensory pseudoneglect exists for the different perspectives on ones own body described in the literature. EP = embodied perspective, DP = disembodied perspective. While graphesthesia tasks on the back are consistently done from a EP perspective, those on the front might either be done from an EP or a DP. B. Lenggenhager et al. / Consciousness and Cognition 42 (2016) 9–14 11 neurological or psychiatric diseases, or of any developmental disorders or substance abuse (Campbell, 2000). Subjects were not reimbursed for their participation and written informed consent was obtained. The study had been done adhering to the conditions laid down in the Declaration of Helsinki. 2.2. Procedure After the assessment of handedness and medical history, the first skin drawing test was administered. This required the writing of the letters A, B, and C (in this order) on a slip of paper (10.5 cm  7.5 cm) to be held by participants with their left hand on one’s front. It was stressed that handwriting quality would not be judged. In fact, it was only recorded whether the writing was left-to-right oriented or mirror-reversed. During the subsequent 20–30 min, subjects solved a computer task unrelated to the present context. They were then administered the somatosensory gap bisection task (see below), followed by a traditional line bisection task in which the midpoint of 12 lines displayed on a A4-sized sheet (lengths 130 mm, 175 mm and 235 mm, respectively) had to be marked with a pencil held in the right hand. The second skin drawing test, identical to the first in all respects, concluded testing. 2.3. Somatosensory gap bisection Subjects were tested in a comfortable sitting positions with their eyes closed. During forehead testing, the examiner softly pressed two metallic styluses (diameter of tips 1 mm) to the skin of the subject’s forehead. These were fixed to a ruler that, held in between the two styluses, guaranteed an equal and constant pressure at the two stimulation points. Three rulers were used, each with a different distance between the two styluses, i.e 50 mm, 70 mm and 90 mm. Styluses protruded 20 mm from the two small-distance rulers and 40 mm from the long-distance ruler. The subject used a metallic rod (length 10 cm, diameter of tip 1 mm) to point to the subjective midpoint between the two felt stimuli with the right hand. There was no time constraint, but subjects were trained in a practice run to respond at a rate that allowed to keep the intertrial interval between 15 and 25 s (Spidalieri & Sgolastra, 1997). Once the skin was touched by the response stylus, subjects were not allowed to reallocate its position, and the deviation from the actual midpoint between the two styluses was measured by the examiner to the nearest mm. To prevent the actual midpoint from being located at similar spots on the skin on consecutive trials and to avoid that it simply coincided with the head midline, a record sheet indicated where to place the ruler, i.e the left stylus had to be aligned with either the outer rim of the left eye, its inner rim, or a point that corresponded as closely as possible with the midpoint of the eye. Fourteen trials were thus administered, locations of the rulers and distances between the styluses alternating pseudorandomly (in an identical sequence for all subjects). During gap bisections on the back of subjects’ head, ruler locations were matched to those used during forehead testing by placing a paper ribbon tightly around the subject’s head on which the eyes were drawn above the real eyes, reflecting an individual’s eye size and distance between the eyes (Fig. 2). By rotating the ribbon 180°, the examiner could apply an analogously randomized stimulation sequence (14 trials) on the back of the head. Response conditions were equal to those described for forehead testing. Half of the subjects of each gender group were tested first on the forehead, the other half first on the back of their head. While physiological constrains for the pointing movement, tactile acuity, or skin properties (e.g. hairiness) are likely to differ between the front and the back of the head, we did not systematically address or correct for these potential confounds. The reason for this was, that we did not expect such mechanisms to bias the participants’ judgments in any systematic way (e.g. always to the left, thus increasing a pseudo-neglect like behavior). 3. Results 3.1. Skin writing task 15 subjects (9 women) produced mirror writing in the skin-writing test on both occasions (‘‘embodied eye”), 22 (10 women) consistently produced a regular writing (‘‘disembodied eye”). The data of the three subjects with an inconsistent skin writing habit (2 subjects ‘‘disembodied” on first, ‘‘embodied” on second testing, a third one first ‘‘embodied” then ‘‘disembodied”) were not analyzed. 3.2. Cutaneous gap bisection Averaged over all gaps, the mean deviations from the midline were 3.54 mm (SEM = 0.7) to the right on the front and 2.69 mm (SEM = 1.0) to the left on the back, both of which were significantly different from zero as indicated by a one sample t-test (two-tailed, front: t = 4.99, p < 0.001, back t = 2.62, p = 0.013). For the ANOVA, deviations were calculated as percentages of the respective gap width; positive values suggest a deviation toward the participant’s right ear, negative toward the participant’s left ear. We calculated an ANOVA with the between subject factor GROUP (dis/embodied eye group) and the within-factors LOCATION (forehead vs. back of head) and GAP WIDTH (50 mm, 70 mm, 99 mm). This analysis produced a significant main effect of LOCATION (F(1, 35) = 44.32, p < 0.001) with 12 B. Lenggenhager et al. / Consciousness and Cognition 42 (2016) 9–14 Fig. 2. Results of the somatosensory gab bisection task. Mean displacements (with standard errors of the mean) toward the right ear in somatosensory gap bisections on the forehead (left) and on the back of one’s head (right) for the three gap widths (short = 50 mm, medium = 70 mm, long = 90 mm). Dark gray are the participants with a disembodied, light gray with an embodied perspective according to the skin writing task. positive values (i.e. a pseudoneglect as if seen from a disembodied perspective) in the front and negative values (i.e. a pseudoneglect as seen from an embodied perspective) in the back. It further revealed a main effect of GAP WIDTH (F (1.14, 35) = 10.0, p = 0.01) for which Sidak post-hoc comparisons suggest a stronger leftwards deviation in the short as compared to the medium (p = 0.004) and the long (p = 0.007) gap width. For the interaction effects, the interaction of GAP WIDTH and LOCATION proved significant (F(1.60, 34) = 4.26, p = 0.026) and so did the triple interaction of GROUP, GAP WITH and LOCATION (F(1.60, 34) = 4.17, p = 0.028, corrected for lack of sphericity using a Greenhouse Geisser correction). Fig. 1 shows this 3-way interaction. Posthoc t-test between participants (1-tailed) revealed a significant difference in the predicted direction between the disembodied and the embodied group only in the long gap width (p = 0.035). In order to investigate the relation between gap bisections on the back and on the front, a mean % deviation on the front and on the back were calculated and correlated using a Pearson correlation. The results show a highly significant negative correlation (r = 0.44, p = 0.007). 3.3. Visual line bisection Mean deviation in traditional line bisections was 0.53 mm (SEM = 0.45) to the left of the objective midpoint for the whole group. A one sample t-test suggests no significant difference to zero (t = 1.18, p = 0.25). 4. Discussion This study revealed two main findings. First, and as expected, the results show a systematic bias in a gap bisection task, which speaks for a pseudoneglect in the somatosensory system as it has previously been described for the visual system. Second, if the systematic deviation from the midpoint is taken as an index of a pseudoneglect, it suggest that participants generally spontaneously took a disembodied perspective during the gap bisection task on the forehead while they used an embodied perspective for gap bisection on the back of their head. This effect depended slightly but far less than expected on the spontaneous perspective taken during a graphesthesia task (writing letters on one’s forehead). 4.1. The existence of a somatosensory pseudo-neglect It seems fairly simple to point to the middle of a visually inspected line, or a haptically explored rod or even to bisect the gap between two points. Yet, in healthy, right-handed subjects, line bisections are slightly but consistently shifted toward the left of the objective midpoint. Interpreted as a relative underestimation of the right side of space, this phenomenon was dubbed ‘‘pseudoneglect” by Bowers and Heilman (1980). A similar, albeit smaller effect was found for gap bisection (Bradshaw et al., 1986). This effect is commonly explained by a right hemisphere dominance in the orienting of spatial B. Lenggenhager et al. / Consciousness and Cognition 42 (2016) 9–14 13 attention (e.g. Làdavas, Del Pesce, & Provinciali, 1989), but the exact causes are still poorly understood (Longo, Trippier, Vagnoni, & Lourenco, 2015). The present study was the first to test whether a comparable effect exists in the somatosensory domain when bisecting a spatial extension on one’s own body. Previous related studies have tested the participant’s ability to point to the trunk midline, and the majority of publications have described a leftward shift of the subjective meridian (e.g. Spidalieri & Sgolastra, 1999, but see e.g. Spidalieri & Sgolastra, 2001 for a failed replication). Importantly, these studies tested the midline pointing only on the front and never on the back of the participant’s head. This might be problematic, given the fact that many participants seem to spontaneously take a disembodied perspective on their own body’s front (e.g. Parsons, 1987). The present results suggest a clear leftward bias in the gap bisection task on the back, while the opposite was found for the front. Building on literature using the graphesthetic task (Stracke, 1947), the former finding (see Section 4.2 for the discussion of the latter) suggests that participants may take an embodied perspective when observing their back with their mind’s eye, hence they display a pseudoneglect. In analogy to explanations of pseudoneglect in the visual modality, this result could be explained by assuming a right-hemispheric dominance in spatial attention (e.g. Làdavas et al., 1989) that also comprises visual imagery, possibly automatically triggered by touch on the back of one’s body. In the visual domain, a right hemispheric dominance of near space processing might further strengthen the pseudoneglect, which is evidenced by various studies showing increased leftwards deviations with decreasing distance between the line and the observer (e.g. Longo et al., 2015). In this context, it should be noted, that in the current sample of participants we did not find a pseudoneglect in the classical visual line bisection task. While this could be due to methodological issues (e.g. less experimental trials in the visual task), it could also be speculated that the additional right-hemispheric dominance in body – and especially own body – processing (e.g. Frassinetti, Maini, Romualdi, Galante, & Avanzi, 2008) could have increased the pseudoneglect in the somatosensory as compared to the visual task. Such interpretation of our results might be somewhat in conflict with recent data suggesting a more accurate bisection of body parts compared to objects when presented as visual stimuli in peripersonal space (Bolognini, Casanova, Maravita, & Vallar, 2012; Sposito, Bolognini, Vallar, Posteraro, & Maravita, 2010). Yet, there are too many differences between those experimental setups and the one used here to directly compare the data. Further studies will have to systematically investigate visual as compared to somatosensory bisection of objects and body parts both in the front- and back-space and with various distances to the body. 4.2. Embodied perspective on one’s back and disembodied perspective on one’s front Surprisingly, when asked to bisect a gap of two cutaneous stimuli on the front, participants consistently deviate to their right ear, which would suggest, that they either show a left-sided inattention from the embodied perspective or that they take a disembodied, third-person perspective on their forehead for solving the task and thus a pseudoneglect from that perspective. Albeit currently rather speculative, we judge the latter option as more plausible, as it is further strengthened by the fact that deviations on the back where highly negatively correlated with those on the front. This finding suggests that those participants showing a large pseudoneglect from an embodied perspective when solving the task on the back also show a large pseudoneglect when solving the task on the front from a disembodied perspective. Previous studies using the graphesthesia task vary in the percentage of people using an embodied versus a disembodied perspective for judgments on the front: Most studies found a large majority of participants to spontaneously employ an embodied perspective (Corcoran, 1977; Duke, 1966; Stracke, 1947) while others described a slight predominance of a disembodied perspective (e.g. Ferrè et al., 2014), yet the reason for the interindividual differences are largely unknown. Our data from the letter-writing task showed that about 60% of the participants used a disembodied perspective to solve the task. Against our hypothesis, however, participants who took a third person perspective in the skin drawing test did not generally show a stronger deviation toward the right ear in the gap bisection task (although our respective prediction was confirmed for longest gaps) than participants using an embodied perspective. This suggests that different mechanisms are involved in letter drawing (or letter recognition) and the bisection of experimenter-applied cutaneous distances. Importantly, the perspective spontaneously taken on one’s own body depends not only on the individual but also strongly on the task and the situation (see also (Brugger, 2002) for a discussion on the perception on one’s own body in clinical cases). While it could be expected that the gap bisection task, as a rather non-social task, would be less likely to trigger a disembodied perspective, our data rather suggest the opposite: they suggests that the gap bisection task on the front is generally solved from a third person perspective. Such spontaneous third person perspective taking (i.e., altercentric intrusion) has previously been found in other tasks, most prominently probably in studies looking at reaction times and error rates on perspective taking tasks when a third person with a different view-point is present or absent in a scene (e.g. Kragh Nielsen, Slade, Levy, & Holmes, 2015). Yet, these results and their conclusions are still debated (see e.g. Cole et al., 2016). While we are confident that the gap-bisection task introduced here might importantly add to the literature on perspective taking and its neural and social determinants, future studies should more carefully address the influence of methodological detail. For example, the mere presence of an experimenter in front, who touches the participant, could have affected task performance (see Hass, 1984 for a discussion), as it has been demonstrated in similar visual tasks (e.g. Tversky & Hard, 2009). To conclude, our new experimental paradigm suggests the presence of a tactile pseudoneglect on the surface of one’s own body. Cutaneous gap bisections may help to close the gap between a ‘‘personal geography” (Corcoran, 1977) determined by bodily landmarks and the orientation and navigation in external, interpersonal space. Our skin, long recognized as a social organ (Morrison, Löken, & Olausson, 2009), may hold the key to expand current perspectives on perspective taking: automatic social and empathic perspective taking may depend more on attentional and sensory factors than hitherto assumed. 14 B. Lenggenhager et al. / Consciousness and Cognition 42 (2016) 9–14 References Arnold, G., Spence, C., & Auvray, M. (2016). Taking someone else’s spatial perspective: Natural stance or effortful decentring? Cognition, 148, 27–33. http:// dx.doi.org/10.1016/j.cognition.2015.12.006. Blanke, O. (2012). Multisensory brain mechanisms of bodily self-consciousness. Nature Reviews Neuroscience, 13(8), 556–571. http://dx.doi.org/10.1038/ nrn3292. Bolognini, N., Casanova, D., Maravita, A., & Vallar, G. (2012). Bisecting real and fake body parts: Effects of prism adaptation after right brain damage. Frontiers in Human Neuroscience, 6. http://dx.doi.org/10.3389/fnhum.2012.00154. Bowers, D., & Heilman, K. M. (1980). Pseudoneglect: Effects of hemispace on a tactile line bisection task. Neuropsychologia, 18(4–5), 491–498. Bradshaw, J. L., Bradshaw, J. A., Nathan, G., Nettleton, N. C., & Wilson, L. E. (1986). Leftwards error in bisecting the gap between two points: Stimulus quality and hand effects. Neuropsychologia, 24(6), 849–855. http://dx.doi.org/10.1016/0028-3932(86)90084-9. Brugger, P. (2002). Reflective mirrors: Perspective taking in autoscopic phenomena. Cognitive Neuropsychiatry, 7, 179–194. Campbell, J. J. (2000). Neuropsychiatric assessment. In C. E. Coffey & J. L. Cummings (Eds.), Textbook of Geriatric Neuropsychiatry (2nd ed., pp. 109–124). Washington, DC: American Psychiatric Press. Chapman, L. J., & Chapman, J. P. (1987). The measurement of handedness. Brain and Cognition, 6, 175–183. Cole, G. G., Atkinson, M., Le, A. T. D., & Smith, D. T. (2016). Do humans spontaneously take the perspective of others? Acta Psychologica, 164, 165–168. http:// dx.doi.org/10.1016/j.actpsy.2016.01.007. Corcoran, D. W. (1977). The phenomena of the disembodied eye or is it a matter of personal geography? Perception, 6(3), 247–253. Costantini, M., Committeri, G., & Sinigaglia, C. (2011). Ready both to your and to my hands: Mapping the action space of others. PLoS ONE, 6(4), e17923. http://dx.doi.org/10.1371/journal.pone.0017923. Duke, J. D. (1966). Perception of finger drawings upon the body surface. The Journal of General Psychology, 75(2d Half), 305–314. http://dx.doi.org/10.1080/ 00221309.1966.9710377. Ferrè, E. R., Lopez, C., & Haggard, P. (2014). Anchoring the self to the body: Vestibular contribution to the sense of self. Psychological Science, 25(11), 2106–2108. http://dx.doi.org/10.1177/0956797614547917. Frassinetti, F., Maini, M., Romualdi, S., Galante, E., & Avanzi, S. (2008). Is it mine? Hemispheric asymmetries in corporeal self-recognition. Journal of Cognitive Neuroscience, 20(8), 1507–1516. http://dx.doi.org/10.1162/jocn.2008.20067. Hass, R. G. (1984). Perspective taking and self-awareness: Drawing an E on your forehead. Journal of Personality and Social Psychology, 46(4), 788–798. http:// dx.doi.org/10.1037/0022-3514.46.4.788. Kragh Nielsen, M., Slade, L., Levy, J. P., & Holmes, A. (2015). Inclined to see it your way: Do altercentric intrusion effects in visual perspective taking reflect an intrinsically social process? Quarterly Journal of Experimental Psychology (2006), 68(10), 1931–1951. http://dx.doi.org/10.1080/17470218.2015.1023206. Làdavas, E., Del Pesce, M., & Provinciali, L. (1989). Unilateral attention deficits and hemispheric asymmetries in the control of visual attention. Neuropsychologia, 27(3), 353–366. Longo, M. R., Trippier, S., Vagnoni, E., & Lourenco, S. F. (2015). Right hemisphere control of visuospatial attention in near space. Neuropsychologia, 70, 350–357. http://dx.doi.org/10.1016/j.neuropsychologia.2014.10.035. McIntosh, R. D., McClements, K. I., Dijkerman, H. C., & Milner, A. D. (2004). ‘‘Mind the gap”: The size-distance dissociation in visual neglect is a cueing effect. Cortex, 40(2), 339–346. Morrison, I., Löken, L. S., & Olausson, H. (2009). The skin as a social organ. Experimental Brain Research, 204(3), 305–314. http://dx.doi.org/10.1007/s00221009-2007-y. Natsoulas, T., & Dubanoski, R. A. (1964). Inferring the locus and orientation of the perceiver from responses to stimulation of the skin. The American Journal of Psychology, 77(2), 281–285. http://dx.doi.org/10.2307/1420136. Parsons, L. M. (1987). Imagined spatial transformation of one’s body. Journal of Experimental Psychology General, 116, 172–191. Richard, C., Honoré, J., & Rousseaux, M. (2000). Is there a distortion of body projection in extracorporeal space in neglect patients? NeuroReport, 11(13), 3047–3051. Shimojo, S., Sasaki, M., Parsons, L. M., & Torii, S. (1989). Mirror reversal by blind subjects in cutaneous perception and motor production of letters and numbers. Perception and Psychophysics, 45, 145–152. Spidalieri, G., & Sgolastra, R. (1997). Psychophysical properties of the trunk midline. Journal of Neurophysiology, 78(1), 545–549. Spidalieri, G., & Sgolastra, R. (1999). The head midline as a reliable reference frame for encoding head-on-body orientation. NeuroReport, 10(12), 2473–2476. Spidalieri, G., & Sgolastra, R. (2001). Gravitational cues contribute to accurate localisation of mentally represented cutaneous targets. Experimental Brain Research, 138(2), 274–278. Sposito, A. V., Bolognini, N., Vallar, G., Posteraro, L., & Maravita, A. (2010). The spatial encoding of body parts in patients with neglect and neurologically unimpaired participants. Neuropsychologia, 48(1), 334–340. http://dx.doi.org/10.1016/j.neuropsychologia.2009.09.026. Stracke, J. (1947). Über die Wahrnehmung von Körper, Raum und Bewegung durch die Haut. Der Nervenarzt, 10, 433–437. Tversky, B., & Hard, B. M. (2009). Embodied and disembodied cognition: Spatial perspective-taking. Cognition, 110(1), 124–129. http://dx.doi.org/10.1016/j. cognition.2008.10.008.

Consciousness and Cognition Consciousness and Cognition 14 (2005) 327–350 www.elsevier.com/locate/concog Event-related potential evidence for multiple causes of the revelation effectq P. Andrew Leynesa,*, Joshua Landaub, Jessica Walkerc, Richard J. Addantea a Department of Psychology, The College of New Jersey, P.O. Box 7718, Ewing, NJ 08628, USA b Department of Psychology, York College of Pennsylvania, USA c Department of Psychology, University of California, Los Angeles, USA Received 24 June 2004 Available online 26 October 2004 Abstract Asking people to discover the identity of a recognition test probe immediately before making a recognition judgment increases the probability of an old judgment. To inform theories of this ‘‘revelation effect,’’ event-related potentials (ERPs) were recorded for revealed and intact test items across two experiments. In Experiment 1, we used a revelation effect paradigm where half of the test probes were presented as anagrams (i.e., a related task) and the other items were presented intact. The pattern of ERP results from this experiment suggested that revealing an item decreases initial familiarity levels and caused the revealed items to elicit similar levels of activity. In Experiment 2, half of the probes were preceded by an addition task (i.e., an unrelated task). The pattern of ERP effects in this study were distinct from those observed in Experiment 1. More specifically, revealed item ERPs were more negative than intact ERPs at frontal electrodes and more positive at parietal electrodes early in the interval. Later in the epoch, revealed item ERPs were more negative than intact items. These data suggest that related tasks decrease familiarity and alter the signal-to-noise ratio of old and new items, whereas unrelated tasks affect processing in a different way (perhaps by changing decision processes) q Experiment 1 was supported by a Phi Kappa Phi Student-Faculty Research Fellowship granted to Jessica Walker and Andrew Leynes and was presented at the Eighth Annual Meeting of the Cognitive Neuroscience Society, March 2001. Experiment 2 was supported by a Phi Kappa Phi Student-Faculty Research Fellowship granted to Richard Addante and Andrew Leynes. Andrew Leynes was also supported by The College of New JerseyÕs internal research granting program, SOSA. * Corresponding author. Fax: +1 609 637 5178. E-mail address: leynes@tcnj.edu (P.A. Leynes). 1053-8100/$ - see front matter
2004 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2004.08.005 328 P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 that also results in the revelation effect. The implications for current theories of the revelation effect are discussed.
2004 Elsevier Inc. All rights reserved. Keywords: Event-related potentials; Revelation effect; Recognition 1. Introduction In the typical recognition memory experiment, people study a list of items (e.g., words or pictures) and then must identify the studied items on a test that contains the studied items and several new items. Many factors influence the accuracy of recognition judgments including (but not limited to) encoding tasks (Craik & Lockhart, 1972), decision criterion (Green & Swets, 1966), and the composition of the test list (Deese, 1959; Roediger & McDermott, 1995). Recent research demonstrates that another variable, the physical appearance (or initial perception) of the test probe, also affects peopleÕs recognition decisions (Watkins & Peynircioglu, 1990; Whittlesea, Jacoby, & Girard, 1990). For example, when people discover the identity of a test probe before rendering a recognition judgment, they are more likely to call the item ‘‘old’’ as compared with a word presented in its original intact format. This phenomenon, dubbed the revelation effect, is elicited by many tasks. The available evidence shows that people are more likely to call a test probe ‘‘old’’ if it is unfolded one letter at a time (e.g., -T- - - -, -T-E- -, ST-E- -, ST-E-T, STRET, and STREET; LeCompte, 1995; Peynircioglu & Teckan, 1993; Watkins & Peynircioglu, 1990), if the probe is an anagram that must be unscrambled (Peynircioglu & Teckan, 1993; Watkins & Peynircioglu, 1990; Westerman & Green, 1996), if the entire probe or individual letters must be rotated (Peynircioglu & Teckan, 1993; Watkins & Peynircioglu, 1990), if the probe is degraded (Luo, 1993), if an unrelated anagram must be solved before reading the intact probe (Westerman & Green, 1996), and if a numerical addition task must be solved before the test probe (Niewiadomski & Hockley, 2001). One theoretical explanation for the revelation effect draws on the logic of the dual-process theories of recognition (e.g., Jacoby, 1991; Jacoby & Dallas, 1981; Mandler, 1980). By this account, people make a recognition decisions based on two separate pieces of evidence. On the one hand, people can claim to recognize an item if they actually recollect the circumstance and details of their original encounter with the item. On the other hand, recognition may be based solely on a strong feeling that they encountered the item earlier (i.e., the item feels familiar). In terms of the revelation effect, some have suggested that the process of discovering the identity of the test probe temporarily increases its familiarity (Luo, 1993; Peynircioglu & Teckan, 1993) or the global activation (Westerman & Green, 1998). People then mistake this increased familiarity as evidence that the item originated from the study list and this produces more ‘‘old’’ responses for both old items (targets) and new items (lures). Although this increment-to-familiarity account has intuitive appeal and some experimental support, this account fails to explain all of the experimental findings. For example, Westerman and Green (1996) found a revelation effect when the revealed item and the recognition test probe were different words. Similarly, Niewiadomski and Hockley (2001) found a revelation effect when P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 329 participants completed an unrelated nonverbal task (i.e., numerical addition) immediately before judging the test probe. Consequently, the increment-to-familiarity account has difficulty explaining how solving an anagram or solving a nonverbal task increases the familiarity of an unrelated test probe. Another piece of evidence that is difficult to explain with the increment-to-familiarity account comes from Hicks and Marsh (1998) revelation experiments where they examined the format of the recognition test. In their study, after studying a list of words, participants saw two test probes (i.e., a two-alternative forced choice, 2AFC, recognition test) and were instructed to select which one they studied. One of the test probes was revealed, and the other was presented intact. If a temporary increase in familiarity produces the revelation effect, then this paradigm should produce a revelation effect similar to the one where people judged each item separately. However, Hicks and Marsh (1998) found that people were actually less likely to call the revealed items old. To explain the revelation effect, Hicks and Marsh appealed to a signal detection theory (SDT) of recognition memory and suggested that as people try to decode the test probe, this temporarily activates a variety of alternative words in memory. This decoding process decreases the signal-to-noise ratio because the average familiarity values of the old and new words become more similar. In this situation, people relax their decision criteria and this artificially increases the number of hits and false alarms. Niewiadomski and Hockley (2001) proposed the criterion flux account which also posits a criterion shift explanation for the revelation effect, (see also Hockley & Niewiadomski, 2001). According to both the criterion flux and the decrement-to-familiarity (Hicks & Marsh, 1998) accounts, the revelation process causes people to adopt a more liberal decision criterion. The criterion flux account proposes that any task that disrupts working memory processes during the recognition decision process (e.g., a numerical problem or solving an anagram) causes a temporary loss of list information which results in participants adopting a more liberal decision criterion. Conversely, Hicks and Marsh argued that people relax their decision criteria because of a decrease in familiarity (i.e., the signal-to-noise ratio) of the revealed test probes. Thus, both accounts agree that the revelation effect results from a shift in criterion but they differ on the genesis of the criterion shift. Whittlesea and Williams (2001a, 2001b) have proposed an alternative explanation for the revelation effect that we will address in Section 4. Because many of these theoretical explanations make strikingly similar predictions about the revelation effect and they seem to be able to explain the same behavioral data, we believe that cognitive neuroscience may help to uncover the cognitive mechanisms that produce the revelation effect. Specifically, direct measures of brain activity can provide insight into the mental processes that contribute to memory, and, in this case, event-related potentials (ERPs) appear to be particularly well suited to the task. ERPs measure brain activity time-locked to the appearance of stimuli and provide excellent temporal resolution of moment-to-moment changes in cognitive processing. Previous studies indicate that memory processes affect ERP activity. For example, studied items tend to elicit more positive ERP activity than unstudied items (this is sometimes referred to as an old/new ERP effect; see Rugg, 1995 for a review). Researchers have used ERPs to examine predictions made by theories of recognition memory. For example, research indicates that the types of subjective bases described in the dual process theory appear to elicit different patterns of old/new ERP differences (see Curran, 2000 or Mecklinger, 2000 330 P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 for a review). Familiarity is typically observed as an old/new ERP difference at frontal electrode sites that begins approximately 300–500 ms after the onset of the test probe. In contrast, recollection has been linked with an old/new ERP difference at parietal electrode sites approximately 400–600 ms after the onset of the probe. In addition, frontal ERP effects that occur later in the course of memory processing (approximately 800 ms after stimulus onset) seem to reflect more difficult decisions, such as discriminating between different sources (e.g., Leynes, 2002; Leynes, Bink, Marsh, Allen, & May, 2003; Ranganath & Paller, 2000; Wilding & Rugg, 1996). Based on the idea that familiarity and recollection produce different ERPs, Azimian-Faridani and Wilding (in press) recorded ERPs when words were presented intact or preceded by an anagram unrelated to the test probe (i.e., the Westerman & Green, 1996 paradigm). They found that intact item ERPs were more positive than the revealed ERPs 300–500 ms after the onset of the test probe. Because this effect is in the same temporal period as the familiarity related ERP effects, they interpreted this result as evidence that revelation is associated with a decrease in familiarity, supporting the decrement-to-familiarity account (Hicks & Marsh, 1998). Although Azimian-Faridani and WildingÕs data are valuable, there are a number of difficulties with interpreting these findings. First, in their analysis of the ERPs they collapsed across response type and analyzed old and new item ERPs without regard for the subjectÕs response. They claimed this procedure eliminated criterion influences on ERP data because they were not creating ERPs based on responses. Unfortunately, the decision criterion is set for both old and new items; therefore, it is unclear how this procedure eliminated any criterion differences between the judgments for the intact and revealed items. More importantly, because the revelation effect is usually more robust for false alarms, examining ERP differences between hits and false alarms would be informative. Another problem with this study is that it only examined one of many paradigms that elicit the revelation effect. This issue is important because Verde and Rotello (2004) provide evidence indicating that there may be more than one cause for the revelation effect. Consequently, there is a clear need to explore the generality of these findings before theories of the revelation effect can be advanced. The purpose of Experiment 1 was to extend the ERP literature on the revelation effect by comparing ERP activity elicited by discovered test probes (i.e., a related task) to the activity elicited by intact test probes. Importantly, ERP differences have been observed when the form of the information is changed from study to test indicating that the ERP differences are not merely produced by identical perceptual representations of the information (Kazmerski & Friedman, 1997; Paller & Gross, 1998). In addition, ERPs elicited by accurate and inaccurate responses were examined to provide a detailed examination of ERP activity. Any ERP differences between these items could provide important evidence about how the perceptual differences from revealing items at test alter the cognitive processing on a memory test. Experiment 2 capitalized on the strengths of the ERP technique versus traditional behavioral measures. Recent behavioral evidence has suggested that related and unrelated tasks produce different changes that ultimately cause the revelation effect (Verde & Rotello, 2004). ERPs can provide more detailed evidence of changes in cognitive processing than behavioral measures; therefore, the purpose of Experiment 2 was to determine if unrelated tasks (i.e., an addition problem) produce different patterns of ERPs, which would be additional evidence that there are multiple causes of the revelation effect. P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 331 2. Experiment 1 2.1. Method 2.1.1. Participants Twenty-four students from The College of New Jersey (7 males, 17 females), aged 18–22, volunteered in exchange for course credit. In addition, an incentive of $25 was offered to the participant who had the best memory and fastest reaction time among those tested. All participants were right handed (Oldfield, 1971), reported that they had normal or corrected-to-normal vision, and did not have a history of neurological disease. 2.1.2. Apparatus The participants sat in a chair located in an electrically shielded chamber facing a VGA computer monitor located 110 cm away. Each character presented on the monitor extended 0.4 of visual angle vertically and horizontally. Participants were monitored by closed circuit television with a two-way communication system and they responded by pressing keys on a computer keyboard. 2.1.3. Stimuli Four hundred eight letter words served as the stimuli in the experiment (word frequency was 10 or more occurrences per million; Kucera & Francis, 1967). Of these 400 words, 200 were randomly chosen for each participant by the software to be targets (i.e., old words) during the test phase. In addition, half of the old and half of the new words were randomly chosen by the software to be revealed as anagrams during the recognition test. All anagrams were formed by switching the third and sixth letters of the word (e.g., RATIONAL vs. RANIOTAL). 2.1.4. Procedure The experimental session consisted of a study phase and a test phase. Participants studied 200 words for a nonspecific memory test. Each study word was presented one at a time in the center of the computer screen, remaining on the screen for only 500 ms with a 500 ms time lapse between each word. We selected this study procedure because rapid presentation increases participantÕs reliance on familiarity and tends to produce a more robust revelation effect (Landau, 2001). The entire study phase lasted approximately 3.5 min. After a brief delay (approximately 5 min), each participant then completed a recognition memory test consisting of 400 words, 200 target words and 200 new words (i.e., lures). In addition, half of the targets and lures were presented as anagrams (i.e., revealed), and the other half were presented in their intact format. Before the test, participants were told that half of the words on the test (including both intact words and anagrams) were old and half were new. They were generally instructed to determine if each word was old or new and to register their recognition responses on the keyboard. Furthermore, participants were instructed that some of the items would appear as anagrams on the test. On these trials, participants were told to mentally solve the anagram (by switching the 3rd and 6th letters) and then to determine if the item was old or new. The computer proceeded to the next trial after the recognition response was recorded (for both intact and revealed items). 332 P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 2.1.5. ERP recording procedures Potentials were sampled at a rate of 150 Hz from 20 Ag/AgCl electrodes mounted in an elastic cap (Neuromedical Supplies) and referenced to the left mastoid online and referenced to the average of the left and right mastoids offline. Electrodes were placed over the frontal lobes (Fp1, Fp2, F7, F3, Fz, F4, F8, FC3, FCz, and FC4), temporal lobes (FT7, FT8, T7, T8, TP7, and TP8), parietal lobes (CP3, CPz, CP4, P7, P3, Pz, P4, and P8), occipital lobes (O1 and O2), and at the central position on the scalp (C3, Cz, and C4). Vertical electrooculogram (vEOG) was recorded bipolarly using two Ag/AgCl electrodes affixed above and below the subjectÕs left pupil. Horizontal electrooculogram (hEOG) was recorded bipolarly from identical electrodes and attached to the outer canthi of both eyes. Interelectrode impedance was below 5 kX. EEG and EOG signals were recorded with a Contact Precision Instruments amplifier with a 0.01–40 Hz bandpass (
3 dB attenuation). During the test phase, EEG and EOG were sampled for 300 ms before the probe word and for 2700 ms after the probe word. The data were digitally filtered off-line using a 30 Hz lowpass filter (
3 dB/oct). Occular artifacts were corrected using the algorithm developed by Semlitsch, Anderer, Schuster, and Presslich (1986). Trials on which ERP amplitudes exceeded ±150 lV were excluded from the analyses. 2.2. Results and discussion Unless specified otherwise, all results reported in this article are significant at the a = 0.05 level. 2.2.1. Behavioral response data As is conventional in the revelation effect literature, we analyzed the proportion of old items called old (i.e., hits) and new items incorrectly identified as old (i.e., false alarms) as a function of item status (revealed or intact). A 2 · 2 within-subjects analysis of variance with factors of item status (intact vs. revealed) and item type (hits vs. false alarms) indicated that participants registered more hits than false alarms, F (1, 23) = 131.0. Although overall responses did not vary across intact and revealed items (F (1, 23) = 3.36, p = .08), the interaction between item status and item type was significant, F (1, 23) = 6.08. Post hoc analyses found that the hit rates for the intact (M = .65, SD = .07) and revealed targets were not different (M = .66, SD = .10), t (23) < 1. However, people were more likely to call revealed lures old (M = .51, SD = .11) as compared with intact lures (M = .45, SD = .10), t (23) = 2.57. Failure to find a reliable revelation effect for the targets is consistent with the previous literature. Hicks and MarshÕs (1998) meta-analysis of the revelation effect literature showed that the revelation effect is typically larger for lures than targets and that the targets sometimes fail to produce a significant revelation effect. In general, revealed items tend to produce a decreased measure of responder sensitivity (d 0 ) and a more relaxed decision criteria (i.e., more negative C values). For comparison, we also analyzed these signal detection measures. Analysis of d 0 found that sensitivity was lower for the revealed (M = .43, SD = .21) items as compared with the intact items (M = .54, SD = .28); F (1, 23) = 4.59. The decision criterion analysis found that participants were slightly more liberal for revealed items (M =
.24, SD = .26) than intact items (M =
.15, SD = .19); however, this difference was only marginally significant, F (1,23) = 3.56, p = 0.07. These findings are consistent with the results from previous revelation effect investigations. P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 333 Reaction times for correct responses are presented here for comparison to other ERP investigations of memory. Intact hits (M = 1584, SD = 403) were identified faster than revealed hits (M = 2638, SD = 828; F (1, 23) = 71.93), and intact correct rejections (M = 1886, SD = 492) were also identified faster than revealed correct rejections (M = 3668, SD = 1627; F (1, 23) = 42.60). 2.2.2. ERP data The first comparison of interest examined the traditional old/new effects reported in many ERP studies of recognition; thus, ERPs elicited by hits and correct rejections were analyzed for intact and revealed items separately. This initial comparison was followed by additional comparisons (described later) in an effort to elucidate how revelation affects ERP activity. To quantify the ERP effects in all comparisons, eight amplitude measures were computed as the average activity relative to a 300 ms pre-test word baseline over consecutive 300 ms intervals beginning 300 ms after the onset of the test probe (i.e., 300–600, 600–900, 900–1200, 1200–1500, 1500–1800, 1800–2100, 2100–2400, and 2400–2700 ms). These amplitude measures were selected to be consistent with prior ERP investigations of recognition memory and to allow for a sensitive analysis of ERP effects given the lack of comparable ERP data for revealed items. The eight intervals were selected to capture ERP activity leading up to the average response time to judge old revealed items (i.e., 2600 ms). Each of the eight intervals were analyzed at frontal (i.e., F3, Fz, F4, FC3, FCz, and FC4) and parietal electrode regions (i.e., CP3, CPz, CP4, P3, Pz, and P4) using a within-subjects analysis of variance with factors of item type, anterior/posterior (frontal and parietal), and electrode site. The electrode sites were selected based upon the apparent differences in the grand average ERP data and previous ERP studies of recognition. Interactions involving item type and electrode sites, which indicate a difference in ERP topography across electrode sites, were evaluated after after ERPs were scaled to eliminate amplitude differences between the conditions (McCarthy & Wood, 1985). When appropriate, analyses incorporated the Geisser-Greenhouse correction for nonsphericity (corrected degrees of freedom are reported). 2.2.3. Traditional old/new ERP effects The top of Fig. 1 plots the grand average ERPs elicited by hits and correct rejections at left and right frontal and parietal electrodes. The bottom of Fig. 1 plots the ERPs elicited by false alarms and misses. The ERPs elicited by the revealed items are very similar, whereas the intact ERPs elicited by false alarms and hits appear to be more positive than both miss and correct rejection ERPs. 2.2.3.1. Intact hits vs. correct rejections. The main effect of item type (hit vs. CR) was significant for intact items during the 300–600 ms, (F (1, 23) = 4.89), 600–900 ms (F (1, 23) = 7.62), 900– 1200 ms (F (1, 23) = 13.21), and 1500–1800 ms intervals (F (1, 23) = 5.49). It was also marginally significant during the 1200–1500 ms (F (1, 23) = 3.57, p = .07), 1800–2100 ms (F (1, 23) = 4.15, p = .05), and 2400–2700 ms intervals (F (1, 23) = 3.58, p = .07). No significant interactions involving the factor of item type were observed for intact items during any of the intervals. To further explore this result, we performed follow-up analyses that compared ERP amplitudes in the frontal and parietal electrode regions separately. The post hoc analyses revealed a significant main effect of item type at frontal electrode sites for each of the seven intervals tested [300– 600 ms: F (1, 23) = 5.43; 600–900 ms: F (1,23) = 8.46; 900–1200 ms: F (1, 23) = 19.94; 1200– 334 P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 Fig. 1. Experiment 1 grand average ERPs recorded at left and right frontal electrode sites (F3 and F4) and left and right parietal electrode sites (P3 and P4). ERPs elicited by intact items appear in the left half of the figure, whereas ERPs elicited by revealed items appear in the right half of the figure. Hit and correct rejection ERPs are plotted in the top half of the figure, and false alarm and miss ERPs are plotted in the bottom half of the figure. Positive voltage is plotted upward in all graphs. The zero point on the x-axis corresponds to the onset of the test probe and axis ticks are placed at 600 ms increments. 1500 ms: F (1,23) = 5.80; 1500–1800 ms: F (1, 23) = 5.59; 1800–2100 ms: F (1, 23) = 4.71; 2400– 2700 ms: F (1, 23) = 4.73]. The post hoc analyses of parietal ERPs revealed a significant main effect of item type for only the 600–900 ms (F (1, 23) = 5.58) and 900–1200 ms intervals (F (1, 23) = 5.47). None of the interactions involving item type were significant for either analysis. 2.2.3.2. Revealed hits vs. correct rejections. The analysis of item type (hits vs. CRs) for revealed items did not yield any significant main effects, largest F = 2.33, ps > .05. In addition, no significant interactions involving the factor of item type were found for the revealed items (the largest F = 1.64, ps > .05). The analysis of old and new ERPs for the intact and revealed items indicates that intact hits elicited more positive ERPs than intact correct rejections. This old/new ERP effect at the frontal electrode sites began approximately 300 ms after the presentation of the probe and lasted for approximately 1800 ms. A second old/new ERP effect was found at the parietal electrode sites P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 335 but this effect began about 600 ms after the onset of the probe and lasted for 600 ms. Neither of these old/new effects were observed for the revealed items. Because the revelation effect was significant for the lures only, the absence of old/new ERP differences in the revelation condition may have been produced by a selective increase in the ERP amplitudes elicited by new items. Although most ERP studies report differences between old and new items, recent evidence shows that false alarms can also elicit more positive ERPs. Curran (2000) reported that false alarms to lures (which differed in their plurality from studied words) elicited more positive ERPs than correct rejections suggesting that the false alarms possessed higher familiarity levels than correctly rejected lures. Thus, the revelation manipulation in the present experiment may have increased the familiarity levels of new words without affecting the familiarity of old items. Keeping with this logic, we compared false alarm ERPs to those elicited by correct rejections under the assumption that false alarms would possess more familiarity than new words called ‘‘new.’’ Thus, ERP differences in the revealed condition might be limited to a difference between false alarms and correct rejections if revelation specifically affected familiarity of new items. 2.2.3.3. Intact false alarms vs. correct rejections. The results of the ERP analysis of false alarms and correct rejections was very similar to those reported for hits versus correct rejections. The main effect of item type (FA vs. CR) was significant for intact items beginning 300 ms after the onset of the probe and ending several hundred milliseconds later [300–600 ms: F (1, 23) = 7.50; 600–900 ms: F (1, 23) = 4.80; 900–1200 ms: F (1, 23) = 7.35; 1200–1500 ms: F (1, 23) = 9.99; 1500– 1800 ms: F (1, 23) = 4.67; 1800–2100 ms: F (1, 23) = 8.04; 2400–2700 ms: F (1, 23) = 5.36]. No significant interactions involving the factor of item type were observed for intact items during any of the intervals. To further explore the results, we followed these analyses with subsidiary analyses that compared ERP amplitudes in the frontal and parietal regions separately. The post hoc analyses revealed a significant main effect of item type at frontal electrode sites for six of the seven intervals tested [300–600 ms: F (1, 23) = 6.36; 900–1200 ms: F (1, 23) = 7.14; 1200–1500 ms: F (1, 23) = 9.63; 1500–1800 ms: F (1, 23) = 5.22; 1800–2100 ms: F (1, 23) = 7.92; 2400–2700 ms: F (1, 23) = 5.25]. Additionally, the main effect of item type was marginally significant for the 600-900 ms interval, F (1, 23) = 3.91, p = .06. The post hoc analyses of parietal ERPs found a significant main effect of item type for six of the seven intervals tested [300–600 ms: F (1, 23) = 7.33; 600–900 ms: F (1, 23) = 5.36; 900–1200 ms: F (1, 23) = 6.51; 1200–1500 ms: F (1, 23) = 8.16; 1800– 2100 ms: F (1, 23) = 7.18; 2400–2700 ms: F (1, 23) = 4.85]. The main effect of item type was marginally significant during the 1500–1800 ms interval, F (1, 23) = 3.47, p = .08. None of the interactions were significant for the analyses of either the frontal or parietal electrode sites. 2.2.3.4. Revealed false alarms vs. correct rejections. The analysis of item type (FA vs. CR) for revealed items did not reveal any significant main effects nor any significant interactions involving the factor of item type, largest F = 3.21, all ps > .05. To complete the ERP analyses, we compared ERPs elicited by intact and revealed items with the same response. Thus, ERPs elicited by hits were compared to false alarms and in a second comparison correct rejections were compared to misses. None of the analyses for either condition (revealed or intact items) achieved significance at the a = 0.05 level. 336 P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 2.2.4. ERP comparisons between revealed and intact items One possible explanation for the ERP effects reported earlier is that revealing items introduces a latency jitter that blurs the traditional old/new ERP effects observed on most recognition tests. According to this explanation, revealing items at test introduces more variance in the time that it takes to identify the probes. This variation causes variation in the onset of memory related ERP effects. Therefore, the typical old/new ERP effects are randomly distributed across the epoch, and the jitter would produce the null effects observed for revealed item ERPs. To address this potential explanation, we compared hits and correct rejections across conditions. If revealing items does, in fact, blur old/new ERP effects, then we expected the ERP amplitudes for intact hits to be more positive than revealed hits. At the very least, we might observe different amplitudes as a function of time. Table 1 presents the mean ERP amplitudes at the midline electrodes for hits (top half of the table) and correct rejections (bottom half of the table). Examination of the peak amplitudes (indicated in bold type) for hits reveals a pattern that could result from jitter. ERPs elicited in both conditions peaked first at parietal electrodes and later at frontal electrodes. However, intact hit ERPs peaked earlier and appear to be greater than revealed hit ERPs, which is the pattern expected from latency jitter. Although intact correct rejection ERPs show a similar pattern as hits, ERPs for revealed correct rejections peak first at central sites and appear to be greater than intact correct rejection amplitudes. This pattern of results is inconsistent with a jitter explanation. These effects are more clearly observed in Fig. 2 which contrasts intact and revealed ERPs for hits, correct rejections, and false alarms. These descriptive observations were followed with statistical analyses to further explore this alternative argument. Table 1 Mean ERP amplitudes (lV) at the midline electrodes for intact and revealed items Time interval Intact Revealed Pz CPz FCz Hits 600–900 900–1200 1200–1500 1500–1800 1800–2100 2100–2400 2400–2700 5.594 4.523 3.445 3.45 2.264 1.366 1.152 4.947 4.954 4.066 3.808 2.352 1.181 0.964 1.357 3.075 3.237 3.073 1.709 0.625 0.567 Correct rejections 600–900 900–1200 1200–1500 1500–1800 1800–2100 2100–2400 2400–2700 3.791 3.179 2.344 1.634 0.547
.099
.325 3.118 3.151 2.905 1.959 0.524
.003
.089
.590 0.432 1.271 0.576
.554
.859
1.260 Fz Pz CPz FCz Fz 1.091 3.067 3.371 3.055 1.835 0.652 0.548 3.477 4.402 4.399 3.968 3.341 2.865 1.880 3.019 5.018 5.460 5.154 4.372 3.769 2.386
.282 4.155 4.149 4.155 3.205 2.906 1.962
.600 4.485 4.082 4.485 3.572 3.074 2.170
.404 0.842 1.695 1.084
.310
.544
1.373 2.741 4.050 4.441 4.460 3.712 2.58 1.551 1.988 4.271 5.083 5.010 4.458 3.503 2.155
1.274 2.044 3.426 3.817 3.345 2.522 1.367
1.278 2.089 3.978 4.618 4.149 3.196 1.867 P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 337 Fig. 2. Experiment 1 grand average ERPs at left and right frontal electrode (F3 and F4) and left and right parietal electrodes (P3 and P4) are plotted for intact (solid line) and revealed items (dotted line). ERPs elicited by hits, correct rejections, and false alarms are displayed in the graphs in the upper left hand, upper right hand, and bottom of the figure, respectively. 2.2.4.1. Intact hits vs. revealed hits. The main effect of Condition (Intact vs. Revealed) was significant for hits during the first two ERP intervals [300–600 ms: F (1, 23) = 10.00; 600–900 ms: F (1, 23) = 8.55], which indicated that intact ERPs were reliably more positive than revealed items. Later in the interval, the opposite pattern emerged in that revealed hits were more positive than intact hits [1800–2100 ms: F (1, 23) = 4.73; 2100–2400 ms: F (1, 23) = 7.33; 2400–2700 ms, Condition · Site: F (2.95, 67.85) = 3.01]. Thus, the hit ERP amplitudes were larger for the intact items, and the peak amplitudes for the revealed condition appear to occur later in the epoch—a pattern which is consistent with the jitter argument. 2.2.5. Intact CR vs. revealed CR The analyses comparing intact and revealed correct rejections revealed a significant main effect of Condition during the first interval [300–600 ms: F (1, 23) = 5.52], reflecting the more positive intact ERPs very early. However, revealed ERPs began to be more positive than intact ERPs (indicated by significant main effects of condition) beginning approximately 1200 ms into the 338 P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 epoch and continuing for the duration of the epoch [1200–1500 ms: F (1, 23) = 10.26; 1500– 1800 ms: F (1, 23) = 11.97; 1800–2100 ms: F (1, 23) = 15.68; 2100–2400 ms: F (1, 23) = 16.51; 2400–2700 ms: F (1, 23) = 8.47]. As a result, these effects do not appear to be a simple case of latency jitter. If that was the case, then revealed correct rejections should elicit smaller amplitudes or simply be shifted back in time, as was the case for hits. Instead, revealed correct rejections elicited more positive amplitudes during the time intervals that typical old/new effects were observed for intact items. Thus, a latency jitter argument simply cannot explain the observed pattern of effects. 2.2.5.1. Intact FAs vs. revealed FAs. One last comparison of interest is the contrast of ERPs elicited by false alarms in both condition. This is an important comparison given the fact that the revelation effect often manifests itself as an increase in false alarms for revealed items (Hicks & Marsh, 1998). As is evident in Fig. 2, the pattern of differences for false alarm ERP data was similar to that observed for hits and correct rejections. Early in the interval, intact items elicited more positive ERP amplitudes than revealed items [300–600 ms: F (1, 23) = 8.87; 600–900 ms: F (1, 23) = 6.99]. Later in the epoch, revealed items began to be more positive than intact items [Condition · Site: F (2.75, 63.25) = 3.49]. This effect was weaker than that observed in the comparison of hits and correct rejections because the main effect of condition was only marginally significant during the late intervals [1800–2100 ms: F (1, 23) = 3.38, p = .079; 2100–2400 ms: F (1, 23) = 4.05, p = .056]. 2.3. Experiment 1 discussion In Experiment 1, we measured ERPs as people made recognition judgments for intact and revealed items. The behavioral data indicate that our subjects produced the typical revelation effect, and the signal detection measures indicated that the revelation process decreased responder sensitivity and produced a more liberal decision criterion, which is consistent with the effects reported in the revelation effect literature (cf., Hicks & Marsh, 1998). One of the goals of Experiment 1 was to measure ERPs in a paradigm that used a related task to generate the revelation effect to provide information that would help differentiate among the various theoretical accounts of the revelation effect. Our examination of the more traditional recognition memory ERP effects in the present data indicated that intact items elicited an old/new ERP difference that was not present for the revealed items. At first glance, the absence of old/ new ERP effects for revealed items appears to contradict the recognition ERP literature; however, the old/new effect is not always observed (e.g., Jordan & Thomas, 1999; Rugg, Cox, Doyle, & Wells, 1995; Rugg & Doyle, 1992). Our analysis of the behavioral data indicated a robust revelation effect for the lures but not in the targets. We examined the possibility that revealing items only affected lures and their associated ERP activity while not affecting hits by averaging and analyzing ERPs elicited by false alarms. This analysis yielded results similar to the hits and correct rejections comparison. False alarms elicited more positive ERPs than correct rejections in the intact condition regardless of whether the item was old or new (cf., Curran, 2000). These results did not extend to the revealed items because there were no ERP differences among the four types of items in the revealed condition. The absence of ERP differences in the revealed condition resulted from a selective increase in ERPs elicited by the new items. These ERP data indicate that the revelation process caused these items to elicit similar levels of ERP activity regardless of the itemÕs P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 339 true status. Importantly, the revealed item ERPs appeared to be functionally equivalent to ERPs elicited by intact old items. Direct comparisons of hits, correct rejections, and false alarms elicited by intact and revealed item ERPs provided additional evidence that the revelation process alters the usual cognitive processing that people engage during a recognition decision. Across all comparisons, we observed two general ERP differences between intact and revealed items. Early in the recording interval (300– 600 ms), intact items elicited more positive ERPs than revealed items. Later in the interval, revealed item ERPs were more positive than intact item ERPs. ERP studies of recognition have identified two ERP old/new differences that differ in time and location. Some have suggested that the early (300–500 ms) frontal old/new ERP difference reflects familiarity, whereas recollection is a later (600–900 ms) old/new difference at the parietal sites (Curran, 2000; Mecklinger, 2000). The early ERP effects between revealed and intact items are similar to the familiarity-related difference (‘‘FN400’’) reported by Curran (1999, 2000). In our experiment, revealed items elicited more negative ERPs than intact items beginning 300 ms after the onset of the test probe at electrodes across the scalp. Assuming that early ERPs reflect the initial assessment of familiarity, this means that revealed items do not elicit the initial feelings of familiarity elicited by intact items. This difference is most likely due to their altered form at test. Instead, it appears that revealing items at test actually serves to decrease familiarity (cf., Azimian-Faridani & Wilding, in press; Hicks & Marsh, 1998). Although the early ERP differences appear to reflect a change in familiarity, ERP differences that occur later in the epoch have been connected to recollection or more elaborate decision processes (see Mecklinger, 2000). In the present study, revealed items elicited more positive ERPs than intact items late in the recording epoch. These results mirror those observed when people identify incomplete line drawings (Cycowicz & Friedman, 1999), and collectively, they suggest that discovering the identity of test probes produces positive ERPs across the scalp. These ERPs are somewhat similar to those elicited by old items on a recognition test; however, the widespread topography suggests that it does not specifically reflect an increase in either familiarity or recollection that produce more localized ERP activity. Taken together, the ERP data from the Experiment 1 suggest that revealing an item initially decreases familiarity. Later in the course of processing, the revealed items (regardless of item type and decision) elicit similar brain activity (as indicated by the absence of traditional old/ new effects). Hicks and Marsh (1998) argued that the revelation process activates a variety of alternative memory traces and this decreases the level of familiarity. This decreased familiarity makes the recognition judgment more difficult because it is now more difficult to distinguish between old and new items. To differentiate between old and new items, people adopt a more liberal decision criterion, increasing the number of ‘‘old’’ responses. Thus, the ERP results of Experiment 1 are entirely consistent with the decrement-to-familiarity account because revealed items caused an initial decrease in familiarity followed by similar levels of brain activation for all item types. 3. Experiment 2 Recently, Verde and Rotello (2004) provided evidence that the revelation effect is not the function of a single cognitive process. More specifically, they found that the revelation effect causes a 340 P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 shift in sensitivity when the revealed items are identical to the test probes. This idea is consistent with the results from Experiment 1, which had people solve anagrams of the test probe before they rendered a recognition judgment. However, Verde and RotelloÕs evidence indicates that the revelation effect is produced by a shift in response bias when the revealed items are unrelated to the test probe. These observations appear to be inconsistent with another ERP study that used unrelated anagrams to elicit the revelation effect (Azimian-Faridani & Wilding, in press). In that study, Azimian-Faridani and Wilding present evidence that they argued was consistent with the decrement-to-familiarity account. Therefore, the purpose of Experiment 2 was to measure ERPs using an unrelated task. We used the paradigm developed by Niewiadomski and Hockley (2001) because they found a robust revelation effect for probes that were preceded by a numerical addition problem. This paradigm has the added advantage of generating the revelation effect using a non-verbal task. If Verde and RotelloÕs hypothesis is correct, then we should observe a different pattern of ERP effects when the task is unrelated to the probe compared to when the task is related to the probe. Alternatively, if we see ERP effects similar to those in Experiment 1, this would be support for the decrement-tofamiliarity account. 3.1. Method 3.1.1. Participants Thirty students from The College of New Jersey (17 males, 13 females), aged 18–22, volunteered in exchange for course credit. In addition, an incentive of $25 was offered to the participant who had the best memory and fastest reaction time among those tested. All participants were right handed (Oldfield, 1971), reported that they had normal or corrected-to-normal vision, and did not have a history of neurological disease. 3.1.2. Apparatus and procedure The materials and procedures were the same as those used in Experiment 1 with the following exceptions. Items were not revealed at test. Instead, an addition problem immediately preceded half of the test probes. The problems involved adding three digit numbers whose sum did not exceed 999. The software randomly determined the numbers for each participant. This paradigm was pilot tested to ensure that behavioral performance would be similar to that observed in Experiment 1, and this testing indicated that the procedures used in Experiment 1 would need to be altered in order to adapt this paradigm. The addition problems added substantial length to the experiment; therefore, we reduced length of the study and test phases of the experiment. For each participant, the computer randomly selected 170 words from the pool of 400 items to be studied. Half of the targets and half of the lures were then randomly designated as ‘‘revealed items’’ (i.e., preceded by an addition task). This modification resulted in 85 items of each type (i.e., intact targets, intact lures, revealed targets, and revealed lures). The study time was increased from 500 ms in experiment 1 to 3500 ms in the present experiment. These changes created conditions that were adequate for ERP recording and created behavioral responding that was nearly equivalent to that observed in Experiment 1. P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 341 3.2. Results 3.2.1. Behavioral response data The proportion of old items called old (i.e., hits) and new items incorrectly identified as old (i.e., false alarms) as a function of item status (revealed or intact) were analyzed with a 2 · 2 withinsubjects analysis of variance. A significant main effect of item type (hits vs. false alarms) indicated that there were more hits than false alarms (F (1, 29) = 160.87) and a significant main effect of item status (intact vs. revealed) indicated there were more old responses for revealed items, F (1, 29) = 31.43. Importantly, these effects were qualified by a significant interaction between item status and item type (F (1, 29) = 9.58), which indicated that revelation effect was larger for lures than for targets. Post hoc analyses found that the hit rates for revealed targets (M = .77, SD = .15) were higher than for intact targets (M = .70, SD = .14; F (1, 29) = 15.83), and the false alarm rate was higher for revealed lures (M = .46, SD = .17) than intact lures (M = .33, SD = .13), F (1, 29) = 36.00. Thus, the revelation effect was observed for both targets and lures in the present experiment. Analysis of responder sensitivity (d 0 ) indicated that the ability to detect old and new items did not differ between intact (M = 1.04, SD = .46) and revealed items (M = 0.96, SD = .48; F (1, 29) = 2.96, p = .10; cf., Verde & Rotello, 2004). The analysis of decision criterion (C) indicated that participants were more liberal when items were revealed (M =
.36, SD = .42) than when items were presented intact (M =
.06, SD = .33; F (1, 29) = 29.11). The analysis of reaction times to correct responses indicated that intact hits (M = 1220, SD = 257) were identified faster than revealed hits (M = 1421, SD = 357; F (1,29) = 34.41), and intact correct rejections (M = 1390, SD = 309) were also identified faster than revealed correct rejections (M = 1743, SD = 389; F (1, 29) = 118.19). 3.2.2. ERP data The ERP data were analyzed in the same manner as in Experiment 1, with the following exception. Five amplitude measures were computed as the average activity relative to a 300 ms pre-test word baseline over consecutive 300 ms intervals beginning 300 ms after the onset of the test probe (i.e., 300–600, 600–900, 900–1200, 1200–1500, and 1500–1800 ms). The later intervals were not analyzed in this experiment because the overall response times were faster in this experiment. 3.2.3. Traditional old/new ERP effects The top of Fig. 3 plots the grand average ERPs for intact and revealed items. ERPs elicited by hits and correct rejections are displayed in the top half of the figure, whereas the bottom of Fig. 3 plots the ERPs elicited by false alarms and misses. 3.2.3.1. Intact hits vs. correct rejections. Hit ERPs were more positive than correct rejection ERPs during the 300–600 ms interval, Item Type: F (1, 29) = 6.36. During the 900–1200 ms interval a significant Item Type · Electrode Site interaction was also detected, (F (2.5, 72.5) = 6.51). No significant interactions involving the factor of item type were observed for intact items during any of the intervals. Post hoc analyses compared ERP amplitudes in the frontal and parietal electrode regions separately. During the 300–600 ms interval, a significant main effect of item type was observed at both 342 P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 Fig. 3. Experiment 2 grand average ERPs elicited by intact and revealed items are plotted separately. This figure uses the same format as Fig. 1. frontal (F (1, 29) = 5.17) and parietal electrodes (F (1, 29) = 7.90) that indicated hits elicited more positive ERPs than correct rejections. In addition, a significant item type by electrode interaction was observed at parietal electrodes (F (2.15, 62.35) = 4.23) that indicated that the old/new ERP difference was larger at left parietal electrode sites. During the 900–1200 ms interval, the item type main effect was only marginally significant at the frontal electrodes, F (1, 29) = 3.81, p = .06. However, the significant item type by electrode site interaction at parietal electrodes (F (2.65, 76.85) = 9.39) indicated that the old/new effect continued to be larger at left parietal electrodes. 3.2.3.2. Revealed hits vs. correct rejections. The analyses revealed significant three-way interactions between Item Type (hit vs. CR), Anterior/Posterior electrode position (frontal vs. parietal), and Electrode Site (left vs. right hemisphere) for the 300–600 ms (F (4.05, 117.45) = 2.99) and 600– 900 ms intervals, F (3.4, 98.6) = 3.97. No other effects were significant. These three-way interactions reflect the tendency for correct rejection ERPs to be slightly more positive than intact ERPs at left frontal electrode sites, whereas there is little difference in the ERPs recorded from the right P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 343 hemisphere electrode sites. This trend did not reveal any significant differences when post hoc tests were performed; therefore, the effects in the omnibus analysis appear to be the result of very subtle differences rather than a clear difference between hit and correct rejection ERPs. 3.2.3.3. Additional ERP comparisons of intact and revealed words. To be consistent with Experiment 1, several comparisons were analyzed for intact and revealed items separately. More specifically, false alarm ERPs were compared to correct rejections, hit ERPs were compared to false alarms, and correct rejection ERPs were compared to misses. None of these comparisons revealed any significant effects for either intact or revealed items. 3.2.4. ERP comparisons between revealed and intact items Fig. 4 contrasts intact and revealed ERPs for hits, correct rejections, and false alarms. As in Experiment 1, there are apparent differences between ERPs elicited by intact and revealed items; thus, each comparison was subjected to statistical analyses to further explore the ERP effects observed in this experiment. Fig. 4. Experiment 2 grand average ERPs elicited by intact and revealed items are contrasted in the graphs. This figure uses the same format as Fig. 2. 344 P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 3.2.4.1. Intact hits vs. revealed hits. During the 300–600 ms interval, the condition (Intact vs. Revealed) factor interacted with electrode site positions, Condition · Anterior/Posterior · Site: F (1.85, 53.65) = 10.16. The nature of this effect is apparent in Fig. 4 in that intact hits elicited more positive ERPs at frontal electrodes but more negative ERPs at parietal electrodes. During the subsequent intervals, intact hits elicited more positive ERPs than revealed hits as indicated by significant main effects of condition [600–900 ms: F (1,29) = 12.71; 900–1200 ms: F (1, 29) = 14.74; 1200– 1500 ms: F (1, 29) = 4.56]. However, these main effects were accompanied by significant three-way interactions (Condition · Anterior/Posterior · Site) [600–900 ms: F (2.05, 59.45) = 12.82; 900– 1200 ms: F (2.75, 79.75) = 6.87; 1200–1500 ms: F (3.15, 91.35) = 4.01], which indicated that that this difference was larger at frontal and left hemisphere electrodes. 3.2.4.2. Intact CR vs. revealed CR. A similar pattern of results was observed for the correct rejections. During the first time interval, intact correct rejection ERPs were more positive at frontal electrodes but more negative ERPs at parietal electrodes than revealed correct rejection ERPs, Condition · Anterior/Posterior · Site: F (2.9, 84.1) = 9.15. Intact correct rejections elicited more positive ERPs than revealed correct rejections as indicated by significant main effects of condition for the following two time intervals [600–900 ms: F (1, 29) = 4.55; 900–1200 ms: F (1, 29) = 4.52]. However, these main effects were accompanied by significant interactions of electrode position during the 600–900 ms [Condition · Anterior/Posterior, F (1, 29) = 5.09; Condition · Site: F (1.75, 50.75) = 4.81] and 900–1200 ms intervals [Condition · Anterior/Posterior, F (1, 29) = 28.33; Condition · Site: F (1.6, 46.4) = 7.62]. During the last two time intervals, significant interactions involving electrode position were detected for the 1200–1500 ms [Condition · Anterior/Posterior, F (1, 29) = 17.63; Condition · Site: F (1.9, 55.1) = 7.33] and 1500– 1800 ms intervals [Condition · Anterior/Posterior, F (1, 29) = 8.51; Condition · Site: F (2.65, 76.85) = 5.58]. These condition by electrode position interactions indicated that intact correct rejection ERPs were more positive than revealed at frontal and left hemisphere electrodes. 3.2.4.3. Intact FA vs. revealed FA. The pattern of results for false alarm ERPs was similar to that observed for hits and correct rejections. During the first time interval, intact false alarm ERPs were more positive at frontal electrodes but more negative at parietal electrodes than revealed false alarm ERPs, Condition · Anterior/Posterior · Site: F (2.4, 40.8) = 5.03. During the 900– 1200 ms interval, intact false alarm ERPs were more positive than revealed false alarm ERPs, Condition: F (1, 17) = 6.21. Significant interactions involving electrode position indicated that this difference was larger at frontal and left hemisphere electrodes Condition · Anterior/Posterior, F (1, 17) = 53.73; Condition · Site: F (1.6, 27.2) = 8.58]. These basic effects persisted during the last two time intervals [1200–1500 ms: Condition · Anterior/Posterior · Site: F (3.15, 53.55) = 4.85; 1500–1800 ms: Condition · Anterior/Posterior · Site: F (3.6, 61.2) = 3.91]. 3.3. Experiment 2 discussion The purpose of Experiment 2 was to observe the revelation effect and the concomitant changes in ERP activity in a paradigm that utilized an unrelated task to generate the revelation effect. We adopted the paradigm developed by Niewiadomski and Hockley (2001) in which addition prob- P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 345 lems preceded half of the test probes (i.e., the revealed items). Analysis of the behavioral data indicated a revelation effect for both lures and targets. Responder sensitivity (d 0 ) was not affected by the math problems; however, people adopted a more liberal criterion for the items preceded by an addition problem. These behavioral data are entirely consistent with existing studies on the revelation effect (cf., Hicks & Marsh, 1998; Verde & Rotello, 2004). As with Experiment 1, old/new ERP differences were observed for intact items but not for revealed items. No additional ERP differences were observed from comparisons across different types of items in the revealed and intact conditions. Although this pattern differs from the results of Experiment 1, the recognition ERP literature provides mixed results (when ERPs elicited by errors have been examined) in that sometimes errors elicit similar old/new effects (e.g., Curran, 2000; Finnigan et al., 2002) and other times they do not (e.g., Neville, Kutas, Chesney, & Schmidt, 1986; Wilding & Rugg, 1997). Three separate analyses contrasted ERPs elicited by hits, correct rejections, and false alarms. These analyses yielded similar results regardless of item type and response. Early in the interval, intact ERPs were more positive than revealed ERPs at frontal electrode sites. The opposite pattern was observed at parietal electrode sites. In subsequent intervals, intact ERPs were more positive than revealed and this difference tended to be larger at frontal and left hemisphere electrodes. In sum, we observed the same general effects on traditional old/new ERP effects as in Experiment 1. However, direct comparisons of intact and revealed item ERPs presented a different pattern of ERP differences than those observed in Experiment 1. 4. General discussion Across two experiments, we measured ERPs as people made recognition judgments for intact and ‘‘revealed’’ test probes. Items were revealed in Experiment 1 by presenting the test probe as an anagram that people had to solve before they made a recognition judgment. A numerical addition task was completed prior to judging half of the recognition probes (i.e., revealed items) in Experiment 2. A summary of the basic findings from each of these experiments is presented first followed by a discussion of the implications of these findings for theories of the revelation effect. 4.1. Summary of findings In Experiment 1, people were more likely to call the revealed items (i.e., anagrams) old as compared with the intact items. A closer analysis indicated that this effect was quite large for the lures but not for the targets. Signal detection measures indicated that the revelation process decreased responder sensitivity and produced a slightly liberal shift in decision criterion. The traditional old/ new ERP effects (i.e., old item ERPs are more positive than new item ERPs) were observed for intact items, but these effects were absent for revealed items. Direct ERP comparisons of intact and revealed items indicated that intact items elicited higher levels of familiarity very early in the recording epoch; however, revealed items elicited more positive ERPs later in the epoch. Because the activity of all revealed items was similar, these results appear to be consistent with the decrement-to-familiarity account of the revelation effect (Hicks & Marsh, 1998). In Experiment 2, 346 P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 the revelation effect was detected for both targets and lures. Using numerical addition problems to elicit the revelation effect did not affect responder sensitivity but it did produce a more liberal decision criterion for the revealed items. Examination of traditional old/new ERP effects was consistent with the results of Experiment 1 in that old/new effects were observed for intact but not revealed items. Importantly, comparisons of the ERPs for intact and revealed items uncovered a different pattern of results than those observed in Experiment 1. Although intact items elicited more positive ERPs at frontal electrode sites early in the epoch (consistent with Experiment 1), the opposite pattern was observed at parietal electrode sites where the revealed ERPs were more positive than intact ERPs. Furthermore, intact ERPs were more positive than revealed ERPs for the duration of the epoch, which is the opposite of the effects observed in Experiment 1. There were a number of similarities across Experiments 1 and 2. First, the old/new ERP effects were observed for intact but not revealed items. Second, comparison of the intact and revealed ERPs produced the same pattern of differences regardless of the type of item contrasted (i.e., hits, correct rejections, and false alarms). These similarities suggest that the pattern of results was not influenced by the fact that the revelation effect was limited to lures in Experiment 1 and was present in targets and lures in Experiment 2. If this had an impact on ERPs, then logic dictates a different pattern of differences for items that show the revelation effect (i.e., lures in Experiment 1) versus items that do not show the revelation effect (i.e., targets in Experiment 1). 4.2. Implications for theories of the revelation effect Although there are several possible explanations for the revelation effect, a complete theoretical interpretation of the effect has been elusive. Some have argued that discovering the identity of the revealed test probes temporarily increases the feeling of familiarity associated with the test probe (Luo, 1993; Peynircioglu & Teckan, 1993). Despite this contention, Verde and Rotello (2003) have presented ROC curves suggesting that changes in the familiarity of the test probes, by itself, is not sufficient to account for the revelation effect. In contrast, Hicks and Marsh (1998) believe that the revelation process activates alternative memory traces and this decreases the level of familiarity of the test probe (the decrement-to-familiarity account). This decreased familiarity makes the recognition judgment more difficult because the average familiarity values of the studied and non-studied words become more similar. Because this increases the difficulty of distinguishing between old and new words, people adopt a more liberal decision criterion and this increases the number of ‘‘old’’ responses. According to the criterion flux account (Hockley & Niewiadomski, 2001; Niewiadomski & Hockley, 2001) tasks that precede the recognition decision cause people to temporarily become confused about the appropriate decision criteria for the judgment and people become more liberal with their recognition decisions. Recently, Verde and Rotello (2004) suggested that the revelation effect arises from a confluence of factors. When the revelation task requires people to unscramble and then judge the same item (a related task) a decrease in sensitivity or changes in familiarity produce the revelation effect. However, if the task that precedes the recognition decision is unrelated to the identity of the test probe, then the revelation effect arises solely from a shift in response bias. The goal of these two experiments was to measure ERPs to provide new information that might help differentiate the various theoretical accounts of the revelation effect. Collectively, the results from the two present ERP studies support Verde and RotelloÕs (2004) argument that the revelation effect is not a sin- P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 347 gular phenomenon. The results from Experiment 1 demonstrate that the traditional revelation effect paradigm decreased fluency for the revealed items. The change in fluency was accompanied by similar levels of ERP activity for all of the revealed items. Thus, this evidence supports a decrement-to-familiarity account of the revelation effect (Hicks & Marsh, 1998) and suggests that changes in sensitivity produce the revelation effect when revealed items are related to the probe (Verde & Rotello, 2004). However, Experiment 2 indicates that a simple change in fluency does not occur when the task is unrelated to the test probe because different patterns of ERP amplitudes were observed at frontal and parietal electrodes during the time in which familiarity-related ERP effects occur. This conclusion is based on the observation that changes in fluency appear to influence ERP activity that occurs early in the epoch (approximately 300–500 ms after the probe). In Experiment 1, revealed items elicited more negative ERP amplitudes than intact items at frontal and parietal electrode sites early in the epoch, whereas revealed items elicited more positive ERP amplitudes at parietal electrode sites in Experiment 2. As a result, the different patterns of ERP activity early in the epoch between the two experiments favor the interpretation that unrelated tasks are not solely caused by changes in fluency (Verde & Rotello, 2004). Instead, the revelation effect from unrelated tasks might result from a shift in response bias (Niewiadomski & Hockley, 2001). Unfortunately, the conclusion that unrelated tasks produce a shift in response bias cannot be strengthened by a comparison to previous ERP studies that have manipulated decision criterion during recognition because this data is not available. Thus, this claim will need to be reevaluated after direct investigations of decision criterion and recognition-related ERP activity emerge. Importantly, our data provide a more general context to interpret data from another recent ERP study of the revelation effect (Azimian-Faridani & Wilding, in press). Azimian-Faridani and Wilding presented unrelated anagrams before half of the test probes, and found that revelation decreases familiarity. These results are consistent with Experiment 1 results, but are inconsistent with Verde and RotelloÕs (2004) argument that unrelated tasks affect decision criterion rather than familiarity. Verde and Rotello noted that related tasks tend to decrease sensitivity, which serves as the basis for their argument that related tasks affect familiarity. In Experiment 1 we found that a related task decreased sensitivity and influenced familiarity-related ERP activity, whereas Experiment 2 presented evidence that an unrelated task does not alter sensitivity and causes ERP differences that cannot be specifically connected to changes in familiarity. Unfortunately, Azimian-Faridani and Wilding did not analyze sensitivity, but d 0 calculations based upon mean response proportions suggest that their revelation manipulation also decreased sensitivity and affected familiarity-related ERP effects. These observations suggest that the task employed by Azimian-Faridani and Wilding is more akin to a related task than an unrelated task. Perhaps, the verbal nature of the anagram task affected familiarity, whereas unrelated tasks that are nonverbal (e.g., an addition task) tend to influence other recognition factors, such as decision criterion. As researchers collect more ERP and behavioral data on the revelation effect, we believe that Whittlesea and WilliamsÕs discrepancy-attribution hypothesis (Whittlesea & Williams, 2001a, 2001b) might be best suited to explain the various causes of the revelation effect. By their account, as people make recognition judgments, as well as other types of judgments, the discrepancy between how they expect to judge a particular item and how they subjectively experience that item influences their decisions. When an item is processed fluently and people cannot point to any obvi- 348 P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 ous alternative explanation for this fluency, they interpret the surprising feeling of fluency as evidence that they previously encountered the item (called ‘‘surprising redintegration’’ by Whittlesea & Williams, 2001b, p. 27). However, people will attribute the feeling to this other causal agent and subsequently judge the item as new if there is some other plausible explanation for the fluent processing. In addition to this cause of familiarity, Whittlesea and Williams have identified other situations that produce familiarity. For example, the feeling of familiarity can arise when there is a burst of processing fluency in a context that does not appear to create the surprising fluency. Regarding the revelation effect, the related tasks (as in our Experiment 1) produce familiarity because the probe is initially disguised as an anagram and this produces low levels of fluency. As people solve the anagram, they experience an increase in fluency, which in turn, creates a distinct feeling of familiarity that they attribute to a previous encounter with the item earlier in the experimental session. Unrelated tasks produce the feeling of familiarity via a different process. In this case, the difficult processing in the unrelated task is immediately followed by fluent processing of the probe, and people misattribute the fluent processing to the experience that the stimulus was studied earlier (Whittlesea & Williams, 2001a). Thus, the discrepancy-attribution hypothesis articulates different causes of familiarity that can affect memory judgments. This alternative explanation might explain why the related and unrelated tasks produce different revelation effects and different patterns of ERP activity. The data from these two experiments demonstrate that ERPs are a useful tool for differentiating the theoretical explanations of the revelation effect. Additionally, these data support the argument that there are distinct kinds of revelation effects. Despite these advances, no single study can answer all the questions. There is a clear need for additional ERP studies of the revelation effect to determine the generality of the findings reported here and to provide additional evidence about how different tasks affect familiarity. The ERP effects observed in Experiment 2 suggest that unrelated tasks produce different revelation effects than related probes. The significance of the ERP effects in Experiment 2 await further clarification from more general studies of recognition memory that examine the influence of decision criteria and sensitivity on ERP activity. For example, these data might indicate that an unrelated task simply causes a liberal shift in decision criterion if studies that directly manipulate decision criteria on recognition tests produce similar results. Alternatively, an unrelated task might cause a temporary loss of criterion setting (Hockley & Niewiadomski, 2001; Niewiadomski & Hockley, 2001) or change the context that is used to evaluate fluency (Whittlesea & Williams, 2001a) and either one of these phenomena might produce the ERP effects observed in Experiment 2. Thus, understanding how these phenomena affect ERP activity can inform our understanding of how unrelated tasks produce the revelation effect. References Azimian-Faridani, N., & Wilding, E. L. (in press). An event-related potential study of the revelation effect. Psychonomic Bulletin & Review. Craik, F. I. M., & Lockhart, R. S. (1972). Levels of processing: A framework for memory research. Journal of Verbal Learning and Verbal Behavior, 11, 671–684. Curran, T. (1999). The electrophysiology of incidental and intentional retrieval: ERP old/new effects in lexical decision and recognition memory. Neuropsychologia, 37, 771–785. P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 349 Curran, T. (2000). Brain potentials of recollection and familiarity. Memory & Cognition, 28(6), 923–938. Cycowicz, Y. M., & Friedman, D. (1999). ERP recordings during a picture fragment completion task: Effects of memory instructions. Cognitive Brain Research, 8, 271–288. Deese, J. (1959). On the prediction of occurrence of particular verbal intrusions in immediate recall. Journal of Experimental Psychology, 58, 17–22. Finnigan, S., Humphreys, M. S., Dennis, S., & Geffen, G. (2002). ERP Ôold/newÕ effects: Memory strength and decisional factor(s). Neuropsychologia, 40, 2288–2304. Green, D. M., & Swets, J. A. (1966). Signal detection theory and psychophysics. New York: Wiley. Hicks, J. L., & Marsh, R. L. (1998). A decrement-to-familiarity interpretation of the revelation effect from forcedchoice tests of recognition memory. Journal of Experimental Psychology: Learning Memory, and Cognition, 24(5), 1105–1120. Hockley, W. E., & Niewiadomski, M. W. (2001). Interrupting recognition memory: Tests of a criterion-change account of the revelation effect. Memory & Cognition, 29(8), 1176–1184. Jacoby, L. L. (1991). A process dissociation framework: Separating automatic from intentional uses of memory. Journal of Memory and Language, 30, 513–541. Jacoby, L. L., & Dallas, M. (1981). On the relationship between autobiographical memory and perceptual learning. Journal of Experimental Psychology: General, 110, 306–340. Jordan, T. R., & Thomas, S. M. (1999). Memory for normal and distorted pictures: Modulation of the ERP repetition effect. Journal of Psychophysiology, 13, 224–233. Kazmerski, V. A., & Friedman, D. (1997). Old/new differences in direct and indirect memory tests using pictures and words in within- and cross-form conditions: Event-related potential and behavioral measures. Cognitive Brain Research, 5, 255–272. Kucera, H., & Francis, W. N. (1967). Computational analysis of present-day American English. Providence, RI: Brown University Press. Landau, J. D. (2001). Altering the balance of recollection and familiarity influences the revelation effect. American Journal of Psychology, 114, 425–437. LeCompte, D. D. (1995). Recollective experience in the revelation effect: Separating the contributions of recollection and familiarity. Memory & Cognition, 23(3), 324–334. Leynes, P. A. (2002). The effect of specific queries on source-monitoring event-related potentials. Brain and Cognition, 50, 218–233. Leynes, P. A., Bink, M. L., Marsh, R. L., Allen, J. D., & May, J. C. (2003). Test modality affects source monitoring and event-related potentials. American Journal of Psychology, 116(3), 389–413. Luo, C. R. (1993). Enhanced feeling of recognition: Effects of identifying and manipulating test items on recognition memory. Journal of Experimental Psychology: Learning Memory, and Cognition, 19, 405–413. Mandler, G. (1980). Recognizing: The judgment of previous occurrence. Psychological Review, 87, 252–271. McCarthy, G., & Wood, C. C. (1985). Scalp distributions of event-related potentials: An ambiguity associated with analysis of variance models. Electroencephalography and Clinical Neurophysiology, 62, 203–208. Mecklinger, A. (2000). Interfacing mind and brain: A neurocognitive model of recognition memory. Psychophysiology, 37, 565–582. Neville, H., Kutas, M., Chesney, G., & Schmidt, A. L. (1986). Event-related potentials during initial encoding and recognition of congruous and incongruous words. Journal of Memory and Language, 25, 75–92. Niewiadomski, M. W., & Hockley, W. E. (2001). Interrupting recognition memory: Tests of familiarity-based accounts of the revelation effect. Memory & Cognition, 29(8), 1130–1138. Oldfield, R. C. (1971). The assessment and analysis of handedness: The Edinburgh Inventory. Neuropsychologia, 9, 97–113. Paller, K. A., & Gross, M. (1998). Brain potentials associated with perceptual priming vs. explicit remembering during the repetition of visual word-form. Neuropsychologia, 36, 559–571. Peynircioglu, Z. F., & Teckan, A. I. (1993). Revelation effect: Effort or priming does not create the sense of familiarity. Journal of Experimental Psychology: Learning, Memory, and Cognition, 19, 382–388. Ranganath, C., & Paller, K. A. (2000). Neural correlates of memory retrieval and evaluation. Cognitive Brain Research, 9, 209–222. 350 P.A. Leynes et al. / Consciousness and Cognition 14 (2005) 327–350 Roediger, H. L., & McDermott, K. B. (1995). Creating false memories: Remembering words not presented in lists. Journal of Experimental Psychology: Learning, Memory, and Cognition, 21, 803–814. Rugg, M. D. (1995). ERP studies of memory. In M. D. Rugg & M. G. H. Coles (Eds.), Electrophysiology of mind: Event-related brain potentials and cognition. New York, NY: Oxford University Press. Rugg, M. D., Cox, C. J. C., Doyle, M. C., & Wells, T. (1995). Event-related potentials and the recollection of low and high frequency words. Neuropsychologia, 33(4), 471–484. Rugg, M. D., & Doyle, M. C. (1992). Event-related potentials and recognition memory for low- and high-frequency words. Journal of Cognitive Neuroscience, 4(1), 69–79. Semlitsch, H. V., Anderer, P., Schuster, P., & Presslich, O. (1986). A solution for reliable and valid reduction of ocular artifacts, applied to the P300 ERP. Psychophysiology, 23, 696–703. Verde, M. F., & Rotello, C. M. (2004). ROC curves show that the revelation effect is not a single phenomenon. Psychonomic Bulletin & Review, 11, 560–566. Verde, M. F., & Rotello, C. M. (2003). Does familiarity change in the revelation effect?. Journal of Experimental Psychology: Learning, Memory, and Cognition, 29, 739–746. Watkins, M. J., & Peynircioglu, Z. F. (1990). The revelation effect: When disguising test items induces recognition. Journal of Experimental Psychology: Learning, Memory, and Cognition, 16, 1012–1020. Westerman, D. L., & Green, R. L. (1996). On the generality of the revelation effect. Journal of Experimental Psychology: Learning, Memory, and Cognition, 22, 1147–1153. Westerman, D. L., & Green, R. L. (1998). The revelation that the revelation effect is not due to revelation. Journal of Experimental Psychology: Learning, Memory, and Cognition, 24, 377–386. Whittlesea, B. W. A., Jacoby, L. L., & Girard, K. (1990). Illusions of immediate memory: Evidence of an attributional basis for feeling of familiarity and perceptual quality. Journal of Memory and Language, 29, 716–732. Whittlesea, B. W. A., & Williams, L. D. (2001a). The discrepancy-attribution hypothesis: I. The heuristic basis of feelings of familiarity. Journal of Experimental Psychology: Learning, Memory, and Cognition, 27, 3–13. Whittlesea, B. W. A., & Williams, L. D. (2001b). The discrepancy-attribution hypothesis: II. Expectation, uncertainty, surprise, and feelings of familiarity. Journal of Experimental Psychology: Learning, Memory, and Cognition, 27, 14–33. Wilding, E. L., & Rugg, M. D. (1996). An event-related potential study of recognition memory with and without retrieval of source. Brain, 119, 889–905. Wilding, E. L., & Rugg, M. D. (1997). Event-related potentials and the recognition memory exclusion task. Neuropsychologia, 35, 119–128.

Consciousness and Cognition 22 (2013) 920–930 Contents lists available at SciVerse ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Short Communication The nature of the memory buffer in implicit learning: Learning Chinese tonal symmetries Feifei Li a, Shan Jiang b, Xiuyan Guo c,a,⇑, Zhiliang Yang a, Zoltan Dienes d a School of Psychology and Cognitive Science, East China Normal University, Shanghai, China School of Social Administration, Shanghai University of Political Science and Law, Shanghai, China c Shanghai Key Laboratory of Magnetic Resonance, East China Normal University, Shanghai, China d Sackler Centre for Consciousness Science and School of Psychology, University of Sussex, Brighton, UK b a r t i c l e i n f o Article history: Received 11 August 2012 Available online 14 July 2013 Keywords: Implicit learning Nonlocal dependencies Inversion Retrograde Structural knowledge Subjective measures Context-free grammar a b s t r a c t Previous research has established that people can implicitly learn chunks, which (in terms of formal language theory) do not require a memory buffer to process. The present study explores the implicit learning of nonlocal dependencies generated by higher than finitestate grammars, specifically, Chinese tonal retrogrades (i.e. centre embeddings generated from a context-free grammar) and inversions (i.e. cross-serial dependencies generated from a mildly context-sensitive grammar), which do require buffers (for example, last in-first out and first in-first out, respectively). People were asked to listen to and memorize artificial poetry instantiating one of the two grammars; after this training phase, people were informed of the existence of rules and asked to classify new poems, while providing attributions of the basis of their judgments. People acquired unconscious structural knowledge of both tonal retrogrades and inversions. Moreover, inversions were implicitly learnt more easily than retrogrades constraining the nature of the memory buffer in computational models of implicit learning. Ó 2013 Elsevier Inc. All rights reserved. 1. Introduction Humans are equipped with powerful learning mechanisms for acquiring unconscious knowledge of structural regularities (Dienes, 2012; Reber, 1989; for a different view on the knowledge being unconscious, see Shanks, 2005; for a somewhat intermediate position, see Cleeremans, 2006). Such implicit learning plays a major role in different areas of human cognition, including music (e.g. Rohrmeier & Rebuschat, 2012; Rohrmeier, Rebuschat, & Cross, 2011; Tillman, Bharucha, & Bigand, 2000), perceptual-motor skills (e.g. Reed, McLeod, & Dienes, 2010), and language acquisition (e.g. Chen et al., 2011; Guo et al., 2011; Leung & Williams, 2011; Poletiek, 2002; Saffran, Newport, Aslin, Tunick, & Barrueco, 1997; Williams, 2009). One of the key questions in implicit learning has focused on the contents of the acquired knowledge. Reber (1967) initially claimed that participants’ knowledge could take the form of abstract rules, for example rules that distinguished terminal elements (the elements that actually appear in the string) from non-terminal symbols (e.g. classes of such elements, like word classes); and rules that are about nonlocal rather than adjacent elements (Manza & Reber, 1997). However, some have argued that implicit learning in more general domains may merely involve learning of allowable chunks of successive terminals (e.g. Perruchet & Vinter, 1998) or specific sequences of terminals found in learned exemplars (e.g. Brooks & Vokey, 1991; Jamieson & Mewhort, 2009). ⇑ Corresponding author. Address: School of Psychology and Cognitive Science, East China Normal University, North Zhongshan Road 3663, Shanghai 200062, China. Fax: +86 21 67644201. E-mail addresses: xyguo@psy.ecnu.edu.cn, xyguo2006@gmail.com (X. Guo). 1053-8100/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.concog.2013.06.004 F. Li et al. / Consciousness and Cognition 22 (2013) 920–930 921 There is good evidence that both chunks and specific encountered patterns are learned in implicit learning paradigms (e.g. Pothos & Bailey, 2000; Scott & Dienes, 2008), but linguists long ago rejected chunking as an explanation of language acquisition (e.g. Chomsky, 1959). They argued that natural language can only be acquired and processed by a mechanism that was able to deal with grammars more complex than finite state (and even finite-state grammars can involve more than chunking) (e.g. Gazdar, Klein, Pullum, & Sag, 1985; Joshi, Vijay-Shanker, & Weir, 1991; Steedman, 2000). In Chomsky’s (1959) hierarchy, finite-state grammars, context-free grammars, context-sensitive grammars and general phrase-structure grammars constitute an inclusion hierarchy. That is, each grammar in the hierarchy involves rules with restrictions, the restrictions being lifted as one goes up the hierarchy, so that grammars higher up can produce structures impossible lower down. For instance, with no length restrictions, context-free grammars, unlike finite-state grammars, can generate sentences where the last half is the reverse of the first (e.g., AAB-BAA, cf. Chomsky, 1956). With no length restrictions, context-sensitive grammars, unlike context-free grammars, can generate sentences where the last half is a copy of the first (e.g., AAB-AAB, cf. Chomsky, 1956). Copying and reversing are types of symmetries. Thus, symmetry has a structural complexity beyond finite-state. The Chomsky hierarchy is just one way of specifying complexity (see e.g., Van den Bos & Poletiek, 2008, 2010, for other measures of complexity in artificial grammar learning). It remains an open issue whether the Chomsky hierarchy happens to measure complexity in a psychologically relevant way, an issue we will be addressing by using symmetries (cf. Dienes & Longuet-Higgins, 2004; Westphal-Fitch, Huber, Gómez, & Fitch, 2012). The grammars above finite-state in the Chomsky hierarchy uniquely produce various symmetries. Symmetry occurs when transformation leaves a structure invariant; a mirror symmetry occurs when the transformation is reflection. For example, in virtue of exhibiting mirror symmetries, musical structures are analogous to certain linguistic structures. A retrograde symmetry in a melody, such as CEB-BEC (Balch, 1981; think of the music score for the first half being reflected in a vertical mirror to obtain the second half), corresponds to centre embedding in natural language (e.g. ‘‘The bamboo the panda ate was fresh’’, cf. Dienes, Kuhn, Guo, & Jones, 2012), and to a context-free grammar in Chomsky’s hierarchy (Chomsky, 1959; Fitch & Friederici, 2012; Hopcroft, Motwani, & Ullman, 2000; i.e. a level above finite state). People can acquire retrograde structures in at least one domain (natural language) and even other animals may be able to: Starlings might (Gentner, Fenn, Margoliash, & Nusbaum, 2006; but contrast e.g. Swaddle & Ruff, 2004); baboons might (Rey, Perruchet, & Fagot, 2012). However pigeons appear not to learn retrograde symmetries at all (Huber et al., 1999). So the structure is not consistently easy for any implicit learning mechanism. In situations that may be explicit, people have learned mirror retrogrades of sequences under lab conditions, when they were guided by staged-inputting (Conway, Ellefson, & Christiansen, 2003; Lai & Poletiek, 2011), salient perceptual cues (Mueller, Bahlmann, & Friederici, 2010), or intentional learning (Lai & Poletiek, 2011; Mueller et al., 2010). Distinctively implicit learning of retrograde structures still needs to be demonstrated (cf. Dienes & Longuet-Higgins, 2004, for suggestive evidence; see also Uddén, Ingvar, Hagoort, & Petersson, 2012, discussed below; and see Rohrmeier, Fu, & Dienes, 2012, for evidence of implicit learning of another type of contextfree grammar). In the most convincing evidence to date, Tanaka and Watanabe (in press) showed learning of the retrograde structure on an SRT task, where participants did not report the retrograde nature of the stimuli in post task free report. Another type of symmetry is an inversion, where the elements of a sequence preserve their order but each element is transformed (e.g. to an opposite) (Dienes & Longuet-Higgins, 2004; Jiang et al., 2012; Kuhn & Dienes, 2005). The inversion can be obtained by placing a mirror horizontally below a music score. The inversion corresponds to cross-serial dependencies in some natural languages, where a sequence of nouns is followed by a sequence of verbs in corresponding order (e.g., ‘‘Aad heft Jantje de lerares de knikkers laten helpen opruimen’’ in Dutch, literal: ‘‘Aad has Jantje the teacher the marbles let help collect up’’, gloss: ‘‘Aad let Jantje help the teacher collect up the marbles’’, cf. Christiansen & Chater, 1999); both inversions and cross-serial dependencies can be generated by a mildly context-sensitive grammar (Fitch & Friederici, 2012; Hopcroft et al., 2000; i.e. a level just above context free but not fully context sensitive, a term introduced by Joshi et al., 1991, to unite abstractly many formalisms emerging to describe natural language, e.g. Gazdar, 1988; Steedman, 2000). Kuhn and Dienes(2005) showed that participants learnt to like tunes instantiating a musical inversion, though they could not as sensitively classify the same tunes as rule governed or not. Thus, people can implicitly learn more than chunks of adjacent elements, and perhaps even acquire inversions per se (though Kuhn & Dienes, 2008, found a Simple Recurrent Network could learn the same material by learning a fixed length long distance association, a simpler structure than an inversion per se; cf. also Desmet, Poulin-Charronnat, Lalitte, & Perruchet, 2009, who raised possible confounds, albeit not ones that removed the learning effect of inversions when statistically controlled). Jiang et al. (2012) found that, controlling both chunks and repetition patterns (and the possible confounds raised by Desmet et al.), people could implicitly learn to discriminate nonlocal tonal inversions from non-inversions in artificial Chinese poetry. Jiang et al. thus provide a paradigm where highly controlled apparent implicit learning of symmetries can be found. The first aim of the present study was to investigate the implicit learning of retrograde structures using the Jiang et al. (2012) artificial Chinese poetry paradigm. The poetry they used is tonal. Chinese is a tonal language that uses four tones to signal different meanings; for example, the syllable ‘‘ma’’ pronounced in tone 1 means ‘‘mother’’, but ‘‘horse’’ when in tone 3. Tone 1, tone 2, tone 3, and tone 4 indicate flat, rising, falling–rising and falling phonetic characteristics in pitch respectively. Tone 1 and tone 2 are categorized into ping (level) tones, while tone 3 and tone 4 are categorized into ze (oblique) tones for the purposes of Chinese poetry. By virtue of the rising and falling intonation in words, Chinese is figuratively depicted as ‘‘the small waves adding on the large waves’’ (Chao, 1933), where each tone superimposes on the overall intonation pattern of a sentence. Tones are closely intertwined with meanings to achieve a musical and esthetic effect. In Jiang et al.’s paradigm, participants are asked to memorize artificial poems, constructed so that the Chinese tones in successive lines bear 922 F. Li et al. / Consciousness and Cognition 22 (2013) 920–930 a symmetry relation to each other. For an inversion relation, if a tone for a syllable in a certain position is one category (e.g. ping) in the first line, the syllable in the same position in the second line would be in the other category (e.g. ze). For a retrograde relation, the order of tone categories would be reversed in successive lines. After memorizing poems, participants are then informed that the poems were constructed according to a rule and asked to classify new poems as well formed or not, with half of poems instantiating the relevant symmetry. In Jiang et al. (2012), in order to assess the conscious status of the knowledge acquired, the ‘‘structural knowledge attributions’’ of Dienes and Scott (2005) were used. Specifically, after each classification decision, subjects indicated if the decision was based on a pure guess, intuition (they had some confidence but have no idea why), memory (they recollected or failed to recollect a sequence) or rules. A subject may learn that the lines of poetry are inversions, as shown by their tendency to classify new poetry as well-formed according to this feature, but not be aware that they knew this feature. Such a subject may insist their classification was just a guess, or based on intuition. By contrast, if subjects were aware of the basis of their classification, they could claim they followed memory or a rule. That is, unconscious knowledge of relevant structure prima facie exists when people say they are guessing or using intuition (implicit attributions); conscious knowledge of structure exists when people say they are using rules or recollection (explicit attributions) (see Dienes, 2012, for a review and evaluation of this method). Exploring the relative difficulty of learning different symmetries is important for evaluating models of implicit learning. Dienes and Longuet-Higgins (2004) discussed how, by contrasting the implicit learning of retrogrades and inversions, researchers could investigate the nature of the memory buffer required for processing structures beyond finite-state (which, by definition, require a buffer; e.g. Chomksy, 1963). A first in-first out buffer, which outputs material in the same order it was inputted, will facilitate detecting inversions, whereas a last in-first out buffer, which outputs material in the reverse order to which it was inputted, will facilitate detecting retrogrades (Christiansen & Chater, 1999; Dienes & Longuet-Higgins, 2004; Kuhn & Dienes, 2008; Uddén et al., 2012). Thus, investigating implicit learning of retrogrades and inversions is theoretically valuable for modelling in terms of constraining the computational properties of the memory buffer involved in the implicit learning of sequences. Whatever the exact architecture of the memory buffer, differences in difficulty between inverses and retrogrades may be explained in terms of the relative memory cost in processing them (compare natural language parsing for which centre embeddings (retrogrades) have been argued to be especially difficult because of memory cost, in syntactic prediction locality theory, Gibson, 1998). Uddén et al. (2012) demonstrated that Dutch participants performed better on materials instantiating inversions than retrogrades in an artificial grammar learning paradigm involving strings of letters (in which each letter in one section of the string was paired with a corresponding letter in another section; namely, F was paired with L and D with P). Nonetheless, the authors made no claims about the knowledge being of symmetrical structures (i.e., cross-serial dependencies/inversions or centre embeddings/retrogrades) per se, nor of the knowledge being unconscious. Specifically, although the associative chunk strength (ACS) for the grammatical and non-grammatical test strings was controlled, the repetition structures were not controlled, but the authors pointed out, differed between retrogrades and inversions. A repetition pattern is the pattern of letter repeats across a string, e.g. the pattern MTVTX can be represented as 12324 indicating that the second letter is repeated in the fourth position but all other letters are unique (Brooks & Vokey, 1991). Thus, participants may have simply memorized repetition patterns, a strategy for which there is evidence in artificial grammar learning (e.g. Tunney & Altmann, 2001). Participants did not report the symmetry rules in post task report; but perhaps they did not learn symmetry patterns, implicitly or explicitly. Thus, we controlled both chunks and repetition patterns as Jiang et al. (2012) did, to allow more focused interpretations of the content of the acquired knowledge. We also took trial by trial attribution ratings (specifically, those used by e.g. Chen et al., 2011; Dienes, Baddeley, & Jansari, 2012; Dienes & Scott, 2005; Guo et al., 2011; Jiang et al., 2012; Kemeny & Luckacs, 2013; Kiyokawa, Dienes, Tanaka, Yamada, & Crowe, 2012; Mealor & Dienes, 2012; Neil & Higham, 2012; Rebuschat, 2008; Rebuschat, Hamrick, Sachs, Riestenberg, & Ziegler, 2014; Wan, Dienes, & Fu, 2008) to sensitively measure the conscious status of knowledge on the fly. In sum, the present study presented participants with artificial poems to remember, where successive lines of the poems, for different groups, instantiated either retrogrades or inversions of the sequences of successive Chinese tones. Implicit learning was established using subjective measures of the conscious status of the structural knowledge used by participants (see Dienes, 2012, for the argument that these measures separate different knowledge types in theoretically expected ways). That is, subjects indicated whether the basis of their judgment on each trial was a guess, intuition, memory, or rules. The first aim was to establish whether retrogrades can be implicitly learned at all. The second aim was to establish whether retrogrades or inversions are easier, in order to explore the functional properties required of the memory buffer in implicit learning. 2. Method 2.1. Participants Ninety-four volunteers (70 females, M = 21.88, SD = 3.76) from East China Normal University took part in the experiment in exchange for credits or 20 RMB. All the participants were native Chinese speakers and none of them reported a history of hearing difficulties. They were randomly allocated to one of four groups, with 25 in the retrograde experimental (trained) F. Li et al. / Consciousness and Cognition 22 (2013) 920–930 923 group, 22 in the retrograde control (untrained) group, 25 in the inversion experimental (trained) group and 22 in the inversion control (untrained) group. 2.2. Materials Two grammars were used in this experiment: retrogrades (centre embedding) and inversions (cross-serial dependencies) based on two tone types (ping and ze tones). Specifically, a total of 20 tonal syllables were selected. The tone type of 10 tonal syllables was ping: ‘‘can1, chao1, hui1, ju1, shen1, di2, fo2, lai2, ping2, qin2’’, and the tone type of other 10 tonal syllables was ze: ‘‘er3, guo3, mai3, ye3, zhan3, jun4, kan4, tu4, wei4, zou4’’. Each string consisted of 10 different tonal syllables where the tone types (pings or zes) of the first five tonal syllables predicted the tone types of the following five tonal syllables by forming a retrograde or an inversion (see Figs. 1 and 2). Two sets of stimuli were developed according to the retrograde and inversion rules. For each rule, 8 grammatical tone type strings were generated, 4 of which served as training strings and 4 as test strings. In addition, for the retrograde rule, 4 ungrammatical tone type strings were created for the test phase by exchanging the tone type of the second element for that of the seventh element of the grammatical test strings, producing violations in the second (and ninth) and fourth (and seventh) positions (see Appendix A). For the inversion rule, 4 ungrammatical tone type strings were created for the test phase by exchanging the tone type of the seventh element for that of the ninth element of the grammatical test strings, producing the same violating positions as for the retrograde rule (see Appendix A). Each training tone type string was shown twelve times with different tonal syllables each time, with 48 training tonal syllable strings in all. Each test tone type string was shown six times with different tonal syllables each time, with 48 test tonal syllable strings in all (e.g., ungrammatical tone type string ‘‘ping ze ze ze ping - ping ping ze ping ping’’ generated tonal syllable strings ‘‘lai2 jun4 kan4 zhan3 shen1 di2 ju1 wei4 can1 hui1’’, ‘‘qin2 er3 wei4 zhan3 fo2 - hui1 ju1 kan4 di2 ping2’’ and so on.). None of the tonal syllable strings had a clear semantic interpretation. Although 32 tone type strings can be created by using all possible permutations of pings and zes to form a retrograde, we chose 8 grammatical tone type strings according to the following criteria: (1) The first element of the first half was not the inverse of first element of the second half; similarly, the last element of the first half was not the inverse of the last element of the second half, to avoid the tone type string forming an inversion, even a partial inversion. For example, the tone type string ‘‘ping Fig. 1. Examples of grammatical strings of the retrograde and inversion. Each string consists of 10 different tonal syllables. The tone types (pings or zes) of the previous five tonal syllables predict the following five tonal syllables by forming a retrograde or an inversion. (a) For retrograde, if the tone type of the first tonal syllable is ze, then the tone type of the tenth tonal syllable must be ze, and if the tone type of the second tonal syllable is ze, then the tone type of the ninth tonal syllable must be ze, and so on. (b) For inversion, if the tone type of the first tonal syllable is ping, then the tone type of the sixth tonal syllable must be ze, and if the tone type of the second tonal syllable is ping, then the tone type of the seventh tonal syllable must be ze, and so on. A1 A2 A3 A4 A5 B5 B4 B3 B2 B1 Retrograde A1 A2 A3 A4 A5 B1 B2 B3 B4 B5 Inversion Fig. 2. Retrograde and inversion rules used in this experiment. A and B could be any ping tones or ze tones. (a) For retrograde, any ping tones could go with any other ping tones in an AB pair, and any ze tones could go with any other ze tones in an AB pair. If A1 is ping, then B1 must be ping, and if A2 is ze, then B2 must be ze, and so on. (b) For inversion, any ping tones could go with any other ze tones in an AB pair, and any ze tones could go with any other ping tones in an AB pair. If A1 is ping, then B1 must be ze, and if A2 is ze, then B2 must be ping, and so on. 924 F. Li et al. / Consciousness and Cognition 22 (2013) 920–930 Table 1 Mean MFF and ACS for grammatical and ungrammatical strings for retrograde and inversion (M ± SD). Rule Measure Tone type Tones 1–4 Tonal syllable G UG G UG G UG Retrograde MFF GACS AACS 720.00 ± 0.00 224.75 ± 7.79 36.00 ± 0.00 720.00 ± 0.00 233.75 ± 0.00 36.00 ± 0.00 360.15 ± 1.48 49.08 ± 4.33 9.16 ± 2.51 360.20 ± 1.84 50.84 ± 3.00 9.38 ± 2.89 72.26 ± 2.09 2.00 ± 0.69 0.34 ± 0.54 72.05 ± 1.71 1.70 ± 0.60 0.50 ± 0.68 Inversion MFF GACS AACS 720.00 ± 0.00 224.75 ± 0.00 36.00 ± 0.00 720.00 ± 0.00 224.75 ± 7.79 36.00 ± 0.00 360.77 ± 2.21 48.91 ± 1.98 9.50 ± 2.96 360.20 ± 2.16 48.70 ± 2.36 9.31 ± 2.75 71.58 ± 2.48 1.67 ± 0.62 0.47 ± 0.58 71.95 ± 1.95 1.71 ± 0.67 0.31 ± 0.44 Note. G = grammatical, UG = ungrammatical; MFF = mean feature frequency, GACS = global associative chunk strength, AACS = anchor associative chunk strength. ze ping ping ze - ze ping ping ze ping’’ was excluded. (2) The sequence of tone types in the first half could not be the same as in the second half. For example, ‘‘ping ze ze ze ping - ping ze ze ze ping’’ was excluded, because it might be easily detected. Similarly, the criterion of choosing 8 grammatical tone type strings for the inversion rule was same as that of retrograde. For the materials for both retrograde and inversion strings, we controlled both repetition structure and chunks. None of the test strings had the same repetition structures, in terms of being a succession of tone types, as any of the training strings. Furthermore, none of the test grammatical strings had the same repetition structures, in terms of tones 1–4, as any of the training strings. With respect to tonal syllables, because each string consisted of 10 different tonal syllables, the repetition structure of tonal syllable strings was the same for all retrogrades and non-retrogrades and inversions and non-inversions. Furthermore, mean feature frequency (MFF), global associative chunk strength (GACS) and anchor associative chunk strength (AACS) were counterbalanced between grammatical and ungrammatical test tone type strings, ps > .05 (see Table 1) (see Knowlton & Squire, 1994: GACS is the average frequency with which each chunk in a test string appeared in training strings; AACS is the average frequency with which each chunk at the beginning and end of a test string appeared at the beginning and end of training strings). Grammatical and ungrammatical strings were also balanced along the same dimensions in terms of chunks of tonal syllables and also separately for tones 1–4 rather than tone types (see Table 1). The 20 tonal syllables, each lasting for 450 ms, were created by Chinese pronunciation software (Xunfei interphonic, cf. Jiang et al., 2012). For each tonal syllable string, a 600 ms interval was interposed between the fifth and sixth tonal syllables to create a perceptual gap between the first half of the tonal syllable string and its retrograde or inversion in the final half (cf. Mueller et al., 2010). Thus, each tonal syllable string lasted for 5100 ms. 2.3. Procedure There were two phases: training phase and test phase. Only the two experimental groups received the training phase; all the participants received the test phase (cf. Dienes & Altmann, 2003; Jiang et al., 2012). 2.3.1. Training phase Participants in the experimental groups were asked to listen to 144 tonal syllable strings in all, which consisted of 48 grammatical tonal syllable strings presented three times in a different random order for each participant. In each trial, a warning tone was presented for 500 ms, followed by a 5100 ms tonal syllable string and a 5000 ms blank. Participants were instructed to listen to each tonal syllable string carefully and silently repeat it during the 5000 ms delay before the next trial. The training phase lasted about 30 min. 2.3.2. Test phase During the test phase, participants were informed that the tonal syllable strings that they heard in the training phase were generated using a specific rule and were asked to listen to 48 new tonal syllable strings presented in a random order. For each tonal syllable string, they were asked to judge whether the given string was grammatical and attribute their decision basis to four categories (guess, intuition, memory and rule). As defined to participants, ‘‘Guess’’ indicated that the judgment was based on nothing at all, it could just as well be based on a toss of a coin; ‘‘Intuition’’ indicated that the judgment was based on a hunch or feeling that could not be explicated further, i.e. there was confidence in the judgment but the person had no idea why the judgment was right; ‘‘Memory’’ indicated that the judgment was based on a recollection; ‘‘Rule’’ indicated that the judgment was based on a rule that could be stated if asked. 3. Results 3.1. Proportion of correct responses C þ0:5 The proportion of correct response was calculated by NNþ1 (NC being the number of correct responses; and N the total number of responses), the correction corresponding to a Bayesian prior of chance performance worth just one observation, useful when some participants have low N for some conditions (as used in e.g. Dienes & Scott, 2005; cf. Baguley, 2012, p. 83). 925 F. Li et al. / Consciousness and Cognition 22 (2013) 920–930 For the retrograde groups, the classification performance of experimental and control groups were 0.52 (SD = 0.06) and 0.49 (SD = 0.05), respectively. Participants in the experimental group performed significantly better than the control group, t(45) = 2.31, p < .05, d = 0.67, with sequential Bonferroni correction. For the inversion groups, the classification performance of experimental and control groups were 0.56 (SD = 0.05) and 0.50 (SD = 0.06), respectively. Participants in the experimental group performed significantly better than the control group, t(45) = 3.84, p < .001, d = 1.12, with sequential Bonferroni correction, see Fig. 3. For the control groups, the classification performance of retrograde and inversion groups were not significantly different, t(42) = 0.68, p > .05, but for the experimental groups, the difference was significant, t(48) = 2.37, p = .022, d = 0.67, and remained significant after Bonferronni correction, see Fig. 3. Learning the inversion was easier than learning the retrograde. 3.2. Unconscious structural knowledge of tonal retrograde and inversion The implicit nature of the knowledge can be assessed by the ‘‘structural knowledge attributions’’ of Dienes and Scott (2005). Specifically, after each classification decision, subjects indicated if the decision was based on a pure guess, intuition (they have some confidence but have no idea why), rule, or memory. A subject may learn that the lines of poetry are inversions, as shown by their tendency to classify new poetry as well-formed according to this feature, but not be aware that they know this feature. Such a subject may insist their classification was just a guess, or based on intuition. By contrast, if subjects were aware of the basis of their classification, they could claim they followed memory or a rule. That is, unconscious knowledge of relevant structure prima facie exists when people say they are guessing or using intuition (implicit attributions); conscious knowledge of structure when people say they are using rule or recollection (explicit attributions) (see Dienes, 2012, for a review and evaluation of this method). According to the practice of Dienes and Scott (2005), guess and intuition attributions were combined as indicators of unconscious structural knowledge (implicit attributions), and memory and rule attributions were combined as indicators of conscious structural knowledge (explicit attributions). The response proportions of each attribution for retrograde and inversion are shown in Table 2. Because only four participants chose the memory category and none of the participants chose the rule category, we only analyzed the performance of implicit attributions rather than performance of explicit attribuC þ0:5 tions.The proportion of correct response for implicit attributions was calculated by the formula NNþ1 (NC being the number of correct responses when the participants chose implicit attributions; and N the total number of implicit attributions). For retrogrades, the classification performance of implicit attributions for experimental and control groups were 0.52 (SD = 0.06) Experimental group Control group Percentage of correct responses 0.7 0.6 ** * * 0.5 0.4 0.3 0.2 0.1 0.0 Inversion Retrograde Fig. 3. Percentage of correct responses for the four groups on overall performance (⁄p < .05, ⁄⁄p < .01). Error bars indicate standard error of the mean. Table 2 Response proportions of each attribution for the four groups (M ± SD). Rule Group Implicit attribution Explicit attribution Guess Intuition Memory Rule Retrograde Experimental Control 0.34 ± 0.27 0.40 ± 0.27 0.66 ± 0.27 0.60 ± 0.27 0.00 ± 0.01 0.00 ± 0.00 0.00 ± 0.00 0.00 ± 0.00 Inversion Experimental Control 0.23 ± 0.23 0.60 ± 0.34 0.77 ± 0.23 0.39 ± 0.34 0.00 ± 0.01 0.01 ± 0.02 0.00 ± 0.00 0.00 ± 0.00 Note. Only four participants chose memory (two in retrograde experimental group, one in inversion experimental group and one in inversion control group, in ten trials in all); none of the participants chose rule. 926 F. Li et al. / Consciousness and Cognition 22 (2013) 920–930 Experimental group Control group Percentage of correct responses 0.7 0.6 ** * * 0.5 0.4 0.3 0.2 0.1 0.0 Inversion Retrograde Fig. 4. Percentage of correct responses for the four groups on performance of implicit attributions (⁄p < .05, ⁄⁄p < .01). Error bars indicate standard error of the mean. and 0.49 (SD = 0.05), respectively. Implicit knowledge for the experimental group differed significantly from the control group, t(45) = 2.17, p < .05, d = 0.63, indicating unconscious structural knowledge of the tonal retrograde. For inversions, the classification performance of implicit attributions for experimental and control groups were 0.56 (SD = 0.05) and 0.49 (SD = 0.06), respectively. Implicit knowledge for the experimental group differed significantly from the control group, t(45) = 3.87, p < .001, d = 1.13, indicating unconscious structural knowledge of the tonal inversion. For the control groups, the classification performance of implicit attributions for retrogrades and inversions was not significantly different, t(42) = 0.57, p > .05, but for the experimental groups, the difference was significant, t(48) = 2.37, p = .022, d = 0.67, see Fig. 4, indicating implicit learning was easier for inversions than retrogrades. 4. Discussion The aim of the present study was to investigate the implicit learning of Chinese tonal retrogrades (centre embedding) and inversions (cross-serial dependencies). We provided clear evidence that people can acquire unconscious structural knowledge of retrogrades and inversions. Crucially we also showed that implicit learning was easier for inversions than retrogrades. The results coincide with those of previous studies arguing that people can go beyond learning chunks and repetition patterns and implicitly learn to detect patterns instantiating symmetries, i.e. nonlocal patterns produced by recursive rules (Dienes & Longuet-Higgins, 2004; Jiang et al., 2012; Kuhn & Dienes, 2005; Uddén et al., 2012). We provided evidence for the knowledge being unconscious by taking trial by trial attributions for the basis of the judgments. Almost all attributions were guess or intuition, i.e. claims by the participant that either the judgment had no basis or else they had no idea what it was. Dienes (2012) reviewed evidence that such attributions, at least in the case of the finite state grammars typically used in the implicit learning literature, separate qualitatively different types of knowledge, types that differ in theoretically expected ways. For example, implicit attributions are more resistant to concurrent executive-demanding secondary tasks than explicit attributions. Thus, the attributions have evidential support as a means for measuring the conscious status of knowledge. The retrograde and inversion rules in the present study take the form of what Marcus (2001) calls ‘‘operations over variables’’. That is, an inversion or a retrograde is an operation that applies to a vector, and can in principle apply to a vector of any length (Dienes & Longuet-Higgins, 2004). We used stimuli of fixed length, so people may not have learnt the inversion or symmetry per se (especially given the low level of learning for retrogrades). For example, for the inversion, people may have learnt an association between corresponding positions in successive lines (Kuhn & Dienes, 2008). For the retrograde, people may have just learned to associate a specific position in one line (e.g. the 4th) with a specific position in the next line (e.g. the 7th). Such knowledge would enable accurate classification if the position which was learnt was one of the positions we altered as violations (i.e., the 2nd, 4th, 7th and 9th positions, but not otherwise). Thus, such long distance associations could explain the advantage of learning inversions over retrogrades (because in inversions, but not retrogrades, the associations occur at the same fixed distance for all positions), without people having actually learnt the symmetries per se. Nevertheless, the learning we have demonstrated still goes beyond what has been shown in artificial grammar learning before (but see especially Remillard, 2010, for learning long distance contingencies in the SRT task). In order to show that people have learnt the symmetries per se, the next step would be to show people can generalize their knowledge to poems with different length lines. It would also be important to show that retrograde performance can be increased to a more impressive level, for example by repeating the training phase, to increase confidence it was the retrograde that was learnt and not a structure that happened to correlate with it in the current materials (Dulany, 1962). F. Li et al. / Consciousness and Cognition 22 (2013) 920–930 927 If people have learnt the symmetries per se, then they found the grammar higher in the Chomsky’s hierarchy (inversions, mildly context sensitive) easier than the grammar lower down (retrogrades, context free). The hierarchy is just a way of classifying grammars formally, though it is meant as a useful measure of complexity (Fitch & Friederici, 2012); our results tentatively suggest the measure is not psychologically realistic. Dienes and Longuet-Higgins (2004) pointed that a first in-first out buffer would facilitate detecting inversion rather than retrograde. Working memory appears in general to be first in-first out (backward span is harder than forward span). The actual buffer used need not be strictly either first in-first out, or last infirst out; for example the buffer in a Simple Recurrent Network (SRN) is computationally flexible with its exact properties constructed according to the error landscape it is trained in. Nonetheless, a given memory system may function more like a first in-first out or last in-first out buffer, or find it easier to be trained to act like one of those rather than the other. In this sense, our results suggest that functionally the memory buffer used in implicit learning is more like a first in-first out buffer than a last in-first out buffer, which is consistent with Uddén et al. (2012) in this respect. The exact architecture of the memory buffer involved in implicit learning, however, is an open question. Kuhn and Dienes (2008) have investigated the nature of the buffer involved in implicit learning of nonlocal dependencies by modelling with the SRN (cf. e.g. Carting, 2008, for SRN modelling of learning phrase-structure grammars). In as yet unpublished work we have found that the SRN can learn the two structures and it replicates the advantage of inversions over retrogrades (Li, 2013). Thus, functionally the SRN acts, or finds it easier to be trained to act, more like a first in-first out buffer than a last in-first out buffer. An even more flexible type of memory (such as the random access memory of a desktop computer) can act as a first in-first out buffer or a last in-first out buffer with equal facility, so it cannot by its architectural properties explain an advantage of inverses over retrogrades. Future research is needed to determine whether the SRN-type buffer really matches the computational properties of the buffer used in implicit learning (cf. Cleeremans, 1993; Jones & Mclaren, 2009; see also French, Addyman, & Mareschal, 2011, for why the SRN-type buffer needs to be changed in some learning contexts). It is not clear whether prior expectations of structure instantiating inversions versus retrogrades could over-ride the effect of the type of buffer the system uses. Backward span is harder than forward span in standard working memory tasks even when people fully expect to perform the task (e.g. Robinson, Mervis, & Robinson, 2003). Nonetheless, Chinese Tang poetry uses an inversion and not retrograde relation between tone classes, knowledgeour participants were likely exposed to as children. That prior knowledge was not sufficient for them to perform well in the untrained control groups. But there is evidence that implicit learning is sensitive to prior knowledge (e.g. Chen et al., 2011; Leung & Williams, 2011; Ziori & Dienes, 2008), and knowledge of Tang poetry may have biased the implicit learning mechanism to be sensitive to pick up inversions rather than retrogrades. Future research should investigate the role of prior knowledge in the relative ease of inversions versus retrogrades. The symmetries were defined over non-terminal symbols, i.e. over the ping-ze classes of tones (tones being the terminals). However, we cannot be sure based on the data presented here that people induced the rules over those classes or rather learnt specific associations between the constituent tones. The classes constitute likely prior knowledge for participants. Jiang (2012) used an arbitrary classification of the four tones into two categories and found no learning of the inversion rule. Thus, the use of ping-ze classes appears an important contribution to learning, a claim we plan to explore further with experiments and modelling. In our materials we used a correspondence between same elements for retrogrades (ping was mapped to ping) and a correspondence between different elements for inversions (ping was mapped to ze). Thus, the rules differed not only in instantiating centre embeddings versus cross-serial dependencies, but also in terms of how corresponding elements were mapped (i.e. as same versus different). However, the extra difference in mapping different rather than same elements should make the inverse more difficult than the retrograde; yet the results showed that inverses were learnt more easily than retrogrades (compare Dowling, 1972, who found the same for explicit judgments in music). In music, transposes (i.e. copies of intervals) are considerably easier to detect than inverses (Krumhansl, Sandell, & Sergeant, 1987), although both instantiate a cross-serial dependency structure. Indeed, there is evidence that people have a general facility for detecting sameness, at the level of well-used perceptual categories (e.g. Endress, Dehaene-Lambertz, & Mehler, 2007). As we found cross-serial dependencies to be easier than centre embeddings, in spite of the former using a mapping between different elements, our claim for their relative ease of learning stands. Future research could use a copy rather than an inversion to instantiate a cross-serial dependency (so ping would map to ping and ze to ze); or an inverse retrograde (a retrograde where pings are transformed to zes and vice versa) for a centre embedding, to test our claim further. To conclude, controlling both chunks and repetition patterns, we showed that people could implicitly learn Chinese tonal retrogrades and inversions, abstract nonlocal rules. The results suggest that the buffer used in implicit learning functions more like a first in-first out buffer than a last in-first out buffer. Together with Jiang et al. (2012), we provide a paradigm where the nature of the memory buffer in implicit learning can be explored. Acknowledgments This research was supported by Academic Scholarship for Doctoral Candidates in ECNU (XRZZ2011008), National Natural Science Foundation of China (31271090), Key Project Foundation for Research and Innovation of Shanghai Municipal Education Commission (12ZS046), and Interuniversity Attraction Poles Program of the Belgian Federal Science Policy Office (grant 7/33). 928 F. Li et al. / Consciousness and Cognition 22 (2013) 920–930 Appendix A. Tone type strings of the retrograde and inversion. Sort Retrograde Inversion 1 1 1 1 2 2 2 2 3 3 3 3 pppzp-pzppp pzppp-pppzp zzzpz-zpzzz zpzzz-zzzpz ppzzp-pzzpp pzzpp-ppzzp zzppz-zppzz zppzz-zzppz pzzzp-ppzpp ppzpp-pzzzp zpppz-zzpzz zzpzz-zpppz pppzp-zzzpz pzppp-zpzzz zzzpz-pppzp zpzzz-pzppp ppzzp-zzppz pzzpp-zppzz zzppz-ppzzp zppzz-pzzpp ppzzp-zppzz pzzpp-zzppz zzppz-pzzpp zppzz-ppzzp Note. p = ping, z = ze; 1 = grammatical tone type strings in the training phase, 2 = grammatical tone type strings in the test phase, and 3 = ungrammatical tone type strings in the test phase. References Baguley, T. (2012). Serious stats: A guide to advanced statistics for the behavioral sciences. Basingstoke, England: Palgrave Macmillan. Balch, W. R. (1981). The role of symmetry in the good continuation ratings of two-part tonal melodies. Perception and Psychophysics, 29, 47–55. http:// dx.doi.org/10.3758/BF03198839. Brooks, L. R., & Vokey, J. R. (1991). Abstract analogies and abstracted grammars: Comments on Reber (1989) and Mathews et al. (1989). Journal of Experimental Psychology: General, 120, 316–323. http://dx.doi.org/10.1037/0096-3445.120.3.316. Carting, B. (2008). On the implicit acquisition of a context-free grammar by a simple recurrent neural network. Neurocomputing, 71, 1527–1537. http:// dx.doi.org/10.1016/j.neucom.2007.05.006. Chao, Y. (1933). A preliminary study of English intonation and its Chinese equivalents. BIHP Supplement, 1. Chen, W., Guo, X., Tang, J., Zhu, L., Yang, Z., & Dienes, Z. (2011). Unconscious structural knowledge of form-meaning connections. Consciousness and Cognition, 20, 1751–1760. http://dx.doi.org/10.1016/j.concog.2011.03.003. Chomsky, N. (1956). Three models for the description of language. IRE Transactions on Information Theory, 2(3), 113–124. http://dx.doi.org/10.1109/ TIT.1956.1056813. Chomsky, N. (1959). On certain formal properties of grammars. Information and Control, 2, 137–167. http://dx.doi.org/10.1016/S0019-9958(59)90362-6. Chomksy, N. (1963). Formal properties of grammars. In R. D. Luce, R. R. Bush, & E. Galanter (Eds.). Handbook of mathematical psychology (Vol. II, pp. 323–418). New York: Wiley. Christiansen, M. H., & Chater, N. (1999). Toward a connectionist model of recursion in human linguistic performance. Cognitive Science, 23, 157–205. http:// dx.doi.org/10.1207/s15516709cog2302_2. Cleeremans, A. (1993). Mechanisms of implicit learning: Connectionist models of sequence processing. Cambridge, MA: MIT Press. Cleeremans, A. (2006). Conscious and unconscious cognition: A graded, dynamic, perspective. In Q. Jing, M. R. Rosenzweig, G. d’Ydewalle, H. Zhang, H.-C. Chen, & K. Zhang (Eds.), Progress in psychological science around the world. Neural, cognitive and developmental issues (Vol. I, pp. 401–418). Hove, England: Psychology Press. Conway, C. M., Ellefson, M. R., & Christiansen, M. H. (2003). When less is less and when less is more: Starting small with staged input. In Proceedings of the 25th annual conference of the cognitive science society (pp. 270–275). Mahwah, NJ: Lawrence Erlbaum. Desmet, C., Poulin-Charronnat, B., Lalitte, P., & Perruchet, P. (2009). Implicit learning of nonlocal musical rules: A comment on Kuhn and Dienes (2005). Journal of Experimental Psychology: Learning, Memory, and Cognition, 35, 299–305. http://dx.doi.org/10.1037/a0014278. Dienes, Z. (2012). Conscious versus unconscious learning of structure. In P. Rebuschat & J. Williams (Eds.), Statistical learning and language acquisition (pp. 337–364). Mouton de Gruyter. Dienes, Z., & Altmann, G. (2003). Measuring learning using an untrained control group: Comment on R. Reber and P. Perruchet. The Quarterly Journal of Experimental Psychology A, 56, 117–123. http://dx.doi.org/10.1080/02724980244000305. Dienes, Z., Baddeley, R. J., & Jansari, A. (2012). Rapidly measuring the speed of unconscious learning: Amnesics learn quickly and happy people slowly. PLoS ONE, 7, e33400. http://dx.doi.org/10.1371/journal.pone.0033400. Dienes, Z., Kuhn, G., Guo, X., & Jones, C. (2012). Communicating structure, affect and movement: Commentary on Bharucha, Curtis & Paroo. In P. Rebuschat, M. Rohrmeier, J. A. Hawkins, & I. Cross (Eds.), Language and music as cognitive systems (pp. 156–168). Oxford, England: Oxford University Press. Dienes, Z., & Longuet-Higgins, C. (2004). Can musical transformations be implicitly learned? Cognitive Science, 28, 531–558. http://dx.doi.org/10.1016/ j.cogsci.2004.03.003. Dienes, Z., & Scott, R. (2005). Measuring unconscious knowledge: Distinguishing structural knowledge and judgment knowledge. Psychological Research, 69, 338–351. http://dx.doi.org/10.1007/s00426-004-0208-3. Dowling, W. J. (1972). Recognition of melodic transformations: Inversion, retrograde and retrograde inversion. Perception and Psychophysics, 12, 417–421. http://dx.doi.org/10.3758/BF03205852. Dulany, D. E. (1962). The place of hypotheses and intentions: An analysis of verbal control in verbal conditioning. In C. W. Eriksen (Ed.), Behavior and awareness (pp. 102–129). Durham, NC: Duke University Press. Endress, A. D., Dehaene-Lambertz, G., & Mehler, J. (2007). Perceptual constraints and the learnability of simple grammars. Cognition, 105, 577–614. http:// dx.doi.org/10.1016/j.cognition.2006.12.014. Fitch, W. T., & Friederici, A. D. (2012). Artificial grammar learning meets formal language theory: An overview. Philosophical Transactions of the Royal Society B, 367, 1933–1955. http://dx.doi.org/10.1098/rstb.2012.0103. French, R. M., Addyman, C., & Mareschal, D. (2011). TRACX: A recognition-based connectionist framework for sequence segmentation and chunk extraction. Psychological Review, 118, 614–636. http://dx.doi.org/10.1037/a0025255. F. Li et al. / Consciousness and Cognition 22 (2013) 920–930 929 Gazdar, G. (1988). Applicability of indexed grammars to natural languages. In U. Reyle & C. Rohrer (Eds.), Natural language parsing and linguistic theories (pp. 69–94). Dordrecht Neth: Reidel Publishing. Gazdar, G., Klein, E., Pullum, G. K., & Sag, I. A. (1985). Generalized phrase structure grammar. Cambridge, MA: Harvard University Press. Gentner, T. Q., Fenn, K. M., Margoliash, D., & Nusbaum, H. C. (2006). Recursive syntactic pattern learning by songbirds. Nature, 440, 1204–1207. http:// dx.doi.org/10.1038/nature04675. Gibson, E. (1998). Linguistic complexity: Locality of syntactic dependencies. Cognition, 68, 1–76. http://dx.doi.org/10.1016/S0010-0277(98)00034-1. Guo, X., Zheng, L., Zhu, L., Yang, Z., Chen, C., & Zhang, L. (2011). Acquisition of conscious and unconscious knowledge of semantic prosody. Consciousness and Cognition, 20, 417–425. http://dx.doi.org/10.1016/j.concog.2010.06.015. Hopcroft, J. E., Motwani, R., & Ullman, J. D. (2000). Introduction to automata theory, languages and computation (2nd ed.). Reading, MA: Addison-Wesley. Huber, L., Aust, U., Michelbach, G., Olzant, S., Loidolt, M., & Nowotny, R. (1999). Limits of symmetry conceptualization in pigeons. The Quarterly Journal of Experimental Psychology B, 52, 351–379. http://dx.doi.org/10.1080/027249999393040. Jamieson, R. K., & Mewhort, D. J. K. (2009). Applying an exemplar model to the artificial-grammar task: Inferring grammaticality from similarity. The Quarterly Journal of Experimental Psychology, 62, 550–575. http://dx.doi.org/10.1080/17470210802055749. Jiang, S. (2012). Implicit learning of Chinese tonal regularity (Unpublished doctoral dissertation). Shanghai, China: East China Normal University. Jiang, S., Zhu, L., Guo, X., Ma, W., Yang, Z., & Dienes, Z. (2012). Unconscious structural knowledge of tonal symmetry: Tang poetry redefines limits of implicit learning. Consciousness and Cognition, 21, 476–486. http://dx.doi.org/10.1016/j.concog.2011.12.009. Jones, F. W., & McLaren, I. P. L. (2009). Human sequence learning under incidental and intentional conditions. Journal of Experimental Psychology: Animal Behavior Processes, 35, 538–553. http://dx.doi.org/10.1037/a0015661. Joshi, A. K., Vijay-Shanker, K., & Weir, D. J. (1991). The convergence of mildly context-sensitive formalisms. In P. Sells, S. M. Shieber, & T. Wasow (Eds.), Processing of linguistic structure (pp. 31–81). Cambridge, MA: MIT Press. Kemeny, F., & Luckacs, A. (2013). Self-insight in probabilistic category learning. The Journal of General Psychology, 140, 57–81. http://dx.doi.org/10.1080/ 00221309.2012.735284. Kiyokawa, S., Dienes, Z., Tanaka, D., Yamada, A., & Crowe, L. (2012). Cross cultural differences in unconscious knowledge. Cognition, 124, 16–24. http:// dx.doi.org/10.1016/j.cognition.2012.03.009. Knowlton, B. J., & Squire, L. R. (1994). The information acquired during artificial grammar learning. Journal of Experimental Psychology: Learning, Memory, and Cognition, 20, 79–91. http://dx.doi.org/10.1037/0278-7393.20.1.79. Krumhansl, C. L., Sandell, G. J., & Sergeant, D. C. (1987). The perception of tone hierarchies and mirror forms in twelve-tone serial music. Music Perception, 5, 31–78. http://dx.doi.org/10.2307/40285385. Kuhn, G., & Dienes, Z. (2005). Implicit learning of nonlocal musical rules: Implicitly learning more than chunks. Journal of Experimental Psychology: Learning, Memory, and Cognition, 31, 1417–1432. http://dx.doi.org/10.1037/0278-7393.31.6.1417. Kuhn, G., & Dienes, Z. (2008). Learning non-local dependencies. Cognition, 106, 184–206. http://dx.doi.org/10.1016/j.cognition.2007.01.003. Lai, J., & Poletiek, F. H. (2011). The impact of adjacent-dependencies and staged-input on the learnability of center-embedded hierarchical structures. Cognition, 118, 265–273. http://dx.doi.org/10.1016/j.cognition.2010.11.011. Leung, J., & Williams, J. (2011). Constrains on implicit learning of form-meaning connections. Language Learning, 62, 634–662. http://dx.doi.org/10.1111/ j.1467-9922.2011.00637.x. Li, F. (2013). Implicit learning of Chinese tonal symmetry rules and the neural network simulations (Unpublished doctoral dissertation). Shanghai, China: East China Normal University. Manza, L., & Reber, A. S. (1997). Representing artificial grammars: Transfer across stimulus forms and modalities. In D. Berry (Ed.), How implicit is implicit learning? (pp 73–106). Oxford, England: Oxford University Press. Marcus, G. (2001). The algebraic mind. Cambridge, MA: MIT Press. Mealor, A. D., & Dienes, Z. (2012). Conscious and unconscious thought in artificial grammar learning. Consciousness and Cognition, 21, 865–874. http:// dx.doi.org/10.1016/j.concog.2012.03.001. Mueller, J. L., Bahlmann, J., & Friederici, A. D. (2010). Learnability of embedded syntactic structures depends on prosodic cues. Cognitive Science, 34, 338–349. http://dx.doi.org/10.1111/j.1551-6709.2009.01093.x. Neil, G. J., & Higham, P. A. (2012). Implicit learning of conjunctive rule sets: An alternative to artificial grammars. Consciousness and Cognition, 21, 1393–1400. http://dx.doi.org/10.1016/j.concog.2012.07.005. Perruchet, P., & Vinter, A. (1998). PARSER: A model for word segmentation. Journal of Memory and Language, 39, 246–263. http://dx.doi.org/10.1006/ jmla.1998.2576. Poletiek, F. H. (2002). Learning recursion in an artificial grammar. Acta Psychologica, 111, 323–335. http://dx.doi.org/10.1016/S0001-6918(02)00057-4. Pothos, E. M., & Bailey, T. M. (2000). The role of similarity in artificial grammar learning. Journal of Experimental Psychology: Learning, Memory, and Cognition, 26, 847–862. http://dx.doi.org/10.1037/0278-7393.26.4.847. Reber, A. S. (1967). Implicit learning of artificial grammars. Journal of Verbal Learning and Verbal Behaviour, 6, 855–863. http://dx.doi.org/10.1016/S00225371(67)80149-X. Reber, A. S. (1989). Implicit learning and tacit knowledge. Journal of Experimental Psychology: General, 118, 219–235. http://dx.doi.org/10.1037//00963445.118.3.219. Rebuschat, P. (2008). Implicit learning of natural language syntax (Unpublished doctoral dissertation). Cambridge, England: University of Cambridge. Rebuschat, P., Hamrick, P., Sachs, R., Riestenberg, K., & Ziegler, N. (2014). Implicit and explicit knowledge of form-meaning connections: Evidence from subjective measures of awareness. In J. Bergsleithner, S. Frota, & J. K. Yoshioka (Eds.), Noticing: L2 studies and essays in honor of Dick Schmidt. Honolulu, HI: University of Hawaii at Manoa, National Foreign Language Resource Center. Reed, N., McLeod, P., & Dienes, Z. (2010). Implicit knowledge and motor skill: What people who know how to catch don’t know. Consciousness and Cognition, 19, 63–76. http://dx.doi.org/10.1016/j.concog.2009.07.006. Remillard, G. (2010). Implicit learning of fifth- and sixth-order sequential probabilities. Memory and Cognition, 38, 905–915. http://dx.doi.org/10.3758/ MC.38.7.905. Rey, A., Perruchet, P., & Fagot, J. (2012). Centre-embedded structures are a by-product of associative learning and working memory constraints: Evidence from baboons (Papio Papio). Cognition, 123, 180–184. http://dx.doi.org/10.1016/j.cognition.2011.12.005. Robinson, B. F., Mervis, C. B., & Robinson, B. W. (2003). The roles of verbal short-term memory and working memory in the acquisition of grammar by children with Williams syndrome. Developmental Neuropsychology, 23, 13–31. http://dx.doi.org/10.1080/87565641.2003.9651885. Rohrmeier, M., Fu, Q., & Dienes, Z. (2012). Implicit learning of recursive, hierarchical grammatical structures. PLoS ONE, 7, e45885. http://dx.doi.org/10.1371/ journal.pone.0045885. Rohrmeier, M., & Rebuschat, P. (2012). Implicit learning and acquisition of music. Topics in Cognitive Science, 4, 525–553. http://dx.doi.org/10.1111/j.17568765.2012.01223.x. Rohrmeier, M., Rebuschat, P., & Cross, I. (2011). Incidental and online learning of melodic structure. Consciousness and Cognition, 20, 214–222. http:// dx.doi.org/10.1016/j.concog.2010.07.004. Saffran, J. R., Newport, E. L., Aslin, R. N., Tunick, R. A., & Barrueco, S. (1997). Incidental language learning: Listening (and learning) out of the corner of your ear. Psychological Science, 8, 101–105. http://dx.doi.org/10.1111/j.1467-9280.1997.tb00690.x. Scott, R. B., & Dienes, Z. (2008). The conscious, the unconscious, and familiarity. Journal of Experimental Psychology: Learning, Memory, and Cognition, 34, 1264–1288. http://dx.doi.org/10.1037/a0012943. Shanks, D. R. (2005). Implicit learning. In K. Lamberts & R. Goldstone (Eds.), Handbook of cognition (pp. 202–220). London, England: Sage. Steedman, M. J. (2000). The syntactic process. Cambridge, MA: MIT Press. 930 F. Li et al. / Consciousness and Cognition 22 (2013) 920–930 Swaddle, J. P., & Ruff, D. A. (2004). Starlings have difficulty in detecting dot symmetry: Implications for studying fluctuating asymmetry. Behavior, 141, 29–40. http://dx.doi.org/10.1163/156853904772746583. Tanaka, K., & Watanabe, K. (in press). Implicit transfer of reversed temporal structure in visuomotor sequence learning. Cognitive Science. Tillman, B., Bharucha, J. J., & Bigand, E. (2000). Implicit learning of tonality: A self-organizing approach. Psychological Review, 107, 885–913. http://dx.doi.org/ 10.1037//0033-295X.107.4.885. Tunney, R. J., & Altmann, G. T. M. (2001). Two modes of transfer in artificial grammar learning. Journal of Experimental Psychology: Learning, Memory, and Cognition, 27, 614–639. http://dx.doi.org/10.1037/0278-7393.27.3.614. Uddén, J., Ingvar, M., Hagoort, P., & Petersson, K. M. (2012). Implicit acquisition of grammars with crossed and nested non-adjacent dependencies: Investigating the push-down stack model. Cognitive Science, 36, 1078–1101. http://dx.doi.org/10.1111/j.1551-6709.2012.01235.x. Van den Bos, E. J., & Poletiek, F. H. (2008). Effects of grammar complexity on artificial grammar learning. Memory and Cognition, 36, 1122–1131. http:// dx.doi.org/10.3758/MC.36.6.1122. Van den Bos, E. J., & Poletiek, F. H. (2010). Structural selection in implicit learning of artificial grammars. Psychological Research, 74, 138–151. http:// dx.doi.org/10.1007/s00426-009-0227-1. Westphal-Fitch, G., Huber, L., Gómez, J. C., & Fitch, W. T. (2012). Production and perception rules underlying visual patterns: Effects of symmetry and hierarchy. Philosophical Transactions of the Royal Society B, 367, 2007–2022. http://dx.doi.org/10.1098/rstb.2012.0098. Wan, L., Dienes, Z., & Fu, X. (2008). Intentional control based on familiarity in artificial grammar learning. Consciousness and Cognition, 17, 1209–1218. http:// dx.doi.org/10.1016/j.concog.2008.06.007. Williams, J. N. (2009). Implicit learning. In W. C. Ritchie & T. K. Bhatia (Eds.), New handbook of second language acquisition (pp. 319–353). Bingley, England: Emerald Group Publishing Ltd.. Ziori, E., & Dienes, Z. (2008). How does prior knowledge affect implicit and explicit concept learning? The Quarterly Journal of Experimental Psychology, 61, 601–624. http://dx.doi.org/10.1080/17470210701255374.

Consciousness and Cognition 42 (2016) 15–25 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Continuity of phenomenology and (in)consistency of content of meaningful autobiographical memories Martina Luchetti a,⇑, Nicolino Rossi b, Ornella Montebarocci b, Angelina R. Sutin a a b Florida State University College of Medicine, Department Behavioral Sciences and Social Medicine, 1115 West Call Street, Tallahassee, FL 32306, USA University of Bologna, Department of Psychology, Viale Berti Pichat 5, Bologna 40126, Italy a r t i c l e i n f o Article history: Received 14 October 2015 Revised 13 January 2016 Accepted 14 February 2016 Keywords: Autobiographical memory Continuity Memory Experiences Questionnaire Phenomenology Test-retest interval a b s t r a c t Phenomenology is a critical component of autobiographical memory retrieval; it reflects both (a) memory-specific features and (b) stable individual differences. Few studies have tested phenomenology longitudinally. The present work examined the continuity of memory phenomenology in a sample of Italians adults (N = 105) over a 4-week period. Participants retrieved two ‘key’ personal memories, a Turning Point and an Early Childhood Memory, rated the phenomenology of each memory, and completed measures of personality, psychological distress and subjective well-being. Phenomenological ratings were moderately stable over time (median correlation >.40), regardless of memory content. Personality traits, psychological distress and well-being were associated with phenomenology cross-sectionally and with changes in phenomenology over time. These results suggest that how individuals re-experience their most important personal memories is relatively consistent over time and shaped by both trait and state aspects of psychological functioning. Ó 2016 Elsevier Inc. All rights reserved. 1. Introduction Subjective experience (phenomenology) is a critical component of autobiographical memory retrieval (Sutin & Robins, 2007; Tulving, 2002). When an individual remembers an important event from his/her life (e.g., the birth of a child, wedding day, etc.), he/she may see the scene as it is re-happening and re-experience the emotions felt at the time. An individual may consistently rate one particular memory as extremely vivid and emotionally intense (or, likewise, consistently dim and fragmented) either because of the visual clarity and emotional charge of that particular event or because all of his/her memories tend to be vivid and emotionally intense. When an individual retrieves a memory, its phenomenology reflects both (a) memory- and event-specific features and (b) stable individual characteristics (Singer & Salovey, 1993). Phenomenology is a dynamic feature of autobiographical recollection that may fluctuate based on the importance of the event in the memory, the relevance to the individual’s current goals, and/or the emotional state of the individual at the time of retrieval. Few studies, however, have assessed autobiographical memory longitudinally. Most longitudinal work has focused more on the consistency and accuracy of memory content than how phenomenology fluctuates over time. Hirst et al. (2015), for example, examined memories of the attack of September 11, 2001 at 1 week and at 1, 3 and 10 years after the attack and found marked inconsistencies for canonical features (e.g., how did you first learn about the attack?) across the follow-ups. The content of memories often changes over time, especially for specific information (e.g. locations, activities, ⇑ Corresponding author. E-mail addresses: martina.luchetti@med.fsu.edu (M. Luchetti), nicolinocesare.rossi@unibo.it (N. Rossi), ornella.montebarocci@unibo.it (O. Montebarocci), angelina.sutin@med.fsu.edu (A.R. Sutin). http://dx.doi.org/10.1016/j.concog.2016.02.011 1053-8100/Ó 2016 Elsevier Inc. All rights reserved. 16 M. Luchetti et al. / Consciousness and Cognition 42 (2016) 15–25 etc.; Drivdahl & Hyman, 2014). Even when reporting the most meaningful personal life events, less than 20% of specific events are repeated across interviews (McAdams et al., 2006; Thorne, Cutting, & Skaw, 1998). Still, many individuals identify the same experience as a ‘key event’ even after several years (e.g., Bauer, Tasdemir-Ozdes, & Larkina, 2014; Leonard & Burns, 2006). Whether or not an individual recalls the same or a different experience, relatively consistent narrative themes are observed across sessions (McAdams et al., 2006; Thorne et al., 1998) and the emotions and motivations retrieved tend to be consistent across memories and over time (Sutin & Robins, 2005). Consistency may extend to phenomenological characteristics of the memory (e.g., vividness, emotional intensity, etc.). Rubin, Schrauf, and Greenberg (2004), for example, examined participants’ ratings of several phenomenological qualities over a short-term interval. Participants (N = 30 students) recalled and rated 20 events twice over a two-week period. The consistency of phenomenology was moderate (r = .50) within participant ratings of the 20 events and the stability of memory phenomenology was high (r = .80) across the two sessions. Recently, Thomsen, Hammershøj Olesen, Schnieber, Jensen, and Tønnesvang (2012) showed that memories for important life events maintained or slightly increased in their levels of pleasantness, emotional intensity, and importance compared to trivial control memories over a 3-month period, and were still rated as important after three years (Thomsen et al., 2015). Meaningful memories tend to remain vividly detailed even after several years (Luchetti, Montebarocci, Rossi, Cutti, & Sutin, 2014), though the specific details may shift (Hirst et al., 2015). There is also stability in the proportion of specific versus overgeneral memories retrieved by participants in response to cue-words up to several years (Sumner et al., 2014). Less work has examined the full range of phenomenological qualities to address whether some aspects of phenomenology are more stable than others. How individuals retrieve and re-experience their memories may be shaped, in part, by trait (e.g., underlying personality dispositions) and state (e.g., symptoms of anxiety and depression) psychological functioning (e.g., Rasmussen & Berntsen, 2010; Rubin, Boals, & Berntsen, 2008; Rubin & Siegler, 2004; Sutin & Robins, 2010). For instance, individuals who score high on Openness to experience—the tendency to explore inner worlds through ideas, fantasies and feelings—tend to report stronger reliving qualities when recalling an autobiographical event (Rasmussen & Berntsen, 2010; Rubin & Siegler, 2004). For example, open individuals tend to retrieve memories that are perceived as vivid and coherent and central to their life stories (Rasmussen & Berntsen, 2010). Individuals high in Neuroticism, who have a negative worldview, tend to retrieve memories that are more negative in valence and emotionally intense and psychologically distance themselves from their more positive memories (Sutin, 2008). For individuals high in Conscientiousness, their tendency to be organized extends to how they retrieve their past experiences; that is, they tend to perceive their memories as vivid and coherent (Sutin, 2008). There are no consistent associations between Extraversion and Agreeableness and phenomenology, and no studies have examined the five traits in relation to changes in phenomenology. The phenomenology of autobiographical memory has been linked to psychological distress (e.g., Sutin & Gillath, 2009) and clinical disorders (see Sumner, 2012; Williams et al., 2007 for reviews). Some phenomenological dimensions (e.g., specificity/coherence) play a critical role in the development and maintenance of clinical conditions, such as depression and posttraumatic stress disorder (e.g., Boelen, Huntjens, & van den Hout, 2014; Sumner, Mineka, & McAdams, 2013). Symptoms of distress are generally associated with the retrieval of less phenomenologically intense memories (less vivid, coherent and detailed memories; Luchetti & Sutin, 2015). Recollection of personal experiences that are phenomenologically-rich, in contrast, may increase an overall sense of well-being and life satisfaction (Latorre et al., 2013; Sutin, 2008; Waters & Fivush, 2014). The aim of the present work is to examine the stability of the phenomenology of autobiographical memory across memories and over time. Specifically, participants were asked to retrieve two ‘key’ moments of their life—an important life-changing event (Turning Point Memory) and a remote event (Early Childhood Memory)—and rate the affect and phenomenology of each memory at two sessions separated by 4 weeks. In addition, we examine the relation between participants’ phenomenological experience of their memories and five-factor personality traits, psychological distress and well-being and whether changes in memory phenomenology are associated with these aspects of psychological functioning. 2. Method 2.1. Participants and procedure The study utilized an on-line survey created with Qualtrics software (www.qualtrics.com) and distributed by electronic mail to a community sample from Italy. Participants provided informed consent in person at the time of recruitment. They were asked to complete a memory task (Block 1)—i.e. a Turning Point and an Early Childhood Memory—and a series of selfreport questionnaires (Block 2) on two separate occasions. Once the first survey was completed (Time 1), a second link was sent again to each respondent after 4 weeks (Time 2). On average, it took about an hour to complete the survey. The order of memory recall and the order of survey blocks were counterbalanced across participants.1 Ethical approval for the study was 1 Forty-three percent of the sample completed the Memory Task (Block 1) first and 57% of the sample completed the Questionnaires (Block 2) first. Half of the participants recalled the Turning Point Memory first, while the other half recalled the Early Childhood Memory first. Independent t-tests were performed to test order effects on the memory measures; only 8 of 56 tests were significant. M. Luchetti et al. / Consciousness and Cognition 42 (2016) 15–25 17 obtained from the ethical committee of the University of Bologna, and the Institutional Review Board at the Florida State University also approved the analyses. A total of 105 participants responded at Time 1 (38.1% males; age M = 28.7, SD = 3.9, range 18–39; years of education M = 17.7, SD = 2.9); 72.4% (N = 76) responded at Time 2. The time (days) between the two sessions varied across participants (M = 39.4, SD = 12.8). Due to missing data, 100 participants answered at least one of the memory requests at Time 1, and 69 answered at least one memory at Time 2. A total of 65 and 61 participants had valid Turning Point and Childhood Memory, respectively, at both sessions. Those who did not respond at Time 2 tended to be less educated (t = 2.57, p = .012; M = 16.5, SD = 2.5 vs. M = 18.1, SD = 2.9 years of education) than those tested at both time points; no differences were observed in the proportion of males (40%), workers (50%) and married participants (5%) between respondents and non-respondents. Non-respondents also had shorter narratives for turning point memories (t = 2.26, p = .027; M = 77.8, SD = 62.4 vs. M = 117.3, SD = 107.3 words) and more accessible childhood memories (t = 2.59, p = .011; M = 4.3, SD = 0.8 vs. M = 3.7, SD = 1.0). 2.2. Memory task Using an adaptation of the Life Story Interview instructions (McAdams, 2008), participants were asked to recall two ‘key’ moments from their lives. Specifically, they were asked to write about an important life-changing event (i.e., Turning Point Memory): ‘‘In looking back over your life, it may be possible to identify certain key moments that stand out as turning points— episodes that marked an important change in you or your life story. Please identify a particular episode in your life story that you now see as a turning point in your life. If you cannot identify a key turning point that stands out clearly, please describe some event in your life wherein you went through an important change of some kind. Please describe what happened, where and when, who was involved, and what you were thinking and feeling.” And, to describe an important event from their early childhood years (i.e., Early Childhood Memory): ‘‘Think about an event from your early childhood that stands out in some way. Please describe this event in detail. What happened, where and when, who was involved, and what were you thinking and feeling?” Memory requests were translated into Italian using the standard translation/back-translation method. At both Time 1 and Time 2 sessions, participants were asked to rate the affective and phenomenological experience of each memory and report their age at the time of the recalled event. The instructions at Time 2 were identical to those used at Time 1; participants were not told whether or not to write about the same experience. The events in the memories reported at Time 2 were coded as either the ‘same’ or ‘different’ from the events reported at Time 1. A memory was considered to refer to the same event when the core experience described in the memory was the same (e.g., ‘‘moving to Florida”). A second judge independently coded whether the event was the same or not for a subset (20%) of the memories; there was high agreement with the first judge (k P .90 for both memory types). 2.2.1. Affect After describing each memory, participants rated their emotions during recall on a 5-point response scale (from 1 = very slightly or not at all to 5 = extremely). They rated six positive (proud, inspired, exited, strong, determined, enthusiastic) and six negative emotions (upset, scared, ashamed, hostile, guilty, and distressed) drawn from the Positive and Negative Affect Schedule (PANAS; Watson, Clark, & Tellegen, 1988). The individual emotions were composited into Positive Affect (PA) and Negative Affect (NA) scales for each memory. At both time points, alpha reliabilities of PA and NA were above .80 for both memories. 2.2.2. Phenomenology Participants completed the 31-item short form of the Memory Experiences Questionnaire (MEQ; Luchetti & Sutin, 2015) that assessed Vividness, Coherence, Accessibility, Time Perspective, Sensory Details, Visual Perspective, Emotional Intensity, Sharing, Distancing, and Valence on 5-point response scale (from 1 = strongly disagree to 5 = strongly agree) for each memory type. At Time 1, alpha reliability ranged from .20 (Accessibility) to .94 (Valence) for the Turning Point Memory (median = .73) and from .60 (Sensory Details) to .90 (Valence) for the Childhood Memory (median = .77). At Time 2, alpha reliability ranged from .34 (Sensory Details) to .91 (Sharing, Distancing and Valence) for the Turning Point Memory (median = .76) and from .60 (Sensory Details) to .94 (Valence) for the Childhood Memory (median = .81). 2.2.3. Content Narratives were coded into several category based on the nature and the valence of the recalled event. Specifically, we coded the memory for: (1) interpersonal conflicts (e.g., breakups, conflicts with close others, etc.), (2) death of a close other, (3) accident, injury or illness, (4) accident, injury or illness of a close other, (5) physical assaults (e.g., being attacked by strangers), (6) failures or struggles in a skill-related or personal domain (e.g., failing a course), (7) geographic separations (e.g., moving away from close others, living alone for university), (8) positive relationships (e.g., dating, falling in love, marriage, positive moments with close others), (9) recreation or exploration (e.g., hobbies, travel experiences, vacations), (10) skillrelated achievements or attaining personal goals (e.g., completing a degree, receiving recognition or an award), or (11) other (k P 74 for both memories). We also coded the memories for positive, negative, or mixed valence (i.e., not clearly positive or negative) (k P 62 for both memories). 18 M. Luchetti et al. / Consciousness and Cognition 42 (2016) 15–25 2.3. Other measures 2.3.1. Personality At Time 1, participants completed the Italian translation of the Big Five Inventory (BFI; Fossati, Borroni, Marchione, & Maffei, 2011; John & Srivastava, 1999), a 44-item measure of Neuroticism, Extraversion, Openness to experience, Agreeableness, and Conscientiousness. Respondents were asked to rate each item (e.g. ‘‘I see Myself as Someone Who. . .” ‘‘Is depressed, blue”) on a 5-point response scale (from 1 = strongly disagree to 5 = strongly agree). Alpha reliabilities ranged from .69 (Agreeableness) to .83 (Conscientiousness and Openness). 2.3.2. Psychological distress At both time points, participants completed the Patient Health Questionnaire depression module (PHQ-9; Mazzotti et al., 2003; Spitzer, Kroenke, Williams, & the Patient Health Questionnaire Study Group, 1999). They rated the frequency of 9 depressive symptoms (e.g. ‘‘feeling tired or having little energy”) on a 4-point response scale (from 0 = not at all to 3 = nearly every day). 21.2% of the present sample scored P the threshold score of 10 for moderate to severe symptomatology at Time 1 (Kroenke, Spitzer, & Williams, 2001); alphas were P.80 at both Time 1 and Time 2. Participants also completed an 11-item measure of loneliness rated on a 3-point response scale (from 1 = hardly ever or never to 3 = often), which has been used in large sample surveys (see Smith et al., 2013); alpha reliabilities were P.90 at both time points. At the end of each session, they completed a multiple-choice checklist to indicate any potential stressful events (i.e. natural disaster, accident, lifethreatening illness or injury, etc.) that occurred within the last month. A yes/no variable was subsequently created to indicate the presence/absence of any of the listed events. 2.3.3. Subjective well-being Participants answered the 8-item Flourishing Scale (FS; Diener et al., 2009) and a single item on life satisfaction (i.e. ‘‘I am satisfied with my life”) on a 5-point response scale (from 1 = strongly disagree to 5 = strongly agree). The FS provides an overall index of well-being and positive functioning; higher scores reflect respondents’ positive perception of themselves in various areas of functioning—e.g. experiencing positive relationships and having purpose in life. Alphas were above P.80 at both time points. 3. Results 3.1. Continuity across memories at Time 1 We first compared characteristics of the Turning Point and Childhood Memory at Time 1. The average number of words for each memory was not significantly different (about 100 words for each memory, p > .05). Participants recalled turning points that were closer to their current age (at about 20 years of age) than their early childhood memories. Geographic separations (e.g., living alone for university) were the most common turning point events (22%), followed by personal achievements (17%), failures (or struggles) in a skill-related domain (14%), and interpersonal conflicts (11%); the other events were less common (Table 1). Participants’ childhood memories were generally from their early childhood (at about 5 years of age); most of the sample recalled a positive moment with close others (35.8%), followed by interpersonal conflicts (15.8%). Overall, turning points tended to be more mixed in valence than the early memories (34.0% vs. 15.8%, respectively). A repeated measure Analysis of Variance (ANOVA) was performed on the memory ratings with Type of Event (Turning Point vs. Early Childhood Memory) as the within-subject factor (Table 2).2 Participants tended to report more PA (but not NA) in the turning point than in the early memories [F(1, 94) = 12.73, p = .001, partialg2 = .12]. Turning points were also rated higher in phenomenology than the childhood memories: They were rated as more vivid [F(1, 94) = 63.19, p < .001, partialg2 = .40], coherent [F (1, 94) = 11.21, p = .001, partialg2 = .11], accessible [F(1, 94) = 20.57, p < .001, partialg2 = .18], emotionally intense [F(1, 94) = 20.97, p < .001, partialg2 = .18], as having more sensory details [F(1, 94) = 16.69, p < .001, partialg2 = .15] and a clearer time perspective [F (1, 94) = 81.27, p < .001, partialg2 = .46], and as being more likely to be shared with others [F(1, 94) = 16.21, p < .001, partialg2 = .15] and retrieved from a first-person visual perspective [F(1, 94) = 6.72, p = .011, partialg2 = .07]. However, when recency (i.e. the difference in years between age at the time of Turning Point and age at the time of Early Childhood Memory) was included as a covariate in the analyses, the observed differences were reduced to non-significance. To test the continuity of memory affect and phenomenology, each of the PANAS and MEQ scales were correlated with its respective scale across the two memories (see Table 2). Correlations were generally positive and moderate in size. PA reported in the turning point memory was significantly correlated with PA in the early childhood memory; there was no correlation between NA across the two memories. Vividness and Emotional Intensity were the phenomenological dimensions most consistent across the memories, followed by Sensory Details and Distancing. This pattern suggests that phenomenological dimensions such as Vividness or Emotional Intensity were less memory-specific than person specific. That 2 When examining gender as between-subject factor, no significant main effect was observed on MEQ ratings; however, a significant Type of Event  Gender interaction was found for Coherence [F(1, 93) = 9.08, p = .003, partialg2 = .09] and Time Perspective [F(1, 93) = 4.68, p = .033, partialg2 = .05]. That is, for turning points, females tended to score higher than men on these scales. In contrast, males tended to report higher scores on NA scale than females for the childhood memories [F(1, 93) = 9.87, p = .002, partialg2 = .10]. 19 M. Luchetti et al. / Consciousness and Cognition 42 (2016) 15–25 Table 1 Percentages of event type and valence for Turning Point and Early Childhood Memory. TP (N = 100) EC (N = 95) Event type 1. Interpersonal conflicts (e.g., breakups, conflicts with close others) 2. Death (e.g., death of a close other by illness, murder, or suicide) 3. Accidents, injuries or illnesses 4. Close others’ accidents, injuries, or illnesses 5. Physical assaults (e.g., by strangers) 6. Failure or struggles in skill-related domains or in attaining personal goals (e.g., failing a course, getting fired) 7. Geographic separation from close others (e.g., moving away from close others, living along for university) 8. Positive relationships (e.g., dating, falling in love, marriage, positive moments with family and friends) 9. Recreation or exploration (e.g., hobbies, travel experiences, vacations) 10. Skill-related achievements or attaining personal goals (e.g., completing a degree) 11. Other experiences 11.0 10.0 3.0 1.0 – 14.0 22.0 8.0 7.0 17.0 7.0 15.8 5.3 6.3 5.3 – 7.4 2.1 35.8 7.4 9.5 5.3 Event valence Positive Negative Mixed 31.0 35.0 34.0 46.3 37.9 15.8 Note. TP = Turning Point Memory, EC = Early Childhood Memory. Table 2 Affect and phenomenology across memories and over time. Variables Time 1 Time 2 TP Age at the event No. of words PA NA Vividness Coherence Accessibility Time perspective Sensory details Visual perspective Emotional intensity Sharing Distancing Valence TP vs ECa EC TP EC M (SD) a M (SD) a Cohen’s d r M (SD) a M (SD) a 20.6 107.9 3.0 1.6 4.2 4.2 4.4 3.9 3.6 3.8 4.2 2.9 3.2 3.7 (5.5) (100.8) (1.2) (0.8) (0.8) (0.9) (0.6) (0.9) (0.8) (1.1) (0.9) (1.3) (1.2) (1.4) – – .88 .82 .67 .72 .20 .57 .51 .84 .75 .87 .89 .94 5.8 102.0 2.5 1.6 3.3 3.7 3.8 2.7 3.1 3.4 3.7 2.3 3.0 3.7 (2.4) (85.0) (1.1) (0.9) (1.0) (1.0) (1.0) (1.2) (0.9) (1.1) (1.0) (1.1) (1.2) (1.4) – – .86 .85 .80 .73 .73 .77 .60 .78 .76 .86 .86 .90 3.51*** 0.06 0.45*** 0.03 0.96*** 0.47*** 0.68*** 1.21*** 0.52*** 0.34* 0.48*** 0.55*** 0.11 0.06 .25* .63*** .23* .06 .32** .07 .08 .14 .24* .17 .47*** .11 .24* .03 20.3 90.6 2.8 1.6 3.9 3.8 4.2 3.6 3.3 4.0 3.9 2.9 3.4 3.7 (5.4) (78.0) (1.2) (0.8) (0.9) (0.9) (0.8) (0.9) (0.7) (1.0) (0.8) (1.2) (1.2) (1.5) – – .88 .84 .83 .65 .60 .65 .34 .88 .70 .91 .91 .91 6.2 95.2 2.3 1.4 3.3 3.7 3.8 2.7 3.1 3.6 3.3 2.0 3.1 3.7 (2.9) (81.4) (1.1) (0.7) (1.1) (1.0) (0.9) (1.2) (0.8) (1.1) (1.1) (1.1) (1.2) (1.5) – – .86 .81 .80 .72 .77 .76 .60 .91 .82 .87 .87 .94 Note. TP = Turning Point Memory, EC = Early Childhood Memory. At Time 1, 100 participants had valid TP and 95 valid EC; 65 and 61 had respectively valid TP and EC at both time points. a ds for mean differences between TP and EC were computed; in addition, each PANAS and MEQ scale was correlated (Pearson’s r) with its respective scale across the two memories. * p < .05. ** p < .01. *** p < .001. is, participants with highly vivid and emotional memories may retrieve any given memory as visually clear and emotionally charged, regardless of the recency or the type of the retrieved event. 3.2. Continuity of the events recalled and phenomenological ratings over time At Time 2, participants tended to report a different experience than at Time 1: 49.2% of the sample wrote about a different turning point and 68.9% wrote about a different childhood event. We tested for mean differences between Time 1 and Time 2 taking into account whether or not participants wrote about a ‘same’ or ‘different’ event: A 2 (Time 1 vs. Time 2)  2 (Same vs. Different Event) repeated measures Analysis of Variance (ANOVA) was performed on the memory ratings.3 Given the number of 3 Only part of the sample reported to had experienced stressful events in the month immediately before the beginning of the study (37.6%), or during the time interval (7.6%). When examining the presence of stress as between-subject factor, a significant main effect of stress was found on coherence ratings of the turning point events [F(1, 59) = 6.52, p = .013, partialg2 = .10]; that is, those who experienced a recent stressful event reported a more specific turning point. A significant Time  Stress interaction was observed for Emotional Intensity of turning points [F(1, 59) = 6.68, p = .012, partialg2 = .10]; those without a recent stress (but not those with recent stress) decreased in emotional intensity at Time 2. No main effects of stress or interactive effects of stress with time were found for childhood events. 20 M. Luchetti et al. / Consciousness and Cognition 42 (2016) 15–25 days between sessions varied across participants, we controlled for the interval length by including the number of days between Time 1 and Time 2 as a covariate in the analyses. For the Turning Point Memory, no effects of Time were found; however, a significant interaction Time  Same/Different Event was observed for Accessibility [F(1, 61) = 4.48 p = .038, partialg2 = .07], Time Perspective [F(1, 61) = 4.02, p = .049, partialg2 = .06] and Emotional Intensity [F(1, 61) = 6.82, p = .011, partialg2 = .10]. When a different Turning Point was retrieved at Time 2, participants rated their second memory as less accessible, temporally clear, and emotionally intense. For the Early Childhood Memory, instead, we found a significant effect of Time for Sharing [F(1, 57) = 11.75, p = .001, partialg2 = .17] and Distancing [F(1, 57) = 6.51, p = .013, partialg2 = .10]; that is, there was a decrease in Sharing and an increase in Distancing at Time 2 [Cohen’s d = 0.28 and 0.19, respectively]. A significant effect of Same/Different Event was found for Time [F(1, 57) = 10.39, p = .002, partialg2 = .15] and Visual Perspective [F(1, 57) = 4.31, p = .042, partialg2 = .07]; those who wrote about the same events had a clearer time perspective and retrieved their memories from a first-person perspective [Cohen’s d = 0.78 and 0.47 for Time and Visual Perspective, respectively]. A significant Time  Same/Different Event interaction was also found for Distancing [F(1, 57) = 8.30, p = .006, partialg2 = .13]; in this case, though lower at Time 1, Distancing ratings tended to increase for those who reported the same Childhood Memory at Time 2. Despite the observed mean-level differences, test–retest correlations for the PANAS and MEQ scales were moderately strong and generally significant (median correlation = .41 for turning point and =.45 for early childhood memories), especially for those who wrote about the same experience at Time 2 (Table 3; median correlation for reporting the same memory = .71; median correlation for reporting a different memory = .27). Even if the content changed, memory affect and phenomenology remained fairly stable over time. Of the 10 phenomenology dimensions, Emotional Intensity and Distancing were the most stable over time (rs > 40 for turning point and P50 for childhood memories), followed by Sharing and Valence (rs P 53 for turning point and P44 for childhood memories). The least stable dimensions were Accessibility (for both Turning Point and Early Childhood Memory) and Coherence and Visual Perspective (for Early Childhood Memory only). 3.3. Correlates of phenomenology 3.3.1. Personality Correlations between personality and the memory variables for the full sample (N = 105) are shown in Table 4. Given the age and gender differences in personality and phenomenology, correlations were corrected for these variables (zero-order correlations are reported in the online Supplementary Material). There were a few significant correlations between personality, PANAS and MEQ scores. Participants who scored higher in Conscientiousness tended to retrieve turning point memories that had more positive and less negative emotions and that were vivid, shared with others, and positively valenced; this trait was also associated with more vivid and sensory detailed early memories. Participants who scored high in Neuroticism reported more negative emotions, tended not to share their turning points with others, and retrieved a more negatively valenced event; Neuroticism was unrelated to phenomenology in the childhood memory. Of the remaining traits, Extraversion was associated with more positive affect and less negative affect in the turning point memory but not the childhood memory, and Openness was correlated with Sensory Details of the childhood memory but not the turning point memory. Finally, individuals who scored higher in Agreeableness recalled less vivid and coherent turning point memories that were positively valenced; Agreeableness was also negatively related to Sharing in the early memory. 3.3.2. Psychological distress PA and NA of the Turning Point Memory (but not of Early Childhood Memory) showed, respectively, negative and positive associations with the distress measures. Participants with more depressive symptoms tended to rate their turning points as negatively valanced, to see them from an observer prospective and to share them less frequently with others. Those who felt lonely also reported less frequently shared memories and more negative life-changing experiences. The only significant association between distress and the memory variables for the Early Childhood Memory was a negative correlation between Loneliness and Sharing. 3.3.3. Subjective well-being PA of Turning Point Memory was positively correlated with subjective well-being, while NA was negatively associated with the well-being measures. Participants with high scores on positive functioning and life satisfaction rated their turning points as accessible and positive and tended to share them often. The only association observed for the Early Childhood Memory was a positive correlation between positive functioning and Sharing. 3.3.4. Predicting change in memory affect and phenomenology over time To test whether personality was associated with change in affect and phenomenology over time, each personality trait was correlated with the residual change scores, computed by predicting the Time 2 memory ratings from the corresponding Time 1 ratings and saving the residuals. For the Turning Point Memory, Neuroticism and Conscientiousness were inversely associated with change in Distancing (r = .26, p = .043 for Neuroticism; r = .29, p = .020 for Conscientiousness), which indicated that individuals high in Neuroticism and low in Conscientiousness tended to progressively distance themselves from their life-changing experiences. For the Early Childhood Memory, participants high in Neuroticism tended to report less PA (r = .26, p = .047), Sharing (r = .33, p = .010) and Valence (r = .35, p = .007) in their second memories than in their first 21 M. Luchetti et al. / Consciousness and Cognition 42 (2016) 15–25 Table 3 Correlations of memory variables over time. Dimension Age at the event No. of words Positive affect Negative affect Vividness Coherence Accessibility Time perspective Sensory details Visual perspective Emotional intensity Sharing Distancing Valence Sample size TP EC Overall Same Different Overall Same Different .49*** .54*** .51*** .65*** .31* .35** .11 .34** .42*** .32* .47*** .62*** .41*** .53*** 65 .99*** .62*** .80*** .84*** .55*** .57*** .31 .73*** .44* .32 .73*** .67*** .57*** .91*** 33 .08 .41* .13 .42* .20 .19 .01 .09 .41* .32 .26 .55*** .28 .14 32 .80*** .73*** .41*** .47*** .48*** .26* .23 .62*** .41*** .17 .50*** .46*** .59*** .44*** 61 .96*** .76*** .75*** .86*** .82*** .71*** .48* .86*** .61** .38 .72*** .80*** .66** .85*** 19 .61*** .68*** .29 .28 .36* .11 .14 .40** .38* .10 .44** .25 .62*** .27 42 Note. TP = Turning Point Memory, EC = Early Childhood Memory. A total of 65 and 61 participants had respectively valid Turning Point and Early Childhood Memory at both sessions. * p < .05. ** p < .01. *** p < .001. memories. Conscientiousness was also associated with an increase in Valence (r = .29, p = .026) in the childhood memories. There were no other significant correlations between the personality traits and the residual change scores. Whether or not the memory reported at Time 2 was different from the memory reported at Time 1 did not moderate any of these effects. To test whether changes in phenomenology were associated with changes in psychological distress and subjective wellbeing, change scores for memory ratings were correlated with the residual change scores of each of the psychological measures. Increases in depressive symptoms over time were associated with decreases in Positive Affect (r = .31, p = .016) and Valence (r = .26, p = .039) for the Turning Point Memory (but not Early Childhood Memory). Increases in loneliness were also associated with increases in NA (r = .30, p = .019) for the Turning Point Memory. Finally, increases in positive functioning and life satisfaction were associated with increases in PA (r = .35, p = .007 and r = .28, p = .036, respectively) but only for the Early Childhood Memory. 4. Discussion The present study examined the continuity of autobiographical memory phenomenology across memories and over time. Participants retrieved two meaningful memories, a Turning Point and an Early Childhood Memory and rated the subjective experience of each memory on two occasions separated by 4 weeks. Results showed moderate stability in phenomenology ratings comparable to what was found in previous studies (Rubin et al., 2004). Dimensions such as Emotional Intensity were more consistent than dimensions such as Visual Perspective across memories and over time. In addition to content and memory-specific features (e.g., recency), individual differences in psychological functioning were also associated with the subjective experience of memories. That is, personality, psychological distress and well-being were associated with phenomenology and change in phenomenology over the retest interval. While some studies have shown a high percentage of similar events reported over time (Bauer et al., 2014; Jack & Hayne, 2010), others have found low consistency (McAdams et al., 2006; Sutin & Robins, 2005; Thorne et al., 1998). In the present study, participants were just as likely to retrieve a memory of a different event as the same one at Time 2. As suggested by Drivdahl and Hyman (2014), consistency rates may depend on the type of event remembered. For example, in their study on relationship memories, participants were fairly consistent in the recollections of first meetings and first dates, perhaps because early events are more frequently rehearsed in the context of a couples’ story. Even when recalling the same event, however, details may change: Memories are malleable, and personal narratives are frequently updated with additional information from new experiences (Drivdahl & Hyman, 2014). There was relative continuity in memory affect and phenomenology regardless of whether participants reported a memory of the same or different event at Time 2 than at Time 1: The median overall test–retest coefficient was above .40 for both the Turning Point and Early Childhood memories. When participants reported the same events at both time points, the retest correlations approached the magnitude typically found for stable individual differences (median r was .71). Correlations were moderately strong even when memories of different events were retrieved across the two sessions (i.e. median r was .27 for ‘different’ event), with only a few coefficients approaching zero: Accessibility (for both Turning Point and Early Childhood Memory) and Coherence and Visual Perspective (for Early Childhood Memory only). As such, some phenomenology dimensions were rated high (or low) independent of the particular event recalled in the memory. Any given memory was 22 M. Luchetti et al. / Consciousness and Cognition 42 (2016) 15–25 Table 4 Partial correlations of autobiographical memory with personality, psychological distress and well-being variables controlling for age and sex (total N = 105). Memory variables Age at the event No. of words Positive affect Negative affect Vividness Coherence Accessibility TP TP TP TP TP TP TP EC EC EC EC EC EC EC Personality Neuroticism Extraversion Openness Agreeableness Conscientiousness .03 .19 .13 .11 .09 .01 .02 .15 .03 .02 .12 .13 .21* .11 .04 .01 .06 .18 .25* .06 .15 .22* .14 .15 .36*** .06 .06 .07 .10 .16 .20* .23* .07 .09 .32*** .11 .01 .15 .15 .08 .03 .19 .01 .21* .22* .14 .11 .02 .10 .30** .19 .06 .05 .27** .00 .09 .02 .08 .01 .17 .12 .18 .02 .05 .11 .20 .13 .19 .02 .08 Psychological distress Depression Loneliness .15 .03 .11 .13 .07 .09 .02 .03 .34*** .31** .09 .05 .35*** .44*** .02 .05 .06 .01 .15 .09 .16 .08 .11 .07 .08 .12 .10 .08 Subjective well-being Positive functioning Life satisfaction .10 .11 .06 .02 .04 .10 .07 .05 .38*** .38*** .01 .05 .36*** .40*** .08 .10 .02 .05 .09 .08 .10 .08 .15 .02 .23* .20* .18 .14 Note. TP = Turning Point Memory, EC = Early Childhood Memory. A total of 100 and 95 participants had respectively valid Turning Point and Childhood Memory at Time 1; 104 participants completed the questionnaires section. * p < .05. ** p < .01. *** p < .001. likely retrieved as emotionally intense, for example, regardless of the recency or the type of event recalled. This pattern suggests stable individual differences in the way individuals retrieve their memories. It is worth noting that even though the alpha reliabilities of some of the MEQ scales—in particular, Accessibility, Sensory Details and Time Perspective—were poor for the Turning Point Memory, the magnitude of retest correlations of these scales was adequate when participants recalled the ‘same’ event at Time 2 (>.30 for Accessibility, >.40 for Sensory Details and >.70 for Time Perspective). In the case of Sensory Details, for example, items heterogeneity likely affected internal consistency: The items of this scale assess different aspects of the construct such as the presence of sounds, smells, and tastes in the memory, and the reliving sensation associated with it. Nonetheless, these same items were scored similarly when re-administered at Time 2: Participants consistently experienced the recalled memories as high (or low) in Sensory Details over time. These results suggest that participants’ responses were unlikely to be random. To date, relatively few studies have explored the association between personality and memory phenomenology (though see Rasmussen & Berntsen, 2010; Rubin & Siegler, 2004). And, to our knowledge, none has correlated personality with phenomenology changes. Of the five traits, Conscientiousness and Neuroticism had the most associations with the PANAS and MEQ dimensions. These traits are generally linked with measures of episodic memory and cognitive functioning (Luchetti, Terracciano, Stephan, & Sutin, 2015; Wilson et al., 2015) and are associated with psychological outcomes related to memory characteristics (e.g., depression; Kotov, Gamez, Schmidt, & Watson, 2010). As expected (see Sutin, 2008), Conscientiousness was related to the rehearsal of more positive and detailed autobiographical recollections; and, consistent with previous studies (Rubin et al., 2008), Neuroticism was associated with the retrieval of more negative events. Individuals low in Conscientiousness and high in Neuroticism tended to distance more from their life-changing experiences at follow-up, perhaps finding it difficult to integrate their turning-point experience with other self-relevant information. Conscientiousness and Neuroticism were also related to changes in affect of early childhood memories. These two personality traits have been linked to the construction of positive vs. negative life stories (see Thomsen, Hammershøj Olesen, Schnieber, & Tønnesvang, 2014), and also seem to drive the subjective experience of life-story memories. The present data did not replicate previous findings for Openness (Rasmussen & Berntsen, 2010). There was only one significant correlation between Openness and phenomenology (i.e. more Sensory Details in the Early Childhood Memory), and it was unrelated to changes in phenomenology. Methodological differences—in the sample size and composition, and in the type of memory assessed (i.e., word-cued vs. life-story memories)—may account for the inconsistent results. Consistent with their tendency to experience positive emotions (Watson & Clark, 1992), individuals high in Extraversion reported more positive affect and less negative affect in their Turning Point Memory. Surprisingly, Extraversion was unrelated to Sharing. Autobiographical memories are often used to create intimacy with others through the sharing of important experiences (Alea & Bluck, 2003; see McLean & Pasupathi, 2006). The present results suggest that individuals who are more introverted are just as likely to use this strategy as individuals who are more extraverted. Only a few phenomenological dimensions of the Turning Point Memory (i.e. Valence but not Coherence) showed significant correlations with psychological distress and subjective well-being. Psychological distress (i.e. depression and loneliness) was associated with the retrieval of more negative and less frequently shared turning points, whereas overall wellbeing was linked with retrieval of more positive events. Although low memory coherence is commonly associated with psy- 23 M. Luchetti et al. / Consciousness and Cognition 42 (2016) 15–25 Memory variables Time perspective Sensory details Visual perspective Emotional intensity TP TP TP TP EC EC EC Sharing EC TP EC Distancing Valence TP TP EC EC .09 .18 .06 .19 .17 .12 .17 .08 .09 .18 .12 .16 .08 .14 .14 .05 .14 .27* .02 .28** .05 .05 .02 .14 .06 .15 .15 .01 .09 .17 .16 .05 .18 .03 .09 .18 .03 .09 .14 .16 .26* .16 .12 .02 .36*** .13 .04 .04 .21* .14 .09 .02 .16 .12 .01 .03 .04 .03 .19 .11 .23* .11 .07 .21* .30** .13 .14 .06 .09 .02 .17 .14 .09 .15 .01 .08 .04 .02 .22* .07 .07 .19 .01 .00 .17 .08 .47*** .20* .18 .22* .02 .08 .07 .06 .33*** .34*** .15 .15 .07 .12 .04 .07 .07 .03 .09 .03 .02 .11 .07 .07 .02 .09 .00 .04 .31** .33*** .21* .20 .04 .04 .08 .00 .39*** .46*** .08 .18 chological distress (e.g., Sutin & Gillath, 2009), other studies have reported significant associations with posttraumatic stress for dimension such as emotional intensity but not coherence (e.g., Rubin, Boals, & Klein, 2010; Rubin, Dennis, & Beckham, 2011; Rubin et al., 2008). Moreover, in the present work, results only tentatively suggested a relation between change in psychological symptoms and memory ratings. However, the high test–retest correlations observed over the short interval for the state aspects of psychological functioning (i.e. current distress and well-being) limited the ability to predict change. There are some limitations in the present study that could be improved in future work. First, the re-test sample was relatively small. Second, the question of continuity was assessed only for two memories, but additional key events can be of particular interest to study (e.g. low point, high point, etc.). We did not constrain the valence or the domain, (e.g., work, love, etc.) of the memories. Positive vs. negative memories may maintain their qualities or fade differently over time (Walker & Skowronski, 2009). Moreover, the interrelation between memory, personality, and psychological outcomes may vary depending on how central the event is to participants’ life story and identity (e.g., Berntsen & Rubin, 2007). Third, the current sample was composed of participants aged 18–39 years. One possible explanation for the percentage of ‘different events’ reported at Time 2 may be related to participants’ young age. Young adults are especially open to variation and reinterpretation of their life story compared to older adults (McAdams & Olson, 2010). Stability coefficients of phenomenology may vary as well as a function of age. As such, other age groups should be included in the research. Lastly, we considered a short time interval. A longer follow-up (i.e. several months, years) will allow for more changes in memory and personality, and their reciprocal relation. Personality may shape memory content and the interpretation of past events, which in turn are aspects potentially implicated in personality development and change over time (Sutin, Costa, Wethington, & Eaton, 2010; Sutin & Robins, 2005). It will be also interesting to explore the association between phenomenology and the trajectories of both psychological distress and subjective well-being over long-term periods. Given that memory phenomenology is particularly important in the clinical domain, further insights should be obtained by examining memory continuity in clinical populations. Autobiographical memories constitute the daily material to work on during clinical interactions (Singer & Bonalume, 2010; Singer & Conway, 2011). Writing about a negative (potentially traumatic) event has been found to have beneficial effects (Pennebaker, 1997). Rubin et al. (2010) suggested that only the act of retrieving and rating (thinking about) a stressful memory leads to significant reductions in event-related distress; the repeated reliving of the event in a safe environment leads to a reduction of negative valence, emotional intensity and availability of the memory. Adler et al. (2015) observed that the way individuals construe personal narratives, and in particular narratives of life-challenging experiences, shapes the trajectory of mental health over years. 5. Conclusions There is a need for longitudinal studies that examine not only memory content and accuracy but also other important memory components, such as phenomenology. The present results suggest that the subjective experience of personal memories is relatively consistent over a short-term interval and shaped by both trait and state aspects of psychological functioning. It would be interesting to extend the present findings to examine the continuity (and changes) of phenomenology, its relation with personality and psychological symptoms, in both clinical and non-clinical samples, varying time-intervals between sessions. The way an individual construes life narratives and experiences meaningful memories, can be considered an expression of personality functioning that in part contributes to long-term adaptive/maladaptive outcomes (Singer, Blagov, Berry, & Oost, 2013). 24 M. Luchetti et al. / Consciousness and Cognition 42 (2016) 15–25 Acknowledgments There were no funding sources for this study. Part of the data reported here was presented in a doctoral dissertation completed by Dr. Luchetti at the University of Bologna, Department of Psychology. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.concog. 2016.02.011. References Adler, J. M., Turner, A. F., Brookshier, K. M., Monahan, C., Walder-Biesanz, I., Harmeling, L. H., ... Oltmanns, T. F. (2015). Variation in narrative identity is associated with trajectories of mental health over several years. Journal of Personality and Social Psychology, 108, 476–496. http://dx.doi.org/10.1037/ a0038601. Alea, N., & Bluck, S. (2003). Why are you telling me that? A conceptual model of the social function of autobiographical memory. Memory, 11, 165–178. http://dx.doi.org/10.1080/741938207. Bauer, P. J., Tasdemir-Ozdes, A., & Larkina, M. (2014). Adults’ reports of their earliest memories: Consistency in events, ages, and narrative characteristics over time. Consciousness and Cognition, 27, 76–88. http://dx.doi.org/10.1016/j.concog.2014.04.008. Berntsen, D., & Rubin, D. C. (2007). When a trauma becomes a key to identity: Enhanced integration of trauma memories predicts posttraumatic stress disorder symptoms. Applied Cognitive Psychology, 21, 417–431. http://dx.doi.org/10.1002/acp.1290. Boelen, P. A., Huntjens, R. J., & van den Hout, M. A. (2014). Concurrent and prospective associations of habitual overgeneral memory and prospection with symptoms of depression, general anxiety, obsessive compulsiveness, and post-traumatic stress. Memory, 22, 747–758. http://dx.doi.org/10.1080/ 09658211.2013.824985. Diener, E., Wirtz, D., Tov, W., Kim-Prieto, C., Choi, D., Oishi, S., & Biswas-Diener, R. (2009). New measures of well-being: Flourishing and positive and negative feelings. Social Indicators Research, 39, 247–266. http://dx.doi.org/10.1007/978-90-481-2354-4_12. Drivdahl, S. B., & Hyman, I. E. Jr., (2014). Fluidity in autobiographical memories: Relationship memories sampled on two occasions. Memory, 22, 1070–1081. http://dx.doi.org/10.1080/09658211.2013.866683. Fossati, A., Borroni, S., Marchione, D., & Maffei, C. (2011). The Big Five Inventory (BFI): Reliability and validity of its Italian translation in three independent nonclinical samples. Hirst, W., Phelps, E. A., Meksin, R., Vaidya, C. J., Johnson, M. K., Mitchell, K. J., ... Olsson, A. (2015). A ten-year follow-up of a study of memory for the attack of September 11, 2001: Flashbulb memories and memories for flashbulb events. Journal of Experimental Psychology: General, 144, 604–623. http://dx.doi. org/10.1037/xge0000055. Jack, F., & Hayne, H. (2010). Childhood amnesia: Empirical evidence for a two-stage phenomenon. Memory, 18, 831–844. http://dx.doi.org/10.1080/ 09658211.2010.510476. John, O. P., & Srivastava, S. (1999). The Big-Five trait taxonomy: History, measurement, and theoretical perspectives. In L. A. Pervin & O. P. John (Eds.). Handbook of personality: Theory and research (Vol. 2, pp. 102–138). New York: Guilford Press. Kotov, R., Gamez, W., Schmidt, F., & Watson, D. (2010). Linking ‘‘big” personality traits to anxiety, depressive, and substance use disorders: A meta-analysis. Psychological Bulletin, 136, 768–821. Kroenke, K., Spitzer, R. L., & Williams, J. B. W. (2001). The PHQ-9: Validity of a brief depression severity measure. Journal General Internal Medicine, 16, 606–613. http://dx.doi.org/10.1046/j.1525-1497.2001.016009606.x. Latorre, J. M., Ricarte, J. J., Serrano, J. P., Ros, L., Navarro, B., & Aguilar, M. J. (2013). Performance in autobiographical memory of older adults with depression symptoms. Applied Cognitive Psychology, 27, 167–172. http://dx.doi.org/10.1002/acp.2891. Leonard, R., & Burns, A. (2006). Turning points in the lives of midlife and older women: Five-year follow-up. Australian Psychologist, 41, 28–36. http://dx.doi. org/10.1080/00050060500391902. Luchetti, M., Montebarocci, O., Rossi, N., Cutti, A. G., & Sutin, A. R. (2014). Autobiographical memory and psychological distress in a sample of upper-limb amputees. PLoS ONE, 9, e99803. http://dx.doi.org/10.1371/journal.pone.0099803. Luchetti, M., & Sutin, A. R. (2015). Measuring the phenomenology of autobiographical memory: A short-form of the Memory Experiences Questionnaire. Memory. http://dx.doi.org/10.1080/09658211.2015.1031679. Luchetti, M., Terracciano, A., Stephan, Y., & Sutin, A. R. (2015). Personality and cognitive decline in older adults: Data from a longitudinal sample and metaanalysis. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences. http://dx.doi.org/10.1093/geronb/gbu184. Mazzotti, E., Fassone, G., Picardi, A., Sagoni, E., Ramieri, L., Lega, I., ... Pasquini, P. (2003). The Patient Health Questionnaire (PHQ) for the screening of psychiatric disorders: a validation study versus the Structured Clinical Interview for DSM-IV Axis I (SCID-I). Italian Journal of Psychopathology, 9, 235–242. McAdams, D. P. (2008). The life story interview. Retrieved from: <http://www.sesp.northwestern.edu/docs/LifeStoryInterview.pdf>. McAdams, D. P., Bauer, J. J., Sakaeda, A. R., Anyidoho, N. A., Machado, M. A., Magrino-Failla, K., ... Pals, J. L. (2006). Continuity and change in the life story: A longitudinal study of autobiographical memories in emerging adulthood. Journal of Personality, 74, 1371–1400. http://dx.doi.org/10.1111/j.14676494.2006.00412.x. McAdams, D. P., & Olson, B. D. (2010). Personality development: Continuity and change over the life course. Annual Review of Psychology, 61, 517–542. http:// dx.doi.org/10.1146/annurev.psych.093008.100507. McLean, K. C., & Pasupathi, M. (2006). Collaborative narration of the past and extraversion. Journal of Research in Personality, 40, 1219–1231. http://dx.doi. org/10.1016/j.jrp.2005.11.006. Pennebaker, J. W. (1997). Writing about emotional experiences as a therapeutic process. Psychological Science, 8, 162–166. http://dx.doi.org/10.1111/j.14679280.1997.tb00403.x. Rasmussen, A. S., & Berntsen, D. (2010). Personality traits and autobiographical memory: Openness is positively related to the experience and usage of recollections. Memory, 18, 774–786. http://dx.doi.org/10.1080/09658211.2010.514270. Rubin, D. C., Boals, A., & Berntsen, D. (2008). Memory in posttraumatic stress disorder: Properties of voluntary and involuntary, traumatic and nontraumatic autobiographical memories in people with and without PTSD symptoms. Journal of Experimental Psychology: General, 137, 591–614. http://dx. doi.org/10.1037/a0013165. Rubin, D. C., Boals, A., & Klein, K. (2010). Autobiographical memories for very negative events: The effects of thinking about and rating memories. Cognitive Therapy and Research, 34, 35–48. http://dx.doi.org/10.1007/s10608-008-9226-6. Rubin, D. C., Dennis, M. F., & Beckham, J. C. (2011). Autobiographical memory for stressful events: The role of autobiographical memory in posttraumatic stress disorder. Consciousness and Cognition, 20, 840–856. http://dx.doi.org/10.1016/j.concog.2011.03.015. Rubin, D. C., Schrauf, R. W., & Greenberg, D. L. (2004). Stability in autobiographical memories. Memory, 12, 715–721. http://dx.doi.org/10.1080/ 09658210344000512. M. Luchetti et al. / Consciousness and Cognition 42 (2016) 15–25 25 Rubin, D. C., & Siegler, I. C. (2004). Facets of personality and the phenomenology of autobiographical memory. Applied Cognitive Psychology, 18, 913–930. http://dx.doi.org/10.1002/acp.1038. Singer, J. A., Blagov, P., Berry, M., & Oost, K. M. (2013). Self-defining memories, scripts, and the life story: Narrative identity in personality and psychotherapy. Journal of Personality, 81, 569–582. http://dx.doi.org/10.1111/jopy.12005. Singer, J. A., & Bonalume, L. (2010). Autobiographical memory narratives in psychotherapy: A coding system applied to the case of Cynthia. Pragmatic Case Studies in Psychotherapy, 6, 134–188. Singer, J. A., & Conway, M. A. (2011). Reconsidering therapeutic action: Loewald, cognitive neuroscience and the integration of memory’s duality. The International Journal of Psychoanalysis, 92, 1183–1207. http://dx.doi.org/10.1111/j.1745-8315.2011.00415.x. Singer, J. A., & Salovey, P. (1993). The remembered self.New York: The Free Press. Smith, J., Fisher, G., Ryan, L., Clarke, P., House, J., & Weir, D. (2013). HRS Psychosocial and Lifestyle Questionnaire 2006–2010: Documentation Report. Retrieved from <http://hrsonline.isr.umich.edu/sitedocs/userg/HRS2006-2010SAQdoc.pdf>. Spitzer, R. L., Kroenke, K., & Williams, J. B. W., and the Patient Health Questionnaire Study Group. (1999). Validity and utility of a self-report version of PRIME-MD: the PHQ Primary Care Study. The Journal of the American Medical Association, 282, 1737–1744. http://dx.doi.org/10.1001/jama.282.18.1737. Sumner, J. A. (2012). The mechanisms underlying overgeneral autobiographicalmemory: An evaluative review of evidence for the CaR-FA-X model. Clinical Psychology Review, 32, 34–48. http://dx.doi.org/10.1016/j.cpr.2011.10.003. Sumner, J. A., Mineka, S., & McAdams, D. P. (2013). Specificity in autobiographical memory narratives correlates with performance on the Autobiographical Memory Test and prospectively predicts depressive symptoms. Memory, 21, 646–656. http://dx.doi.org/10.1080/09658211.2012.746372. Sumner, J. A., Mineka, S., Zinbarg, R. E., Craske, M. G., Vrshek-Schallhorn, S., & Epstein, A. (2014). Examining the long-term stability of overgeneral autobiographical memory. Memory, 22, 163–170. http://dx.doi.org/10.1080/09658211.2013.774021. Sutin, A. R. (2008). Autobiographical memory as a dynamic process: Autobiographical memory mediates basic tendencies and characteristic adaptations. Journal of Research in Personality, 42, 1060–1066. http://dx.doi.org/10.1016/j.jrp.2007.10.002. Sutin, A. R., Costa, P. T., Jr., Wethington, E., & Eaton, W. (2010). Turning points and lessons learned: Stressful life events and personality trait development across middle adulthood. Psychology and Aging, 25, 524–533. http://dx.doi.org/10.1037/a0018751. Sutin, A. R., & Gillath, O. (2009). Autobiographical memory phenomenology and content mediate attachment style and psychological distress. Journal of Counseling Psychology, 56, 351–364. http://dx.doi.org/10.1037/a0014917. Sutin, A. R., & Robins, R. W. (2005). Continuity and correlates of emotions and motives in self-defining memories. Journal of Personality, 73, 793–824. http:// dx.doi.org/10.1111/j.1467-6494.2005.00329.x. Sutin, A. R., & Robins, R. W. (2007). The phenomenology of autobiographical memory: The Memory Experiences Questionnaire. Memory, 15, 390–411. http:// dx.doi.org/10.1080/09658210701256654. Sutin, A. R., & Robins, R. W. (2010). Correlates and phenomenology of first and third person memories. Memory, 18, 625–637. http://dx.doi.org/10.1080/ 09658211.2010.497765. Thomsen, D. K., Hammershøj Olesen, M., Schnieber, A., Jensen, T., & Tønnesvang, J. (2012). What characterizes life story memories? A diary study of Freshmen’s first term. Consciousness and Cognition, 21, 366–382. http://dx.doi.org/10.1016/j.concog.2012.01.006. Thomsen, D. K., Hammershøj Olesen, M. H., Schnieber, A., & Tønnesvang, J. (2014). The emotional content of life stories: Positivity bias and relation to personality. Cognition and Emotion, 28, 260–277. http://dx.doi.org/10.1080/02699931.2013.815155. Thomsen, D. K., Jensen, T., Holm, T., Hammershøj Olesen, M. H., Schnieber, A., & Tønnesvang, J. (2015). A 3.5 year diary study: Remembering and life story importance are predicted by different event characteristics. Consciousness and Cognition, 36, 180–195. http://dx.doi.org/10.1016/j.concog.2015.06.011. Thorne, A., Cutting, L., & Skaw, D. (1998). Young adults’ relationship memories and the life story: Examples or essential landmarks? Narrative Inquiry, 8, 237–268. http://dx.doi.org/10.1075/ni.8.2.02tho. Tulving, E. (2002). Episodic memory: From mind to brain. Annual Review of Psychology, 53, 1–25. http://dx.doi.org/10.1146/annurev. psych.53.100901.135114. Walker, W. R., & Skowronski, J. J. (2009). The Fading Affect Bias: But what the hell is it for? Applied Cognitive Psychology, 23, 1122–1136. http://dx.doi.org/ 10.1002/acp.1614. Waters, T. E., & Fivush, R. (2014). Relations between narrative coherence, identity, and psychological well-being in emerging adulthood. Journal of Personality. http://dx.doi.org/10.1111/jopy.12120. Advance online publication. Watson, D., & Clark, L. A. (1992). On traits and temperament: General and specific factors of emotional experience and their relation to the five-factor model. Journal of Personality, 60, 441–476. http://dx.doi.org/10.1111/j.1467-6494.1992.tb00980.x. Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54, 1063–1070. http://dx.doi.org/10.1037/0022-3514.54.6.1063. Williams, J. M. G., Barnhofer, T., Crane, C., Herman, D., Raes, F., Watkins, E., & Dalgleish, T. (2007). Autobiographical memory specificity and emotional disorder. Psychological Bulletin, 133, 122–148. http://dx.doi.org/10.1037/0033-2909.133.1.122. Wilson, R. S., Boyle, P. A., Yu, L., Segawa, E., Sytsma, J., & Bennett, D. A. (2015). Conscientiousness, dementia related pathology, and trajectories of cognitive aging. Psychology and Aging, 30(1), 74–82. http://dx.doi.org/10.1037/pag0000013.

Consciousness and Cognition 22 (2013) 388–401 Contents lists available at SciVerse ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Lay psychology of the hidden mental life: Attribution patterns of unconscious processes Ofri Maor, David Leiser ⇑ Department of Psychology, Ben-Gurion University of the Negev, PO Box 653, Beer Sheva 84105, Israel a r t i c l e i n f o Article history: Received 2 April 2012 Keywords: Unconscious Theory of mind Social representations Intentionality Lay theories Folkpsychology a b s t r a c t In spite of extensive research on theory of mind, lay theories about the unconscious have scarcely been investigated. Three questionnaire studies totaling 689 participants, examined to what extent they thought that a range of psychological processes could be unconscious. It was found that people are less willing to countenance unconscious processes in themselves than in others, regardless of the time period considered – present, past or future. This is especially true when specific experience-like situations are envisioned, as opposed to considering the question in abstract or generic terms. In addition, the notion of unconscious psychological processes is resisted for certain domains in particular: intending, sensing, believing, and thinking. We interpret this pattern by positing the existence of two conceptions about the unconscious that are differentially applied according to circumstances: one originating in prevalent social representations about the unconscious, the other based on self-model of the person as an intentional actor. Ó 2013 Elsevier Inc. All rights reserved. 1. Introduction People entertain theories about the mind. Those theories come naturally to people who develop normally, and people use them to understand others and themselves. Over the past decades, countless studies have been devoted to describing and analyzing theories of mind, its development, and its impairment in pathology (e.g. Dimaggio & Lysaker, 2010; Malle, 2004; Wellman, Cross, & Watson, 2001; Ziv, Leiser, & Levine, 2011). Theory of mind (ToM) is a complex, interconnected set of concepts relating to mental activities, such as beliefs and desires, used to explain a person’s behavior. It enables people to make sense of past and present behaviors and allows predictions about the future (Bartsch & Wellman, 1995). Strikingly, whereas many studies have explored theories of the mind in its conscious aspects (e.g. Wellman, 1990), everyday theories about the unconscious mind have been neglected. The concept of the unconscious has changed much since Freud’s time (1915/1957), but its importance has certainly not abated (for reviews from different perspectives see Bargh & Morsella, 2008; Ekstrom, 2004; Hassin, Uleman, & Bargh, 2005). While Uleman (2005) claimed that the psychoanalytic unconscious is to most laypeople the only unconscious, little is actually known about how it is conceived by the general public today. Lay beliefs do not necessarily resemble theoretical-academic models, whose influence is slow and indirect (Keil, 2010; Moscovici & Hewstone, 1983). Theoretical-academic models about the unconscious could therefore not serve as a neutral starting point to our exploration. Further, while countless studies have also investigated people’s awareness of their own mental functioning (e.g. Baumeister, Masicampo, & Vohs, 2011), these studies cannot enlighten us regarding naïve theories of unconscious processes. The present study is an initial attempt to fill this lacuna. ⇑ Corresponding author. Fax: +972 86428348. E-mail addresses: ofri.maor@gmail.com (O. Maor), dleiser@exchange.bgu.ac.il (D. Leiser). 1053-8100/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.concog.2012.12.007 O. Maor, D. Leiser / Consciousness and Cognition 22 (2013) 388–401 389 We set out to investigate how the general public uses the concept of unconscious. Our first descriptive task was to map what aspects of mental life can be unconscious. To do this, we explored the psychological components (thought, emotions, memory, etc.) that may be unconscious, according to folkpsychology. We also delved deeper, and asked several questions about the unconscious that have been investigated in other studies about the way people understand the human mind: Does the unconscious dimension primarily apply to mental processes or to contents? Do naïve psychologists think more readily about unconscious mental life as abstract theoretical statements, or in terms of everyday concrete examples? Are unconscious processes equally attributed to everyone or is there a disparity between self and others? These interrelated questions helped us explicate how unconscious processes figure in contemporary folk psychology in the Western world. Before we can review studies regarding psychological components and other important questions mentioned above, it will be helpful to briefly survey several relevant literatures about unconscious mental life and public perceptions. 1.1. Naïve psychological conceptions about the unconscious In daily life, we often find it difficult to understand our acts, feelings or thoughts. Events that should cause happiness may trigger sadness or irritation instead; there are memories we cannot retrieve and that later pop up, without apparent explanation. Since commonsense and folk psychology is influenced by the scientific realm (e.g. Kelley, 1992; Moscovici & Hewstone, 1983), we may expect people to know about unconscious processes and contents and to consider them as appropriate, or even compelling, explanations for certain everyday occurrences. On the other hand, Wegner (2002) who studied consciousness and free will claimed that people understand minds around the concept of an agent. This idea, as he develops it, center around the ideal of conscious agency, which involves both intention and conscious will. The experience of consciously intending an action we then perform isn’t a direct or reliable indication that our conscious thought caused the action (e.g. Gazzaniga, 2011; Mele, 2009; Wegner, 2002). There can be illusions of conscious will. According to Wegner, people hold onto the illusion of conscious will because knowing what you do and why you do it, particularly in pursuit of a goal, is highly valued. This suggests that people refrain from attributing their behavior and their mental life to an unconscious source. We should therefore also ask specifically to what extent there is room, within naïve psychology, for unconscious processes and contents, and how widely they are applied. Few studies have examined people’s stated perceptions and understandings of the unconscious mental world. Flavell, Green, Flavell, and Lin (1999) studied the development of children’s knowledge of the unconscious, but their definition of unconsciousness referred to a state of deep dreamless sleep. Two other studies examined daily conception of the unconscious as part of Freudian and psychoanalytic theory. Moscovici (1961/2008) studied the social representation of psychoanalysis in France. He found a common conscious-unconscious model that was based on psychoanalytic conceptions, though rather simplified. In this model Moscovici found that the unconscious was a part of the human psyche, but also an autonomous agency, and a force that was in conflict with another force, consciousness. Moscovici argued that, for the public, the unconscious functions as a symbol of an unwanted force that can impose itself on our autonomous or independent personalities, interfering with the lives and free will of individuals. A second, more recent study was conducted by Schomerus, Matschinger, and Angermeyer (2008), and assessed the extent to which the public has incorporated Freudian theory in its understanding of mental illness. It asked narrowly about the meaning of unconscious conflicts. They found that while the term ‘‘unconscious’’ is used, people had difficulties to explain it and their understandings about unconscious had little in common with the psychoanalytic one (a finding typical for lay theories and conceptions, see Leiser, 2001). There are also some empirical studies about naïve understanding of concepts associated with unconscious mental life, such as dreams (e.g. Morewedge & Norton, 2009) and defense mechanisms (e.g. Cramer & Brilliant, 2001). However, these studies are too narrow to stand for the whole of unconscious mental activity and cannot serve as a starting point to our study. To summarize, past research about the general public’s understanding of unconscious mental life relied on a specific theoretical perspective (usually Freudian), or focused on limited mental activities that are associated with the unconscious. We will now turn to general models of naïve understanding about the human mind, as a different potential source of information. 1.2. Models of the mind and unconscious mental life Three main models of healthy adult conceptions of mental life stand out: the D’Andrade (1987, 1995) folk model of the mind; Wellman (1990) theory of mind; and Malle (2004) folk theory of mind. Each model deals with naïve conceptions about unconscious mental life differently. Relying on qualitative methodologies, D’Andrade (1987, 1995) described a folk model of the mind, used in Western cultures. This model includes five major parts: perceptions (senses), thoughts (e.g. beliefs, reason), feelings/ emotions (e.g. love, lust, feeling tired), wishes (e.g. desire, hope, need), and intentions (e.g. decide, plan). D’Andrade (1987) claimed that in the folk model of the mind there is an important differentiation between mental contents and mental processes, to which we return below. Comparing his folk model with psychoanalytic and academic models of the mind, D’Andrade (1987) concluded that according to the folk model people act primarily in light of their conscious feelings and thoughts. Some mental processes and contents may be unconscious, but these are considered atypical, would be treated as a problem, that can be resolved by turning one’s full attention to the situation. 390 O. Maor, D. Leiser / Consciousness and Cognition 22 (2013) 388–401 Another, very influential, way of regarding the ability to reason about the human mental world was labeled the Theory of Mind (ToM) (Gopnik & Wellman, 1994; Wellman, 1990). ToM refers to the tendency to understand people in terms of their traits and mental states (Premack & Woodruff, 1978), and emphasizes the importance of the person’s beliefs, desires, goals (e.g. Wellman, 1990) and emotions (Fonagy, Gergely, Jurist, & Target, 2002). Influenced by D’Andrade’s (1987) pioneering work, Wellman (1990) presented a comprehensive model of mind as entertained by adults, with a theoretical sketch that did not emerge from empirical research. Wellman depicted belief-desire reasoning schemes comprised of familiar basic elements: Perception (e.g. seeing, hearing), sensation (e.g. dizziness, pain), physiological emotions (e.g. hunger) and basic emotions (e.g. love, fear), cognitive emotions (e.g. boredom), thinking (e.g. dreaming, learning) and belief, desire and intentions (e.g. decide, plan), actions and reactions; in a more elaborated scheme, the effect of traits was added. Complex connections obtain between those basic elements. Wellman (1990) did allow for unconscious explanations: ‘‘. . . naive adults understand that people at times act from such underground, nonconscious, irrational processes, processes hard to fit within the cognitive rational machinery of belief-desire psychology’’ (p. 312). However, those remarks regarding the unconscious mind were neither theoretically elaborated nor empirically tested. And last, Malle (2004) originated an influential, empirically-based model of folk theory of mind and explanations that focuses on the complex concept of intentionality. According to his model, behaviors that relate to mental states break down into two fundamentally different types: intentional action and unintentional action. Awareness of an action is only one condition for considering it intentional, and consciousness is not a main concern for the model. Although the models about naïve understanding of the mind presented here did not study unconscious perception in any detail, they can serve as a starting point to our study. More specifically, Wellman (1990) and D’Andrade’s (1987, 1995) elaboration of mental components or domains allows us to explore whether, for our informants, those domains can also exist in an unconscious modality. We next present theoretical and empirical background regarding two important questions we raised above, the abstractconcrete and self-other dimensions. 1.3. Abstraction, naïve knowledge and unconscious mental life Lay knowledge can take the form of general abstract statements, or of specific concrete instances (Keil, 2006; Leiser, 2001). Since no autonomous spontaneous organizing process exists to coordinate knowledge originating from different sources (Leiser, 2001), it is important to investigate both forms of knowledge. For instance, there could be a difference between thinking that people can have feelings they are not aware of, and applying that notion to a specific case where John’s unexpected behavior on his brother’s wedding would be considered as an expression of unconscious jealousy. If knowledge about unconscious mental life is mostly acquired from practical experience, we may expect people to attribute unconscious contents and processes more readily when asked about specific cases with concrete features than when asked abstractly framed questions. The opposite would be true if they acquire that knowledge primarily via cultural osmosis (including lexical assimilation), and transmission of scientific notions to commonsense discourse through media, art, etc. (Moscovici, 1961/2008; Moscovici & Hewstone, 1983; Sperber, 1990; Wagner, 2007). 1.4. Self-Other differences regarding perceptions of mental life Reviewing theoretical debates, clinical evidence and neurocognitive empirical data, Dimaggio, Lysaker, Carcione, Nicolò, and Semerari (2008) concluded that knowing oneself and knowing others are largely independent competences. This conclusion is in line with data from social psychology research that found differences between the way people perceive, understand and judge self (actor) and other (observer). Two lines of evidence support this self-other distinction: research concerning self-knowledge, and studies about unwanted influences on judgments and evaluations. If self-knowledge is, at least to some extent, the opposite of lack of awareness, then self-knowledge studies are relevant. Pronin, Kruger, Savitsky, and Ross (2001) found that people judged their own degree of accuracy in judging others to be greater than that of their peers. Importantly to us, they also found that people think their own self-knowledge is greater than that of their peers. That is, people are not only convinced they know their peers better than their peers know them; they also feel they know their peers better than their peers know themselves. As we mentioned above, naïve psychological theory uses the unconscious as a symbol for limiting, unacceptable, interfering forces (Moscovici, 1961/2008), or regards it as an atypical, problematic state that should be resolved (D’Andrade, 1987). Therefore, research about unwanted mental influences could shed light on the way unconscious processes and contents are perceived. Wilson and Brekke (1994) who reviewed and re-conceptualized research dealing with unwanted influences on judgments and evaluations defined mental contamination as a process whereby a person has an unwanted response because of mental processing that is unconscious or uncontrollable. They suggested that people underestimate their own susceptibility to mental contamination, even as they expect that others would be more biased than they would. This tendency to see biases in others, while being blind to it in oneself has been demonstrated across a range of motivational and cognitive biases (for a review, see Pronin, Gilovich, & Ross, 2004). This could be due to self-enhancement motives, since people are more likely to evaluate the self in more favorable terms than they evaluate other people (e.g. Brown, 1986). Alternatively, this gap between self and other could be explain by the introspection illusion: a hypothesized phenomenon whereby people overvalue introspection information (e.g. thoughts, feelings, and other conscious mental contents), relative to other O. Maor, D. Leiser / Consciousness and Cognition 22 (2013) 388–401 391 available information (e.g. behavior, naïve theories, and population base rates), when assessing their own actions, motives, and preferences, but not when assessing others (Pronin, 2006; Pronin et al., 2004). Interestingly, this phenomenon was found to be time-dependent, as people often refer to their past and future selves as though they were other people (e.g. Pronin, Olivola, & Kennedy, 2008; Pronin & Ross, 2006). In conclusion, to the extent that unconscious processes and contents are seen as unwanted features or biases (D’Andrade, 1987; Moscovici, 1961/2008), they will be attributed to other people more than to oneself, so that people would believe that unconscious processes and contents are more likely to be found in other people minds than in their owns. The present study aims to map the psychological components that may be unconscious and their features, according to folkpsychology. We do not ask to what extent people are in fact aware and accurate about their own and others mental life in particular cases, but focus on their beliefs about these matters. To this end we examined specific properties of everyday perceptions about the unconscious and asked three questions: (1) Do naïve psychologists think about unconscious mental life as abstract statements or in terms of everyday concrete examples? (2) Is the unconscious mind attributed more to others than to self and would this be affected by the time-frame? (3) Which aspects or domains of the human mind (thought, emotions, memory, etc.) are considered as possibly unconscious? The remainder of this paper is organized as follows: first, we present a brief overview of the survey materials and subjects. This is followed by the report on three studies that address the first two questions presented above. Next, we look again at the data from all three studies and, contrasting the ten ToM domains we tested, examine to which aspects of the human mind unconsciousness is more readily attributed. Lastly, the general discussion advances a framework of daily dual-theory model about the mind that accounts for the pattern of our findings. 2. Overview of materials and subjects Our three studies all use questionnaires we developed on the basis of the theory of mind model presented by Wellman (1990) and the work of D’Andrade’s (1987, 1995). Wellman’s ToM model details a wide range of components of the naïve declarative understanding of the human mind. His scheme for belief-desire reasoning (p. 109; 115, Wellman, 1990), served as a handy starting point for detailing domains in our questionnaires, allowing us to explore a potential unconscious dimension of components of the ToM model. The domains about which we asked whether they could be unconscious were: thinking, believing, desiring, intending, sensing, perceiving, trait, behaving, feeling and knowing. We made several adjustments to best serve our purposes: (1) Wellman (1990) distinguished between cognitive emotions, physiological emotions and basic emotions. We unified these terms: Cognitive emotions and basic emotions were included in feeling, and we combined his physiological emotions (such as hunger and thirst) with sensing. (2) A separate domain of knowing was added, in view of its importance to the naïve theories research. (3) Believing was separated from thinking. It was important to formulate the questions broadly enough, in a way that could apply to all the relevant domains. Previous research used different terms to ask about the unconscious: for instance, D’Andrade used ‘‘aware’’ and ‘‘know’’ alternately. Absent useable data about folk models of the unconscious and in order to keep our questions’ language close to everyday spoken Hebrew, we used the verb to know (‘‘lada’at’’) in our questionnaire (the usual Hebrew term for the un/conscious shares the same root). For example: ‘‘I have beliefs that I don’t know about’’. Further, each questionnaire began with the following orienting statement: ‘‘Some people believe in the existence of the unconscious, and think there are complex processes and contents in our mind’’, in order to prompt them to interpret ‘‘not knowing’’ in the intended sense, of not consciously knowing. As discussed above, the connection between unconscious and naïve models of the mind raises several important issues. The first of these is the abstraction level of knowledge about unconscious mental life. We modified the questionnaire to include questions at different levels of abstraction. Abstract questions were general statements about each domain (e.g. ‘‘I feel something and I don’t know that I feel it’’), whereas concrete questions were statements about specific situations (e.g. ‘‘I am angry at my best friend and I don’t know I’m angry at him’’). In order to investigate this variable, we used Wellman (1990) model, where ToM’s domains are introduced as a collection of specific cases gathered under a more abstract definition. For instance, according to Wellman (1990, p. 109) the domain of thinking includes to reason, learn, imagine, and remember; that of perception includes to see, hear, taste, and smell. This issue is initially explored by Study 1, and in more detail in Study 2. The second variable, object, concerns the role of the target (self vs. other people) in the attribution of unconscious processes or contents. Each of the three studies included different versions of the questionnaires, designed to address this issue in various ways. Finally, we examined the distinction between unconscious mental states and mental processes (e.g. D’Andrade, 1987; Uhlmann, Pizarro, & Bloom, 2008). Do people make a distinction between unconscious ToM processes (e.g. thinking, believing) and the mind as a container of unconscious mental contents (e.g. thoughts, beliefs), and if so, which one predominates? We labeled this variable dynamism, and explored it in Study 1. Since our research was aimed at exploring perceptions of adult Israeli naïve psychologists about the unconscious, we excluded from our samples participants younger than eighteen years, and those who had formally studied psychology – whether in high school, university or any other formal frameworks. In addition, in order to ensure that our participants 392 O. Maor, D. Leiser / Consciousness and Cognition 22 (2013) 388–401 were sufficiently fluent in Hebrew and familiar with Israeli culture, we excluded those who lived in Israel less than ten years. 3. Study 1 In this first study, we sought to test three general hypotheses about perception of the unconscious. First, we examined whether unconsciousness is attributed equally to self and to others. We also investigated whether people are more disposed to ascribe unconscious contents or unconscious processes. D’Andrade (1987) claimed that in the folk model of the mind, the mind may be regarded as a collection of internal states (a container), as opposed to a set of internal processes (a processor limited to small concurrent actions). He found that the distinction between processes and states influences how people represent the mind and its operation. The difference between processes and contents was also found relevant for empirical research about unconscious social cognition (Gawronski, Hofmann, & Wilbur, 2006; Uhlmann et al., 2008). Finally, we made a preliminary examination of the level of abstraction effect, to test the need for a more thorough one. 3.1. Method 3.1.1. Participants and design Two hundred and sixteen adults (108 men and 108 women) who had not studied psychology volunteered to complete our survey. They were randomly assigned to a 2  2  2 mixed factorial design in which Object (self-vs.-other) and Dynamism (contents-vs.-processes) served as between-subjects factors and Abstraction level (abstract-vs.-concrete) as a within-subjects factor. Participants were 18–83 years old (M = 30.21, SD = 11.91), 47% had academic education and the remainder high-school education. 3.1.2. Procedure and materials A female experimenter approached passengers as they waited for a train or in the course of a train journey, and asked them to complete a psychology survey. All adult passengers present were invited to participate. The questionnaire started with a short general description of the research as ‘‘related to understanding the human mind’’, followed by a consent form. Following an introductory statement about the unconscious (presented above), participants were asked to judge the possibility of an unconscious process (or content) using a scale from 1 (impossible) to 6 (always possible). They did so for different psychological domains and with reference to their own/others’ mind. Participants were randomly assigned to one out of four conditions, as determined by the version of the questionnaire they received (see Table 1): (1) Self-Contents: participants were asked to rate the possibility of unconscious ToM’s contents, within their own mind (n = 52). (2) Self-Processes: participants were asked about processes, within their own mind (n = 56). (3) Other-Contents: who were asked about unconscious contents, within other people’s mind (n = 56). (4) Other-Processes: who answered about unconscious processes, within other people’s mind (n = 52). The questions were all couched in the present tense, and covered seven domains: thinking, believing, desiring, intending, behaving, feeling and knowing. The questionnaires had two parts, abstract and concrete: The first, abstract part asked seven questions, one about each domain; the second consisted of questions that present examples of the abstract psychological domains. The concrete level of the questionnaires regarding other people was preceded by the precision: ‘‘Assume that the people described are healthy of body and mind’’. We asked a single concrete question for each domain, with one exception: we asked two concrete questions about the feeling domain (jealousy and fear). As will be discussed later, this was done to check the possibility that different likelihood levels are attributed to different specific unconscious cases. The reasoning was that, should we find differences between cases of the same general domain, it would demonstrate the importance of the specific example used, implying that another concrete example would affect the response pattern. Table 1 gives an example of questions about the intending domain, with the concrete example of planning taken from Wellman (1990). The original (Hebrew) questions were matched in length for each abstract/concrete question in the different questionnaire versions: self vs. other and processes vs. contents. The questionnaire concluded with demographic questions, the last question related to formal training in psychology. Table 1 Questions about intending domain by condition in Study 1. Is it possible that. . . Self-content Self-process Other-content Other-process Abstract I have intentions that I don’t know about I have a plan to leave my neighborhood and I don’t know this. I intend something and I don’t know I intend it. I plan to leave my neighborhood and I don’t know this. A person has intentions that he doesn’t know about Tomer has a plan to leave his neighborhood and he doesn’t know this. A person intends something and he doesn’t know he intends it. Tomer plans to leave his neighborhood and he doesn’t know this. Concrete O. Maor, D. Leiser / Consciousness and Cognition 22 (2013) 388–401 393 3.2. Results A mixed design analysis of variance (GLM) with 2 (Object: self-vs.-other)  2 (Dynamism: processes-vs.-contents) between variables  2 (Abstraction level: abstract-vs.-concrete) within variable, showed two main effects: Abstraction level and Object. The findings support our hypothesis about a self-other asymmetry: Participants attributed a higher probability of unconscious processes and contents to other people (M = 3.95, SD = 0.09) than to themselves (M = 3.11, SD = 0.09), F(1, 212) = 48.32, p < 0.0001, g2p ¼ 0:19. Regarding the effect of abstractness, participants believed in the existence of unconscious contents and processes when the ToM domains were presented abstractly (M = 3.82, SD = 1.02), more than when they considered concrete examples (M = 3.26, SD = 1.11), F(1, 212) = 108.4, p < 0.0001, g2p ¼ 0:34. A weak trend to accept the existence of unconscious contents (M = 3.64, SD = 0.08) more than of unconscious processes (M = 3.43, SD = 0.09) failed to reach significance, F(1, 212) = 2.9, p = 0.09, g2p ¼ 0:01 (although this pattern repeated with all seven domains presented). No other significant effect was found. As mentioned in the materials presentation, we included questions about two different feelings, and we analyzed the differences between them as a planned comparison, in a repeated measures analysis of variance for the feeling domain (abstract feelings, fear, and jealousy). This showed the same significant trend of accepting the unconscious when presented abstractly (M = 4.02, SD = 1.47), rather than as a specific emotion (M = 3.38, Sd = 1.37), F(1, 215) = 42.03, p < 0.0001, g2p ¼ 0:16. More important is the significant difference between fear (M = 2.91, SD = 1.59) and jealousy (M = 3.85, SD = 1.59), F(1, 215) = 74.55, p < 0.0001, g2p ¼ 0:26, and the proximity of jealousy to the abstract question about feelings. These findings suggest variability in the approach, in keeping with to the nature of each concrete case, hence the importance to consider more than one representative of each domain. 3.3. Discussion Study 1 yields a differentiated picture of lay perceptions about unconscious. The findings provide initial evidence supporting our self-other difference hypothesis. People believed that unconscious contents and processes are more likely to be found in other people’s minds than in their own, though the cause of this tendency remains unclear: is it due to self-enhancement motivations? To the introspection illusion (Pronin et al., 2004)? Or perhaps to some other potential influence? Study 3 will speak to these questions. Study 1 also provides an indication about the way explicit judgments about unconscious mental life are produced. Participants found abstract level statements about unconscious ToM to be more likely than concrete cases. Several reasons associated with learning processes about the unconscious and with the ways the unconscious is seen by the general public could explain this abstract-concrete difference. For instance, one important conclusion could be that commonsense psychology is formed through internalization of the knowledge and theories of experts, rather than on the basis of personal action and experience (Moscovici & Hewstone, 1983). However, we first need to consider a simpler explanation for the abstraction effect, related to the design of our study. Abstract questions cover a range of examples, of which some are more readily than others susceptible of an explanation in terms of unconscious processes or contents. When participants were asked whether unconscious contents or processes could be attributed to certain domain abstractly they might have thought spontaneously about concrete examples and used them to answer the abstract question. Thus, one specific case contained in the domain could tip the balance in favor of unconscious attributions. For example, if someone were asked about unconscious feelings, and in answering this thought about some feelings that she believed to be mostly conscious (such as sadness or happiness), and also about one emotion that is often unconscious (e.g. jealousy), she might conclude by declaring that feelings are likely to be unconscious. Indeed, we did find a significant difference between the two feelings we asked about in the feeling domain, indicating that the use of a single example to represent each of the other six domains may not do justice to potentially rich and fine distinctions people may make. It is also conceivable that our questions did not include the highest scored examples the respondent had in mind, and this would be enough to generate the effect abstract-concrete observed. Study 2 was designed to correct this crucial flaw. A last lesson from Study 1 is that we cannot ignore the difference between contents and process. While the trend to accept unconscious contents more than processes did not reach significance, it is possible that this difference is not negligible since it repeated within the different ToM domains. We will not investigate this distinction here, but the methodological implication is clear: we must avoid mixing the two formulations as though they are equivalent. A real difference may exist between the two ways of conceiving the unconscious, even if it did not express itself strongly in probability attribution. Accordingly, we always used unconscious mental processes for wording of the questions in the coming studies. 4. Study 2 Study 2 was designed to further our understanding about the relation of specificity and abstraction to the attribution of unconscious processes. Study 1 suggested that people accept the existence of the unconscious more when it in an abstract manner, but we also saw that this finding may be dependent on the specific question we asked in each domain. For that reason, in Study 2 participants were asked several concrete questions about each of the ToM domains. For two of the domains (thinking and feeling) the questionnaire contained a large number of concrete examples, to allow us to examine the 394 O. Maor, D. Leiser / Consciousness and Cognition 22 (2013) 388–401 abstraction level effect in depth, as we will see below. We added another abstraction level through a single very general question about the existence of the unconscious. For this study, we relied on a different population of participants – internet users. The use of an online survey allowed for a longer questionnaire than did the train passengers condition in Study 1. We also extended the number of domains from seven to ten, improving the fit with Wellman’s model (Wellman, 1990): thinking, believing, desiring, intending, sensing, perceiving, trait, behaving, feeling and knowing. 4.1. Method 4.1.1. Participants and design One hundred and sixty participants (109 women and 51 men), without formal training in psychology, volunteered to complete our survey on the internet. They were randomly assigned to a 2  3 mixed factorial design in which Object (self-vs.-other) served as the between-subjects factor and Abstraction level (general-abstract-concrete) served as a within-subjects factor. Participants were 18–65 years old (M = 31.65, SD = 10.76). Most of our sample had university education (80%), and the rest had high school education. 4.1.2. Procedure and materials We distributed a research questionnaire labeled ‘‘Understanding mental processes’’, through private email addresses and social networking, and also published an invitation to participate on various online forums. The questionnaire was administered through a commercial platform (surveymonkey.com). Participants were randomly assigned to one of the two research conditions: (1) Self: questions about the possibility of unconscious processes within their own mind (n = 81). (2) Other: questions about the possibility of unconscious processes within other people’s minds (n = 79). This design enabled us to study possible self-other asymmetric influences on abstraction. The questionnaires of Study 2 were the same as in Study 1, apart from a few changes concerning the additional elaboration on abstraction levels, the introduction of three additional ToM domains, and some adaptations for the online version as explained below. The same orienting statement (‘‘Some people believe in the existence of the unconscious. . . etc.’’) served as an introduction for the questionnaire, and provided the framework for the study. Participants were then asked to determine whether they deem the existence of an unconscious process in their own/other’s mind possible on a 6 point scale. Questions were in the present tense. After participants were directed to think about the unconscious concept, both Self and Other questionnaires started with eleven abstract questions: one question for every ToM domain as mentioned, and a general abstract question about the possibility of unconscious (general-self: ‘‘There are things I don’t know about myself’’; general-other: ‘‘There are things people don’t know about themselves’’). Then followed 33 questions pertaining to concrete examples about the unconscious for the different ToM domains: knowing (2 questions. For example: ‘‘I know how to write a text message using my phone with my eyes closed and I don’t know I know this’’), feeling (8), thinking (5), believing (2), desiring (3), intending (3), sensing (3), perceiving (2), trait (2) and behaving (3). The examples used for each domain were adapted from Wellman (1990, p. 108)1 whenever available and, when not available, novel ones were prepared. The survey software presented the questions in a different random order for each participant. 4.2. Results The data were analyzed in a mixed design analysis of variance (see Fig. 1). As in Study 1, participants attribute unconscious processes to others (M = 4.4, SD = 0.11) more than to self (M = 3.55, SD = 0.1), F(1, 158) = 32.4, p < 0.0001, g2p ¼ 0:17. Planned comparisons showed this self-other difference for all three levels of abstraction used in Study 2 [General: other (M = 5.35, SD = 0.97), self (M = 4.62, SD = 1.44), F(1, 158) = 14.35, p < 0.001, g2p ¼ 0:08; Abstract: other (M = 4.2, SD = 0.96), self (M = 3.4, SD = 1.25), F(1, 158) = 20.51, p < 0.0001, g2p ¼ 0:11; Concrete: other (M = 3.65, SD = 0.92), self (M = 2.63, SD = 0.99), F(1, 158) = 45.17, p < 0.0001, g2p ¼ 0:22]. Participants were less willing to acknowledge unconscious processes when presented in the abstract that than when presented in general, and even less willing to do so when referring to concrete cases (General: M = 4.98, SD = 1.28; Abstract: M = 3.79, SD = 1.18; Concrete: M = 3.13, SD = 1.08). This was shown by a set of planned comparisons that showed both a significant difference between general and the mean of abstract and concrete, F(1, 158) = 335.29, p < 0.0001, g2p ¼ 0:68, and between abstract and concrete, F(1, 158) = 133.49, p < 0.0001, g2p ¼ 0:46. The interaction between the two variables – Abstraction level and Object – was not significant, F(2, 316) = 1.73, p = 0.18, g2p ¼ 0:01. Additional analyses are presented in Table 2. We ran planned comparisons contrasting abstract and concrete for each of the ten domains. In all cases, except for thinking, the abstract question about the domain scored significantly higher than the mean of concrete instances in the domain. 1 We examined the emotions domain in more detail through multiple concrete questions. Concrete questions concerned the six basic emotions: love, surprise, happiness, anger, sadness and fear (Shaver, Schwartz, Kirson, & O’Connor, 2001), as well as two of Wellman’s cognitive feelings: jealousy and boredom. 395 O. Maor, D. Leiser / Consciousness and Cognition 22 (2013) 388–401 Fig. 1. Attribution of unconscious to self and others by three abstraction levels: general, abstract and concrete, in Study 2. Table 2 Comparison mean (SD) for abstract and concrete, and planned comparison analysis (effect size) for their difference, within the ten ToM domains in Study 2. a ** Behaving Perceiving Trait Knowing Desiring Feeling Intending Sensing Believing Thinking Abstract: Mean (SD) 4.33 (0.12) 4.23 (0.12) 4.22 (0.12) 4.22 (0.12) 3.93 (0.12) 3.82 (0.13) 3.56 (0.13) 3.29 (0.13) 3.18 (0.12) 3.14 (0.13) Concrete: Mean (SD) 3.78 (0.11) 3.42 (0.11) 3.65 (0.12) 3.2 (0.11) 2.85 (0.1) 3.12 (0.1) 2.48 (0.12) 2.78 (0.09) 2.76 (0.11) 3.29 (0.09) F(1, 159) ðg2p Þa 27.16** (0.15) 49.08** (0.24) 34.75** (0.18) 70.25** (0.31) 98.22** (0.38) 48.08** (0.23) 87.04** (0.36) 8.13** (0.1) 13.01** (0.08) 1.82 (0.01) Planned comparisons for abstract-concrete differences. p < 0.01. Table 3 Mean (SD) for abstract level and highest concrete example. Behaving Perceiving Trait Knowing Desiring Feeling Intending Sensing Believing Thinking Mean (SD) abstract 4.33 (0.12) 4.23 (0.12) 4.22 (0.12) 4.22 (0.12) 3.93 (0.12) 3.82 (0.13) 3.56 (0.13) 3.29 (0.13) 3.18 (0.12) 3.14 (0.13) Mean (SD) highest example 4.16 (0.14) 3.5 (0.14) 3.86 (0.14) 3.23 (0.14) 3.6 (0.13) 3.42 (0.14) 2.81 (0.12) 3.39 (0.13) 2.98 (0.14) 4.67 (0.12) To further our understanding about the abstract level effect, we carried out post hoc comparisons (Scheffé) to test the difference between the highest concrete example and the abstract level for each domain (see Table 3). In no domain (except for thinking) did any of the concrete instances score significantly higher than the abstract question about the domain (and for the domains perceiving, trait, knowing, and intending, the abstract question scored significantly lower, p’s < 0.01). To illustrate, in the Feeling domain: even the concrete emotion with the highest value (envy: M = 3.42, SD = 0.14) did not score significantly higher than the abstract question concerning the feeling domain (M = 3.82, SD = 0.13). It is interesting to note that this pattern where the lower abstraction level cases didn’t score as high as the more abstract question repeated itself with the general – abstract domains difference, F(10, 1590) = 44.73, p < 0.001, g2p ¼ 0:22. Here, post hoc comparisons (Scheffé) reveled that even the highest scored abstract domain (behaving, M = 4.33, SD = 0.1) scored significantly lower than the general question about the possibility of unconscious processes (M = 4.98, SD = 0.12), p < 0.01. 396 O. Maor, D. Leiser / Consciousness and Cognition 22 (2013) 388–401 4.3. Discussion Study 2 replicated findings from Study 1 regarding the effect of abstraction level on perception of the unconscious. Findings with the more elaborate questionnaire demonstrate an abstraction effect regarding the possibility of unconscious processes. People tend to accept the possibility that unconscious processes occur more readily when those processes are presented abstractly. Our analysis of the data shows that this effect is not due to rational averaging processes. The consistent pattern whereby more abstract statements scored higher than any concrete example imply that this effect cannot be explained by supposing that participants thought about several extreme examples that sway their answer when answering the abstract questions (especially since the abstract questions were always presented before the concrete example). Only one domain, thinking, included concrete examples that scored higher than the abstract question. This domain was also the only one with no significant abstraction level gap, and therefore may be consider as atypical, for reasons unknown. Moreover, as with Study 1, Study 2 demonstrated Self-Other differences regarding the unconscious. As noted, the questions in Study 1 and 2 asked about the present. Therefore, another potential explanation for the abstraction effect stem from the very definition of unconscious processes – namely that one is not aware of them as they occur. Study 3 was designed to address this question as well as to examine other potential explanations to the Self-Other differences. 5. Study 3 Study 1 and Study 2 supported our initial hypothesis that people believe unconscious processes are more likely to exist in other people’s minds than in their own. Study 3 was designed to shed light on potential explanations for this self-other asymmetry, using past theoretical understanding about self over time and different types of others. People often perceive and treat their past and future selves as though they are other people (e.g. Pronin & Ross, 2006; Pronin et al., 2008). Pronin and her colleagues linked this tendency to people’s inattention to the internal states of their past and future selves, and their lack of focus or sensitivity to their mental life, as experienced in the past, or likely to experience in the future. Further, people tend to see themselves as improving over time, or evaluate their past selves in ways that make them feel good about their present selves (Wilson & Ross, 2001). To address these issues the self condition was split into three conditions: self-past, self-present and self-future. We argue that insofar as people perceive unconscious processes as reflecting unfavorably on them, they should treat their past and future selves as other people (Pronin & Ross, 2006; Pronin et al., 2008) and therefore find fewer unconscious processes in their present selves than in their past and future selves. The same prediction could also be the effect of the defining quality of unconscious processes – namely, that people are not aware of them as they occur. Though we might expect that past and future selves would be treated as other persons, it is important to ask: which kind of others? Close others are perceived in ways more similar to self than more distant others (Prentice, 1990), especially when access to and appreciation of the internal mental world of those close others is available (Liberman, Trope, & Stephan, 2007). To address these questions about different objects, we split the other condition into close friend and average person. Given past findings, we assume that a person’s close friend is seen as more similar to self than is an average person both from motivational (Prentice, 1990) and cognitive considerations (e.g., see the review in Liberman et al., 2007). 5.1. Method 5.1.1. Participants and design Three hundred thirteen students (182 female and 131 male), who had no formal training in psychology, volunteered to complete our survey on the internet or while they waited for their class to begin at the University. (No significant betweensamples differences were found and the responses from 112 internet users and 201 campus students were combined in our analysis.) Participants were randomly assigned to one of five time-subject combinations conditions (see below). Mean age was 24 (SD = 4). 5.1.2. Procedure and materials In its pen and paper version, the experimenters asked students to participate in a psychology survey. In the electronic version, the survey was published on online forums for students, similar to Study 2. We designed the study to investigate the two dimensions of interest, time and personal distance in a within subjects design. This allowed us to compare three levels for time, regarding the self: present, future and past, and also to compare three levels of social distance (in the present): self, close friend and the average person. Since self/present is common to both comparisons, we had five conditions in total. Having all these conditions within the confines of a single design enabled us to compare the two dimensions of interest, as seemed desirable in view of the literature we surveyed above. We used a temporal distance of five years. The span of five years was chosen since our sample includes many students and we wanted them to think about their adult or almost adult years, rather than their childhood; and importantly Pronin and Ross (2006) did find temporal effects when they used such time spans. After a short introduction as in the previous studies (consent form and a general statement about the unconscious), participants were instructed to answer questions about a single subject at one specific time, according to one of the five O. Maor, D. Leiser / Consciousness and Cognition 22 (2013) 388–401 397 conditions: (1) Self-Present: Questions about the possibility of unconscious processes in their own mind, as it is at the present (n = 62). (2) Self-Past: Questions about the possibility of unconscious processes in their own mind, as they were five years ago (n = 66). (3) Self-Future: Questions about the possibility of unconscious processes in their own mind, five years from now (n = 60). (4) Average person-Present: Questions about the possibility of unconscious processes in the mind of the average person, in the present (n = 64). (5) Close friend-Present – Questions about the possibility of unconscious processes in the mind of a close friend, in the present (n = 61). For example, the ‘Self-Past’ questionnaire directed the participants as follows: ‘‘In answering, please consider the way you think and feel about yourself, as you were five years ago’’ (highlighted in the original). Participants were asked about unconscious processes within the ten domains: thinking, believing, desiring, intending, sensing, perceiving, trait, behaving, feeling and knowing. The questions were all formulated at the abstract level of Study 2. The questions in the five conditions were identical, except for the differences defining the conditions (e.g. Self-Past: ‘‘I felt something and I didn’t know I felt it’’; Self-Future: ‘‘I will feel something and I won’t know I feel it’’; Close friend-Present: ‘‘My friend feels something and doesn’t know he feels it’’; and so forth). The answer scale was the same as in Study 1 and Study 2, and the demographic questions were the same too. 5.2. Results A mixed design analysis (GLM) of 5 (object-time conditions)  10 (ToM domains) revealed an overall significant main effect of object-time conditions, F(4, 308) = 13.84, p < 0.0001, g2p ¼ 0:15. However, as illustrated in Fig. 2, the specific pattern of results was different from that expected, mainly regarding the temporal-self conditions. The planned comparison established that there was no significant difference between the probability of unconscious processes in one’s mind in the present (M = 3.21, Sd = 0.13), and in the future and the past (combined) (M = 3.27, Sd = 0.12), F(1, 308) = 0.17, ns, g2p < 0:001. These unexpected results are not due to the effect of different perspective about the unconscious within different ToM domains, since there was no interaction between the object-time condition and the ToM domains, F(36, 2772) = 0.13, ns, g2p ¼ 0:02. Therefore the object-time conditions were tested as post hoc comparisons (Scheffé test), that revealed that the difference between average person and close friend is statistically significant (p = 0.008). The average person is also different from all the conditions involving self: past, present and future (all p’s < 0.0001). No other difference was significant. A less conservative post hoc comparison with Fisher LSD test found close friend also to be different from the three conditions involving self (all p’s < 0.05), but still no temporal effect regarding the self. As expected, then, respondents judged their close friend to resemble them more than did the average person, with respect to the incidence of unconscious processes. Yet surprisingly, there are no significant differences between past and future selves and the present self, while past and future selves are different from other people (close friend or average person). 5.3. Discussion These unanticipated results of Study 3 can be illuminating regarding the theoretical explanations that we considered. While some theories, presented above, can explain why unconscious processes are attributed to the average person more I represent confidence level at 95%. * Significant difference by Scheffé test. {*}Significant difference by Fisher LSD test. Fig. 2. Probability of unconscious processes for the five time  object conditions in Study 3. 398 O. Maor, D. Leiser / Consciousness and Cognition 22 (2013) 388–401 than to a close friend or to oneself, the uniqueness of our topic will become apparent when we consider the total lack of temporal differences regarding self and the significant difference between the average person condition and past and future selves conditions. Let us consider theoretical avenues to the interpretation our findings. Self-enhancement motivations can explain why close friends were judged more like the self than did the average person. If people feel that their friends represent something about their identity (Brown, 1986), and if unconscious influences are seen as unwanted interference, then it is clear why they would be more willing to attribute unconscious processes to unspecific others than to their close friends. This tendency can also explain the weak difference found between self and close friend. However, the lack of difference regarding self over time cannot be as easily explained by such motivational explanations, especially since people tend to see themselves as improving over time (Wilson & Ross, 2001). In general, one should be wary of interpreting null results, but note that in Study 3, the absence of temporal differences coexists with the difference as a function of social distance within the same experimental design. Another possible interpretation of the social distance effect is better familiarity with our friend’s mind than with that of other people (Prentice, 1990). It is harder to explain the lack of a marked difference between self and close friend in terms of familiarity: the person with whose mental processes one is most familiar is presumably oneself, and one might be expected to recognize one’s own mental processes accurately, including the occurrence of unconscious ones. However, this weak difference could be due to the basic feature of unconscious processes, namely that, by definition, one is not aware of them as they take place. But this logical limitation is removed as time passes, and one should be able to recognize at least some of one’s own past unconscious mental processes more readily than in the present. Since no temporal differences were found, those explanations cannot account for the whole picture. Likewise, the introspection illusion (Pronin, 2006; Pronin et al., 2004) cannot offer a satisfactory explanation for our findings. The introspection illusion model explains the self-others differences by a tendency to overvalue introspection information relative to other available information, when assessing self, but not when assessing others (for review: Pronin et al., 2004). However, Pronin and her colleagues also found that people tend to perceive and treat their past and future selves as though they are other people (e.g. Pronin & Ross, 2006; Pronin et al., 2008). Using methods similar to theirs, (time differences of five years and today) our results display a different pattern in which past and future selves are not considered as others. Construal Level Theory (Liberman & Trope, 1998; Liberman et al., 2007) faces a similar difficulty. What then makes attribution of unconscious mental life to the person himself so unique? What could explain the lack of time differences? Before offering our view, we still need to examine our last question: in which domains were unconscious processes considered possible, and for which domains was the notion resisted. 6. Unconscious ToM’s domains In this section, we compare the ten unconscious ToM domains that were studied: knowing, feeling, thinking, believing, desiring, intending, sensing, perceiving, traits and behaving, by reviewing the pattern of our findings across the three different studies. 6.1. Data In order to pool the data from the three studies, we used the abstract questions concerning unconscious mental processes that occurred in all of them. Specifically, from Study 1, we used the seven abstract questions in conditions Self-Processes (n = 52) and Other-Processes (n = 52). From Study 2, the ten abstract questions in both conditions: Self (n = 79) and Other (n = 81). And from Study 3 we used data from participants who answered questions about either Self-Present (n = 62) or about Average person-Present (n = 64). This adds up to 390 respondents for the domains: knowing, feeling, thinking, believing, desiring, intending and behaving; and to 286 respondents for the remaining domains: traits, sensing, perceiving. 6.2. Results and discussion Before we can use the pooled answers from our three studies for our purposes, we need to test whether there are differences between the studies. A mixed design analysis of variance checked for differences between the 3 Studies  7 Domains included in all three (knowing, feeling, thinking, believing, desiring, intending and behaving). No interaction was found between domains and studies, F(12, 2322) = 0.68, ns, g2p ¼ 0:003; nor was there a main effect of study, F(2, 387) = 0.03, ns, g2 < 0.001. A second mixed design analysis that examined the 3 Domains not included in Study 1: 3 Domains (traits, sensing, perceiving)  2 Studies (Studies 2 and 3) also found neither interaction between Domains and Studies, F(2, 568) = 1.45, p = 0.23, g2p ¼ 0:005; nor a main effect of Study, F(1, 284) = 1.34, p = 0.25, g2p ¼ 0:004. Accordingly, we pooled all the participants from the different studies. We submitted the domains to a hierarchical clustering algorithm (see Fig. 3). Cluster analysis involves two parameters: the definition of distance between items (here, the domains), and the clustering method. We used the most common definition of distance, namely, the geometric distance between them in the multidimensional space spanned by the participants. Clustering was performed according to ‘‘Ward’s method’’ (Statsoft Inc., 2010). This method is a kind of reverse analysis of O. Maor, D. Leiser / Consciousness and Cognition 22 (2013) 388–401 399 Fig. 3. Clustering of the Unconscious ToM domains, Ward’s clustering method, City-block (Manhattan) distances, with data from Studies 1 to 3. variance, and attempts to minimize the sum of squares of any two clusters that can be formed at each step. The meaning of clustering, with this method, is readily intelligible. This procedure leads to a tree that in our case clearly groups the indices in two distinct clusters: (1) behaving (M = 4.39, SD = 1.47), trait (M = 4.23, SD = 1.43), knowing (M = 4.25, SD = 1.45), perceiving (M = 4.09, SD = 1.58), desiring (M = 3.94, SD = 1.55), and feeling (M = 3.83, SD = 1.59). (2) Intending (M = 3.53, SD = 1.61), sensing (M = 3.21, SD = 1.58), believing (M = 3.2, SD = 1.55), and thinking (M = 3.16, SD = 1.59). In sum, three different samples showed a consistent pattern of results regarding the possibility that some mental domains could be unconscious while others are less likely. Behaving, traits, knowing, perceiving, desiring and feelings were all considered as more likely to include unconscious cases than intending, sensing, believing and thinking. 7. General discussion Let us begin by summarizing our findings once more. People are less willing to countenance unconscious processes in themselves than in others, regardless of the time period considered – present, past or future. This is especially true when specific experience-like situations are envisioned, as opposed to considering the question in abstract or generic terms. In addition, the notion of unconscious psychological processes is resisted for certain domains in particular: intending, sensing, believing, and thinking. In presenting the findings, we made reference to some important theories, which could not explain the complete picture revealed. Below, we present our interpretation of this novel set of findings. However, before turning to it, we will briefly review those theories and their lack of correspondence with our data. 7.1. A look at existing theories The abstraction level effect is consistent with Moscovici’s view that knowledge about unconscious mental life is acquired primarily through transmission of scientific notions to commonsense discourse, whether through media, art, and so on (Moscovici, 1961/2008; Moscovici & Hewstone, 1983) or through language and conceptual learning (Sperber, 1990). However, this account does not fit the pattern of differential attribution of unconscious mental life as a function of social distance and time. The introspection illusion is another social-cognitive explanatory approach that fails to account for the entire picture. According to the introspection illusion perspective, people overvalue introspection information relative to other available information when assessing their own actions, motives and preferences, but not when assessing that of others (Pronin, 2006; Pronin et al., 2004). By their nature, unconscious mental processes and contents are outside our awareness and not available for introspection, at least in the present. When people are asked whether they have some unconscious thought at the present, even if such a thought exists, it is not available as introspective information. Accordingly, the explanation runs, people attribute unconscious mental activity to others more than to themselves. The findings from Study 3 were inconsistent with the introspection illusion explanation. The lack of temporal differences regarding unconscious processes, given the significant difference between an average person and past and future selves, does not match the introspection illusion literature, that finds that past and future selves are perceived in a way equivalent to other people (e.g. Pronin & Ross, 2006; Pronin et al., 2008). Moreover, the introspection illusion is presumably operative even when people have detailed access to others’ introspections (Pronin & Kugler, 2007), as would be the case with close friends. This is at variance with our finding that close friends were seen as almost as fully conscious of their mental lives as the self. 400 O. Maor, D. Leiser / Consciousness and Cognition 22 (2013) 388–401 Self-enhancement motivation may also be considered as a factor, since it seems that people perceive unconscious processes as undesirable, and people tend to evaluate themselves in more favorable terms than they do other people (e.g. Brown, 1986). As unconscious mental processes and contents were attributed to others more than to oneself, this may indicate that the unconscious mind is conceived negatively (as was indeed found by D’Andrade, 1987; Moscovici, 1961/2008). However, whereas findings concerning self-enhancement show that people tend to see themselves as improving over time (Wilson & Ross, 2001), we found no evidence of temporal differences regarding one’s own unconscious mental processes. Evidently, people prefer to distance themselves from unconscious mental activities, at any given time period, over presenting themselves as improving over time. Does this discrepancy imply that self-enhancement motives are irrelevant? We don’t think so, but these motives cannot tell the whole story. 7.2. The dual-theories model We suggest our findings can best be explained by postulating the existence of different conceptual models about the mind. In thinking about the unconscious, contemporary laypeople have two competing conceptions or theories that originate in two different sources. Freudian terminology has penetrated everyday language, and casual mention of unconscious motivations or behavior is commonplace, even if the theoretical concepts referenced are misunderstood (Moscovici, 1961/2008; Schomerus et al., 2008). Further, the mass media frequently echo findings about unconscious influence on decision making. It would require a very deliberate and principled attitude for any contemporary to counteract the effect of this influence. Moreover, the prevalence of such terminology and the related social representations occurred for good reason. The explanatory power of unconscious motivations, or unconsciously performed activities, is indisputable. People occasionally observe others being jealous or angry, for example, yet unaware of their own jealousy or in denial of their anger. It is therefore not surprising that respondents to a large extent endorse this social representation. Alongside this explanation stands a much older and more fundamental model, the (implicit) self-model of the actor. In order to function, people have a model of themselves as sensible intentional agents (Mele, 2009). This self-model promotes coherent goal-directed activity and is based on concepts of knowledge, desire and belief (Wellman, 1990); deliberation and intention (Malle, 2004); and conscious will (Gazzaniga, 2011; Wegner, 2002). It is buttressed by conscious experience (all the mental processes you ever encounter and control are conscious), but its logical status is that of a core cognition (Carey, 2009). This model is as natural as it is compelling when we consider our own actions, and it provides the explanation of choice for our own behavior, present, past and future. Crucially, this conceptual framework does not include an unconscious dimension. D’Andrade (1987, 1995) remarked that the folk model of the mind can include unconscious properties, but that these are considered as marginal oddities, caused by inessential and quickly corrected conditions such as not paying properly attention. When considering the question ‘‘Can X be unconscious?’’ one or the other of these two models will be preponderant. If the question is phrased in the abstract, about generic mental activities concerning people in general, the social representation of the unconscious is deployed. But when it refers to the self or close associates, and especially if the question relates to the core features of an ideal, sensible, goal-directed actor (sensation, intention, belief and thinking), then that representation is avoided. Interesting, the last three are also the ones for whom Baumeister et al. (2011) found an actual role of consciousness. The two models, internal versus external conceptions, can coexist since no spontaneous organizing process exists to coordinate different sources of knowledge regarding any particular phenomenon or expertise (Leiser, 2001). Some significant limitations of our study should be acknowledged. First, the interpretation of our findings would be strengthened by experiments examining whether the self-other and concrete-abstract differences are specific to mental contents. Further, we relied on closed questionnaires, and it is imperative to use complementary methods. These would include more open ones, such as eliciting associations, and more directive ones, asking about theoretically based topics such as dreams, conflicts, intuitions, and automatic cognitive procedures. The present study is but a first step in the study of lay understanding of the unconscious mind and the role of the unconscious in folkpsychology. Future work should take into account individual differences relevant to the questions examined and could examine potential influences of theoretical understandings about naïve conceptions of the unconscious. Individual differences in the ability of people to understand mental complexity may relate to the way they conceptualize mental life. Automatic and unconscious abilities in ToM and mentalization (such as sense of agency and reflective functions) influence and constrain how people understand the social world and participate in it. It is likely that these differences, both within the normal range and in pathology, also express themselves when people are confronted to questions such as those in our study. By the same token, personality differences related to the discernment of social reality and mental processes (such as selfesteem and locus of control) may profoundly influence the answers to our questionnaires, which in turn may prove valuable in analyzing them. Our research shows that people may accept the unconscious human mind abstractly for certain mental domains, but that when it comes to their own daily life, they tend to deny its influence. These findings, if confirmed and developed, have significance for a wide range of philosophical, theoretical and practical areas of psychology, in particular, the study of theory of mind, meta-cognition, moral responsibility, self-awareness, and psychodynamic psychotherapy. O. Maor, D. Leiser / Consciousness and Cognition 22 (2013) 388–401 401 Acknowledgments We are grateful to Yoav Bar-Anan, Kathleen Vohs and John Bargh for helpful discussions, to Tali Petel and Ben Peled for research assistance, and to Anne Dubitzky for her editorial help. References Bargh, J. A., & Morsella, E. (2008). The unconscious mind. Perspectives on Psychological Science, 3, 73–79. Bartsch, K., & Wellman, H. M. (1995). Children talk about the mind. Oxford, England: Oxford University Press. Baumeister, R. F., Masicampo, E. J., & Vohs, K. D. (2011). Do conscious thoughts cause behavior? Annual Review of Psychology, 62, 331–361. Brown, J. D. (1986). Evaluations of self and others: Self-enhancement biases in social judgments. Social Cognition, 4, 353–376. Carey, S. (2009). The origin of concepts. New York: Oxford University Press. Cramer, P., & Brilliant, M. (2001). Defense use and defense understanding in children. Journal of Personality, 69, 291–321. D’Andrade, R. (1995). The development of cognitive anthropology. Cambridge, England: Cambridge University Press. D’Andrade, R. (1987). A folk model of the mind. In D. Holland & N. Quinn (Eds.), Cultural models in language and thought (pp. 112–148). Cambridge, England: Cambridge University Press. Dimaggio, G., & Lysaker, P. H. (2010). Metacognition and severe mental disorders: From research to treatment. London: Routledge. Dimaggio, G., Lysaker, P. H., Carcione, A., Nicolò, G., & Semerari, A. (2008). Know yourself and you shall know the other. . . to a certain extent: Multiple paths of influence of self-reflection on mindreading. Consciousness and Cognition, 17, 778–789. Ekstrom, S. R. (2004). The mind beyond our immediate awareness: Freudian, Jungian and cognitive models of the unconscious. Journal of Analytical Psychology, 49, 657–682. Flavell, J. H., Green, F. L., Flavell, E. R., & Lin, N. T. (1999). Development of children’s knowledge about unconsciousness. Child Development, 70, 369–412. Fonagy, P., Gergely, G., Jurist, E. L., & Target, M. (2002). Affect regulation, mentalization, and the development of the self. New York: Other Press. Freud, S. (1915/1957). The unconscious. In J. Strachey (Ed.). The standard edition of the complete psychological works of Sigmund Freud (Vol. 14, pp. 159–215). London: Springer Hogarth Press. Gawronski, B., Hofmann, W., & Wilbur, C. J. (2006). Are ‘‘implicit’’ attitudes unconscious? Consciousness and Cognition, 15, 485–499. Gazzaniga, M. S. (2011). Who’s in charge? Free will and the science of the brain. New-York: HarperCollins Publishers. Gopnik, A., & Wellman, H. M. (1994). The theory theory. In L. A. Hirschfield & S. A. Gelman (Eds.), Mapping the mind: Domain specificity in cognition and culture (pp. 257–293). Cambridge, England: Cambridge University Press. Hassin, R., Uleman, J. S., & Bargh, J. A. (Eds.). (2005). The new unconscious. Oxford: Oxford University Press. Keil, F. C. (2006). Explanation and understanding. Annual Review of Psychology, 57, 227–254. Keil, F. C. (2010). The feasibility of folk science. Cognitive Science, 34, 826–862. Kelley, H. H. (1992). Common-sense psychology and scientific psychology. Annual Review of Psychology, 43, 1–23. Leiser, D. (2001). Scattered naïve theories: Why the human mind is isomorphic to internet web. New Ideas in Psychology, 19, 175–202. Liberman, N., & Trope, Y. (1998). The role of feasibility and desirability considerations in near and distant future decisions: A test of temporal construal theory. Journal of Personality and Social Psychology, 75, 5–18. Liberman, N., Trope, Y., & Stephan, E. (2007). Psychological distance. In A. W. Kruglansky & E. T. Higgins (Eds.), Social psychology: Handbook of basic principles (pp. 353–381). New-York: Guilford. Malle, B. F. (2004). How the mind explains behavior: Folk explanations, meaning and social interaction. Cambridge Mass.: MIT Press. Mele, A. R. (2009). Effective intentions: The power of conscious will. New-York: Oxford University Press. Morewedge, C. K., & Norton, M. I. (2009). When dreaming is believing: The (motivated) interpretation of dreams. Journal of Personality and Social Psychology, 96, 249–264. Moscovici, S. (1961/2008). Psychoanalysis: Its image and its public (Trans. D. Macey). Cambridge UK: Polity Press. Moscovici, S., & Hewstone, M. (1983). Social representations and social explanations: From the ‘naive’ to the ‘amateur’ scientist. In M. Hewstone (Ed.), Attribution theory: Social and functional extensions (pp. 98–125). Oxford: Blackwell. Premack, D., & Woodruff, G. (1978). Does the chimpanzee have a theory of mind? Behavioral and Brain Sciences, 1, 515–526. Prentice, D. A. (1990). Familiarity and difference in self- and other-representations. Journal of Personality and Social Psychology, 59, 369–383. Pronin, E. (2006). Perception and misperception of bias in human judgment. Trends in Cognitive Sciences, 11, 37–43. Pronin, E., Gilovich, T., & Ross, L. (2004). Objectivity in the eye of beholder: Divergent perceptions of bias in self versus other. Psychological Review, 111, 781–799. Pronin, E., Kruger, J., Savitsky, K., & Ross, L. (2001). You don’t know me, but I know you: The illusion of asymmetric insight. Journal of Personality and Social Psychology, 81, 639–656. Pronin, E., & Kugler, M. B. (2007). Valuing thoughts, ignoring behavior: The introspection illusion as a source of the bias blind spot. Journal of Experimental Social Psychology, 43, 565–578. Pronin, E., Olivola, C. Y., & Kennedy, K. A. (2008). Doing unto future selves as you would do unto others: Psychological distance and decision making. Personality and Social Psychology Bulletin, 34, 224–236. Pronin, E., & Ross, L. (2006). Temporal differences in trait self-ascription: When self is seen as an other. Journal of Personality and Social Psychology, 90, 197–209. Schomerus, G., Matschinger, H., & Angermeyer, M. C. (2008). Traces of Freud: The unconscious conflict as a cause of mental disorders in the eyes of the general public. Psychopathology, 41, 173–178. Shaver, P., Schwartz, J., Kirson, D., & O’Connor, C. (2001). Emotion knowledge: Further exploration of a prototype approach. In W. G. Parrott (Ed.), Emotions in social psychology (pp. 26–56). Philadelelphia: Psychology Press. Sperber, D. (1990). The epidemiology of beliefs. In C. Fraser & G. Gaskell (Eds.), The social psychological study of widespread beliefs (pp. 25–43). Oxford: Oxford University Press. Statsoft Inc. (2010). Statistica for Windows (Computer program manual). Tulsa: Statsoft. Uhlmann, E. L., Pizarro, D. A., & Bloom, P. (2008). Varieties of social cognition. Journal for the Theory of Social Behavior, 38, 293–322. Uleman, J. S. (2005). Introduction: Becoming aware of the new unconscious. In R. R. Hassin, J. S. Uleman, & J. A. Bargh (Eds.), The new unconscious (pp. 3–15). Oxford: Oxford University Press. Wagner, W. (2007). Vernacular science knowledge: Its role in everyday life communication. Public Understanding of Science, 16, 7–22. Wegner, D. M. (2002). The illusion of conscious will. Cambridge, MA: MIT Press. Wellman, H. M. (1990). The child’s theory of mind. Cambridge, MA: MIT Press. Wellman, H. M., Cross, D., & Watson, J. (2001). Meta-analysis of theory of mind development: The truth about false belief. Child Development, 72, 655–684. Wilson, T. D., & Brekke, N. (1994). Mental contamination and mental correction: Unwanted influences on judgments and evaluations. Psychological Bulletin, 116, 117–142. Wilson, A. E., & Ross, M. (2001). From chump to champ: People’s appraisals of their earlier and present selves. Journal of Personality and Social Psychology, 80, 572–584. Ziv, I., Leiser, D., & Levine, J. (2011). Social cognition in schizophrenia: Cognitive and affective factors. Cognitive Neuropsychiatry, 16, 71–91.

Consciousness and Cognition 43 (2016) 143–151 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Review article Real, rubber or virtual: The vision of ‘‘one’s own” body as a means for pain modulation. A narrative review Matteo Martini School of Psychology, University of East London, Water Lane, London E15 4LZ, UK a r t i c l e i n f o Article history: Received 5 February 2016 Revised 24 May 2016 Accepted 3 June 2016 Keywords: Body ownership Body representation Pain Virtual reality Rubber hand illusion Visual analgesia Virtual hand illusion a b s t r a c t In the last few years a branch of pain research has been focussing on the modulatory effects of the vision of the body on pain perception. So, for instance, the vision of one’s own real body has been proven to induce analgesic effects. On the other hand, bodily illusions such as the rubber hand illusion have provided new tools for the study of perceptual processes during altered body ownership states. Recently, new paradigms of body ownership made use of a technology that is going places both in clinical and in experimental settings, i.e. virtual reality. While the vision of one’s own real body has been proven to yield compelling analgesic effects, slightly more controversial are those attributed to the vision of ‘‘owned” dummy bodies. This review will discuss the studies that examined the effects on pain perception of the vision of the own body, with or without body ownership illusions. Ó 2016 Elsevier Inc. All rights reserved. Contents 1. 2. 3. 4. 5. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The vision of one’s own body in pain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rubber hand illusion and pain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Virtual hand illusion and pain. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 144 145 146 147 149 149 1. Introduction With more than six hundred thousand articles published on the topic, pain is certainly one of the best studied areas in the realm of medicine and neuroscience. Yet, due to its intrinsic subjective nature and the plethora of factors that may modulate it, much is left to be discovered. It is widely accepted that pain is both a sensorial and emotional conscious experience and it has a protective function, steering attention toward a potential threat, thus facilitating the avoidance of dangerous situations. However pain can also occur in the absence of actual tissue damage and ‘‘nociception and pain should not be confused, E-mail address: M.Martini@uel.ac.uk http://dx.doi.org/10.1016/j.concog.2016.06.005 1053-8100/Ó 2016 Elsevier Inc. All rights reserved. 144 M. Martini / Consciousness and Cognition 43 (2016) 143–151 because each can occur without the other” (Loeser & Treede, 2008). At the neural level, a set of brain areas have been identified as being specifically recruited to encode the pain experience. These areas form the so-called ‘‘pain matrix” and are commonly the primary and secondary somatosensory cortices (SI and SII), the insula, the anterior cingulate cortex (ACC), prefrontal cortex (PFC), thalamus, basal ganglia, and cerebellum (Schweinhardt & Bushnell, 2010). The same brain areas have been shown to be active even during the observation of pain in others, in empathy for pain studies (Botvinick et al., 2005; Lamm, Batson, & Decety, 2007; Saarela et al., 2007). Nevertheless, the existence of a specific set of brain areas selectively dedicated to encode the pain experience is still a matter for debate. The ‘‘pain matrix” has been lately challenged in favour of a neural network that mainly responds to salient stimuli, i.e. stimuli capable of engaging one’s attention and motivational status, requiring the subject to make a prompt decision (Apkarian, Hashmi, & Baliki, 2011). So, if on one hand evidence exist in favour of neural structures selectively responding to nociception and pain (Vierck, Whitsel, Favorov, Brown, & Tommerdahl, 2013; Wager et al., 2013), on the other hand studies suggest that the activation of the putative ‘‘pain matrix” is prompted by all salient stimuli, regardless of the sensory channel involved (Mouraux, Diukova, Lee, Wise, & Iannetti, 2011; Mouraux & Iannetti, 2009). In such theoretical framework this salience-detection system would have a protective function detecting and reacting to possible threats, not merely painful, to ensure the physical integrity of the body (Legrain et al., 2012). Further reading on the theoretical frameworks about pain perception can be found in the historical overview of the main pain theories written by Moayedi and Davis (2013). The great interest in pain research can be easily explained by the fact that pain significantly represents a social and economic burden, as well as by the negative impact that it has on the sufferer’s life. Indeed, not only is pain an unpleasant sensory and emotional experience but, in some cases, it can deeply affect person’s life leading to suicidal ideation and behaviour (Campbell et al., 2015). Many studies have therefore attempted to find ways to manage pain states via pharmacological or non-pharmacological interventions. Of particular interest are the non-pharmacological interventions as they bypass the significant adverse side effects reported by conventional drug use (Carter et al., 2014). Belonging to this category are a series of studies that focussed their attention on the analgesic effects of cross-modal perception, for example pain and vision. In particular, from a seminal work with phantom limb pain published twenty years ago (Ramachandran & RogersRamachandran, 1996), a series of studies in the pain research have concentrated their attention on the role played by the vision of one’s own body in the modulation of pain. Bodily illusions like the rubber hand illusion (Botvinick & Cohen, 1998) and its virtual counterpart (Sanchez-Vives, Spanlang, Frisoli, Bergamasco, & Slater, 2010) have provided new avenues for investigating pain perception during body ownership paradigms. Exploiting the principle that, provided synchronous multisensory cues, one can feel a new fake body part as part of his/her own body, the feeling of body ownership can be extended to body parts that differ from the original. However, while the vision of one’s own real body part has been shown to be analgesic (Longo, Betti, Aglioti, & Haggard, 2009), there has been a recent debate on whether the analgesic effects of seeing one’s own body part holds true also during the vision of fake (rubber/virtual) ‘‘owned” body parts (Gilpin, Bellan, Gallace, & Moseley, 2014; Martini, Perez-Marcos, & Sanchez-Vives, 2015). The present work aims to review the research articles that so far have focussed on the effects on pain perception of the vision of one’s own body, either real, rubber or virtual. 2. The vision of one’s own body in pain Cross-modal interactions between the vision of the body and somatosensation have been extensively investigated (Macaluso & Maravita, 2010; Medina & Coslett, 2010; Serino & Haggard, 2010; Wesslein, Spence, & Frings, 2014). A seminal study by Tipper for instance, showed how the vision of one’s own body part influences tactile perception (Tipper et al., 1998). However, regarding pain perception, it was only 10 years after Tipper’s work that the possible effects of the vision of the body were investigated in healthy subjects. In 2008, Valeriani and co-workers found that when their participants observed clips of another’s hand receiving painful stimuli, while they concomitantly were getting painful laser stimulations on their hands, the early nociceptive-related neural processing was modulated, compared to the observation of the controlled stimuli (Valeriani et al., 2008). A step forward on this line was taken by Longo and collaborators in another laser-evoked potentials (LEPs) study (Longo, Betti, et al., 2009). In their work, these authors reported the first evidence that the vision of one’s own body part in pain is analgesic. In three different experiments they showed how while their participants were looking at their own painfully stimulated hand (but still keeping vision non-informative of the painful stimulation that was occurring), they felt less pain compared to when they were looking at a box or even at somebody else’s hand. The analgesic effect related to the vision of their own body part was also accompanied by a reduction of the late N2/P2 components of the LEPs (Longo, Betti, et al., 2009). The authors proposed that this effect was likely to be due to a visually-induced activation of the inhibitory GABAergic interneurons in the somatosensory areas. In support of this idea are the findings from a somatosensory evoked potentials investigation. Here Cardini and coworkers found that the vision of the hand, compared to the vision of a box, produced a suppression of the early somatosensory potential when two fingers were stimulated at the same time, thus revealing an augmented inhibitory interneuronal activity within the somatosensory cortex (Cardini, Longo, & Haggard, 2011). This result was later supported by another EEG study showing that the vision of the body, compared to the vision of a neutral object, increased noxious-related beta oscillatory activity bilaterally in the sensorimotor cortices, likely reflecting cortical inhibitory activity of nociceptive stimuli processing (Mancini, Longo, Canzoneri, Vallar, & Haggard, 2013). Furthermore, in a following neuroimaging study, it was found that the vision of the body part subjected to painful stimulations increased M. Martini / Consciousness and Cognition 43 (2016) 143–151 145 the functional coupling between areas of the so-called ‘‘pain matrix” and the visual body network areas in the posterior parietal cortex and occipito-temporal areas (Longo, Iannetti, Mancini, Driver, & Haggard, 2012b). Also, the vision of the hand led to a reduction in the activation of the primary somatosensory cortex and the operculo-insular cortex following painful stimuli (Longo, Iannetti, Mancini, Driver, & Haggard, 2012a). Notably, not only the analgesic effects of the vision of the body are restricted to the vision of the own body, but they also seem to be site-specific, so that less pain is felt only when looking at the body site where pain occurs (Diers et al., 2013). In addition, the visual modification (in size) of the body part observed, shapes the pain modulation according to whether the seen body part is enlarged or shrunk (Mancini, Longo, Kammers, & Haggard, 2011). However, results coming from studies that used visual manipulations of the body part, are contradictory. For instance, it has been shown how the vision of a tinier hand seen through a convex mirror leads to a diminished heat pain threshold, while the vision of a bigger hand through a concave mirror increases it (Mancini et al., 2011). Likewise a recent study showed that the visual magnification of the stimulated hand led to a reduced unpleasantness related to the painful stimulation, accompanied by a reduced physiological response (SCR) (Romano & Maravita, 2014). Nevertheless, a study conducted on chronic pain patients reported the opposite results. Here the vision of their magnified hand, led patients with chronic hand pain to feel more pain during the movement of their limb, while the vision of their shrunk hand reduced it (Moseley, Parsons, & Spence, 2008). On the same line Diers and coworkers showed that the observation by chronic pain patients of their own downscaled back reduced their pain, while no effect was reported for the vision of an enlarged back (Diers et al., 2013). A similar result has been found by Ramachandran in a case study with a patient with phantom limb pain: a minimizing lens that gave the illusion that the lost left forearm was back there, although shrunk, reduced the patients’ pain, while a magnifying lens provided no pain modulation (Ramachandran, Brang, & McGeoch, 2009). Conversely, regardless of the type of visual manipulation, osteoarthritis patients seem to benefit from both an illusory shrinkage or stretch of their hand in pain (Preston & Newport, 2011). These contradictory results reported by studies on healthy subjects, as well as chronic pain patients, would point at a complex relationship between pain and the neural representation of the body. A possible explanation could reside in the altered neural representation of the body in chronic pain patients (Gilpin, Moseley, Stanton, & Newport, 2015). So, a visual manipulation of the size of the body would induce a different effect in healthy subjects, with an intact body representation, compared to patients who have a representation of the body distorted (Tsay, Allen, Proske, & Giummarra, 2015). Furthermore, it has been recently discovered that, to be effective in lowering pain intensity, the visual feedback relative to the body has to be ‘‘live” as in a real time video rather than static as in a picture, at least with chronic lower back pain patients (Diers, Löffler, Zieglgänsberger, & Trojan, 2015). Altogether these findings disclose an important modulatory effect of the vision of one’s own body part while being in pain, both in healthy subjects during acute pain models and in patients with chronic pain. Remarkably, the representation of the body can be challenged experimentally. For instance it is possible to incorporate a prosthetic arm in one’s body schema (Mayer, Kudar, Bretz, & Tihanyi, 2008). Moreover, fake body parts can be easily replaced with different versions of it when it comes, for example, to their appearance. Therefore the question arises on whether the analgesic effect of the vision of one’s own body are transferrable to new dummy ‘‘owned” body parts, and if and how the vision of an altered body shapes the pain experience. 3. Rubber hand illusion and pain With a simple experiment, Botvinick and Cohen demonstrated that intermodal matching between a touch felt on a hidden hand and a concomitant tactile stimulation seen on a rubber hand, led the majority of their participants to experience the socalled rubber hand illusion (RHI) (Botvinick & Cohen, 1998). In subjective terms, this illusion brings about the feeling that the rubber hand belongs to one’s own body (body ownership) (Botvinick & Cohen, 1998). At the neural level such experience is reflected by an increased blood oxygen level–dependent (BOLD) signal in the bilateral premotor cortex (Ehrsson, Spence, & Passingham, 2004). Interestingly, such illusion is not constricted to the hand only, but it can be transferred to the whole body (Petkova & Ehrsson, 2008; Petkova, Khoshnevis, & Ehrsson, 2011). Probably the first study to use the principles of the RHI on pain perception has been the one carried out by ValenzuelaMoguillansky, Bouhassira, and O’Regan (2011). In two separate experiments the authors showed how, immediately after the RHI, pain ratings to painful heat stimuli decreased, while the mere observation of the rubber hand without any concomitant tactile stimulation did not yield such analgesic effect. Conversely, in the second experiment the authors found that the pain ratings were higher right after synchronous visuo-tactile stimulation (RHI), than after the asynchronous visuo-tactile condition, although this difference was not significant between conditions. They interpreted these controversial results taking into account different possible factors such as, among others, the ‘violation of the expectation’ taking place during the asynchronous condition (Lewis & Lloyd, 2010), which would lead to disturbing and ambiguous sensations, possibly drawing more attention than during the normal RHI. Also, the authors called into play the role of different types of changes in the body schema of their participants, namely incorporating the rubber hand and disowning the real one. In this case, the vision of an ‘‘owned” rubber hand, painfully stimulated, would bring about higher pain ratings (Valenzuela-Moguillansky et al., 2011). Body ownership illusions have not been limited to a rubber limb, but they have also been extended to whole bodies. Based on a slightly different type of experimental paradigm, Hansel and co-workers reported how by stroking the participants’ back and a mannequin’s back placed in front of them, the participants’ pressure pain threshold increased as compared 146 M. Martini / Consciousness and Cognition 43 (2016) 143–151 to when the participants looked at a non-corporeal object; moreover, the stronger the self-identification with the mannequin the higher the pain threshold (Hänsel, Lenggenhager, von Känel, Curatolo, & Blanke, 2011). However, it should be pointed out that in this experiment there was a lack of correspondence between the body part seen (the mannequin’s back) and the real body part receiving painful stimuli (the index finger). Also, the increased pain threshold was reported regardless of the type of tactile stimulation used (synchronous/asynchronous). In addition, in the paradigm used by Hansel and co-workers the dummy body is seen from a third person perspective and this, along with multisensory correlations, can lead to the socalled Out of Body Experience (OBE) phenomenon (Ehrsson, 2007; Lenggenhager, Tadi, Metzinger, & Blanke, 2007). It could be that, under these circumstances, the analgesic effects of seeing the own body may be mistaken for other analgesiainducing processes, likely related to attentional mechanisms or to the disownership (of the real body) taking place during the OBE (Guterstam & Ehrsson, 2012). Actually, the absence of analgesic processes during body ownership paradigms was reported by Mohan and colleagues in a RHI study. In two different experiments conducted by two independent labs, these authors reported that during the RHI, their participants failed to show a significant modulation of the heat pain as assessed by a visual analogue scale (VAS). Also their participants did not report a modulation of either heat or cold pain threshold (Mohan et al., 2012). However, in disagreement with Mohan’s finding are the results of another RHI-pain study. Using three conditions in a within subjects design, Hegedus and collaborators found that both RHI and vision of the own hand resulted in slightly (but significantly) higher pain thresholds, compared to the asynchronous stroking control condition (Hegedüs et al., 2014). Similar findings were found in a study investigating the effects of the RHI on the perceived discomfort caused by cold stimuli. In their experiment, Siedlecka and colleagues applied ice on the hand of their participants, right after the induction of the RHI (Siedlecka, Klimza, Łukowska, & Wierzchoń, 2014). Two distinct groups of participants were considered, one experimental group where the stroking to induce the illusion was timely and spatially synchronized, and a control group where the stroking was spatially incongruent. The authors found that resistance time to cold stimulation increased and the sensation of discomfort decreased in the experimental group, compared to the resistance time and discomfort levels reached in the control group. Furthermore, the strength of body ownership illusion correlated positively with cold resistance time and with the experienced unpleasantness. These results would support Hegedus and colleagues’ findings, according to which during the vision of an owned rubber hand the thermal pain threshold increases, similarly to what happens during the vision of one’s own real body part in pain. Further support to these findings derive from an experiment with the RHI and the cold pressor test carried out by Giumarra and coworkers. In their experiment these authors found that pain tolerance significantly increases following the RHI compared to a control visuo-tactile asynchronous stimulation (Giummarra, Georgiou-Karistianis, Verdejo-Garcia, & Gibson, 2015). It has been suggested that, by promoting the feeling that a rubber limb is one’s own real limb, the RHI fosters a reorganization of the body image with consequences on the perceptual processes (Longo, Schüür, Kammers, Tsakiris, & Haggard, 2009). Indeed, the physical appearance of the rubber hand seems to play a major role in driving the modulatory effect of the vision of the body on pain perception. So, the analgesic effect of seeing the ‘‘own” body can be abolished, or can even lead to hyperalgesic effects (i.e., higher pain/increased sensitivity to pain), if the ‘‘owned” prosthetic hand looks injured. For instance, capsaicin-induced pain is worsened by the vision of an ‘‘owned” rubber hand that has a scar (Giummarra et al., 2015). Similarly, Osumi and colleagues found that the heat pain threshold significantly decreases when the ‘‘owned” rubber hand looks injured, not only compared to when it looks normal, but also when it is abnormally bent or very hairy (Osumi, Imai, Ueta, Nobusako, & Morioka, 2014). The majority of studies reported here suggest an analgesic effect of the vision of a rubber hand when the hand is incorporated into one’s own body image. Further, the visual characteristics of the rubber hand can determine the presence of either more or less pain, depending on the type of visual manipulation that takes place. Nonetheless, further studies are needed to verify the presence of any causal relationship between the affective dimension of body image and pain perception. This could be facilitated by the usage of between- instead of within-subjects designs and by relying on statistical models that address causation, for instance, path analysis and structural equation modelling (Leppink, 2015). Research manipulating affect without changing visual appearance, and measuring the relation between affect and pain, independent of visual appearance, could also help explore whether there are any causal relations between the affective dimension of body image and pain perception. 4. Virtual hand illusion and pain The term virtual reality (VR) was possibly used for the first time by Antonin Artaud in 1938 in dramaturgy (Artaud, 1938). However, the term as we currently mean it was first introduced in 1989 by Lanier, Minsky, Fisher, and Druin (1989). Since then, many technological advances have been made in terms of computer processing speed, video quality and tracking systems. Nowadays, VR is no longer a prerogative of the entertainment field only, but it has finally found useful applications in teleoperation, psychotherapy, rehabilitation and behavioural neurosciences (Tarr & Warren, 2002). Indeed, Immersive Virtual Reality (IVR) technology represents a powerful tool to generate sensory environments that can be replicated almost identically and that are under the full control of the experimenter (Sanchez-Vives & Slater, 2005). At the beginning of this century, VR has been successfully introduced in pain management (Hoffman, Doctor, Patterson, Carrougher, & Furness, 2000). For instance, it has been shown how playing a video game in VR makes adolescent and adult patients with burn wounds feel less pain during their treatments (Hoffman, Doctor, et al., 2000; Hoffman, Patterson, & M. Martini / Consciousness and Cognition 43 (2016) 143–151 147 Carrougher, 2000). As objective evidence of VR analgesia, Hoffman and colleagues found in an fMRI brain scan study that VR greatly and significantly reduced pain related brain activity (Hoffman et al., 2004). A second fMRI study showed that the amount of pain reduction from VR was comparable to the analgesia from a moderate dose of hydromorphone pain medication (Hoffman et al., 2007). The analgesic properties of VR in this case have been largely attributed to its powerful distractive capacity, which would be higher than other similar non-VR interventions (i.e. video games on traditional PC screens), and would endorse an effective management of mild and severe pain states (Hoffman et al., 2011, 2014). In an elegant and comprehensive review by Malloy and Milling on the effectiveness of VR interventions for pain relief, it has been pointed out that immersive VR technology is more likely than non-immersive set ups to promote analgesia (Malloy & Milling, 2010). The disparity between these two technologies in terms of pain outcomes would be greatly due to the distractibility given by the sense of presence, that refers to the feeling of being actually immersed in the virtual world, which would be much higher in immersive scenarios (Sanchez-Vives & Slater, 2005). The analgesic effectiveness of the immersive scenarios would be further improved by the usage of High-Tech helmets, which enables a wider field-of-view, likely boosting the sense of presence and drawing more attentional resources (Hoffman et al., 2006). Additionally, not only the sense of presence influences the effectiveness of VR through distraction, but also anxiety as well as positive emotions such as fun, may affect the experience of pain during the VR exposure (for an extensive review on the topic, see Triberti, Repetto, & Riva, 2014). For example, Maani and co-workers found that patients (soldiers with severe combat-related burn injuries) reported significantly less pain and rated fun during burn wound care significantly higher during VR (Maani et al., 2011). However, not only VR can be used to display virtual interactive scenarios that promote sense of presence and attentional engagement, but it has also been used effectively in the induction of the illusion of ownership over a virtual body (Bergström, Kilteni, & Slater, 2016; GonzálezFranco, Peck, Rodríguez-Fornells, & Slater, 2014; Kilteni, Grau-Sánchez, Veciana De Las Heras, Rodríguez-Fornells, & Slater, 2016; Kilteni, Groten, & Slater, 2012; Kilteni, Maselli, Kording, & Slater, 2015; Kilteni, Normand, Sanchez-Vives, & Slater, 2012; Maselli & Slater, 2013; Slater, Perez-Marcos, Ehrsson, & Sanchez-Vives, 2008; Slater, Perez-Marcos, Ehrsson, & Sanchez-Vives, 2009). Similar to what happens in the aforementioned RHI, during this illusion, participants feel a virtual body part as belonging to their own body, and this feeling can be extended to the entire virtual body (Maselli & Slater, 2013). Such process can lead to profound psychological (Peck, Seinfeld, Aglioti, & Slater, 2013) and perceptual (Banakou, Groten, & Slater, 2013) consequences, as people react to it as if it was their own real body (Slater, Spanlang, SanchezVives, & Blanke, 2010). On this premise, Martini and colleagues tested the effect of virtual body ownership on pain perception. The authors showed how only during the vision of a virtual arm there was a substantial increase in the heat pain threshold as compared to the vision of either a virtual or a real object. Notably, the effect was not attributable to the mere observation of the virtual arm but, rather, the feeling of ownership over the virtual limb was crucial for the analgesic effect to emerge (Martini, Perez-Marcos, & Sanchez-Vives, 2014). Therefore it seems that the analgesic effect of the vision of one’s own body part in pain holds true even during the vision of a virtual body, provided that there is a feeling of ownership over it. Once again, the analgesic effect seems to be strictly connected with the visual properties of the body. So for instance, if the body visually fades away, also the pain sensation will be affected by it. This has been shown in a study where, in different conditions, the avatar’s body was rendered with different levels of transparency. Here, the more the participants felt that the transparent virtual arm was their own limb, the lower was their heat pain threshold (Martini, Kilteni, Maselli, & SanchezVives, 2015). Therefore, similar to what has been found for rubber arms, the vision of an altered body can disrupt the emergence of the analgesic effects and it could even make pain worse. What seems to be a key factor in driving the modulation of pain is the expectation created by the body-related visual input. For instance, in an experiment using visual manipulations of the virtual body and heat pain threshold, changes in colour of the virtual arm differently drove the pain threshold of the participants according to the kind of colour displayed: the vision of a bluish ‘‘cold” arm led to a higher pain threshold as compared to a reddish ‘‘hot” arm, while a neutral green arm stood right in the middle (Martini, Perez-Marcos, & Sanchez-Vives, 2013). Exploiting the powerful association that colours have with temperatures, the vision of one’s own arm changing colour triggered top-down processes that differently modulated pain perception (Martini et al., 2013). Recently Romano and co-workers reported that physiological responses to pain, as measured via SCR, are reduced when the participants look at the avatar’s body from a first person perspective as compared to the vision of the avatar’s body turned 90 degrees from the real. Moreover, in the same study the authors show how the physiological response is negatively correlated with the size of the virtual body, so that the bigger the body the lower the SCR (Romano, Llobera, & Blanke, 2015). This result would support the analgesic effects of the vision of the own magnified body part on pain, which in healthy subjects would boost analgesia (Mancini et al., 2011; Romano & Maravita, 2014). 5. Conclusions Body and pain are strictly related, since the former is where the latter ‘happens’ and is at the same time perceived. Even the neural representation of the body seems to be quite critical in shaping the pain experience. For instance, many chronic pain patients report having a distorted representation of their affected body part (Gilpin et al., 2015; Melzack, 1990) and cortical misrepresentations of the body are associated with chronic pain states (Lotze & Moseley, 2007; Preston & Newport, 2011). However, it is not completely clear whether the vision of the body promotes a correct representation of the body via online visual feedback. Additionally, it is known how pain can be modulated by different psychological factors. In pain experiments 148 M. Martini / Consciousness and Cognition 43 (2016) 143–151 involving the usage of a dummy body, these factors interact with the illusion of body ownership and with the physical properties of that body, modulating the pain output (Fig. 1). For instance patients with schizophrenia, as well as healthy subjects taking psychoactive drugs mimicking its symptoms, are more likely to experience the illusion of ownership over a rubber hand (Morgan et al., 2011; Peled, Ritsner, Hirschmann, Geva, & Modai, 2000). At the same time anxiety, depression and anger are known to increase pain, while positive emotions usually decrease perceived pain (Peters, 2015). Cognitive factors such as attention, expectancy, appraisal and multisensory integrations can either increase or decrease pain and interact with emotional factors (see Peters, 2015; Senkowski, Höfle, & Engel, 2014; Triberti et al., 2014, as recent reviews on these topics). The physical appearance of the body may also play a significant role. Indeed the vision of a distorted body appearance may modulate the body ownership levels (Martini et al., 2015), but it may also affect psychological factors such as body image (Osumi, Imai, Ueta, Nobusako, et al., 2014) and modulate pain perception during illusory body ownership paradigms (Martini et al., 2013, 2015; Osumi, Imai, Ueta, Nakano, et al., 2014; Osumi, Imai, Ueta, Nobusako, et al., 2014) or in studies involving the vision of one’s own real body (Mancini et al., 2011; Romano & Maravita, 2014). Attention is certainly one of the most important factors in the modulation of pain (Legrain et al., 2009). Yet, the analgesic effect linked to the vision of one’s own body is far from merely being due to distraction mechanisms. Interestingly, it has been proposed that such effects entail the shaping of the somatosensory maps in the primary somatosensory cortex, through an increase in the intracortical inhibition (Haggard, Iannetti, & Longo, 2013). Studies relying on bodily illusions, either taking into account rubber hands or virtual limbs, suggest that this analgesic effect can be transferrable to the vision of fake dummy bodies, provided that these are perceived as belonging to one’s own body. Nonetheless, further studies are needed to disclose the full potential that bodily illusions might have on pain, especially with the use of IVR. For instance, future investigations making use of dummy bodies or virtual reality set-ups may want to consider much larger sample sizes to gain a greater statistical power, to be able to better understand the contribution that the subjective feeling of ownership over fake body parts has in pain modulation. How the vision of the body and emotional/cognitive factors interact must also be better elucidated. In a recent research article using two different heat-pain intensities, it has been shown that viewing their own body, as compared to viewing a non-corporeal object (a foam block), brought about a report bias toward higher pain judgments in the participants, as well as a reduction in the discriminability between the stimulation intensities. However, seeing the thermode probe approaching the body eliminated the reduction in discriminability (Beck, Làdavas, & Haggard, 2016). Importantly, the authors reported that such difference was not accounted for differences in stimulation predictability. These findings show the importance played by psychological factors, such as task-related goals and the visual context, in shaping the pain experience during the observation of the body (Beck et al., 2016). At last but not least, despite the abundance of neuroimaging studies on pain perception, the majority of the studies on pain and the vision of the body during illusory body ownership were conducted without using neuroimaging techniques. Future studies on this field may consider the utilization of brain stimulation (for ex. TMS, tDCS) and/or brain imaging techniques (for ex. fMRI, EEG, fNIRS) to shed further light on the identification of a neural web underlying this complex phenomenon. Fig. 1. Schematic representation of the factors involved in pain modulation during body ownership (B.O.) paradigms: multisensory correlations facilitate the onset of B.O. over a fake body part, but afferent multisensory cues may also modulate pain perception. The appearance of the body may regulate B.O. levels but also affect psychological factors such as body image and expectation, all of which may affect pain perception. Psychological factors embrace both cognitive (e.g. attentional resources, expectancy and appraisal) and emotional factors (e.g. positive or negative mood). Both body appearance and multisensory processing are strictly related to the affective/cognitive domain of the ‘‘psychological factors” group, but here are depicted as separate factors for the sake of clarity, because they are often studied as separate factors in B.O. experiments and may contribute to shape the pain experience independently. M. Martini / Consciousness and Cognition 43 (2016) 143–151 149 Acknowledgments The author is grateful to Tasha Oneida Lawson for her useful comments. References Apkarian, A. V., Hashmi, J. A., & Baliki, M. N. (2011). Pain and the brain: Specificity and plasticity of the brain in clinical chronic pain. Pain, 152(3 Suppl.), S49–S64. http://dx.doi.org/10.1016/j.pain.2010.11.010. Artaud, A. (1938). The theatre and its double. Translated by Mary Press., Richards (Vol. 958). New York: Grove. Banakou, D., Groten, R., & Slater, M. (2013). Illusory ownership of a virtual child body causes overestimation of object sizes and implicit attitude changes. Proceedings of the National Academy of Sciences of the United States of America, 110(31), 12846–12851. http://dx.doi.org/10.1073/pnas.1306779110. Beck, B., Làdavas, E., & Haggard, P. (2016). Viewing the body modulates both pain sensations and pain responses. Experimental Brain Research. http://dx.doi. org/10.1007/s00221-016-4585-9. Bergström, I., Kilteni, K., & Slater, M. (2016). First-person perspective virtual body posture influences stress: A virtual reality body ownership study. PLoS ONE, 11(2), e0148060. http://dx.doi.org/10.1371/journal.pone.0148060. Botvinick, M., & Cohen, J. (1998). Rubber hands ‘‘feel” touch that eyes see. Nature, 391(6669), 756. http://dx.doi.org/10.1038/35784. Botvinick, M., Jha, A. P., Bylsma, L. M., Fabian, S. A., Solomon, P. E., & Prkachin, K. M. (2005). Viewing facial expressions of pain engages cortical areas involved in the direct experience of pain. NeuroImage, 25(1), 312–319. http://dx.doi.org/10.1016/j.neuroimage.2004.11.043. Campbell, G., Bruno, R., Darke, S., Shand, F., Hall, W., Farrell, M., & Degenhardt, L. (2015). Prevalence and correlates of suicidal thoughts and suicide attempts in people prescribed pharmaceutical opioids for chronic pain. The Clinical Journal of Pain. http://dx.doi.org/10.1097/AJP.0000000000000283. Cardini, F., Longo, M. R., & Haggard, P. (2011). Vision of the body modulates somatosensory intracortical inhibition. Cerebral Cortex (New York, NY: 1991), 21 (9), 2014–2022. http://dx.doi.org/10.1093/cercor/bhq267. Carter, G. T., Duong, V., Ho, S., Ngo, K. C., Greer, C. L., & Weeks, D. L. (2014). Side effects of commonly prescribed analgesic medications. Physical Medicine and Rehabilitation Clinics of North America, 25(2), 457–470. http://dx.doi.org/10.1016/j.pmr.2014.01.007. Diers, M., Löffler, A., Zieglgänsberger, W., & Trojan, J. (2015). Watching your pain site reduces pain intensity in chronic back pain patients. European Journal of Pain (London, England). <http://doi.org/10.1002/ejp.765>. Diers, M., Zieglgänsberger, W., Trojan, J., Drevensek, A. M., Erhardt-Raum, G., & Flor, H. (2013). Site-specific visual feedback reduces pain perception. Pain, 154(6), 890–896. http://dx.doi.org/10.1016/j.pain.2013.02.022. Ehrsson, H. H. (2007). The experimental induction of out-of-body experiences. Science (New York, NY), 317(5841), 1048. http://dx.doi.org/ 10.1126/science.1142175. Ehrsson, H. H., Spence, C., & Passingham, R. E. (2004). That’s my hand! Activity in premotor cortex reflects feeling of ownership of a limb. Science (New York, NY), 305(5685), 875–877. http://dx.doi.org/10.1126/science.1097011. Gilpin, H. R., Bellan, V., Gallace, A., & Moseley, G. L. (2014). Exploring the roles of body ownership, vision and virtual reality on heat pain threshold. European Journal of Pain, 18(7), 900–901. http://dx.doi.org/10.1002/j.1532-2149.2014.483.x. Gilpin, H. R., Moseley, G. L., Stanton, T. R., & Newport, R. (2015). Evidence for distorted mental representation of the hand in osteoarthritis. Rheumatology (Oxford, England), 54(4), 678–682. http://dx.doi.org/10.1093/rheumatology/keu367. Giummarra, M. J., Georgiou-Karistianis, N., Verdejo-Garcia, A., & Gibson, S. J. (2015). Feeling the burn: When it looks like it hurts, and belongs to me, it really does hurt more. Consciousness and Cognition, 36, 314–326. http://dx.doi.org/10.1016/j.concog.2015.07.010. González-Franco, M., Peck, T. C., Rodríguez-Fornells, A., & Slater, M. (2014). A threat to a virtual hand elicits motor cortex activation. Experimental Brain Research, 232(3), 875–887. http://dx.doi.org/10.1007/s00221-013-3800-1. Guterstam, A., & Ehrsson, H. H. (2012). Disowning one’s seen real body during an out-of-body illusion. Consciousness and Cognition, 21(2), 1037–1042. http:// dx.doi.org/10.1016/j.concog.2012.01.018. Haggard, P., Iannetti, G. D., & Longo, M. R. (2013). Spatial sensory organization and body representation in pain perception. Current Biology: CB, 23(4), R164–R176. http://dx.doi.org/10.1016/j.cub.2013.01.047. Hänsel, A., Lenggenhager, B., von Känel, R., Curatolo, M., & Blanke, O. (2011). Seeing and identifying with a virtual body decreases pain perception. European Journal of Pain (London, England), 15(8), 874–879. http://dx.doi.org/10.1016/j.ejpain.2011.03.013. Hegedüs, G., Darnai, G., Szolcsányi, T., Feldmann, A., Janszky, J., & Kállai, J. (2014). The rubber hand illusion increases heat pain threshold. European Journal of Pain (London, England). http://dx.doi.org/10.1002/j.1532-2149.2014.00466.x. Hoffman, H. G., Chambers, G. T., Meyer, W. J., 3rd, Arceneaux, L. L., Russell, W. J., Seibel, E. J., ... Patterson, D. R. (2011). Virtual reality as an adjunctive nonpharmacologic analgesic for acute burn pain during medical procedures. Annals of Behavioral Medicine: A Publication of the Society of Behavioral Medicine, 41(2), 183–191. http://dx.doi.org/10.1007/s12160-010-9248-7. Hoffman, H. G., Doctor, J. N., Patterson, D. R., Carrougher, G. J., & Furness, T. A. (2000). Virtual reality as an adjunctive pain control during burn wound care in adolescent patients. Pain, 85(1–2), 305–309. Retrieved from <http://www.ncbi.nlm.nih.gov/pubmed/10692634>. Hoffman, H. G., Meyer, W. J., Ramirez, M., Roberts, L., Seibel, E. J., Atzori, B., ... Patterson, D. R. (2014). Feasibility of articulated arm mounted oculus rift virtual reality goggles for adjunctive pain control during occupational therapy in pediatric burn patients. Cyberpsychology, Behavior, and Social Networking, 17 (6), 397–401. http://dx.doi.org/10.1089/cyber.2014.0058. Hoffman, H. G., Patterson, D. R., & Carrougher, G. J. (2000). Use of virtual reality for adjunctive treatment of adult burn pain during physical therapy: A controlled study. The Clinical Journal of Pain, 16(3), 244–250. Retrieved from <http://www.ncbi.nlm.nih.gov/pubmed/11014398>. Hoffman, H. G., Richards, T. L., Coda, B., Bills, A. R., Blough, D., Richards, A. L., & Sharar, S. R. (2004). Modulation of thermal pain-related brain activity with virtual reality: Evidence from fMRI. NeuroReport, 15(8), 1245–1248. Retrieved from <http://www.ncbi.nlm.nih.gov/pubmed/15167542>. Hoffman, H. G., Richards, T. L., Van Oostrom, T., Coda, B. A., Jensen, M. P., Blough, D. K., & Sharar, S. R. (2007). The analgesic effects of opioids and immersive virtual reality distraction: Evidence from subjective and functional brain imaging assessments. Anesthesia and Analgesia, 105(6), 1776–1783. http://dx. doi.org/10.1213/01.ane.0000270205.45146.d (table of contents). Hoffman, H. G., Seibel, E. J., Richards, T. L., Furness, T. A., Patterson, D. R., & Sharar, S. R. (2006). Virtual reality helmet display quality influences the magnitude of virtual reality analgesia. The Journal of Pain: Official Journal of the American Pain Society, 7(11), 843–850. http://dx.doi.org/10.1016/j. jpain.2006.04.006. Kilteni, K., Grau-Sánchez, J., Veciana De Las Heras, M., Rodríguez-Fornells, A., & Slater, M. (2016). Decreased corticospinal excitability after the illusion of missing part of the arm. Frontiers in Human Neuroscience, 10, 145. http://dx.doi.org/10.3389/fnhum.2016.00145. Kilteni, K., Groten, R., & Slater, M. (2012). The sense of embodiment in virtual reality. Presence: Teleoperators and Virtual Environments, 21(4), 373–387. http:// dx.doi.org/10.1162/PRES_a_00124. Kilteni, K., Maselli, A., Kording, K. P., & Slater, M. (2015). Over my fake body: Body ownership illusions for studying the multisensory basis of own-body perception. Frontiers in Human Neuroscience, 9, 141. http://dx.doi.org/10.3389/fnhum.2015.00141. Kilteni, K., Normand, J.-M., Sanchez-Vives, M. V., & Slater, M. (2012). Extending body space in immersive virtual reality: A very long arm illusion. PLoS ONE, 7 (7), e40867. http://dx.doi.org/10.1371/journal.pone.0040867. Lamm, C., Batson, C. D., & Decety, J. (2007). The neural substrate of human empathy: Effects of perspective-taking and cognitive appraisal. Journal of Cognitive Neuroscience, 19(1), 42–58. http://dx.doi.org/10.1162/jocn.2007.19.1.42. Lanier, J., Minsky, M., Fisher, S., & Druin, A. (1989). Virtual environments and interactivity: Windows to the future. In ACM SIGGRAPH panel proceedings. 150 M. Martini / Consciousness and Cognition 43 (2016) 143–151 Legrain, V., Damme, S. Van., Eccleston, C., Davis, K. D., Seminowicz, D. A., & Crombez, G. (2009). A neurocognitive model of attention to pain: Behavioral and neuroimaging evidence. Pain, 144(3), 230–232. http://dx.doi.org/10.1016/j.pain.2009.03.020. Legrain, V., Mancini, F., Sambo, C. F., Torta, D. M., Ronga, I., & Valentini, E. (2012). Cognitive aspects of nociception and pain. Bridging neurophysiology with cognitive psychology. Neurophysiologie Clinique/Clinical Neurophysiology, 42(5), 325–336. http://dx.doi.org/10.1016/j.neucli.2012.06.003. Lenggenhager, B., Tadi, T., Metzinger, T., & Blanke, O. (2007). Video ergo sum: Manipulating bodily self-consciousness. Science (New York, NY), 317(5841), 1096–1099. http://dx.doi.org/10.1126/science.1143439. Leppink, J. (2015). On causality and mechanisms in medical education research: An example of path analysis. Perspectives on Medical Education, 4(2), 66–72. http://dx.doi.org/10.1007/s40037-015-0174-z. Lewis, E., & Lloyd, D. M. (2010). Embodied experience: A first-person investigation of the rubber hand illusion. Phenomenology and the Cognitive Sciences, 9 (3), 317–339. http://dx.doi.org/10.1007/s11097-010-9154-2. Loeser, J. D., & Treede, R.-D. (2008). The kyoto protocol of IASP basic pain terminology. Pain, 137(3), 473–477. http://dx.doi.org/10.1016/j.pain.2008.04.025. Longo, M. R., Betti, V., Aglioti, S. M., & Haggard, P. (2009). Visually induced analgesia: Seeing the body reduces pain. The Journal of Neuroscience: The Official Journal of the Society for Neuroscience, 29(39), 12125–12130. http://dx.doi.org/10.1523/JNEUROSCI.3072-09.2009. Longo, M. R., Iannetti, G. D., Mancini, F., Driver, J., & Haggard, P. (2012a). Linking pain and the body: Neural correlates of visually induced analgesia. The Journal of Neuroscience: The Official Journal of the Society for Neuroscience, 32(8), 2601–2607. http://dx.doi.org/10.1523/JNEUROSCI.4031-11.2012. Longo, M. R., Iannetti, G. D., Mancini, F., Driver, J., & Haggard, P. (2012b). Linking pain and the body: Neural correlates of visually induced analgesia. The Journal of Neuroscience: The Official Journal of the Society for Neuroscience, 32(8), 2601–2607. http://dx.doi.org/10.1523/JNEUROSCI.4031-11.2012. Longo, M. R., Schüür, F., Kammers, M. P. M., Tsakiris, M., & Haggard, P. (2009). Self awareness and the body image. Acta Psychologica, 132(2), 166–172. http:// dx.doi.org/10.1016/j.actpsy.2009.02.003. Lotze, M., & Moseley, G. L. (2007). Role of distorted body image in pain. Current Rheumatology Reports, 9(6), 488–496. Retrieved from <http://www.ncbi.nlm. nih.gov/pubmed/18177603>. Maani, C. V., Hoffman, H. G., Morrow, M., Maiers, A., Gaylord, K., McGhee, L. L., & DeSocio, P. A. (2011). Virtual reality pain control during burn wound debridement of combat-related burn injuries using robot-like arm mounted VR goggles. The Journal of Trauma, 71(1 Suppl.), S125–S130. http://dx.doi. org/10.1097/TA.0b013e31822192e2. Macaluso, E., & Maravita, A. (2010). The representation of space near the body through touch and vision. Neuropsychologia, 48(3), 782–795. http://dx.doi.org/ 10.1016/j.neuropsychologia.2009.10.010. Malloy, K. M., & Milling, L. S. (2010). The effectiveness of virtual reality distraction for pain reduction: A systematic review. Clinical Psychology Review, 30(8), 1011–1018. http://dx.doi.org/10.1016/j.cpr.2010.07.001. Mancini, F., Longo, M. R., Canzoneri, E., Vallar, G., & Haggard, P. (2013). Changes in cortical oscillations linked to multisensory modulation of nociception. The European Journal of Neuroscience, 37(5), 768–776. http://dx.doi.org/10.1111/ejn.12080. Mancini, F., Longo, M. R., Kammers, M. P. M., & Haggard, P. (2011). Visual distortion of body size modulates pain perception. Psychological Science, 22(3), 325–330. http://dx.doi.org/10.1177/0956797611398496. Martini, M., Kilteni, K., Maselli, A., & Sanchez-Vives, M. V. (2015). The body fades away: Investigating the effects of transparency of an embodied virtual body on pain threshold and body ownership. Scientific Reports, 5, 13948. http://dx.doi.org/10.1038/srep13948. Martini, M., Perez-Marcos, D., & Sanchez-Vives, M. V. (2013). What color is my arm? Changes in skin color of an embodied virtual arm modulates pain threshold. Frontiers in Human Neuroscience, 7, 438. http://dx.doi.org/10.3389/fnhum.2013.00438. Martini, M., Perez-Marcos, D., & Sanchez-Vives, M. V. (2014). Modulation of pain threshold by virtual body ownership. European Journal of Pain (London, England). http://dx.doi.org/10.1002/j.1532-2149.2014.00451. Martini, M., Perez-Marcos, D., & Sanchez-Vives, M. V. (2015). Author’s reply to the commentary by Gilpin et al.. European Journal of Pain (London, England), 19 (1), 143–144. http://dx.doi.org/10.1002/ejp.606. Maselli, A., & Slater, M. (2013). The building blocks of the full body ownership illusion. Frontiers in Human Neuroscience, 7, 83. http://dx.doi.org/10.3389/ fnhum.2013.00083. Mayer, A., Kudar, K., Bretz, K., & Tihanyi, J. (2008). Body schema and body awareness of amputees. Prosthetics and Orthotics International, 32(3), 363–382. http://dx.doi.org/10.1080/03093640802024971. Medina, J., & Coslett, H. B. (2010). From maps to form to space: Touch and the body schema. Neuropsychologia, 48(3), 645–654. http://dx.doi.org/10.1016/j. neuropsychologia.2009.08.017. Melzack, R. (1990). Phantom limbs and the concept of a neuromatrix. Trends in Neurosciences, 13(3), 88–92. Moayedi, M., & Davis, K. D. (2013). Theories of pain: From specificity to gate control. Journal of Neurophysiology, 109(1), 5–12. http://dx.doi.org/10.1152/ jn.00457.2012. Mohan, R., Jensen, K. B., Petkova, V. I., Dey, A., Barnsley, N., Ingvar, M., ... Ehrsson, H. H. (2012). No pain relief with the rubber hand illusion. PLoS ONE, 7(12), e52400. http://dx.doi.org/10.1371/journal.pone.0052400. Morgan, H. L., Turner, D. C., Corlett, P. R., Absalom, A. R., Adapa, R., Arana, F. S., ... Fletcher, P. C. (2011). Exploring the impact of ketamine on the experience of illusory body ownership. Biological Psychiatry, 69(1), 35–41. http://dx.doi.org/10.1016/j.biopsych.2010.07.032. Moseley, G. L., Parsons, T. J., & Spence, C. (2008). Visual distortion of a limb modulates the pain and swelling evoked by movement. Current Biology: CB, 18 (22), R1047–R1048. http://dx.doi.org/10.1016/j.cub.2008.09.031. Mouraux, A., Diukova, A., Lee, M. C., Wise, R. G., & Iannetti, G. D. (2011). A multisensory investigation of the functional significance of the ‘‘pain matrix”. NeuroImage, 54(3), 2237–2249. http://dx.doi.org/10.1016/j.neuroimage.2010.09.084. Mouraux, A., & Iannetti, G. D. (2009). Nociceptive laser-evoked brain potentials do not reflect nociceptive-specific neural activity. Journal of Neurophysiology, 101(6), 3258–3269. http://dx.doi.org/10.1152/jn.91181.2008. Osumi, M., Imai, R., Ueta, K., Nakano, H., Nobusako, S., & Morioka, S. (2014). Factors associated with the modulation of pain by visual distortion of body size. Frontiers in Human Neuroscience, 8, 137. http://dx.doi.org/10.3389/fnhum.2014.00137. Osumi, M., Imai, R., Ueta, K., Nobusako, S., & Morioka, S. (2014). Negative body image associated with changes in the visual body appearance increases pain perception. PLoS ONE, 9(9), e107376. http://dx.doi.org/10.1371/journal.pone.0107376. Peck, T. C., Seinfeld, S., Aglioti, S. M., & Slater, M. (2013). Putting yourself in the skin of a black avatar reduces implicit racial bias. Consciousness and Cognition, 22(3), 779–787. http://dx.doi.org/10.1016/j.concog.2013.04.016. Peled, A., Ritsner, M., Hirschmann, S., Geva, A. B., & Modai, I. (2000). Touch feel illusion in schizophrenic patients. Biological Psychiatry, 48(11), 1105–1108. Retrieved from <http://www.ncbi.nlm.nih.gov/pubmed/11094144>. Peters, M. L. (2015). Emotional and cognitive influences on pain experience. Modern Trends in Pharmacopsychiatry, 30, 138–152. http://dx.doi.org/10.1159/ 000435938. Petkova, V. I., & Ehrsson, H. H. (2008). If I were you: Perceptual illusion of body swapping. PLoS ONE, 3(12), e3832. http://dx.doi.org/10.1371/journal. pone.0003832. Petkova, V. I., Khoshnevis, M., & Ehrsson, H. H. (2011). The perspective matters! Multisensory integration in ego-centric reference frames determines fullbody ownership. Frontiers in Psychology, 2, 35. http://dx.doi.org/10.3389/fpsyg.2011.00035. Preston, C., & Newport, R. (2011). Analgesic effects of multisensory illusions in osteoarthritis. Rheumatology (Oxford, England), 50(12), 2314–2315. http://dx. doi.org/10.1093/rheumatology/ker104. Ramachandran, V. S., Brang, D., & McGeoch, P. D. (2009). Size reduction using Mirror Visual Feedback (MVF) reduces phantom pain. Neurocase, 15(5), 357–360. http://dx.doi.org/10.1080/13554790903081767. Ramachandran, V. S., & Rogers-Ramachandran, D. (1996). Synaesthesia in phantom limbs induced with mirrors. Proceedings of the Royal Society B: Biological Sciences, 263(1369), 377–386. http://dx.doi.org/10.1098/rspb.1996.0058. M. Martini / Consciousness and Cognition 43 (2016) 143–151 151 Romano, D., Llobera, J., & Blanke, O. (2015). Size and viewpoint of an embodied virtual body impact the processing of painful stimuli. The Journal of Pain: Official Journal of the American Pain Society. http://dx.doi.org/10.1016/j.jpain.2015.11.005. Romano, D., & Maravita, A. (2014). The visual size of one’s own hand modulates pain anticipation and perception. Neuropsychologia, 57, 93–100. http://dx. doi.org/10.1016/j.neuropsychologia.2014.03.002. Saarela, M. V., Hlushchuk, Y., Williams, A. C. de C., Schürmann, M., Kalso, E., & Hari, R. (2007). The compassionate brain: Humans detect intensity of pain from another’s face. Cerebral Cortex (New York, NY: 1991), 17(1), 230–237. http://dx.doi.org/10.1093/cercor/bhj141. Sanchez-Vives, M. V., & Slater, M. (2005). From presence to consciousness through virtual reality. Nature Reviews. Neuroscience, 6(4), 332–339. http://dx.doi. org/10.1038/nrn1651. Sanchez-Vives, M. V., & Slater, M. (2005). From presence to consciousness through virtual reality. Nature Reviews Neuroscience, 6(4), 332–339. Sanchez-Vives, M. V., Spanlang, B., Frisoli, A., Bergamasco, M., & Slater, M. (2010). Virtual hand illusion induced by visuomotor correlations. PLoS ONE, 5(4), e10381. http://dx.doi.org/10.1371/journal.pone.0010381. Schweinhardt, P., & Bushnell, M. C. (2010). Pain imaging in health and disease–How far have we come? The Journal of Clinical Investigation, 120(11), 3788–3797. http://dx.doi.org/10.1172/JCI43498. Senkowski, D., Höfle, M., & Engel, A. K. (2014). Crossmodal shaping of pain: A multisensory approach to nociception. Trends in Cognitive Sciences, 18(6), 319–327. http://dx.doi.org/10.1016/j.tics.2014.03.005. Serino, A., & Haggard, P. (2010). Touch and the body. Neuroscience and Biobehavioral Reviews, 34(2), 224–236. http://dx.doi.org/10.1016/j. neubiorev.2009.04.004. Siedlecka, M., Klimza, A., Łukowska, M., & Wierzchoń, M. (2014). Rubber hand illusion reduces discomfort caused by cold stimulus. PLoS ONE, 9(10), e109909. http://dx.doi.org/10.1371/journal.pone.0109909. Slater, M., Perez-Marcos, D., Ehrsson, H. H., & Sanchez-Vives, M. V. (2008). Towards a digital body: The virtual arm illusion. Frontiers in Human Neuroscience, 2, 6. http://dx.doi.org/10.3389/neuro.09.006.2008. Slater, M., Perez-Marcos, D., Ehrsson, H. H., & Sanchez-Vives, M. V. (2009). Inducing illusory ownership of a virtual body. Frontiers in Neuroscience, 3(2), 214–220. http://dx.doi.org/10.3389/neuro.01.029.2009. Slater, M., Spanlang, B., Sanchez-Vives, M. V., & Blanke, O. (2010). First person experience of body transfer in virtual reality. PLoS ONE, 5(5), e10564. http:// dx.doi.org/10.1371/journal.pone.0010564. Tarr, M. J., & Warren, W. H. (2002). Virtual reality in behavioral neuroscience and beyond. Nature Neuroscience, 5(Suppl.), 1089–1092. http://dx.doi.org/ 10.1038/nn948. Tipper, S. P., Lloyd, D., Shorland, B., Dancer, C., Howard, L. A., & McGlone, F. (1998). Vision influences tactile perception without proprioceptive orienting. NeuroReport, 9(8), 1741–1744. Retrieved from <http://www.ncbi.nlm.nih.gov/pubmed/9665593>. Triberti, S., Repetto, C., & Riva, G. (2014). Psychological factors influencing the effectiveness of virtual reality-based analgesia: A systematic review. Cyberpsychology, Behavior and Social Networking, 17(6), 335–345. http://dx.doi.org/10.1089/cyber.2014.0054. Tsay, A., Allen, T. J., Proske, U., & Giummarra, M. J. (2015). Sensing the body in chronic pain: A review of psychophysical studies implicating altered body representation. Neuroscience and Biobehavioral Reviews, 52, 221–232. http://dx.doi.org/10.1016/j.neubiorev.2015.03.004. Valenzuela-Moguillansky, C., Bouhassira, D., & O’Regan, J. K. (2011). Role of body awareness. Journal of Consciousness Studies (9–10), 110–142. Retrieved from <http://www.ingentaconnect.com/content/imp/jcs/2011/00000018/f0020009/art00006>. Valeriani, M., Betti, V., Le Pera, D., De Armas, L., Miliucci, R., Restuccia, D., ... Aglioti, S. M. (2008). Seeing the pain of others while being in pain: A laser-evoked potentials study. NeuroImage, 40(3), 1419–1428. http://dx.doi.org/10.1016/j.neuroimage.2007.12.056. Vierck, C. J., Whitsel, B. L., Favorov, O. V., Brown, A. W., & Tommerdahl, M. (2013). Role of primary somatosensory cortex in the coding of pain. Pain, 154(3), 334–344. http://dx.doi.org/10.1016/j.pain.2012.10.021. Wager, T. D., Atlas, L. Y., Lindquist, M. A., Roy, M., Woo, C.-W., & Kross, E. (2013). An fMRI-based neurologic signature of physical pain. The New England Journal of Medicine, 368(15), 1388–1397. http://dx.doi.org/10.1056/NEJMoa1204471. Wesslein, A.-K., Spence, C., & Frings, C. (2014). Vision affects tactile target and distractor processing even when space is task-irrelevant. Frontiers in Psychology, 5, 84. http://dx.doi.org/10.3389/fpsyg.2014.00084.

Consciousness and Cognition 19 (2010) 745–750 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Approaching self-deception: How Robert Audi and I part company q Alfred Mele Florida State University, Department of Philosophy, Tallahassee, FL 32306-1500, United States a r t i c l e i n f o Article history: Available online 7 July 2010 Keywords: Belief Deception Experimental philosophy Self-deception a b s t r a c t This article explores fundamental differences between Robert Audi’s position on selfdeception and mine. Although we both depart from a model of self-deception that is straightforwardly based on stereotypical interpersonal deception, we differ in how we do that. An important difference between us might be partly explained by a difference in how we understand the kind of deceiving that is most relevant to self-deception. Ó 2010 Elsevier Inc. All rights reserved. 1. Introduction Robert Audi and I have respectfully disagreed about self-deception for a long time (see Mele, 1982). In fact, Audi’s work on self-deception is what first attracted my attention to the topic. In the present article, I return to my roots in a way. My point of departure is a relatively recent exchange of ideas between Audi and me on self-deception (Audi, 2007; Mele, 2007, pp. 252– 254). In Mele, 2007, I concentrate on one of the novel ideas in Audi’s work on self-deception, his position on the relationship between self-deception and delusion. Here I focus on a more basic disagreement between us – a disagreement about what self-deceived people do and do not believe. In my view, at least in typical cases of self-deception, people believe a false proposition regarding which they are self-deceived (Mele, 2001). In Audi’s view, although the self-deceived person sincerely avows a false proposition p, he does not actually believe that p (see below). Some philosophers defend positions on this particular point of disagreement that are more similar to Audi’s than to mine (Bach, 1981; Funkhouser, 2005; Gendler, 2007; Rey, 1988); and the converse is true of many other philosophers. However, my concern here is my disagreement with Audi in particular. 2. Orienting models and a basic disagreement Audi writes: Given our wide agreement on data, why do we differ on self-deception? I credit the influence of a kind of difference more common in philosophy than many people realize: a difference in orienting models. Mele’s model is apparently the act of deceiving, in which the deceived forms a false belief and so does not see the truth in question; mine is the state of being deceived – altered to apply, so far as possible, to a person who is both deceived and deceiver and hence, apart from inconsistent beliefs, must see the truth (2007, p. 253). There is something to this. But, in my opinion, our most basic disagreement about self-deception is not explained by the difference in orienting models Audi describes here. A distinctive feature of Audi’s view is that no one who is self-deceived with q This article is part of a special issue of this journal on Self, Other and Memory. E-mail address: amele@fsu.edu 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.06.009 746 A. Mele / Consciousness and Cognition 19 (2010) 745–750 respect to a false proposition, p, actually believes that p (1982, p. 147); instead, the self-deceived person ‘‘sincerely” (p. 137) – or ‘‘non-lyingly” (p. 139) – avows p or is disposed so to avow it (also see Audi, 1985, pp. 174–175 and 1997a, p. 144). I find this highly counterintuitive (see below), and I take Audi’s thesis that no one who is self-deceived with respect to a false proposition, p, actually believes that p to be the most basic point on which we disagree. Elsewhere, I write: ‘‘Stock examples of self-deception, both in popular thought and in the literature, feature people who falsely believe – in the face of strong evidence to the contrary – that their spouses are not having affairs, or that their children are not using illicit drugs, or that they themselves are not seriously ill” (Mele, 2001, p. 9). If Audi is right, the cases at issue, as I describe them, are not actually cases of self-deception. In his view, because these people have the false beliefs I mentioned, they are deluded and not self-deceived (In Mele, 1982, I examine arguments presented in Audi, 1982 for this thesis; interested readers may consult both articles.). Consider the following four cases, all of which I offer elsewhere as examples of ways in which desiring that p can contribute to believing that p in instances of self-deception:1 CASE 1. Don just received a rejection notice on a journal submission. He hopes that his article was wrongly rejected, and he reads through the comments offered. Don decides that the referees misunderstood a certain crucial, complex point and that their objections consequently do not justify the rejection. He believes that the paper ought to have been accepted. In fact, however, the referees’ criticisms are warranted; and a few days later, when Don rereads his paper and the comments in a more impartial frame of mind, it is clear to him that this is so. CASE 2. Sid is very fond of Roz, a college classmate with whom he often studies. Wanting it to be true that Roz loves him, he interprets her refusing to date him and her reminding him that she has a steady boyfriend as an effort on her part to ‘‘play hard to get” in order to encourage Sid to continue to pursue her and prove that his love for her approximates hers for him. As Sid interprets Roz’s behavior, not only does it fail to count against the hypothesis that she loves him, it is evidence for the truth of that hypothesis. This contributes to his believing, falsely, that Roz loves him. CASE 3. Beth’s father died a short time ago, not long after her twelfth birthday. Owing partly to her desire that she was her father’s favorite, she finds it comforting to attend to memories and photographs that place her in the spotlight of her father’s affection and unpleasant to attend to memories and photographs that place a sibling in that spotlight. Accordingly, she focuses her attention on the former and is inattentive to the latter. This contributes to Beth’s coming to believe – falsely – that she was her father’s favorite child. In fact, Beth’s father much preferred the company of her brothers, a fact that the family photo albums amply substantiate. CASE 4. Betty, a political campaign staffer who thinks the world of her candidate, has heard rumors from the opposition that he is sexist, but she hopes he is not. That hope motivates her to scour his voting record for evidence of political correctness on gender issues and to consult people in her own campaign office about his personal behavior. Betty misses rather obvious and weighty evidence that her boss is sexist – which he in fact is – even though she succeeds in finding less obvious and less weighty evidence for her favored view. As a result, she comes to believe that her boss is not sexist. I do not offer these cases as the best or clearest examples of self-deception featuring a false belief that p. Their purpose is to illustrate how desiring that p can contribute to believing that p in cases of self-deception. But, as I see it, each of these four cases is a case of self-deception. And on Audi’s view, none of these cases is a case of self-deception: more precisely, Don is not is not self-deceived about whether his paper was wrongly rejected, Sid is not self-deceived about whether Roz loves him, and so on. Both Audi and I reject the idea that self-deceived people believe that p is true while also believing that p is false. And each of us might try to explain why the other is confused about what the self-deceived person does and does not believe. Each of us might think that the other has certain theoretical commitments that lead him astray or a flawed orienting model. Other philosophers have weighed in on our views of self-deception, and we have replied. But I know of no published work on what nonspecialists think about cases such as the ones above. Audi says that he has tried ‘‘to do justice to self-deception as pre-theoretically understood” (2007, p. 252). So I decided to ‘‘poll the folk,” as experimental philosophers say. I conducted two studies. In study 1, participants were 89 students in a basic philosophy class at Florida State University. Self-deception was not on the course agenda. Participants were given the following written instructions: ‘‘We’re interested in how you understand the expression ‘self-deception.’ Please read the four cases below and circle your answers to the questions about them. As you’ll see, 1 is a strong yes and 7 is a strong no.” Participants read cases 1 through 4 above. After each case, they were presented with the following text: Is this an example of self-deception? Yes, definitely 1 2 3 4 5 6 7 No, definitely not The cases appeared in four different orders: ABDC, BCAD, CDBA, DACB, where A is the story about Don, B is the story about Sid, and so on. The results were as follows: 1 Cases 1 and 2 derive from Mele, 1983, p. 369. Cases 3 and 4 derive from Mele, 2001, pp. 26–27. A. Mele / Consciousness and Cognition 19 (2010) 745–750 CASES MEANS Don Sid Beth Betty 4.38 2.35 2.52 2.88 747 These results provide evidence that ‘‘self-deception as pre-theoretically understood” (Audi, 2007, p. 252) is not restricted to cases in which people lack the pertinent false belief. Given Audi’s concern ‘‘to do justice to self-deception as pre-theoretically understood” (2007, p. 252), these results should worry him.2 Some readers may wonder – as I did – why the mean response to Don’s story differed markedly from the mean responses to the other stories. In the end (that is, ‘‘a few days later”), Don sees the truth. My hunch was that this fact accounts for at least most of the difference: many readers may have judged that Don’s story is not an example of self-deception because, in the end, he is not self-deceived. In study 2, I tested this hypothesis with the following story: CASE 1a. Don just received a rejection notice on a journal submission. He hopes that his article was wrongly rejected, and he reads through the comments offered. Don decides that the referees misunderstood a certain crucial, complex point and that their objections consequently do not justify the rejection. He believes that the paper ought to have been accepted. In fact, however, the referees’ criticisms are warranted and Don should have known that his paper deserved to be rejected. Participants were 40 students in another basic philosophy class at Florida State University. Self-deception was not on the course agenda. Participants were presented with case 1a alone, and the instructions varied accordingly from those in study 1. The mean response was 3.07. Counting answers of 1, 2, and 3 as signifying agreement and answers of 5, 6, and 7 as signifying disagreement, 27 participants agreed that case 1a is an example of self-deception and 7 disagreed. A majority of these participants, unlike Audi, evidently do not restrict cases of self-deception to cases in which people lack the pertinent false belief. If I were to be persuaded that I am wrong about the meaning of ‘‘self-deception,” that the great majority of the respondents in my studies misapply the concept of self-deception to the cases they are asked about, and that Audi is right, I doubt that I would be terribly disappointed. In my work on self-deception (or what I think of as self-deception), my primary concern is to explain how ordinary instances of it happen. I have suggested that the following conditions are jointly sufficient for entering self-deception in acquiring a belief that p. 1. The belief that p which S acquires is false. 2. S treats data relevant, or at least seemingly relevant, to the truth value of p in a motivationally biased way. 3. This biased treatment is a nondeviant cause of S’s acquiring the belief that p. 4. The body of data possessed by S at the time provides greater warrant for p than for p. (Mele, 2001, pp. 50–51).3 Drawing on empirical work (in Mele, 2001 and elsewhere), I have articulated and defended a theory about how it happens that people satisfy this collection of conditions (and, of course, I have offered guidance on interpreting conditions 2 and 3). If Audi is right, a person’s satisfying the first of these conditions is incompatible with his being in self-deception with respect to p; again, he holds that the self-deceived person does not actually believe the false proposition with respect to which he is self-deceived. On Audi’s view, a person who satisfies my four conditions may be deluded, but he is not in self-deception with respect to p. If I were to be persuaded that he is right, I might just say that I have learned that, really, what I had been talking about all along was entering self-delusion in acquiring a belief that p; and I might leave it at that. But, of course, partly in light of the data I discussed, I remain unpersuaded that Audi is right about the meaning of ‘‘self-deception.” Indeed, I am persuaded that he is wrong about it. 3. History and more Audi says that the apparent difference between our respective orienting models is related to our disagreement about whether self-deception is an essentially historical phenomenon (2007, p. 254). I believe that he is right about this. After providing a bit more background, I will explain why. I have already reported my proposed set of sufficient conditions for entering self-deception in acquiring a belief that p. As I observe elsewhere (Mele, 1987, p. 131 and 2001, pp. 56–57), people can also enter self-deception in retaining a belief. Here is an example from Mele, 1987 (pp. 131–132). Sam has believed for many years that his wife, Sally, would never have an 2 A referee expressed the worry that study 1 is not a proper test because it includes no control (or contrast) group. However, the point of the study is to gather evidence about whether the claim that being self-deceived with respect to a false proposition, p, requires not actually believing that p – that alleged necessary condition on self-deception – does ‘‘justice to self-deception as pre-theoretically understood” (Audi, 2007, p. 252). Control and contrast groups are not needed for this purpose. Incidentally, in this article I am not concerned to argue that self-deception entails false belief. It is the claim of Audi’s on which I am focusing that concerns me. 3 Obviously, this set of alleged sufficient conditions does not commit me to the view that self-deception requires false beliefs. 748 A. Mele / Consciousness and Cognition 19 (2010) 745–750 affair. In the past, his evidence for this proposition was very good. Sally obviously adored him, she never displayed a sexual interest in another man, she condemned extramarital sexual activity, she was secure, she was happy with her family life, and so on. However, things recently began to change. Sally is now arriving home late from work on the average of two nights a week, she frequently finds excuses to leave the house alone after dinner and on weekends, and Sam has been informed by a close friend that Sally has been seen in the company of a certain Mr. Jones at a theater and a local lounge. Nevertheless, Sam continues to believe that Sally would never have an affair. But he is wrong. Her relationship with Jones is by no means platonic. On the model of my proposed set of sufficient conditions for entering self-deception in acquiring a belief, I propose that the following conditions are conceptually sufficient for entering self-deception in retaining a belief that p: 1. The belief that p which S retains is false. 2. S treats data relevant, or at least seemingly relevant, to the truth value of p in a motivationally biased way. 3. This biased treatment is a nondeviant cause of S’s retaining the belief that p. 4. The body of data possessed by S at the time provides greater warrant for p than for p. Whenever one enters self-deception in acquiring or retaining a belief that p, there is a transition from one’s not being selfdeceived in believing that p to one’s being self-deceived in believing that p. In my view, an essential feature of the transition is that it includes one’s deceiving oneself, a process that takes time. This is part of what Audi refers to as my ‘‘orienting model.” (Readers who believe that deceiving is always intentional will be misled by what I have just said. I believe that much deceiving is not intentional, and I argue that self-deceivers typically do not intentionally deceive themselves; see Mele, 2001, chapter 1.) Furthermore, as I see it, one is not in self-deception unless one entered self-deception. If this is right and if I am right about what entering self-deception requires, then being in self-deception has an essential historical feature: one is not in self-deception unless one deceived oneself, and deceiving oneself takes time.4 Some things have essential historical features. For example, a burn is not a sunburn unless it was caused by exposure to the sun, and a piece of paper is not a genuine US dollar bill if it was not produced by the US Treasury Department. I see a person’s being in a state of self-deception as similar to a patch of skin’s being is a sunburned state. Audi asks (2007, p. 254): ‘‘Must we deny that if I am perfectly duplicated, then, if I am now in self-deception, a full, purely psychological description (apart from listing self-deception) would ground (or even entail) my being in self-deception?” Well, must we deny that if I am perfectly duplicated, then, if I now have a sunburn on my back, my duplicate has a sunburn on his back too, even though his burn was produced by a heat lamp? My answer to the second question is yes, we must deny it. My answer to the first question is the same (assuming that the purely psychological description includes nothing historical), but I am open to being talked out of it. As Audi observes, he has ‘‘argued that ‘self-deception with respect to p is a state in which S [as deceiver] unconsciously knows (or has some reason to believe, and unconsciously and truly believes) that [p], [yet, as deceived] sincerely avows or is disposed so to avow, that p, and has at least one want which in part explains why the belief that p is unconscious’” (2007, p. 252; the brackets are Audi’s and the embedded quotation is from Audi, 1982). This is a complex and distinctive state, and I can see why someone who thinks of self-deception along these lines would think that nothing more – including any fact about how the person came to be in this state – is required for being in self-deception with respect to p (at least for many p’s) than being in this state. In my view, the purely here-and-now state that a self-deceived person is required to be in is simpler. I do not require the unconscious true belief, and so, of course, I also do not require the presence of a want that partly explains why the belief is unconscious. Perhaps it is partly because my view is simpler on this person-internal front that I am inclined to see something historical as essential to being ‘‘in self-deception.” The expression ‘‘in self-deception” is not one for which I have much use. As Audi mentions (2007, p. 252), my focus has been on entering self-deception. However, I am happy to say that someone who has entered self-deception in acquiring or retaining a belief that p is in self-deception with respect to p, as long as the person continues to be self-deceived in believing that p. One thing that satisfaction of this last clause requires, of course, is that the person continues to believe that p. But that is not sufficient for satisfying the clause. The person’s evidence may have come to be such that he is now warranted in believing that p. In some such case, it may be true that although the person was self-deceived in believing that p and still believes that p (a false proposition), he is no longer self-deceived in believing it (and is no longer in self-deception with respect to p). Audi writes: ‘‘Self-caused deception need not be self-deception; and even an analysis of self-caused deception may not suffice for an adequate account of self-deception” (2007, p. 253). I agree entirely. As I see it, causing oneself to be deceived is a necessary condition for self-deception, not a sufficient condition. And, as I point out in Mele, 1987, ‘‘if Jones, due to unmotivated carelessness, misreads a clearly printed word in a news article, thereby causing himself to have a false belief (and, hence, to be deceived), neither the deceiving of himself, nor the condition of deception which he has produced in himself, is the sort of thing which we ordinarily call self-deception” (pp. 124–125). This is an example of self-caused deception without self-deception. Audi contends that his model ‘‘does justice to the dissociation self-deception apparently entails and places the element of deception in the tendency to make sincere avowals of propositions one unconsciously knows to be false yet – in misleading 4 As I understand deceiving oneself, it entails causing oneself to have a false belief. (Whether most lay folk understand deceiving oneself as I do is testable.) A. Mele / Consciousness and Cognition 19 (2010) 745–750 749 but natural terms – ‘consciously believes’ to be true. By virtue of the dissociation, the sincere avowal, which would normally imply belief, manifests a kind of deception” (2007, p. 253). In a related vein, some philosophers, including Audi (1997b), have argued that my proposed sufficient conditions for entering self-deception in acquiring a belief that p are not sufficient because they do not capture a kind of tension that is necessary for self-deception (see Mele, 2001, pp. 52–56 for references and discussion). The quartet of conditions at issue certainly does not entail that there is no tension in self-deception. Nor do I claim that self-deception normally is tension-free. Satisfying my four conditions may often involve considerable psychic tension. The present question is whether tension or dissociation is conceptually necessary for entering self-deception in acquiring a belief that p. My answer is no. Given the details of Sid’s case, for example, even if he is free of psychic conflict during the process of acquiring the belief that Roz loves him and while that belief is in place, he is self-deceived and he enters selfdeception in acquiring that belief. In Audi’s view, again, ‘‘self-deception with respect to p is a state in which”: (1) ‘‘S [as deceiver] unconsciously knows (or has some reason to believe, and unconsciously and truly believes) that [p],” (2) ‘‘[yet, as deceived] sincerely avows or is disposed so to avow, that p,” (3) ‘‘has at least one want which in part explains why the belief that p is unconscious” (2007, p. 252), and (4) does not believe that p. I believe that self-deception is pretty common (see Mele, 2001). How common are cases that satisfy Audi’s four conditions? Suppose we were scientists trying to answer this question. How would we proceed? One thing we would want to be able to do is to figure out a way to tell whether someone who sincerely avows that p also believes that p. Because this would be very difficult, it would be very difficult to be confident, regarding any apparently ordinary instance of self-deception, that it satisfies both condition 2 and condition 4. When I add to that the obstacles to confidence about what alleged self-deceivers unconsciously know or believe, I find that I am reluctant even to guess about the frequency of cases satisfying Audi’s four conditions. One might propose that we arrive at the postulation of the satisfaction of Audi’s four conditions in cases of apparent selfdeception by an inference to the best explanation. We look at what the person does and says and judge that the best explanation of his deeds and words is that he satisfies Audi’s four conditions. Here is an alternative explanatory hypothesis about tension-involving cases of the kind on which Audi focuses: what the person avows – namely, p – he believes; and rather than unconsciously knowing or believing that p, he consciously believes that there is a significant chance that p. Readers are encouraged to make up their own minds about which hypothesis is more plausible. For a defense of the hypothesis I just reported, see Mele, 2001, pp. 67–73. Someone might claim that, owing to the nature of deceiving, self-deception requires that one know (or at least believe) the truth. Recall, in this connection, Audi’s report that his state model of self-deception is meant to apply ‘‘to a person who is both deceived and deceiver and hence, apart from inconsistent beliefs, must see the truth” (2007, p. 253). Obviously, it is not the fact that the person is deceived that is meant to motivate the claim that he ‘‘must see the truth”; the fact that he is a ‘‘deceiver” is meant to do this. Consider the following two propositions: 1. By definition, person A deceives person B (where B may or may not be the same person as A) into believing that p only if A knows, or at least believes truly, that p and causes B to believe that p. 2. By definition, deceiving is an intentional activity: nonintentional deceiving is conceptually impossible. In various places, including chapter 1 of Mele, 2001, I argue that both propositions are false and that acceptance of either of them easily leads to confusion about self-deception. I cannot repeat the arguments here, but I will say something about this issue (drawing on Mele, 2001, pp. 8–9). In a standard use of ‘deceived’ in the passive voice, we properly say such things as ‘‘Unless I am deceived, I left your article on my desk.” Here ‘deceived’ means ‘mistaken’. There is a corresponding use of ‘deceive’ in the active voice. In this use, to deceive is ‘‘to cause to believe what is false” (as the Oxford English Dictionary reports). Plainly, one can intentionally or unintentionally cause someone to believe what is false; and one can cause someone to acquire the false belief that p even though one does not oneself believe that p. Yesterday, mistakenly believing that my daughter’s car keys were on my dresser, I told her they were there. In doing so, I caused her to believe a falsehood. I deceived her, in the sense identified; but I did not do so intentionally, nor did I cause her to believe something I disbelieved. Stereotypical instances of deceiving someone else into believing that p are instances of intentional deceiving in which the deceiver knows or believes truly that p. The point made in the preceding paragraph has little significance for self-deception, if paradigmatic instances of it have the structure of stereotypical instances of interpersonal deception. However, like Audi, I argue that they do not. Of course, we differ in how we depart from a model of self-deception that is straightforwardly based on stereotypical interpersonal deception. Audi observes that he has a state model of self-deception whereas I have an act model (2007, p. 253); but his model seems to be influenced by an assumption about acts of deceiving that I reject. At any rate, one important difference between us is that he requires the self-deceived person to know or believe the pertinent true proposition, and I do not. This difference might be partly explained by a difference in how we understand the kind of deceiving that is most relevant to self-deception. I close with a comment on the survey studies I reported. Although some philosophers find experimental philosophy irritating, I believe that – whether it is irritating or not – its techniques can be useful in some philosophical debates. Here my studies were put to a very limited use. As I reported, Audi says that he has tried ‘‘to do justice to self-deception as pre-theoretically understood” (2007, p. 252). Now, I know too much about motivationally biased belief (see Mele, 2001) to be 750 A. Mele / Consciousness and Cognition 19 (2010) 745–750 confident about my own untested hunches regarding how self-deception is pre-theoretically understood. And just as my hunches about this may be biased partly by my theory, the same is true of Audi. What I tested was the hypothesis that self-deception, as pre-theoretically understood, is such that no one who is self-deceived regarding a false proposition p actually believes that p. And the results definitely do not favor the hypothesis. Although, as I have noted, I am inclined to believe that every case of self-deception involves deceiving and that the kind of deceiving at issue entails causing a false belief, this inclination of mine was not examined in this article and I have not tested the hypothesis that my inclination is in line with majority judgments of lay folk about cases. Others are free to conduct such tests. The results may prove interesting. References Audi, R. (1982). Self-deception, action, and will. Erkenntnis, 18, 133–158. Audi, R. (1985). Self deception and rationality. In M. Martin (Ed.), Self deception and self understanding (pp. 169–194). Lawrence: University of Kansas Press. Audi, R. (1997a). Self-deception, rationalization, and the ethics of belief. In R. Audi (Ed.), Moral knowledge and ethical character (pp. 131–156). New York: Oxford University Press. Audi, R. (1997b). Self-deception vs. self-caused deception: A comment on Professor Mele. Behavioral and Brain Sciences, 20, 104. Audi, R. (2007). Belief, intention, and reasons for action. In M. Timmons, J. Greco, & A. Mele (Eds.), Rationality and the good (pp. 248–259). New York: Oxford University Press. Bach, K. (1981). An analysis of self-deception. Philosophy and Phenomenological Research, 41, 351–370. Funkhouser, E. (2005). Do the self-deceived get what they want? Pacific Philosophical Quarterly, 86, 295–312. Gendler, T. (2007). Self-deception as pretense. Philosophical Perspectives, 21, 231–258. Mele, A. (1982). Self-deception, action, and will: Comments. Erkenntnis, 18, 159–164. Mele, A. (1983). Self-deception. Philosophical Quarterly, 33, 365–377. Mele, A. (1987). Irrationality: An essay on akrasia, self-deception, and self-control. New York: Oxford University Press. Mele, A. (2001). Self-deception unmasked. Princeton: Princeton University Press. Mele, A. (2007). Self-deception and three psychiatric delusions: On Robert Audi’s transition from self-deception to delusion. In M. Timmons, J. Greco, & A. Mele (Eds.), Rationality and the good (pp. 163–175). New York: Oxford University Press. Rey, G. (1988). Toward a computational account of Akrasia and self-deception. In A. Rorty & B. McLaughlin (Eds.), Perspectives on self-deception (pp. 264–296). Berkeley: University of California Press.

Consciousness and Cognition 19 (2010) 1107–1109 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Reply Beyond inattentional blindness and attentional misdirection: From attentional paradigms to attentional mechanisms q Daniel Memmert ⇑, Philip Furley German Sport University Cologne, Institute of Cognitive and Team/Racket Sport Research, Am Sportpark Müngersdorf 6, 50933 Köln, Germany a r t i c l e i n f o Article history: Received 7 May 2010 Available online 29 June 2010 Keywords: Overt attention Covert attention Working memory Perceptual load Representative task designs Sport a b s t r a c t Memmert (2010) tried to foster the development of attentional research by discussing four differences between attentional misdirection (AM) and inattentional blindness (IB). Considering this goal, the comment was received in the intended way by the comments of Most (2010) and Moran and Brady (2010) who make a number of highly valuable suggestions for further progress. As initially suggested by Memmert (2010) this dialog should help unravel the underlying attentional mechanisms of different paradigms. Therefore, we first discuss the suggested distinction between central and spatial IB by Most (2010). Second, we argue that working memory and perceptual load research seem particularly interesting in this regard and should be taken into consideration when conducting future research along the lines of IB and AM. Third, representative task designs can be an important mosaic piece in across-the-board attention theories and highly useful for deriving further testable hypothesis in naturalistic settings. The most important claim of all commentaries in this issue is that the proposed ideas can all be empirically tested and thereby contribute to the advancement of an unified theoretical framework incorporating IB, AM in consideration of overt and covert attention mechanisms. Ó 2010 Elsevier Inc. All rights reserved. 1. Introduction In a series of studies Kuhn and colleagues (e.g., Kuhn & Findlay, in press) described a novel attentional misdirection (AM) approach to investigate overt and covert attention mechanisms in connection with inattentional blindness (IB). Memmert (2010) tried to foster the development of the link between both approaches by discussing four differences between AM and IB which concern the conceptual aspects of the unexpected object and methodological aspects of the task design. The central point of the commentary of Memmert (2010) is that extreme caution is required when comparing theoretical discussions and empirical evidence from both paradigms. The main aim behind this effort was to initiate further theory development in the area of attentional research. Considering this goal, the comment was received in the intended way by the first excellent comments by Most (2010) and Moran and Brady (2010) who make a number of highly valuable suggestions for further progress. As initially suggested by Memmert (2010) this dialog should help unravel the underlying attentional mechanisms of the different paradigms. Therefore, several preliminary suggestions for future research lines incorporating IB and AM are put forth in order to study selective attention processes in more depth. Most (2010) and Moran and Brady (2010) q Reply to Commentaries on Most, S. B. (2010). What’s inattentional about inattentional blindness? Consciousness and Cognition, 19, 1102–1104; Moran, A., & Brady, N. (2010). Mind the gap: Misdirection, inattentional blindness and the relationship between overt and covert attention. Consciousness and Cognition, 19, 1105–1106. ⇑ Corresponding author. Fax: +49 221 4995 637. E-mail address: memmert@dshs-koeln.de (D. Memmert). 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.06.005 1108 D. Memmert, P. Furley / Consciousness and Cognition 19 (2010) 1107–1109 elaborate on this by stating further interesting ideas and advancements of the initial ideas that are intended to bridge the gap between AM and IB. Subsequently we want to comment on the following topics to steadily promote this process. 2. Sub-types of IB The differentiation of (at least) two sub-types of IB by Most (2010) which he assumes to be driven by distinct attentional mechanisms – (a) covert allocation of spatial attention; (b) distraction or preoccupation of non-spatial selection mechanisms stemming from late-stage bottlenecks – seems highly valuable, not only for providing a basis of convergence of various attention paradigms (e.g., object substitution masking; Enns & Di Lollo, 1997), but also allowing a substantial discussion of older and more current attentional frameworks. Thereby, one needs to investigate if these sub-types can be empirically separated, how many sub-types of IB actually exist and respectively integrate these in overarching attentional frameworks while considering both covert and overt attentional processes as initiated by Most (2010). Another interesting question arising from the possible distinction of central and spatial IB is whether there might be dichotomous or gradual boundaries. Most’s (2010) starting point for the distinction between central and spatial IB is the comparison of results found in Mack and Rock’s (1998) static studies and results found in Most, Simons, Scholl, and Chabris (2000) dynamic monitoring task. Surprisingly Mack and Rock (1998, p. 56 and p. 69) find contrary results regarding possible central and spatial IB within their static line judging paradigm. In different conditions of their experiments a black circle (=unexpected object) appeared either parafoveally with the primary attention demanding task at fixation or at fixation with the attention demanding task in the periphery. Interestingly only 25% of the participants are inattentionally blind in the parafoveal condition (in contrast to 73% in the far condition by Most et al., 2000) whereas 89% fail to notice the unexpected object at fixation (47% in the on-line condition by Most et al., 2000). This finding is somewhat counterintuitive to Most et al. (2000) and seemingly does not support the spatial vs. central IB distinction, which is probably due to the fact that the two paradigms differ in major aspects. Therefore, these findings require further research and show that extreme caution is required when comparing results from different IB tasks. Thus, it is necessary to investigate a possible distinction within a single study. Two bodies of literature that are addressed in the following chapter seem particularly interesting in this regard and should be taken into consideration when conducting future research along the lines of IB and AM. 3. Working memory and perceptual load A recent study by Fougnie and Marois (2007) linked working memory (WM) load to IB by showing that executive processing can result in IB. Referring to results by Most et al. (2001), Most, Scholl, Clifford, and Simons (2005) they conclude that both visuospatial and executive information processing can result in IB and thus support Most’s (2010) distinction of central and spatial IB. When reading the distinction between spatial and central IB the long-standing debate of early (e.g., Broadbent, 1958; Neisser & Becklen, 1975) versus late selection (e.g., Deutsch & Deutsch, 1963; Tipper, 1985) views comes to mind for which Lavie (1995) put forth a possible resolution within her elegant perceptual load (PL) model. According to her model, processing proceeds from relevant to the irrelevant items until capacity runs out. Under conditions of low perceptual load spare capacity inevitably spills over and irrelevant information is processed, whereas irrelevant processing can be prevented when a high load in relevant processing exhausts capacity. Cartwright-Finch and Lavie (2007) could find evidence for the load theory using a modification of Mack and Rock’s (1998) IB task. Both the studies by Fougnie and Marois (2007) and Cartwright-Finch and Lavie (2007) demonstrate that attentional theory development benefit from integrating different paradigms with one another. Nevertheless, Fougnie and Marois (2007) explicitly state that extreme caution has to be warranted when comparing results from different paradigms as their pattern of results appears to be inconsistent with findings from the perceptual load paradigm, which they state is due to the slight difference that in one paradigm the distractor is expected and in the other one not. Memmert (2010) makes a similar claim when comparing IB with AM. Thus, our main argument is that theory development can substantially benefit from the integration of different paradigms within a single study but one needs to be cautious when comparing results from different experimental settings. This exemplary discussion emphasizes how the integration of the IB, AM, WM, and PL paradigms can be important mosaic pieces in across-the-board attention theories and highly useful for deriving further testable hypothesis. 4. Representative task designs We welcome that Moran and Brady (2010) encourage our view (Memmert, 2010) that ‘‘recent research on inattentional blindness and attentional misdirection has shown that these paradigms have great potential for the development of more differentiated attention frameworks leading to new testable assumptions of attention mechanisms’’. As Moran and Brady state the field of sports seems a fruitful area for studying complex human behavior in a complex context by providing test procedures with context specific performance data (representative task designs: Brunswik, 1956; see also, Dicks, Davids, & Button, 2009). In the meantime new data (Furley, Memmert, & Heller, 2010) could address one consideration of Moran and Brady (2010) by showing that level of expertise affected the occurrence of IB in a representative basketball task amongst adults (for expertise effects in the general IB paradigm; see also Memmert, 2006). In combination with the further proposed lines of research by Moran and Brady (2010) we think that future research can benefit from incorporating individual differ- D. Memmert, P. Furley / Consciousness and Cognition 19 (2010) 1107–1109 1109 ence variables (such as working memory capacity, attentional capacity) in representative task designs in order to foster the illumination of underlying covert and overt attentional mechanisms. 5. Outlook The most important and redundant claim of all commentaries in this issue is that the proposed ideas can all be empirically tested and thereby contribute to the advancement of an unified theoretical framework incorporating IB, AM in consideration of overt and covert attention mechanisms. All of the commentaries encourage the integration of currently isolated attention paradigms and theories. The vivid discourse in this issue appears to be a promising attempt in doing so. Acknowledgments Special thanks go to Matt Dicks for many inspirations and comments on earlier versions of this commentary. References Broadbent, D. E. (1958). Perception and communication. London: Pergamon Press. Brunswik, E. (1956). Perception and the representative design of psychological experiments (2nd ed.). Berkeley: University of California Press. Cartwright-Finch, U., & Lavie, N. (2007). The role of perceptual load in inattentional blindness. Cognition, 102, 321–340. Deutsch, J., & Deutsch, D. (1963). Attention: Some theoretical considerations. Psychological Review, 70, 80–90. Dicks, M., Davids, K., & Button, C. (2009). Representative task designs for the study of perception and action in sport. International Journal of Sport Psychology, 40, 506–524. Enns, J. T., & Di Lollo, V. (1997). Object substitution: A new form of visual masking in unattended visual locations. Psychological Science, 8, 135–139. Fougnie, D., & Marois, R. (2007). Executive working memory load induces inattentional blindness. Psychonomic Bulletin & Review, 14, 142–147. Furley, P., Memmert, D., & Heller, C. (2010). The dark side of visual awareness in sport – Inattentional blindness in a real-world basketball task. Attention, Perception & Psychophysics, 72, 1327–1337. Kuhn, G., & Findlay, J. (in press). Misdirection, attention and awareness. Inattentional blindness reveals temporal relationship between eye movements and visual awareness. Quarterly Journal of Experimental Psychology. Lavie, N. (1995). Perceptual load as a necessary condition for selective attention. Journal of Experimental Psychology—Human Perception and Performance, 21, 451–468. Mack, A., & Rock, I. (1998). Inattentional blindness. Cambridge, MA: The MIT Press. Memmert, D. (2006). The effects of eye movements, age, and expertise on inattentional blindness. Consciousness and Cognition, 15, 620–627. Memmert, D. (2010). The gap between inattentional blindness and attentional misdirection. Consciousness and Cognition, 97, 1097–1101. Moran, A., & Brady, N. (2010). Mind the gap: Misdirection, inattentional blindness and the relationship between overt and covert attention. Consciousness and Cognition, 97, 1105–1106. Most, S. B., Simons, D. J., Scholl, B. J., & Chabris, C. F. (2000). Sustained inattentional blindness: The role of location in the detection of unexpected dynamic events. Psyche, 6(14). <http://psyche.cs.monash.edu.au/v6/psyche-6-14-most.html>. Most, S. B., Scholl, B. J., Clifford, E. R., & Simons, D. J. (2005). What you see is what you set: Sustained inattentional blindness and the capture of awareness. Psychological Review, 112, 217–242. Most, S. B., Simons, D. J., Scholl, B. J., Jimenez, R., Clifford, E., & Chabris, C. F. (2001). How not to be seen: The contribution of similarity and selective ignoring to sustained inattentional blindness. Psychological Science, 12, 9–17. Most, S. B. (2010). What’s inattentional about inattentional blindness? Consciousness and Cognition, 97, 1102–1104. Neisser, U., & Becklen, R. (1975). Selective looking: Attending to visually specified events. Cognitive Psychology, 7, 480–494. Tipper, S. P. (1985). The negative priming effect: Inhibitory effects of ignored primes. The Quarterly Journal of Experimental Psychology, 37A, 571–590.

Consciousness and Cognition 19 (2010) 1097–1101 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog The gap between inattentional blindness and attentional misdirection Daniel Memmert * German Sport University Cologne, Institute of Cognitive and Team/Racket Sport Research, Am Sportpark Müngersdorf 6, 50933 Köln, Germany a r t i c l e i n f o Article history: Received 2 June 2009 Available online 1 February 2010 Keywords: Overt attention Covert attention Functionality Awareness a b s t r a c t Kuhn and colleagues described a novel attentional misdirection approach (deliberate diversion of attention away from a visually salient stimulus) to investigate overt and covert attention mechanisms in connection with inattentional blindness (not being able to perceive something that is plainly visible because one’s attention has not been focused on it). This misdirection paradigm is valuable to study the temporal relationship between eye movements and visual awareness. Although, as put forth in this comment, the link between attentional misdirection and inattentional blindness needs to be developed further. There are at least four differences between the two paradigms which concern the conceptual aspects of the unexpected object and the methodological aspects of the task design. This highlights the need for a broader theoretical framework incorporating inattentional blindness and overt and covert attention mechanisms. Two possible research lines focusing on the orienting attention research and the ‘‘selection-for-action” paradigm are discussed. Ó 2010 Elsevier Inc. All rights reserved. 1. Introduction Kuhn and colleagues examined in a series of interesting experiments how misdirection can prevent observers from perceiving an unexpected salient object (Kuhn, Amlani, & Resnsink, 2008; Kuhn & Tatler, 2005; Kuhn, Tatler, Findlay, & Cole, 2008). They described a novel experimental method to investigate overt and covert attention mechanisms in an ecologically valid situation by showing observers a magic trick in which the magician made an obviously visually salient lighter and a cigarette disappear by manipulating the attentional focus of the observers (e.g., moving the arm/head in the opposite direction).1 For example, in one of the standard video clips by Kuhn and colleagues, a magician makes both a lighter and a cigarette disappear by intentionally misdirecting the attentional focus of the observer away from his hand (besides covering the objects with this hand) from which he drops both of the objects into his lap. In this way the magician draws attention away from the hand that is relevant for the trick towards the hand that is not relevant. With that misdirection paradigm Kuhn and colleagues can investigate the link between fixation and visual awareness. Kuhn and colleagues stated in a number of publications that the misdirection trick is analogous to inattentional blindness (Kuhn, Amlani, et al., 2008; Kuhn & Tatler, 2005; Kuhn, Tatler, et al., 2008). For example: ‘‘The misdirection employed by magicians parallels inattentional-blindness paradigms” (Kuhn & Tatler, 2005, p. 1156), ‘‘our chosen magic trick employs similar principles to those used in recent inattentional-blindness studies” (Kuhn & Tatler, 2005, p. 1156), or ‘‘this misdirection trick is analogous to inattentional blindness” (Kuhn & Findlay, in press). In this comment we will argue that the link between the attentional misdirection paradigm and the inattentional blindness paradigm needs to be developed further. * Fax: +49 221 4995 637. E-mail address: memmert@dshs-koeln.de 1 Overt attention processes imply that visible shifts of attention are taking place, which is generally monitored by eye tracking. Covert attention processes aim at capturing non-visible (cannot measured by eye tracking) allocations of attention (for a discussion, see Findlay, 2004; Henderson, 2003). 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.01.001 1098 D. Memmert / Consciousness and Cognition 19 (2010) 1097–1101 For this purpose a definition of the ‘‘inattentional blindness” paradigm is needed: ‘‘When attention is diverted to another object or task, observers often fail to perceive an unexpected object, even if it appears at fixation – a phenomenon termed inattentional blindness” (Mack & Rock, 1998, p. 14). For example, in the famous video clip by Simons and Chabris (1999), the observers had to watch a basketball game involving six players, three wearing a white shirt and three wearing a black shirt. They were instructed to watch only the players in black and to count the number of passes and bounces made by the attended team. Because of this attention-demanding counting-task, many observers failed to perceive the unexpected gorilla walking straight through the scene. Without this attention-demanding counting-task, nearly all participants noticed the unexpected object in the same video clip. The major finding is that conscious perception seems to require attention (Becklen & Cervone, 1983; Mack & Rock, 1998; Neisser & Becklen, 1975, see also Most, Scholl, Clifford, & Simons, 2005). There are at least four differences between the attentional misdirection and inattentional blindness paradigm. These differences concern conceptual and methodological aspects that show that extreme caution is required when comparing theoretical discussions and empirical evidence. To state it clearly here, the attentional misdirection paradigm is by itself valuable to explore the relationship between overt and covert attention. Nevertheless we will argue why one has to be careful comparing the two paradigms. Four differences between these paradigms will be discussed one by one in the following paragraphs. The first three disconnections raise possibilities suggesting how inattentional blindness might be a different phenomenon than attentional misdirection. The fourth disconnection focuses more on the type of stimulus material used in both paradigms. 2. Four arguments for the disconnection between attentional misdirection and inattentional blindness 2.1. Disconnection 1: definition of unexpected object – foreshadowing vs. no foreshadowing In the attentional misdirection paradigm, observers might anticipate an unexpected event (e.g., drop of the cigarette) because they know they are watching a magician perform a magic trick, whereas they do not expect to perceive unexpected objects in the inattentional blindness paradigm while performing a task. For example, in the aforementioned study by Mack and Rock (1998), the participants were asked to judge which of the two arms of a briefly displayed cross is longer. In the critical trial, nobody anticipated an unexpected object in the form of a geometric shape which appeared at a nearby position for the same duration. The same is true for the participants of Simons and Chabris’ (1999) famous basketball experiments, who did not expect a man in a gorilla suit or a woman with an umbrella moving through the group while they counting the number of passes made between three basketball players. Thus, the conceptualization of the unexpected object in the attentional misdirection and inattentional blindness paradigm differ to a certain degree. This argument concerns the mind-set (=what the participants think that will happen next) of the observers. The significant role of the mind-set in perceiving unexpected objects could be revealed under laboratory conditions (Most et al., 2005) and in realistic scenarios (Most & Astur, 2007). To sum up, the first disconnection concerns the fact that both paradigms differ concerning the foreshadowing or no-foreshadowing of the unexpected object by the participants. This raises questions on the specific overt and covert attention mechanisms present in both paradigms. 2.2. Disconnection 2: control task – no control vs. control The inattentional blindness paradigm includes a full-attention trial in order to ensure that all participants perceive the unexpected object without the primary attention-demanding task. In this full-attention trial the participants were told to simply watch the screen without performing the primary attention-demanding task (e.g., counting the passes in the gorilla video). Observers who failed to report the unexpected object in this trial were rejected from the data analysis. The full-attention trial can be seen as a kind of control task for two reasons: first, it can be used to ensure that the participants followed task instructions, and second, that the participants consciously noticed the unexpected object under full-attention conditions. A full-attention trial does not exist in the misdirection paradigm. Therefore, it remains unclear if the observers perceive the unexpected object under full-attention conditions without misdirection. On the whole, the difference between having a control condition vs. having no control condition creates an important difference between the two paradigms. This distinction has to be taken into consideration if one attempts to compare and discuss results from both paradigms concerning the relationship between fixation and detection. 2.3. Disconnection 3: attentional workload of the task – no distractor vs. distractor In the attentional misdirection paradigm there is no primary attention-demanding task (‘‘distractor task”) in contrast to the inattentional blindness paradigm (e.g., counting passes in the gorilla video). Hence it remains unclear how much conscious overt attention the participants employ on the whole task during the magic trick. This point was recognized by Kuhn and Findlay (in press) as well. ‘‘In the misdirection trick, participants’ attentional resources are constrained through the systematic, but implicit orchestration of attention”. In this context, the attention demanding primary task in the inattentional blindness paradigm does not necessarily lead to a kind of attentional misdirection. For example, a series of studies used a D. Memmert / Consciousness and Cognition 19 (2010) 1097–1101 1099 dynamic monitoring inattentional blindness task by Most, Simons, Scholl and Chabris (2000), in which the participants had to count the total number of times that letters cross a horizontal line in the middle of a display. During a critical trial, an unexpected object moved horizontally over different distances (on-line, near, far, very far) from the line passing through the center of the display. Less than half the observers noticed the unexpected object, even though the object always stayed on what was presumably the focus of attention and was clearly visible for 5 s. In this experimental setting of the inattentional blindness paradigm, the unexpected object appeared in the area of focal attention of the participants. In the attentional misdirection paradigm by Kuhn and colleagues, the unexpected object (e.g., drop of a cigarette) was not directly in the attentional focus of the observers because the magician drew the observers’ attention away from the cigarette. Thus, the drop of the cigarette was indeed unexpected but not according to the definition of inattentional blindness at fixation of the observers (see Mack & Rock, 1998, p. 14). In two experiments, Koivisto, Hyönä, and Revonsuo (2004) and Memmert (2006) demonstrate this point empirically. They show that no time differences in holding fixation on the unexpected object were evident between the subjects who perceived this object and the ones who did not perceive it. For the misdirection task, the critical event is transient (e.g., the drop) rather than sustained, and the critical object (e.g., a cigarette) is absent rather than present. To sum up, the inattentional blindness paradigm includes a distractor in the form of a primary task while there is no distractor task in the misdirection paradigm because no demanding primary task is given. This manipulation allows for experimental control over the amount of attention deployed by the participants in the inattentional blindness paradigm (by varying the difficulty of the primary task) and not in the misdirection paradigm. Thus, the misdirection paradigm seems more like change blindness than inattentional blindness in some ways. Change blindness refers to the failure to perceive something different about a display whereas inattentional blindness refers to a failure to notice something present in a display (e.g., Rensink, 2002). For example, in the change blindness task by Simons, Chabris, Schnur, and Levin (2002), a basketball is removed from a scene and the participants have to notice that unexpected disappearance. 2.4. Disconnection 4: functionality of the unexpected object – relevant vs. not relevant In the inattentional blindness paradigm, the unexpected objects are not important and therefore irrelevant for the task of the participants (for a discussion of attentional capture and inattentional blindness, see Simons, 2000). For example, it is not necessary to notice the man in the gorilla suit in the Simons and Chabris (1999) study while counting accurately in the primary passing task. This is highly significant when considering that various results from inattentional blindness literature (Most et al., 2005) suggest that limited attentional resources—beside the attentional set—are a main factor for missing the irrelevant object (for a similar interpretation in the area of the perceptual load paradigm, see Cartwright-Fincha & Lavie, 2007; Lavie, 1995). In line with this argument is evidence showing that experienced observers in the inattention tasks have a greater possibility of perceiving the unexpected object because of a reduction of the attentional demands of the primary task leaving more attentional resources for the irrelevant object (Memmert, 2006; Neisser & Dube, 1978, cited in Neisser (1979)). However, this argument stands in contrast to the misdirection paradigm that suggests that more attention is paid to the unexpected object when it has a function with regard to the task at hand. This was the case in the Kuhn paradigm where the unexpected object becomes relevant because the participants were asked how the magician performs the trick. On the whole, the usual unexpected stimuli used in the inattentional blindness paradigm and in almost all studies (exceptions: Haines, 1991; Memmert & Furley, 2007; Most & Astur, 2007; Strayer, Drews, & Johnston, 2003), with no functional relevance for the task (e.g., perceiving the man in the gorilla suit), stand in contrast to the unexpected stimuli used in the misdirection paradigm with functional relevance for the task (e.g., drop of a cigarette). 3. Beyond awareness, inattention and misdirection The inattentional blindness paradigm and the attentional misdirection paradigm are valuable on their own to investigate attentional processes. But the four aforementioned disconnections between inattentional blindness and attentional misdirection suggest that these paradigms are not equivalent. Thus, the relevant findings cannot be compared with each other without considerable caution. The recent literature demonstrates the need for a unified theoretical framework incorporating inattentional blindness and overt and covert attention mechanisms. In order to initiate such a framework, insights from one line of literature must be tested in the other. This comment aims to initiate this development. Consequently, two possible avenues of future research emerge from the described disconnections and will be discussed next. We focused on outlining first thoughts for further research avenues in line with disconnections 1 (definition of unexpected object – foreshadowing vs. no foreshadowing) and 4 (functionality of the unexpected object – relevant vs. not relevant). The misdirection trick could be a valuable tool for a first research line which could be connected to disconnections 1. Kuhn and colleagues developed an experimental setting to investigate overt and covert attention mechanisms with a paradigm where the participants pay attention to a trick performed by a magician. With the misdirection approach and data from eye movement behavior they showed that conscious perception is not related to where the participants were looking at the time of the event, and thus demonstrate how overt and covert attention can be spatially dissociated (Kuhn & Findlay, in press). In future work, this result by Kuhn and colleagues could be discussed in closer connection with the selective attention and orienting attention research by Posner (1980). Selective attention is closely linked to attentional orienting, because both sub- 1100 D. Memmert / Consciousness and Cognition 19 (2010) 1097–1101 processes are involved in directing attention (steering) to certain areas (Awh, Armstrong, & Moore, 2006). According to Posner (1980), orienting of attention in the visual field facilitates the processing of the information present in the attended location and inhibits the processing of information present in the unattended location. The cueing paradigm by Posner (1980; see also Posner & Peterson, 1990) is usually used to examine the costs and benefits of orienting attention in the visual field. In general, Posner and Peterson (1990) have shown that performance in signal detection tasks is enhanced by pre-cueing the location where the target stimulus is likely to appear. The head start given to the attentional system leads to better reaction times and response accuracies if the target appears at the location indicated by the precue (valid condition). Performance decreases if the target appears at an uncued location (invalid condition) (Gottlob, Cheal, & Lyon, 1999). The experimental settings used by Posner and colleagues seem valuable in order to establish a closer link between attentional misdirection and the established overt and covert attention mechanism. The inattentional blindness approach could be a valuable tool for a second research line, in which the inattentional blindness paradigm could profit from the kind of unexpected object in the attentional misdirection paradigm (see Section 2.4). Here, the cigarette had a functional meaning in the attention test. Prior studies have already put a stronger emphasis on the importance of functionality in selective attention processes (Neumann, 1987; Neumann, Van der Heijden, & Allport, 1986). Allport (1987) suggests within his principle ‘‘selection-for-action” that perceptual selection is a necessary precondition for any action in the environment. So far, the unexpected object in the inattentional blindness paradigm had no functional relevance for the primary task and therefore was irrelevant for the given task. A future line of research could be to develop real world scenarios of inattentional blindness in which the unexpected object has a functional role in the attention task. We are aware of only a few experimental approaches in the area of inattentional blindness utilizing a ‘‘functional” inattentional blindness task. In the area of flying, Haines (1991) investigated pilots in a flight simulation system. The task for the experienced pilots was to land a plane safely (primary performance task) while monitoring a superimposed head-up display (secondary attention-demanding task). While they were landing the plane virtually, another plane was placed directly on the runway (unexpected object) that the pilots often failed to notice. The airplane on the runway could be interpreted as a functional object, because security was obviously important for the primary task. In the area of driving, Most and Astur (2007) as well as Strayer et al. (2003) provide evidence for inattentional blindness in a driving simulation system. Strayer et al. (2003) demonstrated that participants failed to notice a running child (unexpected object) crossing the street while driving a car (primary performance task) and speaking on the telephone (secondary attention-demanding task). The crossing child could again be interpreted as a functional element because security was clearly important for the primary task. In the area of sport, Memmert and Furley (2007) showed that inattentional blindness exists among skilled athletes who failed to detect a free team mate (unexpected object) when attention was diverted to the direct opponent (primary performance task: name the position of your direct opponent player). If attention was not engaged in the primary performance task (full-attention trial) all participants passed to the obviously unmarked team-member. The open player can be interpreted as a functional object because it was obviously the best solution (controlled for by expert ratings) in the secondary attention-demanding task. Both directions of research are intended to give preliminary suggestions on how future studies could incorporate inattentional blindness and attentional misdirection in order to study selective attention processes in more depth. The research could focus on the targeted intentional or coincidental unintentional misdirection of attention in dealing with a primary task, when new, albeit task-relevant, functional stimuli emerge. These results would also directly lead to practical implications for complex real life scenarios (e.g., security when driving cars). Recent research on inattentional blindness and attentional misdirection has shown that these paradigms’ have great potential for the development of more differentiated attention frameworks leading to new testable assumptions of attention mechanisms. Acknowledgments Special thanks go to Daniel Simons, Philip Furley, and three anonymous reviewers for many inspirations and comments on earlier versions of this manuscript. References Allport, D. A. (1987). Selection-for-action: Some behavioural and neurophysiological considerations of attention an action. In H. Heuer & A. F. Sanders (Eds.), Perspectives on perception and action. Hillsdale, NJ: Erlbaum. Awh, E., Armstrong, K. M., & Moore, T. (2006). Visual and oculomotor selection: Links, causes and implications for spatial attention. Trends in Cognitive Sciences, 10, 124–130. Becklen, R., & Cervone, D. (1983). Selective looking and the noticing of unexpected events. Memory and Cognition, 11, 601–608. Cartwright-Fincha, U., & Lavie, N. (2007). The role of perceptual load in inattentional blindness. Cognition, 102, 321–340. Findlay, J. M. (2004). Eye scanning and visual search. In J. M. Henderson & F. Ferreira (Eds.), The interface of language. Vision and action: Eye movements and the visual world (pp. 135–159). New York: Psychology Press. Gottlob, L. R., Cheal, M., & Lyon, D. R. (1999). Attention operating characteristics in a location-cueing task. The Journal of General Psychology, 126, 271–287. Haines, R. F. (1991). A breakdown in simultaneous information processing. In G. Obrecht & L. W. Stark (Eds.), Presbyopia research: From molecular biology to visual adaptation (pp. 171–175). New York: Plenum. Henderson, J. M. (2003). Human gaze control during real-world scene perception. Trends in Cognitive Sciences, 7, 498–504. Koivisto, M., Hyönä, J., & Revonsuo, A. (2004). The effects of eye movements, spatial attention, and stimulus features on inattentional blindness. Vision Research, 44, 3211–3221. D. Memmert / Consciousness and Cognition 19 (2010) 1097–1101 1101 Kuhn, G., Amlani, A. A., & Resnsink, R. A. (2008). Towards as science of magic. Trends in Cognitive Science, 12, 349–354. Kuhn, G., & Findlay, J. (in press). Misdirection, attention and awareness. Inattentional blindness reveals temporal relationship between eye movements and visual awareness. Quarterly Journal of Experimental Psychology. Kuhn, G., & Tatler, B. W. (2005). Magic and fixation: Now you don’t see it, now you do. Perception, 34, 1153–1161. Kuhn, G., Tatler, B. W., Findlay, J. M., & Cole, G. G. (2008). Misdirection in magic: Implications for the relationship between eye gaze and attention. Visual Cognition, 16, 391–405. Lavie, N. (1995). Perceptual load as a necessary condition for selective attention. Journal of Experimental Psychology: Human Perception and Performance, 21, 451–468. Mack, A., & Rock, I. (1998). Inattentional blindness. Cambridge: MIT Press. Memmert, D. (2006). The effects of eye movements, age, and expertise on inattentional blindness. Consciousness and Cognition, 15, 620–627. Memmert, D., & Furley, P. (2007). ‘‘I spy with my little eye!” – Breadth of attention, inattentional blindness, and tactical decision making in team sports. Journal of Sport & Exercise Psychology, 29, 347–365. Most, S. B., & Astur, R. S. (2007). Feature-based attentional set as a cause of traffic accidents. Visual Cognition, 15, 125–132. Most, S. B., Scholl, B. J., Clifford, E. R., & Simons, D. J. (2005). What you see is what you set: Sustained inattentional blindness and the capture of awareness. Psychological Review, 112(1), 217–242. Most, S. B., Simons, D. J., Scholl, B. J., & Chabris, C. F. (2000). Sustained inattentional blindness: The role of location in the detection of unexpected dynamic events. PSYCHE, 6(14). Neisser, U. (1979). The control of information pickup in selective looking. In A. D. Pick (Ed.), Perception and its development: A tribute to Eleanor J. Gibson (pp. 201–219). Hillsdale: Lawrence Erlbaum. Neisser, U., & Becklen, R. (1975). Selective looking: Attending to visually specified events. Cognitive Psychology, 7, 480–494. Neumann, O. (1987). Beyond capacity: A functional view of attention. In H. Heuer & A. F. Sanders (Eds.), Perspective on perception and action (pp. 361–394). Hillsdale, NJ: Erlbaum. Neumann, O., Van der Heijden, A. H. C., & Allport, D. A. (1986). Visual selective attention: Introductory remarks. Psychological Research, 48, 185–188. Posner, M. I. (1980). Orienting of attention. Quarterly Journal of Experimental Psychology, 32, 3–25. Posner, M. I., & Peterson, S. E. (1990). The attention system of the human brain. Annual Review of Neuroscience, 13, 25–42. Rensink, R. A. (2002). Change detection. Annual Review of Psychology, 53, 245–277. Simons, D. J. (2000). Attentional capture and inattentional blindness. Trends in Cognitive Sciences, 4, 147–155. Simons, D. J., & Chabris, C. F. (1999). Gorillas in our midst: Sustained inattentional blindness for dynamic events. Perception, 28, 1059–1074. Simons, D. J., Chabris, C. F., Schnur, T. T., & Levin, D. T. (2002). Evidence for preserved representations in change blindness. Consciousness and Cognition, 11, 78–97. Strayer, D. L., Drews, F. A., & Johnston, W. A. (2003). Cell phone induced failures of visual attention during simulated driving. Journal of Experimental Psychology: Applied, 9, 23–32.

Consciousness and Cognition 19 (2010) 731–744 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Self-deception as pseudo-rational regulation of belief q Christoph Michel *, Albert Newen 1 Ruhr-Universität Bochum, Institut für Philosophie, GA3/141, Universitätsstraße 150, D-44780 Bochum, Germany a r t i c l e i n f o Article history: Received 28 December 2009 Available online 23 August 2010 Keywords: Self-deception Rationality Belief-revision Self-immunization a b s t r a c t Self-deception is a special kind of motivational dominance in belief-formation. We develop criteria which set paradigmatic self-deception apart from related phenomena of automanipulation such as pretense and motivational bias. In self-deception rational subjects defend or develop beliefs of high subjective importance in response to strong counter-evidence. Self-deceivers make or keep these beliefs tenable by putting prima-facie rational defense-strategies to work against their established standards of rational evaluation. In paradigmatic self-deception, target-beliefs are made tenable via reorganizations of those belief-sets that relate relevant data to target-beliefs. This manipulation of the evidential value of relevant data goes beyond phenomena of motivated perception of data. In selfdeception belief-defense is pseudo-rational. Self-deceivers will typically apply a dual standard of evaluation that remains intransparent to the subject. The developed model of self-deception as pseudo-rational belief-defense is empirically anchored. So, we hope to put forward a promising candidate. Ó 2010 Elsevier Inc. All rights reserved. 1. A distinction: auto-manipulation vs. self-deception 1.1. Pluralism Peter believes that he and John, a famous and celebrated actor and (in reality) a quite distant friend of Peter, are involved in a very deep friendship. They do know each other, and John has even invited Peter to some of his parties. But Peter has never even come close to John’s inner circle. Nevertheless their recent interaction has motivated Peter to believe that he is John’s favorite. However, no one else, including John himself, would come to the conclusion that he and Peter are close friends. Examples like this illustrate that self-deception is a familiar phenomenon. But the analysis of self-deception has proven to be a tricky endeavor. There is little consensus on the proper analysis of the cognitive state and the cognitive dynamics of selfdeception. The two basic reasons for the high diversity in theory are, firstly, that there are various ways in which motivation can influence acceptance and, secondly, that the pre-analytic folk-intuition of self-deception is heavily under-determined. To put things right, we suggest a distinction between motivational dominance on acceptance in general and a more constrained notion of self-deception in particular. Most people are willing to call a subject S a ‘self-deceiver’ in loose, ordinary folk-talk, if the following is true of his or her mind and behavior: q This article is part of a speical issue of this journal on Self, Other and Memory. * Corresponding author. Fax: +49 234 32 14963. E-mail addresses: Christoph.Michel@rub.de (C. Michel), Albert.Newen@rub.de (A. Newen). 1 Fax: +49 234 32 14963. 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.06.019 732 C. Michel, A. Newen / Consciousness and Cognition 19 (2010) 731–744 1.1.1. Pattern of Motivational Dominance (PMD)2 (1) S is of normal intelligence and has normal rational capacities. (2) S is strongly motivated to evaluate p as true. (3) S has access to a sufficient amount of relevant information about the subject matter SM and this information is supportive of not-p. (4) S honestly accepts the proposition p. (5) If S had been neutral towards SM3, the same evidence would have led S to accept not-p instead of p. A few preliminary remarks about what makes those conditions intuitively necessary: Condition (1) says that there can be no general absence or breakdown of S’s rational capacities, including S’s ability to draw simple inferences. S is in that sense sufficiently rational. Condition (2) pragmatically excludes cases in which subjects come to be deceived that p without having any motivation to believe that p. The typical motivational source of self-deception is that the subject matter at stake is of high individual importance to S. Furthermore, as a potential self-deceiver, S cannot be completely blind to the relevant evidence. If S did not access the evidence, S could not be blamed for anything like self-deception, and this is why we need (3). Additionally, S has to accept p. Here ‘acceptance’ is a liberal notion of belief. S’s acceptance of p need, e.g. not entail that S always consistently acts on p. The degree or nature of acceptance is left open here. Finally, (5) claims that it is motivation which causes p-acceptance. S would judge not-p in similar cases when he is neutral with regard to SM.4 Note that (1) and (3) should imply that self-deceivers are evidence-sensitive and in principle capable of rational belief-adjustment with regard to the subject matter. This excludes belief-like phenomena like pathological confabulation and various forms of delusion from the realm of everyday self-deception. PMD provides the necessary and sufficient conditions for a pre-analytic folk-notion of self-deception in a broad sense. But PMD—as it stands—only says that motivation dominates acceptance of p in a rational subject S. PMD is a coarsegrained pattern; it does not specify how the motivational impact on acceptance is accomplished. Therefore PMD is compatible with nearly all the established theoretical analyses of self-deception.5 There are multiple ways motivational dominance in the sense of PMD could be implemented in S. If we do not want all of these equally to go through as ‘selfdeception’, it is clear that the pre-analytic inclination to label a given case of motivational dominance as ‘self-deception’ will not yet tell us whether this type of cases actually captures proper self-deception in contrast to other forms of PMD. The differences between forms of motivational dominance are remarkable. The debate on self-deception has produced various moreor-less appealing psycho-philosophical stories about how motivation contributes to S’s acceptance of p. We can choose freely between models that analyze self-deception as a form of intentional deception, as a case of motivational bias, models which describe self-deception as a gap between belief and avowal, as a form of attention control, as a form of pretense, a form of repression, as a phenomenon related to confabulation or as false meta-belief, just to name a few. This variety of phenomena requires us to specify all of the PMD-conditions given above and see what additional criteria need to be provided. The underdeterminacy of PDM clearly suggests a methodological limitation to a simply example-based approach to self-deception. Example-based approaches look for what is sufficient to explain cases we would pre-theoretically label as ‘self-deception’.6 This method thereby takes it that the pre-analytic intuition is sufficient to identify those cases that are examples of ‘real selfdeception’. However, the above considerations show that all it might finally do is to pick one of various types of PMD-cases. Thus, more work needs to be done. Going beyond PMD we will fix ‘self-deception in a narrow sense’7 by dint of further reasonable constraints. The basic distinction we propose is the distinction between auto-manipulation8 in general and self-deception in particular. PMD is a general model of auto-manipulation in sufficiently rational subjects in all of its varieties, and it is basically an empirical question as to which different forms of motivational dominance actually occur. Auto-manipulation occurs in any case where motivation dominates acceptance of a proposition p. ‘Self-deception’ refers to a class of special phenomena within the broader frame of auto-manipulation. In what follows, we will show that this distinction is useful and argue that being a case of PMD is, while necessary, not sufficient for being a case of self-deception. We intend to show with regard to two examples that selfdeception in a narrow sense is clearly different form other forms of motivational dominance. In the next section, we introduce three fundamental criteria of adequacy for self-deception. 2 Convention for the rest of this paper: ‘‘p” designates the content of the target belief, favored by the subject S but unsupported by evidence; ‘‘not-p” is the content of the belief actually justified by evidence. 3 I.e. if condition (2) is not fulfilled. 4 This comprises the counterfactual claim that, if S had been neutral towards the subject matter, S would have judged not-p, as well as the claim that typically S will judge not-p in parallel cases on the basis of equivalent evidence if those cases are not evaluated by S as being of any subjective importance. 5 In particular, the conditions of PMD do not even address the question that has driven the debate among philosophers: What’s the subject’s or the self’s very own contribution to the manipulation of his or her attitude? PMD, as it stands, is neutral between what the intentionalist and the deflationist camps disagree on. 6 An example-based approach explicitly receives methodological preference by Mele (2001, p. 5f). 7 Simply ‘‘self-deception” in the following. 8 The term ‘auto-manipulation’ is supposed to be neutral between different theories of self-deception and implies no intentionalist preference. C. Michel, A. Newen / Consciousness and Cognition 19 (2010) 731–744 733 1.2. Three constraints on self-deception Giving a satisfactory account of self-deception is a joint-venture of normative and descriptive elements. Basically, an adequate model should incorporate reasonable theoretical and conceptual constraints and capture a real phenomenon. Defining self-deception requires clarifying all five conditions of PMD and providing supplemental conditions. Three criteria of adequacy for an analysis of self-deception suggest themselves from the very start. (i) An adequate model of self-deception must be psychologically viable9 Criterion (i) presupposes that a model is coherent. It is difficult to judge from the armchair what is actually ‘psychologically possible’, but a model of self-deception must, at the very least, be in line with some minimal requirements of rationality, i.e. it should show how self-deception could be plausibly implemented in a sufficiently rational subject. So-called literalist ‘deception-models’10 of self-deception where S deceives himself knowingly and purposefully require that: (a) S believes that not-p, (b) S intends to believe that p, and (c) S comes to believe that p as a result of his intentional efforts to deceive himself. Self-deception, so conceived, involves a simultaneous endorsement of both beliefs, p and not-p, as well as an intention to deceive oneself. This has obvious difficulties. Therefore, refined intentionalism as well as non-intentionalism about self-deception both suggest that to be plausible, a model of everyday self-deception should, firstly, avoid presupposing that self-deceivers must hold incoherent beliefs of the form ‘‘p and not-p”.11 Secondly, self-deception should not require any intention of the form ‘‘Deceive yourself!” since such intentions strongly tend to undermine their accomplishment in sufficiently rational subjects.12 In this sense, the so-called ‘paradoxes of self-deception’ impose limitations on how self-deception can be plausibly characterized as a state and as a process, and virtually all accounts are tailor-made to avoid them. Beyond avoiding the ‘paradoxes’, theorists who aim at describing the paradigmatic garden-variety self-deception would also require the following: (ii) An adequate model of self-deception should provide a description of a phenomenon that is both real and commonplace Condition (ii) says that a model of self-deception need not only be viable but also capture an actual and commonplace phenomenon and explain intuitive examples of garden-variety self-deception. Condition (ii) by itself does not exclude the possibility that forms of auto-manipulation may exist, which are more common or widespread than our target phenomenon. Self-deception occurs frequently enough to be familiar, but it still seems the exception rather than a regular condition of belief. Empirical grounding will be useful to show that a model satisfies (ii). It is, of course, beyond the scope of this article to discuss in detail all intentionalist and non-intentionalist proposals on how to navigate around the cliffs of paradox. And we do not dispute that there are possible cases of intentional auto-manipulation that satisfy the essential requirements of the deception-model. So called ‘mental partitioning’,13 ‘temporal partitioning’,14 or other sophisticated methods of inducing false beliefs in oneself offer ways to retain the essentials of the deceptionmodel. But since a critique of intentionalism is not in the focus of this article, we have to rest content with the observation that intentionalist constructions all share a tendency to satisfy (i) for the price of failing (ii) in different respects.15 We suggest that both forms of partitioning do not describe what one would call intuitive and common examples of self-deception, although they can be seen as actual cases of auto-manipulation. For example if mental partitioning (two subsystems in one mind) is to be the strategy to prevent the paradoxes, it is questionable whether this would capture a commonplace phenomenon, even if such divisions are possible.16 9 For an elucidation of the paradoxes see Section 2.1. In its strongest version the intentionalist ‘deception-model’ of self-deception parallels intentional interpersonal deception. Interpersonal deception requires a deceiver, A, who intentionally leads a finally deceived individual, B, into believing a proposition p that A himself considers false. In the standard case, A believes not-p and leads B to believe that p is true. Similarly, in self-deception S intends to deceive himself. S is motivated to do so due to his desire to evaluate p as true, paired with his belief that not-p. 11 Which is not to say that the mind could not harbor inconsistencies. 12 If A and B are identical, the constitutive asymmetry in knowledge between A and B which enables literal deception gets lost. The deceiver and the deceived both know about the actual facts as well as about the deceptive intention. This spoils self-deception in the same way A will spoil his plan to deceive B by starting out saying, ‘‘Hello, might I deceive you into believing that I’m a professor at Harvard?” In other words, a dynamic paradox is lurking. How could the intention ‘‘Deceive yourself!” bring about a successful belief change when we appear to be required to be ignorant of this very intention as well as of the facts at issue, if we are to become deceived? 13 So-called ‘mental-partitioning’-accounts explain self-deception by postulating more or less independent sub-centers of agency that allow to implement self-deception very close to the model of other-deception. 14 See Bermúdez (1997, 2000). 15 José Bermúdez (2000) has presented a form of self-induced long-term attitude-shift as an instance of intentional self-deception. Brian McLaughlin refers to the ‘diary-case’, a ‘‘memory exploiting stratagem” of self-induced deception (McLaughlin, 1988, Section 31f). A person can mislead herself about the date of an appointment she wants to miss by intentionally making a false entry in her diary and trusting her notoriously bad memory. She’s right about her memories qualities and and misses the appointment, believing that it is set as the diary says. 16 In this article we cannot discuss in detail refined modifications of divisionism like the one provided by Pears (1991). 10 734 C. Michel, A. Newen / Consciousness and Cognition 19 (2010) 731–744 Motivational influence on our beliefs turns deceptive when normally rational subjects manipulate their own attitudes in direct response to strong evidence to the contrary. S can only be blamed for self-deception, if the evidence is strongly asymmetrical with regard to the subject matter at stake. If the body of evidence was really ambiguous, we could hardly blame S for self-deception. If S notices that p has some evidence, if p is integrated in a belief-network and if p has proven pragmatically useful, then S has a perfectly rational motivation not to give up or not to reject his belief that p prematurely. We therefore stress that over and above (i) and (ii), the following condition is a crucial one. (iii) In self-deception rational subjects establish or re-establish p-acceptance in the face of strong evidence to the contrary Self-deceivers must deal with strong evidence against p. We develop the notion of strong counter-evidence and its implications for the dynamics of self-deception in the following section. 2. Criterial evidence and the doxastic dynamics of self-deception From what has been said thus far, a core challenge for a theory of self-deception has become clear. Being a commonplace phenomenon, self-deception is a form of auto-manipulation that has to work in rational and evidence-sensitive subjects. At the same time, self-deception, as a matter of fact, requires that evidence be strongly asymmetrical. So, (iii) will shape how we have to specify PMD-condition (3) in a case of self-deception, i.e. S’s access and treatment of evidence. The point that notp-evidence has to be sufficiently strong before a belief that p can become self-deceptive has been at least implicitly acknowledged by many different approaches to self-deception. These comply with (iii) insofar they presuppose that sufficiently rational self-deceivers do truly believe what is suggested by the evidence, i.e. not-p. We will suggest a weaker alternative: It is sufficient if S considers or suspects that not-p in the face of an undeniable evidential challenge. But let us first evaluate the reasons one may have to ascribe a belief that not-p to self-deceivers. The simple reason for requiring S to believe that not-p is that – since S is basically rational and sensitive to evidence – S will form a belief that accords with the evidence. In addition, behavioral facts have been regarded as requiring S to believe that not-p. Audi (1988) and Rey (1988), among many others, have suggested that a belief that not-p is responsible for supposedly typical forms of self-deceptive behavior such as S’s deliberate avoidance of situations that will confirm not-p to S as well as S’s hesitance to act in accordance with p in some, if not in most or even all situations.17 A different line of motivation for the assumption that not-p is believed by S comes from the intentionalist camp. Davidson (1985, 1998), Pears (1991) and Bermúdez (1997, 2000) claim that S’s belief that not-p plays a crucial role in motivating and initiating the process of self-deception. What else, so one might ask, should move S to deceive himself in the first place, if not his believing that p is false? Thus, in addition to S’s strong motivation to believe that p, S must believe that p is actually false in order to engage in self-deception. Furthermore, the belief that not-p is actually true causes the cognitive dissonance and explains the psychological tension that have been proclaimed a characteristic feature of self-deception by various authors. So, it appears that there could be at least four good reasons to suppose that a person who is self-deceived in accepting p will also believe that not-p: – Rationality: Since S is sufficiently rational and evidence-sensitive, S simply recognizes that not-p. – Behavior: The assumption that not-p is believed by S explains typical self-deception behavior. – Motivation and activation: Only a belief that not-p explains why a process of self-deception is initiated. – Phenomenology: S’s experience of a characteristic psychological tension is caused by his believing that actually not-p. We need not fully endorse all four theoretical motivations for assuming a belief that not-p in order to recognize that they reflect a crucial question which any theory of self-deception has to face. In which way and to what result is not-p-evidence actually accessed and how can it be treated by S? This question is at the very heart of a theory of self-deception, since to answer it is to give a model of the cognitive dynamics of self-deception. Before we can provide such a model, we need to clarify the notion of strong evidence we presuppose in (iii). Counter-evidence is sufficiently strong if, under a neutral evaluation, the not-p-data are clear enough to destroy p-confidence. The kind of evidence we have in mind is best characterized as ‘criterial evidence’ against p (also referred to as criterial negative evidence ‘E ’ in the following). E is directly perceived as evidence against p and is a criterion for believing not-p. If S evaluates information as criterial evidence against p, S entertains a background-belief that this type of information makes the truth of p strongly unlikely. This meta-belief is part of S’s standard-rational evaluation of p-related information and defines what sort of data regularly defeats p. On the basis of this notion of criterial evidence we can develop a preview of what we take to be a very plausible model of the doxastic dynamics of self-deception. The process of self-deception is set off by recognized E in combination with a strong motivation to evaluate p as true. At the outset, if S is rational and sensitive to evidence, S will directly register relevant 17 To give an example: The fact that Peter knows that he and John are not actually best friends not only explains Peter’s absence of trust in John with regard to important matters of privacy but also his systematic avoidance of situations which strengthens the evidence – e.g., Peter now avoids joining when John and some of his friends go hang out, since he had already noticed on several occasions that John treated some of them with more privilege and in a much more attentive manner than him. Whereas psychological literature has investigated on the phenomenon of ’self-immunization’ in this context, we prefer to speak more generally about ’belief-immunization’. C. Michel, A. Newen / Consciousness and Cognition 19 (2010) 731–744 735 data as E , that is, S will perceive and process these data as strong evidence against p. Accessed E triggers a belief-formation tendency towards not-p in S (at least at S’s first encounter with the data), i.e. S’s rational tendency of belief-formation will strongly suggest not-p and undeniably challenge p-confidence. However much S may desire that not-p be false, S is forced to consider or suspect strongly that not-p, if he is capable of a rational evaluation of p-relevant data in the first place. If we tried to imagine S as ‘blinded’ to E or E as being filtered out of S’s overall body of evidence, the conditions of rationality and evidence-sensitivity as well as the access-condition would be implausibly weakened. As described so far, the process is perfectly rational. S enters self-deception only by a subsequent pseudo-rational maneuver in which S attempts to invalidate the undeniable challenge put to p by E . This process has been described as a process of ‘‘immunization” in the psychological literature18 and we will characterize its pseudo-rational quality in Section 4. Assuming that not-p only strongly suggests itself to S, without assuming that S actually believes that not-p has two main advantages. Firstly, the belief-status of p can be maintained, there is no conflict with (i), i.e. there is no threat of a static paradox. Secondly, the aforementioned four reasons for assuming settled not-p beliefs can be accommodated by the alternative attitude without exception. Considering on the basis of accessed E that not-p is very likely true will do. First, in order to account for S’s rationality and evidence-sensitivity, it is sufficient to assume that S suspects not-p on the basis of E . Second, if S is forced to suspect that p is very probably false, this will at some point cause typical behavioral asymmetries between S and S, where S is a self-deceiver and S believes p without strong evidence to the contrary. It will, for example, easily suffice to explain why and how S can identify and avoid the situations he expects to defeat p.19 Thirdly, criterial counter-evidence and S’s subsequent considering of not-p are perfectly sufficient to motivate and initiate a process that can be characterized as selfdeceptive. Lastly, E exerts pressure on S’s desired belief that p. S’s believing p together with his considering not-p is sufficient to evoke cognitive dissonances and psychological tension, if S wants to evaluate p as true. In summary, we suggest that S’s consideration of not-p on the basis of criterial evidence against p is the sufficient and adequate way of spelling out the condition of access to evidence. No settled belief that not-p is necessary. We have argued that a commitment-free consideration of not-p by S is a plausible consequence of S’s access to criterial not-p-evidence. In our approach to self-deception we assume that the only belief self-deceivers need to commit themselves to is p. What needs explanation then is exactly how minimally rational subjects manage to hold or establish non-rationally motivated beliefs against undeniable challenge and to attain confidence in p. As we will argue in the following sections, this is possible by pseudo-rational reorganizations within S’s ego-centric belief-system. These reorganizations serve the purpose of defending p by invalidating the evidential challenge. In a case of self-deception, contrary to rational belief-revision, the desired invalidation of E is achieved by adaptation-strategies that are objectionable from the rational point of view. Pseudo-rational reorganization of p-relevant belief-sets amounts to a rationalization of S’s commitment to p in the face of criterial counter-evidence. Severe evidential challenge forces S to rationalize his commitment to p if S wants to retain or establish p-confidence. We take it as fairly plausible to assume that being a sufficiently rational subject, S’s beliefs are anchored in a rational belief-system, i.e. S takes p to be answerable to evidence and reasons. Even if our reasons for beliefs are not the only or not even the dominating causes of our beliefs that p, it will be rather difficult for us to retain p-confidence if (a) p faces a massive challenge and (b) in facing it, we find ourselves unable to explain how we can commit ourselves to p from a rational point of view. In that sense, self-deceptive rationalization should make p look rationally ‘committable’ (at least) to S, i.e. S should be able to see p as supported by evidence and reasons. Sanford (1988) has developed a rationalization-analysis of self-deception about one’s reasons for actions. Most of us require their reasons for action to be ‘ostensible’20, i.e. these reasons should comply with our self-image as rationally and morally responsible actors. In an analogous sense, beliefs must be committable in order to be embraced by a believer who is minimally ‘epistemically responsible’ about his beliefs. Nonrational commitments to the truth of p that are merely based on desire or emotion would be cases of straightforward irrationality. In self-deception S makes P rationally committable via changes among his background-beliefs. This amounts to a rationalization of S’s commitment to p in the face of E . This motivation-based rationalization will immunize p against E -type evidence. S’s rationalization of his commitment turns into self-deception, (a) if that process puts elements of rational thinking to work against the standard-rational evaluation of the subject matter, (b) if the standard-rational evaluation still remains operative when S evaluates neutral cases21 on the basis of equivalent information, and (c) if rationality-based p-immunization is ad hoc and enjoys priority over the pursuit of truth. We will illustrate in Section 4.2, how self-deceivers characteristically engage in such a dual standard of rationality. In self-deceptive rationalization, S’s capacity of rational thinking cannot only be blinded but must be used as a tool of auto-manipulation. To keep or make p rationally committable, S has to adapt his belief-set in a way that invalidates criterial counter-evidence and saves the desired conclusion. Therefore pseudo-rational adaptation marks a level that can outperform alternative forms of auto-manipulation when counter-evidence is strong. The preceeding sketch of the cognitive dynamics of self-deception shows how we can assume that it is sufficient for selfdeception that S holds only p throughout the whole process (and does not believe the contrary). ‘One-belief views’ avoid the puzzles of self-deception and make it easy to satisfy (ii) by rejecting the two-belief-requirement. In the next section we will examine influential, alternative one-belief views. There are two main types of theories. One claims that in self-deception S 18 See Sections 4.1 and 4.2. This point about suspicion has also been made by van Leeuwen (2007, p. 427). 20 Sanford (1988, p. 157). 21 See PMD condition (5). 19 736 C. Michel, A. Newen / Consciousness and Cognition 19 (2010) 731–744 believes that not-p, while p-acceptance is not situated on the level of belief. We consider Tamar Szabó Gendler’s recent ‘pretense’-model in this context. The other represented most prominently by Alfred Mele, holds that motivationally biased belief-acquisition only involves a belief that p. 3. What do self-deceivers believe? 3.1. The ‘pretense’-view Audi (1982, 1988), Rey (1988) and others have proposed that there is a gap between what self-deceivers believe and what they avow. They claim that self-deceivers actually believe that not-p and that p-acceptance lacks the essential property of belief-states: being action-relevant. Recently, Tamar Szabó Gendler has claimed that self-deceivers do not believe, but merely ‘pretend’ that p. Her view is more intuitive and more adequate with regard to the behavior of self-deceivers than, e.g. the version of Rey. It avoids a strict avowal-action-gap by conceding to ‘pretense’ a scope of action-motivation that is limited, but nonetheless exceeds the level of mere avowal. She characterizes a self-deceptive pretense-mode of mind in the following way: ‘‘A person who is self-deceived about [p]22 pretends (in the sense of makes-believe or imagines or fantasizes) that [p] is the case, often while believing that [not-p] is the case and not believing that [p] is the case. The pretense that [p] largely plays the role normally played by belief in terms of (i) introspective vivacity and (ii) motivation of action in a wide range of circumstances.” (Szabó Gendler, 2007, p. 233f). ‘Pretense’, ‘imagination’, etc. are used in a technical sense here. In contrast to belief, which is defined by Szabó-Gendler as a ‘receptive’ attitude in the service of truth, ‘pretense’ is characterized as ‘projective’. Instead of pursuing truth, self-deceivers mentally escape from the not-p environment and create for themselves a ‘p-world’. This basically makes them inhabitants in two worlds; one in which they rationally believe not-p and another one in which they actively enclose themselves in an imaginary p-world that is buffered against hostile evidence and kept from rational control. Since S cannot believe p, he can just ‘pretend’ or ‘fantasize’ that p. ‘Pretense’ governs action in a kind of encapsulated fantasy-world-mode of self-deception and beliefs govern action in the actual-world-mode. Self-deceivers act on their ‘pretense’ as long as no other motivation overrides the original motivation to stick to p and forces them back into the real-world-belief that not-p. The behavioral profile of ‘pretense’ actually looks very ‘self-deception’-like. And in contrast to the avowal-view, ‘pretense’ clearly satisfies condition (ii); it seems not only real, but also commonplace. Moreover, it addresses (iii), since not-p evidence typically leads to not-p beliefs and p-fantasizing. ‘Pretense’ is a real phenomenon of auto-manipulation. The familiarity of ‘pretense’ (and the fact that ‘pretense’ is clearly not paradoxical) and the similarities between the behavioral patterns of ‘pretense’ and self-deception motivate the general claim: self-deception, if we look at it, is typically only ‘pretense’. However, there are several reasons to be skeptical about this move. Szabó Gendler claims that the view seems ‘‘natural” and she wonders how this obvious explanation has been overlooked. A probable reason is that no one would commonly consider a normal p-pretender to be self-deceived, and therefore, the general claim that all self-deceivers are only engaging in a sort of pretense, imagining or fantasizing is counterintuitive rather than natural. Most, to the contrary, share the intuition that what self-deceivers do, is more like believing p than pretending p. A self-deceptive ‘pretender’ or ‘fantasizer’ is someone who seeks ways of avoiding acknowledging and facing what he knows. He engages in a kind of wishful thinking whereas classically conceived, a self-deceiver seems to be someone who actually gets things wrong and who will defend his doxastic p-commitment, if challenged. Thus, an argument is required to show why it is more intuitive to characterize S as pretending than as believing that p. But apart from the question whether the ‘pretense’-analysis has our intuition on its side, it is not clear whether ’pretense’ is fit to explain self-deception and at the same time to contrast with belief-status. Moreover the phenomenon of ‘pretense’ can easily be acknowledged as a phenomenon of auto-manipulation but remain distinguished from self-deception. In contrast to mere pretenders or imaginative thinkers, Szabó Gendler’s ‘pretenders’ must develop sufficient confidence in p, if they are to avow p sincerely. So, it is obvious that ‘pretense’ must differ from ordinary pretense or fantasizing on the one hand, and from belief on the other. This is the case because: (a) ordinary pretense and fantasizing are by definition confidence-free, non-doxastic states and (b) belief is the thing the account wants to avoid. To play its explanatory role the new concept of ‘pretense’ must be a hybrid that eats its cake and keeps it, too. It has to be belief-like in explaining S’s pbehavior and p-confidence, while being sufficiently imagination-like so as not to conflict with S’s knowledge that p is untrue. This is hard to accept for in imaginative attitudes like ‘pretense’ there is no room for confident and sincere p-avowal as it constitutes the phenomenon of self-deception. Self-deception without p-confidence is not a serious option. A single pretender who’s not a liar, would not display his confidence in defending p, but it is natural to assume that a real self-deceiver would. We assume that it is part of the phenomenon of self-deception that self-deceivers will characteristically be ready to defend p, e.g. when p is directly challenged by other members of S’s community. If we want S to avow p sincerely and develop p-confidence, we should expect that S will explore ways of acquiring p-confidence by invalidating not-p-evidence.23 If S 22 23 Szabó-Gendler’s convention is changed into the convention of this paper, in which ‘‘p” designates the deceived target belief. If the ‘pretense’-mode of mind were to be immune against challenges, evidence-sensitivity and rationality would be lost. C. Michel, A. Newen / Consciousness and Cognition 19 (2010) 731–744 737 turns out to be unable to defend his avowal by rationalizing it, ‘pretense’, and along with it self-deception will either break down or S would have to concede a straightforward irrationality in his p-confidence. So, if Szabó Gendler’s ‘pretense’ is like normal pretense, imagining or fantasizing, it conflicts with the requirement of sincere avowal and p-confidence. Moreover one will not be able to explain how self-deceivers establish and defend p against acute evidence and challenges by others. But if self-deceivers do establish and defend p-confidence in the face of challenge, we leave pretense behind and enter the realm of belief. Motivated ‘pretense’, ‘fantasizing’ and ‘imaging’ are ways to actively avert reality. This can be characterized as an especially strong form of wishful thinking. But averting reality and self-deception can and should be distinguished. They are different phenomena, even if they occasionally exhibit behavioral resemblances. Averting from reality is still is still an option, if self-deception breaks down. The ‘pretense’-model remains a static picture in which evidence and motivation go separated ways. S’s actual beliefs are not subject to motivational influence by definition. ‘Pretense’, if it is different from belief in any significant way, neither alters the subject’s confidence in not-p, not does it influence confidence in p either, since p is never believed. Hence, in the ‘pretense’-view, self-deception is a folk-psychological myth from the very beginning. Szabó Gendler’s initial distinction between belief as a merely ‘receptive’ attitude and ‘pretense’ as a ‘projective’ attitude is also too rigid. It underestimates our actual capacity to see or reinterpret a not-p-world as a p-world under the influence of our motivations. Self-deceivers, as we will see in the next sections, are able to manipulate the evidential value of unwelcome information and develop p-confidence against the evidence. In our view, self-deception is a cognitively dynamic process at the belief-level. This process is not properly described as a process of averting reality, but rather as a process where selfdeceivers incorporate the accessed evidence by making use of their capacity of rational reasoning as a tool of auto-manipulation. One might argue that the presupposed ability and readiness to defend and rationalize p does not necessarily indicate belief-level. But, as a simple matter of fact, rationalized p-commitments will often reach belief-status even against evidence. And there is no threat of paradox in assuming that they would. The condition of rationality and evidence-sensitivity and the condition that not-p-evidence must be strong, are compatible with p-acceptance on belief-level, as we intend to demonstrate in the following sections. Alfred Mele has demonstrated that motivation can be efficacious on the level of belief. We examine his account before developing our model of self-deception as pseudo-rational belief-defense. 3.2. Motivational bias In his influential deflationary analysis of self-deception as ‘motivational bias’, Alfred Mele (1997, 2001) holds that it is sufficient and typical that self-deceivers only acquire a belief in what they desire to be true, namely p, without entertaining any belief that not-p. Mele thinks that not-p-beliefs (or weaker replacements) are completely dispensable when it comes to explaining what happens in garden-variety cases of self-deception. Explaining garden-variety cases requires neither any ‘activator’ nor any sort of intentional effort on the part of the agent—nor should we assume that psychological tension is typical of self-deception. Rather, we can think of real self-deception as working much more smoothly and effortlessly. Beliefformation gets silently biased by our motivation. In his model Mele draws on empirical evidence about stable patterns of biased information processing in subjects, so-called ‘cold biases’: In pragmatic heuristics, for example, the ‘vividness’ and ‘availability’ of p-data, or p-‘confirmation bias’ (a stable inclination to search for confirming rather than disconfirming evidence of hypotheses) generally facilitate p-acceptance at a level far below that of rational scrutiny.24 Cold biasing is a form of silent data-selection or data-structuring that is efficacious entirely at a procedural, sub-intentional level. The essential aspect of self-deception according to Mele is ‘hot bias’, i.e. biased processing driven by motivation. S’s individual motivation to believe that p structures S’s accessing and processing of p-relevant information in a manipulative way and it is further claimed that such motivationally biased belief-formation is sufficient to explain garden-variety cases of self-deception. Thus, if S desires to evaluate p as true and if this desire biases the processing of p-relevant data and S is thereby led into accepting p, S enters self-deception. Mele gives the following jointly sufficient conditions for self-deception: Biased belief-formation 1. The belief that p which S acquires is false. 2. S treats data relevant, or at least seemingly relevant, to the truth value of p in a motivationally biased way. 3. This biased treatment is a non-deviant cause of S’s acquiring the belief that p. 4. The body of data possessed by S at the time provides greater warrant for not-p than for p (Mele, 2001, p. 50f). First, we want to focus on condition (2) and what it tells us about how S accesses and treats evidence. By speaking about ‘‘relevant data” being treated ‘‘in a motivationally biased way” Mele means that S’s motivation to believe that p shapes the body of p-data via biased misinterpretation of p-relevant data and/or via ‘selective focusing and attending’ and/or ‘selective evidence gathering’ (Mele, 2001, pp. 25–31). This way S’s motivation basically functions as a not-p-data filter and as a p-data amplifier: In biased evaluation of the body of p-relevant-data, not-p-supporting data are kept out of focus, get sorted out or downplayed, whereas p-supporting data pass through and receive high promotion. This motivational data-structuring determines S’s perception of the overall evidence and leads S directly into believing that p. 24 Mainly Nisbett and Ross (1980) 738 C. Michel, A. Newen / Consciousness and Cognition 19 (2010) 731–744 The question we would like to stress is: What does S actually ‘see’, when his belief-formation falls victim to such motivationally biased data-processing? What is S’s very own informational basis for confidence in p in that model? Mele explicitly accepts as an exception (2001, p. 52) that the biased mechanisms of data-selection can lead to a misperception of where the weight of the evidence truly lies. In this case, S non-intentionally accumulates an incomplete set of information about the subject matter. The available body of data which supports not-p from unbiased point of view is selected and gathered in a way that S finds the data supporting p since the information S actually has is incomplete. In that case, selective evidence gathering inhibits S’s collecting all relevant not-p-data. Then, contrary to (4), S is actually not in possession of all relevant data and as a consequence, will not access the evidential asymmetry indicated in (4) and consequently, S will not further have to deal with the evidence. Instead, S will develop a belief that accords with a set of data that is incomplete due to manipulation. In this case, S’s belief that p can be described as justified. In Mele’s view, S can nevertheless be self-deceived, even if the relevant not-p information is not in his possession due to motivational influence on the input of data. In the special case just described, the fact that the evidence supports not-p is not accessed by S. We will now elaborate further why in our view motivationally biased data-processing as described in Mele (2001) has general problems with the requirement that the relevant evidence has to be accessed by S. In the special case described above, contrary to condition (4), S is not in the possession of data that speak against p. However, a ‘misperception’ of the evidence seems to be rather typical for biased belief-acquisition, even if S is in possession of all relevant data. According to (4), Mele assumes that typically, S is in possession of the relevant data and treats them in a motivationally biased way. But what we think to be relevant here, is not access to data, but access to (counter)evidence. Even if S has access to all p-relevant data, it is not provided that S himself actually registers the objective evidential trend that the data would suggest to an unbiased subject, if S’s treatment of data is subject to the types of biased processing that Mele mentions. This is possible because data and evidence are situated at different epistemic levels. Two subjects can perceive the same set of data but perceive different evidential values of these data. Only evaluated p-data are evidence for or against p, and one and the same datum or fact can be perceived as having different evidential values. It is not perceptual data that we manipulate but their evidential quality. Desires can doubtlessly influence how we perceive the evidential status of some data. If John desires that Mary be interested in him, a smile from her will be an important datum for John that she might be interested in him, too. If he did not care about her, he might hardly notice it, let alone perceive it as a sign of interest. Motivationally biased perception of data, as characterized by Mele, is situated on the level of data-evaluation. S may collect all the relevant data, but non-intentional mechanisms of biased evaluation of data will highlight p-supportive data and feed them in S’s belief-formation process while data in support of not-p are left aside and downplayed. Therefore, S can access every single datum, while misperceiving25 the evidential tendency of all information. As a consequence, the evidential basis on which S establishes p-confidence is already the result of the biased treatment of data. When being biased, the world as S accesses it is a result of biased treatment of data, i.e. S fails to access the actual evidential value of data due to sub-intentional manipulation. In this sense, S’s motivationally biased treatment of data corrupts the objective weight of evidence and leads S into allocating the weight of evidence to the wrong side. A biased subject can thus collect the relevant data but remain nonetheless blind to the fact that the overall body of data provides greater warrant for not-p than for p. But then, it will be rather the rule than the exception in biased belief-acquisition that S is self-deceived but never aware of the fact that the evidence speaks against p. We think that explaining self-deception against accessed counter-evidence requires a form of treatment of evidence that is not present in Mele’s account. There can hardly be doubted that we possess the capacities of data-manipulation that Mele describes, and it is likely that they influence how we perceive the evidential status of data and that they lead us into false beliefs. But all this only works, as long as the not-p-data S receives is relatively harmless and does not imply a very strong tendency towards not-p. The problem of bias is that a mere biased view on a set of data offers no means to handle instances of criterial counter-evidence.26 Selfdeception, as we have constrained it above, requires a strong asymmetry of evidence, i.e. a strong tendency towards not-p. If that tendency were not strong, ambiguity would remain with regard to the question whether or not p, and hence S could be excused for simply keeping his commitment to p. But S cannot be excused from a rational point of view if he ignores criterial counter-evidence. Criterial counter-evidence, as introduced in Section 2 is directly perceived as evidence against p. Criterial counter-evidence breaks down p-confidence and forces S to consider that not-p is very likely true. Peter has believed that wife Mary is faithful. If he sees Mary passionately embracing John when they think they are unobserved, this will give Peter criterial evidence that he is wrong. Criterial evidence implies that S holds a stable meta-belief such that a perceived event of type X is not compatible with holding the desired belief p. The rule ‘‘If X, then very likely not-p” is a constitutive part of the standard-rationality that Peter endorses in everyday hypothesis-testing. Biased perception-structuring will hardly help him here, because criterial evidence simply cannot be merely misperceived in a biased way. In order to regain confidence in his belief, Peter is forced to update or reorganize his belief-system in a manipulative way if he is to regain confidence in p.27 The task in such reorganization can only be to invalidate the received criterial evidence against p. Hence, if S shall be able to believe that p against criterial counter-evidence, S’s treatment of this evidence will have to consist in the reorganization of his belief-system. 25 With regard to the perception of evidence or of evidential value, we use the term ‘‘perception” in a broad sense. The evidential value attributed to data is the result of an evaluation. In the case of morivational bias this includes top-down influences of motivation on attention and data-interpretation. 26 Similarly, ‘cold bias’ in data-processing will rather facilitate acceptance of p in S in the absence of sufficient evidence than in would be able to establish acceptance of p, if S possesses good evidence against p. 27 Mele’s own examples for ‘‘positive” and ‘‘negative misinterpretation” can also be read as examples of self-serving adaptations in belief-systems, but we don’t have the space to show this here. See Mele (2001, p. 26f). C. Michel, A. Newen / Consciousness and Cognition 19 (2010) 731–744 739 Could not biased evidence filtering alone be strong enough to block even criterial counter-evidence and thereby S’s beliefformation tendency towards not-p? The answer is obviously no, because then, the necessary condition of evidence-sensitivity would not be satisfied, nor would the condition of minimal rationality; believers who are simply insensitive to criterial evidence would be considered victims of mono-thematic irrationality. Cases of lost sensitivity to criterial evidence are located outside the field of garden-variety self-deception even in from a pre-analytic point of view. A strong lack of sensitivity to evidence indicates a phenomenon of delusion rather than a phenomenon of garden-variety self-deception. Thus, biased structuring of evidence-perception as described by Mele is not sufficient to satisfy the criteria of adequacy we have put forward. To achieve that aim we need a reorganization of the subject’s belief-system. Such a reorganization makes believers capable of invalidating criterial evidence. We will explain below, how pseudo-rational reorganization of belief-systems works and why it makes sense to view it as the true criterion of commonplace self-deception. To summarize the argument thus far: Pretense and bias are real cognitive processes. Both of them exhibit considerable similarities to self-deception but both finally fail to describe self-deception in a satisfying way. Both have problems in conceiving of self-deception as a dynamic process of belief, albeit for different reasons. Trying to avoid the classical paradoxes, pretense must deny p-acceptance belief-status. Mele’s account does justice to the intuition that self-deceivers believe that p, but biased evidence-perception cannot handle the constraint that self-deception requires strong counter-evidence. Selfdeceivers are not only biased and they are not only pretenders for they essentially make use of their rational capacity of adapting belief-systems. Neither the ‘motivational bias’-model nor the ‘pretense’-model offers an account of how a strong evidential tendency towards not-p could be handled by self-deceivers. To deny that it could, would amount to denying paradigmatic cases of self-deception. As a commonplace phenomenon, biased perception-structuring will typically contribute to processes of self-deception. But the manipulation of criterial counter-evidence exceeds the level of perception-structuring and exploits our capacity to defend target-beliefs against certain types of undeniable challenge. This process has been described in the empirical literature by Wentura and Greve (2003, 2005). In what follows, we characterize this process as ‘pseudo-rational’ revision in ego-centric belief-systems. 4. Ego-centric belief-systems and dual rationality 4.1. Adaptation in ego-centric belief-systems In this section, we seek to spell out a rationalization model of self-deception and to provide evidence for the claim that it satisfies (i), (ii), and (iii)—namely, that it is free of paradox, that it is an actual and commonplace phenomenon and that it explains how self-deception can cope with criterial counter-evidence, respectively. Our suggestion is to model self-deception as a special class of revisions in ego-centric belief-systems. In ego-centric beliefsystems, beliefs are ranked not only by rational criteria of importance such as their explanatory role, their being highly integrated in a coherent belief-network and their being tested and proven, but also by levels of subjective importance. If a belief has a high subjective importance for S, S has a strong motivation to maintain that belief. According to psychological evidence, motivated belief-formation will be the rule rather than the exception in humans. In everyday belief-revision, there is evidence that the organization of our belief-systems is frequently governed by subjective importance in addition to rational considerations. Typical examples of beliefs of high subjective importance are beliefs about ourselves that are related to self-esteem, but also beliefs that are of emotional value or desirable to S in other ways.28 These may include evaluations relevant to one’s self-image like pride and shame, but also the evaluation of other persons or matters.29 If a belief that p plays a central role in S’s ego-centric belief-system due to its high subjective importance to S, the revision-threshold for p will be correspondingly high. Models of everyday hypothesis-testing30 indicate that, dependent on S’s degree of motivation to believe p, S will in general require more and weightier evidence to reject p than to reject not-p and conversely require less and thinner evidence to embrace p than to embrace not-p.31 This asymmetry in revision-threshold is a first constant of motivated hypothesistesting. Another component of motivated belief-acquisition concerns processes of biased perception-structuring like selective focusing and attending as well as tendentious evaluation of the weight of evidence. They have been described in different ways by theorists like Mele (1997, 2001), Talbott (1995) and van Leeuwen (2008). These processes or strategies serve a non-rational immunization of beliefs of subjective importance. But as long as S is sufficiently rational and sensitive to evidence, we are in need of an additional strategy if we want S to gain p-confidence in the face of criterial counter-evidence. An asymmetrical revision-threshold and biased evaluation of data would not do in cases where the evidence is too plain to be overlooked. We claim that this can only be managed by reorganizing our local belief-sets about the subject matter in question. Self-deceivers characteristically adapt those background-assumptions which constitute the criterial function of E for p. The notion of selfdeception as an immunization of beliefs (typically about ourselves) of high subjective importance against criterial evidence 28 For the role of emotion in self-deception see Sahdra and Thagard (2003) and Mele (2000). For example, if Peter is deceiving himself about Mary, this may be due to the fact that his relationship with Mary deeply affects his self-image or that a certain image of Mary or his relationship to her has high emotional significance to him. Each motivation can be sufficient to trigger a process of self-deception, but they may also coincide. If Mary is Peter’s wife and gives strong evidence that she is cheating on him, Peter’s self-deception that she is faithful can be motivated by his self-esteem as well by the purely emotional significance this fact has to him. 30 Trope and Liberman (1996). 31 In addition, in biased hypothesis-testing S will go for confirming p instead of not-p. See Mele (2001, p. 32). 29 740 C. Michel, A. Newen / Consciousness and Cognition 19 (2010) 731–744 is well-supported by empirical investigation. Experiments by Wentura and Greve (2003, 2005) have convincingly revealed that there is an adaptation of trait-definitions that serves self-immunization. They were able to show that subjects automatically adapt trait-definitions in a self-serving way (like, e.g., ‘being erudite’) if (a), those traits are highly valued parts of their selfimage, and if (b), the subject is confronted with criterial counter-evidence. Via questionnaires and semantic priming tasks, Wentura and Greve confirmed the hypothesis that subjects re-shape trait-definitions according to their actual skills: ‘‘Confronting (student) participants with gaps in their general knowledge will result in a pattern of self-immunization that defines the trait eruditeness in terms of those pieces of knowledge that the participant has successfully demonstrated. Faced with undeniable challenges (failing in a general education quiz), subjects preserve their self-image and incorporate the data by modifying what counts as evidence for ‘being erudite’.” (Wentura & Greve, 2003, p. 33). According to Wentura and Greve, subjects develop idiosyncratic concepts of the trait ‘eruditeness’ in response to what we call criterial counter-evidence. This strategy allows S to keep the valued or esteemed parts of his self-image intact.32 S will adjust his belief-system as a consequence of accessing criterial evidence against beliefs that have high subjective importance. Relevant adjustments will allow S to commit himself to p regardless of the serious challenge posed by counter-evidence. If S has run out of ad hoc-explanations for why p is unaffected by what looks like paradigmatic instances of counter-evidence, S can immunize p against certain types of counter-evidence. In the example above, S has the following doxastic profile before being confronted with his failure: (p) ‘‘I am erudite” (rated by S as an important part of his self-image). (q) ‘‘Knowing about history is necessary for being an erudite person”. (r) ‘‘I have good knowledge of history”. The belief to be preserved is p; the belief that gets refuted by E (the history-test failure) is r. The two beliefs p and r are inferentially related via a connecting belief q. This belief q is the trait definition, i.e. it defines what counts as criterial evidence against p. In this sense, ‘being erudite’ is, as Wentura and Greve note, a theoretical term, and q is part of S’s theory TE about eruditeness in which S fixes all criterial evidence for or against p. 33 In self-deception, S will typically change TE in order to save p. How can S change his theory of eruditeness TE? If S takes r to be defeated by E , S will typically adapt or simply waive q in a way which blocks the inference from not-r to not-p. Additionally, S may introduce supplemental beliefs s and t that justify changing or waiving q. By TE changes, E -type counter-evidence is discarded (e.g. by giving up the claim that historical knowledge is necessary for being erudite). But in large, q can even remain intact (and imply r) if S finds other loopholes for denying that the received information is valid against p. S might say: ‘‘True, being erudite essentially includes knowing about history, but since I’m German, e.g. American history is not part of core-proficiency in history. Besides, knowing about exact dates of battles etc., is not the essence of a true understanding of historical processes at all.” In similar fashion, S can introduce additional conceptual distinctions and subtleties which serve the function of making the inference to not-p questionable: given that E has the effect of establishing not-r, then conceptual distinctions may change the claim in q, thereby restoring confidence in p. In experiments by Wentura and Greve, subjects established congruency between TE and their actual abilities: ‘‘Accordingly, self-immunization means assigning less weight to those observables an individual believes himself to be poor at and more weight to those observables he believes himself to be good at. As a consequence, the immunized concept (‘erudite’) is maintained against a certain failure (‘‘I don’t know when Caesar was murdered”) by reshaping its connotation (‘‘It’s not historical education, but, say, geographical knowledge what counts”).” (Wentura & Greve, 2003, p. 31). Beliefs that are connected to observables only via a set of connecting beliefs about a criterial function of r for p invite selfdeception. S will hardly succeed in deceiving himself about the immediately observable fact that he has failed the history test. In general, q-type connecting beliefs are not as easy to rebut as r-type beliefs. They provide self-deceivers with room for adapting their definitions and for constructing subtleties and ambiguities that they have never stressed before which allows them to reject that the target-belief that p is finally defeated. In our special case, TE, the theory of ‘being erudite’, is notoriously under-determined. TE provides S with the space for re-interpretation that lies at the basis of the stability of self-deception and in practice avoids a regress: TE and q are answerable to evidence, but they contain complex, theory-based terms that are only remotely observational (Newen & Bartels, 2007). Therefore TE-adaptations are able to provide a relative stability for self-deceptive beliefs. S could for example also increase stability of self-deceptive adaptations by conceding to himself superior insights into what true eruditeness does comprise. 32 There is additional empirical evidence for a strong tendency of test-subjects to modify their standards of evaluation in a way that allows them to maintain their self-esteem. See cf. Beauregard and Dunning (2001). If we attribute properties to others like being intelligent or being good in mathematics, we do that according to an idiosyncratic standard that bears on our own abilities. If we are not good in mathematics we establish a low standard for being good in mathematics such that we can self-attribute to share that property if we have advanced competences, we develop a high standard for being good in mathematics such that almost no one besides ourselves passes it. These trait definitions have an egocentric standard and thereby allow us to keep our selfesteem. 33 We want to stress that we do not presuppose a rich concept of ‘theory’ here. A theory T in this case is merely a local set of beliefs with regard to a special subject matter SM. It involves no requirement of strong systematicity. C. Michel, A. Newen / Consciousness and Cognition 19 (2010) 731–744 741 4.2. Dual rationality as a constitutive element of self-deceptive pseudo-rationality Up to this point it is still not entirely obvious whether S is self-deceived. As yet we have developed no criterion as to when the revision of TE is irrational or self-deceptive. It can be rational to defend important beliefs against prima-facie evidence, even if this evidence is criterial according to the old version of TE. What would make S’s move problematic is if S endorses two inconsistent standards of belief revision at the same time. Wentura and Greve (2003, 2005) as well as Beauregard and Dunning (2001) focus on self-related adaptation, but do not consider self-/other-asymmetries. But given the PMD-condition (5)34 and the constraint that self-deceivers endorse standard-rationality, we predict that self-deception will typically involve a dual standard of rationality (where S does not notice the duality). The phenomenon of dual rationality consists in the following disposition to evaluate cases asymmetrically. A revision of the relevant belief-set to neutralize criterial evidence applies only to matters of subjective importance. In dual rationality, subjects will keep their rational standards in general and at the same time introduce a new standard of evaluation with respect to particular matters of subjective importance. With regard to self-esteem, this will typically result in first-/third-person asymmetries. Peter will continue to regard himself as erudite, but the next day during lunch he judges his boss Frank by the old version of TE when Frank shows an embarrassing gap in his knowledge of history. Evidence supports that we are more generous with our ego-centric definitions of desirable traits than we are with the general definition we apply to others. There is evidence for a general tendency in human reasoning that supports the plausibility of such dual rationality in selfdeceivers. Apparently, we do not only use two systems of reasoning but also develop dual standards of evaluation, e.g. when we evaluate human abilities. Beauregard and Dunning (2001) not only prove that there is an ego-centric bias in evaluating abilities like being intelligent but they also make clear that there are two standards of evaluation involved, namely an intersubjective social standard of evaluating mathematical abilities, (which can be characterized by the results of standardized tests) and the ego-centric standard that we use if we evaluate unreflectively. In addition, there is evidence that we can establish dual rationality intrapersonally. There is a strong line of research arguing that people use two systems of reasoning (Evans & Over, 1996; Kahneman & Frederick, 2002, 2005; Reyna & Brainerd, 1994; Sloman, 1996; Stanovich & West, 2000): a primitive, intuitive heuristic, associative reasoning system on the one hand and a cognitively demanding analytic, rule-based reasoning system on the other hand.35 The question remains whether there is evidence that we actually use dual standards internally for evaluating relevantly similar packets of information. Evidence for intrapersonal dual standards of evaluation is for example given by Dunning and Cohen (1992) and Beauregard and Dunning (1998), who have demonstrated that persons refer to different definitions in describing their own performances and the performances of others. Furthermore, if the dual reasoning systems are used for the same task, then we should notice intrapersonal conflicts, since the two systems work independently and in parallel. Hence, they should produce different, competing outcomes. In fact, there is evidence for competing outcomes since measureable monitoring processes are implicitly involved in cases of dual rationality (de Neys & Glumicic, 2008). This supports our analysis nicely, since we have evidence that dual rationality can be established intrapersonally: in the demanding subjectively relevant cases (which are typically first-person), the intuitive, associative reasoning is allowed to triumph over the monitoring process, while in the neutral third-person cases the implicit monitoring process allows the reflective, rule-based reasoning to dominate. The latter especially demands universal employment of rules, while the former remains implicit, involves an ego-centric standard of rationality in subjectively relevant cases and leads to dual rationality. Dual rationality will typically be observable in ‘automatic’ or unreflected adaptation and this is one reason for criticizing S’s activity from a rational point of view. In the case of ‘automatic’ TE-adaptation, S’s own dual standard will not be transparent to S. However, room should be left for the possibility that reflective self-deceivers seek to avoid a dual standard and begin to generalize the revised version of TE also over parallel cases.36 But a strong generalization of TE involves the danger of revealing self-deception. Theoretically, S can iterate self-deception but extending self-deception in iterating the process and creating more dual standards will make the process less economical and make it more and more unlikely that S’s can retain pconfidence. Thus, dual rationality remains characteristic for self-deception, but the dual standard must basically remain opaque to S. What makes TE-change in the above described fashion pseudo-rational (in addition to the typical dual standard of evaluation) is the fact that the change is ad hoc, i.e. it can be identified as made on the basis of the subjective importance of p instead of on the basis of the pursuit of truth. To follow a simplified Quinean standard, rational processes of revision are mind-to-worlddirected, i.e. the primary touchstone for the whole system is the external world. If adaptation-processes obviously violate this primacy and tend towards belief-immunization, the revision will be criticized for following a motivation external to the norms 34 Condition (5) states: If S had been neutral towards SM, the same evidence would have led S to accept not-p instead of p. The normal cognitive development seems to involve a process in which the domain of reasoning is first treated by associative processes before we develop more and more rule-based analytic reasoning. Furthermore, there is evidence that a specific mistaken evaluation in an everyday task called ‘base-rate neglect’ is a consequence of using the associative reasoning system. The effect is reduced if we acquire rule-based reasoning for that domain (Barbey & Sloman, 2007). We can diminish the impact of strong bias—that is constituted by ignoring relevant knowledge and relying on associative reasoning—by learning to reason in a rulebased way. 36 This strategy will include cases which are not immediately self-related into the scope of subjective importance. If not only Peter but also Mary failed the history-test, Peter might generalize TE and thereby also immunize Mary’s status of being an erudite woman against history-test evidence. This strategy is compatible with self-deception but it will be risky for Peter to pursue it too strongly since it direcly exposes revised TE to falsification. If Mary is honest and not willing to participate in collective self-deception, John will encounter immediate counter-evidence against the viability of revised TE. If Peter is insistent, he will be forced to iterate the process and, e.g. discredit Mary’s opinion. Typically, the iterated process will create a dual standard again. 35 742 C. Michel, A. Newen / Consciousness and Cognition 19 (2010) 731–744 of rationality. Moreover, in self-deception the resulting beliefs will typically be false. As a matter of fact, S will typically be wrong about p, since in most cases there is good reason why a connecting beliefs q is part of the standard form of TE. Dual rationality and the tendency towards irrationally motivated immunization justify the characterization of S’s maneuver as ‘pseudo-rational’. 4.3. Conclusion, advantages and scope of the model We have argued that self-deception consists in p-commitments on the basis of pseudo-rational adjustments of the beliefsystem. The adjustments invalidate criterial evidence against p. Self-deception typically involves a dual standard of evaluation. The proposed model reconciles two requirements: Self-deceivers are sufficiently rational and sensitive to evidence and they can manipulate their beliefs in the face of criterial evidence to the contrary. Self-deception can be characterized as a cognitively dynamic process of keeping or establishing belief in p without having to assume that S must believe not-p. As has been pointed out, pseudo-rational adaptation is particularly suitable to meet the strong counter-evidence requirement. But there is one more reason for seeing rational reorganization as a necessary feature of self-deception. As a technique of auto-manipulation, motivated reorganization obviously differs from phenomena like attention control, selection of evidence, control of memory retrieval and control over contents of conscious thought as described by Talbott (1995) and van Leeuwen (2008). It can be questioned whether those revision-free phenomena of belief-control would not perform equally well against criterial counter-evidence in sufficiently rational subjects. We cannot discuss this here. But in our view, there is nevertheless one strong intuitive aspect of the phenomenon of self-deception that makes pseudo-rationality seem indispensable. As indicated in Section 3.1, we see it as part of the phenomenon that self-deceivers defend their target-beliefs in social interaction. If challenged by someone’s pointing out the weak evidence for p, a true self-deceiver will argue for his commitment rather than shy away from it and will defend his confidence. In many cases, self-deceivers might resist giving up p even more than normal unmotivated believers would. If we want self-deceivers to be able to resist rational belief-adaptation, there seems hardly a way around the rationalization strategy. By pseudo-rationality, motivated beliefs can be successfully anchored in rational discourse at least to a limited extent. Pseudo-rationality is not only a successful ‘technique’ to arrive at desired conclusions but also a requirement for self-deception among members of rational discourse. Several forms of auto-manipulation can be active at the same time and lend mutual support in S’s auto-manipulation. However, pseudorationality is necessary for self-deception in rational subjects in addition to lower levels of motivated informationprocessing. The ‘pseudo-rationality’-model satisfies all three strict criteria of adequacy: (i) the developed picture of the cognitive dynamics of self-deception is paradox-free. Pseudo-rational immunization of p against counter-evidence involves neither a dynamic nor a static paradox. (ii) The developed view describes a real psychological phenomenon, as the empirical evidence about the adaptation of trait-definitions and dual rationality suggests. (iii) The model explains how self-deception (in contrast to other forms of auto-manipulation) in rational subjects can operate against criterial evidence. Beyond the strict criteria of adequacy, the model provides ways of doing justice to a set of further intuitions and observations that are connected to self-deception. When engaging in more laborious pseudo-rational defense of p, subjects can feel that self-deception did not just happen to them, but that they have played an active part in a manipulative process. But, as Wentura’s and Greve’s studies indicate, pseudo-rational defense also may proceed in a rather automatic way. Conscious defense of p against p-hostile evidence seems possible, as long as (a) the pseudo-rational quality and dual rationality remain opaque to the subject and (b) as long as S need not think or recognize that his p-commitment is improperly founded.37 We prefer not to characterize self-deception in contrast to other forms of auto-manipulation by requiring it to be intentional. From our point of view, the question of whether self-deception is intentional in any good sense is an issue of secondary importance. All we require for self-deception is pseudo-rational adaptation within a belief-system, no matter whether adaptation is automatized or not. All that has to remain opaque to S is the pseudo-rational quality of his p-commitment. Since the model describes vivid cognitive dynamics, it can account for psychological tension as a typical feature of selfdeception. However, in many cases of self-deception, psychological tension may be low or absent, as Mele suggests. The likely reason for an absence of tension is that self-deception—like any mental activity—becomes habitual. If subjects develop idiosyncratic patterns of interpretation for certain subject matters, they become ‘trained’ self-deceivers. A repeated application of the revised theory of ‘eruditeness’ may reduce tension, e.g. in future history-test failures. Moreover, the fact that subjects register criterial counter-evidence explains why self-deceivers can sometimes be only slightly surprised when self-deception breaks down and they acknowledge that they got it wrong. Finally, the ‘pseudo-rationality’ model even provides a parallel between self- and other-deception: Making p acceptable for others requires, in principle, abilities and strategies similar to those we use for making p committable for ourselves. The model is developed as a model of self-deceptive belief-defense. It is important to note that it can cover self-deceptive belief-acquisition and self-deceptive belief-retention as well. If the matter of p has subjective importance for S, S can develop pconfidence against a not-p environment by a similar strategy in both cases. It is no necessary condition that a belief that p must have been established in S beforehand in order for S to exploit the strategy of pseudo-rational adaptation with regard 37 Talbott (1995) has indicated how far self-deception might involve non-paradoxical intentional aspects that differ from straightforwardly self-deceptive intentions, i.e. intentions with the literal self-deceptive content ‘‘deceive yourself!”. Our account is open for like notions of intentionality. C. Michel, A. Newen / Consciousness and Cognition 19 (2010) 731–744 743 to an issue of subjective importance. Pseudo-rational adaptation can be used to manipulate the evidential value of any data to fit any target-belief, no matter whether this target-belief is to be retained or yet to be established. In both cases self-deception is a rationalization of commitments that S is making on motivational grounds. In an ego-centric belief-system, establishing new beliefs about matters of subjective importance can for example also serve as a stabilization of the self-image, to the effect that single beliefs are acquired in order to stabilize and defend an already established set of motivational beliefs. We have dispositions to embrace certain types of new beliefs and to commit ourselves to p on motivational grounds (e.g. for the purpose of stabilizing an already existing set of motivated beliefs), even if S has never considered p before. If S has a motivation to accept p and p enters consideration, contradicting evidence can be invalidated and p-confidence can be developed on the basis of pseudo-rational adaptation. To demonstrate the flexibility of pseudo-rational adaptation, let’s look at an example of belief-acquisition via Tx-change in which S acquires a belief to the contrary of what he had believed before: John, an advanced graduate student, has never considered himself an even reasonably good philosopher. He has always felt behind his peers in most of the aspects that commonly count as relevant for having any career-prospects in the field. Therefore, he himself has believed for almost all of his graduate period that academic philosophy is definitely not his terrain. However, as John has almost finished his studies, questions about his future become salient and start pressing him. But John is far too lazy to even think about starting anything new and he is basically unwilling to acknowledge that he did not make the right kind of investment for his life. But John also knows well that even if the fact of someone’s being rather weak as a philosopher combined with his having poor talents in explaining philosophical problems to students and his being a fairly misanthrope personality does not make it impossible to get a philosophy job, it will nevertheless for sure mean having an awful life. In his situation, John is strongly motivated to acquire the belief that he is actually an exceptional philosopher, endowed with basically just what it takes to become a good academic philosopher. And in fact, he does change his opinion about what is criterial for someone’s being likely to become a successful academic philosopher and sincerely comes to believe that his great days in philosophy are about to come quite soon. It is possible for him to do so by revising his old theory about the issue, convincing himself that his former belief that he has no talents in the field was the result of a corruption of his mind by misguided standards and unreliable criteria for good philosophy, that his talents and insights have been overlooked because his individual approach does not conform with the expectations, that his achievements have not been properly evaluated due to the mainstream’s lack of sensibility for his approaches. He develops the belief that the whole education- and career-system builds on false priorities, that fundamental mistakes are involved in how philosophy is pursued in departments nowadays, etc. This way John brings his plan into harmony with the evidence about his performance, talents and personal traits. In the absence of job alternatives, John confidently decides to pursue a career as a philosophy professor. Motivated belief-change via change of background-theories is common, and it is the involved pseudo-rationality including dual rationality that will decide whether S’s belief-change is finally to be evaluated as self-deceptive. In principle, the technique of pseudo-rational T-revision seems capable of transforming the evidential status of any information towards a desired result, independent of the subject’s current state of belief and the source of motivation. Neutral data can be transformed to E+ or E , in addition, both can be invalidated to E0 and even E turned into E+ or vice versa38. This flexibility can be used to account for the problematic case of twisted self-deception39 as well. Peter deceives himself into believing that his wife Mary is unfaithful despite having rather good evidence for her faithfulness. It is difficult to clarify the motivational structure of such cases from the armchair, since the acquired belief seems undesirable and it is controversial as to how such cases are correctly characterized from a psychological point of view. But regardless of how the motivational structure of these cases is to be analyzed, Peter could basically proceed as described above. Peter’s revision of his view about what types of behavior indicate unfaithfulness can attribute to neutral or weak data a high evidential status for ‘‘unfaithful”, (e.g. Mary’s admiring a common friend and her engaging in intimate conversation). As a consequence of his new evaluation of evidence, Mary’s in fact unaltered behavior will now be perceived by Peter as indicating unfaithfulness. Another women’s similar behavior would not, since Peter, in line with his old theory, perceives her behavior not as diagnostic of unfaithfulness. Admittedly, twisted cases are tricky but the ‘pseudo-rationality’-model can offer a strategy in case such twisted self-deceivers are sufficiently rational and evidence-sensitive. It seems that the technique of pseudo-rational adaptation is basically available for all kinds of self-deceptive endeavors. Thus, the ‘pseudo-rationality’-model seems to provide an attractive account of self-deceptive auto-manipulation. Acknowledgments The authors wish to thank Alex Byrne, Amber Griffioen, Eduardo García Ramírez, Gottfried Vosgerau and Louise RöskaHardy for very helpful comments on earlier drafts and for discussions on the topic that have led to improvements of the paper. The paper has also benefited much from the remarks of two anonymous reviewers. References Audi, R. (1982). Self-deception, action and will. Erkenntnis, 18(2), 133–158. 38 An example of the latter is Mele’s example of ‘positive misinterpretation’ where Sid’s interpretation turns Roz’ obvious and unambiguous refusal to date him into an encouragement (Mele, 2001, p. 26). 39 The term is introduced by Mele (1999). 744 C. Michel, A. Newen / Consciousness and Cognition 19 (2010) 731–744 Audi, R. (1988). Self-deception, rationalization, and reasons for acting. In A. O. Rorty & B. P. McLaughlin (Eds.), Perspectives on self-deception (pp. 92–120). Berkeley: University of California Press. Barbey, A. K., & Sloman, S. A. (2007). Base-rate respect: From ecological rationality to dual processes. Behavioral and Brain Sciences, 30, 241–297. Beauregard, K. S., & Dunning, D. (1998). Turning up the contrast: Self-enhancement motives prompt egocentric contrast effects in social judgments. Journal of Personality and Social Psychology, 74, 606–621. Beauregard, K. S., & Dunning, D. (2001). Defining self-worth: Trait self-esteem moderates the use of self-serving trait definitions in social judgment. Motivation and Emotion, 25(2), 135–161. Bermúdez, J. (1997). Defending intentionalist accounts of self-deception. Behavioral and Brain Sciences, 20, 107–108. Bermúdez, J. (2000). Self-deception, intentions, and contradictory beliefs. Analysis, 60(4), 309–319. Davidson, D. (1985). Deception and division. In E. LePore & B. McLaughlin (Eds.), Actions and events (pp. 138–148). New York: Basil Blackwell. Davidson, D. (1998). Who is fooled? In J.-P. Dupuy (Ed.), Self-deception and paradoxes of rationality (pp. 1–18). Stanford: CSLI. De Neys, W., & Glumicic, T. (2008). Conflict monitoring in dual process theories of thinking. Cognition, 106, 1248–1299. Dunning, D., & Cohen, G. L. (1992). Egocentric definitions of traits and abilities in social judgment. Journal of Personality and Social Psychology, 63, 341–355. Evans, J. St. B. T., & Over, D. E. (1996). Rationality and reasoning. Hove, UK: Psychology Press. Kahneman, D., & Frederick, S. (2002). Representativeness revisited: Attribute substitution in intuitive judgement. In T. Gilovich, D. Griffin, & D. Kahneman (Eds.), Heuristics and biases: The psychology of intuitive judgement (pp. 49–81). Cambridge, MA: Cambridge University Press. Kahneman, D., & Frederick, S. (2005). A model of heuristic judgement. In K. J. Holyoak & R. G. Morrison (Eds.), The Cambridge handbook of thinking and reasoning (pp. 267–293). Cambridge, MA: Cambridge University Press. McLaughlin, B. P. (1988). Exploring the possibility of self-deception in belief. In A. O. Rorty & B. P. McLaughlin (Eds.), Perspectives on self-deception (pp. 29–62). Berkeley: University of California Press. Mele, A. R. (1997). Real self-deception. Behavioral and Brain Sciences, 20, 91–102. Mele, A. R. (1999). Twisted self-deception. Philosophical Psychology, 12, 117–137. Mele, A. R. (2000). Self-deception and emotion. Consciousness & Emotion, 1(1), 115–137. Mele, A. R. (2001). Self-deception unmasked. Princeton: Princeton University Press. Newen, A., & Bartels, A. (2007). Animal minds and the possession of concepts. Philosophical Psychology, 20(3), 283–308. Nisbett, R., & Ross, L. (1980). Human inference: strategies and shortcomings of social judgments. Englewood Cliffs, N.J.: Prentice Hall. Pears, D. (1991). Self-deceptive belief formation. Synthese, 89, 393–405. Rey, G. (1988). Toward a computational account of akrasia and self-deception. In A. O. Rorty & B. P. McLaughlin (Eds.), Perspectives on self-deception (pp. 264–296). Berkeley: University of California Press. Reyna, V. F., & Brainerd, C. J. (1994). The origins of probability judgment: A review of data and theories. In G. Wright & P. Ayton (Eds.), Subjective probability (pp. 239–272). Oxford, England: John Wiley & Sons. Sahdra, B., & Thagard, P. (2003). Self-deception and emotional coherence. Minds and Machines, 13, 213–231. Sanford, D. H. (1988). Self-deception as rationalization. In A. O. Rorty & B. P. McLaughlin (Eds.), Perspectives on self-deception (pp. 157–169). Berkeley: University of California Press. Sloman, S. A. (1996). The empirical case for two systems of reasoning. Psychological Bulletin, 119, 3–22. Stanovich, K. E., & West, R. F. (2000). Individual differences in reasoning: Implications for the rationality debate. Behavioral and Brain Sciences, 23, 645–726. Szabó Gendler, T. (2007). Self-deception as pretense. Philosophical Perspectives, 21, 231–258. Talbott, W. J. (1995). Intentional self-deception in a single coherent self. Philosophy and Phenomenological Research, 55, 27–74. Trope, Y., & Liberman, A. (1996). Social hypothesis testing: Cognitive and motivational mechanisms. In E. Higgins & E. Kruglanski (Eds.), Social psychology: A handbook of basic principles (pp. 239–270). New York: Guilford Press. Van Leeuwen, N. (2007). The product of self-deception. Erkenntnis, 67, 419–437. Van Leeuwen, N. (2008). Finite rational self-deceivers. Philosophical Studies, 139, 191–208. Wentura, D., & Greve, W. (2003). Who want to be. . . Erudite? Everyone! Evidence for automatic adaptation of trait definitions. Social Cognition, 22(1), 30–53. Wentura, D., & Greve, W. (2005). Assessing the structure of self-concept: Evidence for self-defensive processes by using a sentence priming task. Self and Identity, 4, 193–211.

Consciousness and Cognition 22 (2013) 1074–1081 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Turning back the hands of time: Autobiographical memories in dementia cued by a museum setting Amanda N. Miles ⇑, Lise Fischer-Mogensen, Nadia H. Nielsen, Stine Hermansen, Dorthe Berntsen Department of Psychology and Behavioral Sciences, Center on Autobiographical Memory Research, Aarhus University, Denmark a r t i c l e i n f o Article history: Received 23 January 2013 Available online 15 August 2013 Keywords: Autobiographical memory Aging Dementia a b s t r a c t The current study examined the effects of cuing autobiographical memory retrieval in 12 older participants with dementia through immersion into a historically authentic environment that recreated the material and cultural context of the participants’ youth. Participants conversed in either an everyday setting (control condition) or a museum setting furnished in early twentieth century style (experimental condition) while being presented with condition matched cues. Conversations were coded for memory content based on an adapted version of Levine, Svoboda, Hay, Winocur, and Moscovitch (2002) coding scheme. More autobiographical memories were recalled in the museum setting, and these memories were more elaborated, more spontaneous and included especially more internal (episodic) details compared to memories in the control condition. The findings have theoretical and practical implications by showing that the memories retrieved in the museum setting were both quantitatively and qualitatively different from memories retrieved during a control condition. Ó 2013 Elsevier Inc. All rights reserved. 1. Introduction It is well established that as we age, we begin to show deficits in autobiographical memory. These deficits are even more pronounced in individuals with dementia. While there is an evident drop in the frequency of specific autobiographical memories and a diminishing of episodic details in aging generally, this loss is considerably more pronounced in dementia (Seidl, Lueken, Thomann, Geider, & Schröder, 2011). At the same time several studies show that the distribution of memories across the lifespan seems to remain relatively similar between older adults with dementia and older adults experiencing no signs of dementia in that both show a dominance of memories from young adulthood, also known as the reminiscence bump (Addis & Tippett, 2004; Fromholt & Larsen, 1991; Fromholt et al., 2003; see Seidl et al., 2011, for review). A potential consequence of the autobiographical memory deficits seen in dementia is a loss of identity. This is supported by Addis and Tippett (2004) who found a relationship between some measures of autobiographical memory impairment in dementia and changes in the quality and strength of identity. Complaints from both patients and caregivers also often describe a loss of the person they once knew. A diminishing connection to the past could interrupt the sense of continuity established by the self or one’s identity across the lifespan, as well as diminishing the richness of a person’s past and current quality of life (Conway & Pleydell-Pearce, 2000). ⇑ Corresponding author. Address: Department of Psychology and Behavioral Sciences, Center on Autobiographical Memory Research – CON AMORE, Aarhus University, Bartholins Alle 9, 8000 Aarhus C, Denmark. E-mail address: amanda@psy.au.dk (A.N. Miles). 1053-8100/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.concog.2013.07.008 A.N. Miles et al. / Consciousness and Cognition 22 (2013) 1074–1081 1075 As mentioned previously, event specificity is lacking in many of the memories retrieved by Alzheimer’s disease (AD) patients. Semantic memory or semantic details seem to be left mostly intact until the later stages of the disease while access to the richer episodic details associated with autobiographical memories deteriorates over the course of the disease (Seidl et al., 2011). Episodic details are details pertaining to a specific time and location. Such details are important for a sense of recollecting a specific past event (Levine et al., 2002). Memory retrieval as most frequently studied to date has focused on voluntary (i.e. strategic and controlled) retrieval. However, as dementia progresses the loss of executive functioning makes this type of retrieval more difficult if not impossible. The difficulty in accessing the specific, episodic details of an event may be attributed to the loss of executive functioning/control processes evident in general aging and dementia. Berntsen (2009) postulates that situational cues within the current environment may have a high cue-item discriminability (Rubin, 1995) and therefore lead to involuntary (spontaneous) retrieval of a specific memory, circumventing the strategic (voluntary) retrieval heavily reliant on executive functioning. This is consistent with findings from a diary study showing no age-related deficits in accessing specific episodic memories during involuntary in contrast to voluntary recall (Schlagman et al., 2009). If this is indeed the case, dementia patients may benefit from detailed cues from their past, enhancing the possibility for spontaneous retrieval of episodic details. To date, few studies have attempted to examine situation cued autobiographical memories in dementia patients. Recently, El Haj, Fasotti, and Allain (2012) found that autobiographical memories in dementia patients were retrieved faster and were more specific when retrieved in a music condition than memories evoked in silence. The music played during the session was chosen by the participant (or his or her relatives), providing the participant with cues directly related to their own pasts or goals. In doing so, the experimenters may have created an instance of high cue-item discriminability (e.g., high correspondence between the cue and an event in memory) and increased the retrieval support for the individuals in the music condition. Similarly, some assisted home for living facilities offer reminiscence therapy in which demented older are encouraged to retrieve autobiographical memories in response to items that they may associate with their past (e.g., photographs, old objects). However, little is known as to the effects on autobiographical memory. Here we try to provide a situation with extreme retrieval support by using a museum setting as our experimental condition. We compare frequency and quality of autobiographical memories retrieved in this condition to another (control) condition with similar, but less familiar cues. As such, we predict that the cues given to participants in the experimental condition will evoke more autobiographical memories than cues in the control condition. Moreover, the memories that are retrieved in the experimental condition will be characterized by a greater level of episodic details, suggesting more recollection of specific past events. The experimental setting was a house in the open-air museum of urban history and culture, The Old Town, in Aarhus Denmark. The house consisted of a kitchen, and two adjacent living rooms, all furnished and decorated in early 20th century style. Thus, the Old Town allowed for total immersion into a holistic and historically authentic environment, which reconstructs the cultural and material context of our participants’ youth, allowing cuing of memories along a number of dimensions, including all modalities (i.e., odor, auditory, tactile, etc.) and the reenactment of activities (e.g., singing old songs, setting the coffee table with tablecloth and porcelain, holding and touching the old objects, eating cookies baked from an old recipe), in a style our participants would recognize from their childhood and youth, see Fig. 1. Since 2004, this house in The Old Town has been used for reminiscence therapy with elderly demented people. We therefore took advantage of an already established space and group of reminiscence coordinators. The control condition was a meeting room in an aging activity center with modern furniture and decorations. Fig. 1. Participants in reminiscence therapy at The Old Town. 1076 A.N. Miles et al. / Consciousness and Cognition 22 (2013) 1074–1081 2. Methods 2.1. Participants Fourteen participants were recruited for the study. They were separated into five groups according to five different local community centers of which they were members. Two participants were absent from the first session due to sickness and were excluded from the rest of the study. Therefore, 12 participants were included in the study (M = 87.17 years, SD = 5.18; see Table 1). All participants were recruited via local elderly centers where the local dementia coordinator helped select each participant. While there were no age or gender restrictions for inclusion within the study, there were coincidentally only female participants in the current study. Each participant was diagnosed with dementia (Alzheimer’s disease, vascular, or a combination of both) or they exhibited a reduction in cognitive ability by scoring 23 or less on the Mini Mental State Examination (MMSE; M = 20.58, Range: 14 25; see Table 1) (Folstein, Folstein, & McHugh, 1975). The formal diagnoses of the patients presenting with dementia were made prior to the current study, and these were made via the Danish health care system involving an extensive series of examinations conducted by the patient’s general practitioner and specialists in the area of neurology and psychiatry. Otherwise, the MMSE was used to determine substantial cognitive impairment likely qualifying for a diagnosis of dementia in a comprehensive dementia evaluation. The cut-off used here (MMSE < 24) is conservative yielding less than 1% false positives. Indeed, individuals with higher scores often qualify for a dementia diagnosis (O’Bryant et al., 2008). The MMSE test was administered by a psychologist before either of the sessions took place. While each participant displayed a cognitive deficit, a subjective assessment was made by an assisted home for living staff member as to the participant’s eligibility to participate in the study. Participants were judged to be healthy enough to complete the tests and interviews associated with the study. None of the participants had experienced the loss of a relative within the two months prior to testing and all were physically mobile in order to participate in the session at the Old Town. In order for participants to be included in the study, no other major psychiatric diagnoses were present. 2.2. Design A within subjects design was adopted, in which participants took part in two different sessions. The sessions also took place in two different locations: the Old Town (the experimental condition) and an aging activity center (a control condition in terms of an everyday setting). The order in which the two sessions took place was counterbalanced across groups. 2.3. Materials The cues to be used in the Old Town session were chosen to represent both objects from everyday life and objects for special occasions (e.g., Christmas and confirmation). The objects were very common, and it was therefore assumed that the participants would have encountered the objects previously. Once nine cues for the Old Town were selected, counterparts from modern time were found. These cues represented the same everyday life and special occasion objects but from a more modern period. See Table 2 for the full list of the cues presented in each condition. 2.4. Verbal fluency task Other than the MMSE, participants were also given a category fluency task (Lezak, Howieson, Bigler, & Tranel, 2012; Koroza & Cullum, 1995). Three categories were chosen: animals, things that make people happy and things that make people Table 1 Participant demographics, MMSE and Verbal Fluency scores. P1 P2 P3 P4 P5 P6 P7* P8* P9* P10* P11 P12 Age MMSE VF-animal VF-happy VF-sad 86 88 86 88 76 87 87 82 85 93 96 92 M 87.17 23 20 22 14 22 21 19 25 23 16 20 22 M 20.58 11 8 9 4 5 11 8 5 17 3 8 9 M 8.17 9 7 4 3 4 6 9 4 8 5 5 8 M 6.00 8 6 5 2 1 4 4 2 3 3 0 4 M 3.50 SD 5.18 SD 3.09 SD 3.81 SD 2.13 Note: The participants which were formally diagnosed with dementia prior to the study have been indicated with an asterisk (*). SD 2.20 A.N. Miles et al. / Consciousness and Cognition 22 (2013) 1074–1081 1077 Table 2 Cues presented in both the Old town and Activity center conditions. The Old Town Activity center Direct cues Rick’s coffee substitution Old-fashioned phone (one knob to the central) Christmas decoration (Christmas heart in paper) Scrap books (including scraps) Marbles Small slate and chalk (for school) Hand-held whisk Old-fashioned confirmation card White knitted hand towel for the kitchen Two famous old-fashioned songs: ‘‘Jens Vejmand’’ (1905) and ‘‘Jeg ved en lærkerede’’ (1924) Indirect cues Organic coffee Modern phone (cell phone) Modern Christmas decoration (Santa elf) Collectible cards French fries game in plastic Calculator Electric handmixer Modern confirmation card Black plain terry cloth hand towel for the kitchen Two famous modern songs: ‘‘Kald det kærlighed’’ (1993) and ’’Smuk som et stjerneskud’’ (2000) Old-fashioned china coffee pot (with coffee made in an old-fashioned way) and cream Old-fashioned vanilla-flavored cookies baked in the Old Town’s bakery The reminiscence employee with old-fashioned clothes from 1910 Modern thermos for coffee (with coffee made on a coffee-machine) and low fat milk Cookies with M&M’s from the supermarket The reminiscence employee with normal modern clothes Note: There were a few irregularities in the cues between Group 1 and the rest of the groups. A minor change was that the participants were presented with a Christmas heart in crochet instead of paper. A meat mincer and a hand blender were also presented instead of the hand-held whisk and the electric handmixer the other groups were presented. This particular change was made due to the unwieldiness of the meat mincer. Another minor change involved the snacks the first group received. Instead of old-fashioned vanilla-flavored cookies, they enjoyed crepes. Unfortunately, the crepes were too messy and the switch was made to cookies. All of these changes were made after the first group in order to improve the cues for the study. sad. While the animal category was chosen as a more traditional semantic fluency task, the other two were chosen as they have previously been shown to identify people with severe depression (Fuld, 1980), see Table 1. 2.5. Procedure Each of the 5 groups completed a session in the Old Town and a session in the activity center (as well as four interviews, which will not be presented here) which were separated by approximately one week. The order of the two sessions was counterbalanced across groups with half the groups participating in the Old Town session first, followed by the activity center, while the other half of the groups participated in the activity center session first, and followed by the Old Town. There were no order effects observed (ps > .18). One group showed an advantage for recognition, but this advantage was seen to the same extent in both conditions and thus not reflecting an effect of order. Present at every session were the 2–3 participants, the interviewer, the contact person or dementia coordinator from the local assisted homes for living, and two research assistants. Only the participants and the interviewer took active part in the session while the others observed. The sessions in the activity center took place in a neutral, modern conference room with a table in the middle and chairs around. All items which were not from the current time period (e.g., old paintings) were removed from the room. The session in the Old Town took place in the living room, which had a table in the middle and a sofa and some chairs around it. There was a burning stove and no electrical light in the room. All furniture and decorations belonged to the museum and were original pieces deriving from the beginning of the 20th century (see Fig. 1). The entire setting was an authentic reproduction of an early 20th century home, including the house itself. The participants entered the house through the cobble stone backyard and were guided through the low-ceilinged house to the living room, where the session took place. The sessions in the Old Town and the control condition all had the same course. As the participants arrived and were seated, coffee and cookies were offered. After some introductory small talk the cues were passed around to each participant. The interviewer used open and investigative questions to each cue. For instance: ‘‘Do you know this one?’’ and made sure to ask all participants about the cue if they did not say anything themselves. The interviewer would also encourage the participants to tell more, for instance by saying: ‘‘I am not familiar with that. Could you please tell me some more?’’. The interviewer used the same procedure in both conditions and spoke the same in each condition. After the nine cues had been passed around, the people present (including the observers) sang two songs, after which the session ended. This order of the events mimicked the order of events that has generally been used during reminiscence therapy in the Old Town. We chose to follow this well-established procedure, which had shown to generally work well with demented individuals in this particular context – that is, made them feel at ease and at home in the situation. 2.6. Coding The transcripts of the sessions were coded with a coding scheme adapted from Levine et al. (2002). Each line was given a line number in the transcription. A line could consist of several sentences, as only a change in the person speaking produced a new line. Each line was coded separately. Each variable was coded as dichotomous unless otherwise specified. 1078 A.N. Miles et al. / Consciousness and Cognition 22 (2013) 1074–1081 Table 3 Means and standard deviations of the dependent variables, t-tests and p-values for two-tailed binomial tests. Old Town Memory Recall Recognition Autobiographical Internal External Spontaneity Time (implicit) Time (explicit) Place Elaborated Details * ** Control M SD M SD 120.83 91.92 23.08 99.58 12.08 88.58 66.00 6.67 6.33 14.92 15.93 56.87 55.02 9.03 48.38 7.29 45.41 41.38 4.08 4.62 11.49 14.45 102.67 73.58 22.58 80.58 6.83 76.67 53.42 5.92 4.92 13.42 13.83 65.47 59.24 10.02 56.42 6.73 54.81 48.96 6.71 6.30 13.74 13.37 t Binomial p-value 1.92 2.46* .16 2.16 3.42** 1.38 2.00 .61 .98 .85 2.12 =.039 =.039 =.774 =.039 =.003 =.003 =.003 =.146 =.146 =.388 =.146 p < .05. p < .01. p < .09. Every line was first coded for the presence of a memory. There were three different coding possibilities: 0 if there was no memory content, 1 if the line included recognition, and 2 if the line included recall. If the line was coded as presenting a memory, the line was then coded for the presence or absence of autobiographical memory. Only if the line included autobiographical memory material were the rest of the variables coded. The next four codings further distinguished the type of memory present in each line. Internal events were indicated if a memory included a description of an event that lasted less than a day. An internal detail would be indicated by one of the following characteristics consistent with Levine et al. (2002) previous operationalization: a specific event, an individual or individuals present besides the narrator, weather conditions, physical action/reaction, and emotional action/reaction. External semantic information was indicated if a memory included a description of general autobiographical knowledge, a repeated or general event or a prolonged condition (Levine et al., 2002)1. The next part of the coding concerned the time frame of the memory/memories in each line. A distinction was made between an explicit and implicit description of time. If a memory in the line mentioned a certain year or the age of the participant at the time of the memory (including datable life periods, e.g. confirmation, or historical periods, e.g. the day of the liberation of Denmark after World War II) then the time description was coded as explicit. A coding of an implicit time description would include non-datable but inferred or general timeframe, for instance season or time of year (including formal feasts like Christmas), month, weekday, and time of day. Age at the time of the event was coded if the exact age of the participant during the event was explicit in the line. Period of event was coded if the period of time for an event appeared in the line (i.e., before school age (0–5 years); school age (6 15 years); early adulthood (16 30 years); mid-adulthood (31 50 years); late adulthood (86 95 years); and recent period (86 95 years). Place of the event was coded as present if a memory in the line included an explicit description of one of the following: a city, street, building, room, part of a room, or place related to the residence of others or the participant. In addition, we coded for more elaborated details. These elaborated details were coded as present if a memory in the line included an elaborated characteristic of a person, event, game etc., which was not necessary for understanding the essential part of the memory. Finally, the events were coded for spontaneity. We defined spontaneity as a memory in the line which was produced spontaneously by the participant and not as a response to a direct or indirect question from someone. If an autobiographical memory that was previously described spontaneously was repeated later in the interview but contributed nothing new, then the event was coded as a repeated autobiographical memory. One group (approximately 20% of the participants and groups) was used for developing the current coding manual. The transcriptions of the sessions involving this group were coded by two independent raters. Transcriptions from yet another group were used to establish the interrater reliability between the two coders. The agreement between the raters was 87– 99% across the different variables. Since the raters established such a high level of agreement, they were allowed to code the remaining data independently. 3. Results Because of large variability across the twelve participants, we present the results in terms of two-sided binomial tests, see Table 3. In order to provide more information, the means, standard deviations, and t-tests for the dependent variables are also included in Table 3. The participants had more memories in The Old Town then in the control condition. We found that 10 of the 12 participants had more statements coded as dealing with memories in the Old Town condition than in the control condition, p = .039. A similar effect was found for statements indicating actual recall of remembered material in contrast to 1 If there was a line which included multiple internal or external details, we also coded these in addition to whether or not an internal or external detail were present. However, these occurrences were few and of little theoretical importance. We have therefore decided not to present or discuss them further. A.N. Miles et al. / Consciousness and Cognition 22 (2013) 1074–1081 1079 statements involving simple recognition, for which there were no differences between the Old Town and the control condition (Table 3). Importantly, more memories with autobiographical contents were produced during the sessions in the Old Town than in the control condition, in that, 10 of the 12 participants described more autobiographical memories in the Old Town than in the control condition, p = .039. The qualities and level of details of the memories also differed between the two conditions. As expected, the large majority (11 of 12) participants had more internal (i.e., episodic) details in the Old Town than in the control session, p = .003. A similar effect was found for the number of external details. Furthermore, participants seemed to spontaneously (with no probes) recall more memories in the Old Town session than in the control session. Of the 12 participants, 11 spontaneously retrieved more memories in the Old Town condition than in the control condition, p = .003. It should be mentioned that the participants did not say more in the Old Town session and that the interviewer also did not speak more in the Old Town than in the control session (ps > .10). Therefore, the differences seen between the two conditions were not due to more probing or support from the interviewer, or simply the participants speaking more in one condition than the other. It short, it appeared that the museum setting activated more memories that were autobiographical, contained event specific details and were remembered spontaneously. For examples of autobiographical memories occurring during the session in The Old Town, see Appendix A. The rest of the conversations that were not scored for memory content were comprised of informal chitchat or everyday conversation about the weather or their current lives and future plans. Additionally, though there were few dated events, we examined the temporal distribution of the events that were dateable. The pattern of results was similar between the two conditions with the majority of events occurring during childhood and early adulthood. A total of 92 (Old Town = 47, Control = 45) dateable events from the period 6 15 years of age and another 19 (Old Town = 12, Control = 7) from the period 16 30 years of age were mentioned by the participants. This is in comparison to the periods covering the age 31 95 where only four events in total were mentioned across conditions and participants v2(1) = 99.55, p < .0001. Due to the small sample size, we were unable to statistically examine differences between those with a formal diagnosis of dementia and those without. However, an inspection of the MMSE scores in Table 1 shows that these two groups appeared very similar on this test both concerning mean score and range of the MMSE (Mdiagnosed = 20.75, range 16 25; Mnon-diagnosed = 20.50, range 14 23). Rank order correlation with the MMSE scores and the memory variables in Table 2 showed high correlation coefficients, generally ranging .50–.70 across the variables for both the Old Town and the control conditions. Thus, the MMSE scores clearly appeared to be valid measures of degree of cognitive impairment in the present study. 4. Discussion The Old Town condition showed more memories recalled, more autobiographical memories, more internal (episodic) and external details. There were also more spontaneously recalled memories within the Old Town condition. Thus, overall, there was an improvement in the quality of the memories with the help of the cues provided in the Old Town condition. These findings agree with our hypotheses concerning cue-item discriminability and the spontaneity of the recall. Dementia patients experiencing a decline in executive function would have more difficulty in reaching specific episodic details. With additional retrieval support in the form of specific, multi-sensory, environmental cues with a relevance to their past, participants may circumvent a voluntary or generative mode of retrieval and spontaneously retrieve specific details. These findings agree with previous research showing an improvement in recall in both healthy younger and older adults with additional retrieval support in terms of verbal prompts (e.g., Levine et al., 2002) and in older people with dementia using music cues (El Haj et al., 2012). The present study goes beyond this previous work by providing a total immersion into a holistic and historically authentic environment which reconstructed the environmental context of the participants’ youth and facilitated cuing of memories along a number of dimensions, including all sensory modalities and the reenactment of specific activities. In contrast to previous work, our study also measured memory along a number of different dimensions. Our results show an increase in both the frequency and quality of the memories recalled during the Old Town session in comparison to the session in the control condition. This difference was not simply due to the participants speaking more in the Old Town session or to the interviewer saying more in the Old Town, because no differences were found with regard to number of utterances and words spoken in the two settings. Further, while there were not so many dateable events within the conversations in the present study, we examined the frequency of recorded events across the lifespan. The pattern of results provides further evidence of a temporal gradient in episodic or incident memory in dementia in terms of a marked dominance of memories from childhood and early adulthood in both the Old Town and the control condition (Greene, Hodges, & Baddeley, 1995; Moscovitch et al., 2005; Seidl et al., 2011). These data are consistent with current consolidation theories, such as the multiple trace theory, according to which there are more traces to these older memories as they are more rehearsed (Moscovitch et al., 2005, but see also Squire & Alvarez, 1995 for an alternative explanation). Consistent with previous work, we found an increase in the frequency of events in childhood and young adulthood. This was present in both the Old Town and the control condition. When various sampling methods are employed, autobiographical memories across the lifespan repeatedly have shown a similar temporal pattern, including in depressed and dementia patients (Fromholt et al., 2003). Typically, there is an increase in the number of reported events in young adulthood, during the years of identity formation, which researchers have named the reminiscence bump (Rubin, Wetzler, & Nebes, 1986). There are a few theoretical accounts for the 1080 A.N. Miles et al. / Consciousness and Cognition 22 (2013) 1074–1081 existence and robustness of the bump in young adulthood across varying cuing techniques and participant populations. One such account emphasizes the increased frequency of important transitional events. It is during late adolescence and early adulthood that many individuals experience leaving school, beginning a career, getting married, and having children; all of which may contribute to a person’s identity or centrality to a person’s lifestory further increasing the frequency of rehearsal and likelihood of later recall. The subsequent years to the bump period are characterized by their relative stability, and as such, there is a drop in the number of reported events. Therefore, the lack of differences between the two conditions is unsurprising. The findings from the present study should be evaluated with a number of limitations in mind. First, we had a very small sample size with a relatively wide range in the level of cognitive impairment as evaluated by the MMSE. The latter yielded a great individual variation in memory performance. A larger sample with a smaller range in cognitive impairment might have allowed us to detect even more differences between the two conditions. However, given that the MMSE scores correlated with our memory measures, we also feel that this variability has allowed us a potential insight into the immersions effectiveness on dementia patients; with patients with less cognitive impairment potentially benefiting the most. Second, only four of our participants had a clinical diagnosis of dementia, while the remaining participants were included in the study on the basis of their score on the MMSE. Although we used a conservative cut-off and although our non-diagnosed participants were clearly comparable to the ones with a formal diagnosis both with regard to the mean and the range on the MMSE, further research should aim at replicating and extending the present findings, using a diagnosed group only. Third, there is a general interest for researchers and clinicians about methods capable of dissociating between mild cognitive impairment (MCI) and dementia or predicting dementia from patients currently diagnosed with MCI. The focus of the current study did not touch on this area of interest, but future research utilizing the current paradigm could include a group with MCI as a comparison for the dementia patients we interviewed; potentially leading to a better understanding of the distinctions between MCI and dementia and the progression of the disease. Finally, another potential concern, which is a general drawback to the field of autobiographical memory, is that we have no indication as to the accuracy or frequency of rehearsal for the memories relayed by our participants. Future studies might address the veracity of the participants’ memories based on family accounts. We, however, have no reason to doubt the participants. On the contrary, it is quite likely that some of the memories discussed during the sessions in the current study are highly rehearsed events from each individual’s past. The current study employed a novel paradigm and setting to facilitate the retrieval of autobiographical memories. We replicate previous findings on episodic memory and normal aging and extend this research to dementia and to a more natural environmental setting and cuing paradigm. In spite of the small sample size, the results of the study are promising, by showing a consistent pattern of results along a number of different memory variables. In addition to demented individuals, we believe that the benefits garnered from additional retrieval support in authentic historical environments would also enhance retrieval for other groups having difficulty attaining specificity, including healthy older adults and potentially depressed patients as well. More broadly, the current study has implications for current practice. The techniques and reminiscence setting used in the current study may be applied to dementia patients and assisted home for living facilities. We find that providing retrieval support, in terms of naturalistic historical cues, allows participants to retrieve more specific and detailed memories from their past. In the later years of life, when reminiscence may play a larger role, an increase in quality of life through such reminiscence activities is generally expected, but the effects on autobiographical memory have not been documented before. Further, the present findings may also hold implications more specifically for reminiscence therapy. While memory books and other memory aids have been used in reminiscence work, findings from the present study suggest that immersion into a more holistic and historically authentic environmental cuing may be especially beneficial. Acknowledgments The authors would like to thank the Danish National Research Foundation (DNRF93) and the Danish Council for Independent Research: Humanities for funding. We would also like to thank Tina Jeppesen for her valuable contributions to several phases of the project and the coordinators at each of the local activity centers: Mie Jensen, Susanne Bennekou, Charlotte Sørensen, Helle Caspersen, Hanne Sørensen, and Mette Harby. Finally, we would like to thank Den Gamle By (The Old Town), Lotte Kofoed, Henning Lindberg, Tove Engelhardt Mathiassen and Birgitte Kryger for their continued support and advice. Appendix A. Examples of spontaneously reported memories in the Old Town (translated by the authors, names are fictitious) 2.1. Example 1 In the control condition, Bertha only responds occasionally when there is a conversation about toys. In the Old Town, a scrapbook is circulated around to the participants and a conversation about toys is followed by a memory filled with episodic detail. A.N. Miles et al. / Consciousness and Cognition 22 (2013) 1074–1081 1081 Bertha said: I had a doll that was so .. so big (she holds her hand around 80 cm above the floor to show how big the doll was). And uh. And so it had a porcelain head, and I was so pleased with it, but/then it was broken, and so uh .. Dad uh .. bought a new one, I got it for my birthday. I was so glad it that I danced around with it, and then I fell of course, and then it went to pieces. And then I got a rap on the buttocks. We had a small walking stick one my brothers had received as a birthday gift. And so he bought it, and if we were not completely obedient we got a few raps with it. 2.2. Example 2 In the control condition, another woman (Astrid) tells about not having a phone at home, and that it was possible to borrow a phone at the phone exchange, when she is shown a modern cell phone. The Old Town has an old phone displayed. Astrid tells again about not having a phone at home, but she also adds that she was a maid and functioned as an operator at a telephone exchange. And a memory of a very specific and dramatic day is described. Astrid said: ... As a young girl I worked at a telephone exchange .. and took .. took care of the phone and plugged in all those who wanted to be connected with .. with so and so. ... And then they would ask for such and such numbers, and so I had some such plugs connected to wires, and then the number, which the person asked for, was plugged in and then you said ‘‘Please’’ ..Interviewer: Could you listen?Astrid: (laughs) Yes, one might. ... Well, I must be honest .... There was .. uhh a bad fire. And it was .. it was my in-law’s house. ... And then of course there was a call for the fire-brigade .. and .. and fire engines and .. and the call went through where .. where I sat. But I did not know .. that it was exactly that house that was burning .... So I forwarded the call .. and so .. that way so .. later I got to know that it was my in-law’s house that was burning. References Addis, D. R., & Tippett, L. (2004). Memory of myself: Autobiographical memory and identity in Alzheimer’s disease. Memory, 12, 56–74. Berntsen, D. (2009). Involuntary autobiographical memories. An introduction to the unbidden past. Cambridge, England: Cambridge University Press. Conway, M., & Pleydell-Pearce, C. W. (2000). The construction of autobiographical memories in the self-memory system. Psychological Review, 107, 261–288. Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). ‘‘Mini-mental state’’: A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12, 189–198. Fromholt, P., & Larsen, S. F. (1991). Autobiographical memory in normal aging and primary degenerative dementia (dementia of Alzheimer type). Journal of Gerontology, 46, P85–P91. Fromholt, P., Mortensen, D., Torpdahl, P., Bender, L., Larsen, P., & Rubin, D. (2003). Life-narrative and word-cued autobiographical memories in centenarians: Comparisons with 80-year-old control, depressed, and dementia groups. Memory, 11, 81–88. Fuld, P. A. (1980). Guaranteed stimulusprocessing in the evaluation of memory and learning. Cortex, 16, 255–272. Greene, J. D. W., Hodges, J. R., & Baddeley, A. D. (1995). Autobiographical memory and executive function in early dementia of Alzheimer type. Neuropsychologia, 33, 1647–1670. El Haj, M., Fasotti, L., & Allain, P. (2012). The involuntary nature of music-evoked autobiographical memories in Alzheimer’s disease. Consciousness and Cognition, 21, 238–246. Kozora, E., & Cullum, MC. (1995). Generative naming in normal aging: Total output and qualitative changes using letter and semantic constraints. The Clinical Neuropsychologist, 9, 313–320. Lezak, M. D., Howieson, D. B., Bigler, E. D., & Tranel, D. (2012). Neuropsychological assessment (5th ed.). New York: Oxford University Press. Moscovitch, M., Rosenbaum, R. S., Gilboa, A., Addis, D. R., Westmacott, R., Grady, C., et al (2005). Functional neuroanatomy of remote episodic, semantic and spatial memory: A unified account based on multiple trace theory. Journal of Anatomy, 207, 35–66. Levine, B., Svoboda, E., Hay, J. F., Winocur, G., & Moscovitch, M. (2002). Aging and autobiographical memory: Dissociating episodic from semantic retrieval. Psychology and Aging, 17, 677–689. O’Bryant, S. E., Humphreys, J. D., Smith, G. E., Ivnik, R. J., Graff-Radford, N. R., Petersen, R. C., et al (2008). Detecting dementia with the Mini-Mental State Examination (MMSE) in highly educated individuals. Archives of Neurology, 65(7), 963–967. Rubin, D. C. (1995). Memory in oral traditions. The cognitive psychology of epic, ballads, and counting-out rhymes. New York: Oxford University Press. Rubin, D. C., Wetzler, S. E., & Nebes, R. D. (1986). Autobiographical memory across the life span. In D. C. Rubin (Ed.), Autobiographical memory (pp. 202–221). New York, NY: Cambridge University Press. Schlagman, S., Kliegel, M., Schulz, J., & Kvavilashvili, L. (2009). Differential effects of age on involuntary and voluntary autobiographical memory. Psychology and Aging, 24, 397–411. Seidl, U., Lueken, U., Thomann, P. A., Geider, J., & Schröder, J. (2011). Autobiographical memory deficits in Alzheimer’s disease. Journal of Alzheimer’s Disease, 27, 567–574. Squire, L. R., & Alvarez, P. (1995). Retrograde amnesia and memory consolidation: A neurobiological perspective. Current Opinion in Neurobiology, 5, 169–177.

Consciousness and Cognition 42 (2016) 41–50 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog The influence of retrieval practice on metacognition: The contribution of analytic and non-analytic processes Tyler M. Miller a,⇑, Lisa Geraci b a b Department of Psychology, South Dakota State University, United States Department of Psychology, Texas A&M University, United States a r t i c l e i n f o Article history: Received 22 July 2015 Revised 4 March 2016 Accepted 6 March 2016 Keywords: Metacognition Overconfidence Retrieval practice Inferential processes a b s t r a c t People may change their memory predictions after retrieval practice using naïve theories of memory and/or by using subjective experience – analytic and non-analytic processes respectively. The current studies disentangled contributions of each process. In one condition, learners studied paired-associates, made a memory prediction, completed a short-run of retrieval practice and made a second prediction. In another condition, judges read about a yoked learners’ retrieval practice performance but did not participate in retrieval practice and therefore, could not use non-analytic processes for the second prediction. In Study 1, learners reduced their predictions following moderately difficult retrieval practice whereas judges increased their predictions. In Study 2, learners made lower adjusted predictions than judges following both easy and difficult retrieval practice. In Study 3, judge-like participants used analytic processes to report adjusted predictions. Overall, the results suggested non-analytic processes play a key role for participants to reduce their predictions after retrieval practice. Ó 2016 Elsevier Inc. All rights reserved. 1. Introduction Metacognition refers to people’s awareness, knowledge, and control of their cognitive abilities (Flavell, 1979). Metacognitive awareness, or monitoring, can be measured in a variety of ways (e.g., Nelson & Narens, 1990). For example, the ability to monitor one’s memory can be measured using prospective judgments, in which the person makes a prediction about future memory performance, and using retrospective judgments, in which the person makes a judgment about past performance (Arbuckle & Cuddy, 1969; Hart, 1965). In both cases, the accuracy of these metacognitive monitoring judgments can be assessed by determining how well these judgments correspond with past or future performance using difference scores, gamma correlations, and receiver operating characteristics among other measures (Cheng, 2010; Fleming & Lau, 2014; Nelson, 1996). What is more difficult to determine, though, is how people make these judgments and what information they use to assess how well they will or have performed on a cognitive test (for one account of what information people use to make monitoring judgments see Nelson & Dunlosky, 1991; for a review see Schwartz, 1994). The cues people use to make monitoring judgments can be broadly categorized as either analytic (or theory-based) or nonanalytic (experience-based; Koriat, Bjork, Sheffer, & Bar, 2004; Kelley & Jacoby, 1996). Kelley and Jacoby suggested that analytic cues arise from people’s beliefs or theories about memory and the factors that influence memory performance, whereas nonanalytic cues arise from people’s subjective experiences of performing the task. Similarly, Koriat (1997) offered ⇑ Corresponding author at: Department of Psychology, South Dakota State University, Brookings, SD 57007, United States. E-mail address: tyler.miller@sdstate.edu (T.M. Miller). http://dx.doi.org/10.1016/j.concog.2016.03.010 1053-8100/Ó 2016 Elsevier Inc. All rights reserved. 42 T.M. Miller, L. Geraci / Consciousness and Cognition 42 (2016) 41–50 a framework for describing the types of cues that affect memory predictions. Intrinsic cues were defined as ones that are inherent to the studied items, such as the relationship between studied associates (e.g., nurse–doctor). Extrinsic cues were defined as ones that are not inherent to the studied items that could include a variety of factors such as people’s lay theories of memory (e.g., how repetition influences memory). Mnemonic cues were defined as those that are related to the person’s experience studying the information (thoughts about the items, reminiscence about prior related experiences, etc.). Of course, people likely use all of these types of cues to make monitoring judgments. For example, a person could study a distinctive item (an item printed in large font amongst other small font items), and use the item features and the fluent processing experience, along with a theory about distinctive items or fluent processing being well remembered, to inform their prediction about whether an item will be remembered at a later point. In this paper we will focus on the distinction between analytic and nonanalytic processes. There have been a few attempts to assess the separate contributions of analytic and nonanalytic processes to metacognitive judgments. For example, research shows that participants judge that they will remember words printed in larger fonts better than they will remember words printed in smaller fonts (Rhodes & Castel, 2008; see also Kornell & Bjork, 2009; Kornell, Rhodes, Castel, & Tauber, 2011). In the Rhodes and Castel study, participants made a judgment of learning (JOL) after studying words that were printed in large and small fonts and were later asked to recall these words. Results showed that participants gave higher JOLs for words presented in large fonts compared to those presented in small fonts, whereas recall was not affected by font size. Thus, the authors concluded that the perceptual font information lead to an illusion of memory. One explanation for this illusion is that the font manipulation influenced participants’ non-analytic (or experience-based) processing. The idea is that participants interpreted the relative ease of reading or processing words printed in large font as a benefit for later memory performance. The assumption is that participants have different subjective experiences for larger compared to smaller font words and that they use these experiences as cues to make judgments about the future memorability of these items. But, it could be that people gave higher JOLs to items printed in larger fonts than those printed in smaller fonts because they believed that more salient perceptual features would benefit memory performance, an analytic (or theory-based) process. Recently, evidence has emerged that provides evidence against the non-analytic account—that large fonts are processed more fluently—and supports the analytic interpretation instead (Mueller, Dunlosky, Tauber, & Rhodes, 2014). Mueller et al. showed that items presented in large fonts were not processed more fluently than the items presented in small fonts, as evidence by similar lexical decision and study times for large and small font items, despite being assigned higher JOLs. Further, they found that people predicted that others would recall more large- than small-font items. Here, the possibility of using non-analytic (or experience-based) processes to inform recall predictions was eliminated and yet font size affected predictions, suggesting that people use analytic (theory-based) processes when predicting performance. Finally, participants gave higher JOLs to large-font relative to small-font items before they even saw the items (using a pre-study JOL paradigm; Castel, 2008), again suggesting that people used analytic processes to make JOLs for large and small font items. Thus, this study provided evidence for the role of analytic processes for making JOLs that are based on perceptually salient cues. There have been other attempts to estimate the separate contributions of analytic and non-analytic processes by preventing access to one process. One way to do this is to use a method similar to the one used by Mueller et al. (2014) in which participants make a judgment about another participant’s recall, thus preventing the use of non-analytic (experiencebased) processes from contributing to their judgments(e.g., Matvey, Dunlosky, & Guttentag, 2001; Vesonder & Voss, 1985). In this paradigm some participants are assigned to be learners, and experience the full range of experimental tasks potentially using both analytic and non-analytic processes to assess their learning. Other participants are assigned to be judges, and are simply given information about the learners’ performance, and are asked to make a future performance prediction. By design, the judges can only use analytic processes to make judgments about the learner’s experience and performance, while the learners can use either or both types of processes. In the Matvey et al. (2001) study, learners studied cue-target pairs for a later recognition memory test and were either asked to read the cue and the target item, or generate a rhyming target in response to the cue (e.g., cave - s _ _ _, for cave - save). After each item, learners made a JOL about their future memory performance for that item. Judges saw the outcome of a learner’s attempt to generate the rhyming target (thus removing the experience of attempting to generate the rhyming word, a nonanalytic process), and then made a prediction about how likely the learner would be able to recognize the target. For learners, the longer it took them to generate the target in the rhyme condition the less likely they were to predict that they would recognize the target on a later memory test. Their subjective experience of generating the targets served as a cue for the monitoring judgment. In contrast, the JOLs judges reported for the learners were not influenced by the learner’s time to produce the target item at study. The results demonstrate a case in which people use nonanalytic processes to make JOLs. Thus, there is good evidence from a variety of paradigms showing that participants can use both types of processes to make future memory predictions. In the Mueller et al. study described earlier, participants who relied on analytic cues gave similar monitoring judgments to learners (who used both analytic and non-analytic processes). On the other hand, in the Matvey et al. study, participants who only used analytic processes gave JOLs that were significantly different than the learners’ JOLs. More recently research using item-by-item JOLs indicates that subsequent predictions are affected primarily by non-analytic processes (Serra & Ariel, 2014). Serra and Ariel suggest that participants use a memory for past test (MPT) strategy whereby they make JOLs based on their previous performance on the same to-be-remembered items. Taken together, these studies show that people can use both analytic and non-analytic processes to make JOLs and that the contribution of the two inferential processes can be separated. T.M. Miller, L. Geraci / Consciousness and Cognition 42 (2016) 41–50 43 The goal of the current studies was to examine the contribution of analytic and non-analytic processes to participants’ changes in their memory predictions following retrieval practice. Research shows that people can improve the accuracy of their memory predictions if they participate in retrieval practice (even a very brief amount of practice) prior to making a final test prediction (e.g., Miller & Geraci, 2014). In the Miller and Geraci study, participants studied paired associates, made a global prediction about their future memory performance, attempted to retrieve a sample of the studied items, and finally made a second global performance prediction before taking the criterion memory test. The results indicated that participants’ first performance predictions exhibited overconfidence and that failing to retrieve practice items led to the most change in participants’ second performance predictions, compared to succeeding at retrieval practice. The interpretation was that failure or difficulty retrieving practice items served as a powerful metacognitive cue that allowed participants to improve their subsequent metacognitive predictions. However, the Miller and Geraci (2014) study did not disentangle the contributions of analytic and nonanalytic cues that might have led participants to change their memory performance predictions. For example, it is possible that during retrieval practice, if participants failed to retrieve 2 out of the 4 items, for example, they could use analytic information to inform their following prediction. They might reason that if they don’t remember the items now, they probably won’t remember them on the final test. Or they could use nonanalytic information and reason that if the practice items didn’t come to them very quickly and some didn’t come to them at all, they probably won’t remember them on the final test. Identifying the relative contributions of these analytic and nonanalytic processes is useful for developing a complete theory about how participants change their predictions following retrieval practice and for identifying appropriate interventions to improve people’s metacognitive accuracy. The present study used a learner-judge paradigm to examine the contribution of both types of inferences to changes in global performance predictions following retrieval practice. If the changes participants make to their performance predictions after retrieval practice are due to non-analytic (experience-based) cues, then when this experiential component of retrieval practice is removed, as is the case in the judge condition, people should be insensitive to retrieval practice and show little to no change from original to adjusted prediction. On the other hand, if the changes participants make to their predictions following retrieval practice are due to analytic (theory-based) cues, then judges and learners should both change their performance predictions following retrieval practice. 2. Study 1 2.1. Method 2.1.1. Participants Fifty participants (27 females) aged 18–37 years (M = 19.69) participated in the study for partial course credit. One participant in the judge condition did not report all predictions therefore the judge and the yoked-learner’s data were excluded. All analyses that follow are based on 48 participants. 2.1.2. Design Participants were randomly assigned to condition (Learner or Judge).Each participant made two performance predictions – an Original and an Adjusted performance prediction. The dependent variable was prediction change (Adjusted Original). Therefore, a positive mean prediction change indicated that the participant increased their adjusted prediction and a negative mean prediction change indicated that the participant decreased their adjusted prediction. For example, if the participant originally predicted they would recall 16 (40% of the items or .40) paired associates but then decreased their prediction to 10 (.25) paired associates after retrieval practice, the value for prediction change would be .15. 2.1.3. Materials and procedure A sample of 40 Lithuanian-English paired-associates taken from Grimaldi, Pyc, and Rawson (2010) was used for study. Four items were chosen to be the retrieval practice items. The four items were comprised of two difficult- and two easyto-remember items based on normative data and had a mean recall of .27 after one study session (Grimaldi et al.). We used this composition of retrieval practice items because a previous study suggested that this type of retrieval practice led to improvements in prediction accuracy (Miller & Geraci, 2014). Participants in the Learner condition studied paired-associates presented on the computer one at a time for 8 s per word pair. Following study, learners made an original performance prediction (called the original prediction) and attempted to retrieve the four practice items. Prior to making the second (adjusted) performance prediction, learners read the following instructions – ‘‘Based on your experience attempting to recall the English equivalent of ‘sesuo’, ‘namas’, ‘muilas’, and ‘kamuolys’ please make a new performance prediction so that it can be as accurate as possible. Your prediction about how many word pairs you will be able to remember on the memory test can go up, go down, or stay the same. Enter a number from 0-40 below for your prediction.” Participants in the Judge condition studied paired associates in the same manner as the learners. However, judges did not make an original performance prediction or practice retrieving any items. Rather, they were informed of a yoked learner’s original performance prediction and the results of the learner’s retrieval practice; that is, judges were told how the yoked 44 T.M. Miller, L. Geraci / Consciousness and Cognition 42 (2016) 41–50 learner responded for each practice item. For example, for the first practice item ‘‘sesuo-sister” judges may have read” ‘‘The participant then typed that the English equivalent was sister.” Finally, judges were asked to report an adjusted performance prediction as a whole number for the learner for the upcoming memory test that would contain 40 items. The following is an example instruction that one judge received: You just studied 40 Lithuanian - English word pairs. Another participant was asked to study the same 40 words pairs. Following study, the participant was asked to predict his or her memory performance before the upcoming memory test. The participant predicted that he or she would remember 20 out of 40 items. The participant was then given the opportunity to answer 4 practice items. The participant was given the Lithuanian word and attempted to recall its English equivalent. At this point in the procedure, judges read each of the four practice items (e.g., sesuo – sister) and was given information about how the learner performed on each practice item. Next the judge read: The participant was then asked to adjust his or her prediction based on attempts to answer the practice items. If you were the participant, what would your adjusted performance prediction be from 1-40? 2.2. Results and discussion Results showed that following retrieval practice, learners decreased their predictions by nearly two items (M difference score = .06, SD = .13) whereas judges increased their predictions by nearly three items (M difference score = .08, SD = .12) (see Table 1, Fig. 1). Comparing these difference scores showed that participants were influenced by their assignment to either the Learner or Judge condition (F(1, 46) = 13.40, MSE = .01, p < .001, partial g2 = .23). These results demonstrate that retrieval practice information only reduced performance predictions for learners who had access to non-analytic information. Next we report the retrieval practice performance and monitoring accuracy. Overall, learners recalled fewer than half of the retrieval practice items (M = .37, SE = .06). Learners who recalled half or more of the retrieval practice items (n = 12) did not change their predictions (M difference, adjusted prediction original prediction = .01, SE = .13). In contrast, learners who recalled fewer than half of the retrieval practice items (n = 13) reduced their predictions (M difference, adjusted prediction original prediction = .11, SE = .09). In terms of monitoring accuracy, learners reported more accurate adjusted predictions than original predictions (t(24) = 2.18, p = .039). In fact, after retrieval practice, learners were extremely accurate in their predictions. There was less than a 1% difference between their adjusted predictions and their memory performance, suggesting that learners went from being overconfident about their memory performance to being metacognitively accurate. Learners also reported more accurate adjusted predictions than judges (t(47) = 2.91, p = .006). Study 1 showed that learners reduced their performance predictions following retrieval practice whereas judges did not, indicating that retrieval practice influenced non-analytical processes to which learners had primary access. In Study 2 we manipulated retrieval practice difficulty and examined the relative contribution of analytic and non-analytic processes to participants’ performance predictions. Differences in retrieval practice difficulty may affect subsequent performance predictions because they affect non-analytic processes—retrieval practice creates either an experience of easy retrieval or one of difficult retrieval. Retrieval practice difficulty could also affect subsequent performance predictions because people observe that a person has either difficulty or easy recalling information. If these analytic processes (observing failure) contribute primarily to changes in performance predictions following retrieval practice, then there should be no difference between learners and judges adjusted predictions. If only non-analytic processes contribute to performance predictions following retrieval practice, then only learners should be affected by retrieval practice. And finally if non-analytic processes contribute to changes in performance predictions following retrieval above and beyond analytic processes, then learners should be more affected by retrieval difficulty than judges. Specifically, learners’ adjusted predictions should be lower than judges’ adjusted predictions because they have the additional subjective experience associated with personal retrieval practice. The converse could be true—that analytic processes contribute above and beyond non-analytic processes. However, based on the results from Study 1, we hypothesize that non-analytic processes will play a unique role in influencing predictions following retrieval practice. Table 1 Original and adjusted performance predictions, retrieval practice performance, and memory performance for learners and judges in Study 1 expressed as a proportion of total items. Condition Original prediction Retrieval practice performance Adjusted prediction Memory performance Learner Judge .27 (.03) .27 (.03) .37 (.06) n/a .22 (.04) .35 (.03) .22 (.03) n/a Note. Judges did not engage in retrieval practice or take the final memory test. Standard error values in parentheses. 45 T.M. Miller, L. Geraci / Consciousness and Cognition 42 (2016) 41–50 Fig. 1. Learners’ and judges’ original and adjusted performance predictions as a proportion of total items in Study 1. Error bars represent standard error. 3. Study 2 3.1. Method 3.1.1. Participants One hundred participants (78 females) aged 16–33 years old (M = 18.97) participated in the study for partial course credit. 3.1.2. Design Participants were randomly assigned to condition (Learner or Judge) and retrieval practice difficulty (Easy or Difficult). There were two performance predictions for each participant – an Original and an Adjusted performance prediction. The dependent variable was prediction change (Adjusted Original). 3.1.3. Materials and procedure We used the same materials as those used in Study 1, with the exception that the difficulty of the practice items was varied. We used four practice items that either led to good or poor memory performance based on previous norming (Grimaldi et al., 2010). Previous norming, indicated that the four easy retrieval practice items had a mean recall of .48 and the difficult retrieval practice items had a mean recall of .04 after one study session (Grimaldi). The procedure for Study 2 was nearly identical to the procedure for Study 1 except that there were two levels of retrieval practice difficulty for the learners (Easy or Difficult). Participants were not told they were in the easy or difficult retrieval practice. 3.2. Results and discussion Similar to Study 1, results showed that the way participants changed their predictions was influenced by condition (F(1, 96) = 5.35, MSE = .27, p = .023, partial g2 = .05). Regardless of retrieval practice difficulty, learners decreased their predictions (M = .07, SD = .18) whereas judges increased their predictions (M = .04, SD = .27) (see Table 2, Fig. 2). There was also a main effect of retrieval practice difficulty (F(1, 96) = 7.24, MSE = .37, p = .008, partial g2 = .07). Participants in the Easy retrieval practice condition increased their performance predictions (M = .04, SD = .24) whereas participants in the Difficult retrieval practice condition decreased their performance predictions (M = .08, SD = .22). Finally, the interaction effect was not significant (F < 1, p = .739). Follow up t-tests comparing learners’ and judges’ adjusted predictions showed that learners’ predictions were numerically, although not statistically, lower than judges’ predictions in the easy retrieval practice condition Table 2 Original and adjusted performance predictions, retrieval practice performance, and memory performance for learners and judges in each retrieval practice difficulty condition of Study 2 expressed as a proportion of total items. Condition Retrieval practice performance Adjusted prediction Memory performance Easy retrieval practice Learner .36 (.03) Judge .36 (.03) Original prediction .61 (.05) n/a .36 (.03) .44 (.05) .27 (.03) n/a Difficult retrieval practice Learner .35 (.03) Judge .35 (.03) .18 (.03) n/a .21 (.03) .33 (.04) .34 (.02) n/a Note. Judges did not engage retrieval practice or take the final memory test. Standard error values in parentheses. 46 T.M. Miller, L. Geraci / Consciousness and Cognition 42 (2016) 41–50 Fig. 2. Learners’ and judges’ original and adjusted performance predictions for both easy and difficulty retrieval practice conditions as a proportion of total items in Study 2. Error bars represent standard error. (t(48) = 1.34, p = .185, d = .38) and statistically lower in the difficult retrieval practice condition (t(48) = 2.46, p = .018, d = .69). Both learners and judges were affected by manipulations but learners adjusted predictions were always lower suggesting that non-analytic processes played a critical role above and beyond that of analytic processes in reducing predictions. As in Study 1, we also examined retrieval practice performance and metacognitive monitoring accuracy. Overall, learners in the Easy Retrieval Practice condition recalled more than half of the practice items (M = .61, SE = .05) and learners in the Difficult Retrieval Practice condition recalled less than one item (M = .18, SE = .13). Learners, in either condition, who recalled half or more of the items (n = 24) did not change their predictions (M difference, adjusted prediction original prediction = .02, SE = .03). In contrast, learners who recalled fewer than half of the retrieval practice items (n = 26) reduced their predictions (M difference, adjusted prediction original prediction = .15, SE = .03). For monitoring accuracy, learners’ adjusted predictions were more accurate than judges’ adjusted predictions (F(1, 98) = 4.77, MSE = .06, p = .031, partial g2 = .05). Furthermore, for the learners, difficult retrieval practice led to more underconfidence than learners in the easy retrieval practice condition (F(1, 48) = 22.31, p < .001, partial g2 = .32), which is consistent with previous findings in the literature (i.e., Miller & Geraci, 2014). 4. Study 3 One possible limitation of the previous studies is that participants in the Judge condition studied the Lithuanian-English paired-associates just like participants in the Learner condition. So, even though Judges did not complete any retrieval practice, they could still be using mnemonic cues when they reported an adjusted performance prediction for the yoked learner. To examine this hypothesis, we conducted a study in which we eliminated all experience with the to-be-remembered paired associates. Judges read a short vignette about another fictional participant that had completed the study earlier that day (participants did not know they were reading about a fictional participant). Judges read information about the fictional participant’s original prediction and retrieval practice performance. Then they were asked to report what they believed was the fictional participant’s adjusted prediction. Importantly, all participants in Study 3 read the same vignette, in which the participant predicted 14 items originally and recalled 2 out of 4 (50%) items during retrieval practice. If Judges use only analytic information to make adjusted predictions, then the results of Study 3 should parallel those of the previous studies. That is, judges will report increased adjusted predictions; such a change would not be due to any non-analytic processing. 4.1. Method 4.1.1. Participants Thirty participants participated in the study for partial course credit. Although no data on demographic variables was collected, all participants were recruited from the same participant pool as the previous studies. 4.1.2. Design All participants made up a single-group for a one-sample t-test. Participants’ reported prediction was the dependent variable. Our hypothesis was that judges would use analytic cues when reporting an adjusted prediction. This hypothesis would be supported by a non-statistically significant t-test comparing participants’ prediction to a reference value of 20, which is a direct extrapolation of 50% retrieval practice performance to 50% memory test performance. T.M. Miller, L. Geraci / Consciousness and Cognition 42 (2016) 41–50 47 4.1.3. Materials and procedure In the vignette, participants read that the participant reported an original prediction of 14 items and recalled 2 out of the 4 items (or 50%) during retrieval practice. We chose an original prediction of 14 items and 50% retrieval practice performance because both figures were approximately what participants reported and approximately how they performed during retrieval practice in Study 2 (see Table 2). The vignette each participant read was as follows: Earlier today, a participant studied 40 items of Lithuanian words and their English equivalents. For example, ‘‘sesuosister” was one item the participant studied. After studying the 40 items, the participant predicted he would remember 14 items, or 35%, on a memory test in which he was given the word ‘‘sesuo” and had to write down ‘‘sister”. But before the memory test occurred, the same participant completed a PRACTICE TEST with 4 items that looked exactly the same as what the real memory looked like. So for example, the participant was given ‘‘sesuo” and had to write down ‘‘sister”. In all, the participant completed 4 practice test items and correctly remembered 2 out of the 4 practice items. In other words, he correctly recalled 50% of the practice items. Using his experience from the practice test, if the same participant was asked to make a NEW PREDICTION about his performance on the real memory test, how many items out of 40 do you think he would predict he would remember? Recall that his first prediction was 14 and he correctly remembered 50% of the practice test items. After reading the vignette, participants reported a whole-number adjusted prediction for the fictional participant and provided a rationale for the new prediction. We predicted that removing Judges’ experience with the to-be-remembered items would lead participants to report an adjusted prediction that was higher than the original prediction of 14 items. Finally, participants were asked to provide a rationale for their adjusted predictions. They were asked: Thinking back to what you wrote down in the blank above, why do you think you predicted that the person would get that many items correct on the memory test? Use the space below to explain how you came up with the prediction. 4.2. Results and discussion Participants in the follow-up study reported an average adjusted prediction that was higher than the fictional participant’s original performance prediction (M = 20.9, SE = .87). In fact, 29 out of 30 participants reported an increased adjusted prediction compared to the fictional participant’s original prediction of 14 items. A one sample t-test comparing Study 3 participants’ adjusted predictions to 20 revealed non-significant difference (t(29) = 1.07, p = .30). That participants’ adjusted predictions were so similar to the fictional participant’s retrieval practice performance suggests that the participants were using that information to report an adjusted prediction. Examining participants’ rationale for reporting the adjusted prediction revealed that participants considered the fictional participant’s retrieval practice performance (50%). Twenty-two out of 30 participants wrote something about the fictional participant’s retrieval practice performance. Some participants then reasoned that the fictional participant’s retrieval practice performance was proportionally greater than the original prediction and thought it was reasonable for the adjusted prediction to be increased. In fact, 20 participants explicitly referred to the participant’s 50% retrieval practice performance and many of those directly stated that because the participant recalled 50% of the items on the practice test, he would get 50% correct on the real memory test. For example, one participant wrote ‘‘In the practice test, he got 50% right so I think that he would get 50% on the actual test which is 20.” The pattern of results from the follow-up study is consistent with the pattern of results from both Study 1 and 2. Learners in the previous studies decreased their performance predictions following retrieval practice whereas Judges tended to report increased performance predictions following retrieval practice. In the follow-up study, participants, whose experiences were similar to judges in the previous studies, reported increased performance predictions. Thus, taken together, the results from Studies 1–3 suggest that judges used analytic processes to make their adjusted predictions, and that doing so resulted in a different pattern of adjusted predictions from the learners, who also used non-analytical processes to make their predictions. 5. General discussion Prior research shows that participants decrease their performance predictions following retrieval practice, particularly following difficult retrieval practice (see Miller & Geraci, 2014). This effect could occur because retrieval practice affects people’s beliefs about their memory (an analytic process) or because it affects their subjective experiences with memory (a nonanalytic process). The results of Study 1 showed that learners, who had access to analytic (theory-based) and non-analytic (experience-based) processes, decreased their predictions following retrieval practice. Further, the results of Study 2 indicated that non-analytic processes contributed more to prediction changes following different levels of retrieval practice difficult than did analytic processes. In contrast, judges, who had access only to analytic (theory-based) information, either did not change their predictions or they increased them after reading about the learners’ retrieval practice. Study 3 provided additional evidence that judges used analytic processes to make their adjusted predictions, and that doing so resulted in 48 T.M. Miller, L. Geraci / Consciousness and Cognition 42 (2016) 41–50 a different pattern of adjusted predictions from the learners. Previous work that has investigated these two sources of information has suggested that, while both processes can influence participants’ predictions, analytic processes exert more influence on JOLs (Matvey et al., 2001). Our data demonstrate that following retrieval practice, non-analytic processes exert more influence over subsequent performance predictions. Thus, the current results suggest that people rely on their subjective experiences during retrieval (a non-analytic, experience-based process) when making subsequent performance prediction adjustments. Without direct experience with retrieval practice, judges had to rely on other information to predict performance. For example in Study 1 the learners’ performance on the retrieval practice items was just fewer than half of the items (M = .37). The average adjusted prediction of the judges was very similar (M = .35), suggesting that judges were in fact taking the learners’ retrieval practice performance into account. Similarly in Study 3, the vignette stated the learners recalled half of the items during retrieval practice (i.e. M = .50) and judges reported that the learners would recall about half of the items (M = .52). Again suggesting to us that judges were taking learners’ retrieval practice performance into account. The judges’ adjusted predictions in Study 2 reveal a similar pattern of reporting predictions that are closer to the retrieval practice performance than the original predictions. That is, learners in the easy retrieval practice condition recalled over half of the items (M = .61) and the judges’ average adjusted prediction was M = .44, up from the original prediction (M = .36). Furthermore, learners in the difficult retrieval practice condition recalled under half of the items (M = .18) and the judges’ average adjusted predictions was M = .33, down from the original prediction (M = .35). Our interpretation of the results is that the judges used the learner’s performance on the practice items and their naïve theories, or beliefs, about how memory operates to make their predictions of the learners’ eventual memory performance. Judges may have taken the learner’s retrieval practice performance as an indicator of how the learners would perform on the test. That is, if the learner performed well on the practice test, the judge may have expected the learner to perform as well or even better on the actual memory test. Judges may have considered retrieval practice performance and extrapolated to the memory test by reporting a higher prediction. For example, if the learner correctly retrieved 2 out of the 4 practice items (50%), the judge might have extrapolated 50% performance during retrieval practice to 50% during the memory test. Analyses of the follow-up study results (4.2) converge on this interpretation. That is, after reading the vignette about the fictional participant’s original prediction and 50% retrieval practice performance, almost all of the judge-like participants reported higher adjusted predictions. Furthermore, their reported adjusted performance predictions were not statistically different than 20, or 50%. Therefore, judges were likely considering retrieval practice performance as an analytic cue for eventual memory test performance. We showed that following a short-run of retrieval practice, learners improved their prediction accuracy. In contrast, judges did not improve prediction accuracy, suggesting that the improvement in prediction accuracy following retrieval practice arises from participants’ use of non-analytical processes. Of course we only used four practice test items, and if given many more practice items one would predict that eventually both the learners and the judges would become highly accurate. Dunlosky, Rawson, and McDonald (2002) suggested that for practice tests to be effective for improving monitoring accuracy, they would need to be diagnostic of the final test, something they referred to as the diagnosticity assumption. The most diagnostic practice test would include all 40 items and research shows if participants make retrospective judgments (now called postdictions) their predictions are highly accurate (e.g., Maki & Serra, 1992; Pierce & Smith, 2001). Serra and Ariel (2014) reported that participants used the memory for past test (MPT) heuristic in a multiple study-test trial paradigm. They too, found that non-analytic (experienced) based processes affected performance predictions more so than analytic (theory) based processes. In contrast, in another experiment using pre-study JOLs, participants studied related and unrelated word pairs and made either prestudy JOLs or immediate JOLs. Participants gave higher estimates for related than for unrelated pairs, suggesting that participants’ beliefs at least partially drive the relatedness effect on JOLs (Mueller, Tauber, & Dunlosky, 2013). With a learner/judge paradigm, participants have access to different information but they are also making slightly different judgments (either self- or other-relevant judgments). It could be that the simple act of making a judgment about oneself leads people to try to be more accurate (and lower their predictions in this case). It might also be that, when making selfpredictions, people are motivated to show self-consistency, and not change their predictions, or to believe in themselves and maintain high performance predictions. Future research should examine the effect of making a self- versus other-judgment on participants’ willingness or ability to adjust their predictions following an intervention. We do not think that simply making the self-judgment (vs. other judgment) is sufficient to lead people to lower their predictions, as we know that this doesn’t occur in other studies without some intervention (see control condition of Miller & Geraci, 2014). But, it could be that following an intervention people are more or less willing to change their predictions when they are predicting their own versus another’s performance. Exactly how learners translated their retrieval practice performance into an adjusted performance prediction is unknown. In Study 1, learners recalled fewer than half of the 4 retrieval practice items (M = .37) and reduced their subsequent performance predictions. In Study 2, learners in the Easy retrieval practice condition recalled over half the items (M = .61) and made very little change to their subsequent predictions while those in the Difficult retrieval practice conditions had poor performance (M = .18) and reduced their predictions by 40%. These results suggest that there is not a one-to-one correspondence between retrieval practice performance and adjusted performance predictions. One possibility is that people focused on any failure during retrieval practice and this failure had a disproportionally large influence on their subsequent performance prediction. Such a disproportionate influence of negative information on experience is consistent with several findings from the negativity bias literature. For example, much has been written about the balance of positive and negative T.M. Miller, L. Geraci / Consciousness and Cognition 42 (2016) 41–50 49 emotions that leads to happiness and fulfillment (c.f. Frederickson, 2013; Frederickson & Losada, 2005). This research suggests that negative emotions have a larger influence on people compared to positive emotions. Similarly, we know that people are highly attentive to negative feedback and predictive of whether or not that person will make the same mistake on a future trial (Gehring, Liu, Orr, & Carp, 2012; Van der Helden, Boksem, & Blom, 2009). Indeed, errors are memorable and may be crucial for learning. People are more likely to respond correctly on subsequent trials if they had previously responded incorrectly with high- rather than low-confidence, a finding known as the hyper-correction effect (Butler, Fazio, & Marsh, 2011; Metcalfe & Finn, 2011). The evidence supporting the disproportional influence of negative over positive aspects of one’s experience has led some to suggest that the imbalance ‘‘may in fact be a general principle or law of psychological phenomena” (Baumeister, Bratslavsky, Finkenauer, & Vohs, 2001; p. 323). Thus, our data, showing that a small amount of retrieval failure has a relatively large effect on subsequent performance predictions, could be seen as consistent with these various literatures. While learners were able to decrease their performance predictions following retrieval practice, in some cases, such as when following moderate or easy retrieval practice, judges actually increased their subsequent performance predictions relative to the baseline prediction. Why did they increase their predictions relative to baseline? The current study cannot say for sure, but there is good evidence that some people overestimate their own performance and underestimate others’ performance (Hartwig & Dunlosky, 2014; Kruger & Dunning, 1999). Thus, it is possible that while performing poorly on a practice test led learners to lower their subsequent performance predictions (because they were overconfident to begin with), the same amount of practice success (or failure) may be interpreted differently by judges. If people generally expect others to perform poorly, then it is possible that judges may have expected the learners to perform more poorly than they did on the practice items, and might consider even modest success to be surprising and indicative of greater future success. This hypothesis is speculative and awaits future testing. For now, our primary finding is that learners were able to use retrieval practice experiences to reduce their overconfidence. 5.1. Conclusions The current studies add to the literature by demonstrating that retrieval failure is beneficial for metacognitive monitoring. Most importantly, the results demonstrate that while both types of processes – analytic and non-analytic – can contribute to changes in memory predictions following retrieval practice, the subjective experience of retrieval practice plays a key role in leading participants to reduce their performance predictions. References Arbuckle, T. Y., & Cuddy, L. L. (1969). Discrimination of item strength at time of presentation. Journal of Experimental Psychology, 81, 126–131. Baumeister, R. F., Bratslavsky, E., Finkenauer, C., & Vohs, K. D. (2001). Bad is stronger than good. Review of General Psychology, 5, 323–370. Butler, A. C., Fazio, L. F., & Marsh, E. J. (2011). The hypercorrection effect persists over a week, but high confidence errors return. Psychonomic Bulletin & Review, 18, 1238–1244. Castel, A. D. (2008). Metacognition and learning about primacy and recency effects in free recall: The utilization of intrinsic and extrinsic cues when making judgments of learning. Memory & Cognition, 36, 429–437. Cheng, C. (2010). Accuracy and stability of metacognitive monitoring: A new measure. Behavior Research Methods, 42, 715–732. Dunlosky, J., Rawson, K. A., & McDonald, S. L. (2002). Influence of practice tests on the accuracy of predicting memory performance for paired associates, sentences, and text material. In T. J. Perfect & B. L. Schwartz (Eds.), Applied metacognition (pp. 68–92). Cambridge, UK: Cambridge University Press. Flavell, J. (1979). Metacognition and cognitive monitoring: A new area of cognitive developmental inquiry. American Psychologist, 34, 906–911. Fleming, S. M., & Lau, H. C. (2014). How to measure metacognition. Frontiers in Human Neuroscience, 8. Published online 2014 Jul 15. Frederickson, B. L. (2013). Updated thinking on positivity ratios. American Psychologist, 68, 814–822. Frederickson, B. L., & Losada, M. F. (2005). Positive affect and the complex dynamics of human flourishing. American Psychologist, 60, 678–686. Gehring, W. J., Liu, Y., Orr, J. M., & Carp, J. (2012). The error-related negativity (ERN/Ne). In S. J. Luck & E. Kappenman (Eds.), Oxford handbook of event-related potential components (pp. 231–291). New York: Oxford University Press. Grimaldi, P. J., Pyc, M. A., & Rawson, K. A. (2010). Normative multitrial recall performance, metacognitive judgments, and retrieval latencies for LithuanianEnglish paired associates. Behavior Research Methods, 42, 634–642. Hart, J. T. (1965). Memory and feeling-of-knowing experience. Journal of Educational Psychology, 56, 208–216. Hartwig, M. K., & Dunlosky, J. (2014). The contribution of judgment scale to the unskilled-and-unaware phenomenon: How evaluating others can exaggerate over- (and under-) confidence. Memory & Cognition, 42, 164–173. Kelley, C. M., & Jacoby, L. L. (1996). Adult egocentrism: Subjective experience versus analytic bases for judgment. Journal of Memory and Language, 35, 157–175. Koriat, A. (1997). Monitoring one’s own knowledge during study: A cue-utilization approach to judgments of learning. Journal of Experimental Psychology: General, 126, 349–370. Koriat, A., Bjork, R. A., Sheffer, L., & Bar, S. K. (2004). Predicting one’s own forgetting: The role of experience-based and theory-based processes. Journal of Experimental Psychology: General, 133, 643–656. Kornell, N., & Bjork, R. A. (2009). A stability bias in human memory: Overestimating remembering and underestimating learning. Journal of Experimental Psychology: General, 138, 449–468. Kornell, N., Rhodes, M. G., Castel, A. D., & Tauber, S. K. (2011). The ease of processing heuristic and the stability bias: Dissociating memory, memory beliefs, and memory judgments. Psychological Science, 22, 787–794. Kruger, J., & Dunning, D. (1999). Unskilled and unaware of it: How difficulties in recognizing ones’ own incompetence lead to inflated self-assessments. Journal of Personality and Social Psychology, 77, 1121–1134. Maki, R. H., & Serra, M. (1992). Role of practice tests in the accuracy of text predictions on text material. Journal of Educational Psychology, 84, 200–210. Matvey, G., Dunlosky, J., & Guttentag, R. (2001). Fluency of retrieval at study affects judgments of learning (JOLs): An analytic or nonanalytic basis for JOLs? Memory & Cognition, 29, 222–233. Metcalfe, J., & Finn, B. (2011). People’s hypercorrection of high-confidence errors: Did they know it all along? Journal of Experimental Psychology: Learning, Memory, and Cognition, 37, 437–448. 50 T.M. Miller, L. Geraci / Consciousness and Cognition 42 (2016) 41–50 Miller, T. M., & Geraci, L. (2014). Improving metacognitive accuracy: How failing to retrieve practice items reduces overconfidence. Consciousness and Cognition, 29, 131–140. Mueller, M. L., Dunlosky, J., Tauber, S. K., & Rhodes, M. G. (2014). The font-size effect on judgments of learning: Does it exemplify fluency effects or reflect people’s beliefs about memory? Journal of Memory and Language, 70, 1–12. Mueller, M. L., Tauber, S. K., & Dunlosky, J. (2013). Contributions of beliefs and processing fluency to the effect of relatedness on judgments of learning. Psychonomic Bulletin & Review, 20, 378–384. Nelson, T. O. (1996). Gamma is a measure of the accuracy of predicting performance on one item relative to another item, not of the absolute performance of an individual item: Comments on Schraw (1995). Applied Cognitive Psychology, 10, 257–260. Nelson, T. O., & Dunlosky, J. (1991). When people’s judgments of learning (JOLs) are extremely accurate at predicting subsequent recall: The ‘‘delayed-JOL” effect. Psychological Science, 2, 267–270. Nelson, T. O., & Narens, L. (1990). Metamemory: A theoretical framework and new findings. In G. H. Bower (Ed.). The psychology of learning and motivation (Vol. 26, pp. 125–173). New York: Academic Press. Pierce, B. H., & Smith, S. M. (2001). The postdiction superiority effect in metacomprehension of text. Memory & Cognition, 29, 62–67. Rhodes, M. G., & Castel, A. D. (2008). Memory predictions are influenced by perceptual information: Evidence for metacognitive illusions. Journal of Experimental Psychology: General, 137, 615–625. Schwartz, B. L. (1994). Sources of information in metamemory: Judgments of learning and feelings of knowing. Psychonomic Bulletin & Review, 1, 357–375. Serra, M. J., & Ariel, R. (2014). People use the memory for past-test heuristic as an explicit cue for judgments of learning. Memory & Cognition, 42, 1260–1272. Van der Helden, J., Boksem, M. A. S., & Blom, J. H. G. (2009). The importance of failure: Feedback-related negativity predicts motor learning efficiency. Cerebral Cortex, 20, 1596–1603. Vesonder, G. T., & Voss, J. F. (1985). On the ability to predict one’s own responses while learning. Journal of Memory and Language, 24, 363–376.

Consciousness and Cognition 43 (2016) 66–74 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Attentional cueing induces false memory Kiyofumi Miyoshi a,b,⇑, Hiroshi Ashida a a b Graduate School of Letters, Kyoto University, Sakyo, Kyoto 6068501, Japan Japan Society for the Promotion of Science, Kyoto University, Sakyo, Kyoto 6068501, Japan a r t i c l e i n f o Article history: Received 28 March 2015 Revised 15 December 2015 Accepted 16 May 2016 Keywords: Recognition memory Memory illusion Processing fluency Attentional cueing a b s t r a c t The fluency of stimulus processing significantly contributes to recognition memory judgments. We investigated the effect of processing fluency induced by attentional cueing on recognition judgments. Participants performed a Remember/Know recognition test, while their spatial attention was manipulated in the test session. Stimulus location was either predicted (congruent condition) or unpredicted (incongruent condition) using an arrow cue. The results revealed that familiarity-based false recognition increased in the incongruent condition wherein the participants may have attributed part of the perceived disfluency to the attentional cue, and they may have overestimated the fluency for the stimulus, leading to increased false recognition. However, in the congruent condition, the participants may have attributed some parts of the perceived fluency to the attentional cue and underestimated the fluency for the stimulus, leading to decreased false recognition. In sum, stimulus-irrelevant attentional cueing induces unintentional processing about the source of fluency and biases recognition memory. Ó 2016 Elsevier Inc. All rights reserved. 1. Introduction Human memory has been extensively investigated using recall and recognition tests (see Yonelinas, 2002 for a review). While participants in recall tests retrieve internal memory representations, participants in recognition tests compare the perceptual input of the stimulus with their internal memory representations. Thus, recognition tests involve perceptual as well as memory processes. The perception of the stimulus in recognition memory tests can affect whether the stimulus is recognized as old or new. One excellent example of this is ‘‘memory illusion,” reported by Jacoby and Whitehouse (1989). They demonstrated that a stimulus preceded by a masked unconscious prime was more likely to be recognized as old, thus, suggesting that increased processing fluency generates increased feelings of familiarity for the stimulus, leading to increased ‘‘old” responses. Furthermore, Jacoby and Whitehouse (1989) reported that participants were more likely to recognize a stimulus as new when it was preceded by a visible prime. They explained this reversion of the priming effect using the idea that participants attribute some parts of perceived fluency to a visible prime and discount fluency for the stimulus. This reversion of the priming effect has been replicated in several studies, but the ‘‘fluency discounting” account has been challenged (Higham & Vokey, 2000; Huber, Clark, Curran, & Winkielman, 2008; Klinger, 2001). Huber et al. (2008) proposed that the reversal of the priming effect on recognition judgments is explicable without fluency discounting. Their ‘‘fluency/disfluency” account suggests that a short prime duration leads to priming and makes stimulus processing fluent, which results in increased ‘‘old” responses. In contrast, a longer prime duration leads to habituation and makes stimulus processing disfluent, which results in decreased ‘‘old” responses. In short, Jacoby and Whitehouse (1989) assumed that a conscious awareness of the prime ⇑ Corresponding author. E-mail addresses: miyoshi80@gmail.com (K. Miyoshi), ashida@psy.bun.kyoto-u.ac.jp (H. Ashida). http://dx.doi.org/10.1016/j.concog.2016.05.006 1053-8100/Ó 2016 Elsevier Inc. All rights reserved. K. Miyoshi, H. Ashida / Consciousness and Cognition 43 (2016) 66–74 67 induces fluency discounting, which is critical for the reversal of priming, whereas Huber et al. (2008) assumed that stimulus fluency itself is critical, regardless of awareness of the prime. One reason for this controversy may be that the visual awareness of the prime and fluency for the stimulus covary according to the manipulation of prime duration. Thus, alternative methods to manipulate fluency are likely to offer new insights into this topic. One potent method to manipulate fluency is attentional cueing (e.g., Posner, 1980; Shepherd, Findlay, & Hockey, 1986; Tipples, 2002). The processing of the target stimulus becomes more fluent when the stimulus location is predicted by attentional cueing. Using attentional cueing, one can manipulate the fluency/disfluency of stimulus processing with keeping a visual awareness of the cue constant. We employed an arrow as an attentional cue because even a non-predictive arrow cue automatically affects participants’ attention (e.g., Ristic & Kingstone, 2006; Tipples, 2002). We set cue validity at 50% to equalize the number of predicted trials and unpredicted trials. It is important to take note of the results of a study by Rajaram and Geraci (2000). In this study, participants performed a Remember/Know recognition memory test (Gardiner, 1988; Tulving, 1985) while stimulus fluency was manipulated by conceptual priming. The results demonstrated that priming selectively affected ‘‘know” responses but did not affect ‘‘remember” responses. These results provided strong support for the dual-process theory of recognition memory, which assumes that recollection and familiarity contribute to recognition memory (Atkinson & Juola, 1974; Hintzman & Curran, 1994; Jacoby & Dallas, 1981; Mandler, 1980; Wixted, 2007). The selective effect of fluency manipulation on know responses is compatible with two influential models of recognition memory. Under the dual-process signal detection (DPSD) model (Yonelinas, 1994), recollection is considered an all-or-none threshold process, which is not affected by fluency manipulation. However, familiarity is considered a continuous signal detection process that can be affected by fluency manipulation. Alternatively, under the continuous dual-process (CDP) model (Wixted & Mickes, 2010), both recollection and familiarity are considered continuous variables, additively combined into a single unidimensional memory strength distribution, on which old/new responses depend. Furthermore, the CDP model assumes that remember responses are based on a recollection distribution, and know responses are based on a familiarity distribution. The CDP model can be compatible with the results in Rajaram and Geraci (2000), if it is assumed that fluency manipulation affects familiarity distribution but leaves recollection distribution relatively unaffected. We return to this point in Section 4. In the present study, we investigated the effect of attentional cueing on recollection and familiarity using the Remember/ Know procedure. If the stimulus-irrelevant attentional cue biases memory, this will be a matter of significant importance in the theoretical understanding of human memory and in our lives outside the laboratory. 2. Preliminary experiment We conducted a preliminary study to explore appropriate sample size and other task settings (e.g., set size of stimuli, and duration of attentional cue). We expected that attentional cueing would affect participants’ recognition bias but not recognition accuracy. Dissimilar to the main experiment, participants were explicitly instructed to attend to the direction of the cue. 2.1. Methods 2.1.1. Participants Twenty-two students from Kyoto University aged between 18 and 25 (M = 21.5, SD = 1.97) participated in the preliminary experiment and were paid according to the Kyoto University standard. This sample size was determined by reference to related studies (14 in Dew & Cabeza, 2013; 20–24 in Lucas, Taylor, Henson, & Paller, 2012; 24 in Woollams, Taylor, Karayanidis, & Henson, 2008; 22 in Taylor, Buratto, & Henson, 2013; 14–20 in Whittlesea & Williams, 2000; 13–21 in Whittlesea & Williams, 2001). Eleven participants were male and 11 were female. All participants had normal color vision. Informed written consent was obtained from participants before the experiment. All data were collected in accordance with the ethical principles of the Japanese Psychological Association. Data from two participants were excluded because one participant exhibited recognition accuracy below chance and the other exhibited no false alarms at all. 2.1.2. Stimuli The stimuli consisted of 320 color pictures of everyday objects; the pictures were obtained from the Massachusetts Institute of Technology website (http://cvcl.mit.edu/MM/index.html). These stimuli were randomly assigned to each experimental condition (congruent/new, congruent/old, incongruent/new, and incongruent/old) for each participant. The stimuli were displayed on a dark background on a 23-inch computer monitor (I-O DATA, LCD-MF234XPBR) using the software Presentation (Neurobehavioral Systems). Each stimulus was framed in a white window measuring 6.8
6.8 cm. A white arrow measuring 1.8
9.4 cm was used as an attentional cue. The distance between the monitor and participants was 50 cm. 2.1.3. Procedure Before the experiment described below, each participant completed a practice session, which included four study trials and eight test trials. During the study session, participants were instructed to memorize 160 stimuli presented on the 68 K. Miyoshi, H. Ashida / Consciousness and Cognition 43 (2016) 66–74 computer monitor. Each stimulus was presented for 1000 ms with an inter-trial interval of 1500 ms, during which a central fixation cue was presented (Fig. 1). After the study session, participants conducted an arithmetic task for 2 min. During the test session, 160 new stimuli and 160 old stimuli were presented in random order. Participants judged whether each stimulus was old or new and reported their awareness of memory according to the Remember/Know procedure (Gardiner, 1988). ‘‘Remember” responses indicated that recognition was accompanied by a conscious recollection of the specific details of the stimulus or through contextual information from the study session. ‘‘Know” responses indicated that the recognition was supported by a vague feeling of familiarity without conscious recollection of stimulus details or contextual information. Before the experiment, participants were instructed that half of the stimuli in the test session were new and that the other half were old. This instruction was given to prevent participants from being too conservative in judging whether the stimuli were old or new. During each test trial, after the presentation of the fixation cross for 1500 ms, an arrow was presented for 500 ms as an attentional cue, followed by the stimulus (Fig. 1). In the preliminary study, participants were explicitly instructed to attend to the direction of the attentional cue. The stimulus continued to be presented until participants responded. The stimulus was presented either to the left or the right of the monitor, and its center was 8.2 cm away from the center of the monitor. If participants provided an ‘‘old” response, an alternative (R or K) was presented and participants made a resultant metamemory judgment. The arrow was equally likely to appear toward the left or the right and was non-predictive as to either the location or the study status (old/new) of the stimulus. Participants experienced a total of 320 trials (80 congruent/new, 80 congruent/old, 80 incongruent/new, and 80 incongruent/old trials) in random order. After the experiment, participants completed a questionnaire in which they were asked whether they assumed any relationships to exist between the attentional condition (congruent/incongruent) and the study status of the stimulus (old/new). 2.1.4. Data analysis Recollection and familiarity are considered independent or redundant, not mutually exclusive (see Yonelinas, 2002 for a review). However, in the Remember/Know procedure, participants make a know response when a stimulus is familiar and not recollected. Therefore, the raw proportion of know responses underestimates the actual familiarity (Yonelinas & Jacoby, 1995). To overcome the problem, recollection and familiarity were estimated under the independent Remember/Know assumption (IRK) (Yonelinas & Jacoby, 1995). Under this assumption, familiarity is calculated by dividing the proportion of know responses by 1 minus the proportion of remember responses [IRK familiarity = pKnow/(1  pRemember)]. These estimates were calculated separately for each condition. Under the IRK assumption, the estimate of recollection and familiarity can have a mutually independent value. Thanks to this, we calculated recognition accuracy (Pr) and recognition bias (Br) (Snodgrass & Corwin, 1988) separately for recollection and familiarity. Pr is defined as the proportion of hits minus the proportion of false alarms [Pr = pHit  pFA] and sometimes called the ‘‘corrected recognition score.” Br is calculated by dividing the proportion of false alarms by 1 minus Pr [Br = pFA/(1  Pr)]. Furthermore, under the IRK assumption, metamemory (recollection/IRK familiarity) can be used as a factor in ANOVA analysis because independence of data is assured between factor levels (see Scariano & Davenport, 1987 with regards to the assumption of independence in ANOVA analysis). Repeated-measures ANOVAs were conducted separately on the false recognition of new stimuli (false alarm), correct recognition of old stimuli (hit), recognition accuracy (Pr), and recognition bias (Br). Partial g2 was reported to represent effect size. According to Cohen’s criteria, partial g2 of 0.01 is small, 0.06 is moderate, and 0.14 is large (Cohen, 1992). (Please see Richardson (2011) as to the applicability of Cohen’s criteria to partial g2.) According to a kind suggestion by the editor, when we observed a null result in a conventional statistical test, we conducted a Bayesian analysis of a null hypothesis (H0) versus an alternative hypothesis (H1) and reported the Bayes factor (BF) [p(D|H0)/p(D|H1)]. To estimate the BF, we used the software More Power 6.0 (Campbell & Thompson, 2012), which uses the Bayesian information criterion (BIC) to approximate the BF (Raftery, 1995, 1999). Table 1 provides a scale of evidence for interpreting the BF (Jeffreys, 1961). Fig. 1. Schematic diagrams of the experimental procedure. 69 K. Miyoshi, H. Ashida / Consciousness and Cognition 43 (2016) 66–74 Table 1 Evidence scale for interpreting Bayes factors (Jeffreys, 1961). Bayes factor Evidence for H0 over H1 1–3.2 3.2–10 10–100 >100 Not worth more than a bare mention Positive Strong Very strong 2.2. Results and discussion The post-experiment questionnaire showed that no participants assumed the existence of any relationships between the attentional condition and the study status of the stimulus. Recognition memory performances and reaction times for old/ new responses in each condition are summarized in Table 2. We conducted an ANOVA on the false recognition of new stimuli with metamemory (recollection/IRK familiarity) and attentional condition (congruent/incongruent) as within-participant factors. There was a significant main effect for metamemory (F(1, 19) = 51.90, p < 0.001, partial g2 = 0.732) but no significant main effect for attentional condition (F(1, 19) = 0.22, p = 0.645, partial g2 = 0.011, BF = 4.0). An interaction was nonsignificant (F(1, 19) = 0.03, p = 0.865, partial g2 = 0.002, BF = 4.4). These results suggest that attentional cueing did not affect the false recognition of new stimuli. The RT for old/new responses for new stimuli was significantly faster in the congruent condition than in the incongruent condition (t(19) = 3.95, p < 0.001, d0 = 0.87). Next, we conducted an ANOVA on the correct recognition of old stimuli with metamemory and attentional condition as within-participant factors. There was no significant main effect for metamemory (F(1, 19) = 0.34, p = 0.568, partial g2 = 0.017, BF = 3.8) but a significant main effect for attentional condition (F(1, 19) = 6.28, p = 0.021, partial g2 = 0.248) was observed. An interaction was nonsignificant (F(1, 19) = 0.25, p = 0.620, partial g2 = 0.013, BF = 3.9). The results indicate that the correct recognition of old stimuli was more frequent in the incongruent condition than in the congruent condition. The RT for old/new responses for old stimuli was significantly faster in the congruent condition than in the incongruent condition (t(19) = 5.02, p < 0.001, d0 = 1.12). We also analyzed recognition accuracy (Pr). An ANOVA showed a significant main effect for metamemory (F(1, 19) = 7.61, p = 0.013, partial g2 = 0.286) and attentional condition (F(1, 19) = 4.58, p = 0.045, partial g2 = 0.194). An interaction was nonsignificant (F(1, 19) = 0.25, p = 0.626, partial g2 = 0.013, BF = 3.9). These results suggest that recognition accuracy was higher in the incongruent condition than in the congruent condition. Finally, we conducted an ANOVA on recognition bias (Br) with metamemory and attentional condition as withinparticipant factors. A significant main effect for metamemory was found (F(1, 19) = 20.64, p < 0.001, partial g2 = 0.521), but no significant main effect was found for attentional condition (F(1, 19) = 0.571, p = 0.459, partial g2 = 0.029, BF = 3.3). An interaction was nonsignificant (F(1, 19) = 0.14, p = 0.711, partial g2 = 0.007, BF = 4.2). These results suggest that attentional cueing did not affect participants’ recognition bias. Notably, however, additional analyses without the IRK assumption showed a different pattern of results than that described above (see Appendix A for details). Simply stated, attentional cueing did not have significant effects on participants’ raw responses but did have effects on recognition measures estimated from those responses under the IRK assumption. The present findings must be interpreted carefully. In any case, contrary to our expectation, attentional cueing did not affect participants’ recognition bias. Although it is unclear why this pattern of results arose, one possibility is that the participants might try to remain unaffected by attentional manipulation in recognition judgments when they consciously shift their attention on their own. To explore this possibility, we conducted the main experiment, in which participants were not explicitly instructed to attend to the direction of the arrow. As several significant effects of attentional cueing were observed in the preliminary study, we used the same task settings (e.g., number of participants, set size of stimuli, duration of attentional cue) in the main experiment. Table 2 Mean recognition performances and mean reaction time in the preliminary experiment. Congruent Recollection pKnow IRK familiarity RT (ms) Incongruent Old stimuli New stimuli Pr Br Old stimuli New stimuli Pr Br 0.53 (0.18) 0.24 (0.12) 0.49 (0.12) 1325 (258) 0.03 (0.04) 0.12 (0.07) 0.13 (0.08) 1367 (225) 0.50 (0.20) 0.06 (0.07) 0.07 (0.15) 0.20 (0.10) 0.02 (0.03) 0.12 (0.07) 0.13 (0.08) 1505 (277) 0.53 (0.20) 0.36 (0.16) 0.55 (0.19) 0.23 (0.10) 0.53 (0.12) 1429 (278) 0.40 (0.15) 0.20 (0.10) SD is shown in parenthesis. 70 K. Miyoshi, H. Ashida / Consciousness and Cognition 43 (2016) 66–74 3. Main experiment 3.1. Methods 3.1.1. Stimuli and procedure The stimuli and procedure were the same as those in the preliminary experiment, except that no instruction was provided regarding the attentional cue; participants were not explicitly instructed to attend to the direction of the cue. Nevertheless, the arrow cue was shown to automatically change the attentional states of participants (e.g., Bayliss, Paul, Cannon, & Tipper, 2006; Tipples, 2002). 3.1.2. Participants In all, 22 students from Kyoto University aged between 19 and 24 years (M = 21.2, SD = 1.47) participated in the experiment and were paid according to the Kyoto University standard. Fifteen participants were male and 7 were female. All participants had normal color vision. Informed written consent was obtained from participants before the experiment. All data were collected in accordance with the ethical principles of the Japanese Psychological Association. 3.2. Results The post-experiment questionnaire showed that no participants assumed the existence of any relationships between the attentional condition and the study status of the stimulus. Recognition memory performances and reaction times for old/ new responses in each condition are summarized in Table 3. Recollection and familiarity were estimated under the IRK assumption. First, we conducted an ANOVA on the false recognition of new stimuli with metamemory (recollection/IRK familiarity) and attentional condition (congruent/incongruent) as within-participant factors (Fig. 2A). A significant main effect was found for metamemory (F(1, 21) = 82.78, p < 0.001, partial g2 = 0.798), but no significant main effect was found for attentional condition (F(1, 21) = 3.25, p = 0.086, partial g2 = 0.134, BF = 1.0). An interaction between these two factors was significant (F (1, 21) = 4.44, p = 0.047, partial g2 = 0.175). A simple main effect analysis revealed that the estimate of IRK familiarity was significantly higher in the incongruent condition than in the congruent condition (F(1, 42) = 7.38, p = 0.0096, partial g2 = 0.260), which suggests that participants exhibited familiarity-based false recognition more frequently in the incongruent condition. A post hoc power analysis revealed that the power to detect this simple main effect was 0.98 under the current settings (sample size = 22 and significance level = 0.05). No significant difference between the estimates of recollection in the congruent condition and in the incongruent condition was seen (F(1, 42) = 0.02, p = 0.896, partial g2 = 0.001, BF = 4.6). In Table 3 Mean recognition performances and mean reaction time in the main experiment. Congruent Recollection pKnow IRK familiarity RT (ms) Incongruent Old stimuli New stimuli Pr Br Old stimuli New stimuli Pr Br 0.50 (0.17) 0.25 (0.13) 0.49 (0.17) 1231 (249) 0.03 (0.04) 0.14 (0.07) 0.15 (0.07) 1268 (235) 0.47 (0.18) 0.05 (0.06) 0.06 (0.06) 0.23 (0.12) 0.03 (0.03) 0.16 (0.07) 0.17 (0.07) 1335 (278) 0.49 (0.21) 0.35 (0.15) 0.51 (0.20) 0.26 (0.13) 0.53 (0.19) 1281 (267) 0.36 (0.19) 0.28 (0.13) SD is shown in parenthesis. Fig. 2. (A) Estimates of false recollection and false IRK familiarity. (B) Estimates of correct recollection and correct IRK familiarity. Error bars indicate the standard error of the mean. ** Indicates p < 0.01. K. Miyoshi, H. Ashida / Consciousness and Cognition 43 (2016) 66–74 71 addition, the old/new response time for new stimuli was significantly faster in the congruent condition than in the incongruent condition (t(21) = 3.45, p = 0.002, d0 = 0.74). Second, we conducted an ANOVA on the correct recognition of old stimuli with metamemory and attentional condition as within-participant factors (Fig. 2B). No significant main effect was found for metamemory (F(1, 21) = 0.002, p = 0.965, partial g2 = 0.0001, BF = 4.7), nor was a significant main effect found for attentional condition (F(1, 21) = 1.23, p = 0.281, partial g2 = 0.055, BF = 2.5). An interaction between metamemory and attentional condition was also nonsignificant (F(1, 21) = 1.41, p = 0.249, partial g2 = 0.063, BF = 2.3). These results suggest that attentional cueing did not affect the correct recognition of old stimuli. The old/new response time for old stimuli was significantly faster in the congruent condition than in the incongruent condition (t(21) = 2.94, p = 0.008, d0 = 0.63). Third, we analyzed recognition accuracy (Pr) under the IRK assumption (Table 3). We conducted an ANOVA on Pr with metamemory and attentional condition as within-participant factors. A significant main effect for metamemory was found (F(1, 21) = 6.78, p = 0.017, partial g2 = 0.244), but no significant main effect was found for attentional condition (F(1, 21) = 0.21, p = 0.651, partial g2 = 0.010, BF = 4.2). Furthermore, no significant interaction between metamemory and attentional condition was seen (F(1, 21) = 0.02, p = 0.893, partial g2 = 0.001, BF = 4.6). The results suggest no significant effect of attentional cueing on recognition accuracy. Finally, we analyzed recognition bias (Br) under the IRK assumption (Table 3). We conducted an ANOVA on Br with metamemory and attentional condition as within-participant factors. A significant main effect for metamemory was found (F(1, 21) = 77.70, p < 0.001, partial g2 = 0.787), but no significant main effect was found for attentional condition (F(1, 21) = 2.79, p = 0.110, partial g2 = 0.117, BF = 1.2). The interaction between metamemory and attentional condition approached significance (F(1, 21) = 3.33, p = 0.082, partial g2 = 0.137, BF = 0.9). A simple main effect analysis revealed that familiaritybased Br was higher in the incongruent condition (F(1, 42) = 5.63, p = 0.022, partial g2 = 0.211), suggesting that familiarity-based recognition was more liberal in the incongruent condition. A post hoc power analysis revealed that the power to detect this simple main effect was 0.95. Attentional condition did not significantly affect recollection-based Br (F(1, 42) = 0.24, p = 0.624, partial g2 = 0.011, BF = 4.2). In summary, the present results reveal that familiarity-based false recognition increased by incongruent cueing in testing. This increased false recognition was associated with participants’ liberal recognition bias, not with decreased recognition accuracy. Thus, increased familiarity-based false recognition in the incongruent condition cannot be attributed to the possibility that participants could not accurately recognize the stimulus due to inattention. Note that additional analyses without the IRK assumption showed a consistent pattern of results with the above-mentioned main results (see Appendix B for details). 4. General discussion The main experiment provides the novel finding that stimulus-irrelevant attentional cues bias recognition memory judgments and induce familiarity-based false recognition. Although we did not explicitly instruct participants to attend to the direction of the arrow cue, it automatically affected the fluency of stimulus processing (reflected in significant RT difference). In the incongruent condition, participants may attribute part of the perceived disfluency to the attentional cue and overestimate the fluency of the stimulus, thus leading to more liberal familiarity-based recognition judgments. On the other hand, in the congruent condition, participants may attribute some parts of the perceived fluency to the attentional cue rather than to the study session. As a result, they might underestimate the fluency of the stimulus and exhibit more conservative familiarity-based recognition judgments. Unlike the prime used in previous studies, the attentional cue in the present study was not perceptually or semantically related to the stimulus. Nevertheless, participants may still attribute the perceived fluency or disfluency to the cue. We did not provide any instruction about the attentional cue to participants. Moreover, participants reported in the questionnaire that they did not assume any relationships to exist between the attentional condition (congruent/incongruent) and the study status of the stimulus (old/new). Therefore, attentional cueing might induce unintentional processing about the source of fluency, consequently causing bias in participants’ recognition memory. This explanation accords with the fluency discounting account introduced by Jacoby and Whitehouse (1989) but is inconsistent with the fluency/disfluency account proposed by Huber et al. (2008). If stimulus fluency itself is critical for memory bias, recognition judgments would have been more liberal in the congruent condition; however, the opposite occurred in this study. The present findings may also be explained in the context of the discrepancy-attribution theory (e.g., Whittlesea & Williams, 1998, 2000, 2001). This theory assumes that a feeling of familiarity arises when actual fluency is discrepant from expected fluency (e.g., seeing one’s spouse in the subway rather than in the kitchen). In the present study, expected fluency may be lower in the incongruent condition than in the congruent condition. Unexpectedly high fluency in the incongruent condition may lead to a strong feeling of familiarity, leading to participants exhibiting more old responses. Another possible speculation stems from the effect of selective attention. Fischer and Whitney (2009) reported that spatially directed attention narrows the tuning of population-coded position representations in the primary visual cortex. The narrowed representation of the stimulus may inhibit the diffusional activation of delusive information and may reduce familiarity-based false recognition in the congruent condition. However, this account cannot explain why attentional cueing induces an overall shift of recognition bias in familiarity-based recognition judgments. 72 K. Miyoshi, H. Ashida / Consciousness and Cognition 43 (2016) 66–74 Attentional cueing only affected the estimate of familiarity; the estimate of recollection was not affected (this pattern was positively supported by the Bayesian analysis). As we mentioned in the introduction, the pattern of the results is compatible with two influential dual-process models of recognition memory. In the DPSD model (Yonelinas, 1994), recollection is an allor-none threshold process, whereas familiarity is a continuous process that can be influenced by fluency manipulation. In the CDP model (Wixted & Mickes, 2010), recollection and familiarity are considered to depend on two respective signal distributions. If we assume that attentional cueing affects familiarity distribution with recollection distribution relatively unaffected, this model is also compatible with the present results. Nevertheless, as recollection is understood as a continuous process under the CDP model, there is room for recollection to be affected by fluency manipulation. Consistent with this, Lucas et al. (2012) reported that masked priming affected both remember and know responses. Furthermore, Kurilla and Westerman (2008) reported that masked priming affected individual ratings of recollection and familiarity. However, these empirical findings do not instantly allow for rejection of the DPSD model (Yonelinas, 1994), because Remember/Know responses and individual ratings may not be process-pure measures of recollection and familiarity; these putative measures of recollection could partially reflect familiarity and thus be affected by fluency manipulation. Furthermore, attentional cueing only affected the false familiarity for new stimuli; the correct familiarity for old stimuli was not affected. In accordance with this evidence, several studies demonstrated that the fluency effect on recognition judgments was more robust for new rather than old stimuli (e.g., Tunney & Fernie, 2007; Westerman, 2008; Whittlesea & Williams, 2001). One possibility is that the estimate of familiarity for old stimuli may be to some extent contaminated with recollection from the study session, and thus less affected by attentional cueing. Several studies support the idea that know responses entail some amount of recollection (Hicks, Marsh, & Ritschel, 2002; Johnson, McDuff, Rugg, & Norman, 2009; Wais, Mickes, & Wixted, 2008). However, the pattern of results is not consistent among the literature reporting that fluency manipulation affects responses for both old and new stimuli (e.g., Dew & Cabeza, 2013; Rajaram & Geraci, 2000). The present study shows that stimulus-irrelevant attentional cueing biases recognition memory. This has significant implications for everyday situations, in that the incidental manipulation of attention could occur in many situations in our daily lives. Imagine that something distracts your attention from a to-be-perceived situation. In this case, you may unintentionally overestimate the perceived fluency for the situation and falsely recognize it as one you have experienced before. Déjà vu may be one manifestation of this bias. Moreover, this kind of memory bias is of serious concern in some cases, such as eyewitness testimony. To explore the generality and applicability of the effect reported here, additional studies using variable materials would be helpful. In particular, the stimuli of social importance, such as human faces or lexical words, may be suitable for further investigation. Lastly, it is important to note several limitations in the present study. One issue regards the discrepancy in results between the preliminary and main experiments. The preliminary results showed that the attentional cue did not change participants’ recognition bias when they explicitly shifted their attention. One could speculate that unintentional processing about the source of fluency might not occur when attention is voluntarily shifted. It might also be possible that participants try to remain unaffected by attentional manipulation when they consciously shift their attention on their own, and intentional memory processing takes precedence over unintentional processing, regarding the source of fluency. Consistent with these speculations, participants exhibited slower RTs in the preliminary experiment than in the main experiment (though this is a between-participant comparison), suggesting that they depended more heavily on intentional memory processing. Future research is needed to explore the factors that modulate the effect of attentional cueing on recognition judgments. It also deserves consideration that the preliminary and main results were consistent, in that old responses were increased in the incongruent condition, though this pattern was prominent for old stimuli in the preliminary experiment, but it was for new stimuli in the main experiment. Possibly, future research with stronger statistical power could detect significant effects of attentional cueing on both old and new stimuli and solve the present discrepancy. Another issue is that there were no control conditions and thus we could only conduct relative comparisons between the congruent and incongruent conditions. Establishing an appropriate control condition enables us to separately assess the effect of congruent and incongruent cues. Finally, it should be noted that the mechanism for attentional cueing to affect processing fluency may differ in quality from that for other experimental manipulations (e.g., repetition priming, conceptual priming, changes in perceptual clarity). Differences and similarities of the effect of these qualitatively different experimental manipulations on recognition judgments remain open for further investigation. 5. Conclusion Attentional cueing biases familiarity-based recognition judgments. This effect is well explained by the hypothesis regarding unintentional processing about the source of fluency. As attentional cueing is a simple yet powerful method for manipulating fluency, it would be of great service for future research to further elucidate the contribution that processing fluency makes to memory processes. Acknowledgments This work was supported by a JSPS Grant-in-Aid for scientific research (22220003) and a Grant-in-Aid for JSPS Fellows (15J06742). K. Miyoshi, H. Ashida / Consciousness and Cognition 43 (2016) 66–74 73 Appendix A. Additional analyses for the preliminary experiment Additional analyses were conducted using the raw proportion of remember and know responses without the IRK procedure (Table 2). An ANOVA on the proportion of responses for new stimuli was conducted with metamemory response (remember/know) and attentional condition (congruent/incongruent) as within-participant factors. A significant main effect was found for metamemory response (F(1, 19) = 51.31, p < 0.001, partial g2 = 0.730), but no significant main effect was found for attentional condition (F(1, 19) = 0.14, p = 0.71, partial g2 = 0.007, BF = 4.2). An interaction between these two factors was not significant (F(1, 19) = 0.002, p = 0.963, partial g2 = 0.0001, BF = 4.5). Next, we conducted an ANOVA on the proportion of responses for old stimuli. A significant main effect was found for metamemory response (F(1, 19) = 22.91, p < 0.001, partial g2 = 0.547) and a main effect for attentional condition was approaching significance (F(1, 19) = 3.39, p = 0.081, partial g2 = 0.152, BF = 0.9). An interaction between these two factors was nonsignificant (F(1, 19) = 1.60, p = 0.222, partial g2 = 0.078, BF = 2.0). Further, we analyzed signal detection measures (d0 and c) for old/new responses pooled across metamemory responses. Response accuracy (d0 ) in the congruent condition (M = 1.90, SD = 0.68) was not significantly different from that in the incongruent condition (M = 2.02, SD = 0.77) (t(19) = 1.30, p = 0.211, d0 = 0.29, BF = 1.9). Moreover, response criterion (c) in the congruent condition (M = 0.17, SD = 0.23) was not significantly different from that in the incongruent condition (M = 0.13, SD = 0.25) (t(19) = 1.40, p = 0.176, d0 = 0.31, BF = 1.7). In short, these additional analyses showed that attentional cueing did not have significant effects on participants’ raw responses without the IRK procedure. Appendix B. Additional analyses for the main experiment To follow up the main results, additional analyses were conducted using the raw proportion of remember and know responses without the IRK procedure (Table 3). An ANOVA on the proportion of responses for new stimuli was conducted with metamemory response (remember/know) and attentional condition (congruent/incongruent) as within-participant factors. A significant main effect was found for metamemory response (F(1, 21) = 75.63, p < 0.001, partial g2 = 0.783), but no significant main effect was found for attentional condition (F(1, 21) = 3.48, p = 0.076, partial g2 = 0.142, BF = 0.9). An interaction between these two factors was nearly significant (F(1, 21) = 4.44, p = 0.054, partial g2 = 0.166, BF = 0.6). A simple main effect analysis revealed that the proportion of know responses was significantly higher in the incongruent condition than in the congruent condition (F(1, 42) = 7.50, p = 0.009, partial g2 = 0.263). No significant difference between the proportion of remember responses in the congruent condition and that in the incongruent condition was seen (F(1, 42) = 0.02, p = 0.892, partial g2 = 0.001, BF = 4.6). Next, we conducted an ANOVA on the proportion of responses for old stimuli. A significant main effect was found for metamemory response (F(1, 21) = 16.29, p = 0.001, partial g2 = 0.437). There was no significant main effect for attentional condition (F(1, 21) = 0.80, p = 0.382, partial g2 = 0.037, BF = 3.1). An interaction between these two factors was nonsignificant (F(1, 21) = 0.01, p = 0.906, partial g2 = 0.001, BF = 4.6). To further support the main findings, we analyzed signal detection measures (d0 and c) for old/new responses pooled across metamemory responses. Response accuracy (d0 ) in the congruent condition (M = 1.75, SD = 0.53) was not significantly different from that in the incongruent condition (M = 1.73, SD = 0.69) (t(21) = 0.28, p = 0.782, d0 = 0.06, BF = 4.5). However, response criterion (c) in the incongruent condition (M = 0.06, SD = 0.26) was marginally significantly liberal than that in the congruent condition (M = 0.16, SD = 0.25) (t(21) = 1.93, p = 0.067, d0 = 0.41, BF = 0.8). In sum, these additional results are consistent with the main results under the IRK assumption. References Atkinson, R. C., & Juola, J. F. (1974). Search and decision processes in recognition memory. In D. H. Krantz, R. C. Atkinson, & P. Suppes (Eds.), Contemporary developments in mathematical psychology (pp. 243–290). San Francisco: Freeman. Bayliss, A. P., Paul, M. A., Cannon, P. R., & Tipper, S. P. (2006). Gaze cuing affective judgments of objects: I like what you look at. Psychonomic Bulletin & Review, 13(6), 1061–1066. Campbell, J. I. D., & Thompson, V. A. (2012). More Power 6.0 for ANOVA with relational confidence intervals and Bayesian analysis. Behavior Research Methods, 44(4), 1255–1265. Cohen, J. (1992). A power primer. Psychological Bulletin, 112(1), 155–159. Dew, I. T. Z., & Cabeza, R. (2013). A broader view of perirhinal function: From recognition memory to fluency-based decisions. The Journal of Neuroscience, 33 (36), 14466–14474. Fischer, J., & Whitney, D. (2009). Attention narrows position tuning of population responses in V1. Current Biology, 19(16), 1356–1361. Gardiner, J. M. (1988). Functional aspects of recollective experience. Memory & Cognition, 16(4), 309–313. Hicks, J. L., Marsh, R. L., & Ritschel, L. (2002). The role of recollection and partial information in source monitoring. Journal of Experimental Psychology: Learning, Memory, and Cognition, 28(3), 503–508. Higham, P. A., & Vokey, J. R. (2000). Judgment heuristics and recognition memory: Prime identification and target-processing fluency. Memory & Cognition, 28(4), 574–584. Hintzman, D. L., & Curran, T. (1994). Retrieval dynamics of recognition and frequency judgments: Evidence for separate processes of familiarity and recall. Journal of Memory and Language, 33(1), 1–18. Huber, D. E., Clark, T. F., Curran, T., & Winkielman, P. (2008). Effects of repetition priming on recognition memory: Testing a perceptual fluency–disfluency model. Journal of Experimental Psychology: Learning, Memory, and Cognition, 34(6), 1305–1324. 74 K. Miyoshi, H. Ashida / Consciousness and Cognition 43 (2016) 66–74 Jacoby, L. L., & Dallas, M. (1981). On the relationship between autobiographical memory and perceptual learning. Journal of Experimental Psychology: General, 110(3), 306–340. Jacoby, L. L., & Whitehouse, K. (1989). An illusion of memory: False recognition influenced by unconscious perception. Journal of Experimental Psychology: General, 118(2), 126–135. Jeffreys, H. (1961). Theory of probability.Oxford, U.K.: Oxford University Press. Johnson, J. D., McDuff, S. G. R., Rugg, M. D., & Norman, K. A. (2009). Recollection, familiarity, and cortical reinstatement: A multivoxel pattern analysis. Neuron, 63(5), 697–708. Klinger, M. R. (2001). The roles of attention and awareness in the false recognition effect. The American Journal of Psychology, 114(1), 93–114. Kurilla, B. P., & Westerman, D. L. (2008). Processing fluency affects subjective claims of recollection. Memory & Cognition, 36(1), 82–92. Lucas, H. D., Taylor, J. R., Henson, R. N., & Paller, K. A. (2012). Many roads lead to recognition: Electrophysiological correlates of familiarity derived from short-term masked repetition priming. Neuropsychologia, 50(13), 3041–3052. Mandler, G. (1980). Recognizing: The judgment of previous occurrence. Psychological Review, 87(3), 252–271. Posner, M. I. (1980). Orienting of attention. Quarterly Journal of Experimental Psychology, 32(1), 3–25. Raftery, A. E. (1999). Bayes factors and BIC: Comment on ‘‘A critique of the Bayesian information criterion for model selection”. Sociological Methods and Research, 27(3), 411–427. Raftery, A. E. (1995). Bayesian model selection in social research. In P. V. Marsden (Ed.), Sociological methodology 1995 (pp. 111–196). Cambridge, MA: Blackwell. Rajaram, S., & Geraci, L. (2000). Conceptual fluency selectively influences knowing. Journal of Experimental Psychology: Learning, Memory, and Cognition, 26 (4), 1070–1074. Richardson, J. T. E. (2011). Eta squared and partial eta squared as measures of effect size in educational research. Educational Research Review, 6(2), 135–147. Ristic, J., & Kingstone, A. (2006). Attention to arrows: Pointing to a new direction. Quarterly Journal of Experimental Psychology, 59(11), 1921–1930. Scariano, S. M., & Davenport, J. M. (1987). The effects of violations of independence assumptions in a one-way ANOVA. The American Statistician, 41(2), 123–129. Shepherd, M., Findlay, J. M., & Hockey, R. J. (1986). The relationship between eye movements and spatial attention. Quarterly Journal of Experimental Psychology Section A: Human Experimental Psychology, 38(3), 475–491. Snodgrass, J. G., & Corwin, J. (1988). Pragmatics of measuring recognition memory: Applications to dementia and amnesia. Journal of Experimental Psychology: General, 117(1), 34–50. Taylor, J. R., Buratto, L. G., & Henson, R. N. (2013). Behavioral and neural evidence for masked conceptual priming of recollection. Cortex, 49(6), 1511–1525. Tipples, J. (2002). Eye gaze is not unique: Automatic orienting in response to uninformative arrows. Psychonomic Bulletin & Review, 9(2), 314–318. Tulving, E. (1985). Memory and consciousness. Canadian Psychology, 26(1), 1–12. Tunney, R. J., & Fernie, G. (2007). Repetition priming affects guessing not familiarity. Behavioral and Brain Functions, 3(1), 40–46. Wais, P. E., Mickes, L., & Wixted, J. T. (2008). Remember/know judgments probe degrees of recollection. Journal of Cognitive Neuroscience, 20(3), 400–405. Westerman, D. L. (2008). Relative fluency and illusions of recognition memory. Psychonomic Bulletin & Review, 15(6), 1196–1200. Whittlesea, B. W. A., & Williams, L. D. (1998). Why do strangers feel familiar, but friends don’t? A discrepancy-attribution account of feelings of familiarity. Acta Psychologica, 98(2–3), 141–165. Whittlesea, B. W., & Williams, L. D. (2000). The source of feelings of familiarity: The discrepancy-attribution hypothesis. Journal of Experimental Psychology: Learning, Memory, and Cognition, 26(3), 547–565. Whittlesea, B. W., & Williams, L. D. (2001). The discrepancy-attribution hypothesis: I. The heuristic basis of feelings of familiarity. Journal of Experimental Psychology: Learning, Memory, and Cognition, 27(1), 3–13. Wixted, J. T. (2007). Dual-process theory and signal detection theory of recognition memory. Psychological Review, 114(1), 152–176. Wixted, J. T., & Mickes, L. (2010). A continuous dual-process model of remember/know judgments. Psychological Review, 117(4), 1025–1054. Woollams, A. M., Taylor, J. R., Karayanidis, F., & Henson, R. N. (2008). Event-related potential associated with masked priming of test cues reveal multiple potential contributions to recognition memory. Journal of Cognitive Neuroscience, 20(6), 1114–1129. Yonelinas, A. P. (1994). Receiver-operating characteristics in recognition memory: Evidence for a dual-process model. Journal of Experimental Psychology: Learning, Memory, and Cognition, 20(6), 1341–1354. Yonelinas, A. P. (2002). The nature of recollection and familiarity: A review of 30 years of research. Journal of Memory and Language, 46(3), 441–517. Yonelinas, A. P., & Jacoby, L. L. (1995). The relation between remembering and knowing as bases for recognition: Effects of size congruency. Journal of Memory and Language, 34(5), 622–643.

Consciousness and Cognition 19 (2010) 1105–1106 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Commentary Mind the gap: Misdirection, inattentional blindness and the relationship between overt and covert attention q Aidan Moran *, Nuala Brady School of Psychology, University College, Dublin, Ireland a r t i c l e i n f o Article history: Available online 3 April 2010 Keywords: Attention Overt attention Covert attention Misdirection Inattentional blindness a b s t r a c t The present commentary addresses two issues arising from Memmert’s (2010) paper. First, can the ‘misdirection’ and ‘inattentional blindness’ paradigms provide important insights into the relationship between ‘overt’ and ‘covert’ attentional processes? Second, what are the most fruitful directions for research that seeks to combine these attentional paradigms in ecologically valid settings? We argue that although Memmert’s (2010) paper postulates several important differences between the misdirection and inattentional blindness paradigms, it may not emphasise sufficiently strongly the significant insights into attention that have been yielded by the former approach. To illustrate, we discuss the utility of the misdirection paradigm in providing an ecologically valid method to investigate the relationship between overt and covert attentional processes. Such naturalistic methods are required to ensure optimal integration of the misdirection and inattentional blindness paradigms within a general theory of attention. Ó 2010 Elsevier Inc. All rights reserved. Commentaries on commentary papers run the risk of losing sight of the original topic of interest. To avoid this problem, the present commentary addresses two issues arising from Memmert’s (2010) paper. First, can the ‘misdirection’ and ‘inattentional blindness’ paradigms provide important insights into the relationship between ‘overt’ and ‘covert’ attentional processes? Second, what are the most fruitful directions for research that seeks to combine these attentional paradigms in ecologically valid settings? Before tackling these questions, however, some background information is required. Recently, neuroscientists (e.g., Kuhn, Amlani, & Rensink, 2008; Macknik et al., 2008) have shown considerable interest in the use of magicians’ tricks to provide valuable insights into the dynamic nature of attentional processes. This interest stems from the fact that the performance of a magic trick requires a method (how the trick works) to achieve an effect (what the spectator perceives). Typically, magicians use ‘misdirection’ or the strategy of combining cues to divert ‘‘the spectator’s attention away from a secret action” (Macknik et al., 2008, p. 872) – towards the effect and away from the method (Lamont & Wiseman, 1999). Inspired by this idea, Kuhn and his colleagues (e.g., Kuhn & Findlay, 2010; Kuhn, Amlani, et al., 2008) devised a misdirection paradigm to investigate the relationship between visual attention and awareness. In one version of this paradigm, a magician appears to make a cigarette and a lighter ‘disappear’ by diverting observers’ attention away from the hand with which he actually drops the two objects into his lap and towards the hand that is not relevant to the trick. According to Kuhn and Findlay (2010), misdirection is ‘‘analogous to inattentional blindness” (p. 138) – a phenomenon in which people who are engaged in attentionally demanding tasks ‘‘often fail to perceive an unexpected object, even if it appears at fixation” (Mack & Rock, 1998, p. 14). q Commentary on Memmert, D. (2010). The gap between inattentional blindness and attentional misdirection. Consciousness and Cognition, 19, 1097–1101. * Corresponding author at: School of Psychology, University College, Dublin, Belfield, Dublin 4, Ireland. E-mail address: Aidan.Moran@ucd.ie (A. Moran). 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.03.011 1106 A. Moran, N. Brady / Consciousness and Cognition 19 (2010) 1105–1106 Although Memmert’s (2010) paper is valuable in postulating important differences between the misdirection and inattentional blindness paradigms (two of which are acknowledged by Kuhn and Findlay (2010)), it may not emphasise strongly enough the significant insights into attention already yielded by the former approach. For example, consider Kuhn, Tatler, Findlay and Cole’s (2008) discovery that spectators’ eye positions at the moment of the cigarette drop in the misdirection paradigm were not related to whether nor not they realised how the trick was done. Instead, those spectators who, after the drop, shifted their fixation quickly to the empty cigarette hand were more likely to report having seen the cigarette fall than those who missed it. From such evidence, Kuhn, Tatler, et al. (2008) concluded that in a magic trick, it is not really where the spectators’ gaze is directed that matters – but instead, where they focus their attention. Clearly, this conclusion highlights the need to distinguish between overt attention (the shifting of attention in space by means of saccades and eye fixations) and covert attention (orienting to stimuli using internal neural adjustments without the movement of the eyes) (Wu & Remington, 2003). Recently, Kuhn and Findlay (2010) used eye tracking technology to compare the scan paths of those who detect the lighter drop (thereby missing the magic) with those who do not. Thirteen of twenty participants (65%) in the ‘real’ trick condition – where the drop, made with the magician’s left hand, was visible on video – reported that they saw the drop, whereas no participant in the ‘fake’ trick condition, where the drop was edited out of the video, reported seeing the drop. Of the group which detected the drop, only four participants were fixating in the vicinity of the lighter when it dropped whereas nine participants were fixating on the magician’s face or in the vicinity of his right hand which was engaged in misdirecting the observers’ attention away from the drop. These nine observers presumably detected the drop covertly using peripheral vision. Interestingly, these ‘peripheral detectors’ were not distinguished from the ‘non-detectors’ in the ‘real’ trick condition where they were fixating at the time the lighter was dropped. Indeed, on average, the two groups fixated equally far away from the region of the drop. However, the nine ‘peripheral detectors’ were distinguished by the pattern of their fixations subsequent to the drop. Specifically, they moved rapidly to the region where the lighter was dropped, fixating this region significantly earlier in the video sequence than those who missed the drop. This suggests that the detection of the event led to the subsequent fixation of the region, and not that the fixation of the region led to the detection of the drop. Such a conclusion is strengthened by the fact that those who missed the drop in the ‘real’ trick condition, and those who had no opportunity to see the drop in the ‘fake’ trick condition, eventually fixated the region of the drop at a comparable point in the video sequence to each other, and significantly later than the ‘peripheral detectors’. Overall, Kuhn and Findlay (2010) interpret their result as showing a close relationship between covert and overt attention. The fact that the overt shift in attention follows the covert detection of the lighter drop in the misdirection paradigm supports the idea that covert and overt attention are linked. For example, in the premotor theory of attention covert attention is assumed to arise from, and reflect the planning of, a saccade (Rizzolatti, Riggio, Dascola, & Umilta, 1987). Turning to the second topic of this commentary, Memmert (2010) identifies some naturalistic, dual-task environments (e.g., flying, driving) where attentional diversion could be investigated using the misdirection and inattentional blindness paradigms. Of particular interest in this regard is the domain of competitive sport which, when studied using head-mounted eye-trackers, offers a rich and dynamic natural laboratory for the study of visual attention–action relationships (Moran, 2009). A key question here is whether or not expert athletes – who are known to differ from relative novices in a host of cognitive processes such as anticipation skills (Williams & Ford, 2008; Yarrow, Brown, & Krakauer, 2009) – are as susceptible to inattentional blindness manipulations as are less skilled counterparts. Surprisingly, Memmert and Furley (2007) found that even skilled athletes may fail to detect an unmarked team-mate when their attention is diverted towards an immediate opponent. Clearly, this finding could be investigated further by including such variables as type of sport (team versus individual), level of expertise of the performer, and type of distractions encountered (predictable versus unpredictable). Finally, we welcome Memmert’s (2010) conclusion that, despite their differences, the misdirection and inattentional blindness paradigms can be combined fruitfully to investigate overt and covert attentional processes in naturalistic settings. References Kuhn, G., Amlani, A. A., & Rensink, R. A. (2008). Towards a science of magic. Trends in Cognitive Sciences, 12, 349–354. Kuhn, G., & Findlay, J. M. (2010). Misdirection, attention and awareness: Inattentional blindness reveals temporal relationship between eye movements and visual awareness. The Quarterly Journal of Experimental Psychology, 63, 136–146. Kuhn, G., Tatler, B. W., Findlay, J. M., & Cole, G. C. (2008). Misdirection in magic: Implications for the relationship between eye gaze and attention. Visual Cognition, 16, 391–405. Lamont, P., & Wiseman, R. (1999). Magic in theory. Hatfield, Herts: University of Hertfordshire Press. Mack, A., & Rock, I. (1998). Inattentional blindness. Cambridge, MA: MIT Press. Macknik, S. L., King, M., Randi, J., Robbins, A., Teller Thompson, J., et al (2008). Attention and awareness in stage magic: Turning tricks into research. Nature Reviews: Neuroscience, 9, 871–879. Memmert, D. (2010). The gap between inattentional blindness and attentional misdirection. Consciousness and Cognition, 19, 1097–1101. Memmert, D., & Furley, P. (2007). ‘I spy with my little eye’ – Breadth of attention, inattentional blindness and tactical decision making in team sports. Journal of Sport & Exercise Psychology, 29, 365–381. Moran (2009). Cognitive psychology in sport: Progress and prospects. Psychology of Sport and Exercise, 10, 420–426. Rizzolatti, G., Riggio, L., Dascola, I., & Umilta, C. (1987). Reorienting attention across the horizontal and vertical meridians: Evidence in favour of a premotor theory of attention. Neuropsychologia, 25, 31–40. Williams, A. M., & Ford, P. R. (2008). Expertise and expert performance in sport. International Review of Sport & Exercise Psychology, 1, 4–18. Wu, S., & Remington, R. (2003). Characteristics of covert and overt visual orienting: Evidence from attentional and oculomotor capture. Journal of Experimental Psychology: Human Perception and Performance, 29, 1050–1067. Yarrow, K., Brown, P., & Krakauer, J. W. (2009). Inside the brain of an elite athlete: The neural processes that support high achievement in sports. Nature Reviews: Neuroscience, 10, 585–596.

Consciousness and Cognition 19 (2010) 1102–1104 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Commentary What’s ‘‘inattentional” about inattentional blindness? q Steven B. Most Department of Psychology, University of Delaware, 108 Wolf Hall, Newark, DE 19716-2577, United States a r t i c l e i n f o Article history: Received 19 January 2010 9 September 2010 Available online 23 February 2010 Keywords: Inattentional blindness Spatial attention Awareness Attentional misdirection Magic Attention Perception a b s t r a c t In a recent commentary, Memmert critiqued claims that attentional misdirection is directly analogous to inattentional blindness (IB) and cautioned against assuming too close a similarity between the two phenomena. One important difference highlighted in his analysis is that most lab-based inductions of IB rely on the taxing of attention through a demanding primary task, whereas attentional misdirection typically involves simply the orchestration of spatial attention. The present commentary argues that, rather than reflecting a complete dissociation between IB and attentional misdirection, this difference highlights potential grounds for delineating mechanistically distinct forms of IB: spatial inattentional blindness, which stems from the covert misallocation of spatial attention, and central inattentional blindness, which stems from disruption or preoccupation of perceptual mechanisms that interface with higher-level processes such as working memory. Recognition of such distinctions can help situate theoretical understanding of IB more firmly within the context of the broader attention literature. Ó 2010 Elsevier Inc. All rights reserved. Recently, Kuhn and colleagues developed a line of investigation into the relationship between misdirection of spatial attention and participants’ failures to notice salient objects (e.g., Kuhn & Findlay, 2010; Kuhn & Tatler, 2005; Kuhn, Tatler, Findlay, & Cole, 2008). In doing so, they suggested that the consequences of such attentional misdirection (and similar misdirection used by stage magicians; see Macknik et al. (2008)) is directly analogous to ‘‘inattentional blindness” (IB). In a valuable commentary, Memmert (2010) critiques this claim and cautions against assuming too close a similarity between these phenomena. Among his reasons, he highlights the following: typically, IB is induced by engaging participants in an attentionally demanding primary task, whereas the attentional misdirection procedure simply manipulates where participants aim their ‘‘spotlight” of spatial attention. Memmert’s point is well taken; taking it a step further, it might be that buried within this distinction lie grounds for delineating mechanistically different forms of IB. Inattentional blindness refers to the common failure to notice plainly visible items when attention is otherwise preoccupied, even though people look directly at them (e.g., Mack & Rock, 1998; Most, Scholl, Clifford, & Simons, 2005; Most et al., 2001; Neisser & Dube, 1978, cited in Neisser, 1979; Simons & Chabris, 1999). In many cases, IB stems from people ‘‘covertly” (i.e., independent of eye movements) directing spatial attention away from the item in question. For example, in one series of experiments, participants judged the relative lengths of the arms of a briefly presented cross and an unexpected additional item could appear in one of the cross’s quadrants (Mack & Rock, 1998). With this small spatial separation between the cross’s arms and the unexpected object, about 25% of people failed to detect the unexpected item. In a different variation, participants tracked a subset of items moving randomly within a computerized display and counted the number of times that these items touched a horizontal line bisecting the display (Most, Simons, Scholl, & Chabris, 2000). On a critical trial, an unexpected cross traveled horizontally through the display at a variable distance from the line, which was presumably the focus of q Commentary on Memmert, D. (2010). The gap between inattentional blindness and attentional misdirection. Consciousness and Cognition, 19, 1097–1101. E-mail address: most@psych.udel.edu 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.01.011 S.B. Most / Consciousness and Cognition 19 (2010) 1102–1104 1103 spatial attention. The further the cross appeared from the line, the less likely people were to notice it. That is, IB was more likely to occur when spatial attention was directed away from the unexpected item (see also Mack & Rock, 1998; Newby & Rock, 1998). Note that this particular demonstration of IB is indeed analogous to attentional misdirection. However, IB can also stem from aspects of selection that are independent of where people attend. For example, in the aforementioned study, fewer than half of the participants noticed the unexpected cross even when it traveled on the line at the focus of spatial attention (Most et al., 2000). Similarly, in other studies, people failed to notice the critical stimulus even though it intermingled with and often overlapped the items that people were tracking (Becklen & Cervone, 1983; Most et al., 2001, 2005; Neisser & Becklen, 1975; Neisser & Dube, 1978, cited in Neisser, 1979; Simons & Chabris, 1999). Eye-tracking studies (which employ an overt, if not covert, index of attention) confirm that patterns of fixation do not differentiate between people who notice and people who fail to notice unexpected objects (Koivisto, Hyönä, & Revonsuo, 2004; Memmert, 2006), and evidence from beyond the inattentional blindness literature converges to suggest that although spatial attention enhances the quality of information available to later stages of perception (e.g., Carrasco, Ling, & Read, 2004; Carrasco, Williams, & Yeshurun, 2002), it is not sufficient by itself to support visual awareness (Kentridge, Heywood, & Weiskrantz, 1999, 2004; Lambert, Naikar, McLachlan, & Aitken, 1999; McCormick, 1997; Woodman & Luck, 2003). Direct evidence that IB can arise from bottlenecks independent of the locus of spatial attention comes from studies in which IB was induced simply by placing participants under heavy working memory and/or executive load (Fougnie & Marois, 2007; Todd, Fougnie, & Marois, 2005). Thus, even when an object is the focus of spatial attention, failures to see it can occur due to preoccupation of more central, late-stage bottlenecks critical to perception. In short, the overarching term ‘‘inattentional blindness” likely obscures mechanistic distinctions between at least two different sub-types, one driven by covert allocation of spatial attention and a second driven by preoccupation or disruption of non-spatial selection mechanisms that make the difference between what Block has called ‘‘phenomenal consciousness” and ‘‘cognitive access” to perceptual representations (Block, 2007). Although phenomenologically similar, these two types of IB link mechanistically to somewhat different literatures. IB driven by spatial attention, which could be termed spatial inattentional blindness, connects naturally with the spatial cueing literature (e.g., Posner, 1980). A second type of IB, which could be termed central inattentional blindness, links more closely with phenomena such as the attentional blink, repetition blindness, and object substitution masking, which reveal failures of visual awareness stemming from late-stage bottlenecks, such as those that interface with visual working memory and contribute to the individuation of object representations (e.g., Chun & Potter, 1995; Di Lollo, Enns, & Rensink, 2000; Enns & Di Lollo, 1997; Kanwisher, 1987; Moore & Lleras, 2005; Raymond, Shapiro, & Arnell, 1992).1 Some manipulations that have been found to modulate rates of IB, such as participants’ expectations about the number of items that will appear in a display (White & Davies, 2008) or prioritization of items on the basis of semantic meaning (Koivisto & Revonsuo, 2007), might operate though their impact on mechanisms related to central-IB.2 Notably, the distinction between spatial- and central-IB has potential implications for the notion of ‘‘inattentional amnesia”, the suggestion that IB reflects not a failure of perception, but rather the rapid forgetting of information at a post-perceptual stage of processing (Wolfe, 1999). Whereas central-IB might very well reflect the type of memorial processes implicated in this construct, spatial-IB might reflect a more profound failure of initial perceptual encoding. This distinction also has potential relevance to a second disconnection highlighted by Memmert: IB typically occurs only for unexpected stimuli, whereas attentional misdirection can induce failures to notice even expected stimuli. One possibility is that unexpectedness is more critical for the induction of central-IB than of spatial-IB. Note, however, that this distinction might be more practical than theoretical in nature. If it were possible to ensure that central resources were 100% preoccupied, then one might find high rates of IB even for stimuli that are expected. To conclude, the term ‘‘inattentional blindness” has to date been used largely to refer to a phenomenologically related family of instances where people fail to see objects and events due to a preoccupation of attention. At a mechanistic level, however, although the term captures the essence of the experience, it may obscure subtleties in the meaning of ‘‘inattentional”. Memmert is right to urge caution when drawing too direct an analogy between IB and the misdirection of spatial attention. However, it might be that such caution is warranted not because the two are entirely dissociable, but because the orchestration of spatial attention represents only one tile in the mosaic of inattentional blindness phenomena. Acknowledgments Thanks to Dan Simons and Adam Grant for feedback on early drafts of this manuscript. 1 Central-IB might conceivably be sub-divided further into additional sub-types stemming from different perceptual bottlenecks. For example, although the attentional blink and object substitution masking are both thought to arise from mechanisms beyond the allocation of spatial attention, evidence suggests that undetected items in the former elicit neural signatures of semantic activation and category-specific identification (Luck, Vogel, & Shapiro, 1996; Marois, Yi, & Chun, 2004), whereas those in the latter do not (Reiss & Hoffman, 2006, 2007). 2 IB also depends robustly on other non-spatial factors, such as how people ‘‘tune” attention to prioritize certain visual features (i.e., on their feature-based attentional set), with stimuli that match one’s attentional set more likely to reach awareness than those that do not (Most & Astur, 2007; Most et al., 2001, 2005; Simons & Chabris, 1999). Although such attentional sets might impact mechanisms related to central-IB, it is alternatively possible that they reveal yet a third class of IB phenomena: feature-based IB, which could reflect the filtering of stimuli at a stage too early to qualify as central-IB, but which nevertheless depends on factors that are partially or wholly independent of spatial attention (e.g., see Desimone & Duncan, 1995). 1104 S.B. Most / Consciousness and Cognition 19 (2010) 1102–1104 References Becklen, R., & Cervone, D. (1983). Selective looking and the noticing of unexpected events. Memory & Cognition, 11, 601–608. Block, N. (2007). Consciousness, accessibility, and the mesh between psychology and neuroscience. Behavioral and Brain Sciences, 30, 481–548. Carrasco, M., Ling, S., & Read, S. (2004). Attention alters appearance. Nature Neuroscience, 7, 308–313. Carrasco, M., Williams, P. E., & Yeshurun, Y. (2002). Covert attention increases spatial resolution with or without masks: Support for signal enhancement. Journal of Vision, 2(6), 467–479. Chun, M. M., & Potter, M. C. (1995). A two-stage model for multiple target detection in raped serial visual presentation. Journal of Experimental Psychology: Human Perception and Performance, 21, 109–127. Desimone, R., & Duncan, J. (1995). Neural mechanisms of selective visual attention. Annual Review of Neuroscience, 18, 193–222. Di Lollo, V., Enns, J. T., & Rensink, R. A. (2000). Competition for consciousness among visual events: The psychophysics of reentrant visual processes. Journal of Experimental Psychology: General, 129, 481–507. Enns, J. T., & Di Lollo, V. (1997). Object substitution: A new form of masking in unattended visual locations. Psychological Science, 8, 135–139. Fougnie, D., & Marois, R. (2007). Executive working memory load induces inattentional blindness. Psychonomic Bulletin & Review, 14(1), 142–147. Kanwisher, N. (1987). Repetition blindness: Type recognition without token individuation. Cognition, 27, 117–143. Kentridge, R. W., Heywood, C. A., & Weiskrantz, L. (1999). Attention without awareness in blindsight. Proceedings of the Royal Society of London, Series B, 266, 1805–1811. Kentridge, R. W., Heywood, C. A., & Weiskrantz, L. (2004). Spatial attention speeds discrimination without awareness in blindsight. Neuropsychologia, 42, 831–835. Koivisto, M., Hyönä, J., & Revonsuo, A. (2004). The effects of eye movements, spatial attention, and stimulus features on inattentional blindness. Vision Research, 44, 3211–3221. Koivisto, M., & Revonsuo, A. (2007). How meaning shapes seeing. Psychological Science, 18, 845–849. Kuhn, G., & Findlay, J. M. (2010). Misdirection, attention and awareness: Inattentional blindness reveals temporal relationship between eye movements and visual awareness. Quarterly Journal of Experimental Psychology, 63, 136–146. Kuhn, G., & Tatler, B. W. (2005). Magic and fixation: Now you don’t see it, now you do. Perception, 34, 1153–1161. Kuhn, G., Tatler, B. W., Findlay, J. M., & Cole, G. G. (2008). Misdirection in magic: Implications for the relationship between eye gaze and attention. Visual Cognition, 16, 391–405. Lambert, A., Naikar, N., McLachlan, K., & Aitken, V. (1999). A new component of visual orienting: Implicit effects of peripheral information and subthreshold cues on covert attention. Journal of Experimental Psychology: Human Perception and Performance, 25, 321–340. Luck, S. J., Vogel, E. K., & Shapiro, K. L. (1996). Word meanings can be accessed but not reported during the attentional blink. Nature, 383, 616–618. Mack, A., & Rock, I. (1998). Inattentional blindness. Cambridge, MA: MIT Press. Macknik, S. L., King, M., Randi, J., Robbis, A., Teller Thompson, J., et al (2008). Attention and awareness in stage magic: Turning tricks into research. Nature Reviews Neuroscience, 9, 871–879. Marois, R., Yi, D. J., & Chun, M. M. (2004). The neural fate of perceived and missed events in the attentional blink. Neuron, 41(3), 465–472. McCormick, P. A. (1997). Orienting attention without awareness. Journal of Experimental Psychology: Human Perception and Performance, 23, 168–180. Memmert, D. (2006). The effects of eye movements, age, and expertise on inattentional blindness. Consciousness and Cognition, 15, 620–627. Memmert, D. (2010). The gap between inattentional blindness and attentional misdirection. Consciousness and Cognition, 19, 1097–1101. Moore, C. M., & Lleras, A. (2005). On the role of object representations in substitution masking. Journal of Experimental Psychology: Human Perception and Performance, 31, 1171–1180. Most, S. B., & Astur, R. S. (2007). Feature-based attentional set as a cause of traffic accidents. Visual Cognition, 15, 125–132. Most, S. B., Scholl, B. J., Clifford, E., & Simons, D. J. (2005). What you see is what you set: Sustained inattentional blindness and the capture of awareness. Psychological Review, 112, 217–242. Most, S. B., Simons, D. J., Scholl, B. J., & Chabris, C. F. (2000). Sustained inattentional blindness: The role of location in the detection of unexpected dynamic events. Psyche, 6(14). <http://journalpsyche.org/ojs-2.2/index.php/psyche/article/view/2551/2471>. Most, S. B., Simons, D. J., Scholl, B. J., Jimenez, R., Clifford, E., & Chabris, C. F. (2001). How not to be seen: The contribution of similarity and selective ignoring to sustained inattentional blindness. Psychological Science, 12, 9–17. Neisser, U. (1979). The control of information pickup in selective looking. In A. D. Pick (Ed.), Perception and its development: A tribute to Eleanor J. Gibson (pp. 201–219). Hillsdale, NJ: Erlbaum. Neisser, U., & Becklen, R. (1975). Selective looking: Attending to visually specified events. Cognitive Psychology, 7, 480–494. Newby, E. A., & Rock, I. (1998). Inattentional blindness as a function of proximity to the focus of attention. Perception, 27, 1025–1040. Posner, M. I. (1980). Orienting of attention. Quarterly Journal of Experimental Psychology, 32, 3–25. Raymond, J. E., Shapiro, K. L., & Arnell, K. M. (1992). Temporary suppression of visual processing in an RSVP task: An attentional blink? Journal of Experimental Psychology: Human Perception and Performance, 18, 849–860. Reiss, J. E., & Hoffman, J. E. (2006). Object substitution masking interferes with semantic processing: Evidence from ERPs. Psychological Science, 12, 1015–1020. Reiss, J. E., & Hoffman, J. E. (2007). Disruption of early face recognition processes by object substitution masking. Visual Cognition, 15, 789–798 (Shapiro et al., 1997). Simons, D. J., & Chabris, C. F. (1999). Gorillas in our midst: Sustained inattentional blindness for dynamic events. Perception, 28, 1059–1074. Todd, J. J., Fougnie, D., & Marois, R. (2005). Visual short-term memory load suppresses temporo-parietal junction activity and induces inattentional blindness. Psychological Science, 16(12), 965–972. White, R. C., & Davies, A. A. (2008). Attention set for number: Expectation and perceptual load in inattentional blindness. Journal of Experimental Psychology: Human Perception and Performance, 34, 1092–1107. Wolfe, J. M. (1999). Inattentional amnesia. In V. Coltheart (Ed.), Fleeting memories: Cognition of brief visual stimuli (pp. 71–94). Cambridge, MA: MIT Press. Woodman, G. F., & Luck, S. J. (2003). Dissociations among attention, perception, and awareness during object-substitution masking. Psychological Science, 14, 605–611.

Consciousness and Cognition Consciousness and Cognition 14 (2005) 390–425 www.elsevier.com/locate/concog Common fronto-parietal activity in attention, memory, and consciousness: Shared demands on integration? Hamid Reza Naghavia,b,*, Lars Nybergb a Psychiatry and Clinical Psychology Research Center, Tehran University of Medical Sciences, Roozbeh Hospital, South Kargar Street, 13185/1741 Tehran, Iran b Department of Psychology, Umeå University, S-901 87 Umeå, Sweden Received 21 May 2004 Available online 8 December 2004 Abstract Fronto-parietal activity has been frequently observed in fMRI and PET studies of attention, working memory, and episodic memory retrieval. Several recent fMRI studies have also reported fronto-parietal activity during conscious visual perception. A major goal of this review was to assess the degree of anatomical overlap among activation patterns associated with these four functions. A second goal was to shed light on the possible cognitive relationship of processes that relate to common brain activity across functions. For all reviewed functions we observed a consistent and overlapping pattern of brain activity. The overlap was most pronounced for the bilateral parietal cortex (BA 7 and BA 40; close to the intraparietal sulcus), and dorsolateral prefrontal cortex (right BA 9 and left BA 6). The common fronto-parietal activity will be discussed in terms of processes related to integration of distributed representations in the brain. Ó 2004 Elsevier Inc. All rights reserved. Keywords: Conscious perception; Visual consciousness; Attention; Working memory; Episodic memory retrieval; Integration; Frontoparietal network; Functional neuroimaging * Corresponding author. E-mail address: HamidReza.Naghavi@psy.umu.se (H.R. Naghavi). 1053-8100/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2004.10.003 H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 391 1. Introduction Investigations into the neural correlates of consciousness are based on a variety of techniques, including EEG (Desmedt & Tomberg, 1995; Engel & Singer, 2001; Gaetz, Weinberg, Rzempoluck, & Jantzen, 1998; Kaernbach, Schroger, Jacobsen, & Roeber, 1999; Koivisto & Revonsuo, 2003; Ojanen, Revonsuo, & Sams, 2003; Rodriguez et al., 1999; Summerfield, Jack, & Burgess, 2002; Wilenius-Emet, Revonsuo, & Ojanen, 2004), MEG (Gaetz et al., 1998; Joliot, Ribary, & Llinas, 1994; Mäkinen, May, & Tiitinen, 2004; Strüber & Herrmann, 2002; Vanni, Revonsuo, Saarinen, & Hari, 1996; Walla et al., 2002), TMS (Overgaard, Nielsen, & Fuglsang-Frederiksen, 2004; Ro, Breitmeyer, Burton, Singhal, & Lane, 2003), single-cell recordings (Leopold & Logothetis, 1996; Logothetis, 1998; Sengpiel, Blakemore, & Harrad, 1995; Sheinberg & Logothetis, 1997), and studies of brain-damaged patients (Deouell, 2002; Driver & Vuilleumier, 2001; Farah, OÕReilly, & Vecera, 1997; ffytche et al., 1998; Sahraie et al., 1997; Vuilleumier et al., 2002; Whatham, Vuilleumier, Landis, & Safran, 2003). In the present article we will mainly be concerned with results from studies using functional magnetic resonance imaging (fMRI) and to a lesser degree the related method positron emission tomography (PET). With fMRI we specifically mean the common type of fMRI which uses blood oxygenation level dependent (BOLD) contrast, a method for indicating changes in neural activity through measurement of changes in the level of blood oxygenation. In PET changes in blood flow are examined by marking the blood with a radioactive tracer. The results from both fMRI and PET only indirectly reflect neural activity and the methods suffer from limitations in terms of spatial and temporal resolution (Logothetis & Wandell, 2004; Raichle, 1998). Still, fMRI and PET have been widely used for studying cognitive functions and seem to give robust and reproducible results. Several fMRI studies have focused on the neural correlates of awareness of specific phenomenal content in the human visual system. A number of different experimental methods have been used in these studies for inducing changes in visual awareness, such as binocular rivalry (Lee & Blake, 2002; Lumer & Rees, 1999; Lumer, Friston, & Rees, 1998; Polonsky, Blake, Braun, & Heeger, 2000; Tong, Nakayama, Vaughan, & Kanwisher, 1998) and bistable perception (Kleinschmidt, Buchel, Zeki, & Frackowiak, 1998; Sterzer, Russ, Preibisch, & Kleinschmidt, 2002). The results from multiple studies converge to show that different areas of ventral visual cortex are related to visual awareness (see Rees, Kreiman, & Koch, 2002 for a review). However, Rees and colleagues concluded that although activity in ventral visual cortex is a consistent neural correlate of visual awareness, an additional contribution from parietal and prefrontal loci seems necessary. The nature of the cognitive processes that relate to fronto-parietal activity is not clear, but Rees et al. noted that this activation pattern has strong similarities with that associated with various forms of attention. They therefore encouraged work that explored similarities in functional activity and cognitive processes related to awareness and attention. Fronto-parietal activity is also a consistent neural correlate of other cognitive processes, notably working memory and episodic memory (Cabeza & Nyberg, 2000). Several recent publications have been concerned with similarities in brain activity associated with working memory and episodic memory (Braver et al., 2001; Cabeza, Dolcos, Graham, & Nyberg, 2002; Nyberg, Forkstam, Petersson, Cabeza, & Ingvar, 2002, 2003; Ranganath, Johnson, & DÕEsposito, 2003; for reviews, see Cabeza & Nyberg, 2002; Fletcher & Henson, 2001; Nyberg & Cabeza, in press) and also 392 H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 attention and memory (Cabeza et al., 2003; LaBar, Gitelman, Parrish, & Mesulam, 1999; Marklund et al., in press). In the present study we review fMRI and PET activation studies of attention, working memory, episodic memory retrieval, and conscious perception with a special focus on fronto-parietal activity. To select the studies on each of the first three functions, first, we prepared a list of fMRI and PET studies published after 1999 by an electronic search within the SCI-Extended database of ISI Web of Knowledge. The reason why we only included publications after 1999 was that previous reviews (Cabeza & Nyberg, 2000) have covered that literature and revealed an association between fronto-parietal activity and the cognitive functions of interest here (attention, working memory, and episodic retrieval). Then, we excluded publications that: (1) were not peer-reviewed, (2) studied special groups such as young, elderly or abnormal populations, (3) did not examine the whole brain or (4) did not mention Talairach coordinates (Talairach & Tournoux, 1988). Of the remaining studies, for each function, we selected 12 publications that, in our view, were designed to study the most general aspects of the function of interest. Thus, we included studies that varied with regard to the type of test (e.g., recall/recognition) and/or materials (e.g., visual/auditory stimuli); however, we did not include studies that were focused on highly specific aspects of the function. For example, we did not consider reports of differential activity associated with spatial versus auditory attention; as such, a contrast could be expected to exclude common activity related to both attention conditions. While ROI studies were not included, we did consider whole-brain studies that used functional ROIs for data analysis. It should be emphasized that although our review is focused on fronto-parietal areas, the studies were not selected because they showed a specific pattern of activation such as fronto-parietal activity. For conscious perception we followed the same search process and the same inclusion and exclusion criteria as described above, except that all publication years were considered. As it turned out, most of the included functional imaging studies of conscious perception were indeed published after 1999. To limit the scope of the review we left out functional imaging studies of special forms of conscious experience such as mental imagery and perceptual after-effects as well as altered states of consciousness. A major goal of the review was to more precisely assess the degree of anatomical overlap among activation patterns associated with attention, working memory, episodic retrieval, and conscious perception. A second goal was to shed light on the possible cognitive relationship between processes engaged by these functions. The results will show overlapping fronto-parietal activity across functions that will be discussed in relation to shared demands on integration of multiple representations. 2. Selective review of fMRI and PET studies 2.1. Attention The concept of attention refers to one of the basic characteristics of cognition, namely the capacity to voluntarily or involuntarily give priority to some parts of the information that is available at a given moment. This cognitive ability has a critical role in the guidance of our behavior. During the last decade, numerous fMRI and PET studies have tried to shed light on the H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 393 localization as well as the nature of underlying neural mechanisms of attention. Studies up to 1999 showed that fronto-parietal activity was the most consistent activation pattern for attention (Cabeza & Nyberg, 2000). In our review of more recent papers we find that this impression still holds (see also Pessoa, Kastner, & Ungerleider, 2003). The results from recent studies of attention are presented in Table 1. Most of the studies have revealed a distributed system of brain regions that control attention by enhancing the processing of attended aspects of information. Frontal and parietal regions, including areas in the superior parietal lobule, the intraparietal sulcus, and the frontal eye field have consistently been activated in various tasks involving spatially directed attention (Beauchamp, Petit, Ellmore, Ingeholm, & Haxby, 2001; Corbetta, Kincade, & Shulman, 2002; Gitelman et al., 1999; Hopfinger, Buonocore, & Mangun, 2000; Kim et al., 1999; Rosen et al., 1999). Activations in the middle and inferior frontal gyrus, the inferior parietal lobule, and the anterior cingulate cortex have also been observed. There appears to be a general spatial attention network that operates independently of the specific task requirements, although additional brain regions may be involved in each specific task. Evidence for a fronto-parietal network of regions involved in attentional control comes from a wide variety of imaging studies. Gitelman et al. (1999) used stringent control to eliminate the contributions of motor output, visual fixation, inhibition of eye movements, and working memory, and found activations in the frontal eye fields and the posterior parietal cortex. Similarly, Beauchamp et al. (2001) observed that both overt (eye movements towards attended location) and covert (no eye movements) shifts of visuospatial attention induced activations in the precentral sulcus, intraparietal sulcus, and lateral occipital cortex. Endogenous (voluntarily) and exogenous (stimulus-driven) types of attention also yielded comparable patterns of activations. Rosen et al. (1999) found that both exogenous and endogenous orienting activated bilateral parietal and dorsal premotor regions, including the frontal eye fields. Interestingly, the spatial extent and intensity of activation was greatest for the endogenous condition, consistent with the hypothesis that endogenous orienting is more effortful (less automatic) than exogenous orienting. Moreover, the right dorsolateral prefrontal cortex (DLPFC), BA 46, was activated selectively by the endogenous condition. In a similar study, Kim et al. (1999) found that both endogenous and exogenous types of attention activated the frontal eye fields, DLPFC, and posterior parietal cortex, as well as some other regions. Activations in the parietal and frontal areas have been reported not only for visual attention but also in several studies of attention in other modalities. Zatorre, Mondor, and Evans (1999) used PET to investigate whether similar neural systems are involved in attending to spectral and spatial features of sounds. In addition to bilateral auditory cortex they observed increases in the right superior parietal, right dorsolateral frontal, and right premotor regions. In another PET study of auditory spatial attention, Lipschutz, Kolinsky, Damhaut, Wikler, and Goldman (2002) found that when listeners are involved in auditory spatial attention tasks an interacting network of frontal, temporal, and parietal regions is activated. These results are consistent with the idea of a multimodal large-scale attentional network. Some other studies have also investigated the multimodality of attentional neural systems. Shaywitz et al. (2001) used tasks involving processing of printed and/or spoken words and found that selective attention resulted in increased activation in left parietal sites as well as inferior frontal sites. Divided attention resulted in additional increases in activation in the same left hemisphere sites and was also uniquely associated with increased activation of homologous sites in 394 Table 1 H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 Attention H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 Symbols and abbreviations: (s) left lateral; (d) right lateral; ( ) bilateral lateral; (h) left medial; (j) right medial; ( ) bilateral midline; L, left; R, right; ant., anterior; post., posterior; vent., ventral; SMA, supplementary motor area. 395 396 H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 the right hemisphere. Macaluso, Frith, and Driver (2002) studied the spatial and intermodal selection during stimulation of vision and touch, and observed that in the intraparietal sulcus, spatial attentional effects were multimodal, independent of the modality attended. There are still other studies that have investigated specific types of attention such as tactile attention or attention to action. Burton et al. (1999) used PET to study cortical regions mediating tactile attention. In addition to somatosensory foci, attention conditions activated several nonsensory foci in frontal cortex, and bilateral superior parietal cortex. Posterior parietal regions observed in that study were 1.5 cm anterior to the most anterior site described in visuospatial experiments. Burton et al. suggested the posterior parietal region may be subdivided into modality-specific subregions, each of which processes information needed to attend to a specific modality. Rowe, Friston, Frackowiak, and Passingham (2002) investigated attentional selection in the motor system. They found that attention to action increased activity in prefrontal, premotor, and parietal cortex, compared with unattended performance of the same movements. Using event-related fMRI, Hopfinger et al. (2000) investigated the neural sources of top-down attentional control. They found that a distributed network, including the superior frontal gyrus, midfrontal gyrus, superior parietal lobule, and intraparietal sulcus, as well as superior temporal gyrus is involved in the control of spatial attention. Hopfinger et al. suggested that activity in these areas biases the neural activity in sensory brain structures coding the spatial locations of upcoming target stimuli. Corbetta, Kincade, Ollinger, McAvoy, and Shulman (2000) observed that maintaining attention toward a relevant location before target presentation is associated with activity in the intraparietal sulcus, suggesting that the intraparietal sulcus is a top-down source of biasing signals observed in visual cortex. In a recent fMRI study, Corbetta and Shulman (2002), exploring neural substrates of spatial attention, found that a dorsal fronto-parietal system, including superior parietal lobule, intraparietal sulcus, and the frontal eye field, is involved in the generation of attentional sets associated with goal-directed stimulus–response selection. A second, ventral system was involved in detecting behaviorally relevant stimuli. This latter network, which was strongly lateralized to the right hemisphere, involved the temporoparietal junction (at the intersection of the inferior parietal lobule and the superior temporal gyrus) and the middle and inferior frontal gyri. Yantis et al. (2002) have used event-related fMRI to study the dynamics of attentional control, and observed that shifting the location of attention produces a rapid, transient increase in superior parietal lobule activation. A similar transient activity in superior parietal lobule has also been reported when subjects switch their attention from the house stream to the face stream or vice versa while viewing a stream of spatially superimposed houses and faces (Yantis & Serences, 2003). Yantis and colleagues suggest that the transient activity of superior parietal lobule provides a biasing signal to visual cortex that affects on-going processing to favor the attended location or object. Taken together, imaging studies provide evidence that a distributed fronto-parietal network is generally involved in attentional tasks and may function to generate top-down biasing signals that modulate activity in lower-level sensory systems (Fig. 1A). 2.2. Working memory Working memory refers to a system for temporary storage of information relevant to on-going activities, as well as to the processing operations that make use of this information for guidance of H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 397 Fig. 1. Frequency of regional brain activity within different Brodmann areas associated with: (A) attention, (B) working memory, (C) episodic retrieval, and (D) conscious perception in the studies listed in Tables 1–4, respectively. 398 H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 behavior. Baddeley and Hitch (1974) proposed a multicomponent model of working memory that has been influential within cognitive psychology and neuroscience. The basic model consisted of two independent ‘‘slave systems’’ for the storage of verbal and visuospatial information and a ‘‘Central Executive’’ to distribute attentional control. Working memory is viewed as a fundamental cognitive capacity that contributes critically to many high-level cognitive functions including learning, reasoning, and language comprehension (Baddeley, 1992; Jonides, 1995). Neuroimaging studies have used paradigms such as delayed match to sample and n-back to investigate the neural bases of working memory. In delayed match to sample, which is primarily designed for evaluating maintenance of information in working memory, subjects are shown a set of stimuli (e.g., letters or objects), which disappears after 5–10 s. This is followed by one (or several) single stimuli and the subjectÕs task is to judge whether the matching stimulus was contained in the initial set. n-back tasks can be used for evaluating both maintenance and manipulation of information in working memory. The subject is presented with a series of individual stimuli and has to respond when a stimulus is identical to one that was presented n positions earlier (usually two or three). In line with evidence from brain lesion studies (Luria, 1966) as well as single-cell studies in monkeys (Fuster & Alexander, 1971), most functional imaging studies have focused on the role of the frontal lobes in working memory. However, numerous fMRI and PET studies up to 1999 revealed consistent involvement of not only frontal (especially DLPFC) but also parietal areas (specifically BA 7 and BA 40) in a wide variety of working memory tasks using different kinds of stimuli and focusing on various processing components (Cabeza & Nyberg, 2000; Hartley & Speer, 2000; Smith & Jonides, 1999). More recent imaging studies, which are selectively included in our review and summarized in Table 2, tend to show the same pattern (see also Collette & Van der Linden, 2002). The most obvious trend in recent imaging studies of working memory has been to try to associate specific brain regions with different kinds of or various processing components of working memory based on material-specific (e.g., visuospatial vs. phonological) or process-specific (e.g., maintenance vs. manipulation) dissociations (Hartley & Speer, 2000). Material- versus processbased distinctions have often been seen as conflicting positions, although they are not mutually exclusive [the hybrid model of Postle and DÕEsposito (2000) accounts for both kinds of effects]. 2.2.1. Material-specific dissociations Using PET, Smith, Jonides, and Koeppe (1996) found a clear-cut hemispheric double dissociation, with the right dorsal prefrontal cortex (BA 46) subserving spatial working memory and the left ventral prefrontal cortex (BA 44) as well as left premotor cortex (BA 6) subserving active maintenance within verbal working memory. The left inferior parietal cortex (BA 40) was proposed as a neural substrate for the passive phonological short-term store. However, two subsequent fMRI investigations failed to support the notion of working memory related hemispheric specificity. Activation of left inferior frontal cortex was not observed, and right prefrontal cortex was activated in spatial as well as in verbal working memory (DÕEsposito, Ballard, Aguirre, & Zarahn, 1998), or both regions were active bilaterally without showing any significant dominance effects (Nystrom et al., 2000). In a recent study, Zurowski et al. (2002) found a common frontoparietal network associated with both spatial and phonological working memory in contrast to comparable judgment tasks (without working memory component). Bilateral superior frontal Table 2 Working H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 399 (continued on next page) 400 Symbols and abbreviations: (s) left lateral; (d) right lateral; ( ) bilateral lateral; (h) left medial; (j) right medial; ( ) bilateral midline; L, left; R, right; ant., anterior; post., posterior; vent., ventral; SMA, supplementary motor area; WM, working memory. H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 Tabel 3 (continued) H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 401 sulcus and adjacent gyri, posterior parietal cortex, precuneus, and, less consistently, middle frontal gyrus were involved in the common network. Moreover, they observed that bilateral anterior and posterior intraparietal sulcus, as well as right superior frontal sulcus, exhibited dominance for spatial working memory. However, while robust activation within the left inferior frontal gyrus was present during both phonological working memory and phonological judgment, no region specific for phonological working memory was found. Zurowski et al. (2002) concluded that left prefrontal lateralization for verbal working memory probably reflects more general phonological processing strategies that are not specific for working memory. Several studies have investigated the neural bases of specific types of phonological (verbal) and/ or visuospatial working memory, or even working memory for other modalities such as olfaction. Crottaz-Herbette, Anagnoson, and Menon (2004) examined similarities and differences between visual verbal working memory (vis-VWM) and auditory verbal working memory (aud-VWM) using fMRI and identical stimuli. There was extensive overlap of activation bilaterally in the dorsolateral and ventrolateral prefrontal cortex (VLPFC), intraparietal sulcus, supramarginal gyrus, and the basal ganglia. However, the left posterior parietal cortex, primarily along the intraparietal sulcus, showed greater responses during vis-VWM whereas the left DLPFC showed greater responses during aud-VWM. They also detected bilateral suppression of the superior and middle temporal cortex during vis-VWM, and of the occipital cortex during aud-VWM, suggesting that cross-modal inhibitory processes may help to provide preferential access to high-order heteromodal association areas. These results indicate that although similar prefrontal and parietal regions are involved in aud-VWM and vis-VWM, there are modality differences in the way neural signals are generated, processed, and routed during verbal working memory. Focusing on two specific types of visuospatial working memory, Manoach et al. (2004) investigated spatial and shape working memory using tasks with minimal manipulation demands. The tasks had identical stimuli and required the same motor response. Subjects maintained either the location or the shape of the targets that appeared in particular locations in working memory and responded to each probe by indicating whether it was a target. During a control task, subjects indicated whether the probe appeared on the right or left side of the screen. Both spatial and shape working memory were correlated with activations in bilateral VLPFC, right DLPFC, frontopolar prefrontal cortex, lateral premotor areas, superior parietal lobule, and inferior parietal lobule; however, activity was generally greater in the right and left hemisphere for spatial and shape working memory, respectively. In direct comparisons, spatial working memory was associated with increased right ventrolateral and frontopolar prefrontal cortex activation, whereas shape working memory was associated with increased left VLPFC activity. Although, these findings were interpreted by the authors as supporting hemispheric specialization for spatial versus shape working memory, the results also revealed common involvement of fronto-parietal areas in both types of working memory. Similarly, Mecklinger, Bosch, Gruenewald, Bentin, and von Cramon (2000) used fMRI to examine putative content-specific organization of working memory. In one experiment, either unfamiliar geometrical objects or their spatial locations had to be memorized, whereas in the other experiment either unfamiliar faces or biological objects (butterflies) were memorized. All tasks activated a similar cortical network including the intraparietal sulcus, premotor cortex, the inferior frontal sulcus, and pre-supplementary motor area. For geometrical objects and faces this activation was larger in the left than in the right hemisphere, whereas a bilateral or right dominant distribution was obtained for butterflies and spatial locations. These results suggest that the 402 H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 underlying neural systems activated in working memory can flexibly be used across a variety of informational domains. Involvement of modality-specific posterior areas during maintenance of information in working memory might be regarded as a solution for this flexibility. Accordingly, Druzgal and DÕEsposito (2001) have provided evidence for involvement of fusiform gyrus in face working memory in an n-back task for gray-scale faces. In addition, they found consistent linear load-dependent increases in right middle frontal gyrus, bilateral supplementary motor area, and right superior parietal lobule. Further support for involvement of an amodal or multimodal fronto-parietal network during working memory processes comes from Hautzel et al. (2002). Using fMRI in a two-back task they investigated commonalties and differences of verbal, spatial, real object, and shape working memory. Several regions, including DLPFC, VLPFC, superior parietal lobule, and inferior parietal lobule, were conjointly activated across all stimulus types. They observed no stimulus-specific differences in the activation patterns in the prefrontal cortex. However, extra-frontal regions specialized in feature processing and involved in the preprocessing of the stimuli were selectively activated by these different subtypes of working memory. Dade, Zatorre, Evans, and Jones-Gotman (2001) used PET to examine olfactory working memory in comparison with face working memory. They observed that both olfactory and face working memory engaged dorsolateral and ventrolateral frontal cortex, and inferior parietal areas (BA 40/7) when task requirements were matched. However, only visual working memory showed increased activity within left superior parietal cortex. These results are consistent with the idea of a modality-independent network for working memory, although modality-specific regions (specifically in posterior areas) may be of relevance as well. 2.2.2. Central executive and process-specific dissociations At the core of the working memory model is the central executive. According to Baddeley (2003), the central executive is the most important but least understood component of working memory. The notion of a central executive expands the concept of working memory beyond the borders of memory to a large body of additional mental functions. Baddeley (1986) has suggested that the supervisory attentional system component of the attentional control of action model proposed by Norman and Shallice (1986) might be an adequate approximation of the functions of the central executive system. Therefore, there is an obvious overlap between the concept of working memory and the concept of attention. Furthermore, the central executive is assumed to include several higher mental functions such as planning and decision-making. These functions generally reflect the conscious control of mental processes in contrast to involuntary reflexive cognitive processes. Therefore, there is also considerable overlap with the concept of consciousness. Considering these overlaps, one can expect to observe common features in neuroimaging studies of executive functions, attention, and consciousness. Random number generation has been considered as an example of a task that engages the central executive and could be used as an index of its supervisory capacity (Baddeley, 1986). Several studies have observed activities associated with random number generation within prefrontal and parietal cortices. Recently, Jahanshahi, Dirnberger, Fuller, and Frith (2000) using PET found that relative to a control counting task, random number generation was associated with significant activation of the left DLPFC, the anterior cingulate, the superior parietal cortex bilaterally, the right inferior frontal cortex, and the left and right cerebellar hemispheres. H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 403 An important dissociation of executive functions has been made between active maintenance of information in working memory on the one hand, and updating and manipulating of information within working memory on the other hand. Both active maintenance and updating or manipulating of information have been associated with prefrontal and parietal activations in neuroimaging studies. Manipulating and complex executive functions have usually been associated with activations in DLPFC. Cornette, Dupont, Salmon, and Orban (2001) used PET to identify the neural substrate of maintenance and manipulation of the orientation of a grating presented in central vision. Maintenance of orientations involved a distributed fronto-parietal network including bilateral lateral superior frontal sulcus, VLPFC, and precuneus, as well as right superior parietal lobule. A more medial superior frontal sulcus region was activated during manipulation of orientations in working memory. Moreover, based on functional connectivity analysis, the authors concluded that orientation working memory relies on a coordinated interaction between frontal and parietal regions. Tsukiura et al. (2001) studied processes of maintenance and manipulation in verbal working memory, comparing digit span in contrast with simple number detection. They observed process-specific areas associated with each process as well as overlapping areas. Activated areas specific to maintaining were detected in the right middle (BA 11/10) and medial (BA 6) frontal gyri, the right inferior parietal lobule (BA 40), and the left middle (BA 9) and inferior frontal gyri (BA 44), whereas specific activities in association with manipulating process were identified in the right middle (BA 9/46) and left precentral gyrus (BA 6). The overlapping areas common to the two processes were identified in the right inferior frontal gyrus (BA 47) and the left superior parietal lobule (BA 7). Veltman, Rombouts, and Dolan (2003) investigated neural correlates of maintenance and manipulation processes in an fMRI study. They used parametric versions of both a delayed match to sample task and an n-letter back task as prototypical tasks of maintenance and manipulation, respectively. Load-dependent activities common for both tasks were found in bilateral DLPFC, anterior prefrontal, left VLPFC, and bilateral parietal regions. Moreover, almost all areas that showed workload by task interactions were also activated in a conjunction analysis. The authors concluded that a functional rather than a neuroanatomical distinction exists between maintenance and manipulation. This means that maintenance and manipulation processes differentially activate virtually an identical network including fronto-parietal areas. Clark et al. (2000) studied neural activations associated with updating of verbal working memory with PET. Comparison of a variable target condition with a fixed target condition showed bilateral activation of dorsolateral prefrontal (middle frontal gyrus) and inferior parietal (supramarginal gyrus) cortices. They proposed that during updating process, supramarginal gyrus as an amodal region binds the various modal representations and middle frontal gyrus links posterior representations of the anticipated target stimulus to anterior representations of the planned response. Several other studies have previously found activations associated with updating of verbal working memory in dorsolateral prefrontal and inferior parietal cortex as well as other brain areas including superior parietal cortex, polar frontal region, and anterior cingulate cortex (Cabeza & Nyberg, 2000; Hartley & Speer, 2000). Taken together, working memory tasks tend to activate distributed neural networks including frontal (especially DLPFC) and parietal cortices (Fig. 1B). This is so regardless of whether the focus is on material-specific or process-specific processing in working memory. 404 H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 2.3. Episodic memory retrieval Episodic memory refers to a system for memory of personally experienced events. It is thought to involve an interaction between a Ôretrieval cueÕ (self-generated or provided by the environment) and a memory trace, leading to the reconstruction of some or all aspects of the episode represented by the trace (Rugg & Wilding, 2000; Tulving, 1983). The underlying neural bases of episodic memory retrieval have been extensively investigated with fMRI and PET. A selection of these studies is presented in Table 3. Notably, the presented studies show that across a variety of designs and experimental conditions, involvement of frontal and parietal regions stands out as a consistent neural signature of episodic memory retrieval. This observation is in keeping with several other reviews (for example, see Cabeza & Nyberg, 2000; Mayes & Montaldi, 2001; Rugg & Wilding, 2000). The specific role of fronto-parietal regions in episodic retrieval is not fully understood yet; however, the involvement of several regions within frontal and parietal lobes has been frequently reported in various aspects of episodic retrieval, including retrieval mode, retrieval success, and retrieval effort. Retrieval mode has been proposed as an appropriate cognitive state that is maintained throughout the retrieval period, by which the stimulus event would be treated as an episodic retrieval cue (Tulving, 1983); retrieval effort refers processing resources that are evoked in more demanding retrieval situations; and retrieval success is a term used to denote processes that are associated with, or depend upon, actual recovery of information from memory, which is called ecphory. In a series of early neuroimaging studies, Tulving and colleagues proposed that activation of the right prefrontal cortex (BA 10) is a key neural signature of retrieval mode (Kapur et al., 1995; Nyberg et al., 1995). This proposal is based on findings that reveal invariance of right prefrontal activation, regardless of whether the cue corresponds to a studied or a non-studied item or whether level of retrieval performance is low or high. More recently, in a multistudy analysis of PET experiments, Lepage, Ghaffar, Nyberg, and Tulving (2000) identified regions in bilateral frontal pole (BA 10), bilateral frontal operculum (BA 47/45), and the right DLPFC (BA 8/9), where neuronal activity seems to be correlated with the maintenance of episodic memory retrieval mode. The activations were considerably larger in the right hemisphere. Asymmetrical prefrontal activity in favor of the right hemisphere also has been reported in some other imaging studies of episodic retrieval. For example, Ragland et al. (2000) using PET observed activity in the right anterior (areas 9 and 10) and the right inferior (area 44/45) prefrontal cortex during word recognition tasks; and Bernard, Desgranges, Platel, Baron, and Eustache (2001) in another PET study found activations in the right DLPFC (BA 8/9) associated with stemcued recall task in contrast to a stem completion task. However, several recent studies have found bilateral prefrontal activity during episodic retrieval tasks (Cabeza & Nyberg, 2000; Mayes & Montaldi, 2001). For example, Schmidt et al. (2002) found bilateral activity in the dorsolateral prefrontal and the frontopolar cortices during episodic retrieval across visual and auditory presentation of verbal memory items with both low and high imagery content. It has been proposed that involvement of the left prefrontal cortex in addition to the right side activity may reflect the effect of complexity of retrieval tasks. In a review of the neuroimaging literature, Nolde, Johnson, and DÕEsposito (1998) concluded that relatively simple memory retrieval tasks elicited right prefrontal activation, but more demanding retrieval tasks elicited left prefrontal activation as well (cf., Lundstrom et al., 2003). Table 3 Episodic memory H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 405 (continued on next page) 406 Symbols and abbreviations: (s) left lateral; (d) right lateral; ( ) bilateral lateral; (h) left medial; (j) right medial; ( ) bilateral midline; L, left; R, right; ant., anterior; post., posterior; vent., ventral; SMA, supplementary motor area. H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 Tabel 3 (continued) H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 407 Retrieval tasks that require a higher degree of effort have also been specifically associated with activation of DLPFC. Schmidt et al. (2002), in the above mentioned study, observed differential activity in DLPFC during retrieval of word-pairs with low imagery content. This DLPFC activation in a more effortful task is in line with observations by Nyberg, McIntosh, and Tulving (1998), who proposed that the DLPFC is involved in more difficult retrieval tasks with increased retrieval effort. Successful episodic memory retrieval seems to be critically dependent on contributions of regions in parietal cortex; however, several studies reveal that prefrontal activity may be of relevance in retrieval success as well. Employing rapid event-related fMRI, Konishi, Wheeler, Donaldson, and Buckner (2000) compared hit trials (correctly recognized old items) and correct rejection trials (correctly rejected new items) in an old/new episodic recognition task. The comparison revealed a set of brain regions in lateral and medial parietal cortices as well as prefrontal cortex which was mostly left-lateralized, suggesting that left parietal and frontal regions modulate activity based on the successful retrieval of information from episodic memory. Fronto-parietal activity has also been found correlated with retrieval success in recognition of recently learned faces. Leube, Erb, Grodd, Bartels, and Kircher (2003), in an event-related fMRI study, compared brain activation during correct recognition of the recently learned faces to that observed during correct rejection of unknown faces. They found differences in the left inferior parietal and left medial frontal cortices, suggesting that these two regions are involved in identifying familiarity of faces, while mere processing of personal identity may be associated with more restricted activity in the fusiform gyrus. Correspondingly, Leveroni et al. (2000) used event-related fMRI to compare newly learned, unfamiliar faces with new distractor faces. They observed that retrieval of newly learned faces resulted in a positive signal change in the left inferior parietal, left precuneus, and left medial frontal regions relative to the distractor face condition. Although some researchers have proposed that activation of medial parietal cortex may reflect processing of the imagery content of retrieved information, it seems that successful retrieval of information regardless of imageable characteristics activates parietal cortex (e.g., Krause et al., 1999). Accordingly, in a PET study of episodic memory for music material, Platel, Baron, Desgranges, Bernard, and Eustache (2003) observed that recognition of familiar melodies was associated with activations in precuneus and superior frontal regions, predominantly in the right hemisphere. A further finding of this study was that recognition of newly learned non-familiar melodies produced bilateral activations of frontal areas. This bilateral engagement of the frontal regions may reflect difficulty experienced by the subjects during retrieval of non-familiar material and the occurrence of executive control processes (in keeping with Nolde et al., 1998, as noted above). Herron, Henson, and Rugg (2004), in a recent event-related fMRI study, suggested that successful recognition is only associated with activation of a subset of the regions identified in previous studies, including lateral inferior and medial parietal cortex. According to the authors, in other regions, most notably DLPFC and VLPFC, differences in the activity elicited by old and new items may reflect processes that are contingent upon, rather than in support of, successful retrieval. These contingent processes are postulated to contribute to functions such as target detection, adjustment of expectancies, and overriding prepotent response tendencies. Retrieval tasks that involve elaborated executive processes, such as those required during recollection of highly specific or contextual information of memorized objects, may involve specific 408 Table 4 Consciousness H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 Symbols and abbreviations: (s) left lateral; (d) right lateral; ( ) bilateral lateral; (h) left medial; (j) right medial; ( ) bilateral midline; L, left; R, right; ant., anterior; post., posterior; vent., ventral; SMA, supplementary motor area. H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 409 regions in fronto-parietal areas. Using event-related fMRI, Ranganath, Johnson, and DÕEsposito (2000) found that activity in a region adjacent to left anterior prefrontal cortex (BA 10/46) was modulated by the specificity of information to be retrieved. Activation in this region increased with demands to retrieve perceptually detailed information about studied objects. The authors concluded that the left anterior prefrontal cortex is engaged during the monitoring and evaluation of specific memory characteristics at retrieval. Henson, Shallice, and Dolan (1999) found that demand for recollection of the spatiotemporal contextual information of words presented during a previous study episode activated bilateral dorsolateral, ventrolateral, and anterior prefrontal cortex, as well as superior parietal cortex during recognition; whereas recency judgment without reference to the spatiotemporal context of previously presented words was associated with smaller activations predominantly in right prefrontal cortex. Correspondingly, Dobbins, Rice, Wagner, and Schacter (2003) in an event-related fMRI study found distinct lateral prefrontal and parietal activations that distinguished attempted source recollection from judgments of relative recency. According to Dobbins et al., prefrontal and parietal regions are critical in mediating the manner in which test items are employed as cues for memory. That is, these regions appear to be important for the flexible adoption of decision and monitoring processes that qualitatively differ depending on situational factors such as whether or not item-specific recollection is essential for a given task. Finally, Suzuki et al. (2002) observed different patterns of prefrontal activity during two kinds of temporal context memory tasks: the temporal order of items between lists and within a list. Right prefrontal activity was associated with temporal order judgment between lists, whereas left prefrontal activity was related to temporal order judgment within a list. Taken together, fMRI and PET studies of episodic memory retrieval reveal critical involvement of fronto-parietal regions in episodic retrieval tasks (Fig. 1C). The most consistent finding has been activation of the frontal pole across a wide variety of episodic retrieval conditions, possibly reflecting a role of this region in retrieval mode. Activation of parietal cortex, especially precuneus, exhibits the strongest correlation with successful retrieval of remembered episodes. Finally, several regions in prefrontal and parietal cortices are associated with executive control processes during recollection of information to be retrieved. 2.4. Conscious perception A conscious perception is a complex phenomenon which evolves through several sequential steps; therefore it is conceivably associated with widespread patterns of brain activity. Block (1997) argues for a conceptual distinction between ÔphenomenalÕ and ÔaccessÕ consciousness. While phenomenal consciousness refers to the subjective aspect of experience, access consciousness refers to the direct control of experience through reasoning, reporting, or action.Visual awareness is the most studied type of conscious perception in normal and abnormal humans as well as animals. Some neuroimaging and neurophysiological studies of visual awareness (Logothetis, 1998; Tong et al., 1998) have concentrated on the role of the ventral stream and have not typically considered the potential role of fronto-parietal activations. However, more recent studies suggest that joint activation of category-specific regions in the ventral stream and activity in parietal and prefrontal areas might be crucial for visual awareness (Rees et al., 2002). Our selected fMRI and PET studies of conscious perception, summarized in Table 4, confirm this view. 410 H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 One way to investigate the neural areas associated with visual awareness is to generate bistable percepts such as binocular rivalry or ambiguous figures to dissociate subjective perception from sensory input. Binocular rivalry provides a useful experimental paradigm to study the neural correlates of conscious perception. When dissimilar images are presented to the two eyes, they compete for perceptual dominance so that each image is visible for a few seconds while the other is suppressed. As the alternation occurs in the absence of any changes in the stimulus itself, variation in brain activity can be directly related to conscious perception. Using a binocular rivalry paradigm, Lumer et al. (1998) measured brain activity through fMRI in humans who reported their percepts under two different viewing conditions. In the first condition, subjects viewed dichoptic stimuli consisting of a red-colored drifting grating shown to one eye and a green-colored face shown to the other eye. In the second condition, subjects were exposed to a ‘‘replay’’ of their perception during rivalry. This was achieved by presenting, in a chronology specified by the key reports during rivalry, either the face alone or the grating alone to one eye, and a gray patch of comparable luminance to the other eye. This stimulation was designed to produce a perception that closely mimics rivalry in both quality and timing. During rivalry, compared to the other condition, transient responses associated with shifts of perception were found not only in extrastriate areas, but also in right inferior and superior parietal lobules, bilateral inferior frontal, middle frontal, and insular cortices. These results indicated a specific role for fronto-parietal areas in mediating the perceptual transitions experienced during rivalry. To control the motor effect of subjectsÕ reports, in a second study, Lumer and Rees (1999) examined the neural responses evoked during binocular rivalry in the absence of any motor reports. Consistent with the earlier study, they found that activity in early extrastriate cortex covaried with responses in areas located both at higher levels in the dorsal and ventral visual pathways and in lateral frontal cortex, indicating systematic interactions among these areas. Considering the results of both studies, Lumer and Rees concluded that non-visual cortical areas including prefrontal cortex, acting in concert with functionally specialized visual centers, are associated specifically with conscious visual perception. In ‘‘binocular fusion’’ the two monocular inputs result in the perception of a single fused image instead of conflicting with each other, as it is the case in binocular rivalry. To examine visual consciousness, Moutoussis and Zeki (2002) exploited a form of binocular fusion in which one eye was exposed to a green image on a red background and the other eye was exposed to the same image with a red color on a green background. Fusion of the two images made them invisible at binocular level and the subjects experienced just a uniform yellow field. Moutoussis and Zeki observed that the ‘‘invisible stimuli’’ activated the same stimulus-specific areas in visual cortex that were activated when both eyes had the same stimulation and thereby the stimuli were perceived. Nevertheless, the level of activation was higher with visible stimulation than with invisible stimulation. The researchers concluded that the conscious perception of a stimulus activates the same higher visual areas which are involved in its unconscious processing; however, conscious perception is associated with a higher level of activation in these areas than unconscious processing and also probably with the activation of some other areas. (We did not include the study in Table 4, because although it is a whole-brain study, the authors just mentioned the Talairach coordinates for specific areas of the visual cortex.) Ambiguous figures (such as the Necker cube and Rubins face/vase) produce another type of bistable percepts that has been used for studying visual consciousness. Using fMRI, Kleinschmidt et al. (1998) investigated neural correlates of flips in visual perception by asking human observers H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 411 to look at ambiguous figures. Contrasting perceptual reversals with perceptual stability, they found responses in prestriate visual cortical areas as well as in posterior intraparietal cortex and some frontal areas. Frontal activations were located in ventral prefrontal areas and the frontal eye fields bilaterally. A specific type of ambiguous perceptual conditions that has been studied by researchers is the ‘‘spinning-wheel illusion,’’ a bistable apparent motion stimulus of which both possible percepts correspond to the same object, share the same center, and are perceived as identically patterned stimuli moving at the same speed and changing only in direction. In an fMRI study, Sterzer et al. (2002) analyzed the spatial distribution of event-related activations occurring during spontaneous reversals of perceived direction of motion through the spinning-wheel illusion. In accordance with earlier neuroimaging findings for bistable percepts, Sterzer et al. observed event-related activations in several frontal and parietal areas, including the superior parietal cortex bilaterally, the right inferior parietal cortex, and the premotor and inferior frontal cortex of both hemispheres. These results suggest that, although the activations in functionally specialized extrastriate visual cortex are highly category- or attribute-specific, a fronto-parietal network subserves more general aspects in bistable visual perception. Comparing supraliminal and subliminal perception, for instance by applying masking paradigms, also has been used for investigating neural correlates of visual consciousness. Using fMRI, Bar et al. (2001) examined object recognition with subjects required to recognize masked objects that were presented very briefly. They found that increased awareness of object identity was correlated with increasing activity in ventrotemporal visual regions as well as an anterior shift of activation foci. Furthermore, Bar et al. observed an activity in the inferior frontal gyri which was highest for the recognized masked objects. Dehaene et al. (2001) used fMRI and ERPs to study the cerebral processing of unseen masked words, and found that activation in the ventral visual stream, including fusiform gyrus, can occur without conscious reportability. Interestingly, a remarkable difference between masked and unmasked words was the presence of increased activity at distant parietal, prefrontal, and cingulate sites, only when the words were visible. That is, unmasking the words enabled the propagation of activation of a large-scale correlated cerebral assembly. The authors concluded that their findings are consistent with theories that relate conscious perception to the top-down amplification of sensory information through synchronous coactivation of distant regions including prefrontal cortex (Dehaene et al., 2001). Similarly, using PET, Kjaer, Nowak, Kjaer, Lou, and Lou (2001) compared the neural mechanisms of supraliminal and subliminal visual verbal perception in order to identify the regions differentially supporting awareness of stimuli. The comparison revealed that medial parietal and anterior dorsolateral prefrontal regions in the right hemisphere were selectively activated by aware versus unaware stimulation. Kjaer et al. concluded that these higher order perceptual and executive cortical structures are critical for visual verbal awareness. Two recent studies used a paradigm known as the attentional blink to compare neural activity associated with conscious and unconscious processing of visual stimuli. In this paradigm, while subjects search for two targets presented in a rapid serial visual display of distractor items, they show severe impairment at detecting the second of the two targets when it is presented from 200 to 500 ms after the first target. However, the subjects easily detect the second target when it is presented outside the attentional blink time window. Using fMRI and attentional blink paradigm, Marois, Yi, and Chun (2004) found that although the second target frequently went undetected 412 H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 by the subjects, it nonetheless activated regions of the medial temporal cortex involved in highlevel visual representations. On the other hand, the lateral frontal cortex and anterior cingulate were activated only when the second target was successfully reported, indicating that the frontal activity, in contrast with the temporal activity, is mainly correlated with the subjectsÕ explicit perception of the stimulus rather than its physical presentation. (We did not include this study in Table 4, because significant results were only observed when the authors used an ROI approach. See Section 1.) The other study, which again used fMRI and a version of attentional blink paradigm to study conscious perception, was conducted by Feinstein, Stein, Castillo, and Paulus (2004). They could behaviorally distinguish a subgroup of subjects who successfully identified the second target, the ‘‘Non-Blinkers,’’ from another subgroup who failed to identify the second target at the same condition, the ‘‘Blinkers.’’ Conscious perception of the second target in Non-Blinkers was differentially associated with activation of the anterior cingulate, medial frontal cortex, and frontopolar cortex. Feinstein et al. concluded that the Non-Blinkers could utilize the top-down attentional amplification of the prefrontal regions for transforming the fleeting sensory perception into a conscious awareness. Change detection is another paradigm for studying visual conscious perception. Comparing trials in which subjects consciously detect a change to trials in which they are blind to it can reveal activity due to processes involved in awareness of the change. Using event-related fMRI, Beck, Rees, Frith, and Lavie (2001) examined which neural systems are active when subjects consciously detect a visual change versus when they are functionally blind to an equivalent change. They observed that while some ventral stream activity was also found in situations of change blindness, conscious detection of change evoked activity in both the dorsal and ventral streams. Interestingly, activity in the parietal lobes and DLPFC was exclusively associated with the conscious detection of change. Based on their results, Beck et al. emphasize the importance of dorsal activations for awareness, and in contrast to the traditional account of the ventral and dorsal streamsÕ roles in awareness, by which the ventral pathway is described as ÔconsciousÕ and the dorsal pathway as Ôunconscious,Õ their results suggest that awareness depends on joint activation of dorsal fronto-parietal structures and category-specific regions of the ventral stream. Neuroimaging studies of conscious perception have typically focused on transitions between experiences of different types of phenomenal content, and have occasionally considered sustained perception as a control condition (e.g., Kleinschmidt et al., 1998). Nevertheless, a few studies have specifically investigated putative neuroanatomical dissociation of perceptual transitions from sustained perception. It is conceivable that a sustained percept is implemented simply by enduring activity in the areas responsible for the initial perceptual synthesis. Alternatively, the maintenance of a percept could be mediated by brain areas distinct from those involved in perceptual transitions. In an fMRI study by Portas, Strange, Friston, Dolan, and Frith (2000), this distinction was investigated by using random dot stereograms as stimuli in an object identification task. Within individual trials, brain activity related to identifying the perceptual content of the stereograms was differentiated from sustaining the percepts in mind. The results associated perceptual identification with frontal, parietal, and occipitotemporal regions, whereas sustained perception engaged a distinct dorsolateral prefrontal region as well as the hippocampus. In an fMRI study, Eriksson, Larsson, Åhlström, and Nyberg (2004) have also investigated the possible dichotomy between the neurophysiological bases of perceptual transitions versus sustaining a particular percept over time. Perceptual transitions were indicated by subjects whenever they H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 413 could find the hidden target stimuli while looking at fragmented crowded pictures, and then they were to maintain their percept for a period. Eriksson et al. found common activations in occipitotemporal and fronto-parietal regions for both perceptual transitions and sustained perception. However, regions in the medial temporal lobe (MTL) were activated specifically for transitions, whereas medial and dorsolateral prefrontal regions were activated specifically for sustained perception. These results provided further support for the theory that the initial creation of perceptual awareness and upholding perceptual awareness over time are separate processes involving different brain regions. Eriksson et al. concluded that while MTL activity may be more strongly associated with the initial stages of perceptual awareness; the ventral visual and fronto-parietal regions are associated with both the initial and subsequent stages of perceptual awareness. Finally, we briefly mention a field that is related to conscious perception, although not included in Table 4, namely awareness during learning. A few neuroimaging studies have recently investigated awareness during learning. It has been assumed that learning with awareness (for instance, being aware that a specific tone predicts a visual event) requires the involvement of MTL. However, recent studies suggest that the involvement of MTL is best appreciated by functional connectivity of MTL with other brain areas including dorsolateral prefrontal and parietal regions (McIntosh, Rajah, & Lobaugh, 1999, 2003). Comparing activities associated with sequence learning with and without awareness, Willingham, Salidis, and Gabrieli (2002) found widespread additional activation in fronto-parietal areas when learning was accompanied with awareness. Lastly, studying motor adjustments to changing auditory rhythms, Stephan et al. (2002) observed that while ventral prefrontal areas may be engaged in motor adaptations performed subconsciously, only fully conscious motor control involves DLPFC. Taken together, neuroimaging studies suggest that fronto-parietal activity makes an important contribution to conscious perception (Fig. 1D). Although activity in ventral visual cortex is a consistent neural correlate of visual perception, it might be insufficient to produce awareness without an additional contribution from parietal and prefrontal areas (Crick & Koch, 1995; Rees et al., 2002). Pins and Ffytche (2003) have reached a similar conclusion. They suggested that correlates of consciousness are divided into primary and secondary network nodes, where early activity in the occipital lobe correlates with perceptual processes and later activity in fronto-parietal areas correlates with secondary processes contingent on the outcome of earlier perceptual processing. BlockÕs distinction between phenomenal and access consciousness (Block, 1997) would be of relevance for understanding the putative role of fronto-parietal areas in conscious perception. While phenomenal consciousness might be mainly associated with activation of sensory regions, access consciousness may need involvement of fronto-parietal areas. The specific pattern of fronto-parietal activations during conscious perception might further be differentiated depending on relevant stages and/or types of processing, for instance perceptual transitions versus sustained perception. 3. Discussion The ancient tale ÔThe Elephant in the DarkÕ is about an unknown beast that appeared in a town with only blind people. Three men, who were sent by the King to examine the beast, approached the elephant from different sides. One of them, who touched one ear, said, ‘‘It is large, flat and rough, like a rug.’’ The second, touching the trunk, said, ‘‘It is like a trumpet, but capable of 414 H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 dramatic movement.’’ The third, who touched the legs, shouted, ‘‘No, no, it is mighty and firm like a pillar.’’ The tale illustrates that one and the same object can be seen quite differently depending on oneÕs perspective, and a true interpretation might require integration. Similarly, understanding the contribution of a network in the brain might require integration of results from several different functional paradigms. In particular this is so since the same regions have been activated across various tasks with different cognitive demands (Cabeza & Nyberg, 2000; Culham & Kanwisher, 2001; Duncan & Owen, 2000). A major goal of the present review was to examine similarities in brain activation patterns associated with attention, working memory, episodic memory retrieval, and visual awareness, with the hope that this might constrain the interpretation of the functional significance of involved regions. For all reviewed functions we observed a consistent and overlapping pattern of brain activity (Figs. 2 and 3). The overlap was most pronounced for the bilateral parietal cortex (BA 7 and BA 40; close to the intraparietal sulcus), and DLPFC (right BA 9 and left BA 6). It should be noted that BAs, which are used in this review for comparison of studies, are not characterized based on functional properties, but rather they are defined based on their cytoarchitectonic characteristics. Furthermore, the mere fact that a particular BA is active across multiple studies does not necessarily mean that the same exact region of the brain is active in all the studies. More precisely, while the assessment of overlap has been done between studies in this review, it could be argued that the apparent overlap in activity across functions actually reflects activations of adjacent but distinct regions. However, several within-study fMRI studies have demonstrated overlap in fronto-parietal activation patterns for attention and working memory (LaBar et al., 1999), attention and episodic memory retrieval (Cabeza et al., 2003), and working memory and episodic memory retrieval (Braver et al., 2001; Cabeza et al., 2002; Marklund et al., in press; Nyberg et al., 2002; Ranganath et al., 2003). What is the functional significance of overlapping fronto-parietal activations? Possibly, these activations reflect shared cognitive processes. Several studies converge on attentional and/or working memory processes as the explanation for overlapping fronto-parietal activity patterns. Wagner (1999) has proposed that prefrontal activation during episodic retrieval may be seen as reflecting specific working memory contributions to episodic memory. He argued that material-independent working memory functions contributing to episodic retrieval are mediated by dorsolateral and anterior prefrontal regions. Awh and Jonides (2001) argued that frontal and parietal Fig. 2. The level of common regional brain activity across functions, described as minimum relative frequency of regional brain activity for each Brodmann area. Right and left BA 7, left BA 40, and right BA 9 show the highest level of common regional brain activity across functions. H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 415 Fig. 3. Peaks of activation in fronto-parietal areas associated with: (A) attention, (B) working memory, (C) episodic retrieval, and (D) conscious perception, extracted from the studies listed in Tables 1–4, respectively. Blank areas indicate the regions with highest level of common regional brain activity across functions, namely bilateral BA 7 and BA 40, left BA 6, and right BA 9. Only the peaks in (or adjacent to) these regions are shown. 416 H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 mechanisms involved in spatial working memory represent an attentional circuit that operates in the service of memory. They postulated that spatial attention mechanisms are recruited in the service of a rehearsal-like function to maintain information in working memory. Curtis and DÕEsposito (2003) suggested that DLPFC aids in the maintenance of information by directing attention to internal representations of sensory stimuli and motor plans that are stored in more posterior regions. Cabeza et al. (2003) suggested that common activity in a fronto-parietal-cingulate-thalamic network during episodic retrieval and visual attention reflects general attentional processes. Rees and Lavie (2001) proposed that distributed interactions between modality-specific posterior regions and fronto-parietal areas subserve both visual attention and visual awareness. Several studies have revealed that attention is essential for conscious visual perception. For example, Rees, Russell, Frith, and Driver (1999) in an fMRI study observed that visual recognition completely depends on attention even for highly familiar and meaningful materials. Furthermore, it has been shown that attending to an object amplifies and sharpens the neural representations of the object (Reynolds, Pasternak, & Desimone, 2000), leading to an improved ability to detect the object and report its properties (Carrasco, Penpeci-Talgar, & Eckstein, 2000; see also Carrasco, Ling, & Read, 2004). Attention and working memory are also salient in more general theoretical accounts. One of these has been proposed by Baars (1988, 2002) who coined the concept of Ôglobal workspaceÕ as a mental capacity which represents the dominant information that is widely distributed in the brain at each moment. Baars holds that this makes sense in a nervous system viewed as a massive distributed set of specialized networks, where coordination, control, and problem solving could take place by way of a central information exchange, allowing some regions—such as sensory cortex—to distribute information to the whole. By this view, conscious perception is not restricted to sensory analysis; rather, it enables access to widespread brain sources, whereas unconscious input processing is limited to sensory regions. On the other hand, working memory depends on conscious elements, including conscious perception, inner speech, and visual imagery, each mobilizing widespread functions. Similarly, selective attention enables access to conscious contents, and vice versa. Therefore, several cognitive functions that all correspond to conscious accessibility and conscious control may interact through a unitary functional structure; that is, global workspace. While in some theoretical formulations, the thalamocortical system has been proposed as the neural substrate of global workspace (Baars, 1988; Newman & Baars, 1993), others have emphasized the role of prefrontal cortex (Dehaene & Naccache, 2001). In line with BaarsÕ theory, Dehaene and Naccache postulate that, besides specialized processors, the brain comprises a distributed neural system with long-distance connectivity that can potentially interconnect multiple specialized brain areas in a coordinated manner. Through this workspace, that according to Dehaene and Naccache is mainly mediated by prefrontal cortex, modular systems that do not directly exchange information in an automatic mode can nevertheless gain access to each otherÕs content. The global workspace, by this view, is thought to provide a common communication protocol where multiple input, output, and internal systems can be integrated. Frith and Dolan (1996) have proposed that higher cognitive functions such as working memory, mental imagery, and willed action are all intimately associated with consciousness. They claim that the common process underlying all these functions is that information is ‘‘held in mind’’ for a period of time. This information, which may be about stimuli or responses, can be H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 417 derived from the past or generated for the future. According to Frith and Dolan, brain imaging studies show that ‘‘holding something in mind’’ is associated with activity in an extended system which involves both prefrontal cortex and more posterior areas whose location is determined by the nature of the information being held in mind. Automatic actions and perceptions which do not involve consciousness are associated with activity in the relevant posterior areas, but not in the prefrontal cortex. Baddeley (2000, 2003) has recently proposed the Ôepisodic buffer,Õ as a fourth component of his influential working memory model. The episodic buffer is assumed to be a limited capacity store that binds together information to form integrated episodes. It is assumed to be attentionally controlled by the central executive and accessible to conscious awareness. Its multidimensional coding is thought to allow different systems to be integrated, while conscious awareness provides a convenient binding and retrieval process. Baddeley regards the episodic buffer as a crucial feature of the capacity of working memory to act as a global workspace that is accessed by conscious awareness along the lines suggested by Baars (1988, 2002). This proposal also emphasizes the capacity of working memory to manipulate and create new representations while interacting with longterm memory, rather than simply activating old memories. In an fMRI study, Gruber and von Cramon (2003) have recently identified a material-independent network including DLPFC and the intraparietal sulcus as a putative substrate of the episodic buffer (see also Prabhakaran, Narayanan, Zhao, & Gabrieli, 2000). The latter account offers increased specificity by postulating Ôintegrating within the episodic bufferÕ as a potential explanation for overlapping fronto-parietal activity patterns. Along the theoretical framework proposed by Moscovitch and others (Moscovitch, 1992; Parkin, 1999; Roediger, Buckner, & McDermott, 1999), integration might be conceptualized as a shared cognitive component process that operates in a number of different cognitive domains to bind low-level representations into more elaborated ones. A critical role for integration in conscious perception has been suggested (Edelman & Tononi, 2000). In such tasks multiple perceptual features must be integrated into a unified percept. Integration is also critical in actual recollection of episodic information (e.g., Dobbins, Foley, Schacter, & Wagner, 2002). Episodic representations are believed to be distributed and have to be integrated into a single phenomenological experience at the time of retrieval. Tests of working memory often require that multiple on-line representations are integrated (Irwin & Andrews, 1996; Luck & Vogel, 1997; Wheeler & Treisman, 2002), in particular if different types of information must be integrated (Prabhakaran et al., 2000). Finally, tests of attention often require detection of some kind of change and the integration of such an event with current task goals and response demands. Many tests of sustained attention, when participants are monitoring a stimulus to detect some kind of change, resemble tests of visual awareness in that a searched-for event is suddenly detected and has to be acted upon. Common to all forms of integration is likely that inter-connections between the fronto-parietal network and other parts of the brain are fundamental. Such interactions might be possible to examine through analysis of interregional connectivity patterns (e.g., Friston, 2002; Funahashi, 2001; McIntosh, 2000; Nyberg & McIntosh, 2000). In such an analysis, McIntosh et al. (1999) found that interactions involving left DLPFC and multiple other regions defined awareness in sensory learning. Integration can be thought of as a general concept encompassing several conceptualizations in cognitive psychology and neuroscience, such as grouping, chunking, and binding (Revonsuo, 1999; Treisman, 1996); all of them have been used for referring to reorganization of distributed 418 H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 bits of information and related subjective experiences into more complex but unified structures. Gestalt psychologists dealt with this issue by describing sensory organization in terms of perceptual units. More recently, integration processes have also been reflected in several theories of mind/brain, including TreismanÕs feature integration theory of attention (Treisman, 1988, 1996), DamasioÕs theory of neural convergence zones (Damasio, 1989), and the synchronized oscillation theory of consciousness introduced by Crick and Koch (1990, 1998). Synchronized oscillations have been found to be a neural correlate of integration processes during conscious perception (Engel & Singer, 2001), attention (Niebur, 2002), working memory (Sarnthein, Petsche, Rappelsberger, Shaw, & von Stein, 1998), and recollection of episodic information (Klimesch et al., 2001). Low levels of integration such as binding of simple visual attributes might be mediated by sensory areas of the brain; however, we propose that common fronto-parietal activations associated with conscious perception, attention, working memory and episodic retrieval may reflect a high level of integration among multimodal distributed representations which is the prerequisite for a wide variety of cognitive demands. (In keeping with this hypothesis, findings of a recent fMRI study by Ehrsson, Spence, & Passingham, 2004, confirm a role for parietal and premotor cortices in multisensory integration.) It should be noted that functional neuroimaging techniques can at best specify the coincidence of regional brain activations with specific cognitive demands. These methods cannot determine which brain regions are essential for a specific cognitive process. Thus, to examine the putative role of fronto-parietal network in integration, the results of PET and fMRI studies should be complemented with further studies by electromagnetic and neuropsychological methods. 4. Conclusion Attention, working memory, episodic memory retrieval, and visual awareness engage overlapping patterns of activity in dorsolateral prefrontal and parietal cortex. Multiple sources of evidence link these activation patterns to integration of distributed representations in multiple brain areas. The proposed central role of integration is consistent with general theories of the neural bases of consciousness which emphasize concepts such as global workspace. More generally the results of our review suggest a strong link between consciousness and cognition. Acknowledgments We thank Antti Revonsuo and Johan Eriksson for reading and commenting on the manuscript. We also thank three anonymous reviewers for constructive advice. References Awh, E., & Jonides, J. (2001). Overlapping mechanisms of attention and spatial working memory. Trends in Cognitive Sciences, 5(3), 119–126. Baars, B. J. (1988). A cognitive theory of consciousness. New York: Cambridge University Press. H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 419 Baars, B. J. (2002). The conscious access hypothesis: Origins and recent evidence. Trends in Cognitive Sciences, 6(1), 47–52. Baddeley, A. D. (1986). Working memory. Oxford: Oxford University Press. Baddeley, A. D. (1992). Working memory. Science, 255, 556–559. Baddeley, A. D. (2000). The episodic buffer: A new component of working memory? Trends in Cognitive Sciences, 4(11), 417–423. Baddeley, A. D. (2003). Working memory: Looking back and looking forward. Nature Reviews Neuroscience, 4, 829–839. Baddeley, A. D., & Hitch, G. J. (1974). Working memory. In G. Bower (Ed.). Recent Advances in learning and motivation (Vol. 8). New York: Academic Press. Bar, M., Tootell, R. B. H., Schacter, D. L., Greve, D. N., Fischl, B., Mendola, J. D., et al. (2001). Cortical mechanisms specific to explicit visual object recognition. Neuron, 29, 529–535. Beauchamp, M. S., Petit, L., Ellmore, T. M., Ingeholm, J., & Haxby, J. V. (2001). A parametric fMRI study of overt and covert shifts of visuospatial attention. NeuroImage, 14, 310–321. Beck, D. M., Rees, G., Frith, C. D., & Lavie, N. (2001). Neural correlates of change detection and change blindness. Nature Neuroscience, 4(6), 645–650. Bernard, F., Desgranges, B., Platel, H., Baron, J. C., & Eustache, F. (2001). Contributions of frontal and medial temporal regions to verbal episodic memory: A PET study. NeuroReport, 12, 1737–1741. Block, N. (1997). On a confusion about a function of consciousness. In N. Block, O. Flanagan, & G. Guzeldere (Eds.), The nature of consciousness: Philosophical debates. Cambridge: MIT Press. Braver, T. S., Barch, D. M., Kelley, W. M., Buckner, R. L., Cohen, N. J., Miezin, F. M., et al. (2001). Direct comparison of prefrontal cortex regions engaged by working and long-term memory tasks. NeuroImage, 14, 48–59. Burton, H., Abend, N. S., MacLeod, A.-M. K., Sinclair, R. J., Snyder, A. Z., & Raichle, M. E. (1999). Tactile attention tasks enhance activation in somatosensory regions of parietal cortex: A positron emission tomography study. Cerebral Cortex, 9, 662–674. Cabeza, R., Dolcos, F., Prince, S. E., Rice, H. J., Weissman, D. H., & Nyberg, L. (2003). Attention-related activity during episodic memory retrieval: A cross-function fMRI study. Neuropsychologia, 41(3), 390–399. Cabeza, R., & Nyberg, L. (2000). Imaging cognition II: An empirical review of 275 PET and fMRI studies. Journal of Cognitive Neuroscience, 12, 1–47. Cabeza, R., & Nyberg, L. (2002). Seeing the forest through the trees: The cross-function approach to imaging cognition. In A. Zani & A. M. Proverbio (Eds.), The cognitive electrophysiology of mind and brain. San Diego: Academic Press. Cabeza, R., Dolcos, F., Graham, R., & Nyberg, L. (2002). Similarities and differences in the neural correlates of episodic memory retrieval and working memory. NeuroImage, 16, 317–330. Carrasco, M., Ling, S., & Read, S. (2004). Attention alters appearance. Nature Neuroscience, 7(3), 308–313. Carrasco, M., Penpeci-Talgar, C., & Eckstein, M. (2000). Spatial covert attention increases contrast sensitivity along the CSF: Support for signal enhancement. Vision Research, 40, 1203–1215. Clark, C. R., Egan, G. F., McFarlane, A. C., Morris, P., Weber, D., Sonkkilla, C., et al. (2000). Updating working memory for words: A PET activation study. Human Brain Mapping, 9, 42–54. Collette, F., & Van der Linden, M. (2002). Brain imaging of the central executive component of working memory. Neuroscience and Biobehavioral Reviews, 26, 105–125. Corbetta, M., & Shulman, G. L. (2002). Control of goal-directed and stimulus-driven attention in the brain. Nature Reviews Neuroscience, 3, 201–215. Corbetta, M., Kincade, J. M., & Shulman, G. L. (2002). Neural systems for visual orienting and their relationships to spatial working memory. Journal of Cognitive Neuroscience, 14(3), 508–523. Corbetta, M., Kincade, J. M., Ollinger, J. M., McAvoy, M. P., & Shulman, G. L. (2000). Voluntary orienting is dissociated from target detection in human posterior parietal cortex. Nature Neuroscience, 3, 292–297. Cornette, L., Dupont, P., Salmon, E., & Orban, G. A. (2001). The neural substrate of orientation working memory. Journal of Cognitive Neuroscience, 13(6), 813–828. Crick, F., & Koch, C. (1990). Towards a neurobiological theory of consciousness. Seminars in the Neurosciences, 2, 263–275. 420 H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 Crick, F., & Koch, C. (1995). Are we aware of neural activity in primary visual cortex? Nature, 375, 121–123. Crick, F., & Koch, C. (1998). Consciousness and neuroscience. Cerebral Cortex, 8, 97–107. Crottaz-Herbette, S., Anagnoson, R. T., & Menon, V. (2004). Modality effects in verbal working memory: Differential prefrontal and parietal responses to auditory and visual stimuli. NeuroImage, 21, 340–351. Culham, J. C., & Kanwisher, N. G. (2001). Neuroimaging of cognitive functions in human parietal cortex. Current Opinion in Neurobiology, 11, 157–163. Curtis, C. E., & DÕEsposito, M. (2003). Persistent activity in the prefrontal cortex during working memory. Trends in Cognitive Sciences, 7(9), 415–423. DÕEsposito, M., Ballard, D., Aguirre, G. K., & Zarahn, E. (1998). Human prefrontal cortex is not specific for working memory: A functional MRI study. NeuroImage, 8, 274–282. Dade, L. A., Zatorre, R. J., Evans, A. C., & Jones-Gotman, M. (2001). Working memory in another dimension: Functional imaging of human olfactory working memory. NeuroImage, 14, 650–660. Damasio, A. R. (1989). Time-locked multiregional retroactivation: A systems-level proposal for the neural substrates of recall and recognition. Cognition, 33, 25–62. Dehaene, S., & Naccache, L. (2001). Towards a cognitive neuroscience of consciousness: Basic evidence and a workspace framework. Cognition, 79(1–2), 1–37. Dehaene, S., Naccache, L., Cohen, L., Le Bihan, D., Mangin, J.-F., & Poline, J.-B. (2001). Cerebral mechanisms of word masking and unconscious repetition priming. Nature Neuroscience, 4(7), 752–758. Deouell, L. Y. (2002). Pre-requisites for conscious awareness: Clues from electrophysiological and behavioral studies of unilateral neglect patients. Consciousness and Cognition, 11, 546–567. Desmedt, J., & Tomberg, C. (1995). Neurophysiology of preconscious and conscious mechanisms of the human brain. Electroencephalography and Clinical Neurophysiology/Electromyography and Motor Control, 97, S4. Dobbins, I. G., Foley, H., Schacter, D. L., & Wagner, A. D. (2002). Executive control during retrieval: Multiple prefrontal processes subserve source memory. Neuron, 35(5), 989–996. Dobbins, I. G., Rice, H. J., Wagner, A. D., & Schacter, D. L. (2003). Memory orientation and success: Separable neurocognitive components underlying episodic recognition. Neuropsychologia, 41, 318–333. Driver, J., & Vuilleumier, P. (2001). Perceptual awareness and its loss in unilateral neglect and extinction. Cognition, 79, 39–88. Druzgal, T. J., & DÕEsposito, M. (2001). Activity in fusiform face area modulated as a function of working memory load. Cognitive Brain Research, 10, 355–364. Duncan, J., & Owen, A. M. (2000). Common regions of the human frontal lobe recruited by diverse cognitive demands. Trends in Neurosciences, 23, 475–483. Edelman, G. M., & Tononi, G. (2000). A universe of consciousness: How matter becomes imagination. New York: Basic Books. Ehrsson, H. H., Spence, C., & Passingham, R. E. (2004). ThatÕs my hand! Activity in premotor cortex reflects feeling of ownership of a limb. Science, 305, 875–877. Engel, A. K., & Singer, W. (2001). Temporal binding and the neural correlates of sensory awareness. Trends in Cognitive Sciences, 5, 16–25. Eriksson, J., Larsson, A., Åhlström, K. R., & Nyberg, L. (2004). Visual consciousness: Dissociating the neural correlates of perceptual transitions from sustained perception with fMRI. Consciousness and Cognition, 13, 61–72. Farah, M. J., OÕReilly, R. C., & Vecera, S. P. (1997). The neural correlates of perceptual awareness: Evidence from covert recognition in prosopagnosia. In J. Cohen & J. Schooler (Eds.), Scientific approaches to consciousness. NJ: Lawrence Erlbaum. Feinstein, J. S., Stein, M. B., Castillo, G. N., & Paulus, M. P. (2004). From sensory processes to conscious perception. Consciousness and Cognition, 13, 323–335. ffytche, D. H., Howard, R. J., Brammer, M. J., David, A., Woodruff, P., & Williams, S. (1998). The anatomy of conscious vision: An fMRI study of visual hallucinations. Nature Neuroscience, 1(8), 738–742. Fletcher, P. C., & Henson, R. N. A. (2001). Frontal lobes and human memory: Insights from functional imaging. Brain, 124, 849–881. Friston, K. J. (2002). Bayesian estimation of dynamical systems: An application to fMRI. NeuroImage, 16, 513–530. H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 421 Frith, C., & Dolan, R. (1996). The role of the prefrontal cortex in higher cognitive functions. Cognitive Brain Research, 5(1–2), 175–181. Funahashi, S. (2001). Neuronal mechanisms of executive control by the prefrontal cortex. Neuroscience Research, 39, 147–165. Fuster, J. M., & Alexander, G. E. (1971). Neuron activity related to short-term memory. Science, 173, 652–654. Gaetz, M., Weinberg, H., Rzempoluck, E., & Jantzen, K. J. (1998). Neural network classifications and correlation analysis of EEG and MEG activity accompanying spontaneous reversals of the Necker cube. Cognitive Brain Research, 6, 335–346. Gitelman, D. R., Nobre, A. C., Parrish, T. B., LaBar, K. S., Kim, Y., Meyer, J. R., et al. (1999). A large-scale distributed network for covert spatial Attention: Further anatomical delineation based on stringent behavioural and cognitive controls. Brain, 122, 1093–1106. Gruber, O., & von Cramon, D. Y. (2003). The functional neuroanatomy of human working memory revisited: Evidence from 3-T fMRI studies using classical domain-specific interference tasks. NeuroImage, 19, 797–809. Hartley, A. A., & Speer, N. K. (2000). Locating and fractionating working memory using functional neuroimaging: Storage, maintenance, and executive functions. Microscopy Research and Technique, 51, 45–53. Hautzel, H., Mottaghy, F. M., Schmidt, D., Zemb, M., Shah, N. J., Müller-Gärtner, H. W., et al. (2002). Topographic segregation and convergence of verbal, object, shape and spatial working memory in humans. Neuroscience Letters, 323, 156–160. Henson, R. N. A., Shallice, T., & Dolan, R. J. (1999). Right prefrontal cortex and episodic memory retrieval: A functional MRI test of the monitoring hypothesis. Brain, 122, 1367–1381. Herron, J. E., Henson, R. N. A., & Rugg, M. D. (2004). Probability effects on the neural correlates of retrieval success: An fMRI study. NeuroImage, 21, 302–310. Hopfinger, J. B., Buonocore, M. H., & Mangun, G. R. (2000). The neural mechanisms of topdown attentional control. Nature Neuroscience, 3(3), 284–291. Irwin, D. E., & Andrews, R. V. (1996). Integration and accumulation of information across saccadic eye movements. In T. Inui & J. L. McClelland (Eds.), Attention and performance XVI: Information integration in perception and communication (pp. 125–155). Cambridge, MA: MIT Press. Jahanshahi, M., Dirnberger, G., Fuller, R., & Frith, C. D. (2000). The role of the dorsolateral prefrontal cortex in random number generation: A study with positron emission tomography. NeuroImage, 12, 713–725. Joliot, M., Ribary, U., & Llinas, R. (1994). Human oscillatory brain activity near 40 Hz coexists with cognitive temporal binding. Proceedings of the National Academy of Sciences of the United States of America, 91, 11748–11751. Jonides, J. (1995). Working memory and thinking. In E. E. Smith & D. N. Osherson (Eds.). An invitation to cognitive science: Thinking (Vol. 3). Cambridge, MA: MIT Press. Kaernbach, C., Schroger, E., Jacobsen, T., & Roeber, U. (1999). Effects of consciousness on human brain waves following binocular rivalry. NeuroReport, 10, 713–716. Kapur, S., Craik, F. I., Jones, C., Brown, G. M., Houle, S., & Tulving, E. (1995). Functional role of the prefrontal cortex in retrieval of memories: A PET study. NeuroReport, 6(14), 1880–1884. Kawashima, R., Imaizumi, S., Mori, K., Okada, K., Goto, R., Kiritani, S., et al. (1999). Selective visual and auditory attention toward utterances a PET study. Neuroimage, 10(2), 209–215. Kim, Y. H., Gitelman, D. R., Nobre, A. C., Parrish, T. B., LaBar, K. S., & Mesulam, M. M. (1999). The large-scale neural network for spatial attention displays multifunctional overlap but differential asymmetry. NeuroImage, 9, 269–277. Kjaer, T. W., Nowak, M., Kjaer, K. W., Lou, A. R., & Lou, H. C. (2001). Precuneus–prefrontal activity during awareness of visual verbal stimuli. Consciousness and Cognition, 10, 356–365. Kleinschmidt, A., Buchel, C., Zeki, S., & Frackowiak, R. S. J. (1998). Human brain activity during spontaneously reversing perception of ambiguous figures. Proceedings of the Royal Society of London, 265, 2427–2433. Klimesch, W., Doppelmayr, M., Yonelinas, A., Kroll, N. E. A., Lazzara, M., Rohm, D., et al. (2001). Theta synchronization during episodic retrieval: neural correlates of conscious awareness. Cognitive Brain Research, 12, 33–38. Koivisto, M., & Revonsuo, A. (2003). An ERP study of change detection, change blindness and visual awareness. Psychophysiology, 40, 423–429. 422 H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 Konishi, S., Wheeler, M. E., Donaldson, D. I., & Buckner, R. L. (2000). Neural correlates of episodic retrieval success. NeuroImage, 12, 276–286. Krause, B. J., Schmidt, D., Mottaghy, F. M., Taylor, J., Halsband, U., Herzog, H., et al. (1999). Episodic retrieval activates the precuneus irrespective of the imagery content of word pair associates: A PET study. Brain, 122, 255–263. LaBar, K. S., Gitelman, D. R., Parrish, T. B., & Mesulam, M.-M. (1999). Neuroanatomic overlap of working memory and spatial attention networks: A functional MRI comparison within subjects. NeuroImage, 10, 695–704. Lee, S. H., & Blake, R. (2002). V1 activity is reduced during binocular rivalry. Journal of Vision, 2, 618–626. Leopold, D. A., & Logothetis, N. K. (1996). Activity changes in early visual cortex reflect monkeysÕ percepts during binocular rivalry. Nature, 379, 549–553. Lepage, M., Ghaffar, O., Nyberg, L., & Tulving, E. (2000). Prefrontal cortex and episodic memory retrieval mode. Proceedings of the National Academy of Sciences of the United States of America, 97(1), 506–511. Leube, D. T., Erb, M., Grodd, W., Bartels, M., & Kircher, T. T. J. (2003). Successful episodic memory retrieval of newly learned faces activates a left fronto-parietal network. Cognitive Brain Research, 18, 97–101. Leveroni, C. L., Seidenberg, M., Mayer, A. R., Mead, L. A., Binder, J. R., & Rao, S. M. (2000). Neural systems underlying the recognition of familiar and newly learned faces. The Journal of Neuroscience, 20(2), 878–886. Lipschutz, B., Kolinsky, R., Damhaut, P., Wikler, D., & Goldman, S. (2002). Attention-dependent changes of activation and connectivity in dichotic listening. NeuroImage, 17, 643–656. Logothetis, N. K. (1998). Single units and conscious vision. Philosophical Transactions of the Royal Society of London, B, 353, 1801–1818. Logothetis, N. K., & Wandell, B. A. (2004). Interpreting the BOLD Signal. Annual Review of Physiology, 66, 735–769. Luck, S. J., & Vogel, E. K. (1997). The capacity of visual working memory for features and conjunctions. Nature, 390, 279–281. Lumer, E. D., & Rees, G. (1999). Covariation of activity in visual and prefrontal cortex associated with subjective visual perception. Proceedings of the National Academy of Sciences of the United States of America, 96, 1669–1673. Lumer, E. D., Friston, K. J., & Rees, G. (1998). Neural correlates of perceptual rivalry in the human brain. Science, 280, 1930–1934. Lundstrom, B. N., Petersson, K. M., Andersson, J., Johansson, M., Fransson, P., & Ingvar, M. (2003). Isolating the retrieval of imagined pictures during episodic memory: Activation of the left precuneus and left prefrontal cortex. NeuroImage, 20, 1934–1943. Luria, A. R. (1966). Higher cortical functions. New York: Basic Book. Macaluso, E., Frith, C. D., & Driver, J. (2002). Directing attention to locations and to sensory modalities: Multiple levels of selective processing revealed with PET. Cerebral Cortex, 12, 357–368. Mäkinen, V., May, P., & Tiitinen, H. (2004). Transient brain responses predict the temporal dynamics of sound detection in humans. NeuroImage, 21, 701–706. Manoach, D. S., White, N. S., Lindgren, K. A., Heckers, S., Coleman, M. J., Dubal, S., et al. (2004). Hemispheric specialization of the lateral prefrontal cortex for strategic processing during spatial and shape working memory. NeuroImage, 21(3), 894–903. Marklund, P., Nyberg, L., Fransson, P., Ingvar, M., Petersson, K.M., Cabeza, R. (2003). Similarities and differences in state-and item-related BOLD signal changes during long-term memory tasks: A test of the working-memory hypothesis. Poster presented at the First NorFA Meeting: Functional Neuroimaging Network, Sodergam, Sweden, August 28–31, 2003. Marois, R., Yi, D. J., & Chun, M. M. (2004). The neural fate of consciously perceived and missed events in the attentional blink. Neuron, 41, 465–472. Mayes, A. R., & Montaldi, D. (2001). Exploring the neural bases of episodic and semantic memory: The role of structural and functional neuroimaging. Neuroscience and Biobehavioral Reviews, 25(6), 555–573. McIntosh, A. R. (2000). Towards a network theory of cognition. Neural Networks, 13, 861–870. McIntosh, A. R., Rajah, M. N., & Lobaugh, N. J. (1999). Interactions of prefrontal cortex related to awareness in sensory learning. Science, 284, 1531–1533. McIntosh, A. R., Rajah, M. N., & Lobaugh, N. J. (2003). Functional connectivity of the medial temporal lobe relates to learning and awareness. Journal of Neuroscience, 23, 6520–6528. H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 423 Mecklinger, A., Bosch, V., Gruenewald, C., Bentin, S., & von Cramon, D. Y. (2000). What have Klingon letters and faces in common? An fMRI study on content-specific working memory systems. Human Brain Mapping, 11, 146–161. Moscovitch, M. (1992). Memory and working with memory: A component process model based on modules and central systems. Journal of Cognitive Neuroscience, 4, 257–267. Moutoussis, K., & Zeki, S. (2002). The relationship between cortical activation and perception investigated with invisible stimuli. Proceedings of the National Academy of Sciences of the United States of America, 99, 9527–9532. Newman, J., & Baars, B. J. (1993). A neural attentional model for access to consciousness: A global workspace perspective. Concepts in Neuroscience, 4, 255–290. Niebur, E. (2002). Electrophysiological correlates of synchronous neural activity and attention: A short review. Biosystems, 67(1–3), 157–166. Nolde, S. F., Johnson, M. K., & DÕEsposito, M. (1998). Left prefrontal activation during episodic memory: An eventrelated study. NeuroReport, 9, 3509–3514. Norman, D., & Shallice, T. (1986). Attention to action: Willed and automatic control of behavior. In R. Davidson, G. Schwartz, & D. Shapiro (Eds.). Consciousness and self regulation: Advances in research and theory (Vol. 4, pp. 1–18). New York: Plenum. Nyberg, L., & Cabeza, R. (in press). Similarities in brain activity during working memory, semantic memory, and episodic memory: Implications for theories of memory. In S. Algarabel, A. Pitarque, T. Bajo, S.E. Gathercole, M.A. Conway (Eds.), Theories of memory, Vol. 3. Psychology Press. Nyberg, L., & McIntosh, A. R. (2000). Functional neuroimaging: Network analyses. In R. Cabeza & A. Kingstone (Eds.), Handbook of functional neuroimaging of cognition (pp. 49–72). Cambridge, MA: MIT Press. Nyberg, L., Forkstam, C., Petersson, K.-M., Cabeza, R., & Ingvar, M. (2002). Brain imaging of human memory systems: Between-systems similarities and within-system differences. Cognitive Brain Research, 13, 281–292. Nyberg, L., Marklund, P., Persson, J., Cabeza, R., Forkstam, C., Petersson, K. M., et al. (2003). Common prefrontal activations during working memory, episodic memory, and semantic memory. Neuropsychologia, 41, 371–377. Nyberg, L., McIntosh, A. R., & Tulving, E. (1998). Functional brain imaging of episodic and semantic memory with positron emission tomography. Journal of Molecular Medicine, 76(1), 48–53. Nyberg, L., Tulving, E., Habib, R., Nilsson, L. G., Kapur, S., Houle, S., et al. (1995). Functional brain maps of retrieval mode and recovery of episodic information. NeuroReport, 7(1), 249–252. Nystrom, L. E., Braver, T. S., Sabb, F. W., Delgado, M. R., Noll, D. C., & Cohen, J. D. (2000). Working memory for letters, shapes and locations: fMRI evidence against stimulus-based regional organization in human prefrontal cortex. NeuroImage, 11, 424–446. Ojanen, V., Revonsuo, A., & Sams, M. (2003). Visual awareness of low-contrast stimuli is reflected in event-related brain potentials. Psychophysiology, 40, 192–197. Overgaard, M., Nielsen, J. F., & Fuglsang-Frederiksen, A. (2004). A TMS study of the ventral projections from V1 with implications for the finding of neural correlates of consciousness. Brain and Cognition, 54(1), 58–64. Parkin, A. J. (1999). Component processes versus systems: Is there really an important difference. In J. K. Foster & M. Jelicic (Eds.), Memory: Systems, process, or function? (pp. 273–287). New York: Oxford University Press. Pessoa, L., Kastner, S., & Ungerleider, L. G. (2003). Neuroimaging studies of attention: From modulation of sensory processing to top-down control. The Journal of Neuroscience, 23(10), 3990–3998. Pins, D., & Ffytche, D. (2003). The neural correlates of conscious vision. Cerebral Cortex, 13(5), 461–474. Platel, H., Baron, J.-C., Desgranges, B., Bernard, F., & Eustache, F. (2003). Semantic and episodic memory of music are subserved by distinct neural networks. NeuroImage, 20, 244–256. Polonsky, A., Blake, R., Braun, J., & Heeger, D. J. (2000). Neuronal activity in human primary visual cortex correlates with perception during binocular rivalry. Nature Neuroscience, 3, 1153–1159. Portas, C. M., Strange, B. A., Friston, K. J., Dolan, R. J., & Frith, C. D. (2000). How does the brain sustain a visual percept? Proceedings of The Royal Society of London, 267, 845–850. Postle, B. R., & DÕEsposito, M. (2000). Evaluating models of the topographical organization of working memory function in frontal cortex with event-related fMRI. Psychobiology, 28, 132–145. Prabhakaran, V., Narayanan, K., Zhao, Z., & Gabrieli, J. D. E. (2000). Integration of diverse information in working memory within the frontal lobe. Nature Neuroscience, 3, 85–90. 424 H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 Ragland, J. D., Gur, R. C., Lazarev, M. G., Smith, R. J., Schroeder, L., Raz, J., et al. (2000). Hemispheric activation of anterior and inferior prefrontal cortex during verbal encoding and recognition: A PET study of healthy volunteers. NeuroImage, 11, 624–633. Raichle, M. E. (1998). Behind the scenes of functional brain imaging: A historical and physiological perspective. Proceedings of the National Academy of Sciences of the United States of America, 95, 765–772. Ranganath, C., Johnson, M. K., & DÕEsposito, M. (2000). Left anterior prefrontal activation increases with demands to recall specific perceptual information. Journal of Neuroscience, 20(RC108), 1–5. Ranganath, C., Johnson, M. K., & DÕEsposito, M. (2003). Prefrontal activity associated with working memory and episodic long-term memory. Neuropsychologia, 41, 378–389. Rees, G., Russell, C., Frith, C. D., & Driver, J. (1999). Inattentional blindness versus inattentional amnesia for fixated but ignored words. Science, 286, 2504–2507. Rees, G., & Lavie, N. (2001). What can functional imaging reveal about the role of attention in visual awareness?. Neuropsychologia, 39, 1343–1353. Rees, G., Kreiman, G., & Koch, C. (2002). Neural correlates of consciousness in humans. Nature Reviews Neuroscience, 3, 261–270. Revonsuo, A. (1999). Binding and the phenomenal unity of consciousness. Consciousness and Cognition, 8(2), 173–185. Reynolds, J. H., Pasternak, T., & Desimone, R. (2000). Attention increases sensitivity of V4 neurons. Neuron, 26, 703–714. Ro, T., Breitmeyer, B., Burton, P., Singhal, N. S., & Lane, D. (2003). Feedback contributions to visual awareness in human occipital cortex. Current Biology, 13(12), 1038–1041. Rodriguez, E., George, N., Lachaux, J. P., Martinerie, J., Renault, B., & Varela, F. J. (1999). PerceptionÕs shadow: Long-distance synchronization of human brain activity. Nature, 397, 430–433. Roediger, H. L., Buckner, R. L., & McDermott, K. B. (1999). Components of processing. In J. K. Foster & M. Jelicic (Eds.), Memory: Systems, process, or function? (pp. 31–65). Oxford: Oxford University Press. Rosen, A. C., Rao, S. M., Caffarra, P., Scaglioni, A., Bobholz, J. A., Woodley, S. J., et al. (1999). Neural basis of endogenous and exogenous spatial orienting: A functional MRI study. Journal of Cognitive Neuroscience, 11(2), 135–152. Rowe, J., Friston, K., Frackowiak, R., & Passingham, R. (2002). Attention to action: Specific modulation of corticocortical interactions in humans. NeuroImage, 17, 988–998. Rugg, M. D., & Wilding, E. L. (2000). Retrieval processing and episodic memory: Electrophysiological and neuroimaging evidence. Trends in Cognitive Sciences, 4, 108–115. Sahraie, A., Weiskrantz, L., Barbur, J. L., Simmons, A., Williams, S. C. R., & Brammer, M. J. (1997). Pattern of neuronal activity associated with conscious and unconscious processing of visual signals. Proceedings of the National Academy of Sciences of the United States of America, 94, 9406–9411. Sarnthein, J., Petsche, H., Rappelsberger, P., Shaw, G. L., & von Stein, A. (1998). Synchronization between prefrontal and posterior association cortex during human working memory. Proceedings of the National Academy of Sciences of the United States of America, 95, 7092–7096. Schmidt, D., Krause, B. J., Mottaghy, F. M., Halsband, U., Herzog, H., Tellmann, L., et al. (2002). Brain systems engaged in encoding and retrieval of word-pair associates independent of their imagery content or presentation modalities. Neuropsychologia, 40, 457–470. Sengpiel, F., Blakemore, C., & Harrad, R. (1995). Interocular suppression in the primary visual cortex: a possible neural basis of binocular rivalry. Vision Research, 35, 179–195. Shaywitz, B. A., Shaywitz, S. E., Pugh, K. R., Fulbright, R. K., Skudlarski, P., Mencl, W. E., et al. (2001). The functional neural architecture of components of attention in language-processing tasks. NeuroImage, 13, 601–612. Sheinberg, D. L., & Logothetis, N. K. (1997). The role of temporal cortical areas in perceptual organisation. Proceedings of the National Academy of Sciences of the United States of America, 94, 3408–3416. Smith, E. E., & Jonides, J. (1999). Storage and executive processes in the frontal lobes. Science, 283, 1657–1661. Smith, E. E., Jonides, J., & Koeppe, R. A. (1996). Dissociating verbal and spatial working memory using PET. Cerebral Cortex, 6, 11–20. H.R. Naghavi, L. Nyberg / Consciousness and Cognition 14 (2005) 390–425 425 Stephan, K. M., Thaut, M. H., Wunderlich, G., Schicks, W., Tian, B., Tellmann, L., et al. (2002). Conscious and subconscious sensorimotor synchronization—prefrontal cortex and the influence of awareness. NeuroImage, 15, 345–352. Sterzer, P., Russ, M. O., Preibisch, C., & Kleinschmidt, A. (2002). Neural correlates of spontaneous direction reversals in ambiguous apparent visual motion. NeuroImage, 15, 908–916. Strüber, D., & Herrmann, C. S. (2002). MEG alpha activity decrease reflects destabilization of multistable percepts. Cognitive Brain Research, 14(3), 370–382. Summerfield, C., Jack, A. I., & Burgess, A. P. (2002). Induced gamma activity is associated with conscious awareness of pattern masked nouns. International Journal of Psychophysiology, 44, 93–100. Suzuki, M., Fujii, T., Tsukiura, T., Okuda, J., Umetsu, A., Nagasaka, T., et al. (2002). Neural basis of temporal context memory: A functional MRI study. NeuroImage, 17, 1790–1796. Talairach, J., & Tournoux, P. (1988). Co-planar stereotaxic atlas of the human brain. New York: Theime. Tong, F., Nakayama, K., Vaughan, J. T., & Kanwisher, N. (1998). Binocular rivalry and visual awareness in human extrastriate cortex. Neuron, 21, 753–759. Treisman, A. (1988). Features and objects: The Fourteenth Bartlett Memorial Lecture. Quarterly Journal of Experimental Psychology A, 40, 201–237. Treisman, A. (1996). The binding problem. Current Opinion in Neurobiology, 6, 171–178. Tsukiura, T., Fujii, T., Takahashi, T., Xiao, R., Inase, M., Iijima, T., et al. (2001). Neuroanatomical discrimination between manipulating and maintaining processes involved in verbal working memory; a functional MRI study. Cognitive Brain Research, 11, 13–21. Tulving, E. (1983). Elements of episodic memory. New York: Oxford University Press. Vanni, S., Revonsuo, A., Saarinen, J., & Hari, R. (1996). Visual awareness of objects correlates with activity of right occipital cortex. NeuroReport, 8, 183–186. Veltman, D. J., Rombouts, S. A. R. B., & Dolan, R. J. (2003). Maintenance versus manipulation in verbal working memory revisited: An fMRI study. NeuroImage, 18, 247–256. Vuilleumier, P., Armony, J. L., Clarke, K., Husain, M., Driver, J., & Dolan, R. J. (2002). Neural response to emotional faces with and without awareness: Event-related fMRI in a parietal patient with visual extinction and spatial neglect. Neuropsychologia, 40, 2156–2166. Wagner, A. D. (1999). Working memory contributions to human learning and remembering. Neuron, 22, 19–22. Walla, P., Hufnagl, B., Lehrner, J., Mayer, D., Lindinger, G., Deecke, L., et al. (2002). Evidence of conscious and subconscious olfactory information processing during word encoding: A magnetoencephalographic (MEG) study. Cognitive Brain Research, 14(3), 309–316. Whatham, A. R., Vuilleumier, P., Landis, T., & Safran, A. B. (2003). Visual consciousness in health and disease. Neurologic Clinics, 21(3), 647–686. Wheeler, M. E., & Treisman, A. M. (2002). Binding in short-term visual memory. Journal of Experimental Psychology, 131(1), 48–64. Wilenius-Emet, M., Revonsuo, A., & Ojanen, V. (2004). An electrophysiological correlate of human visual awareness. Neuroscience Letters, 354, 38–41. Willingham, D. B., Salidis, J., & Gabrieli, J. D. E. (2002). Direct comparison of neural systems mediating conscious and unconscious skill learning. Journal of Neurophysiology, 88, 1451–1460. Yantis, S., & Serences, J. T. (2003). Cortical mechanisms of space-based and object-based attentional control. Current Opinion in Neurobiology, 13, 187–193. Yantis, S., Schwarzbach, J., Serences, J. T., Carlson, R. L., Steinmetz, M. A., Pekar, J. J., et al. (2002). Transient neural activity in human parietal cortex during spatial attention shifts. Nature Neuroscience, 5, 995–1002. Zatorre, R. J., Mondor, T. A., & Evans, A. C. (1999). Auditory attention to space and frequency activates similar cerebral systems. NeuroImage, 10, 544–554. Zurowski, B., Gostomzyk, J., Grön, G., Weller, R., Schirrmeister, H., Neumeier, B., et al. (2002). Dissociating a common working memory network from different neural substrates of phonological and spatial stimulus processing. NeuroImage, 15, 45–57.

Consciousness and Cognition 42 (2016) 101–112 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Intentional action processing results from automatic bottom-up attention: An EEG-investigation into the Social Relevance Hypothesis using hypnosis Eleonore Neufeld a,b,c,⇑, Elliot C. Brown d,e, Sie-In Lee-Grimm e, Albert Newen b,c, Martin Brüne c,e a School of Philosophy, USC Dana and David Dornsife College for Letters, Arts and Sciences, University of Southern California, United States Institute of Philosophy II, Faculty of Philosophy and Education, Ruhr-University Bochum, Germany Center for Mind, Brain and Cognitive Evolution, Ruhr-University Bochum, Germany d The Mathison Centre for Mental Health Research and Education, Hotchkiss Brain Institute, University of Calgary, Canada e Research Department of Cognitive Neuropsychiatry, LWL University Hospital Bochum, Ruhr-University Bochum, Germany b c a r t i c l e i n f o Article history: Received 27 August 2015 Revised 10 January 2016 Accepted 2 March 2016 Keywords: Intentional action Social Relevance Hypothesis Bottom-up attention Hypnosis Mu rhythm Social attention Social cognition a b s t r a c t Social stimuli grab our attention. However, it has rarely been investigated how variations in attention affect the processing of social stimuli, although the answer could help us uncover details of social cognition processes such as action understanding. In the present study, we examined how changes to bottom-up attention affects neural EEG-responses associated with intentional action processing. We induced an increase in bottom-up attention by using hypnosis. We recorded the electroencephalographic l-wave suppression of hypnotized participants when presented with intentional actions in first and third person perspective in a video-clip paradigm. Previous studies have shown that the l-rhythm is selectively suppressed both when executing and observing goal-directed motor actions; hence it can be used as a neural signal for intentional action processing. Our results show that neutral hypnotic trance increases l-suppression in highly suggestible participants when they observe intentional actions. This suggests that social action processing is enhanced when bottom-up attentional processes are predominant. Our findings support the Social Relevance Hypothesis, according to which social action processing is a bottom-up driven attentional process, and can thus be altered as a function of bottomup processing devoted to a social stimulus. Published by Elsevier Inc. 1. Introduction Attention plays a key role in most domains of cognition, but it seems to be especially attuned to identify social stimuli. A large body of evidence supports this observation: human infants inherit an attentional preference for social stimuli such as faces (Gliga, Elsabbagh, Andravizou, & Johnson, 2009), we quickly attend to human faces and bodies in naturalistic scenes (Fletcher-Watson, Findlay, Leekam, & Benson, 2008), social stimuli have a higher probability to automatically break through ⇑ Corresponding author at: USC School of Philosophy, 3709 Trousdale Parkway, Los Angeles, CA 90089-045, United States. E-mail address: eneufeld@usc.edu (E. Neufeld). http://dx.doi.org/10.1016/j.concog.2016.03.002 1053-8100/Published by Elsevier Inc. 102 E. Neufeld et al. / Consciousness and Cognition 42 (2016) 101–112 conscious awareness than non-social stimuli (Jiang, Costello, & He, 2007), and direct eye-gaze easily captures attention by both children and adults (Senju & Johnson, 2009) – just to name few examples. Although a complete dichotomous separation of top-down and bottom-up attentional processes is not possible – since they seem to interact with each other in most cognitive tasks (Egeth & Yantis, 1997; Sarter, Givens, & Bruno, 2001; but note Firestone & Scholl, 2015) – we can still ask whether we process social stimuli such as intentional actions in a way that is predominantly guided by automatic, bottom-up processes (cf. also Cook, Barbalat, & Blakemore, 2012). Broadly, automatic bottom-up processes are those processes that are obligatory, and hard to ignore, suppress or alter (Shiffrin & Schneider, 1977).1 An attempt to answer that question is captured by the Social Relevance Hypothesis (SRH) (Oberman, Ramachandran, & Pineda, 2008). According to SRH, various capacities in social cognition crucially depend on social stimuli being automatically assigned a high degree of attentional relevance. SRH has been mainly used to understand Autism Spectrum Disorder (ASD). For example, Oberman et al. (2008) argued that the ‘‘dysfunction reported in individuals with ASD may be the result of lack of social relevance [for them] in the stimuli used”. According to SRH, the low performance of ASD patients in certain social tasks could be due to problems in the way their attention latches on to social stimuli, rather than in their social competence per se. Chevallier, Kohls, Troiani, Brodkin, and Schultz (2012) also defend SRH mainly in the context of ASD. They point out that, because social interactions and cooperative behavior lead to important fitness benefits, social stimuli are automatically prioritized. According to their theory, this social attention system is malfunctioning in ASD patients, and impedes performance in subsequent social tasks and processes. Support for this causal relation came from experiments investigating the pattern of the EEG l-rhythm in ASD patients. The l-rhythm is an EEG rhythm recorded over sensorimotor cortices that oscillates in the alpha (8–12 Hz) and beta (15– 25 Hz) frequency band (Hari, 2006; Pineda, 2005). Recent evidence furthermore suggests distinguishable subcomponents of the l-rhythm in the lower and upper alpha range (Dumas, Soussignan, Hugueville, Martinerie, & Nadel, 2014; Sebastiani et al., 2014). The l is selectively suppressed when intentional actions are performed or observed by an agent, which is thought to reflect event-related desynchronization of motor cortices (Hari, 2006; Pineda, 2005). Due to its functionally related oscillation pattern, the l is commonly assumed to index intentional action processing (Hari, 2006; Kilner, Marchant, & Frith, 2006; Oberman et al., 2005; Perry & Bentin, 2009; Perry, Stein, & Bentin, 2011; Pineda, 2005). Oberman et al. (2005) found that ASD patients showed a clear suppression of the l when they themselves executed motor actions, but lacked the suppression pattern healthy adults display when they observed actions of others (see also Bernier, Dawson, Webb, & Murias, 2007; Raymaekers, Wiersema, & Roeyers, 2009). A subsequent study by Oberman et al. (2008), however, qualified this effect: when ASD patients watched intentional actions of familiar persons rather than of strangers, their l-pattern was normalized. According to the authors, these results suggest that adequate social processing depends on the relevance or salience which our attentional system grants to social stimuli. Some preliminary support for SRH also comes from studies with psychologically unaffected individuals. Kilner and Lemon (2013) tested whether, if the type of action is fixed, the perspective from which it is viewed can affect the MEG lsuppression. Their idea was that watching an action from the front should capture attention to a higher degree (the action could directly affect you) than watching actions from the back (the action need not directly concern you), and that this difference in attentional relevance affects action processing. As predicted, the l-suppression was stronger when observing actions from the front compared to when observing them from the back, suggesting that actions with different degrees of relevance are represented differently. Brown, Wiersema, Pourtois, and Brüne (2013) showed that the strength of the lsuppression was modulated by the degree of reward associated with actions, suggesting that we devote a higher amount of attentional resources to actions that are more rewarding (cf. also Trilla Gros, Panasiti, & Chakrabarti, 2015). These findings motivate the idea that the degree of relevance of the social stimuli we are confronted with is a key element for social processing. Now, evidence for SRH primarily stems from experiments that focus on the effects on social performance of attentional manipulations in ASD patients. However, although SRH’s predictions about socio-pathological patient groups such as ASD are important and interesting, SRH is a general hypothesis about the mechanisms underlying social processing. SRH claims that, under normal conditions, social stimuli come with a high degree of behavioral, and thus attentional relevance. Since stimuli that usually display a high degree of biological and behavioral relevance should be processed efficiently in an automatic and bottom-up manner (Eastwood, Smilek, & Merikle, 2001; Egeth & Yantis, 1997; Geng & Mangun, 2009; Parkhurst, Law, & Niebur, 2002; Peck, Jangraw, Suzuki, Efem, & Gottlieb, 2009; Pratto & John, 1991; Summerfield & Egner, 2009; Treue, 2003; Wang, Cavanagh, & Green, 1994), a prediction of SRH is that normally, social stimuli capture our attention automatically and are thus processed in a bottom-up driven way. Consequently, the proper processing of such stimuli depends on that bottom-up automaticity. It should then generally be possible to boost social cognition performance by enhancing bottom-up attention to social stimuli. To test this prediction, we conducted an EEG experiment in which we examined whether boosting attentional bottom-up processing increases l-suppression, taken as a signal for intentional action processing. 1 We want to clarify that we focus on automatic processes triggered by bottom-up attention. This presupposes the well-accepted distinction between bottom-up and top-down attentional mechanisms, but does not imply that automatic processes triggered by bottom-up attention are independent from complex cognitive processes. Bottom-up attention might still be subject to cognitive penetration, which is the modification of our perceptual experience by higher-level cognition (Vetter & Newen, 2014). The important distinction we draw on in this paper is between automatic processes which are triggered only by bottom-up attention, and those non-automatic processes which also involve top-down attentional processes (e.g., higher-level control processes). E. Neufeld et al. / Consciousness and Cognition 42 (2016) 101–112 103 To boost automatic bottom-up processes, we used hypnosis. Hypnosis is now seen as a powerful tool to influence both bottom-up and top-down attentional processes (Egner, Jamieson, & Gruzelier, 2005; Egner & Raz, 2007; Iani, Ricci, Baroni, & Rubichi, 2009; Lifshitz, Aubert Bonn, Fischer, Kashem, & Raz, 2013; Raz, 2011; Raz, Fan, & Posner, 2005, 2006; Raz, Shapiro, Fan, & Posner, 2002), and has the advantage of altering attentional processes of subjects ‘‘from inside” instead of exogenously by changing the presented stimuli. Consider as an illustration the famous Stroop test (Stroop, 1935), the ‘‘hallmark” (Crawford, 1994) of attention research. It requires subjects to name the ink color of words whose meaning picks out an incongruent color (e.g., ‘‘GREEN” presented in red ink). The lexical retrieval of the word is automatized to such an extent that it interferes dramatically with top-down attentional control processes that are required for the correct task performance. Accordingly, participants are usually slow and not very accurate in naming the colors of the words. An array of studies demonstrated that when highly suggestible subjects are put into neutral hypnotic trance, i.e., when they are not given any explicit hypnotic instructions, participants’ performance in the Stroop task becomes worse, with significant increases in reaction times or deteriorating accuracy rates (Jamieson & Sheehan, 2004; Kaiser, Barker, Haenschel, Baldeweg, & Gruzelier, 1997; Kallio, Revonsuo, Hämäläinen, Markela, & Gruzelier, 2001; Nordby, Hugdahl, Jasiukaitis, & Spiegel, 1999; Sheehan, Donovan, & MacLeod, 1988). These findings robustly demonstrate that hypnotic induction without any specific instructions inhibits top-down control processes, giving way for automatic bottom-up processes. The bottom-up enhancing and cognitive control inhibiting effect of hypnosis is further substantiated by evidence from electrophysiological (Kaiser et al., 1997) and functional neuroimaging (Egner et al., 2005) studies, as well as behavioral tests assessing frontal functions such as verbal fluency (Gruzelier & Warren, 1993; Kallio et al., 2001; cf. also Farvolden & Woody, 2004). For example, Kaiser et al. (1997) registered a significant reduction in the positive error ERP component during neutral hypnosis in the Stroop paradigm, a component typically seen as indexing error evaluation. This reduction might reflect a failure of top-down inhibition of the automatic Stroop response, and a weaker concern of the participants to inhibit errors (cf. also Gruzelier, 1998). Kallio et al. (2001) tested the effect of hypnosis on behavioral tasks functionally related to anterior control, and found negative effects on word fluency tasks for participants with high suggestibility values when under hypnosis (cf. also Gruzelier, 1998). In an experiment deploying event-related fMRI and EEG, Egner et al. (2005) tested whether hypnosis generates changes in brain areas usually associated with attentional control, and found that hypnosis leads to a decoupling of conflict-monitoring activity of the ACC and frontal activity related to attentional control, accompanied by a decrease of functional connectivity between frontal midline and left lateral scalp sites as indexed by the EEG gamma-band (see also Cardeña, Jönsson, Terhune, & Marcusson-Clavertz, 2013). Importantly, those results are generally limited to highly suggestible participants. Within hypnosis research, the dissociation of high and low suggestibility groups is standard procedure, since experimental evidence manifests inter-individual variations between both groups in various cognitive domains (Oakley & Halligan, 2013). Consequently, experimental hypotheses about the effects of hypnosis are usually restricted to highly suggestible participants. Findings of that sort have corroborated the wide-spread view that hypnosis leads to an inhibition of frontal control processes (Crawford & Gruzelier, 1992; Farvolden & Woody, 2004; Gruzelier, 1998, 2000; Kihlstrom, 2013) or dissociation of frontally mediated top-down control processes from other processes they usually cooperate with (Egner & Raz, 2007; Egner et al., 2005; Hilgard, 1977; Jamieson & Sheehan, 2004; Jamieson & Woody, 2007; Woody & Bowers, 1994; cf. also Landry, Appourchaux, & Raz, 2014). As a result, highly suggestible individuals are seen as less likely to produce internally generated second order thoughts, but more prone to process stimuli that capture attention automatically in a bottom-up way (Egner & Hirsch, 2005; Egner & Raz, 2007; Egner et al., 2005; Jamieson & Sheehan, 2004; Jamieson & Woody, 2007). That effect, i.e., the enhancement of bottom-up processing by neutral hypnosis, is precisely what we aimed to make use of in our experiment, extended to the domain of social cognition. We presented hypnotized participants videos in which they observed intentional actions in a first (self) and third-person (other) perspective without giving them any specific hypnotic instruction, aiming to trace how intensified bottom-up processing affects the l as a signal for intentional action processing. We hypothesized that if intentional action processing is a bottom-up driven, automatic process – as predicted by SRH – lsuppression should increase in highly suggestible individuals when under neutral hypnosis. It should also be noted that so far, no study genuinely investigated the effect of increased bottom-up processing on the EEG l-oscillation, so our study can provide interesting insights about the influencing factors of the phenomenology of the l-wave. 2. Methods 2.1. Participants Nineteen (15 female) healthy right-handed participants were included in the EEG study, following recruitment from the Ruhr-University Bochum and following pre-selection, with a mean age of 22.53 years (SD 3.74). Participants with psychiatric disorders were excluded, according to the Mini-DIPS (Diagnostisches Interview bei Psychischen Storungen: Kurzfassung; Margraf, 1994) diagnostic questionnaire. All participants were informed of all procedures and consent was obtained from all. 104 E. Neufeld et al. / Consciousness and Cognition 42 (2016) 101–112 2.2. Procedure 2.2.1. Behavioral measures Participants were given the German version of the Short Suggestibility Scale (SSS; Kotov, Bellman, & Watson, 2007), which comprises 21 items of the Multidimensional Iowa Suggestibility Scale (MISS; Kotov et al., 2007), to determine the individual degree of suggestibility. Empathy was also assessed with the German version on the Interpersonal Reactivity Index (IRI; Davis, 1983); the Saarbrückener Personality Questionnaire (SPF; Paulus, 2009), which rates four empathy factors: fantasy, perspective taking, empathetic concern and personal distress. These questionnaires were all completed before the EEG testing session began. 2.2.2. Pre-selection A professional hypnotist was employed to perform the pre-selection procedure as well as the main hypnotic induction. The pre-selection procedure was used to identify participants that reached a sufficiently deep state of hypnotic trance and were therefore deemed suitable to be included in the EEG testing session. This procedure involved groups of 10 potential candidates that were administered a collective hypnosis. The hypnotist introduced a general hypnotic trance using rhythmic and calm speech, while giving explicit, but collective hypnotic suggestions. Hypnotic suggestions included instructions such as to forget their name and to imagine themselves playing in an orchestra. Based on the degree to which candidates submitted to the suggestions, the hypnotist selected suitable participants according to his personal judgment and experience. Following this collective hypnosis and pre-selection, the hypnotic state was terminated. 2.2.3. Hypnotic induction Only pre-selected participants were recruited for the EEG testing session. Shortly before the start of the EEG experiment, the hypnotist induced a five minute mini-hypnotic state to facilitate the hypnotic procedure in the main testing phase. All participants performed the EEG paradigm twice in a single session, and, to control for potential order effects, half of the participants performed the task first under the hypnotic state, and then off hypnosis (i.e. the control condition), and the other half followed the opposite order. Participants were randomly assigned to either having the control condition first or the hypnosis first. For the main hypnotic induction for the on hypnosis condition of the EEG testing, the hypnotist induced the full state of neutral hypnotic trance. He engaged in rhythmic, calm speech, telling the subject that they feel tired, feel their eyes becoming heavy, and told them to imagine falling deeper and deeper. The induction period lasted for approximately 3– 5 min. Testing with the EEG paradigm started immediately after hypnotic induction. After every main block of trials, the hypnotist engaged in further hypnotic speech with the participant to maintain a sufficient degree of hypnotic trance. After the EEG paradigm was complete, which lasted approx. 30 min, the hypnotist terminated the hypnotic trance. To do this, he told participants that he would slowly count backwards from ten to one, instructing participants that they should become awake and clear when he reached number one. For the control condition (i.e. off hypnosis), participants completed the same EEG paradigm, but without the hypnotic induction. 2.2.4. EEG procedure and paradigm Throughout the task, participants were seated with their hands flat on the table in front of the computer screen. They were presented with a series of video sequences displaying two persons sitting at a table and transferring coins from one bowl to one of three other colored bowls (orange, blue and green) in the middle of the table. The paradigm consisted of self and other conditions. In the ‘‘other” condition, the participant saw the person sitting facing them (i.e. third-person perspective) performing the action of transferring a coin from one bowl to another with their right hand. In the ‘‘self” condition, the participant saw the person facing the same direction as the participant (i.e. first-person perspective) performing the action of transferring a coin from one bowl to another. Before the video sequences began, the test persons were explicitly instructed to imagine these hands to be their own hands. Each trial consisted of four parts, lasting a total of 11s, as shown in Fig. 1. A block of 8 practice trials came before the 6 blocks (20 trials each block) of main trials with a single video used in each trial. Videos were presented in a pseudorandom order. Participants were asked to count the number of coins put into each of the bowl. After each of the main blocks, a short break was given and participants were asked how many coins they had counted for each bowl. The counting was only intended to ensure participants were paying attention to the stimuli. 2.3. EEG recording and analysis 2.3.1. EEG data acquisition EEG was recorded at a sampling rate of 512 Hz from 32 channels with a Brain Products BrainAmp system, using passive AgCl electrodes held in a BrainCap elasticated EEG cap (from EasyCap). The 32 electrode positions were distributed over the scalp according to the international 10–20 EEG system. An additional electrode was placed below the left eye (EOG) to capture eye blinks. According to the EasyCap 32-channel standard electrode configuration, the ground electrode was located along the midline, anterior to electrode Fz. The reference electrode was located along the midline, more posteriorly, between electrodes Cz and Fz. Following acquisition, the raw data were processed offline using EEGLab, a MATLAB based open-source toolbox (Delorme & Makeig, 2004). Firstly, the data was visually inspected for segments with obvious artefacts that were E. Neufeld et al. / Consciousness and Cognition 42 (2016) 101–112 105 Fig. 1. Experimental design of EEG paradigm, showing the action from (a) the third-person perspective (i.e. other condition), and (b) the first-person perspective (self condition). then removed. The data was re-referenced to the mastoids (TP9 and TP10) and submitted to a band-pass filter of 0.1–100 Hz, with a 50 Hz notch filter applied. Ocular correction was performed using independent component analysis. 2.3.2. l-rhythm extraction The l-suppression was extracted from the central electrodes overlaying sensorimotor cortex; electrode positions C3, Cz, and C4. Baseline and action observation epochs were first determined. For the baseline for the l-extraction, the 1 s epoch preceding video onset (when the fixation cross was seen) was used as the baseline for l-extraction. The 8 s epoch for each trial (including 1 s fixation cross before video onset, 1 s resting phase before action onset, and 6 s action epoch) was segmented into 1 s segments, and further analysis was done with these 1 s segments. A Fast Fourier Transform (FFT) was performed separately on each of the 1 s epochs and an average was then taken for each condition, and consequently power values in the lower (8–10 Hz) and upper (10–12 Hz) alpha and beta frequency bands (15–25 Hz) were extracted. To calculate the l-suppression, first the ratios of the action observation condition over the baseline condition epochs were taken using power values of lower and upper alpha and beta bands. Then a natural log transform (ln) was performed on these ratios to produce the l-suppression values (Oberman et al., 2005; Raymaekers et al., 2009). 2.3.3. Statistical analysis Firstly, low and high suggestibility groups were determined according to the scores from the SSS (median split). Table 1 shows means and standard deviations of SSS scores for each group. Statistical analyses for the l-rhythm at lower alpha, upper alpha and beta frequency bands were done separately. For each frequency band, the first step of the analysis of the EEG data used a repeated measures ANOVA with l-rhythm values, with hypnosis (2 levels: on and off), perspective (2 levels: ‘‘self” and ‘‘other”) and electrode (3 levels: C3, Cz and C4) as within subjects factors, and group (2 levels: low and high suggestibility groups) as the between subjects factor. Post-hoc comparisons were used with significant effects from the ANOVA. Following all ANOVAs and post-hoc comparisons, Spearmans correlation analyses were then performed with all participants to investigate associations between the different EEG l-rhythm frequency bands and the measure of empathy with the IRI. Correlation analyses were also performed between behavioral measures to investigate relationships between suggestibility, on the SSS, and empathy traits, on the IRI. 3. Results 3.1. EEG data ANOVA results from the lower alpha band showed main effects of electrode (F(2, 34) = 101.78, p < 0.001, g2p = 0.857) and perspective (F(1, 17) = 8.54, p = 0.009, g2p = 0.334). We saw no main effect of group and no main effect of hypnosis. However, we did see a significant three-way interaction between perspective, electrode and group (F(2, 34) = 3.59, p = 0.038, g2p = 0.174). Post-hoc comparisons revealed no significant group differences between electrodes and perspectives. Although interestingly, when looking at all participants, we see significant differences in the l-suppression in the low alpha band 106 E. Neufeld et al. / Consciousness and Cognition 42 (2016) 101–112 Table 1 Demographics and means and SDs for scores on the Short Suggestibility Scale (SSS) and Interpersonal Reactivity Index (IRI), as well as statistical results from group comparisons of the low and high suggestibility groups. Female/male Age Suggestibility score (SSS) IRI Fantasy IRI Perspective taking IRI Empathetic concern IRI Personal distress * *** Low suggestibility group High suggestibility group Test statistic 6/4 24.0 (4.8) 40.6 (7.1) 12.7 (3.8) 14.0 (3.2) 13.6 (3.0) 9.4 (3.0) 9 female 20.9 (1.4) 54.3 (7.8) 14.7 (2.8) 13.9 (1.5) 16.2 (1.9) 11.8 (3.6) – t(17) = 1.88 t(17) = 4.02*** t(17) = 1.27 t(17) = 0.10 t(17) = 2.27* t(17) = 1.50 p < 0.05. p < 0.001. Fig. 2. Bar chart showing l-rhythm suppression in the low alpha band (mean for all participants) in ‘‘self” vs. ‘‘other”, for each electrode (⁄p < 0.05, ⁄⁄ p < 0.01). Error bars represent one standard error of the mean. between ‘‘self” and ‘‘other” perspectives at electrodes C3 (t(18) = 2.44, p = 0.025) and Cz (t(18) = 3.41, p = 0.003), in which there is greater l-suppression in the other condition, as seen in Fig. 2. ANOVA results for upper alpha band activity shows main effects of electrode (F(2, 34) = 132.932, p<0.001, g2p = 0.887). We see no main effect of group and no main effect of hypnosis. There were also no significant interaction effects. ANOVA results for beta band activity shows main effects of electrode (F(2, 34) = 48.81, p < 0.001, g2p = 0.742) and a main effect of group (F(1, 17) = 11.39, p = 0.004, g2p = 0.401). Furthermore, the two way interactions found between hypnosis and group was approaching significance (F(1, 17) = 3.80, p = 0.068, g2p = 0.183). Post hoc comparisons revealed significant differences between low and high suggestibility groups when on hypnosis at electrodes C3 (t(17) = 2.92, p = 0.010), C4 (t(17) = 4.36, p < 0.001) and Cz (t(17) = 2.90, p = 0.010). No significant group differences were found for beta band activity when participants were off hypnosis (all p > 0.05). Bar charts displaying l-rhythm values are shown in Fig. 3. As we can see from the figure, under hypnosis, participants with high hypnotic suggestibility had significantly greater l-suppression in the beta band than those with a low hypnotic suggestibility. It is worthwhile noting that when comparing the l-rhythm suppression on and off hypnosis within groups, with a paired samples t-test, we see a trend level difference in the high suggestibility group at electrodes C3 and C4 (C3 (t(8) = 2.04, p = 0.076), C4 (t(8) = 2.03, p = 0.077)), whereas the low suggestibility group showed clearly no difference in <mu>-rhythm suppression between on and off hypnosis (p > 0.05). E. Neufeld et al. / Consciousness and Cognition 42 (2016) 101–112 107 Fig. 3. Bar charts showing l-rhythm suppression in the beta band, comparing low and high suggestibility participants, on and off hypnosis. (⁄⁄p = <0.01, ⁄⁄⁄ p < 0.001). Error bars represent one standard error of the mean. 3.2. Behavioral measures Correlations between the overall mean l-suppression in the lower alpha band activity (pooled across conditions) and the behavioral measures revealed significant associations with IRI fantasy scale (r = 0.625, p = 0.006), IRI perspective taking (r = 0.552, p = 0.018) and IRI empathetic concern (r = 0.621, p = 0.006). Scatter plots of these relationships are shown in Fig. 4. No significant correlations were found between l-suppression in the upper alpha or the beta frequency bands and behavioral measures of empathy on the IRI. The behavioral measures also showed associations with each other. When looking at all participants (i.e. including low and high suggestibility participants), we see significant correlations between hypnotic suggestibility and IRI fantasy (r = 0.512, p = 0.030), and between hypnotic suggestibility and IRI empathetic concern (r = 0.733, p = 0.001). Scatterplots of these relationships are shown in Fig. 5. In terms of relationships between factors of the IRI, we only saw significant correlations between IRI fantasy and empathetic concern (r = 0.572, p = 0.013). 4. Discussion In the present study, we examined how attentional bottom-up variation by means of hypnosis affects the l-suppression of healthy participants when watching intentional action scenes. We predicted that the l-rhythm suppression is greater in the hypnosis condition for highly suggestible individuals. In support of our prediction, under hypnosis the l-suppression of the high suggestibility group was significantly higher than the l-wave suppression of the low suggestibility group. This difference is due to an increase in l-wave suppression within the high suggestibility group: whereas the difference of lsuppression for the high suggestibility group between on and off hypnosis was approaching significance, there was no difference for the low suggestibility group between the on and off hypnosis condition. Our findings support the Social Relevance Hypothesis (Chevallier et al., 2012; Oberman et al., 2008), according to which social processing depends on social stimuli being processed in an automatic, bottom-up driven manner. Our results verify a central prediction of this view: if social cognition processes are highly dependent on attentional factors such as relevance, saliency, and automaticity, then, since those factors are the main determinants of bottom-up attentional processes, it should be possible to boost processing of social stimuli by boosting their attentional bottom-up valence (Eastwood et al., 2001; Egeth & Yantis, 1997; Geng & Mangun, 2009; Parkhurst et al., 2002; Peck et al., 2009; Pratto & John, 1991; Summerfield & Egner, 2009; Treue, 2003; Wang et al., 1994). Accordingly, when we increased bottom-up processing of social stimuli, the neural signal associated with social action processing increased. Our finding, in conjunction with findings of studies with ASD patients, provides important empirical building blocks for a theory of social action processing. Whereas past studies already showed that ASD patients’ neural responses and behavioral outcomes can be normalized as a function of such attentional factors as relevance, saliency and automaticity, our study adds that under normal conditions, intentional action 108 E. Neufeld et al. / Consciousness and Cognition 42 (2016) 101–112 Fig. 4. Scatterplots showing relationship between l-suppression in the lower alpha band (8–10 Hz) and the (a) fantasy (R2 = 0.391), (b) perspective taking (R2 = 0.304) and (c) empathetic concern (R2 = 0.385) factors of the Interpersonal Reactivity Index (IRI; Davis, 1980), a measure of empathy. Fig. 5. Scatterplots showing relationship between hypnotic suggestibility and the (a) fantasy (R2 = 0.262), and (b) empathetic concern (R2 = 0.538) factors of the Interpersonal Reactivity Index (IRI; Davis, 1980), a measure of empathy. processing is predominantly automatic and bottom-up driven. Taken together, the studies show that adequate processing of intentional action stimuli depends on bottom-up and automatic processes. Oberman et al. (2005) showed that if these are disabled, as in the case of ASD patients, the neural response associated with intentional action processing diminishes, and we E. Neufeld et al. / Consciousness and Cognition 42 (2016) 101–112 109 showed that making bottom-up processes prevalent in healthy participants is followed by an increase of the neural signal linked to social action processing. As expected, the influence of attentional variation on social stimuli processing was only observable in the high suggestibility group. The dissociation between high and low suggestibility groups is standard procedure within hypnosis research. Experimental evidence manifests inter-individual variations between high and low suggestible subjects on various levels: behaviorally (Braffman & Kirsch, 1999; Oakley & Halligan, 2013), neurally (Dienes & Hutton, 2013; Hoeft et al., 2012; Horton, Crawford, Harrington, & Downs, 2004; Kihlstrom, 2013; Naish, 2010), and genetically (Lichtenberg, BachnerMelman, Gritsenko, & Ebstein, 2000; Raz, 2005). Correspondingly, high/low suggestibility membership leads to different predictions about task outcomes. In our case, previous findings showed that only highly suggestible participants are more prone to automatic bottom-up processes when under hypnosis. Accordingly, our hypothesis predicted that only highly suggestible participants display higher l-suppression due to increased attentional bottom-up processing of social stimuli. Our data also revealed higher l-suppression in the lower alpha band in the third as compared to the first-person condition when averaged over all participants and both hypnosis conditions (‘‘on” and ‘‘off” hypnosis). This is in tension with many other studies showing a higher l-suppression for ‘‘self” conditions (Hoenen, Schain, & Pause, 2013; Oberman et al., 2005; Woodruff & Maaske, 2010; Woodruff, Martin, & Bilyk, 2011), and with theories predicting facilitated processing of stimuli that depict ourselves (Jeannerod & Anquetil, 2008). Our theoretical framework of social relevance, in contrast, predicts this outcome. In a natural environment, visually attending to social stimuli depicting other persons is arguably more relevant for fitness benefits than visual attendance to stimuli depicting actions of ourselves. Stimuli of other persons might be important perceptual signals indexing potential rewards or harms. In addition, somatosensory feed-forward loops provide us with feedback of our own body postures (David, Newen, & Vogeley, 2008; Tsakiris & Haggard, 2005; Tsakiris, Longo, & Haggard, 2010), so visual attention devoted to them would be superfluous. In the case of first-person agency (i.e., registering that I am the agent of an action) we are provided with a feeling of agency and a judgment of agency (Synofzik, Vosgerau, & Newen, 2008), while there is no need for an additional perception of one’s own agency. In the case of registering the other person’s intentional action, we have a bottom-up attention-moduled perception of third-person intentional actions (we may add a perception-based judgment, but this was not focus of this investigation). Assuming that attentional bottom-up factors such as relevance increase the amount of social action processing devoted to stimuli, l-suppression should then be higher in the third than in the first-person condition. It is also important to note that studies that found higher l-suppression for the ‘‘self” perspective contrasted a first-person motor execution condition with a third-person observation condition (Hoenen et al., 2013; Oberman et al., 2005; Woodruff & Maaske, 2010; Woodruff et al., 2011), whereas our study compared the perspective difference in two purely perceptual conditions. However, we cannot rule out that the higher l-suppression in the third-person perspective was caused by a higher amount of visual information given in this condition. In line with this interpretation, Oberman et al. (2008) found higher (although not significant) l-suppression in the observational ‘‘self” condition when the amount of visual information was identical. Future research has to settle this debate. We also examined the relationship between empathy measures and suggestibility. In line with other findings (Cardeña, Terhune, Lööf, & Buratti, 2009; Wickramasekera & Szlyk, 2003), we found a positive relationship between the empathy markers fantasy and empathetic concern and suggestibility values. So suggestibility correlates both with higher values in certain empathy traits and with l-suppression when under hypnosis. Interestingly, Wickramasekera and Szlyk (2003) additionally found significant correlations between empathy and hypnotic absorption. This opens questions for theories about how hypnosis happens: speculatively, empathetic personality traits could be important factors for the formation of hypnotic experience. This interpretation seems particularly reasonable when considering the social basis of the hypnotic induction phase. The fantasy scale on the IRI measures ‘‘the tendency to identify with characters in movies, novels, plays and other fictional situations”. This ability is generally required when the hypnotist induces hypnosis. Participants are usually given instructions to put themselves imaginatively into novel situations to facilitate relaxation and exercise attentional absorption (e.g., ‘‘your feet have grown into the floor”, ‘‘you are walking down a long stairway”). Also the empathetic concern measure is defined as ‘‘a tendency for the respondent to experience feelings of warmth, compassion and concern for others undergoing negative experiences”; in short, the tendency to identify with and share regard for other persons. A possible interpretation of our finding could thus be that the ability to identify and empathetically resonate with other persons also facilitates the establishment of the social connection to the hypnotist, which in turn makes the adaptation of the hypnotist’s instructions easier. Interestingly, however, our results also displayed an inverse relationship between several measures of empathy (fantasy score, perspective taking, and empathetic concern) and the l-rhythm suppression in the lower alpha band: participants with greater l-rhythm suppression in the lower alpha band exhibited less empathy traits. While also other studies report an inverse relationship between empathy markers and l-suppression (Horan, Pineda, Wynn, Iacoboni, & Green, 2014; Perry, Troje, & Bentin, 2010; cf. also Woodruff et al., 2011), this finding seems to be in tension with our other results. Suggestibility values correlated both with higher l-suppression and stronger empathy traits, but l-suppression simultaneously correlates negatively with the mentioned empathy traits. How can we best account for this observation? The difference in correlation patterns might be explained by the difference of l-wave frequency bands. Whereas suggestibility values correlate with lsuppression measured in the beta frequency band, the empathy traits correlate negatively with l-suppression in the lower alpha band. Although studies generally do not functionally distinguish between different parts of the alpha and beta bands of the l-wave (Hari & Kujala, 2009; Hari et al., 2014; Ritter, Moosmann, & Villringer, 2009), more recent studies point at a possible functional, temporal and spatial dissociation between lower and higher alpha as well as beta elements (Dumas et al., 110 E. Neufeld et al. / Consciousness and Cognition 42 (2016) 101–112 2014; Sebastiani et al., 2014). In our study, the fact that suggestibility correlates with l-suppression in the beta band and empathy traits, while empathy traits correlate negatively with l-suppression in the lower alpha band, also speaks in favor of an underlying functional dissociation. A task for future studies is to address the functional dissociations underlying different components of the l-rhythm. Our study provides another example of how hypnosis can be used to study cognition. However, why is it the case that highly suggestible individuals are more prone to bottom-up, automatic processes when under neutral hypnotic trance? Egner and Raz’s (2007) ‘‘impaired attention model” provides an answer to that question. Building on the ‘‘dissociated control theory” of hypnosis by Woody and Bowers (1994), they propose that hypnosis ‘dissociates’ attentional control from automatic processes. This is why the internal generation of thoughts and task strategies is impaired. Under neutral hypnosis, bottom-up processes become predominant – there is no interference from higher-order thoughts. Accordingly, Jamieson and Sheehan (2004) found that when asked for their task strategies, participants under hypnosis report more often to ‘‘just letting things happen”. In addition, Egner et al. (2005) found functional neuroimaging evidence that frontal conflict monitoring processes are decoupled from cognitive control mechanisms of the brain when under neutral hypnosis. Dissociation of automatized perceptual processes from higher-order, top-down thoughts is a plausible candidate to explain the increased focus on inherently salient stimuli. Our results also hint at other directions of research. Chevallier et al. (2012) suggested that abnormalities in ASD patients when processing social stimuli do not necessarily result from impairments of their social competence, but are a consequence of attentional bottom-up deviations. Our study shows that the method of hypnosis can be successfully used in healthy participants to change attentional bottom-up processing of social stimuli. An interesting question for future research is whether hypnosis could be used as a therapeutic means in socio-pathological patient groups such as ASD to normalize attentional processing of social stimuli and subsequently influence social task outcomes positively. To conclude, our study shows that attentional bottom-up variation can have a high impact on the neural responses linked to social action observation. This teaches us a valuable lesson for research of higher-capacities such as mindreading. When trying to model the cognitive operations involved in such psychological tasks, it seems to make good sense to focus on basic cognitive phenomena – in our case, specific kinds of attentional processes – to find the cognitive elements required for the task’s computation. We found evidence for the view that performance in social action observation tasks strongly depends on bottom-up attentional competences. This suggests that we should take seriously the possibility that other capacities of social cognition also substantially depend on bottom-up processes. Acknowledgements We gratefully acknowledge the assistance of Aaron who hypnotized the study participants and performed the hypnotic suggestion. The article also benefited enormously from constructive comments of two anonymous referees. We are also grateful to Guillermo Del Pinal for insightful discussions on various drafts of the paper. We would also like to thank participants of the conference ‘‘The Nature and Origins of Human Cognition” at the Berlin School of Mind and Brain, who provided valuable feedback on a poster presentation on this topic. References Bernier, R., Dawson, G., Webb, S., & Murias, M. (2007). EEG mu rhythm and imitation impairments in individuals with autism spectrum disorder. Brain and Cognition, 64(3), 228–237. Braffman, W., & Kirsch, I. (1999). Imaginative suggestibility and hypnotizability: An empirical analysis. Journal of Personality and Social Psychology, 77(3), 578–587. http://dx.doi.org/10.1037/0022-3514.77.3.578. Brown, E. C., Wiersema, J. R., Pourtois, G., & Brüne, M. (2013). Modulation of motor cortex activity when observing rewarding and punishing actions. Neuropsychologia, 51(1), 52–58. Cardeña, E., Jönsson, P., Terhune, D. B., & Marcusson-Clavertz, D. (2013). The neurophenomenology of neutral hypnosis. Cortex; A Journal Devoted to the Study of the Nervous System and Behavior, 49(2), 375–385. http://dx.doi.org/10.1016/j.cortex.2012.04.001. Cardeña, E., Terhune, D. B., Lööf, A., & Buratti, S. (2009). Hypnotic experience is related to emotional contagion. The International Journal of Clinical and Experimental Hypnosis, 57(1), 33–46. http://dx.doi.org/10.1080/00207140802463500. Chevallier, C., Kohls, G., Troiani, V., Brodkin, E. S., & Schultz, R. T. (2012). The social motivation theory of autism. Trends in Cognitive Sciences, 16(4), 231–239. http://dx.doi.org/10.1016/j.tics.2012.02.007. Cook, J., Barbalat, G., & Blakemore, S.-J. (2012). Top-down modulation of the perception of other people in schizophrenia and autism. Frontiers in Human Neuroscience, 6, 175. http://dx.doi.org/10.3389/fnhum.2012.00175. Crawford, H. J. (1994). Brain dynamics and hypnosis: Attentional and disattentional processes. International Journal of Clinical and Experimental Hypnosis, 42 (3), 204–232. Crawford, H. J., & Gruzelier, J. H. (1992). A midstream view of the neuropsychophysiology of hypnosis: Recent research and future directions. In E. Fromm & M. R. Nash (Eds.), Contemporary hypnosis research (pp. 227–266). New York: Guilford. David, N., Newen, A., & Vogeley, K. (2008). The ‘‘sense of agency” and its underlying cognitive and neural mechanisms. Consciousness and Cognition, 17(2), 523–534. http://dx.doi.org/10.1016/j.concog.2008.03.004. Davis, M. (1983). Measuring individual differences in empathy: Evidence for a multidimensional approach. J Pers Soc Psychol, 44(1), 113–126. http://dx.doi. org/10.1037//0022-3514.44.1.113. Delorme, A., & Makeig, S. (2004). EEGLAB: An open source toolbox for analysis of single-trial EEG dynamics including independent component analysis. Journal of Neuroscience Methods, 134(1), 9–21. http://dx.doi.org/10.1016/j.jneumeth.2003.10.009. Dienes, Z., & Hutton, S. (2013). Understanding hypnosis metacognitively: rTMS applied to left DLPFC increases hypnotic suggestibility. Cortex, 49(2), 386–392. Dumas, G., Soussignan, R., Hugueville, L., Martinerie, J., & Nadel, J. (2014). Revisiting mu suppression in autism spectrum disorder. Brain Research, 1585, 108–119. http://dx.doi.org/10.1016/j.brainres.2014.08.035. E. Neufeld et al. / Consciousness and Cognition 42 (2016) 101–112 111 Eastwood, J. D., Smilek, D., & Merikle, P. M. (2001). Differential attentional guidance by unattended faces expressing positive and negative emotion. Perception & Psychophysics, 63(6), 1004–1013. http://dx.doi.org/10.3758/BF03194519. Egeth, H. E., & Yantis, S. (1997). Visual attention: Control, representation, and time course. Annual Review of Psychology, 48, 269–297. http://dx.doi.org/ 10.1146/annurev.psych.48.1.269. Egner, T., & Hirsch, J. (2005). Cognitive control mechanisms resolve conflict through cortical amplification of task-relevant information. Nature Neuroscience, 8(12), 1784–1790. Egner, T., Jamieson, G., & Gruzelier, J. (2005). Hypnosis decouples cognitive control from conflict monitoring processes of the frontal lobe. Neuroimage, 27(4), 969–978. Egner, T., & Raz, A. (2007). Cognitive control processes and hypnosis. Hypnosis and Conscious States: The Cognitive Neuroscience Perspective, 29–50. Farvolden, P., & Woody, E. Z. (2004). Hypnosis, memory, and frontal executive functioning. International Journal of Clinical and Experimental Hypnosis, 52(1), 3–26. Firestone, C., & Scholl, B. J. (2015). Cognition does not affect perception: Evaluating the evidence for ’top-down’ effects. The Behavioral and Brain Sciences, 1– 77, 1–77. http://dx.doi.org/10.1017/S0140525X15000965. Fletcher-Watson, S., Findlay, J. M., Leekam, S. R., & Benson, V. (2008). Rapid detection of person information in a naturalistic scene. Perception, 37(4), 571–583. http://dx.doi.org/10.1068/p5705. Geng, J. J., & Mangun, G. R. (2009). Anterior intraparietal sulcus is sensitive to bottom-up attention driven by stimulus salience. Journal of Cognitive Neuroscience, 21(8), 1584–1601. http://dx.doi.org/10.1162/jocn.2009.21103. Gliga, T., Elsabbagh, M., Andravizou, A., & Johnson, M. (2009). Faces attract infants’ attention in complex displays. Infancy, 14(5), 550–562. http://dx.doi.org/ 10.1080/15250000903144199. Gruzelier, J. (1998). A working model of the neurophysiology of hypnosis: A review of evidence. Contemporary Hypnosis, 15(1), 3–21. http://dx.doi.org/ 10.1002/ch.112. Gruzelier, J. H. (2000). Redefining hypnosis: Theory, methods and integration. Contemporary Hypnosis, 17(2), 51–70. http://dx.doi.org/10.1002/ch.193. Gruzelier, J., & Warren, K. (1993). Neuropsychological evidence of reductions on left frontal tests with hypnosis. Psychological Medicine, 23(01), 93. http://dx. doi.org/10.1017/S0033291700038885. Hari, R. (2006). Action–perception connection and the cortical mu rhythm. In Progress in brain research. Event-related dynamics of brain oscillations (Vol. 159, pp. 253–260). Elsevier. Hari, R., Bourguignon, M., Piitulainen, H., Smeds, E., de Tiège, X., & Jousmäki, V. (2014). Human primary motor cortex is both activated and stabilized during observation of other person’s phasic motor actions. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences, 369(1644), 20130171. http://dx.doi.org/10.1098/rstb.2013.0171. Hari, R., & Kujala, M. V. (2009). Brain basis of human social interaction: From concepts to brain imaging. Physiological Reviews, 89(2), 453–479. http://dx.doi. org/10.1152/physrev.00041.2007. Hilgard, E. R. (1977). Divided consciousness.New York: John Wiley. Hoeft, F., Gabrieli, J. D. E., Whitfield-Gabrieli, S., Haas, B. W., Bammer, R., Menon, V., & Spiegel, D. (2012). Functional brain basis of hypnotizability. Archives of General Psychiatry, 69(10), 1064–1072. Hoenen, M., Schain, C., & Pause, B. M. (2013). Down-modulation of mu-activity through empathic top-down processes. Social Neuroscience, 8(5), 515–524. http://dx.doi.org/10.1080/17470919.2013.833550. Horan, W. P., Pineda, J. A., Wynn, J. K., Iacoboni, M., & Green, M. F. (2014). Some markers of mirroring appear intact in schizophrenia: Evidence from mu suppression. Cognitive, Affective & Behavioral Neuroscience, 14(3), 1049–1060. http://dx.doi.org/10.3758/s13415-013-0245-8. Horton, J. E., Crawford, H. J., Harrington, G., & Downs, J. H. (2004). Increased anterior corpus callosum size associated positively with hypnotizability and the ability to control pain. Brain, 127(8), 1741–1747. Iani, C., Ricci, F., Baroni, G., & Rubichi, S. (2009). Attention control and susceptibility to hypnosis. Consciousness and Cognition, 18(4), 856–863. Jamieson, G. A., & Sheehan, P. W. (2004). An empirical test of Woody and Bowers’s dissociated-control theory of hypnosis. International Journal of Clinical and Experimental Hypnosis, 52(3), 232–249. Jamieson, G. A., & Woody, E. (2007). Dissociated control as a paradigm for cognitive neuroscience research and theorizing in hypnosis. Hypnosis and Conscious States: The Cognitive Neuroscience Perspective, 111–131. Jeannerod, M., & Anquetil, T. (2008). Putting oneself in the perspective of the other: A framework for self-other differentiation. Social Neuroscience, 3(3–4), 356–367. http://dx.doi.org/10.1080/17470910701563715. Jiang, Y., Costello, P., & He, S. (2007). Processing of invisible stimuli: Advantage of upright faces and recognizable words in overcoming interocular suppression. Psychological Science, 18(4), 349–355. http://dx.doi.org/10.1111/j.1467-9280.2007.01902.x. Kaiser, J., Barker, R., Haenschel, C., Baldeweg, T., & Gruzelier, J. H. (1997). Hypnosis and event-related potential correlates of error processing in a stroop-type paradigm: A test of the frontal hypothesis. International Journal of Psychophysiology, 27(3), 215–222. http://dx.doi.org/10.1016/S0167-8760(97)00055-X. Kallio, S., Revonsuo, A., Hämäläinen, H., Markela, J., & Gruzelier, J. (2001). Anterior brain functions and hypnosis: A test of the frontal hypothesis. International Journal of Clinical and Experimental Hypnosis, 49(2), 95–108. Kihlstrom, J. F. (2013). Neuro-hypnotism: Prospects for hypnosis and neuroscience. Cortex, 49(2), 365–374. Kilner, J. M., & Lemon, R. N. (2013). What we know currently about mirror neurons. Current Biology, 23(23), R1057–R1062. Kilner, J. M., Marchant, J. L., & Frith, C. D. (2006). Modulation of the mirror system by social relevance. Social Cognitive and Affective Neuroscience, 1(2), 143–148. Kotov, R., Bellman, S., & Watson, D. (2007). Multidimensional Iowa suggestibility scale: Brief manual, Available from: http://www.stonybrookmedicalcenter. org/psychiatry/kotov_r. Landry, M., Appourchaux, K., & Raz, A. (2014). Elucidating unconscious processing with instrumental hypnosis. Frontiers in Psychology, 5, 785. http://dx.doi. org/10.3389/fpsyg.2014.00785. Lichtenberg, P., Bachner-Melman, R., Gritsenko, I., & Ebstein, R. P. (2000). Exploratory association study between catechol-O-methyltransferase (COMT) high/low enzyme activity polymorphism and hypnotizability. American Journal of Medical Genetics, 96(6), 771–774. Lifshitz, M., Aubert Bonn, N., Fischer, A., Kashem, I. F., & Raz, A. (2013). Using suggestion to modulate automatic processes: From Stroop to McGurk and beyond. Cortex, 49(2), 463–473. Margraf, J. (1994). Diagnostisches Kurz-Interview bei psychischen Störungen: Mini-DIPS. Handbuch: mit 15 Tabellen. Springer. Naish, Peter L. N. (2010). Hypnosis and hemispheric asymmetry. Consciousness and Cognition, 19(1), 230–234. http://dx.doi.org/10.1016/ j.concog.2009.10.003. Nordby, H., Hugdahl, K., Jasiukaitis, P., & Spiegel, D. (1999). Effects of hypnotizability on performance of a Stroop task and event-related potentials. Perceptual and Motor Skills, 88(3 Pt 1), 819–830. http://dx.doi.org/10.2466/pms.1999.88.3.819. Oakley, D. A., & Halligan, P. W. (2013). Hypnotic suggestion: Opportunities for cognitive neuroscience. Nature Reviews. Neuroscience, 14(8), 565–576. http:// dx.doi.org/10.1038/nrn3538. Oberman, L. M., Hubbard, E. M., McCleery, J. P., Altschuler, E. L., Ramachandran, V. S., & Pineda, J. A. (2005). EEG evidence for mirror neuron dysfunction in autism spectrum disorders. Cognitive Brain Research, 24(2), 190–198. Oberman, L. M., Ramachandran, V. S., & Pineda, J. A. (2008). Modulation of mu suppression in children with autism spectrum disorders in response to familiar or unfamiliar stimuli: The mirror neuron hypothesis. Neuropsychologia, 46(5), 1558–1565. Parkhurst, D., Law, K., & Niebur, E. (2002). Modeling the role of salience in the allocation of overt visual attention. Vision Research, 42(1), 107–123. http://dx. doi.org/10.1016/S0042-6989(01)00250-4. 112 E. Neufeld et al. / Consciousness and Cognition 42 (2016) 101–112 Paulus, C. (2009). Der Saarbrücker Persönlichkeitsfragebogen SPF (IRI) zur Messung von Empathie: Psychometrische Evaluation der deutschen Version des Interpersonal Reactivity Index. Peck, C. J., Jangraw, D. C., Suzuki, M., Efem, R., & Gottlieb, J. (2009). Reward modulates attention independently of action value in posterior parietal cortex. The Journal of Neuroscience: The Official Journal of the Society for Neuroscience, 29(36), 11182–11191. http://dx.doi.org/10.1523/JNEUROSCI.1929-09.2009. Perry, A., & Bentin, S. (2009). Mirror activity in the human brain while observing hand movements: A comparison between EEG desynchronization in the lrange and previous fMRI results. Brain Research, 1282, 126–132. Perry, A., Stein, L., & Bentin, S. (2011). Motor and attentional mechanisms involved in social interaction–evidence from mu and alpha EEG suppression. NeuroImage, 58(3), 895–904. http://dx.doi.org/10.1016/j.neuroimage.2011.06.060. Perry, A., Troje, N. F., & Bentin, S. (2010). Exploring motor system contributions to the perception of social information: Evidence from EEG activity in the mu/alpha frequency range. Social Neuroscience, 5(3), 272–284. http://dx.doi.org/10.1080/17470910903395767. Pineda, J. A. (2005). The functional significance of mu rhythms: Translating ‘‘seeing” and ‘‘hearing” into ‘‘doing”. Brain Research Reviews, 50(1), 57–68. Pratto, F., & John, O. P. (1991). Automatic vigilance: The attention-grabbing power of negative social information. Journal of Personality and Social Psychology, 61(3), 380–391. http://dx.doi.org/10.1037/0022-3514.61.3.380. Raymaekers, R., Wiersema, J. R., & Roeyers, H. (2009). EEG study of the mirror neuron system in children with high functioning autism. Brain Research, 1304, 113–121. Raz, A. (2005). Attention and hypnosis: Neural substrates and genetic associations of two converging processes. International Journal of Clinical and Experimental Hypnosis, 53(3), 237–258. Raz, A. (2011). Hypnosis: A twilight zone of the top-down variety Few have never heard of hypnosis but most know little about the potential of this mindbody regulation technique for advancing science. Trends in Cognitive Sciences, 15(12), 555–557. http://dx.doi.org/10.1016/j.tics.2011.10.002. Raz, A., Fan, J., & Posner, M. I. (2005). Hypnotic suggestion reduces conflict in the human brain. Proceedings of the National Academy of Sciences of the United States of America, 102(28), 9978–9983. Raz, A., Fan, J., & Posner, M. I. (2006). Neuroimaging and genetic associations of attentional and hypnotic processes. Journal of Physiology-Paris, 99(4), 483–491. Raz, A., Shapiro, T., Fan, J., & Posner, M. I. (2002). Hypnotic suggestion and the modulation of Stroop interference. Archives of General Psychiatry, 59(12), 1155–1161. Ritter, P., Moosmann, M., & Villringer, A. (2009). Rolandic alpha and beta EEG rhythms’ strengths are inversely related to fMRI-BOLD signal in primary somatosensory and motor cortex. Human Brain Mapping, 30(4), 1168–1187. Sarter, M., Givens, B., & Bruno, J. P. (2001). The cognitive neuroscience of sustained attention: Where top-down meets bottom-up. Brain Research Reviews, 35 (2), 146–160. http://dx.doi.org/10.1016/S0165-0173(01)00044-3. Sebastiani, V., de Pasquale, F., Costantini, M., Mantini, D., Pizzella, V., Romani, G. L., & Della Penna, S. (2014). Being an agent or an observer: Different spectral dynamics revealed by MEG. NeuroImage, 102P2, 717–728. http://dx.doi.org/10.1016/j.neuroimage.2014.08.031. Senju, A., & Johnson, M. H. (2009). The eye contact effect: Mechanisms and development. Trends in Cognitive Sciences, 13(3), 127–134. http://dx.doi.org/ 10.1016/j.tics.2008.11.009. Sheehan, P. W., Donovan, P., & MacLeod, C. M. (1988). Strategy manipulation and the Stroop effect in hypnosis. Journal of Abnormal Psychology, 97(4), 455. Shiffrin, R. M., & Schneider, W. (1977). Controlled and automatic human information processing: II. Perceptual learning, automatic attending and a general theory. Psychological Review, 84(2), 127–190. http://dx.doi.org/10.1037/0033-295X.84.2.127. Stroop, J. R. (1935). Studies of interference in serial verbal reactions. Journal of Experimental Psychology, 18(6), 643. Summerfield, C., & Egner, T. (2009). Expectation (and attention) in visual cognition. Trends in Cognitive Sciences, 13(9), 403–409. http://dx.doi.org/10.1016/j. tics.2009.06.003. Synofzik, M., Vosgerau, G., & Newen, A. (2008). Beyond the comparator model: A multifactorial two-step account of agency. Consciousness and Cognition, 17 (1), 219–239. Treue, S. (2003). Visual attention: The where, what, how and why of saliency. Current Opinion in Neurobiology, 13(4), 428–432. http://dx.doi.org/10.1016/ S0959-4388(03)00105-3. Trilla Gros, I., Panasiti, M. S., & Chakrabarti, B. (2015). The plasticity of the mirror system: How reward learning modulates cortical motor simulation of others. Neuropsychologia, 70, 255–262. http://dx.doi.org/10.1016/j.neuropsychologia.2015.02.033. Tsakiris, M., & Haggard, P. (2005). The rubber hand illusion revisited: Visuotactile integration and self-attribution. Journal of Experimental Psychology: Human Perception and Performance, 31(1), 80. Tsakiris, M., Longo, M. R., & Haggard, P. (2010). Having a body versus moving your body: Neural signatures of agency and body-ownership. Neuropsychologia, 48(9), 2740–2749. Vetter, P., & Newen, A. (2014). Varieties of cognitive penetration in visual perception. Consciousness and Cognition, 27, 62–75. Wang, Q., Cavanagh, P., & Green, M. (1994). Familiarity and pop-out in visual search. Perception & Psychophysics, 56(5), 495–500. http://dx.doi.org/10.3758/ BF03206946. Wickramasekera, I. E., & Szlyk, J. P. (2003). Could empathy be a predictor of hypnotic ability? The International Journal of Clinical and Experimental Hypnosis, 51(4), 390–399. http://dx.doi.org/10.1076/iceh.51.4.390.16413. Woodruff, C. C., & Maaske, S. (2010). Action execution engages human mirror neuron system more than action observation. NeuroReport, 21(6), 432–435. http://dx.doi.org/10.1097/WNR.0b013e3283385910. Woodruff, C. C., Martin, T., & Bilyk, N. (2011). Differences in self- and other-induced Mu suppression are correlated with empathic abilities. Brain Research, 1405, 69–76. http://dx.doi.org/10.1016/j.brainres.2011.05.046. Woody, E. Z., & Bowers, K. S. (1994). A frontal assault on dissociated control. In S. J. Lynn & J. W. Rhue (Eds.), Dissociation: Clinical and theoretical perspectives. New York: Guilford Press.

Consciousness and Cognition 19 (2010) 687–689 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Editorial Self, other and memory: A preface q The key topics of this special issue are self-consciousness, understanding others and the role of memory, while the latter is especially investigated with respect to its relation to self-consciousness. It is furthermore a characteristic feature of this special issue that most of the papers are dedicated to an interdisciplinary research program combining the perspectives of philosophy, psychology, neurobiology and cognitive neuroscience. The issue of self-consciousness is firstly investigated concerning the feature of perspective taking (Kockler et al.). This empirical study is complemented by two studies that focus on embodiment: A basic constituent of bodily self-consciousness is surely the anchoring of self-related experiences in one’s own body image (Blanke et al.) and an interesting case is being made of language as being related to embodiment as well (Binkofski et al.). Then three articles are dealing with self-deception including some interconnected discussion (Greve & Wentura; Michel & Newen, Mele). Furthermore, a theoretical investigation of self-consciousness refers to the compositionality of our thoughts (Werning). Investigating the self is contrasted by an empirical study on the perception of others under specific influence of different cultural backgrounds (Bente et al.). With the article of Staniloiu et al. the volume widens the discussion to account for memory as well. Three empirical investigations help to improve our understanding of the role of memory while one theoretical account discusses a core thesis about memory: One important focus is the question what is necessary to establish individual self-consciousness over time (Staniloui et al.): this leads to a discussing of diachronic self-consciousness and autobiographic episodic memory. The development of episodic memory is the focus of the behavioral study of Perner et al. arguing that the ability of mental rotation of figures is initiating a switch from knowledge retrieval to episodic remembering. In addition to the distinction of episodic and semantic memory, Sauvage discusses convincing evidence to support the distinction between familiarity and recollection. The paper of Volz et al. is investigating our subjective experience of intuition as addressed by a fluency heuristic paradigm from the perspective of cognitive neuroscience. Finally, in a philosophical paper Vosgerau argues that, on the basis of a functionalist approach of mental representation, representations only have a specific content if they are actually in use, therefore, stored mental representations do not have a specific content. We shortly characterize the main claims of the papers: One feature of self-consciousness is the capacity to take different perspectives: closely related to an intact body image is our capacity to change perspectives in space. This already ‘‘classical” empirical approach to study spatial perspective taking has been investigated by Kockler et al. for the first time in dynamic environments. In their study ‘‘Visuospatial perspective taking in a dynamic environment” participants were asked to perform left–right-decisions in animated and static virtual environments, both from a first- and third-person-perspective. The results showed a significantly increased activation in the right posterior intraparietal sulcus in conditions in which a dynamic stimulus had to be observed and judged upon from a first-person-perspective. This special case of oneself being involved in object-directed action preparation is interpreted and discussed as the neural mechanism of ‘‘readiness for (re)action”. A very important approach to any cognitive function are disturbances of self-consciousness, this important and sometimes underestimated approach is employed by Blanke et al. in their contribution concerning bodily self-consciousness. The basic key assumption here is that the integration of multisensory bodily signals from the entire body is crucial for bodily self-consciousness. Not only empirically grounded, but also philosophically arguable they refer to a newly developed taxonomy that comprises the three different features of self-location, first person perspective, and self-identification and that allows to categorize body schema disturbances. The authors are focusing on illusory body perceptions of one’s own body as a particularly interesting disturbance of the body schema or body representation. Based on detailed clinical and neuroanatomical data the paper presents two interesting case reports with a disturbance of self-identification and shows their relevance for the neural mechanisms underlying bodily self-consciousness. q This article is part of a special issue of this journal on Self, Other and Memory. 1053-8100/$ - see front matter Ó 2010 Published by Elsevier Ltd. doi:10.1016/j.concog.2010.07.002 688 Editorial / Consciousness and Cognition 19 (2010) 687–689 On the background of such empirical evidence, the concept of embodied cognition has become one the leading frameworks in the study of cognitive processes in general. A very interesting and stimulating approach refers to the hypothesis that also language can be interpreted as embodied cognition. In the study ‘‘Grasping language – a short story on Embodiment.” the authors Binkofski et al. employ a new meta-analytic tool that allows to study neural mechanisms of cognitive functions across different studies. The key finding of cumulative evidence for the recruitment of the somatotopic activation of motor areas clearly supports this hypothesis of language as embodied cognitive faculty and emphasizes the role of sensorimotor areas during language processing. The discussion of self and embodiment is followed by a group of three papers on self-deception. There is few disagreement in the recent philosophical and psychological literature on self-deception about the fact that garden-variety selfdeception differs from stereotypical other-deception. What we call self-deception in everyday contexts can be explained within a framework of a unified self and does not require any partitioning of the self into different sub-centers of agency in order to enable mutual deception. But there is an intense debate about the role of motivational processes in self-deception which is the main focus of the contributions in this volume. Greve & Wentura assess two explanatory virtues that cognitive processes of self-immunization have with regard to selfdeception. The authors approach self-deception from the perspective of cognitive psychology. They draw on empirical evidence for unconscious processes of self-immunization they have provided in a series of studies. Via semantic priming tasks they have tracked automatic and sub-personal processes of self-immunization. On the basis of this evidence Greve & Wentura demonstrate how we balance stability and changes in our self-images in a self-serving way when the pleasure- and the reality-principle have clashed and need to be reconciled. As a sub-personal process, the authors show that self-immunization does not fall victim to the paradoxes of self-deception that arise on a personal level. Michel & Newen show in detail to which degree self-deception remains rational by providing a general model of self-deception that describes self-deception in a proper sense as pseudo-rational regulation of belief-systems. They distinguish self-deception from alternative forms of motivational influence on acceptance (in opposition to the position of Mele) which have been proposed in the recent philosophical literature such as motivational bias and pretense. Self-deceivers violate their own standards of belief-regulation by employing a dual standard of reasoning that remains opaque to them. Drawing on evidence from cognitive psychology on processes of self-immunization and dual rationality the authors demonstrate how self-deception against strong counter-evidence can be reconciled with the subject’s attaining belief-status for the target proposition. The question of belief-status is in the focus of Alfred Mele’s contribution. Mele has put forward an influential analysis which assumes that self-deceivers actually develop a belief of the target proposition. In contrast Robert Audi’s analysis claims that self-deceptive avowals are lacking belief-status. Mele provides evidence for the belief-view which has folk-intuition on its side. He has carried out two surveys that do not support Audi’s claim that a pre-theoretical understanding of the concept of ‘‘self-deception” includes the absence of belief. The results favor the opposite view. Concerning self-consciousness, Werning discusses the question whether it makes sense to presuppose a Cartesian inner self as the bearer of mental content. He argues that we need such an inner self without presupposing Descartes’ dualism. It is argued that the idea of an inner self is unavoidable if we accept two principles: the phenomenal transparency of experience and the semantic compositionality of conceptual content. It is assumed that self-awareness is a second-order state either in the domain of experience or in the domain of thought. In the former case self-awareness turns out empty if experience is transparent. In the latter, it can best be conceived of as a form of mental quotation. Since the only theory of quotation that is compositional is a phonological theory, we have to presuppose an inner self as the bearer of inner speech guaranteeing the compositionality of thought. An innovative transcultural study on the perception of nonverbal behavior that allows to study processes related to person perception and impression formation of others is presented by Bente et al. in their study ‘‘The Others. Universals and Cultural Specificities in the Perception of Status and Dominance from Nonverbal Behavior.” The authors employed a novel methodology that allowed to study nonverbal behavior from different countries in a culture-fair manner by masking to ethnicity and culture, the results of which show that the processing of dominance cues is universal whereas evaluative responses to nonverbal behavior are culture-dependent. This study is an important contribution to the literature on person perception in general and on its cultural influences in particular. In a bundle of four papers memory dimensions are investigated in detail (Staniloiu et al., Perner et al., Sauvage; Vosgerau). From a clinical neuropsychological point of view, Staniloiu et al. study the interesting case ‘‘Psychogenic Amnesia – A Malady of the Constricted Self.” referring to autobiographical–episodic memory as key component of self-consciousness. Autobiographical–episodic memory is understood as the conjunction of subjective time, autonoetic consciousness and the experiencing self. Neurobiologically, the prefrontal cortex and the limbic system are assumed to play key roles. These brain regions appear to be particularly vulnerable to stress-induced brain and are therefore of relevance for disorders of the autobiographical memory like psychogenic amnesia that are linked to environmental influences. The aim of Perner et al. is to investigate the role of the ability of mental rotation for episodic memory. To do that they investigate the common development of children’s ability to ‘‘look back in time” (retrospection, episodic remembering) and to ‘‘look into the future” (prospection). They register free recall, as a measure of episodic remembering (i) with recall of visually experienced events and (ii) with recall of indirectly conveyed events. Quite unexpectedly, ‘‘mental rotators” were markedly worse on indirect items than ‘‘non-rotators”. They speculate that with the ability to rotate children switch from the ability of knowledge retrieval to episodic remembering. Editorial / Consciousness and Cognition 19 (2010) 687–689 689 In her contribution ‘‘ROC in animals: uncovering the neural substrates of recollection and familiarity in episodic recognition memory.” Magdalena Sauvage focuses to the basic neurobiological mechanisms of episodic memory and explores the question whether familiarity and recollection are qualitatively distinct processes that in turn recruit different brain regions or whether they refer to the same process which would let us to assume that shared brain mechanisms are at work. Based on an animal model and innovative research methodologies she favours the view that familiarity and recollection are different processes, from a neurobiological point of view she provides data that lead to the conclusion that the hippocampus supports recollection, but not familiarity. Also related to recognition memory is the employment of simple heuristics that allows decision making on the ground of only sparse information. Volz et al. study the so-called fluency heuristic based on the relative speed with which two different objects are recognized in their study. ‘‘It just felt right: The neural correlates of the fluency heuristic.” Searching for the neural mechanisms of the fluency heuristic showed that the claustrum was recruited during fluency heuristic decisions suggesting that this brain region is responsible for the integration perceptual and memory elements into a conscious ‘‘Gestalt” and, hence, for the subjective experience of fluency. Vosgerau argues that any theory of content has to adopt a ‘‘functionalistic core” according to which representations are defined as substitutes in functions that describe the flexible behavior to be explained by the representation. The content of a representation can thus only be determined if the representation is ‘‘in use”. The stored entities in memory are not in use while they are stored, and hence cannot be assigned a specific content. The term ‘‘template” is introduced to describe stored entities in memory while activated memory is essentially in need of a construction process on the basis of the templates. Early versions of most of the papers were first presented at an international and interdisciplinary conference at the Hanse-Institute of Advanced Study in Delmenhorst (Germany) organized by the editors in summer 2009. This conference was a highlight of a series of ongoing interdisciplinary cooperations including the edition of this volume. We want to thank the VolkswagenStiftung and the Hanse-Institute of Advanced Study, especially its rector Reto Weiler who all supported the interdisciplinary endeavor, not only financially, but also intellectually. The peer-review process was managed at the University of Bochum with important organizational support by Robert Schütze. Finally, we would like to thank the management of the journal, Ann Barajas, and the main editor of the journal, William Banks, for helpful comments and encouraging support. Albert Newen Universität Bochum, Institut für Philosophie, Universitätsstr. 150, 44892 Bochum, Germany E-mail address: albert.newen@rub.de Kai Vogeley University Hospital Cologne, Department of Psychiatry, Kerpener Str. 62, 50924 Cologne, Germany Christoph Michel Universität Bochum, Institut für Philosophie, Universitätsstr. 150, 44892 Bochum, Germany Available online 8 August 2010

Consciousness and Cognition 22 (2013) 975–986 Contents lists available at SciVerse ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Relationships between non-pathological dream-enactment and mirror behaviors q Tore Nielsen a,b,⇑, Don Kuiken c a Dept. Psychiatry, Université de Montréal, Montréal, Québec, Canada Center for Advanced Research in Sleep Medicine, Hôpital du Sacré-Coeur de Montréal, Montréal, Québec, Canada c Dept. Psychology, University of Alberta, Edmonton, Alberta, Canada b a r t i c l e i n f o Article history: Received 19 December 2012 Available online 19 July 2013 Keywords: Dreaming Dream-enacting behaviors Mirror behaviors Mentalizing network Mirror neuron system Empathy Motor-affective resonance Sex differences a b s t r a c t Dream-enacting behaviors (DEBs) are behavioral expressions of forceful dream images often occurring during sleep-to-wakefulness transitions. We propose that DEBs reflect brain activity underlying social cognition, in particular, motor-affective resonance generated by the mirror neuron system. We developed a Mirror Behavior Questionnaire (MBQ) to assess some dimensions of mirror behaviors and investigated relationships between MBQ scores and DEBs in a large of university undergraduate cohort. MBQ scores were normally distributed and described by a four-factor structure (Empathy/Emotional Contagion, Behavioral Imitation, Sleepiness/Anger Contagion, Motor Skill Imitation). DEB scores correlated positively with MBQ total and factor scores even with social desirability, somnambulism and somniloquy controlled. Emotion-specific DEB items correlated with corresponding emotion-specific MBQ items, especially crying and smiling. Results provide preliminary evidence for cross-state relationships between propensities for dream-enacting and mirror behaviors—especially behaviors involving motor-affective resonance—and our suggestion that motor-affective resonance mediates dream-enactment imagery during sleep and emotional empathy during waking. Ó 2013 Elsevier Inc. All rights reserved. 1. Introduction Dream-enacting behaviors (DEBs), or acting out of the fictive movements, speech or emotions of a dream, are prevalent among normal college students (Nielsen, Svob, & Kuiken, 2009). Although most young adults experience them occasionally, frequent and intense DEBs are symptomatic of REM sleep behavior disorder (RBD) (Schenck, Bundlie, Ettinger, & Mahowald, 1986), sleep walking and sleep terrors (Oudiette, Leu, et al., 2009), and other nocturnal anomalies (Ohayon & Schenck, 2010) such as sleep-related eating disorders (Brion et al., 2012) and obstructive sleep apnea (Iranzo & Santamaria, 2005). One hypothesis is that DEBs occur during dreams that are of sufficient perceptual, dramatic, and emotional intensity to over-ride the neuromuscular inhibition of REM sleep. A second (and compatible) hypothesis is that DEBs result from direct disruption Abbreviations: DEBs, dream-enacting behaviors; MBQ, Mirror Behavior Questionnaire; RBD, REM sleep behavior disorder; ToM, theory-of-mind; IFG, inferior frontal gyrus; REM, rapid eye movement; MRI, magnetic resonance imaging; EEG, electroencephalogram; DEBS, dream-enacting behavior scale; SDS, social desirability scale; EMG, electromyogram. q The study was conducted at the Department of psychology of the University of Alberta, Edmonton, Alberta, Canada and the Dream & Nightmare Laboratory, Center for Advanced Research in Sleep Medicine, Hopital du Sacre-Coeur, Montreal, Quebec, Canada. ⇑ Corresponding author. Address: Dream & Nightmare Laboratory, Hôpital du Sacré-Coeur du Montréal, 5400 boul. Gouin Ouest, Montréal, Québec H4J 1C5, Canada. Fax: +1 514 338 2531. E-mail address: tore.nielsen@umontreal.ca (T. Nielsen). 1053-8100/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.concog.2013.06.005 976 T. Nielsen, D. Kuiken / Consciousness and Cognition 22 (2013) 975–986 of the brainstem mechanisms that maintain the atonia of REM sleep. Either dream intensity, atonia disruption, or their combination may explain how the action tendencies embodied in the fictive portrayal of a dreamer’s actions result in overt behaviors that are isomorphic with the actions of the dreamed self. Consistent with this model, among RBD patients, violent actions by the dreamed self frequently culminate in overtly aggressive outbursts that may even injure the patient or his/her sleeping partner (Schenck, Lee, Bornemann, & Mahowald, 2009). Similarly, among healthy subjects, DEBs isomorphic with actions of the dreamed self are evident during nightmares but also during very sad, angry or mirthful dreams—and even intensely erotic ones (Nielsen et al., 2009). However, the isomorphism between DEBs and fictive actions of the dreamed self may be the most salient aspect of an even more inclusive isomorphism. Some evidence suggests not only isomorphism between DEBs and actions of the dreamed self but also between DEBs and actions of other dream figures. For example, although the DEBs of RBD patients are consistently isomorphic with the aggressive actions of the dreamed self, they are regularly isomorphic with the actions of other dream figures as well. That is, their defensive—and enacted—aggression mirrors the aggression perpetrated by other dream figures (Iranzo & Santamaria, 2005; Nielsen, 2011; Schenck & Mahowald, 2002). Similarly, the DEBs of healthy subjects are regularly isomorphic with the actions of the dreamed self, but they are also occasionally isomorphic with the actions of other dream figures. For example, one laboratory subject reported that her dream enactive utterance coincided with the verbalizations of another dream character struggling aggressively to control a horse (Nielsen, McGregor, Zadra, Ilnicki, & Ouellet, 1993). The preceding examples involve the aggression that is especially common in REM dreams (McNamara, McLaren, Smith, Brown, & Stickgold, 2005), but DEBs do not only involve aggression. RBD patients regularly report violent behaviors but also some nonviolent behaviors that are isomorphic with their own dreamed actions, such as digging up dreamed treasure or giving bossy commands (Oudiette, de Cock, et al., 2009). Similarly, NREM somnambulism and sleep terrors involve not only isomorphic aggressive actions (e.g., fighting in self-defence), but also isomorphic avoidant behaviors (e.g., screaming in fear) and expressions of emotion that are independent of fight or flight (e.g., singing dirty or childish songs) (Oudiette, Leu, et al., 2009). In these cases, too, isomorphism between DEBs and the actions of a dreamed self are complemented by occasional isomorphism between DEBs and the actions of other dream characters. For example, the dreamer may enact the growling of a feline dream figure (Eiser & Schenck, 2005); the dreamer’s gestures (pointing fingers, flailing arms) may be synchronized with another character’s speech (Oudiette, de Cock, et al., 2009); or the dreamer may enact another’s physical assault while dreaming about actively protecting the victim (Schenck & Mahowald, 2002). Thus, dreams with DEBs are regularly isomorphic with the actions of the dreamed self, but they are also sometimes isomorphic with the actions of other dream characters. Extension of the DEB isomorphism to include the actions of figures other than the dreamed self is consistent with evidence that sometimes the fictive actions of a dreamed other correspond with the site of somatosensory stimulation administered during REM sleep. For example, Koulack (1969) found that electrical stimulation applied to the wrist during REM sleep influenced not only the actions of the dreamed self but sometimes also the actions of other dreamed characters. Similarly, Nielsen et al. (1993) found that application of pressure stimulation to the legs during REM sleep sometimes influenced the leg sensations and actions of other characters (see dream reports in Table 2). These experimental findings suggest that covert action tendencies resulting from bodily stimulation, like overt DEBs, are isomorphic not only with actions of the dreamed self but also with the actions of other dreamed figures. This more broadly construed isomorphism is reminiscent of the now compelling evidence that similar cortical areas (the mirror neuron system) are activated when a specific action (e.g., grasping something) is executed, when that action is observed, when that action is imagined for oneself (Filimon, Nelson, Hagler, & Sereno, 2007), and when it is imagined for another (Saygin, McCullough, Alac, & Emmorey, 2010; Zwaan, Taylor, & de Boer, 2010). Perhaps the isomorphism between DEBs, the actions of the dreamed self, and the actions of dreamed others depends upon the same motor-affective resonance that mediates personal and interpersonal perception and imagination during waking. The present study examines the correlation between the isomorphism that shapes DEBs and the isomorphism that shapes waking mirror behaviors (e.g., motor mimicry, empathy). Neurophysiological factors responsible for an individual‘s propensity to resonate with the emotions and actions of other characters during dreaming remain unclear, but they may depend upon neural networks that underlie basic social cognition. Two anatomically distinct networks subserving social cognition include the mentalizing network and the mirror neuron system. The mentalizing network supports theory-of-mind (ToM) functions, i.e., the generic ability to recognize and explain mental states (emotions, beliefs, motives) of others (Premack & Woodruff, 1978), including the mental states of fictional others, such as characters in literary texts (Mar & Oatley, 2008). Brain imaging has revealed a consistent ensemble of regions that comprise the mentalizing network: the medial prefrontal cortex, posterior cingulate/precuneus, and bilateral temporal junction. Functioning of the mentalizing network has been explored in dream content, specifically, in the mental features that are attributed to other dreamed characters (Kahn & Hobson, 2005; Kahn, Pace-Schott, & Hobson, 2002; Schweickert & Xi, 2010). An early general assessment of dream character attributes (Kahn et al., 2002), as well as a more focused study of ToM (Kahn & Hobson, 2005), found that participants are frequently aware of what they and other dream characters are feeling in the dream. The dreamer is also frequently interested in other characters’ feelings about him/herself. One study found ToM attributions to be stable in dreams, tending to be maintained once characters are introduced into the narrative—even in the face of character metamorphoses (Schweickert & Xi, 2010). T. Nielsen, D. Kuiken / Consciousness and Cognition 22 (2013) 975–986 977 Recent two-level ToM theories (e.g., Carr, Iacoboni, Dubeau, Mazziotta, & Lenzi, 2003; Spunt & Lieberman, 2012) stipulate that the mentalizing network depends upon activity in the more basic mirror neuron system. The latter becomes active both during observation of an action and performance of that action (di Pellegrino, Fadiga, Fogassi, Gallese, & Rizzolatti, 1992; Rizzolatti & Craighero, 2004) and has been linked to a neural circuit that includes the posterior inferior frontal gyrus (IFG), premotor cortex, and inferior parietal lobe (see Molenberghs, Cunnington, & Mattingley, 2012 for review). Involvement of the mirror neuron system in the orchestration of self- and other-character imagery during dreaming has been proposed (Nielsen, 2007). Such involvement could explain some peculiarities of dream imagery, including the results of the somatosensory stimulation studies just mentioned (Koulack, 1969; Nielsen et al., 1993). Such involvement may also explain the interplay between body image and ‘‘felt presence’’ in sleep paralysis episodes, as well as in other hallucinatory events that involve the fusion of a sense of self and a sense of an other (Nielsen, 2007). According to two-level theories of mentalization, the mirror neuron system is essential for the early stage, preconceptual identification of emotions. Thus, it may mediate affective resonance (‘‘emotion contagion’’), including contagious crying (Geangu, Benga, Stahl, & Striano, 2010), laughing (Sherman, 1975), fear (Zhou & Chen, 2009), anxiety (Eilam, Izhar, & Mort, 2011), and sadness (Papousek, Schulter, & Lang, 2009), but possibly also simple contagious yawning (Cooper, Puzzo, & Pawley, 2008). The mirror neuron system may also underlie basic emotional empathy (Schulte-Ruther, Markowitsch, Fink, & Piefke, 2007). Preconceptual emotion identification involving the mirror neuron system (e.g., the identification of anger) may enable higher-level explanation of the identified emotion through the mentalizing network (e.g., a causal explanation for anger). Functional MRI has revealed components of the mirror neuron system—the right posterior IFG in particular—that become functionally connected with the mentalization network during such higher-level attributions (Spunt & Lieberman, 2012). The right IFG (rIFG) is one component of the mirror system of particular interest in explaining DEBs because its activity may subserve both response inhibition (Chamberlain & Sahakian, 2007) and the development of separate representations of self and other (Spengler, von Cramon, & Brass, 2010). Reduced activity in the rIFG, during overall activation of the mirror system, may explain the release from REM atonia that produces DEBs and the permeable boundaries (isomorphisms) between self and other in the dreams that accompany DEBs. It is noteworthy that rIFG activity is not decreased during normal REM sleep (Braun et al., 1998) when most vivid dreaming with normal muscle atonia occurs and clear separation between self and other in dream characters is evident (cf. Kahn & Hobson, 2005). In summary, both non-pathological DEBs observed in the general population and the pathological DEBs of RBD may reflect the abnormal reduction of rIFG inhibitory activity during overall activation of the mirror system, leading both to a reduction in muscle atonia and to the failure to develop clearly separate representations of self and other. The goal of this study was to examine whether the self-reported frequencies of dream-enacting and mirror behaviors during waking are related. Although a number of methods have been developed for assessing the suggested neurophysiological manifestations of mirror neuron activity (e.g., EEG mu rhythm), we could locate no validated instrument for assessing differences in a subject’s general propensity for expressing different types of mirror behaviors. We therefore constructed an exploratory self-report questionnaire, the Mirror Behavior Questionnaire (MBQ; see Appendix A), to assess several contagious and imitative behaviors. We expected that, if the mirror neuron system is implicated in both DEBs and mirror behaviors, we would observe positive cross-state correlations between DEBS and MBQ scores. We expected that these correlations would be independent of a general tendency to respond to questions with socially desirable responses. Finally, since DEBs are more prevalent among females than males (Nielsen et al., 2009), we expected to see sex differences in mirror behavior scores and in their correlations with DEBs. 2. Methods Subjects were 492 students (188 males; 292 females; 12 not specified) enrolled in a first-year University Psychology course (Mage: 19.1 ± 1.62 yrs; range: 17–29) and receiving partial course credit for participation. They gave informed consent and participated voluntarily; they were free to choose an alternative educational activity for course credit. There were no differences in age between the three groups (M: 19.2 ± 1.73 yrs; F: 19.0 ± 1.55 yrs; NS: 19.1 ± 0.99 yrs; p > .42). All subjects completed an extensive battery of questionnaires as part of a larger research program on personality and dreaming; the specifics of the battery are published elsewhere (Nielsen et al., 2009) and only some results will be described here. Subjects entered their responses to questionnaires on standard, optically scored, answer sheets that were verified manually to remove records with incorrectly coded or out of range responses. Following participation, subjects were given a thorough written debriefing. 2.1. Questionnaire measuresmeasures Dream-Enacting Behaviors Scale (DEBS). Dream-enacting behaviors were assessed with a 6-item self-report scale derived from the results of our previous studies of DEBs in college students (Nielsen et al., 2009); items included speaking out, crying/sobbing, smiling/laughing, bodily fear, anger/defensive behavior, and other movements that often occur during 978 T. Nielsen, D. Kuiken / Consciousness and Cognition 22 (2013) 975–986 Table 1 Mirror Behavior Questionnaire items and categories. Item Content Category MB1 MB2 When you see someone else crying, are you likely to start crying as well? When you see someone else yawning, are you likely to start yawning yourself? When you see someone else sleeping or falling asleep are you likely to feel sleepy yourself? When you see someone expressing anger, are you likely to feel anger yourself? If someone smiles directly at you, are you likely to smile, too? If you see someone smiling at someone else, are you likely to smile, too? Do you laugh out loud when you see someone else laughing? When you see another person’s terrified face, do you feel fear, too? When you are interacting with another person, do you tend to copy their body posture, e.g., folded arms, hands on hips, crossed legs, etc.? When you are interacting with another person, do you tend to copy their body movements, e.g., gesturing with your face or hands, dramatizing by moving around, etc.? When you are interacting with another person who has noticeable ‘nervous motor tics’ (e.g., playing with their hair, rubbing their nose, tapping their foot or fingers, pulling at their clothes), do you start to imitate some of these tics yourself? When you are speaking with someone who has a noticeable accent or dialect, do you tend to imitate features of this accent or dialect? When you are interacting with someone who has noticeable ‘verbal tics’ (e.g., ‘like,’ ‘I mean,’ ‘actually,’ ‘you know,’ etc.) or pauses and hesitations (e.g., ‘um,’ ‘er,’ ‘ah’), do you start using some of these features in your own speech? Do you enjoy imitating the voices of famous people or cartoon characters? When you find yourself together with a young child who is playing a fantasy game, do you join in and play the game with him/her? Are you a ‘physically active’ spectator, i.e., when watching a favorite activity or sport do you get physically involved by copying movements that you see (or would like to see)? I easily learn a new action or skill (e.g., dance style, sports technique, use of a tool) simply by watching someone else performing it I experience a lot of empathy towards others, i.e., I am able to feel more or less what they are feeling Emotional contagion Sleep contagion MB3 MB4 MB5 MB6 MB7 MB8 MB9 MB10 MB11 MB12 MB13a MB14 a MB15 MB16 MB17 MB18 a Sleep contagion Emotional contagion Emotional contagion Emotional contagion Emotional contagion Emotional contagion Communicative imitation Communicative imitation Communicative imitation Communicative imitation Communicative imitation Communicative imitation Communicative imitation Communicative imitation Imitative learning Empathy Item dropped as result of factor analysis. transitions from dreaming to wakefulness. Each item was rated using a 4-point response scale (0 = never; 1 = rarely; 2 = sometimes; 3 = often). A 7th item on sexual arousal assessed by the questionnaire was excluded from this scale for several reasons. Sexual arousal as reflected in objectively measured penile tumescence is associated with dreamed anxiety (Karacan, Goodenough, Shapiro, & Starker, 1979) rather than dream eroticism. Also, item analysis revealed this item to be poorly correlated with other items (all p > .25); its removal from the scale was the only one to increase the scale’s internal consistency (Cronbach’s alpha = 0.76). A principal components analysis (Kaiser normalization, varimax rotation) of the 7 DEBs items produced a single factor (40.7% variance accounted for) on which the sexual arousal item loaded only very weakly (.382; all other items >.682). Finally, sexual arousal was the only item correlated with age (r = .09, p = .04; all other p > .176). Two additional items assessing somnambulism and somniloquy were also assessed to help distinguish true DEBs from these alternative forms of behavioral enactment in sleep. 2.1.1. Mirror Behavior Questionnaire (MBQ) Mirror behaviors were assessed with an 18-item self-report scale (see Table 1). Items were selected to represent common contagious emotions (smiling, laughing), communicative mirroring (speech and motor tics, body movements), imitative learning (learning a new skill, imitating voices) and contagious sleepiness/yawning. It also included an empathy item. Each MBQ item was rated using a 4-point response scale: 0 = never; 1 = rarely; 2 = sometimes; 3 = often. 2.1.2. Social Desirability Scale (SDS) The 13-item short-form of the Marlowe–Crowne Social Desirability Scale was administered (Crowne & Marlowe, 1960; Reynolds, 1982); it measures a form of response bias involving the presentation of self in a socially desirable (or undesirable) light. Internal consistency of the SDS is adequate, from .70 to .79 among student populations (Crino, Svoboda, Rubenfeld, & White, 1983; Tanaka-Matssumi & Kameoka, 1986); scores do not differ for men and women (Loo & Thorpe, 2000; O’Grady, 1988) and did not differ in the present study (males: 4.64 ± 2.72; females: 4.63 ± 2.88; t460 = 0.032, p = .975; SDS forms were incomplete for 31 subjects). 979 T. Nielsen, D. Kuiken / Consciousness and Cognition 22 (2013) 975–986 Table 2 Principal components factor analysis four-factor solution for Mirror Behavior Questionnaire items. Factors: 1 = Empathy/Emotional Contagion; 2 = Behavioral Imitation; 3 = Sleepiness/Anger Contagion; 4 = Motor Skill Imitation. MBQ# Item Factors 1 18 1 6 8 7 5 10 9 11 12 3 4 2 14 17 16 2 3 4 Empathy Cry contagion Smile contagion (at other) Fear contagion Laugh contagion Smile contagion (at you) Copy body movements Copy body posture Copy motor tics Copy dialects Sleep contagion Anger contagion Yawn contagion Imitate famous voices Learn new skill Active spectator 0.656 0.633 0.629 0.540 0.536 0.486 0.237 0.263 0.082 0.032 0.136 0.032 0.388 0.169 0.142 0.204 0.230 0.093 0.084 0.335 0.183 0.086 0.760 0.720 0.650 0.582 0.135 0.172 0.069 0.146 0.036 0.065 0.169 0.070 0.193 0.189 0.109 0.255 0.199 0.187 0.266 0.198 0.750 0.600 0.506 0.130 0.108 0.089 0.047 0.111 0.221 0.068 0.159 0.185 0.003 0.099 0.106 0.292 0.016 0.154 0.115 0.689 0.650 0.624 %Variance explained 15.2 13.5 9.9 9.7 Extraction method: principal component analysis; rotation method: Varimax with Kaiser normalization; rotation converged in 5 iterations. Coefficients in bold indicate items loading >.45 on each of the four factors. 3. Results 3.1. Dream-enacting Behaviors Scale (DEBS) The overall mean DEBS score was 6.63 ± 3.80 with a median of 6, mode of 8, range of 0–18 (out of 18) and internal consistency (Cronbach’s alpha or CA) of .672. Kurtosis of the distribution ( .283) was within the normal range of ±2SD of the SE of kurtosis (.220), i.e., .440 to +.440. Overall, 96.7% of participants had a DEBs score >0. Females had higher DEBS scores (M = 7.20 ± 3.88; CA = .681) than did males (M = 5.80 ± 3.53; CA = .629; t478 = 3.99, p < .0001) and more females (98.3%) than males (94.1%) had a DEBs score > 0. DEBS scores did not correlate with social desirability (SDS scores) for the entire sample (r471 = .023, p = .624) or for males (r178 = .063, p = .399) or females (r280 = .000, p = 1.000) considered separately. 3.2. Mirror Behavior Questionnaire The overall mean MBQ score was 24.1 ± 6.18 with a median of 24, mode of 25, and range of 7–40 (out of 48) and CA of .769. Kurtosis of the distribution ( .105) was within the normal range of ±2SD of the SE of kurtosis (.220). Females had higher MBQ scores (M = 25.2 ± 5.94) than did males (M = 22.6 ± 6.30; t478 = 4.53, p < .0001). MBQ scores did not correlate with SDS scores for the entire sample (r473 = .064, p = .162) or for males (r180 = .015, p = .840), but did so marginally and negatively for females (r209 = .109, p = .068), indicating that the more females possessed a propensity to give socially desirable responses the less they tended to report mirror behaviors. MBQ scores correlated weakly but positively with both somnambulism (r488 = .154, p = .001) and somniloquy (r482 = .151, p = .001). For males both correlations were significant (r186 = .146, p = .046; r185 = .259, p = .0004 respectively), whereas for females a correlation was found for somnambulism (r290 = .153, p = .009) but only marginally for somniloquy (r285 = .089, p = .134; Fisher-z = 1.85, p = .064 versus males). In view of these potential confounds, subsequent correlational analyses involving the DEBS partialed out both variables. Initial principal-components analyses (IBM SPSS Statistics v19.0, 2010) with varimax rotation indicated that two MBQ items (playing with children, copying verbal tics) cross-loaded weakly on more than one factor; they were removed. Reanalysis of the 16 remaining items produced a five-factor solution (54.7% VE), two of which separated negative from positive contagious emotion items but for which the ‘fear contagion’ item loaded on both factors. When constrained to a four-factor solution (Table 2; 48.4% VE), the negative and positive factors and the fear contagion item combined to create a single, Empathy/EmotionalContagion, factor (Table 1 items MB1, MB5-MB8; 15.2% VE). The other 3 factors were the same as those from the five-factor solution; they were labeled Behavioral Imitation (MB9–MB12; 13.5% VE), Sleepiness/AngerContagion (MB2–MB4; 9.9% VE) and Motor Skill Imitation (MB14, MB16, MB18; 9.7% VE). The factor analyses were subsequently repeated using an oblique rotation method (oblimin) and very similar factors were obtained. Correspondingly, inter-component correlations were seen only between factors 1 and 2 (r = .251) and factors 1 and 4 (r = .248; all others: r < .170). The four-factor varimax solution was retained and factor scores generated for use in all further analyses. The substantial loading of yawning on both factors 1 and 3 was further investigated by assessing correlations between yawning and the 7 MBQ emotion contagion items. As shown in Table 3, all correlations were positive and significant for 980 T. Nielsen, D. Kuiken / Consciousness and Cognition 22 (2013) 975–986 Table 3 Pearson correlations between MBQ yawning item and all 7 MBQ emotion contagion items. Empathy Crying Anger Smile at you Smile at other Laugh Fear Male r p .068 .35683 .210 .00382 .181 .01300 .132 .07065 .150 .04010 .170 .01976 .283 .00009 Female r p .200 .00059 .160 .00623 .146 .01280 .193 .00092 .211 .00029 .136 .01997 .246 .00002 Total r p .182 .00005 .279 .00000 .151 .00079 .200 .00002 .200 .00001 .153 .00066 .309 .00000 Fig. 1. Pearson correlations between DEBS, MBQ (salmon bar), MBQ factors (dark blue bars), and somnambulism and somniloquy scores (graded pale blue bars). Horizontal black segments indicate partial correlations after removal of somnambulism and somniloquy scores. *****p < .0000001; ****p < .000001; *** p < .0001; **p < .001; *p < .05. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) the entire sample. When considered separately by sex, this pattern of coefficients was clearly present for females, but for males the correlation between yawning and empathy was not significant. Multivariate comparisons by sex using a one-way MANOVA with factor scores as dependent measures revealed a multivariate effect (T4,472 = 43.70, p < .0001) by which females scored higher than males on factor 1, Empathy/EmotionalContagion (F1,475 = 77.43, p < .0001), and males scored higher than females on factor 4, Motor Skill Imitation (F1,475 = 51.83, p < .0001). Covarying SDS did not diminish the multivariate effect (T4,453 = 42.26, p < .0001) or the effects for factor 1 (F1,456 = 90.51, p < .0001) or factor 4 (F1,456 = 47.64, p < .0001). Covarying somnambulism and somniloquy did not diminish the multivariate effect (T4,459 = 42.04, p < .0001) or the effects for factor 1 (F1,462 = 87.04, p < .0001) or factor 4 (F1,462 = 51.80, p < .0001; all other p = ns). 3.3. Global relationships between DEBS and MBQ scores Pearson correlations between the DEBS and MBQ scores revealed moderate and highly significant, positive relationships for the entire sample (r492 = .346, p < .0001) and for males (r188 = .405, p < .0001) and females (r292 = .284, p < .0001) considered separately. The difference in correlations for the two sexes was not significant when tested with the Fisher r-to-z transformation (z = 1.46, p = .144, 2-tailed). The DEBS  MBQ correlations remained significant when SDS was used as a covariate (total: r490 = .346, p < .0001; males: r177 = .405, p < .0001; females: r279 = .286. p < .0001); they also remained significant when somnambulism and somniloquy were partialed out (total: r477 = .308, p < .0001; males: r180 = .347, p < .0001; females: r281 = .251, p = .0003). For the entire sample, DEBS scores also correlated positively with each of the 4 MBQ factor scores (Fig. 1), especially factor 1, Empathy/EmotionalContagion (r489 = .272, p < .0001). For males, DEBS scores correlated with MBQ factor 1 (r187 = .254; p = .0005), factor 3 (r187 = .256; p = .0004) and factor 4 (r187 = .230; p = .002); for females, they correlated with MBQ factor 1 (r290 = .223; p = .0001) and factor 4 (r290 = .245; p < .0001). Partialing out SDS did not alter this profile of relationships. Also, T. Nielsen, D. Kuiken / Consciousness and Cognition 22 (2013) 975–986 981 when somnambulism and somniloquy were used as covariates, only the correlation between DEBS and factor 2 became insignificant (r475 = .051; p = .270; all other p < .003; Fig. 1). 3.4. Relationships between specific DEBS and MBQ scores To determine the specificity of relationships between dream-enacting and mirror behaviors, correlations were calculated between, on the one hand, individual DEBS items (e.g., crying) and all appropriate MBQ items (e.g., smiling, crying, copy dialects, etc.) and, conversely, individual MBQ items (e.g., crying) and all appropriate DEBs items (e.g., smiling, fear, anger, crying, etc.). These calculations revealed a high degree of specificity for some behaviors (see Table 3 and Fig. 2), especially crying (p = .003), and to a lesser extent laughing (p = .063) (see Table 4). 4. Discussion Results provide support for the expected cross-state relationship between an individual’s propensities for expressing dream-enacting behaviors during sleep and mirror behaviors while awake. The global nature of this relationship is reflected in the fact that total scores on both the DEBS and the MBQ were significantly correlated, that this relationship was observed for both male and female subjects, and that the relationships were largely independent of self-reported somnambulism, somniloquy and a tendency to respond in a socially desirable manner. The generality of the results was also reflected in the finding of generally elevated nonspecific correlations between various mirror behaviors and diverse DEBs. For example, the mirror behaviors of crying and laughing both correlated to some extent with general dreamed movements and other dreamed emotional behaviors. Further, factor analyses confirmed the existence of at least 4 subtypes of mirror behaviors that were associated with DEBs. The subset of mirror behaviors that was most clearly associated with dream-enacting behaviors (i.e., MBQ factor 1) included most of the positive and negative emotion contagion items and the empathy item. This grouping is not only consistent with the existence of a single affective resonance system that includes empathy, but links this system strongly with DEBs. However, the DEBS also was associated with two subsets of items that did not involve emotion (MBQ factor 3, Behavioral Imitation; MBQ factor 4, Motor Skill Imitation). This pattern is consistent with emerging views that the mirror neuron system includes a core circuit and peripheral circuits that subserve separate but interdependent motoric (e.g., movement) and non-motoric (e.g., affective) functions (Molenberghs et al., 2012). Consistent with twolevel theories of mentalization, we suggest that DEBs are mediated by components of the mirror neuron system, especially Fig. 2. Correlations between individual dream-enactment and mirror behavior items. (a) DEBS crying item correlates with several MBQ items but most distinctively with MBQ crying item (p < .000001; graded blue bar); (b) MBQ crying item correlates with several DEBS items but most distinctively with DEBS crying item (p < .000001). (c and d) Similar patterns of correlations were found for DEBS smiling/laughing and MBQ laughing items. Red lines indicate levels of significance for p-values of .01 and .001. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) 982 T. Nielsen, D. Kuiken / Consciousness and Cognition 22 (2013) 975–986 Table 4 Correlations between individual DEBS items and corresponding MBQ items compared with average correlations for all other MBQ items. The DEBS crying item correlates with the corresponding MBQ crying item (r = .310) to a significantly greater degree (p = .003) than it does with all other MBQ items averaged (r = .131). DEBS item Corresponding MBQ item Ave of all other MBQ items Smiling Fear signs Anger Crying Talking Other movement .206 .147 .190 .310 .093 .123 .090 .111 .108 .131 .107 .092 Z p 1.86 0.57 1.31 2.95 0.22 0.49 .063 .569 .190 .003 .926 .624 by the components that mediate early-stage emotional resonance, rather than by the mentalizing system that mediates later-stage emotion attribution and emotion explanation (Spunt & Lieberman, 2012). The specificity of such early-stage emotional resonance is suggested by the relationships observed between selected DEBS and MBQ items. The emotional items crying and smiling on one questionnaire correlated distinctively with their counterparts on the other questionnaire. Thus, individuals who frequently reported crying, smiling or laughing while awakening from their dreams were also more likely to report crying, smiling or laughing when observing similar behaviors by others during wakefulness. These specific correlations may reflect the fundamentally contagious nature of these two particular emotions. Contagious crying can be triggered as early as 18 h following birth (Martin & Clark, 1982), it occurs among 1- to 9-month-old boys and girls, and it does not decrease in prevalence with age (Geangu et al., 2010). Similarly, the infectious nature of laughter (‘group glee’) has been identified in preschool children (Sherman, 1975) and, in adults, mirror reactions to happy facial stimuli, as measured by facial EMG, occur within 500 ms, even when the stimuli are brief, subliminal or in avatar form (Rymarczyk, Biele, Grabowska, & Majczynski, 2011; Weyers, Muhlberger, Hefele, & Pauli, 2006). Thus, the contagious emotions most clearly associated with DEBs are basic, automaticized forms of emotional resonance that developmentally and cognitively precede higher-order mentalizing (e.g., emotional inference and explanation). The clear grouping of the crying, smiling, and laughing emotion items with empathy on Factor 1 of the MBQ is consistent with the notion that such resonance represents basic emotional empathy. Some MBQ items may not have shown clear correspondences across waking and dreaming because they describe behaviors that are less distinctive than crying, smiling, or laughing (e.g., the context specific variations of ‘‘other movement’’). Moreover, several forms of mirror waking behaviors (e.g., imitating famous voices, yawning, learning a new skill) simply do not have dream awakening counterparts. Beyond these methodological constraints, however, the lack of correspondence for some MBQ items may reflect important differences in the social functions of certain emotions. For example, involuntary anger contagion may incite an escalation of anger toward the interaction partner, rather than empathy (Weyers et al., 2006). And fear contagion (Rahko et al., 2010) may initiate joint attention to a potentially shared threat, rather than empathy. Because fear and anger responses to threat predominate in nightmares, while sadness is the predominant emotion in other types of distressing dreams (e.g., existential dreams (Busink & Kuiken, 1996; Kuiken, Chudleigh, & Racher, 2010); post-partum dreams (Nielsen & Paquette, 2007)), such contrasts among these forms—and functions—of emotional contagion warrant closer inspection in future studies. In contrast to the emotion-specific patterns, it is noteworthy that yawning was uniformly correlated with all other emotion contagion items. Yawning is a highly contagious behavior that appears as early as 4 years of age (Helt, Eigsti, Snyder, & Fein, 2010) and is frequently taken both as evidence of empathy (Platek, 2010) and of activity in the mirror neuron system (Cooper et al., 2012; Schurmann et al., 2005). Our results reinforce the notion that yawning has an evolutionarily preserved social function of communicating both sleepiness and emotion (Guggisberg, Mathis, Schnider, & Hess, 2010). 4.1. Motor and emotional resonance Although the mechanisms linking mirror and dream-enacting behaviors remain unclear, it seems likely that behaviors in both domains involve activation of a common neural network that mediates (1) the imitation of others’ movements and emotional expressions during waking; (2) the imagination of an other’s execution of those same movements and emotional expressions during waking; and (3) the imagined presence of characters enacting those activities and expressions during dreaming. 1. Imitative resonance with others’ movements and emotional expressions during waking is evident as surface EMG activity corresponding with the limb or facial muscle movements in an other’s expression of emotion; it is also evident as the vicarious activation of analogous brain areas while observing another person’s actions or emotional expressions (Dimberg, Andreasson, & Thunberg, 2011; Fadiga, Craighero, & Olivier, 2005; Lundqvist, 1995; Roosink & Zijdewind, 2010). 2. Resonance in imagination is evident as the vicarious activation of analogous peripheral and brain circuits while imagining another person’s actions or emotional expressions (Aoyama & Kaneko, 2011; Fourkas, Avenanti, Urgesi, & Aglioti, 2006; Fourkas, Bonavolonta, Avenanti, & Aglioti, 2008; Williams, Pearce, Loporto, Morris, & Holmes, 2012). T. Nielsen, D. Kuiken / Consciousness and Cognition 22 (2013) 975–986 983 3. Resonance in dream imagery may be evident as vicarious movements and emotions, as indicated by facial EMG, limb twitches and activity in motoric brain centers (Gerne & Strauch, 1985; Shimizu & Inoue, 1986). However, to the extent that these phenomena reflect only the dreamer’s movements and emotions or only another dream character’s movements and emotions, they may not be resonance reactions per se. More direct evidence of motor and affective resonance is provided by laboratory studies in which stimulation during REM sleep leads to reports of imagined activity in the analogous limbs of other (non-self) dream characters (Koulack, 1969; Nielsen et al., 1993). It is also evident in research in which, when subjects are asked to rehearse statements about attaining a desired personality attribute, the wished-for attribute manifests as a feature of both dreamed characters and the dreamed self (Cartwright, 1974). It may be concretely evident in the reports of sleep paralysis episodes involving the ‘‘felt presence’’ of another (Nielsen, 2007). Finally, at least two brain regions of the mirror neuron system (the inferior parietal lobule, the inferior frontal gyrus) show reduced activity during wakefulness among RBD patients (Mazza et al., 2006) for whom dream-enacting behaviors are frequent and vigorous. In sum, although the evidence is sparse, it seems increasingly likely that both dream-enacting and mirror behaviors are mediated by vicarious activation of the mirror neural network that accompanies the perception or imagination of others’ movements and emotional expressions. 4.2. Empathy, sex differences and DEBs Individuals vary in the extent to which they exhibit motor and affective resonance. The present results suggest that resonance is particularly likely among those who self-report high levels of empathy. One possibility is that this reflects a high level of familiarity with or skill in the execution of the other’s behaviors. For example, corticospinal facilitations of the muscle activity of expert tennis players are larger when they mentally imagine practicing a tennis forehand stroke than when they imagine other types of movements (Fourkas et al., 2008). Similarly, motor potentials in the arm muscles of frequent ballet spectators are larger when watching ballet than when watching other types of performances (Jola, Abedian-Amiri, Kuppuswamy, Pollick, & Grosbras, 2012). A second possibility is that these individual differences reflect a capacity for basic emotional (but not necessarily cognitive) empathy. For example, compared to low empathic subjects, highly empathic subjects display more facial EMG activity in response to appropriate facial stimuli and rate such facial stimuli as more emotionally intense (Dimberg et al., 2011; Mailhot, Vachon-Presseau, Jackson, & Rainville, 2012). However, it is important to consider a third possibility: because motor mirroring and basic empathy are mediated by the same neural network, the motor skill and empathy interpretations are not mutually exclusive. For example, corticospinal facilitation produced by watching a simple motor task is highly correlated with a measure of empathy (Lepage, Tremblay, & Theoret, 2010). Similarly, corticospinal excitability is higher when observed hand movements have an emotional component (Enticott, Kennedy, Bradshaw, Rinehart, & Fitzgerald, 2011; Montagna, Cerri, Borroni, & Baldissera, 2005). Sex differences also may be implicated in the differential expression of dream-enacting behaviors. For example, new mothers, who are preoccupied with the health and welfare of their newborns, are particularly likely to report dreamenacting behaviors (Nielsen & Paquette, 2007). The gray matter volume of the IFG component of the mirror neuron system is larger in women (Hammers et al., 2007) and correlates with self-reported empathy (Cheng et al., 2009), factors that may directly reflect an elevated capacity for resonant motoric and emotional responding. So may the finding that activity in the key rIFG region is elevated in sensitive mothers while they listen to their own infants cry (Musser, Kaiser-Laurent, & Ablow, 2012). However, such findings more likely reflect this resonance capacity only indirectly in that the rIFG, in particular, is important for inhibiting behavioral responding (Swann, 2012). That it is implicated in both controlled and automatic response inhibition (Lenartowicz, 2011) suggests it may be particularly suited for preventing resonant behaviors triggered elsewhere in the mirror neuron network. This notion is consistent with evidence of substantial rIFG functional plasticity and the suggestion that rIFG size may change to compensate for other cognitive or executive deficits (Hajek et al., 2013). Although the overall MBQ score was higher for females than for males in the present study, females also were higher specifically on the Empathy/EmotionalContagion factor. Together these results are consistent with research demonstrating higher self-reported empathy among females (Baron-Cohen & Wheelwright, 2004; Eisenberg & Lennon, 1983) and objective evidence that females demonstrate greater empathy in appropriate situations than do males. This sex difference in empathy emerges in young children (Auyeung, Allison, Wheelwright, & Baron-Cohen, 2012) and increases with age (Mestre, Samper, Frias, & Tur, 2009). Objective measures indicate that women are more empathic than men in a face-to-face empathy task and display more task-relevant activation of mirror neurons (Schulte-Ruther, Markowitsch, Shah, Fink, & Piefke, 2008; SchulteRuther et al., 2007). Further, both short- and long-latency ERP reactions to pictures of others in pain are more closely correlated with empathy ratings among women than they are among men (Han, Fan, & Mao, 2008). Other research has confirmed sex-related neuroanatomical differences in the human mirror-neuron system (in gray matter volumes) favoring women that are strongly linked to empathy competence (Cheng et al., 2009). It is noteworthy that DEBs were associated with somnambulism and somniloquy in our analyses. Mounting evidence that these sleep behaviors are often accompanied by vivid, emotional, dream imagery (Oudiette, Leu, et al., 2009) suggests that these NREM sleep parasomnias may constitute an alternative expression of mirror neuron activity during sleep, albeit one during which the absence of REM sleep atonia permits a more elaborate—if not dramatic—expression of the imagery. However, that our findings also demonstrate a clear independence between mirror behaviors and DEBs on the one hand 984 T. Nielsen, D. Kuiken / Consciousness and Cognition 22 (2013) 975–986 and somnambulism and somniloquy on the other supports the possibility that the mirror neuron system influences cognitive activity in similar ways throughout sleep. In conclusion, the newly developed MBQ assesses several types of mirror behaviors during waking that are correlated with dream enacting behaviors during sleep. This generic cross-state correlation is independent of social desirability, somnambulism and somniloquy. Moreover, correlations between emotion-specific DEB items and corresponding emotion-specific MBQ items, provide evidence of cross-state relationships especially for behaviors that involve emotional resonance. These emotion-specific correspondences suggest that, rather than implicating a broader mentalizing network concerned with the attribution and explanation of others’ mental states, DEBs reflect the more basic and developmentally earlier resonances mediated by the mirror neuron system. Though preliminary, the findings support our notion that motor and affective resonance mediates both the enactment of dream imagery during sleep and basic emotional empathy during waking. Acknowledgments The research was funded by the Canadian Institutes of Health Research (TN) and the Natural Sciences and Engineering Research Council of Canada (TN). Appendix A. Mirror Behavior Questionnaire (MBQ) The following questions concern reactions of yours that may be influenced by the behaviors of other people. Please rate each one on the 0–3 scale. 1. When you see someone else crying, are you likely to start crying as well? 2. When you see someone else yawning, are you likely to start yawning yourself? 3. When you see someone else sleeping or falling asleep are you likely to feel sleepy yourself? 4. When you see someone expressing anger, are you likely to feel anger yourself? 5. If someone smiles directly at you, are you likely to smile, too? 6. If you see someone smiling at someone else, are you likely to smile, too? 7. Do you laugh out loud when you see someone else laughing? 8. When you see another person’s terrified face, do you feel fear, too? 9. When you are interacting with another person, do you tend to copy their body posture, e.g., folded arms, hands on hips, crossed legs, etc.? 10. When you are interacting with another person, do you tend to copy their body movements, e.g., gesturing with your face or hands, dramatizing by moving around, etc.? 11. When you are interacting with another person who has noticeable ‘nervous motor tics’ (e.g., playing with their hair, rubbing their nose, tapping their foot or fingers, pulling at their clothes), do you start to imitate some of these tics yourself? 12. When you are speaking with someone who has a noticeable accent or dialect, do you tend to imitate features of this accent or dialect? 13. When you are interacting with someone who has noticeable ‘verbal tics’ (e.g., ‘like,’ ‘I mean,’ ‘actually,’ ‘you know,’ etc.) or pauses and hesitations (e.g., ‘um,’ ‘er,’ ‘ah’), do you start using some of these features in your own speech? 14. Do you enjoy imitating the voices of famous people or cartoon characters? 15. When you find yourself together with a young child who is playing a fantasy game, do you join in and play the game with him/her? 16. Are you a ‘physically active’ spectator, i.e., when watching a favorite activity or sport do you get physically involved by copying movements that you see (or would like to see)? Never(0) Rarely(1) Sometimes(2) Often(3) Never(0) Rarely(1) Sometimes(2) Often(3) Never(0) Rarely(1) Sometimes(2) Often(3) Never(0) Rarely(1) Sometimes(2) Often(3) Never(0) Never(0) Rarely(1) Rarely(1) Sometimes(2) Sometimes(2) Often(3) Often(3) Never(0) Never(0) Never(0) Rarely(1) Rarely(1) Rarely(1) Sometimes(2) Sometimes(2) Sometimes(2) Often(3) Often(3) Often(3) Never(0) Rarely(1) Sometimes(2) Often(3) Never(0) Rarely(1) Sometimes(2) Often(3) Never(0) Rarely(1) Sometimes(2) Often(3) Never(0) Rarely(1) Sometimes(2) Often(3) Never(0) Rarely(1) Sometimes(2) Often(3) Never(0) Rarely(1) Sometimes(2) Often(3) Never(0) Rarely(1) Sometimes(2) Often(3) 985 T. Nielsen, D. Kuiken / Consciousness and Cognition 22 (2013) 975–986 17. I easily learn a new action or skill (e.g., dance style, sports technique, use of a tool) simply by watching someone else performing it 18. I experience a lot of empathy towards others, i.e., I am able to feel more or less what they are feeling Never(0) Rarely(1) Sometimes(2) Often(3) Never(0) Rarely(1) Sometimes(2) Often(3) References Aoyama, T., & Kaneko, F. (2011). The effect of motor imagery on gain modulation of the spinal reflex. Brain Research, 1372, 41–48. Auyeung, B., Allison, C., Wheelwright, S., & Baron-Cohen, S. (2012). Brief report: Development of the adolescent empathy and systemizing quotients. Journal of Autism and Developmental Disorders. Baron-Cohen, S., & Wheelwright, S. (2004). The empathy quotient: An investigation of adults with Asperger syndrome or high functioning autism, and normal sex differences. Journal of Autism and Developmental Disorders, 34, 163–175. Braun, A. R., Balkin, T. J., Wesensten, N. J., Gwadry, F., Carson, R. E., Varga, M., et al (1998). Dissociated pattern of activity in visual cortices and their projections during human rapid eye movement sleep. Science, 279, 91–95. Brion, A., Flamand, M., Oudiette, D., Voillery, D., Golmard, J. L., & Arnulf, I. (2012). Sleep-related eating disorder versus sleepwalking: A controlled study. Sleep Medicine, 13, 1094–1101. Busink, R., & Kuiken, D. (1996). Identifying types of impactful dreams – A replication. Dreaming, 6, 97–119. Carr, L., Iacoboni, M., Dubeau, M. C., Mazziotta, J. C., & Lenzi, G. L. (2003). Neural mechanisms of empathy in humans: A relay from neural systems for imitation to limbic areas. Proceedings of the National Academy of Sciences of the United States of America, 100, 5497–5502. Cartwright, R. D. (1974). The influence of a conscious wish on dreams: A methodological study of dream meaning and function. Journal of Abnormal Psychology, 83, 387–393. Chamberlain, S. R., & Sahakian, B. J. (2007). The neuropsychiatry of impulsivity. Current Opinion in Psychiatry, 20, 255–261. Cheng, Y., Chou, K. H., Decety, J., Chen, I. Y., Hung, D., Tzeng, O. J., et al (2009). Sex differences in the neuroanatomy of human mirror-neuron system: A voxelbased morphometric investigation. Neuroscience, 158, 713–720. Cooper, N. R., Puzzo, I., & Pawley, A. D. (2008). Contagious yawning: The mirror neuron system may be a candidate physiological mechanism. Medical Hypotheses, 71, 975–976. Cooper, N. R., Puzzo, I., Pawley, A. D., Bowes-Mulligan, R. A., Kirkpatrick, E. V., Antoniou, P. A., et al (2012). Bridging a yawning chasm: EEG investigations into the debate concerning the role of the human mirror neuron system in contagious yawning. Cognitive, Affective and Behavioral Neuroscience, 12, 393–405. Crino, M. D., Svoboda, M., Rubenfeld, S., & White, M. C. (1983). Data on the Marlowe–Crowne and Edwards social desirability scales. Psychological Reports, 53, 963–968. Crowne, D. P., & Marlowe, D. (1960). A new scale of social desirability independent of psychopathology. Journal of Consulting Psychology, 24, 349–354. di Pellegrino, G., Fadiga, L., Fogassi, L., Gallese, V., & Rizzolatti, G. (1992). Understanding motor events: A neurophysiological study. Experimental Brain Research, 91, 176–180. Dimberg, U., Andreasson, P., & Thunberg, M. (2011). Emotional empathy and facial reactions to facial expressions. Journal of Psychophysiology, 25, 26–31. Eilam, D., Izhar, R., & Mort, J. (2011). Threat detection: Behavioral practices in animals and humans. Neuroscience and Biobehavioral Reviews, 35, 999–1006. Eisenberg, N., & Lennon, R. (1983). Sex differences in empathy and related capacities. Psychological Bulletin, 94, 100–131. Eiser, A. S., & Schenck, C. H. (2005). Dreaming: A psychiatric view and insights from the study of parasomnias. Schweizer Archiv fur Neurologie und Psychiatrie, 156, 440–470. Enticott, P. G., Kennedy, H. A., Bradshaw, J. L., Rinehart, N. J., & Fitzgerald, P. B. (2011). Motor corticospinal excitability during the observation of interactive hand gestures. Brain Research Bulletin, 85, 89–95. Fadiga, L., Craighero, L., & Olivier, E. (2005). Human motor cortex excitability during the perception of others’ action. Current Opinion in Neurobiology, 15, 213–218. Filimon, F., Nelson, J. D., Hagler, D. J., & Sereno, M. I. (2007). Human cortical representations for reaching: Mirror neurons for execution, observation, and imagery. Neuroimage. Fourkas, A. D., Avenanti, A., Urgesi, C., & Aglioti, S. M. (2006). Corticospinal facilitation during first and third person imagery. Experimental Brain Research, 168, 143–151. Fourkas, A. D., Bonavolonta, V., Avenanti, A., & Aglioti, S. M. (2008). Kinesthetic imagery and tool-specific modulation of corticospinal representations in expert tennis players. Cerebral Cortex, 18, 2382–2390. Geangu, E., Benga, O., Stahl, D., & Striano, T. (2010). Contagious crying beyond the first days of life. Infant Behavior and Development, 33, 279–288. Gerne, M., & Strauch, I. (1985). Psychophysiological indicators of affect patterns and conversational signals during sleep. In W. P. Koella, E. Ruther, & H. Schulz (Eds.), Sleep ’84 (pp. 367–369). Stuttgart: Gustav Fischer Verlag. Guggisberg, A. G., Mathis, J., Schnider, A., & Hess, C. W. (2010). Why do we yawn? Neuroscience and Biobehavioral Reviews, 34, 1267–1276. Hajek, T., Cullis, J., Novak, T., Kopecek, M., Blagdon, R., Propper, L., et al (2013). Brain structural signature of familial predisposition for bipolar disorder: Replicable evidence for involvement of the right inferior frontal gyrus. Biological Psychiatry, 73, 144–152. Hammers, A., Chen, C. H., Lemieux, L., Allom, R., Vossos, S., Free, S. L., et al (2007). Statistical neuroanatomy of the human inferior frontal gyrus and probabilistic atlas in a standard stereotaxic space. Human Brain Mapping, 28, 34–48. Han, S., Fan, Y., & Mao, L. (2008). Gender difference in empathy for pain: An electrophysiological investigation. Brain Research, 1196, 85–93. Helt, M. S., Eigsti, I. M., Snyder, P. J., & Fein, D. A. (2010). Contagious yawning in autistic and typical development. Child Development, 81, 1620–1631. Iranzo, A., & Santamaria, J. (2005). Severe obstructive sleep apnea/hypopnea mimicking REM sleep behavior disorder. Sleep, 28, 203–206. Jola, C., Abedian-Amiri, A., Kuppuswamy, A., Pollick, F. E., & Grosbras, M. H. (2012). Motor simulation without motor expertise: Enhanced corticospinal excitability in visually experienced dance spectators. PLoS ONE, 7, e33343. Kahn, D., & Hobson, J. A. (2005). Theory of mind and dreaming: Awareness of feelings and thoughts of others in dreams. Dreaming, 15, 48–57. Kahn, D., Pace-Schott, E., & Hobson, J. A. (2002). Emotion and cognition: Feeling and character identification in dreaming. Consciousness and Cognition, 11, 34–50. Karacan, I., Goodenough, D. R., Shapiro, A., & Starker, S. (1979). Erection cycle during sleep in relation to dream anxiety. In L. A. Gottschalk (Ed.), The content analysis of verbal behavior (pp. 887–893). New York: Spectrum Pub.. Koulack, D. (1969). Effects of somatosensory stimulation on dream content. Archives of General Psychiatry, 20, 718–725. Kuiken, D., Chudleigh, M., & Racher, D. (2010). Bilateral eye movements, attentional flexibility and metaphor comprehension: The substrate of REM dreaming? Dreaming, 20, 227–247. Lenartowicz, A., Verbruggen, F., Logan, G. D., & Poldrack, R. A. (2011). Inhibition-related activation in the right inferior frontal gyrus in the absence of inhibitory cues. Journal of Cognitive Neuroscience, 23, 3388–3399. Lepage, J. F., Tremblay, S., & Theoret, H. (2010). Early non-specific modulation of corticospinal excitability during action observation. European Journal of Neuroscience, 31, 931–937. Loo, R., & Thorpe, K. (2000). Confirmatory factor analyses of the full and short versions of the Marlowe–Crowne Social Desirability Scale. Journal of Social Psychology, 140, 628–635. 986 T. Nielsen, D. Kuiken / Consciousness and Cognition 22 (2013) 975–986 Lundqvist, L. O. (1995). Facial EMG reactions to facial expressions: A case of facial emotional contagion? Scandinavian Journal of Psychology, 36, 130–141. Mailhot, J. P., Vachon-Presseau, E., Jackson, P. L., & Rainville, P. (2012). Dispositional empathy modulates vicarious effects of dynamic pain expressions on spinal nociception, facial responses and acute pain. European Journal of Neuroscience, 35, 271–278. Mar, R. A., & Oatley, K. (2008). The function of fiction is the abstraction and simulation of social experience. Perspectives in Psychological Science, 3, 173–192. Martin, G. B., & Clark, R. D. (1982). Distress crying in neonates: Species and peer specificity. Developmental Psychology, 18, 3–9. Mazza, S., Soucy, J. P., Gravel, P., Michaud, M., Postuma, R., Massicotte-Marquez, J., et al (2006). Assessing whole brain perfusion changes in patients with REM sleep behavior disorder. Neurology, 67, 1618–1622. McNamara, P., McLaren, D., Smith, D., Brown, A., & Stickgold, R. (2005). A ‘‘Jekyll and Hyde’’ within: Aggressive versus friendly interactions in REM and nonREM dreams. Psychological Science, 16, 130–136. Mestre, M. V., Samper, P., Frias, M. D., & Tur, A. M. (2009). Are women more empathetic than men? A longitudinal study in adolescence. Spanish Journal of Psychology, 12, 76–83. Molenberghs, P., Cunnington, R., & Mattingley, J. B. (2012). Brain regions with mirror properties: A meta-analysis of 125 human fMRI studies. Neuroscience and Biobehavioral Reviews, 36, 341–349. Montagna, M., Cerri, G., Borroni, P., & Baldissera, F. (2005). Excitability changes in human corticospinal projections to muscles moving hand and fingers while viewing a reaching and grasping action. European Journal of Neuroscience, 22, 1513–1520. Musser, E. D., Kaiser-Laurent, H., & Ablow, J. C. (2012). The neural correlates of maternal sensitivity: An fMRI study. Developmental Cognitive Neuroscience, 2, 428–436. Nielsen, T. A. (2007). Felt presence: Paranoid delusion or hallucinatory social imagery? Consciousness and Cognition, 16, 975–983. Nielsen, T. (2011). Disturbed dreaming as a factor in medical conditions. In M. Kryger, T. Roth, & W. C. Dement (Eds.), Principles and practice of sleep medicine (5th ed., pp. 1116–1127). New York: Elsevier. Nielsen, T. A., McGregor, D. L., Zadra, A., Ilnicki, D., & Ouellet, L. (1993). Pain in dreams. Sleep, 16, 490–498. Nielsen, T., & Paquette, T. (2007). Dream-associated behaviors affecting pregnant and postpartum women. Sleep, 30, 1162–1169. Nielsen, T., Svob, C., & Kuiken, D. (2009). Dream-enacting behaviors in a normal population. Sleep, 32, 1629–1636. O’Grady, K. E. (1988). The Marlowe–Crowne and Edwards social desirability scales: A psychometric perspective. Multivariate Behavioral Research, 23, 87–101. Ohayon, M. M., & Schenck, C. H. (2010). Violent behavior during sleep: Prevalence, comorbidity and consequences. Sleep Medicine, 11, 941–946. Oudiette, D., de Cock, V. C., Lavault, S., Leu, S., Vidailhet, M., & Arnulf, I. (2009). Nonviolent elaborate behaviors may also occur in REM sleep behavior disorder. Neurology, 72, 551–557. Oudiette, D., Leu, S., Pottier, M., Buzare, M. A., Brion, A., & Arnulf, I. (2009). Dreamlike mentations during sleepwalking and sleep terrors in adults. Sleep, 32, 1621–1627. Papousek, I., Schulter, G., & Lang, B. (2009). Effects of emotionally contagious films on changes in hemisphere-specific cognitive performance. Emotion, 9, 510–519. Platek, S. M. (2010). Yawn, yawn, yawn, yawn; yawn, yawn, yawn! The social, evolutionary and neuroscientific facets of contagious yawning. Frontiers of Neurology and Neuroscience, 28, 107–112. Premack, D., & Woodruff, G. (1978). Does the chimpanzee have a theory of mind? Behavioral and Brain Sciences, 4, 515–526. Rahko, J., Paakki, J. J., Starck, T., Nikkinen, J., Remes, J., Hurtig, T., et al (2010). Functional mapping of dynamic happy and fearful facial expression processing in adolescents. Brain Imaging and Behaviour, 4, 164–176. Reynolds, W. M. (1982). Development of reliable and valid short forms of the Marlowe–Crowne Social Desirability Scale. Journal of Clinical Psychology, 38, 119–125. Rizzolatti, G., & Craighero, L. (2004). The mirror-neuron system. Annual Review of Neuroscience, 27, 169–192. Roosink, M., & Zijdewind, I. (2010). Corticospinal excitability during observation and imagery of simple and complex hand tasks: Implications for motor rehabilitation. Behavioral Brain Research, 213, 35–41. Rymarczyk, K., Biele, C., Grabowska, A., & Majczynski, H. (2011). EMG activity in response to static and dynamic facial expressions. International Journal of Psychophysiology, 79, 330–333. Saygin, A. P., McCullough, S., Alac, M., & Emmorey, K. (2010). Modulation of BOLD response in motion-sensitive lateral temporal cortex by real and fictive motion sentences. Journal of Cognitive Neuroscience, 22, 2480–2490. Schenck, C. H., Bundlie, S. R., Ettinger, M. G., & Mahowald, M. W. (1986). Chronic behavioral disorders of human REM sleep: A new category of parasomnia. Sleep, 9, 293–308. Schenck, C. H., Lee, S. A., Bornemann, M. A., & Mahowald, M. W. (2009). Potentially lethal behaviors associated with rapid eye movement sleep behavior disorder: Review of the literature and forensic implications. Journal of Forensic Sciences, 54, 1475–1484. Schenck, C. H., & Mahowald, M. W. (2002). REM sleep behavior disorder: Clinical, developmental, and neuroscience perspectives 16 years after its formal identification in SLEEP. Sleep, 25, 120–138. Schulte-Ruther, M., Markowitsch, H. J., Fink, G. R., & Piefke, M. (2007). Mirror neuron and theory of mind mechanisms involved in face-to-face interactions: A functional magnetic resonance imaging approach to empathy. Journal of Cognitive Neuroscience, 19, 1354–1372. Schulte-Ruther, M., Markowitsch, H. J., Shah, N. J., Fink, G. R., & Piefke, M. (2008). Gender differences in brain networks supporting empathy. Neuroimage, 42, 393–403. Schurmann, M., Hesse, M. D., Stephan, K. E., Saarela, M., Zilles, K., Hari, R., et al (2005). Yearning to yawn: The neural basis of contagious yawning. Neuroimage, 24, 1260–1264. Schweickert, R., & Xi, Z. (2010). Metamorphosed characters in dreams: Constraints of conceptual structure and amount of theory of mind. Cognitive Science, 34, 665–684. Sherman, L. W. (1975). An ecological study of glee in small groups of preschool children. Child Development, 46, 53–61. Shimizu, A., & Inoue, T. (1986). Dreamed speech and speech muscle activity. Psychophysiology, 23, 210–214. Spengler, S., von Cramon, D. Y., & Brass, M. (2010). Resisting motor mimicry: control of imitation involves processes central to social cognition in patients with frontal and temporo-parietal lesions. Social Neuroscience, 5, 401–416. Spunt, R. P., & Lieberman, M. D. (2012). An integrative model of the neural systems supporting the comprehension of observed emotional behavior. Neuroimage, 59, 3050–3059. Swann, N. C., Cai, W., Conner, C. R., Pieters, T. A., Claffey, M. P., George, J. S., Aron, A. R., & Tandon, N. (2012). Roles for the pre-supplementary motor area and the right inferior frontal gyrus in stopping action: electrophysiological responses and functional and structural connectivity. Neuroimage, 59, 2860–2870. Tanaka-Matssumi, J., & Kameoka, V. A. (1986). Reliabilities and concurrent validities of popular self-report measures of depression, anxiety, and social desirability. Journal of Consulting and Clinical Psychology, 54, 328–333. Weyers, P., Muhlberger, A., Hefele, C., & Pauli, P. (2006). Electromyographic responses to static and dynamic avatar emotional facial expressions. Psychophysiology, 43, 450–453. Williams, J., Pearce, A. J., Loporto, M., Morris, T., & Holmes, P. S. (2012). The relationship between corticospinal excitability during motor imagery and motor imagery ability. Behavioral Brain Research, 226, 369–375. Zhou, W., & Chen, D. (2009). Sociochemosensory and emotional functions: Behavioral evidence for shared mechanisms. Psychological Science, 20, 1118–1124. Zwaan, R. A., Taylor, L. J., & de Boer, M. (2010). Motor resonance as a function of narrative time: Further tests of the linguistic focus hypothesis. Brain and Language, 112, 143–149.

Consciousness and Cognition 43 (2016) 11–26 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Simulated thought insertion: Influencing the sense of agency using deception and magic Jay A. Olson a,⇑, Mathieu Landry b, Krystèle Appourchaux c, Amir Raz a,d a Department of Psychiatry, 1033 Pine Avenue West, McGill University, Montreal, QC H3A 1A1, Canada Integrated Program in Neuroscience, Montreal Neurological Institute, 3801 University Street, Room 141, Montreal, QC H3A 2B4, Canada c Sciences, Normes, Décision, Université Paris–Sorbonne, Maison de la recherche, 28 rue Serpente, Bureau SE 06, 75006 Paris, France d The Lady Davis Institute at the SMDB Jewish General Hospital, Montreal, QC, Canada b a r t i c l e i n f o Article history: Received 24 October 2015 Revised 17 April 2016 Accepted 25 April 2016 Keywords: Sense of agency Thought insertion Volition Deception Magic Phenomenology a b s t r a c t In order to study the feeling of control over decisions, we told 60 participants that a neuroimaging machine could read and influence their thoughts. While inside a mock brain scanner, participants chose arbitrary numbers in two similar tasks. In the Mind-Reading Task, the scanner appeared to guess the participants’ numbers; in the Mind-Influencing Task, it appeared to influence their choice of numbers. We predicted that participants would feel less voluntary control over their decisions when they believed that the scanner was influencing their choices. As predicted, participants felt less control and made slower decisions in the Mind-Influencing Task compared to the Mind-Reading Task. A second study replicated these findings. Participants’ experience of the ostensible influence varied, with some reporting an unknown source directing them towards specific numbers. This simulated thought insertion paradigm can therefore influence feelings of voluntary control and may help model symptoms of mental disorders. Ó 2016 Elsevier Inc. All rights reserved. 1. Introduction People typically believe that they fully control their thoughts and actions. This belief is often mistaken. People can feel control without having it, such as when subtle situational factors heavily influence decisions (Olson, Amlani, Raz, & Rensink, 2015; Thaler & Sunstein, 2008). Conversely, they can have control without feeling it, such as when under hypnosis or when using a Ouija board (Blakemore, Oakley, & Frith, 2003; Gauchou, Rensink, & Fels, 2012). We present a novel method to influence this feeling. The sense of agency refers to the feeling of control over an action or thought. According to recent theories, this sense of agency has two overlapping components: feeling and judgement (Synofzik, Vosgerau, & Newen, 2008). The feeling refers to a low-level classification of whether an action is caused by oneself; the judgement refers to an analogous higher-level classification. Most theories of agency have focused on the feeling component. The comparator model, for example, claims that this feeling arises by comparing the outcome of an action with the initial intention: if they match, one feels a sense of agency (Frith, 2012). Accordingly, people feel more agency over their hand movements while drawing if the outcome of the drawing matches their intention (Synofzik, Thier, & Lindner, 2006). ⇑ Corresponding author. E-mail address: jay.olson@mail.mcgill.ca (J.A. Olson). http://dx.doi.org/10.1016/j.concog.2016.04.010 1053-8100/Ó 2016 Elsevier Inc. All rights reserved. 12 J.A. Olson et al. / Consciousness and Cognition 43 (2016) 11–26 Some theorists argue that these comparator theories of agency better apply to actions than thoughts (Proust, 2009). If they applied to thoughts, one would have to compare the intention and outcome of a thought, which seems unlikely: one does not intend to have a thought before thinking it (Proust, 2009; Synofzik et al., 2008). Nevertheless, the sense of agency over thoughts varies across situations. When Penfield and Roberts (1976) stimulated the brains of their participants, for example, they reported that thoughts occurred without their control. Further, during pre-sleep states, drug experiences, believed spiritual possessions, and hypnosis, thoughts may seem to originate from an external source (Blakemore et al., 2003; Bourguignon, 1976; Masters & Houston, 1966; Mavromatis, 1987). A theory explaining the sense of agency over thoughts would thus have to account for these situations. One related theory claims that feelings of agency are strongest when (a) a thought closely precedes the action, (b) is coherent with that action, and (c) is the only apparent cause (Wegner & Wheatley, 1999). The last condition – that the thought must be the only apparent cause of the action – is called the principle of exclusivity. This principle may also apply to thoughts: believing in an external source of thoughts may reduce agency. Schizophrenics, for instance, often experience thought insertion in which their thoughts seem to originate from a source outside of their own will (Mullins & Spence, 2014). As a result, some schizophrenics conclude there is an influencing machine that can implant their thoughts from a distance (Tausk, 1969). These distortions in the sense of agency over thoughts can be approximated with hypnosis. Walsh, Oakley, Halligan, Mehta, and Deeley (2015) hypnotised suggestible participants and told them that an engineer would insert thoughts into their heads to complete sentences. When participants heard sentence stems, they reported that other words popped into their heads without their control. In the present feasibility study, we attempted to similarly reduce agency by constructing a plausible external source of thoughts, but without hypnotising participants or stimulating their brains. Instead, to create this source of thoughts, we used deception, suggestion, and magic. Mentalism is a branch of magic that mimics unusual mental phenomena such as telepathy and thought insertion. In the context of a magic show, the audience generally knows these apparent abilities are tricks; in other contexts, they may seem more realistic. Indeed, many students cannot distinguish magic tricks from actual abilities and some believe that neuroimaging machines can read minds (Ali, Lifshitz, & Raz, 2014; Benassi, Singer, & Reynolds, 1980; Swiney & Sousa, 2013). Accordingly, we wanted to use magic to convince people that a neuroimaging machine could influence their thoughts, which would then reduce their sense of agency. Being able to experimentally alter this sense of agency would allow researchers to explore the role of higher-level cognition in feelings and judgements of agency (Gallagher, 2007; Synofzik et al., 2008; Vosgerau & Voss, 2013). It would also demonstrate how much deception and suggestion can affect one’s mental experiences. In this paper, we introduce the simulated thought insertion paradigm, which uses deception and magic to influence the sense of agency over thoughts. Study 1 tests whether this paradigm can distort feelings of agency; Study 2 replicates our findings and examines what these distortions feel like experientially. Combined, these studies offer a novel paradigm to study agency by making people believe – and feel – that we are controlling their minds. 2. Study 1: Influencing agency We introduced participants to a brain imaging machine that could ostensibly influence thoughts. We had three hypotheses. First, when people believe a machine is influencing them, they will report less voluntary control over their mental decisions. Second, this apparent influence will affect the decision-making process, reflected by how quickly people make decisions and how often they change their mind. Third, people who tend to feel that external sources influence their lives (i.e., those with an external locus of control; Duttweiler, 1984) will be more suggestible and thus more likely to feel the influence of the machine (Burger, 1981). In short, we expected that manipulating beliefs would cause distortions in feelings and judgements of agency. 2.1. Methods 2.1.1. Participants Thirty-seven undergraduate students from McGill University completed the experiment for course credit; after exclusions, 27 participants remained. They were on average 20.5 years old (SD ¼ 1:8) and all were female. Most of them majored in psychology (78%) and were in the second year of their studies (50%). We chose our sample size based on a power analysis (see Section 2.1.4) while aiming to run as many participants as possible in Studies 1 and 2 over two months. 2.1.2. Procedure Participants completed two comparable tasks (Fig. 1) inside a mock neuroimaging scanner. In the Mind-Reading Task, participants chose arbitrary numbers and the machine appeared to guess them. In the Mind-Influencing Task, the machine chose random numbers and appeared to influence participants to choose them. After each task, we measured the participants’ sense of agency over their decisions. Because we used a high level of deception, a detailed description of the protocol follows; however, readers can skip it without loss of clarity. J.A. Olson et al. / Consciousness and Cognition 43 (2016) 11–26 13 Fig. 1. Overview of tasks. Participants completed 6 trials (3 with feedback) of each task. 2.1.2.1. Briefing (20 min). Each participant met the experimenter (J.O.) at a cognitive neuroscience laboratory in the Montreal Neurological Institute of McGill University. The experimenter asked the participant if she1 had heard of the Neural Activation Mapping Project (as if it were well-known). He explained that the goal was to ‘‘map neural activation patterns onto thoughts and thoughts onto neural activation patterns.” First, during ‘‘calibration,” the participant would think of numbers in the scanner to reportedly map them to her neural activity. The experimenter explained that if this calibration was successful, the participant would complete the Mind-Reading Task. Here, she would silently choose a number while inside the scanner. Based on her brain activity, the machine would then try to infer which number she had chosen. ‘‘As you probably know,” the experimenter explained, ‘‘the machine can only guess the most basic of thoughts: the numbers we calibrated.” This phrasing – ‘‘as you probably know” – was intended to make the statement appear obvious rather than implausible. Indeed, many participants verbally affirmed the statement. The experimenter explained that if the Mind-Reading Task was successful, the participant would complete the MindInfluencing Task. Here, the machine would randomly choose a number. Then, the participant would choose one herself while the machine (ostensibly) tries to influence her to select the pre-chosen number. The machine would accomplish this influence by manipulating the ‘‘natural electromagnetic fluctuations in the brain.” The experimenter assured the participant that the procedure is safe and has no known side effects. If the participant questioned these fluctuations, the technician would give increasingly complex explanations (of standard fMRI functioning) until the participant stopped asking. The participant then read the consent form. Three (male) participants thought they would feel claustrophobic in the scanner and discontinued the study (see Table 1). The rest of the participants completed a ‘‘safety screening” questionnaire. It began with demographic information, followed by a subset of an MRI screening questionnaire,2 then the Internal Control Index (see Section 2.1.3; Duttweiler, 1984). The MRI questionnaire functioned to distract participants from our interest in the Internal Control Index. No participants asked about the latter questionnaire but several mentioned during debriefing that they did not see its relevance. The experimenter checked the questionnaires as if to ensure the participant passed the screening. He then led her to speak with the technician in the control room. The room contained computers showing images of brain scans (Fig. 2a) to increase credibility (cf. McCabe & Castel, 2008). Beside the computers sat a printer for the machine’s output. A monitor above the computers displayed a live video of the adjacent scanner room. The technician (M.L.) then summarised the tasks that the participant would complete and showed how she would later record her chosen numbers. Namely, she would write her number and sign her initials on the machine’s printed output ostensibly to keep an official record of her choices.3 The experimenter then asked the participant to remove any metals from her person. The experimenter and participant entered the scanner room, which contained warning signs about the machine’s magnet. The room contained a mock MRI scanner (Fig. 2b; Psychology Software Tools, Inc., Sharpsburg, PA). 2.1.2.2. Calibration (10 min). The experimenter then explained the calibration procedure. The participant would concentrate on the number that the technician would announce through the intercom. He told her to stay still during the scan, keep her arms and legs uncrossed, and keep her eyes closed (ostensibly to reduce visual cortex noise). 1 All participants were female in our final sample; three male participants chose not to begin the study (see Table 1). This questionnaire came from the Functional MRI Laboratory at the University of Michigan. 3 In reality, this would later give us a plausible reason to show the participant what number the machine had ostensibly guessed. 2 14 J.A. Olson et al. / Consciousness and Cognition 43 (2016) 11–26 Table 1 Sample size and reasons for exclusion. Number 40 3 =37 1 3 6 =27 Note Recruited Did not begin study due to claustrophobia Completed experiment Missed one questionnaire Knew the experimenter was a magician Expressed scepticism or partially guessed our hypothesis Final sample Fig. 2. Control room (A) and scanner room (B). The participant reclined on the table and entered the machine. The machine then made a louder noise (a pre-recorded MRI sound) as if it were working. Over the next five minutes, the technician told the participant to think of specific numbers from 1 to 9. He would pause for a variable amount of time between the numbers and would ask her to think of the numbers 4 and 7 again. The variable pausing and occasional repeats were intended to make the calibration process seem more realistic. 2.1.2.3. Mind-Reading Task (10 min). After calibration, the experimenter described the Mind-Reading Task. The participant would think of any two-digit number (i.e., from 10 to 99) that had no special significance to her (e.g., not her birthday or favourite number). The experimenter stressed that she should think of the number as a unit (e.g., 42) rather than as two digits (e.g., 4 and 2).4 He told the participant she would have 10 s to choose a number inside the scanner. Namely, she would hear a beep, followed by 10 s during which she would choose a number, then another beep, after which she should not change her mind. During the 10 s period, she could change her mind as many times as she wanted. Whenever she thought of a number, or changed her mind about it, she would click a mouse button inside the scanner. The experimenter then explained that afterwards, the machine would examine her brain recording to infer which number she chose. The technician would view this inferred number and write it on a sheet of paper containing the machine’s output, which the experimenter would bring into the scanner room for verification. To begin the Mind-Reading Task, the participant entered the machine. The technician repeated the instructions then started the recording. The machine made louder noises, then beeped, continued its noises for 10 s while the participant silently chose a number, then beeped again and quieted down. The participant exited the scanner. After a few seconds, the technician passed a clipboard containing the machine output to the experimenter, who then asked the participant which number she chose. After her response, he showed her the output to reveal which number the machine had guessed (Fig. 3). The guess matched the participant’s decision. (Unknown to the participant, the experimenter was a magician and the correct guess was accomplished using a magic trick, described later.) The rest of the output contained technical-looking but nonsensical statistics. The participant wrote her chosen number on the paper, which the experimenter returned to the technician. The technician pretended to examine the output then announced whether more trials were needed. This process continued for two more trials, during which the participant chose different numbers. As is common in mentalism, to make the procedure look more realistic, the machine made occasional errors (Benassi et al., 1980; Lamont & Wiseman, 2005). Specifically, on the first trial, the machine’s number was correct but reversed: if the participant chose 42, the machine would guess 24.5 ‘‘Sometimes it takes a few tries to get it,” the experimenter would casually comment. 4 The same example number (42) was given to all participants in case it served as an anchor to influence later decisions. The experimenter also used 10, 20, and 30 as example numbers when explaining that the participant could change her number. 5 If she chose two repeated digits (e.g. 44), the guess would be off by two (e.g. 46). J.A. Olson et al. / Consciousness and Cognition 43 (2016) 11–26 15 Fig. 3. Fake output from the machine. The participant wrote down her chosen number (bottom) and saw that it matched what the technician had written down. ‘‘Try to think of the numbers as a unit rather than individually,” he would suggest, as if her method of concentrating caused the error. While the participant was writing her number, the experimenter notified the technician of the error (‘‘We got a reversal”) through the intercom. This delay gave the participant time to examine the output – and the machine’s guess – more carefully without having to overtly draw her attention to it. As some performers claim, ‘‘When a magician lets you notice something on your own, his lie becomes impenetrable” (Teller, 2012, p. 2). The magic trick used a common method in mentalism (see Corinda, 1968, ‘‘Step One” for a full explanation), though several methods can accomplish the same effect (cf. Rensink & Kuhn, 2014).6 In essence, whichever number the participant reports ends up appearing as if the machine had guessed it earlier. Importantly, the participant could choose any number she wanted: she was not influenced. The method had a high success rate and failed on only 1 of 162 trials. When the machine guessed the numbers correctly, participants acted surprised. Several shrieked, some laughed to themselves, many expressed amusement (one stated, ‘‘Awesome possum!”), some showed confusion (‘‘Interesting. . .”), and one expressed discomfort (‘‘I don’t like this – this is weird”). Two claimed that it is ‘‘like magic”; during debriefing they confirmed they were referring to supernatural magic rather than magic tricks. (None demonstrated knowledge that a magic trick was involved, even during debriefing.) The experimenter validated the participant’s amusement but acted as if these results were common and uninteresting. After three trials with number feedback from the machine, the participant completed three more, during which she stayed inside the scanner.7 After each trial, she would state her number but would not receive feedback about whether it matched the machine’s guess. This lack of feedback made the last three trials preceding the sense of agency measurement identical to the last three of the upcoming Mind-Influencing Task. Thus, any between-task differences in agency were unlikely caused by merely procedural differences between the tasks. The participant then exited the machine and completed two paper questionnaires while alone in the room. The first questionnaire was a distractor which measured stress during the task. The second measured how much volition and effort par- 6 Researchers can contact the authors for more details concerning the magic trick. We used a small number of trials (6) due to time constraints, but this limited the precision of some of our estimates (e.g., reaction time). We also suspected that adding more feedback trials would help participants figure out the deception (cf. Danek, Fraps, von Müller, Grothe, & Ollinger, 2013); however, similar studies have used more or longer trials (Klock, Voss, Weichenberger, Kathmann, & Kühn, in preparation; Swiney & Sousa, 2013) without problematic amounts of suspicion. 7 16 J.A. Olson et al. / Consciousness and Cognition 43 (2016) 11–26 ticipants felt during their decisions (see Section 2.1.3). One participant did not complete all of the questionnaires and so was excluded (see Table 1). 2.1.2.4. Mind-Influencing Task (10 min). Next, the experimenter explained the Mind-Influencing Task, in which the machine would select a random number then influence the participant to choose it.8 At the beginning of each trial, the computer ostensibly generated a random two-digit number which the technician wrote on a sheet of paper in a clipboard. The experimenter brought the closed clipboard into the scanner room in view of the participant. He explained that the participant would again enter the scanner and choose a two-digit number; afterwards, he would check to see if it matched the machine’s randomly selected number in the clipboard. The rest of the instructions matched the Mind-Reading Task. The participant entered the scanner and chose a number. After she exited the machine, the experimenter asked her what number she chose. He then opened the clipboard to check whether it matched the machine’s selected number. On the first trial, it was off by two9; on the second and third, it matched exactly. Again, the participant could choose any number she wanted, and the matching used the same magic trick as before (Corinda, 1968). After these three trials, the participant would complete three more without receiving feedback. She would then complete the same stress and agency questionnaires as before. The entire procedure took one hour. We call this the simulated thought insertion paradigm.10 2.1.2.5. Debriefing (10 min). The experimenter led the participant out of the scanner room for debriefing. He asked her how she found the study. The most common responses were ‘‘cool” and ‘‘interesting.” One claimed it was ‘‘like X-men.” Two reported neutral evaluations (‘‘okay”) and none were negative. The experimenter then asked if the participant felt any difference between the Mind-Reading and Mind-Influencing Tasks, which we discuss in the Results (Section 2.2.5). As a manipulation check, the experimenter then said, ‘‘There’s something we haven’t told you about the experiment yet. Can you guess what it is?” (cf. Mills, 1976). We wanted to exclude participants who doubted whether the task was possible, since we were interested in those who believed the machine could control their thoughts. In total, nine participants showed some suspicion about the task (see Table 1). Three of these were previous students of the experimenter who knew he was a magician and studied deception. The other six showed some suspicion, such as stating that the machine’s task was impossible. Our exclusion criterion was likely liberal: some participants may not have been suspicious until we asked. We chose this criterion before looking at the data and the exclusions did not change any decisions about the null hypotheses in the results (cf. Simmons, Nelson, & Simonsohn, 2011). The experimenter then gave a partial debriefing and explained a complete debriefing would occur after running all of the participants. We continued to conceal the fact that a magic trick was used and that the brain scanner was not functional. After completing data collection we fully debriefed the participants. The experiment conformed to the guidelines of the Jewish General Hospital Research Ethics Committee. 2.1.3. Measures 2.1.3.1. Agency. We were primarily interested in participants’ sense of agency during the tasks. We measured it with the Sense of Agency Rating Scale, a paper questionnaire with 7-point Likert scales (Polito, Barnier, & Woody, 2013). We used a 10-item subset of the questionnaire measuring involuntariness and effortlessness (see Table 2 in Polito et al., 2013). Involuntariness refers to the amount of internal voluntary control felt over the choice of number (e.g., ‘‘I felt that my experiences and decisions were not caused by me”). Effortlessness refers to the perceived active effort in the decision, compared to passive absorption (e.g., ‘‘My experiences and decisions occurred effortlessly”). We used a modified version of the questionnaire that focused on decisions to make it relevant to the tasks (Polito, personal communication, 2014). The involuntariness subscale usually has high internal consistency reliability (Cronbach’s a ¼ :907; Polito et al., 2013); it was similar in our sample (.90; Mind-Reading: .83, Mind-Influencing: .88). Effortlessness usually has lower reliability (Cronbach’s a ¼ :734); it was substantially lower in our sample (.47; Mind-Reading: .36, Mind-Influencing: .54). The latter reliability is insufficient by most guidelines (Peterson, 1994) so we avoid drawing strong conclusions about effortlessness. 2.1.3.2. Response time and choice. Our secondary dependent variables were response time, number choice, and how many decisions were made during the number selection (i.e., how often participants changed their number each trial). Participants clicked a button when they chose a number and each time they changed their mind about it during the 10-s window of number selection inside the scanner. The response time was the duration between the machine’s initial recording beep (signalling that the participant should choose a number) and the last button press of the trial. 2.1.3.3. Locus of control. We had one predictor variable, locus of control, which represents how much the participant believes that her actions control her life situation (Rotter, 1966). We measured locus of control with the Internal Control Index (Duttweiler, 1984), a 28-item paper questionnaire. An example item is: ‘‘If I want something I always work hard to get it.” Higher scores (up to 140) indicate an internal locus emphasising responsibility and autonomy; lower scores (down to 8 The order of the two tasks was counter-balanced across participants. Or, if the number was off by two in the first trial of the Mind-Reading Task, the guessed number would be reversed. Restated, the machine’s number in the first trial of each task would be incorrect in different ways (viz. off by two or reversed). 10 We avoid the unfortunate acronym ‘‘STI”. 9 J.A. Olson et al. / Consciousness and Cognition 43 (2016) 11–26 17 28) indicate an external locus emphasising luck and fate. The scale has fairly high internal consistency reliability (Cronbach’s a ¼ :84; Duttweiler, 1984); it was similar in our sample (.90). Participants had an average Internal Control Index of 104 (SD ¼ 13:26, range: 69–125), which is expected for their age group and education level (Duttweiler, 1984). 2.1.3.4. Stress. Before the Sense of Agency Rating Scale, participants completed the plausibly relevant Short Stress State Questionnaire as a distractor (post-questionnaire items 1–10 in Helton & Näswall, 2014). Participants felt relaxed during the task with an average stress score of 1.53 out of 5. 2.1.4. Analysis Our main analysis involved the questionnaire data. Since we had a repeated-measures design, we tested for mean between-task differences in involuntariness and effortlessness with paired-sample t tests. Given the results of similar studies looking at self-reported changes in control (e.g., Blakemore et al., 2003; Walsh et al., 2015), we expected large effects (Cohen’s d  1). With an a of .05, we had the statistical power to detect differences of at least 0.65 standard deviations 90% of the time. For effect sizes we used dR , a robust version of Cohen’s d, which quantifies mean differences in standard deviations.11 To test whether locus of control predicted differences in involuntariness or effortlessness, we checked for non-linearity then did a Pearson’s r test. Given our sample size, we had the statistical power to detect only large correlations (r P 0:57) 90% of the time. Our secondary analysis examined response times and number choices. We compared the median response times across tasks using the Wilcoxon–Mann–Whitney test, since the distributions were skewed. We excluded trials in which participants received feedback, leaving three trials per task per participant (162 trials in total). We compared the decision count between the tasks using Poisson regression. In particular, we tried to predict the number of times participants pressed the button (signalling that they chose or changed their number) each trial given the task condition. With the low number of trials, we did not have high statistical power for this test. Finally, we simply plot and describe the number choices themselves. All assumptions were reasonable for the tests conducted,12 and all tests were two-tailed. The analyses used R 3.2.4 (R Core Team, 2015), with packages bootES 1.2 for effect sizes and their bootstrapped confidence intervals (Kirby & Gerlanc, 2013), Hmisc 3.17-2 for bootstrapped confidence intervals of raw variables, coin 1.1-2 for the Wilcoxon–Mann–Whitney test (Hothorn, Hornik, van de Wiel, & Zeileis, 2008), and ggplot2 2.0.0 (Wickham, 2009) for graphs. 2.2. Results and discussion 2.2.1. Involuntariness We predicted that people would feel less control over their thoughts when they believed they were being influenced. As predicted, the involuntariness in the Mind-Influencing Task (M ¼ 16:04 ½13:26; 18:67)13 was higher than in the Mind-Reading Task (M ¼ 9:07 ½7:52; 11:11; Fig. 4a). The majority (81%) of participants showed this pattern. The mean difference score was dR ¼ 0:84 ½0:53; 1:48 standard deviations (t26 ¼ 4:44; p < :001). This large effect is consistent with the principle of exclusivity (Wegner & Wheatley, 1999): people feel less voluntary control when there is a plausible external source of the thought. The Mind-Influencing Task involuntariness scores (M ¼ 16:04) resembled those of medium hypnotisable participants under hypnosis (M ¼ 18:83), but were lower than those of high hypnotisable participants (M ¼ 23:56) or schizophrenic patients (M ¼ 23:26; Polito, Langdon, & Barnier, 2015). 2.2.2. Effortlessness We also predicted that people would feel relatively effortless when choosing numbers in the Mind-Influencing Task. Here we saw little difference: the average effortlessness in the Mind-Influencing Task (M ¼ 28:37 ½26:63; 29:89) resembled that of the Mind-Reading Task (M ¼ 29:74 ½28:48; 31:04, t26 ¼ 1:67, p ¼ :11, dR ¼ 0:32 ½0:72; 0:05; see Fig. 4b). The effortlessness differences were unrelated to the involuntariness differences (r 25 ¼ :01 ½:38; :42, p ¼ :96). Recall that the low reliability of the effortlessness measure makes it difficult to draw conclusions here. 2.2.3. Decision We were also interested in whether participants would differ in their decision process, including their response time, how often they changed their mind about the numbers, and which numbers they chose. Participants took longer to choose their number in the Mind-Influencing Task (Mdn ¼ 5:77 ½5:18; 6:38 s) than in the Mind-Reading Task (Mdn ¼ 3:73 ½3:35; 4:62 s; see Fig. 5a). The median difference was 1.83 ½1:01; 2:7 s (Wilcoxon–Mann–Whitney z ¼ 4:07; p < :001). This slower response time may indicate that the decisions felt less fluent. Indeed, several studies have found a negative correlation between the sense of agency and the feeling of fluency during a decision (Chambon, Sidarus, & Haggard, 2014). An exploratory analysis 11 It equals the 20% trimmed mean divided by the 20% Winsorised standard deviation (Algina, Keselman, & Penfield, 2005). The involuntariness and effortlessness distributions may not appear to meet the assumption of normality, but the distributions of their difference scores do, which we used in our tests. For simplicity we plot the original variables rather than the difference scores. 13 Square brackets throughout denote bootstrapped 95% confidence intervals (Cumming, 2014; Kirby & Gerlanc, 2013). 12 18 J.A. Olson et al. / Consciousness and Cognition 43 (2016) 11–26 A B 35 30 30 ● ● 25 20 ● 15 10 Effortlessness 25 Involuntariness 35 20 15 10 ● 5 5 Mind reading Mind influencing Mind reading Mind influencing Fig. 4. Involuntariness (A) and effortlessness (B) ratings by task. Dots show means, width shows frequency, and error bars show 95% bootstrapped confidence intervals. A B 14 12 7 6 5 8 6 4 ● Decision count Response time (s) 10 4 3 ● 2 2 0 1 Mind reading Mind influencing ● ● Mind reading Mind influencing Fig. 5. Response time for last number choice (A) and count of numbers chosen (B) in each trial by task. Dots show medians, width shows frequency, and error bars show 95% bootstrapped confidence intervals. revealed similar correlations here. People who took longer to respond reported increased involuntariness (r 49 ¼ :34 ½:11; :53, p ¼ :02; Fig. 6a) and decreased effortlessness (r49 ¼ :40 ½:60; :17, p ¼ :003; Fig. 6b). People changed their mind the same number of times during each task (Fig. 5b; Poisson regression z = 0.59, p ¼ :56, odds ratio = 1.07). The median count of numbers chosen each trial was 1. Participants also demonstrated biases towards particular numbers. They tended to choose numbers beginning with 4s (Fig. 7a), perhaps due to the example number given (42) which may have served as an anchor. Indeed, 42 seemed to be a relatively common number on later trials. Perhaps participants tried to suppress that number at first, then it later rebounded (cf. Wegner, 1994), or they simply forgot it was mentioned but it remained primed. Participants also tended to choose smaller numbers, perhaps because all of the example numbers mentioned during the briefing were smaller ones (viz. 10, 20, 30, and 42). For the second digit, people commonly chose 7s and 3s (Fig. 7b). These digits are commonly named when asking people to choose a number (Kubovy, 1977; Kubovy & Psotka, 1976; Olson, Amlani, & Rensink, 2012). Overall, the most common selections were 17 and 19 (each chosen 6 times). 19 J.A. Olson et al. / Consciousness and Cognition 43 (2016) 11–26 A B 12 12 ● ● ● 10 ● ● 8 ● ● ● ● ●● ● ● ● ●● 4 ● ● ● ● ● ● ● 6 ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● 2 ● ● 8 ● ● ● ● ● 6 ●● ● ● ● ● ● 4 ● ● ● ●● ● ● ● ● ● ● ● ● ● 2 ● ● ● ● ●● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ● Response time (s) Response time (s) 10 ● ● ● ● ● 7 14 21 28 35 7 14 Involuntariness 21 28 35 Effortlessness Fig. 6. Correlations between involuntariness (A, r ¼ :34; p ¼ :02), effortlessness (B, r ¼ :40; p ¼ :003), and response time for number choice, across both tasks. Dots show averaged response times for each participant and task. Band shows 95% confidence interval. B 25 25 20 20 Count Count A 15 15 10 10 5 5 1 2 3 4 5 6 First digit 7 8 9 0 1 2 3 4 5 6 7 8 9 Second digit Fig. 7. First (A) and second (B) digits in number choices across both tasks. 2.2.4. Locus of control We also expected that locus of control would predict differences in involuntariness and effortlessness between the tasks. Participants with a more external locus of control (i.e., lower Internal Control Index scores) may be more suggestible (cf. Burger, 1981) and may thus experience larger differences between the tasks. Alas, there were no such relationships between locus of control and differences in involuntariness (r 25 ¼ :23 ½:07; :64, p ¼ :25) or effortlessness (r 25 ¼ :17 ½:21; :43, p ¼ :41). 2.2.5. Experience Just before debriefing, we asked participants if they ‘‘noticed any difference between the tasks”. Most of them (72%, n ¼ 18) did. Only in the Mind-Influencing Task, some participants mentioned unusual experiences. One reported that she did not choose the numbers: they ‘‘just came” to her. Another stated that her eyes ‘‘waved back and forth rapidly” as if ‘‘scanning through many numbers” then stopped when she had one. During pilot testing, one participant said that she had a throbbing headache during the Mind-Influencing Task, perhaps from the machine. After giving her a full debriefing of the 20 J.A. Olson et al. / Consciousness and Cognition 43 (2016) 11–26 deception, she reported that the headache suddenly disappeared. These reports motivated us to explore participants’ experiences with more precision in Study 2. 3. Study 2: Replication and qualitative analysis Having established that we can distort the sense of agency over thoughts, we wanted to replicate these findings (cf. Open Science Collaboration, 2015) and examine the subjective experience associated with the distortions. Few researchers have studied the phenomenology of agency and its relation to behavioural data (Appourchaux, 2014; Vosgerau & Voss, 2013). Here, we attempted to complement the results of Study 1 with descriptions of what participants were thinking and feeling while inside the scanner. We used the elicitation interview which aims to gather first-person descriptions of experience (Vermersch, 1994). This technique emphasises the procedural aspects of the experience (the ‘‘how”) rather than evaluations or theoretical knowledge about it (the ‘‘why”). In doing so, it attempts to reduce post hoc rationalisations which are common when explaining decisions (Johansson, Hall, Sikström, & Olsson, 2005; Nisbett & Wilson, 1977; Petitmengin, Remillieux, Cahour, & Carter-Thomas, 2013). As such, this interviewing technique is well-suited for exploring what participants think and feel throughout the decision-making process. 3.1. Methods 3.1.1. Participants Twenty-three undergraduates completed the experiment for course credit; after exclusions, 20 participants remained. They were on average 20.4 years old (SD ¼ 1:8) and most were female (85%). Many of them (45%) majored in psychology and were in the second (45%) or third year (45%) of their studies. The sample thus resembled that of Study 1. 3.1.2. Procedure The procedure mirrored Study 1 except that we interviewed participants about their experience. After each task, the interviewer (K.A.) entered the scanner room. The interviews began with the suggestion that the participant recall sensory cues, such as the technician announcing the last trial and the machine’s beep, to help re-live the experience of the previous decision in detail. This technique helped the participant reach an evocation state in which the experience of choosing a number would become more salient and easier to recall (Vermersch, 1994). The interviewer would then ask participants to describe their experience. The questions were ‘‘content-free” in that they offered little information and instead prompted the participant to continue describing the experience: Interviewer: I’m going to suggest that you take a little time to go back to the moment when you made your choice. Participant: . . . One number popped in and that was . . . in my head and I just couldn’t change it. Before [in the MindReading Task], I would go between [numbers], but [this] one . . . I just didn’t change, I just kept thinking about it. I: Okay, so, if we go back to the first beep, when you hear the first beep, what happens next? P: I couldn’t think of any number, and then . . . it took maybe two seconds and then the number came. And then 71 popped in, and then it just stayed. I: Okay, so two seconds, then the number 71 pops in. How does it manifest itself? P: It just . . . appeared, like I could see . . . the number. Each interview was recorded and lasted from 3 to 12 min (M ¼ 6 min). After the interview, participants completed the same questionnaires as in the previous study. We excluded three participants using the same criteria as in Study 1. One participant knew the experimenter was a magician and two were suspicious about whether the task was possible. In addition, we excluded a fourth participant from the interview analysis alone due to a malfunction during recording. As in Study 1, none of the exclusions changed the decisions about the null hypotheses. 3.1.3. Analysis We ran two analyses. First, we tested the same hypotheses as in Study 1: namely, in the Mind-Influencing Task, people would feel less agency, make slower decisions, and change their mind more often.14 Second, we did an exploratory thematic analysis of the interview data without any a priori themes (Joffe, 2012). This analysis began with transcribing the interviews then looking for recurring themes, of which we found 33 (see Table A1). We gave the list of themes to four blind judges who independently ascribed them to each participant response in the transcripts. For each participant response, we kept only the themes that at least half of the judges agreed upon (without deliberation). After coding the interviews, we compared the frequencies of the themes between the two tasks. Finally, we searched the transcripts for atypical experiences that may not have been captured by assigning individual themes. 14 We did not statistically compare the results of Study 1 and Study 2 for two reasons. First, we lacked statistical power to detect probable differences (viz., small or medium effects). Second, even if we did detect differences, we could not determine if they were caused by the presence of the interviews or trivial changes in the procedure (e.g., more time between the number decisions and questionnaires in Study 2). In any case, the results of the two studies were similar (see Fig. 8). J.A. Olson et al. / Consciousness and Cognition 43 (2016) 11–26 21 3.2. Results and discussion 3.2.1. Behaviour Most of the results resembled those of Study 1 in direction and magnitude (see Fig. 8). Participants felt their decisions were more involuntary in the Mind-Influencing Task (M ¼ 19:4 ½16; 22:75) compared to the Mind-Reading Task (M ¼ 10:35 ½7:95; 12:75; Fig. 9a) and the effect size (dR ¼ 0:95 ½0:61; 1:9) was similar to that of Study 1. These involuntariness scores again resembled those of medium-hypnotisable participants (Polito et al., 2015). Participants made slower decisions in the Mind-Influencing Task (by 1.65 s; Fig. 9c) and changed their mind as frequently in each task (Fig. 9d). There were no correlations between locus of control and agency measures (Fig. 8). Participants again chose lower numbers (Fig. 9e) but, unlike in Study 1, this time they seemed to show a preference for even numbers (Fig. 9f). Unlike in Study 1, participants may have felt more effortless in the Mind-Reading Task (M ¼ 30:1 ½28:75; 31:6) than in the Mind-Influencing Task (M ¼ 28:1 ½26:7; 29:55; dR ¼ 0:72 ½1:28; 0:09; Fig. 9b). However, the reliability of effortlessness was low so it is difficult to draw conclusions (Cronbach’s a ¼ :42; Mind-Reading: .61, Mind-Influencing: .20). The reliability measures of involuntariness (.92; Mind-Reading: .81, Mind-Influencing: .92) and locus of control (:84) were adequate. 3.2.2. Interviews The most common themes reflected what people visualised, heard, or thought of while choosing their numbers (Table A1). The only major difference between the tasks seemed to be how often participants explained the strategy behind their selections. Although the interviewer never asked why they chose their numbers, in the Mind-Reading Task, more participants provided spontaneous explanations (74%, n ¼ 14) than in the Mind-Influencing Task (16%, n ¼ 3). In the MindReading Task, some (n ¼ 8) mentioned choosing a number different from previous selections, such as a number in the 80s because the participant had not ‘‘chosen 80-something in a while.” Others (n ¼ 3) described the associations with their numbers, such as choosing 52 ‘‘as in 52 weeks in a year” or a number close to the participant’s age. One chose 78 because it was a ‘‘good, strong number.” Two others reported choosing the first or last numbers that came to mind. These reasons contrasted with the few explanations given in the Mind-Influencing Task. One participant reported wanting to choose a different number than before; one felt there was nothing ‘‘trying to distract [him] from this number”; and one chose the last number that came to mind. This difference in explanations between the tasks is consistent with theories of agency. Being able to construct a coherent narrative explaining a decision may correlate with agency judgements (Synofzik et al., 2008; Vosgerau & Voss, 2013). Given that the elicitation interview aims to reduce post hoc rationalisations, this difference in explanations may reflect different cognitive strategies used to choose the numbers. Indeed, in the MindReading Task, an exploratory test showed that those who explained their decisions took longer to choose than those who did not (Mdndiff ¼ 1:65 ½0:33; 2:85 s). There were also unusual experiences. In the Mind-Influencing Task, participants reported a variety of agency-related phenomena. Some claimed that the decision did not feel like it was their own: ‘‘It didn’t really feel like I was making the choice – it kind of just happened.” Another reported, ‘‘I kind of felt like [the number] came out of nowhere. So I felt like it . . . wasn’t my choice. I don’t know why I chose it.” Others claimed they could not change the number: ‘‘I thought about trying to change [the number], but then . . . it doesn’t.” Another agreed: ‘‘I almost can’t think of another one.” Some reported feeling that the number came from a source beyond their control. Sometimes the source was the participant’s apparently disobedient brain. One claimed, ‘‘I was going . . . with 34, but my brain just told me no, that’s not the number, so I went. . .: 32, 33, 31, 30.” Another felt that ‘‘it’s almost like my brain is shuffling through numbers until it . . . stuck to one.” Some suggested the source was the machine: ‘‘I can’t put my finger on it – it’s just like . . . once the magnet turned on . . . I got 4, and then I got 7. . .” For others, the source was unknown: ‘‘I felt like I was drawn to [the number].” Another said, ‘‘it really just felt like [the number] kind of came to me [from] somewhere else.” One participant claimed: I feel like it’s a voice . . . dragging me from the number that already exists in my mind. I . . . feel some kind of force, or some kind of . . . image, or [something] trying to distract me from this number, and then I form [another] number. Some participants also felt unusual sensations. One commented, ‘‘I pretty much felt like it was in my brain, . . . I really noticed a kind of a pulsation, . . . almost physical.” Another noted, ‘‘I don’t know why but my face feels really hot, like my head is really hot, but the rest of my body doesn’t feel anything.” In contrast, participants reported almost no unusual experiences or sensations in the Mind-Reading Task. These results match a model of thought insertion which claims that normal thoughts with a reduced sense of agency may feel like inserted ones (Humpston & Broome, 2015). These thoughts often co-occur with ‘‘delusional elaboration,” in which the person tries to identify the source of the thoughts. Similarly, with a reduced sense of agency, some of our participants provided unusual explanations regarding the source of the number decision (see also Swiney & Sousa, 2013). Our paradigm may thus offer a novel way to test these theories of agency. The apparent differences in experience between the two tasks may have been complicated by participants’ intuitive metaphysics. Rose, Buckwalter, and Nichols (2015) found that when people read stories about machines that can predict and influence behaviour, they still interpret these scenarios with the assumption of a free will. In the present study, when participants experienced such scenarios, their probable assumption of a free will may have biased their interpretations and 22 J.A. Olson et al. / Consciousness and Cognition 43 (2016) 11–26 Involuntariness (robust Cohen's d) ● ● ● Effortlessness (robust Cohen's d) ● ● ● Involuntariness and effortlessness correlation (Pearson's r) ● ● ● Locus of control and involuntariness correlation (Pearson's r) Study ● ● ● ● 1 ● 2 ● Combined Locus of control and effortlessness correlation (Pearson's r) ● ● ● Response time (median difference in s) ● ● ● Decision count (odds ratio) ● ● ● −1 0 1 2 3 Fig. 8. Mind-Influencing (compared to Mind-Reading) Task effect sizes between studies. The only inconsistencies between studies involved effortlessness, which had low reliability. Solid vertical lines show null-hypothesised values and error bars show 95% confidence intervals (bootstrapped except for Decision count and Response time). Correlations of involuntariness and effortlessness refer to between-task differences, not the original variables. For the robust Cohen’s d, see Algina et al. (2005). reports. These individual differences may partly account for why some participants reported little difference between the tasks while others reported unusual experiences such as voices influencing their decisions. This exploratory study had several limitations. First, the interviews may have affected participants’ subsequent questionnaire responses, for example by making them more prone to introspection (cf. Petitmengin et al., 2013). The results may therefore be less reliable than those of Study 1. Second, our small sample size prevented us from categorising participants based on their decision-making strategies or experiences (cf. Lutz & Thompson, 2003). Overall, in this study we were again able to influence feelings of agency, but people experienced these influences in a variety of ways. 23 J.A. Olson et al. / Consciousness and Cognition 43 (2016) 11–26 A B 35 25 20 ● 15 ● 10 ● 25 20 15 10 5 5 Mind reading C ● 30 Effortlessness Involuntariness 30 35 Mind influencing Mind reading D 14 Mind influencing 5 4 10 Decision count Response time (s) 12 8 6 ● 4 ● 3 2 2 0 E Mind influencing F 20 ● Mind influencing 20 15 Count 15 Count ● Mind reading 1 Mind reading 10 5 10 5 1 2 3 4 5 6 First digit 7 8 9 0 1 2 3 4 5 6 7 8 9 Second digit Fig. 9. Study 2 results. Involuntariness (A) and effortlessness (B) ratings by task, response time for number choice (C) and count of numbers chosen (D) in each trial by task, and first (E) and second (F) digits in number choices across both tasks. Compare to Figs. 4, 5 and 7. 4. General discussion We introduced and tested the simulated thought insertion paradigm, in which we convinced people that a machine was controlling their thoughts. In doing so, we were able to distort their sense of agency and slow their decision-making process. Their experience of the illusory influence varied, with some reporting a voice or unknown source directing them towards a particular choice. Our findings demonstrate the potency of top-down influences on feelings of agency. Here, participants chose arbitrary numbers after we altered their beliefs about the source of the choice (viz. self-chosen or machine-chosen). Consistent with theories of agency, this change in belief affected the fluency of the decisions (as seen by response time; Chambon et al., 2014), the narratives used to explain them (Bayne & Pacherie, 2007; Synofzik et al., 2008), and the feelings of voluntary control over them.15 15 The data sets (questionnaires, response times, and number choices) are available online at https://osf.io/uc73j/. 24 J.A. Olson et al. / Consciousness and Cognition 43 (2016) 11–26 These results add to the body of evidence suggesting a double dissociation between the sense of agency and actual agency: between the feeling of control and actual control (Wegner, 2003). Consistent with other studies (e.g., Hall, Johansson, & Strandberg, 2012; Johansson et al., 2005), we have previously shown that people can feel control over their decisions without having control (Olson et al., 2015); the present study demonstrates that people also can have control without feeling it. Our paradigm offers a novel way to study this puzzling dissociation. More generally, this paradigm demonstrates the usefulness of magic in experiments. Researchers have previously used magic as a tool to study suggestion (Wilson & French, 2014), decision-making (Johansson et al., 2005), attitude change (Hall et al., 2012), and beliefs (Subbotsky, 2004). In our paradigm, magic allowed us to simulate thought insertion and to demonstrate that this procedure worked. Without it, participants likely would have had less confidence in the procedure (cf. Swiney & Sousa, 2013) which may have reduced our effects. Paradigms similar to ours have been used to model symptoms of mental disorders. For example, researchers have used hypnosis to create the appearance of delusions such as thought insertion in healthy individuals (Connors, 2015; Walsh et al., 2015; Woody & Szechtman, 2011). This method enables researchers to explore such symptoms with a high level of control in more accessible populations (Connors, 2015). Our study demonstrates that some of these symptoms can be partly reproduced with magic tricks rather than with hypnosis. Given that mentalism magic provides the illusion of several unusual mental abilities central to many delusions (e.g., telepathy, psychokinesis, clairvoyance), using magic to simulate these delusions could help model symptoms of mental disorders. This method has several benefits over using hypnosis: it removes the need for time-intensive hypnosis screening and may be more effective on a larger proportion of people.16 However, our paradigm may produce weaker or shorter effects than hypnosis with highly hypnotisable participants (cf. Polito et al., 2015), may be less effective on suspicious individuals, and requires expertise in magic. Further, it is currently unclear whether related paradigms can generalise to thoughts beyond numbers or words (Klock et al., in preparation; Swiney & Sousa, 2013). Our study had other limitations involving experimental control. First, although we compared results between tasks, we did not have a separate control task; for example, we could have had participants choose numbers while the machine was turned off. We assumed that the Mind-Reading Task would not affect agency and therefore could serve as a baseline; as such, we excluded a separate control task, which limited some of our conclusions. Second, we could not isolate the causal role of the magic trick compared to the rest of the deception. Future studies could contrast these by presenting the deception without the magic trick and vice versa. Comparing these conditions would help isolate the source of the changes in agency. Third, perhaps the largest limitation of the study is that participants may have inferred that we were looking for differences between the two tasks and responded accordingly. We attempted to reduce this possibility by giving a cover story, using distractor questionnaires, excluding suspicious participants liberally, and measuring different kinds of variables (subjective and objective). The response times in particular may have been less susceptible to demand characteristics given that we lacked directional predictions about them; participants could not have inferred our hypotheses here and behaved accordingly. Nevertheless, we still saw differences in response times in a consistent direction. Given that we examined how beliefs about the task affect agency, it would have been implausible to reduce demand characteristics by keeping participants blind to their task condition. Indeed, the goal of the paradigm was to create a context in which participants would knowingly disown their thoughts and attribute them to a machine. Our paradigm opens several possibilities for research. If people believe a machine can read their minds, would they more likely tell the truth in a scanner? If people believe it can influence thoughts, would they also believe – and experience – ostensibly implanted feelings or judgements? Do people feel different levels of agency based on the content or emotion of these implanted thoughts (Swiney & Sousa, 2013)? Does simulated thought insertion reduce people’s belief in free will and change their attitudes or behaviour (cf. Vohs & Schooler, 2008)? By adapting established techniques from magicians and deceivers, we hope to enable new experimental methods that can answer these questions. In doing so, we aim to uncover why people so strongly feel that they control their thoughts and actions – even when they do not. Acknowledgements We would like to thank Ilan Goldberg for helping with the design; Josh Laxer and Moriah Stendel for transcribing the interviews; Jason Da Silva Castanheira, Kylar D’Aigle, Madalina Prostean, and Léah Suissa-Rocheleau for coding the interviews; Angela Shen for entering data; Mélanie Bolduc for taking the photos; and Patrick Haggard, Leonie Klock, Michael Lifshitz, Mark Mitton, Claire Petitmengin, Anne Remillieux, Ronald Rensink, Diana Sitoianu, Thomas Strandberg, and three reviewers for helpful comments. J.O. acknowledges the Joseph-Armand Bombardier Scholarship from the Social Sciences and Humanities Research Council of Canada, and funding from the Bial Foundation and the Desjardins Foundation; M.L. acknowledges the Alexander Graham Bell Canada Graduate Scholarship from the Natural Sciences and Engineering Research Council of Canada (NSERC); and A.R. acknowledges the Canada Research Chair program, Discovery and Discovery Acceleration Supplement grants from NSERC, and Canadian Institutes of Health Research. 16 Future studies, though, may benefit by screening participants for schizotypy or mental disorders; simulated thought insertion might be uncomfortable for some individuals. We had not thought of this concern before conducting the study and our sample showed no signs of distress. 25 J.A. Olson et al. / Consciousness and Cognition 43 (2016) 11–26 Appendix A. Interview themes Table A1 lists the interview themes. Table A1 Interview analysis themes, descriptions, and percent of participants reporting them in each task. Only the ‘‘Strategy” theme seemed to show a large difference. Theme Description Mind-Reading % Mind-Influencing % See Strategy Repeat Say Hear Multiple Happened Digits Fast Own Colour Flip Appeared One Feel Random Focus Uncertain NotSee Stuck Write Slow Right Easy Brain NotOwn Pattern Suggestion Certain Confused Suspicious Bold Body Saw the number (in mind’s eye) Explained the reason or strategy for the number choice Thought of number over and over Said number in head Heard number in head Thought of more than one number Number choice ‘‘just happened” seemingly on its own Chose or thought of the two digits separately Number quickly came to mind Felt like participant’s own decision Mentioned the colour of the number Number rapidly changed or ‘‘flipped by” Number just appeared or popped in head Thought of only one number Mentioned what it felt like whilst choosing the number Chose the number randomly Tried to concentrate on the number Did not feel certain about the number Mentioned not seeing the number Number stuck in the head or was hard to change Thought of or visualised writing the number Number slowly came to mind Wondered whether number matched machine’s Choice felt easy Mentioned the brain Did not feel like own choice Numbers followed a pattern (e.g., increasing) Thought she felt something because we told her to Number choice felt certain Seemed confused Seemed suspicious about the experiment Number visually appeared thick or bold Mentioned a bodily sensation 74 74 63 63 58 53 47 42 42 42 37 37 37 37 32 32 32 32 32 21 21 16 16 16 11 11 11 11 5 5 5 5 0 84 16 58 58 37 74 53 37 37 26 53 32 26 26 42 32 21 21 16 26 21 26 16 5 16 16 16 5 11 5 5 0 16 References Algina, J., Keselman, H. J., & Penfield, R. D. (2005). An alternative to Cohen’s standardized mean difference effect size: A robust parameter and confidence interval in the two independent groups case. Psychological Methods, 10(3), 317–328. http://dx.doi.org/10.1037/1082-989X.10.3.317. Ali, S., Lifshitz, M., & Raz, A. (2014). Empirical neuroenchantment: From reading minds to thinking critically. Frontiers in Human Neuroscience. http://dx.doi. org/10.3389/fnhum.2014.0035. Appourchaux, K. (2014). Un nouveau libre arbitre. Paris: CNRS. Bayne, T., & Pacherie, E. (2007). Narrators and comparators: The architecture of agentive self-awareness. Synthese, 159(3), 475–491. Benassi, V. A., Singer, B., & Reynolds, C. B. (1980). Occult belief: Seeing is believing. Journal for the Scientific Study of Religion, 19(4), 337–349. http://dx.doi. org/10.2307/1386128. Blakemore, S. J., Oakley, D. A., & Frith, C. D. (2003). Delusions of alien control in the normal brain. Neuropsychologia, 41(8), 1058–1067. http://dx.doi.org/ 10.1016/S0028-3932(02)00313-5. Bourguignon, E. (1976). Possession. San Francisco: Chandler & Sharp. Burger, J. M. (1981). Locus of control, motivation, and expectancy: Predicting hypnotic susceptibility from personality variables. Journal of Research in Personality, 15(4), 523–537. http://dx.doi.org/10.1016/0092-6566(81)90048-9. Chambon, V., Sidarus, N., & Haggard, P. (2014). From action intentions to action effects: How does the sense of agency come about? Frontiers in Human Neuroscience, 8(320), 1–9. http://dx.doi.org/10.3389/fnhum.2014.00320. Connors, M. H. (2015). Hypnosis and belief: A review of hypnotic delusions. Consciousness and Cognition, 36, 27–43. Corinda (1968). Thirteen steps to mentalism (third). New York: Tannen Magic. Cumming, G. (2014). The new statistics: Why and how. Psychological Science, 25(1), 7–29. http://dx.doi.org/10.1177/0956797613504966. Danek, A. H., Fraps, T., von Müller, A., Grothe, B., & Ollinger, M. (2013). Working wonders? Investigating insight with magic tricks. Cognition, 130(2), 174–185. http://dx.doi.org/10.1016/j.cognition.2013.11.003. Duttweiler, P. C. (1984). The internal control index: A newly developed measure of locus of control. Educational and Psychological Measurement, 44(2), 209–221. http://dx.doi.org/10.1177/0013164484442004. Frith, C. D. (2012). Explaining delusions of control: The comparator model 20 years on. Consciousness and Cognition, 21(1), 52–54. http://dx.doi.org/10.1016/ j.concog.2011.06.010. Gallagher, S. (2007). The natural philosophy of agency. Philosophy Compass, 2(2), 347–357. http://dx.doi.org/10.1119/1.1969491. Gauchou, H. L., Rensink, R. A., & Fels, S. (2012). Expression of nonconscious knowledge via ideomotor actions. Consciousness and Cognition, 21(2), 976–982. 26 J.A. Olson et al. / Consciousness and Cognition 43 (2016) 11–26 Hall, L., Johansson, P., & Strandberg, T. (2012). Lifting the veil of morality: Choice blindness and attitude reversals on a self-transforming survey. PloS One, 7 (9), e45457. http://dx.doi.org/10.1371/journal.pone.0045457. Helton, W. S., & Näswall, K. (2014). Short stress state questionnaire. European Journal of Psychological Assessment, 1, 1–11. http://dx.doi.org/10.1027/10155759/a000200. Hothorn, T., Hornik, K., van de Wiel, M., & Zeileis, A. (2008). Implementing a class of permutation tests: The coin package. Journal of Statistical Software, 28(8), 1–23. Humpston, C. S., & Broome, M. R. (2015). The spectra of soundless voices and audible thoughts: Towards an integrative model of auditory verbal hallucinations and thought insertion. Review of Philosophy and Psychology, 1–19. http://dx.doi.org/10.1007/s13164-015-0232-9. Joffe, H. (2012). Thematic analysis. In Qualitative research methods in mental health and psychotherapy: A guide for students and practitioners (pp. 147–160). John Wiley & Sons. Johansson, P., Hall, L., Sikström, S., & Olsson, A. (2005). Failure to detect mismatches between intention and outcome in a simple decision task. Science, 310 (5745), 116–119. Kirby, K. N., & Gerlanc, D. (2013). BootES: An R package for bootstrap confidence intervals on effect sizes. Behavior Research Methods, 45(4), 905–927. http:// dx.doi.org/10.3758/s13428-013-0330-5. Klock, Voss, Weichenberger, Kathmann, & Kühn (in preparation). Who is thinking? Neural correlates of an altered sense of authorship of thoughts. Kubovy, M. (1977). Response availability and the apparent spontaneity of numerical choices. Journal of Experimental Psychology: Human Perception and Performance, 3(2), 359–364. http://dx.doi.org/10.1037/0096-1523.3.2.359. Kubovy, M., & Psotka, J. (1976). The predominance of seven and the apparent spontaneity of numerical choices. Journal of Experimental Psychology: Human Perception and Performance, 2(2), 291–294. http://dx.doi.org/10.1037/0096-1523.2.2.291. Lamont, P., & Wiseman, R. (2005). Magic in theory: An introduction to the theoretical and psychological elements of conjuring. Bristol: University of Hertfordshire Press. Lutz, A., & Thompson, E. (2003). Neurophenomenology: Integrating subjective experience and brain dynamics in the neuroscience of consciousness. Journal of Consciousness Studies, 10(9–10), 31–52. Masters, R. E. L., & Houston, J. (1966). The varieties of psychedelic experience. Dell. Mavromatis, A. (1987). Hypnagogia: the unique state of consciousness between wakefulness and sleep. London: Routledge & Kegan Paul. McCabe, D. P., & Castel, A. D. (2008). Seeing is believing: The effect of brain images on judgments of scientific reasoning. Cognition, 107(1), 343–352. Mills, J. (1976). A procedure for explaining experiments involving deception. Personality and Social Psychology Bulletin, 2(1), 3–13. Mullins, S., & Spence, S. A. (2014). Re-examining thought insertion: Semi-structured literature review and conceptual analysis. British Journal of Psychiatry, 182, 293–298. http://dx.doi.org/10.1192/bjp.02.322. Nisbett, R. E., & Wilson, T. D. (1977). Telling more than we can know: Verbal reports on mental processes. Psychological Review, 84(3), 231–259. http://dx.doi. org/10.1037/0033-295X.84.3.231. Olson, J. A., Amlani, A. A., Raz, A., & Rensink, R. A. (2015). Influencing choice without awareness. Consciousness and Cognition, 37, 225–236. http://dx.doi.org/ 10.1016/j.concog.2015.01.004. Olson, J. A., Amlani, A. A., & Rensink, R. A. (2012). Perceptual and cognitive characteristics of common playing cards. Perception, 41(3), 268–286. http://dx.doi. org/10.1068/p7175. Open Science Collaboration (2015). Estimating the reproducibility of psychological science. Science, 349(6251), aac4716. http://dx.doi.org/10.1126/science. aac4716. Penfield, W., & Roberts, L. (1976). Speech and brain-mechanisms. New York: Atheneum. Peterson, R. (1994). A meta-analysis of Cronbach’s coefficient alpha. Journal of Consumer Research, 21(2), 381–391. Petitmengin, C., Remillieux, A., Cahour, B., & Carter-Thomas, S. (2013). A gap in Nisbett and Wilson’s findings? A first-person access to our cognitive processes. Consciousness and Cognition, 22, 654–669. http://dx.doi.org/10.1016/j.concog.2013.02.004. Polito, V., Barnier, A. J., & Woody, E. Z. (2013). Developing the Sense of Agency Rating Scale (SOARS): An empirical measure of agency disruption in hypnosis. Consciousness and Cognition, 22(3), 684–696. http://dx.doi.org/10.1016/j.concog.2013.04.003. Polito, V., Langdon, R., & Barnier, A. J. (2015). Sense of agency across contexts: Insights from schizophrenia and hypnosis. Psychology of Consciousness: Theory, Research, and Practice, 2(3), 301–314. http://dx.doi.org/10.1037/cns0000053. Proust, J. (2009). Is there a sense of agency for thought? In L. O’Brien & M. Soteriou (Eds.), Mental actions. Oxford University Press. R Core Team (2015). R: A language and environment for statistical computing. Rensink, R. A., & Kuhn, G. (2014). A framework for using magic to study the mind. Frontiers in Psychology, 5(1508). http://dx.doi.org/10.3389/ fpsyg.2014.01508. Rose, D., Buckwalter, W., & Nichols, S. (2015). Neuroscientific prediction and the intrusion of intuitive metaphysics. Cognitive Science, 1–21. http://dx.doi. org/10.1111/cogs.12310. Rotter, J. B. (1966). Generalized expectancies for internal versus external control of reinforcement. Psychological Monographs: General and Applied, 80(1), 1–28. Simmons, J. P., Nelson, L. D., & Simonsohn, U. (2011). False-positive psychology: Undisclosed flexibility in data collection and analysis allows presenting anything as significant. Psychological Science, 22(11), 1359–1366. http://dx.doi.org/10.1177/0956797611417632. Subbotsky, E. (2004). Magical thinking in judgments of causation: Can anomalous phenomena affect ontological causal beliefs in children and adults? British Journal of Developmental Psychology, 22, 123–152. Swiney, L., & Sousa, P. (2013). When our thoughts are not our own: Investigating agency misattributions using the Mind-to-Mind paradigm. Consciousness and Cognition, 22(2), 589–602. http://dx.doi.org/10.1016/j.concog.2013.03.007. Synofzik, M., Thier, P., & Lindner, A. (2006). Internalizing agency of self-action: Perception of one’s own hand movements depends on an adaptable prediction about the sensory action outcome. Journal of Neurophysiology, 96, 1592–1601. http://dx.doi.org/10.1152/jn.00104.2006. Synofzik, M., Vosgerau, G., & Newen, A. (2008). Beyond the comparator model: A multifactorial two-step account of agency. Consciousness and Cognition, 17 (1), 219–239. http://dx.doi.org/10.1016/j.concog.2007.03.010. Tausk, V. (1969). On the origin of the ‘‘influencing apparatus” in schizophrenia (1919). Psyche, 23(5), 354–384. Teller (2012). Teller reveals his secrets. Smithsonian. Thaler, R. H., & Sunstein, C. R. (2008). Nudge: Improving decisions about health, wealth, and happiness. Yale University Press. Vermersch, P. (1994). L’entretien d’explicitation.Paris: ESF. Vohs, K. D., & Schooler, J. W. (2008). The value of believing in free will: Encouraging a belief in determinism increases cheating. Psychological Science, 19, 49–54. Vosgerau, G., & Voss, M. (2013). Authorship and control over thoughts. Mind & Language, 29(5), 534–565. http://dx.doi.org/10.1111/mila.12065. Walsh, E., Oakley, D. A., Halligan, P. W., Mehta, M. A., & Deeley, Q. (2015). The functional anatomy and connectivity of thought insertion and alien control of movement. Cortex, 64, 380–393. http://dx.doi.org/10.1016/j.cortex.2014.09.012. Wegner, D. M. (1994). Ironic processes of mental control. Psychological Review, 101(1), 34–52. http://dx.doi.org/10.1037/0033-295X.101.1.34. Wegner, D. M. (2003). The illusion of conscious will. MIT Press. Wegner, D. M., & Wheatley, T. (1999). Apparent mental causation. American Psychologist, 54(7), 480–492. Wickham, H. (2009). Ggplot2: Elegant graphics for data analysis. Springer. Wilson, K., & French, C. C. (2014). Magic and memory: Using conjuring to explore the effects of suggestion, social influence and paranormal belief on eyewitness testimony for an ostensibly paranormal event. Frontiers in Psychology, 5(1289). http://dx.doi.org/10.3389/fpsyg.2014.01289. Woody, E. Z., & Szechtman, H. (2011). Using hypnosis to develop and test models of psychopathology. The Journal of Mind-Body Regulaton, 1(1), 4–16.

Consciousness and Cognition Consciousness and Cognition 14 (2005) 474–482 www.elsevier.com/locate/concog The effect of unconscious priming on temporal production q Fuminori Ono *, Jun-ichiro Kawahara Department of Psychology, Hiroshima University, 1-1-1 Kagamiyama, Higashi-Hiroshima 739-8524, Japan Received 15 April 2004 Available online 5 March 2005 Abstract We examined the effects of unconscious priming on temporal-interval production. In Experiment 1, participants were instructed to keep visual displays on a screen for 2500 ms intervals. Half of the displays were repeated across blocks throughout the entire experiment, and the others were newly generated from trial to trial. The displays consisted of patterns so complex that the participants could not intentionally memorize them. The results showed that significantly more time elapsed for old displays than for new displays before participants indicated that a 2500 ms interval had elapsed. Experiment 2 replicated this effect and excluded an alternative account based on perceived pattern complexity. The effect of repetitive presentation was obtained despite the fact the participants did not recognize the repetition, suggesting that unconscious priming increased temporal production. These results suggest that time perception is affected by an unconscious process.
2005 Elsevier Inc. All rights reserved. Keywords: Time perception; Unconscious priming; Perceptual fluency; Temporal production; Unconsciousness 1. Introduction Several factors can increase or decrease time perception. For instance, the perceived duration of briefly presented stimuli is increased as stimulus attributes increase, including their number q This research was supported by a grant from the Research Fellowships of the Japan Society for the Promotion of Science for Young Scientists to the first author. * Corresponding author. Fax: +81 824 24 6763. E-mail address: fuminor@hiroshima-u.ac.jp (F. Ono). 1053-8100/$ - see front matter
2005 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2005.02.001 F. Ono, J. Kawahara / Consciousness and Cognition 14 (2005) 474–482 475 (Mo, 1971, 1974, 1975), size (Mo & Michalski, 1972; Thomas & Cantor, 1975), and complexity (Schiffman & Bobko, 1974). In contrast, perceived duration is reduced as the difficulty of concurrently imposed tasks increases, as in recognition memory tasks (Hicks & Brundige, 1974), mirror drawing tasks (Brown, 1985), and Stroop tasks (Sawyer, Meyers, & Huser, 1994). Matsuda (1996) classified the factors affecting perceived time duration into three categories. The first category is the cognition of the non-temporal attributes of events. Perceived duration increases when observers recognize the stimuli as more numerous, intensive, complex, and/or large. The second category of factors affecting perceived time duration is the level of attention (i.e., awareness) to the lapse of time, where perceived duration increases with increased attention. The third category is neurophysiological excitement, in which the perceived duration increases with higher levels of neurophysiological activity. Most previous research on determinants for modulating the perception of brief events has been concerned with conditions in the first and second categories, which occur above the threshold of consciousness. For example, an observerÕs consciousness is clearly engaged in the cognition of non-temporal attributes of events, and that awareness is reflected when the level of awareness of elapsed time is manipulated. The third category of factors affecting time perception, neurophysiological excitement, can also involve conscious awareness; the feeling of being excited or of having a high fever, for example, involves conscious awareness. In other words, previous studies have examined factors that are at least partly conscious, and have found that they produce robust effects on the perception of temporal duration. It is unknown, however, whether perceived duration is affected by any completely unconscious processes. More specifically, does the perceived duration vary depending on whether the stimuli that were previously exposed and that are stored unconsciously are presented again or a new stimulus is presented? Witherspoon and Allan (1985) demonstrated that time perception is affected by repeated stimulus presentation. They showed that even one prior presentation of a word increased its durationjudgment ratio (DJR; Block, Zakay, & Hancock, 1999; Block, Hancock, & Zakay, 2000). In their experiment, participants read aloud words presented once at the rate of one per second, and then they identified words that were presented for 30 or 50 ms. Some of the words had been read previously; others were new. After each presentation of a word, the participants judged its duration. Results indicated that when a word was presented before making the temporal judgment, participants perceived its duration in a subsequent presentation to be longer than it was. Note that Witherspoon and Allan (1985) did not consider whether the participants were aware of the stimulus repetition. Consequently, it was unclear whether the factor affecting the DJR in their experiment was the repeated presentation of stimuli or awareness of the stimulus repetition. In this study, we resolved this issue by introducing a new paradigm, consisting of a temporal production and unconscious priming task. We tested the effect of unconscious stimulus repetition on temporal production. If the perceived temporal duration of a stimulus varies when the participants are not aware of the repetition of that stimulus, then the variation is the consequence of unconscious priming. 2. Experiment 1 The experiment consisted of three phases. In the practice phase, the participants studied the target interval and practiced the temporal production task. In the temporal-production phase, the 476 F. Ono, J. Kawahara / Consciousness and Cognition 14 (2005) 474–482 participants were asked to keep random-dot pattern displays on the screen for 2500 ms by pressing the mouse button. To examine the effect of unconscious priming on temporal production, half of the displays were repeated across blocks throughout the entire experiment, and the remaining displays were generated anew from trial to trial. This temporal-production phase was followed by the recognition phase, in which the participants judged whether the displays had been presented previously. This phase was designed to ensure that the repetition of specific displays was not consciously encoded. If repetition of specific displays affected temporal production, then there would be differences in the interval for the new and repeated displays. If the displays were stored unconsciously, however, then recognition performance should be at chance levels. 2.1. Method 2.1.1. Participants Sixteen naive students from Hiroshima University volunteered in return for course credit or for pay. The participants were 19- to 30-years-old, and all had normal or corrected-to-normal visual acuity. 2.1.2. Apparatus and stimulus materials The stimuli were displayed on a CRT monitor (GDM-19PS, SONY) controlled by a PC/ATcompatible computer equipped with a frame store (VSG 2/5, Cambridge Research Systems). The stimulus displays consisted of 30 dots and covered approximately 6.0 · 5.0 of visual angle. Each dot covered approximately 0.1 · 0.1 of visual angle. The background was gray. Viewing distance was approximately 60 cm. 2.1.3. Procedure The participants were tested individually. At the beginning of the experiment, the participant was asked to remove his or her watch, and not to count during the temporal-production phase. After instructions were given, there were six alternate practice and temporal-production trials, starting with practice. The practice trials were inserted to stabilize temporal-production performance. At the end of the experiment, the participants were given a recognition test. 2.1.4. Practice phase The displays used in the practice phases contained 30 white dots that appeared within an invisible 6 · 5 grid subtending approximately 6.0 · 5.0 of visual angle. First, to study the target interval, the participant observed the dot pattern three times for 2500 ms. These presentations were followed by the first practice phase. In a practice trial, the participant pressed the right mouse button. After a brief pause of 1000, 1300, or 1600 ms, an array of white dots appeared on the screen and remained until the participant pressed the left mouse button. One of the feedback phrases ‘‘too long,’’ ‘‘too short,’’ or ‘‘correct’’ appeared on the screen to inform the participants of the correctness of their response. ‘‘Correct’’ was displayed when the response fell within a 20% temporal window centered on the target interval of 2500 ms (i.e., 2250–2750 ms). The practice phase continued until a ‘‘correct’’ response was obtained three times in a row. On average, it took about 3 min (15 trials) to complete a practice phase. A practice phase preceded each temporal-production phase. F. Ono, J. Kawahara / Consciousness and Cognition 14 (2005) 474–482 477 2.1.5. Temporal-production phase The two main independent variables were configuration (Old vs. New) and block (1–6). The Old set of dot patterns consisted of six randomly generated configurations that were repeated throughout the entire experiment, once per block. The New set consisted of six different configurations that were generated anew for each block to serve as a control baseline. Separate configurations were generated for each participant. Each session consisted of six blocks of 12 trials each (six Old and six New) for a total of 72 trials. Each display contained 30 dots that appeared within an invisible 12 · 8 grid that subtended approximately 6.0 · 5.0 of visual angle. Each array was heterogeneously colored, consisting of eight red, eight green, seven blue, and seven yellow dots. To ensure that participants would not recognize the repetition, we used different color dots; colors were randomly assigned to each of the dots within the configuration. The color assignments of the dots in the Old configurations were preserved across repetitions. Participants were not told that some of the configurations were displayed repeatedly. Participants pressed the right mouse button to begin each trial. Each trial started with a small fixation dot that appeared in the middle of the screen. After an ISI of 1000, 1300, or 1600 ms, an array of dots appeared on the screen. To prevent participants from conducting the temporal production task without looking at the stimuli, we varied the ISI between the fixation dot and the array. The duration of the ISI was the same each time for a specific configuration. The display remained on the screen until the end of the production period, which was terminated when the participant pressed the left mouse button. The participants were instructed to reproduce the target time interval they had practiced as closely as possible. Each production phase consisted of six blocks of 12 trials each (six Old and six New) for a total of 72 trials. 2.1.6. Recognition phase The purpose of the recognition phase was to confirm that any effects of memory on interval production in the temporal-production phase were unconscious. Six Old sets of dot patterns and six new dot patterns were randomly presented with each display preceded by a 1000, 1300, or 1600 ms delay. The display remained on until there was a response. Participants judged whether the patterns had been presented previously, and responded by pressing a mouse button. 2.2. Results and discussion The variation of the delay before the dot pattern presentation did not affect temporal production, F < 1, ns; thus this factor was collapsed in subsequent analyses. Fig. 1 illustrates the mean Old and New productions in the temporal-production phase. An analysis of variance (ANOVA) was conducted with configuration condition (Old vs. New) and block (2–6) as within-subject variables. As the Old vs. New distinction did not apply in block 1 because all configurations were new, we compared blocks 2–6. The analysis revealed a significant main effect of configuration, F (1, 15) = 6.36, p < .05. The main effect of block was not significant, F < 1, ns. The interaction between condition and block was approached significance, F (4, 60) = 2.06, p < .10. When only the last (Block 6) and the second block (Block 2) were compared, there was a marginally significant interaction between configuration · block F (1, 15) = 3.27, p < .10. Similar pairwise comparison between Block 6 and the remaining blocks (Blocks 3–5) revealed significant interaction effects in the pair of Blocks 3 and 6, F (1, 15) = 4.77, p < .05. A linear and quadratic trend 478 F. Ono, J. Kawahara / Consciousness and Cognition 14 (2005) 474–482 Fig. 1. The mean temporal productions for the Old and New conditions in Experiment 1. analysis revealed no significant trend for block, p > .05, ns, but the linear trend for the interaction of condition and block was marginally significant, F (1, 15) = 3.75, p < .10. These results suggested that temporal production was affected by repeated presentation and, specifically, that repetitive presentation of the visual stimuli increased the interval produced in response. Mean accuracy in the recognition phase was 48.95%, and the hit rate was 43.75%, no different to the false alarm rate of 45.83%, t (15) < 1, ns. No participants reported noticing that certain patterns repeated; therefore, we can assume that unconscious priming affected the interval produced in the temporal-production phase. The combined results demonstrate that unconscious priming of the visual stimuli increased temporal production. In Experiment 2, we tested the effect of unconscious priming on perceived stimulus complexity to the dot pattern, which would explain longer temporal production in the Old condition compared to the New condition. Since Schiffman and Bobko (1974) found that more complex stimuli lengthened perceived duration, perceived stimulus complexity may have mediated the results of Experiment 1. 3. Experiment 2 To test the alternative account of stimulus complexity, we replicated Experiment 1 but asked participants to judge the complexity of the Old and the New patterns at the end of Experiment 2. If complexity is critical, the Old patterns should be judged to be more complex than the New. Experiment 2 was also designed to dismiss three minor concerns. First, since the unconscious priming effect suddenly appeared in the last blocks (the 6th block), we extended the number of blocks to eight to confirm the robustness of the effect. If the effect of block is reliable, temporal production in the Old condition should be greater than that in the New condition even in Blocks 7 and 8. F. Ono, J. Kawahara / Consciousness and Cognition 14 (2005) 474–482 479 Second, the amount of unconscious priming should have increased as blocks progressed, but there was no consistent trend across blocks in Experiment 1. We doubled the number of trials per block to 24 trials per block in Experiment 2, since the effect of pattern repetition was relatively small. Finally, we used a stricter criterion in the practice phase. In the practice phase of Experiment 1, the feedback ‘‘correct’’ was displayed when the response fell within a 20% temporal window centered on the target interval of 2500 ms. In Experiment 2, we narrowed the ‘‘correctness’’ window to an 8% temporal window centered on the target interval of 2500 ms (i.e., 2400–2600 ms). 3.1. Method 3.1.1. Participants Twelve naive students from Hiroshima University volunteered in return for course credit or for pay. The participants were 19- to 23-years-old, and all had normal or corrected-to-normal visual acuity. 3.2. Procedure 3.2.1. Practice phase The practice phase was the same as in Experiment 1, except for two changes. First, the feedback ‘‘correct’’ was displayed when the response fell within an 8% temporal window centered on the target interval of 2500 ms (i.e., 2400–2600 ms). Second, the number of trials in each practice phase was changed to 20. 3.2.2. Temporal-production phase The temporal-production phase was the same as in Experiment 1, except that each production phase consisted of 8 blocks of 24 trials each (12 Old and 12 New), for a total of 192 trials. 3.2.3. Recognition phase The temporal-production phase was the same as in Experiment 1, except that 12 previously presented and 12 newly generated dot patterns were presented. 3.2.4. Complexity estimation phase The purpose of the complexity estimation phase was to examine the possibility that perceived complexity might affect temporal production. Twelve previously presented and 12 newly generated dot patterns were presented in an unpredictable order, with each display preceded by a 1000, 1300, or 1600 ms delay. The display remained on until there was a response. The participant was required to estimate the complexity of each pattern by pressing a designated key on the keyboard (1, very simple; 2, simple; 3, complex; or 4, very complex). 3.3. Results and discussion Varying the delay before the dot pattern presentation did not affect temporal production, F < 1, ns. Fig. 2 presents the mean productions for the Old and New conditions in the temporal-production phase. 480 F. Ono, J. Kawahara / Consciousness and Cognition 14 (2005) 474–482 Fig. 2. The mean temporal productions for the Old and New conditions in Experiment 2. An ANOVA was conducted with configuration condition (Old vs. New) and block (2–8) as within-subject variables. There were no effects of condition (p > .25, ns) or block (p > .06, ns). The interaction between configuration and block, however, was significant, F (6, 66) = 2.76, p < .05. The planned contrast (RyanÕs method) showed that temporal productions in the Old condition were greater than those in the New condition at Blocks 7 and 8, p < .05. A trend analysis indicated a significant linear trend for the interaction of condition and block, F (1, 11) = 15.01, p < .005. These results suggest that temporal production was affected by repeated presentation and, specifically, that repetitive stimulus presentation increased the interval produced in response. Note that the pattern of results across Experiments 1 and 2 seems different. In Experiment 1, the unconscious priming effect became evident at Block 6; while in Experiment 2, the effect reached significance at Blocks 7 and 8. A possible explanation for this delay is that the number of trials per block in Experiment 2 was greater than that in Experiment 1. This means that more trials elapsed until a pattern was repeated in Experiment 2, so it took longer (i.e., a greater number of trials) to obtain the effect. Mean accuracy in the recognition phase was 47% and the hit rate was 22%, no different from the false alarm rate of 25%, t (11) < 1, ns. The mean estimated complexity score for the Old pattern was 2.40, no different from that for the New pattern, 2.56, p > .15, ns. Perceived complexity alone, therefore, cannot account for the effect of unconscious priming on temporal production. 4. General discussion This study investigated whether the length of the temporal interval produced was affected by unconscious priming. Participants maintained visual displays for a 2500 ms period. Half of the displays were repeated across blocks throughout the experiment, and others were generated anew from trial to trial. The displays consisted of patterns so complex that the participants were unable to intentionally memorize them. Results showed that the intervals maintained for repeated displays were longer than those for new displays, despite the fact that participants were unable to recognize the displays, suggesting that unconscious priming increased temporal production. In F. Ono, J. Kawahara / Consciousness and Cognition 14 (2005) 474–482 481 terms of temporal experience, temporal production bears an inverse relationship to the DJR, so underestimating temporal production is subjectively equivalent to overestimating the DJR, and vice versa (Brown, 1997; Zakay, 1993). The present results imply that unconscious priming decreases the DJR. Our results support those of Ono, Kawahara, and Matsuda (2004), who demonstrated that prior subthreshold exposures reduced the perceived duration of a visual display. To explain this result, they provided a speculative interpretation related to the idea of perceptual fluency (Jacoby, 1983) and to Thomas and WeaverÕs (1975) model of time perception. Jacoby (1983) suggested that repeated exposure to a stimulus facilitates subsequent perceptual encoding processes involving that stimulus, despite the fact that the participants are unaware that they have been exposed to the stimulus. Thomas and WeaverÕs model suggests that temporal judgment depends on the encoding processes for non-temporal information. That is, the DJR decreases as encoding processes for non-temporal stimulus features decrease. In conjunction with these ideas, Ono et al. suggested that previous subthreshold exposure to a stimulus increases accessibility to its cognitive representation, while the encoding processes of a non-temporal stimulus feature decrease accessibility, correspondingly reducing the DJR. The results of the present experiments are consistent with that hypothesis (Ono et al., 2004). The present study suggests that, for the displays presented repetitively, better accessibility to their cognitive representations occurred, while at the same time, the encoding processes of the non-temporal stimulus features became less efficient. This resulted in a corresponding reduction of the DJR. In the present experiment, repetitively presenting a difficult-to-memorize stimulus might have increased the perceptual fluency of the stimulus and decreased the processing load needed to encode non-temporal stimulus features, so that the DJR was reduced correspondingly. If we apply this idea to the results of Witherspoon and Allan (1985), then it is conceivable that the factor increasing the DJR was awareness of the stimulus repetition, not the repeated presentation itself. The present study confirmed that repeatedly presenting a stimulus reduced the DJR when participants were unaware of the repetition. There are other two alternatives that could explain the present results. One is the internal pacemaker hypothesis. Treisman (1963) proposed the theoretical concept of an internal pacemaker, where the rate of an internal pacemaker determines subjective duration. If the rate of the pacemaker increases, the model predicts a longer subjective duration. Because the model assumes that non-temporal stimuli affect the pacemaker rate, the more arousing the stimuli, the more the rate increases. In the present study, unconscious priming might have decreased stimulus arousal, so that the pacemaker rate decreased correspondingly and reduced the DJR. The other alternative is that the increased temporal production was due to sensory adaptation. Assume that the repetitive presentation of a pattern leads to a fatigue of the sensory and/or perceptual-representational units that stand for that particular pattern. As a result, the subsequent presentation of the same pattern could cause a delay in the moment when the pattern becomes explicitly visible and thus the temporal production increased. A definitive resolution of these possibilities must await further investigation. In summary, the present results are the first to show the effect of unconscious priming on the perceived interval of a visual display. Specifically, unconscious priming of visual stimuli decreased the DJR. This indicates that unconscious factors should be included in the list of factors that affect time perception. 482 F. Ono, J. Kawahara / Consciousness and Cognition 14 (2005) 474–482 References Block, R. A., Hancock, P. A., & Zakay, D. (2000). Sex differences in duration judgments: A meta-analytic review. Memory & Cognition, 28, 1333–1346. Block, R. A., Zakay, D., & Hancock, P. A. (1999). Developmental changes in human duration judgments: A metaanalytic review. Developmental Review, 19, 183–211. Brown, S. W. (1985). Time perception and attention: The effects of prospective versus retrospective paradigms and task demands on perceived duration. Perception & Psychophysics, 38, 115–124. Brown, S. W. (1997). Attentional resources in timing: Interference effects in concurrent temporal and nontemporal working memory tasks. Perception & Psychophysics, 59, 1118–1140. Hicks, R. E., & Brundige, R. M. (1974). Judgments of temporal duration while processing verbal and physiognomic stimuli. Acta Psychologica, 38, 447–453. Jacoby, L. L. (1983). Perceptual enhancement: Persistent effects of an experience. Journal of Experimental Psychology: Learning, Memory, and Cognition, 9, 21–38. Matsuda, F. (1996). Time estimation. In F. Matsuda, K. Choshi, K. Komura, H. Jingu, K. Yamasaki, & S. Hira (Eds.), Psychological time: A broad and deep mystery (pp. 87–144). Kyoto: Kitaoji-shobou (In Japanese). Mo, S. S. (1971). Judgment of temporal duration as a function of numerosity. Psychonomic Science, 24, 71–72. Mo, S. S. (1974). Comparative judgment of temporal duration as a function of numerosity. Bulletin of the Psychonomic Society, 3, 337–379. Mo, S. S. (1975). Time production of duration as a function of numerosity. Bulletin of the Psychonomic Society, 5, 165–167. Mo, S. S., & Michalski, V. A. (1972). Judgment of temporal duration of area as a function of stimulus configuration. Psychonomic Science, 27, 97–98. Ono, F., Kawahara, J., & Matsuda, F. (2004). Previous subthreshold exposures reduce perceived duration. Current Psychology of Cognition, 22, 27–40. Sawyer, T. F., Meyers, P. J., & Huser, S. J. (1994). Contrasting task demands alter the perceived duration of brief time intervals. Perception & Psychophysics, 56, 649–657. Schiffman, H. R., & Bobko, D. J. (1974). Effects of stimulus complexity on the perception of brief temporal intervals. Journal of Experimental Psychology, 103, 156–159. Thomas, E. A. C., & Cantor, N. E. (1975). On the duality of simultaneous time and size perception. Perception & Psychophysics, 18, 44–48. Thomas, E. A. C., & Weaver, W. B. (1975). Cognitive processing and time perception. Perception & Psychophysics, 17, 363–367. Treisman, M. (1963). Temporal discrimination and the indifference interval: Implications for a model of the ‘‘internal clock’’. Psychological Monographs: General & Applied, 77. Witherspoon, D., & Allan, L. G. (1985). The effect of a prior presentation on temporal judgments in a perceptual identification task. Memory & Cognition, 13, 101–111. Zakay, D. (1993). Time estimation methods—Do they influence prospective duration estimates. Perception, 22, 91–101.

Consciousness and Cognition 42 (2016) 125–134 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Things happen: Individuals with high obsessive–compulsive tendencies omit agency in their spoken language Ela Oren a, Naama Friedmann b, Reuven Dar a,⇑ a b School of Psychological Sciences, Tel Aviv University, Tel Aviv, Israel Language and Brain Lab, School of Education and Sagol School of Neuroscience, Tel Aviv University, Tel Aviv, Israel a r t i c l e i n f o Article history: Received 8 October 2015 Revised 9 March 2016 Accepted 11 March 2016 Keywords: Agency Agent OCD Language a b s t r a c t The study examined the prediction that obsessive–compulsive tendencies are related to an attenuated sense of agency (SoA). As most explicit agency judgments are likely to reflect also motivation for and expectation of control, we examined agency in sentence production. Reduced agency can be expressed linguistically by omitting the agent or by using grammatical framings that detach the event from the entity that caused it. We examined the use of agentic language of participants with high vs. low scores on a measure of obsessive–compulsive (OC) symptoms, using structured linguistic tasks in which sentences are elicited in a conversation-like setting. As predicted, high OC individuals produced significantly more non-agentic sentences than low OC individuals, using various linguistic strategies. The results suggest that OC tendencies are related to attenuated SoA. We discuss the implications of these findings for explicating the SoA in OCD and the potential contribution of language analysis for understanding psychopathology. Ó 2016 Elsevier Inc. All rights reserved. 1. Introduction The sense of agency (SoA) can be broadly defined as ‘‘the registration that I am the initiator of my actions” (Synofzik, Vosgerau, & Voss, 2013). According to a recent model of the SoA (Gentsch & Synofzik, 2014; Synofzik, Vosgerau, & Lindner, 2009; Synofzik, Vosgerau, & Newen, 2008; Synofzik et al., 2013), agency experiences rely on the integration of external cues (e.g., seeing one’s hand moving) with internal cues of proprioception and movement (e.g., feeling the position and the movement of one’s hand) and interoceptive cues, including affective states (e.g., an increase in heart rate and feeling of excitement when performing a personally relevant action). The weight given to these different cues in the integration process that leads to the SoA is believed to depend on the relative reliability of the different cues (Gentsch & Synofzik, 2014). The present study was designed to examine the SoA in obsessive–compulsive disorder (OCD). There are good reasons to believe that OCD is characterized by a distorted SoA (Belayachi & van der Linden, 2010; Gentsch, Schütz-Bosbach, Endrass, & Kathmann, 2012). A deficient SoA is directly implied by the notion of compulsion, which is the experience of individuals with OCD that they do not always choose their actions freely. More generally, a central assumption in Shapiro’s (1965) classic theory of OCD is that obsessive–compulsive (OC) individuals have a deficient sense of autonomy and agency. Surprisingly, very few empirical studies have addressed the SoA in OCD, and only a couple examined basic processes that are believed to contribute to the SoA. Gentsch et al. (2012) examined event-related potentials of OCD and control participants in response to ⇑ Corresponding author at: School of Psychological Sciences, Tel Aviv University, Tel Aviv 69978, Israel. E-mail addresses: ela.oren@gmail.com (E. Oren), naamafr@post.tau.ac.il (N. Friedmann), ruvidar@tauex.tau.ac.il (R. Dar). http://dx.doi.org/10.1016/j.concog.2016.03.012 1053-8100/Ó 2016 Elsevier Inc. All rights reserved. 126 E. Oren et al. / Consciousness and Cognition 42 (2016) 125–134 self-generated vs. externally generated visual stimuli. Self-generated stimuli are typically suppressed in comparison to externally generated stimuli (sensory attenuation), an effect believed to be one source of the SoA. This effect is reflected in EEG studies as a suppression of the N1 component to self-generated, as compared with externally generated, stimuli. In Gentsch et al.’s study, the expected suppression of N1 was reduced in OCD participants. According to the authors, this finding suggests that individuals with OCD fail to predict and to suppress the sensory consequences of their own actions. An earlier study, which compared somatosensory evoked potentials in OCD during relaxation vs. movement, reached a similar conclusion (Rossi et al., 2005). Synofzik and colleagues’ model of the SoA is highly relevant to the SPIS (Seeking Proxies for Internal States) model of OCD (Lazarov, Dar, Liberman, & Oded, 2012a, 2012b; Lazarov, Dar, Oded, & Liberman, 2010; Lazarov, Liberman, Hermesh, & Dar, 2014; Liberman & Dar, 2009). The SPIS model offers an account of OC doubt and ensuing rituals. It postulates that OC individuals are generally uncertain about their internal states, including what they feel, what they know, what they believe, and what they prefer. Therefore, when they must answer questions in regard to their internal states, OC individuals must seek and rely on proxies for these internal states. Proxies in this model are substitutes for the internal state that the individual perceives as more easily discernible or less ambiguous, such as indicators, rules, procedures, behaviors, or environmental stimuli (Liberman & Dar, 2009). The SPIS model implies that to the extent that the SoA in a specific context depends on accurate perception of internal signals, the experience of agency should be attenuated for OC individuals. While the observations and theoretical arguments summarized above indicate a deficient SoA in OCD, other observations appear to lead to the opposite conclusion. The OCD-related construct of inflated responsibility (Salkovskis, Shafran, Rachman, & Freeston, 1999), which is the belief that one is responsible for negative outcomes that are far removed from one’s immediate control, suggests a heightened SoA in this population. Similarly, people with OCD often believe that their thoughts would automatically lead to actions (‘‘thought-action fusion”, Shafran, Thordarson, & Rachman, 1996) or events in the world (‘‘thought-event fusion”, Gwilliam, Wells, & Cartwright-Hatton, 2004), which also appears to suggest an elevated SoA. In line with these observations, Reuven-Magril, Dar, and Liberman (2008) found elevated judgments of perceived control in participants with high OC tendencies (compared to those with low OC tendencies) and in OCD participants (compared to nonclinical controls). Notably, these subjective judgments reflected an illusory sense of control, as in reality, events in the experimental procedure used by the authors were entirely uncontrollable. A similar duality was observed in the study by Gentsch et al. (2012): whereas the EEG indices in OCD participants suggested a deficiency in the process believed to lead to the SoA, when participants rated the relation between their actions and the visual stimuli, these explicit judgments of agency were higher in OCD participants compared to controls and were correlated with the severity of OC symptoms. One way to conceptualize the apparent inconsistency noted above is to suggest that the explicit measures of the SoA used in previous studies reflect attempts to compensate for the experience of a deficient SoA seen in implicit measures. In this view (e.g., Reuven-Magril et al., 2008; Shapiro, 1965), the OC person experiences an attenuated sense of autonomy and will, two psychological constructs that are highly correlated with self-agency and a sense of control (de Haan, Rietveld, & Denys, 2013). This deficit motivates compensatory efforts to control all actions, thoughts, impulses and emotions. Such compensation mechanism is exhibited when a person with OCD tries to prevent negative events, on which snhe has no control, by controlling what snhe does, thinks, desires, or feels. This causal model implies that explicit measures of agency in OCD, such as the illusion of control (Reuven-Magril et al., 2008) and OCD core beliefs in inflated responsibility and control (Rachman, 1993; Salkovskis et al., 1999; Shafran, 1997) might miss a more basic experience of reduced agency in this population. Therefore, in the present study, we measured the SoA indirectly through the use of language. The language people use can be informative in regard to many psychological processes (e.g., Semin & Fiedler, 1988) and systematic analysis of language can serve as an implicit measure of psychological constructs, including attitudes (e.g., Sekaquaptewa, Vargas, & Von Hippel, 2010) and the level of agency (Duranti, 2004). Specifically, reduced agency can be expressed by omission of the agent altogether or by using alternative grammatical framings that detach the event from the entity that might have caused it. Clinical experience with OCD suggests that indeed, their language often conveys reduced agency. For example, a client might say ‘‘there is the thought that. . .” rather than ‘‘I think that. . .” or ‘‘you feel that. . .” rather than ‘‘I feel that. . .” or ‘‘the clothes go into the washer” rather than ‘‘I put the clothes in the washer”. Yamamoto (2006) suggested that the mitigation of self-agency is motivated by the wish to reduce the responsibility and guilt attributed to the agent. A good illustration for this is the comparison between ‘‘police shot 100 demonstrators” and ‘‘100 demonstrators were killed” (Fairclough, 1992). Another classic example is the rhetorical phrase ‘‘mistakes were made” used by President Nixon, presumably expressing reluctance to take responsibility for his actions (Kuha, 2007). Fausey and Boroditsky (2010) recently demonstrated that subtly different linguistic descriptions of accidents influence how much people blame and punish those involved. They altered the level of agency expressed in descriptions of accidents, using either agentic (e.g., ‘‘she had ignited the napkin”) or non-agentic (e.g., ‘‘the napkin had ignited”) language. In accordance with their hypothesis, agentic descriptions led participants to attribute more blame and request higher financial penalties than did non-agentic descriptions. For OC individuals, the use of non-agentic language may not only reflect a diminished SoA, but might also serve to reduce the responsibility and guilt, which are central concerns in this disorder (e.g., Lopatka & Rachman, 1995; Mancini, D’Olimpio, & Cieri, 2004). As shown above, previous studies found that explicit measures of SoA often diverted from implicit measures and actually showed elevated SoA in OCD. These explicit measures assessed the SoA in the specific tasks used in those studies, all of which were intentionally vague in order to allow for the perception of elevated SoA or illusory control (e.g., Gentsch et al., 2012; Reuven-Magril et al., 2008). We reasoned that if we assess the general SoA of OC individuals, as distinct from the specific E. Oren et al. / Consciousness and Cognition 42 (2016) 125–134 127 experimental task, we might be able to tap the core diminished SoA and bypass the hypothesized compensation mechanism that results in elevated estimates of control. Therefore, in addition to the implicit measure, namely the spoken language of participants, we also used a novel self-report scale of the general SoA (see below in Materials and Procedure). Based on the rationale elaborated above, we hypothesized that OCD and OC tendencies would be associated with a general attenuation of access to internal states, which would lead to an impairment in the sense of agency. In this study, we tested our hypothesis in the context of language by analyzing indices of agency in sentences that individuals with high vs. low OC tendencies produced. We predicted that high OC tendencies would be associated with reduced linguistic expression of agency, compared to low OC tendencies. In addition, we hypothesized that high OC tendencies, compared to low OC tendencies, would be associated with reduced SoA as measured directly using a self-report scale. 2. Method 2.1. Participants 64 Hebrew-speaking participants were recruited for the study: 30 low OC participants and 34 high OC participants. The participants were students who had completed the Obsessive–Compulsive Inventory-Revised (see Materials description below) prior to the experiment and scored in the top and bottom 25% of the distribution of responders, respectively. Using ‘‘analog” samples of high and low scorers on measures of OCD has proven very useful (e.g., Lazarov et al., 2010, 2012a, 2012b; Soref, Dar, Argov, & Meiran, 2008; for review see Abramowitz et al., 2014). Three participants were excluded from the study due to the fact that when they completed the OCI-R again, as part of the experiment itself, their answers no longer placed them in the top or bottom 25% of the distribution. Another participant was excluded from the study because her age and education were considerably higher than all other participants, who were undergraduate students. The remaining sample of 60 participants consisted of 28 low OC participants (60% women, mean age of 22.9, SD = 2.13, range: 18–28) and 32 high OC participants (81% women, mean age of 22.9, SD = 1.77, range: 19–28). All participants reported that their native language was Hebrew and that they had normal hearing and normal (or corrected) vision. All participants signed an informed consent prior to participation, received payment or course credit for their participation and were fully debriefed after the completion of the study. The experimental protocol was approved by the local ethics committee. 2.2. Materials and procedure 2.2.1. Self-report questionnaires The Obsessive–Compulsive Inventory-Revised (OCI-R, Foa, Kozak, Salkovskis, Coles, & Amir, 1998); The OCI-R lists 18 characteristic symptoms of OCD. Each symptom is followed by a 4-point Likert scale ranging from 0 (not at all) to 4 (extremely), on which participants indicate the symptom’s prevalence during the last month. The OCI-R has been shown to have good validity, test–retest reliability, and internal consistency in both clinical (Foa et al., 2002) and non-clinical samples (Hajack, Huppert, Simons, & Foa, 2004). The participants completed the OCI-R in the screening phase and again in the last part of the experiment itself. The Sense of Agency Scale (SoAS, Ťápal, Oren, Dar, & Eitam, 2016); The SoAS lists 13 statements reflecting cognitions which are related to the SoA (for example: ‘‘the decision whether and when to act is within my hands”, ‘‘I am the origin of my actions”). Each statement is followed by a 7-point Likert scale ranging from 0 (do not agree at all) to 7 (absolutely agree), on which participants indicate their agreement or disagreement with the sentence. Seven of the items were reverse worded to avoid response bias (for example: ‘‘my actions just happen without my intention”, see Appendix C). Data collected from 408 participants using an on-line Israeli panel has shown good internal consistency for the entire scale (a = .84). In addition, in another pool of 236 Israeli participants, the SoAS has shown good test–retest reliability; a sub-group (N = 91) of the original sample completed the scale again two months after the first wave of data collection. Test–retest reliability was r(90) = .67 for the total score. In our study, the participants completed the SoAS in the context of another experiment conducted in our lab, several weeks after they participated in the current experiment. 2.2.2. Sentence production tasks We used two structured test methodologies to elicit sentence production in a natural, conversation-like setting: a preference task and a question-about-a-picture task. 2.2.2.1. The preference task. In the preference task (based on the task developed and described in Friedmann & Szterman, 2006; Friedmann, Yachini, & Szterman, 2015; Novogrodsky & Friedmann, 2006), the experimenter presented 30 questions in which participants were asked to choose between two alternatives, for example: ‘‘There are two men. The artist is drawing one man and painting one man. Which man do you prefer to be?” In Hebrew, such questions typically elicit relative clauses (‘‘I prefer to be the man whom the artist is drawing”). The formation of these relative clauses can convey differing levels of agency. For example, ‘‘the man that the artist is drawing” expresses a high SoA, as the agent himnherself is mentioned, whereas ‘‘the drawn man” or ‘‘the one who is being drawn” express low SoA, as the agent is omitted. Because relative clauses are sensitive markers for syntactic impairments (see Friedmann & Novogrodsky, 2004, 2007; 128 E. Oren et al. / Consciousness and Cognition 42 (2016) 125–134 Friedmann et al., 2015 for relative clauses in developmental syntactic impairments in Hebrew), the elicitation of relative clauses also allowed us to rule out possible language impairments of the participants. Women participants were presented with situations with female protagonists and men participants were presented with masculine protagonists. To examine whether the identity of the agent affects participants’ avoidance of agents, items included either human or inanimate protagonists, and the human protagonists were either the participant himnherself or another person. The different types of items were presented in semi-random order, so that the same type of items were not presented successively. To minimize repetitive answers with a fixed structure and to encourage spontaneous speech, we also included filler questions (see Appendix A for examples and details). 2.2.2.2. The question-about-a-picture task. The second task elicited the production of sentences using 28 questions about pictures. To avoid fixed patterns of answers, filler questions were also included in this task, in mixed order, as described in detail in Appendix B. The first block of items comprised eight questions about pairs of pictures, modelled after Friedmann and Szterman (2006). One picture in each pair showed one figure performing an action on the other, and the other picture showed the same figures in reversed roles. The experimenter described the pictures using simple sentences and then asked about one of the figures (‘‘Here are two pictures. In one picture, the boy is tickling the grandfather, and in one picture, the grandfather is tickling the boy. Which grandfather is this?”). The target responses were object relatives, allowing differing levels of agency, as in the preference task: An answer expressing agency would be one in which the agent is mentioned (e.g., ‘‘This is the grandfather that the boy is tickling”), whereas an answer omitting agency would be one in which the agent is omitted (e.g., ‘‘This is the grandfather who’s being tickled”). The second block included 20 pictures depicting one figure performing an action on another (for example: a boy covering, washing, splashing or scratching a girl), modelled after Costa, Lobo, Carmona, and Silva (2008) and Varlokosta et al. (2015). The experimenter described each picture in a simple sentence and then asked the participants a why-question (‘‘The grandmother is covering the girl and now the girl is happy. Why is the girl happy?”). Like the object relative clauses described above, the elicited because-clauses in this block allowed participants to express or omit agency. For example, ‘‘The girl is happy because the grandmother is covering her” expresses a high SoA, as the agent himnherself is mentioned, whereas ‘‘The girl is happy because she has a blanket” expresses a low SoA, as the action is not mentioned and hence the agent is omitted. To examine possible effects of content, we included both negative (mostly aggressive) and non-negative pictures (see Appendix B for details and examples). No time limit was imposed during testing and no response-contingent feedback was given by the experimenter. All the responses of the participants were transcribed during the sessions. Participants were told that the experiment was designed to test students’ preferences and viewpoints, so they were unaware of the fact that our interest was in fact their spoken language. After completing these tasks, participants completed the OCI-R online using the Qualtrics website (http://telaviv.eu. qualtrics.com). 2.2.3. Coding of responses The level of the SoA in all tasks was measured for each participant based on the frequency of linguistic features that could be used to exercise, attribute, or deny agency in her/his responses. These included the following: (1) Passive without a by phrase: Using sentences with passive verbs without the by phrase (‘‘the floor is being washed,” not ‘‘the floor is being washed by a man”) instead of active ones (‘‘a man is washing the floor”) allows the speaker to omit the agent; (2) Adjectival passive: In addition to verbal passives, one can also use adjectival passive to omit the agent (‘‘The baked cake”); (3) Unaccusativity: This is a grammatical device for encoding agentless actions by only describing the undergoer of the action to whom something is happening, and not the agent conducting the action (‘‘the vase broke”); (4) Inanimate subjects: In everyday spoken (and written language), it is not uncommon to encounter sentences such as this: ‘‘Rents jumped to record highs in Southland. . .” (Duranti, 2004). This represents a way to omit agency by substituting human agents with inanimate ones, thereby attributing the agency to an intermediate agent (sometimes referred to as ‘causer’, Alexiadou & Schäfer, 2006); (5) Arbitrary pro: in Hebrew, passive voice is rarely used; Instead, by using a generic third person plural pronoun, which is unpronounced but is seen in the inflection of the verb agreeing with the subject of the clause, the actual agent is left unmentioned. For instance, ‘‘Ha-yeled she-menashkim oto” [‘‘the boy that (they are) kissing-3rd-plural-masculine him”]; (6) Avoidance of relative clauses (RC): in the context of our tasks, producing a relative clause calls for an agent, thus avoidance of its production is another way of omitting the agent, for example: ‘‘I prefer the shirt from the dryer”, instead of ‘‘I prefer the shirt that I dried”; (7) Avoidance of agency by cancellation of the action: Replies with a because-clause allow responses that describe the setting without mentioning the original action in the sentence, and thus the agent himnherself. For example: Experimenter: ‘‘The boy is washing the giraffe and the giraffe is wet. Why is the giraffe wet?” participant: ‘‘Because of the water and the soap”, instead of ‘‘Because the boy is washing it”. Each response was coded according to the presence or absence of each one of these features. We predicted that high OC tendencies would be associated with attenuated linguistic expression of agency, as assessed by increased usage of the linguistic tools that allow omission of agency. Specifically, we hypothesized that high OC participants would use non-agentic speech more than low OC participants in both the preference task and the question-about-a-picture task, regardless of the content of the questions. E. Oren et al. / Consciousness and Cognition 42 (2016) 125–134 129 3. Results Within each task, the level of agency was measured for each participant based on the frequency of linguistic features that can be used to exercise, attribute, or deny agency, as detailed above. Our main dependent measure in each task was the percentage of answers in which the agent was omitted by each participant, using one of the possible seven strategies. The distribution of this percentage for each group in each task did not differ significantly from normality as assessed by measures of skewness and the kurtosis, which did not differ significantly from zero in any of the experimental conditions, allowing analysis with parametric statistics. 3.1. The preference task We used an alpha level of .05 for all statistical tests. In order to test the general hypothesis that high OC participants would be more inclined than low OC participants to omit the agent of the action, we conducted a two-tail independent sample t-test on the proportion of non-agentic responses. Consistent with our prediction, high OC participants produced a larger percentage of non-agentic responses (M = 26.9%, SD = 18.9) than the low OC group (M = 16.8%, SD = 13.1), t(58) = 2.43, p = .02, Cohen’s d = 0.63. As Fig. 1 illustrates, high OC participants had a higher number of non-agentic responses in all the strategies that they used. We conducted additional analyses to test whether the content of the items affected participants’ tendency to express or omit agency. There was no difference between human and inanimate agents, and no interaction between the type of agent (human vs. inanimate) and OC tendencies (high vs. low), p > .27. In contrast, sentences in which the subject was another person were characterized by less agentic speech, compared to sentences in which the subject was the participant himnherself, F(1, 58) = 8.74, p = .004, gp2 = .13 (Mself = 12.08%, SD = 23.24%, Manother person = 21.25%, SD = 29.41%). Importantly, however, the interaction between the type of agent (the participant himnherself vs. another person) and OC tendencies (high vs. low) was not statistically significant, F(1, 58) = 3.09, p = .08 (Mhigh OCnmyself ¼ 15:62%, SD = 26.75%, Mlow OCnmyself ¼ 8:03%, SD = 18.07%, Mhigh OCnanother person ¼ 29:69%, SD = 34.45%, M low OCnanother person ¼ 11:61%, SD = 18.61%). Finally, overall and regardless of OC tendencies, the average frequency of sentences with agents (M = 23.4, SD = 5.1) was significantly larger than that of sentences without agents (M = 6.7, SD = 5.1), v2 = 4.65, p = .03. 3.2. The question-about-a-picture task We conducted a two-tail independent sample t-test to examine the hypothesis that high OC participants would be more inclined than low OC participants to omit the agent conducting the action. Consistent with our prediction and with the preference task’s results, the high OC group produced over three times more responses without an agent (M = 35.2%, SD = 25.4) than did the low OC group (M = 10.3%, SD = 13.9), t(58) = 4.77, p < .001, Cohen’s d = 1.25. As Fig. 2 illustrates, high OC participants produced a higher number of responses without agents in all relevant strategies. We also conducted additional analyses in order to see whether the type of content (negative vs. non-negative) affected participants’ tendency to express or omit agency. There was no main effect of the type of content and no interaction between the content and OC tendencies (high vs. low), all p > .52. As in the preference task, a Chi-square goodness of fit test revealed that the average frequency of sentences with an agent in the question-about-a-picture task was larger than that of sentences without an agent, v2 = 3.91, p = .04 (Magent = 21.4, SD = 6.8, Mno agent = 6.6, SD = 6.8). Concerning both the preference task and the question-about-a-picture task, notice that the reduction of agency cannot be ascribed to avoidance of complex syntactic structures: some of the ways to avoid agency actually resulted in a more complex syntactic structure. For example, sentences with passive are relatively complex morphologically and syntactically (e.g., Belletti, 2008; Collins, 2005). In other cases, the response did not differ from an agentive target response in terms of syntactic complexity, as is the case with the use of the cancellation of the action in the because-clauses in the question-about-apicture task. 3.3. The Sense of Agency Scale (SoAS) Due to the fact that the participants completed the SoAS several weeks after they participated in the current experiment, nine participants dropped out and did not complete the SoAS. Cronbach’s alpha for the remaining sample of 51 participants was .87, indicating good internal consistency of the scale. We conducted a two-tail independent sample t-test to examine the hypothesis that high OC tendencies would be associated with low agency related cognitions. Consistent with our prediction, high OC individuals reported less agency related cognitions, as indicated by lower scores on the SoAS (M = 4.9, SD = 0.9), compared to the low OC individuals (M = 5.4, SD = 0.6), t(49) = 2.10, p = .04, Cohen’s d = .65. 4. Discussion This study examined the prediction that the sense of agency (SoA), as manifested through spoken language, would be attenuated in individuals with high obsessive–compulsive (OC) tendencies, as compared to individuals with low OC tendencies. We examined this hypothesis using two sentence elicitation tasks, in which relative clauses and because-clauses were 130 E. Oren et al. / Consciousness and Cognition 42 (2016) 125–134 Fig. 1. Average sum of non-agentic answers by OC tendencies and linguistic strategy in the preference task. Fig. 2. Average sum of non-agentic answers by OC tendencies and linguistic strategy in the question-about-a-picture task. elicited as responses to simple questions presented to high and low OC participants. The results showed that both high and low OC participants used agentic speech more than they used non-agentic speech. This is not surprising; As Bandura (2001) has asserted, the capacity to exercise control over the nature and quality of one’s life is the essence of humanness, hence it was reasonable to assume that most of the participants’ answers would reflect a sense of agency. Importantly, the difference between the two OC groups was in line with our predictions: high OC participants used non agentic speech more than low OC participants did. This non agentic speech included the omission of the agent altogether or the production of alternative grammatical framings that detach the event from the entity that caused it. This was found in the two main methodologies we used: the preference task and the question-about-a-picture task. Unexpectedly, there two OC groups did not differ in their tendency to omit agency between sentences in which the agent was the participant himnherself or another person. Moreover, OC individuals’ tendency to omit agency was not altered by whether the actions had positive or negative valence and whether the agent was human vs. inanimate.1 This pattern may imply that OC individuals tend to perceive events in the world as happening, not as being caused, which may be related to a 1 Note that our findings may indicate that the SoA, at least as expressed in spoken language, represents a general sense of a causer of change, be it a person with mental states and intentions or an inanimate cause. In that sense, the notion of agent that is relevant to the SoA as expressed through language is similar to the notion of the ‘‘Causer” in linguistic terms (Everaert, Marelj, & Siloni, 2012; Reinhart, 2003). E. Oren et al. / Consciousness and Cognition 42 (2016) 125–134 131 reduced sense of predictability and control. Such an experience of the world may account for the low sense of control in individuals with OCD and their attempt to compensate for it by constructing rigid rules and rituals (Moulding & Kyrios, 2007; Pacherie, 2008; Reuven-Magril et al., 2008). This (admittedly speculative) hypothesis may tie into recent findings suggesting deficient sensory integration in OCD (e.g., Dar, Kahn, & Carmeli, 2012; Lewin, Wu, Murphy, & Storch, 2014; Rossi et al., 2005), as sensory integration is believed to depend on accurate perception of causality (Lochmann & Deneve, 2011; Parise, Spence, & Ernst, 2012). Future studies are needed to examine whether indeed OCD tendencies are associated with deficits in perceiving causal relationships. The SoA is considered an important building block for our concept of free choice and beliefs about our influence on our actions and environment (Bandura, 1986; Deci & Ryan, 1985; Haggard, 2005; Jeannerod, 2003). As documented in the Introduction, OCD presents a complex picture in regard to the SoA. Two previous evoked potential studies (Gentsch et al., 2012; Rossi et al., 2005) have indicated a diminished SoA in OCD, a finding that is consistent with the sense of compulsion that characterizes this disorder. Conversely, when high OC individuals are asked to estimate their SoA, these explicit estimates tend to be higher than those of controls participants (Gentsch et al., 2012; Reuven-Magril et al., 2008). These inflated estimates of control or agency are consistent with beliefs in one’s ability to control events in the world through rituals or thoughts (e.g., ‘‘thought-action fusion”, Shafran et al., 1996), which are characteristic of OCD. As explicated earlier above, this apparent contradiction may be resolved if we conceptualize explicit measures of the SoA and beliefs in magical power in OCD as attempts to compensate for a more basic experience of a deficient SoA that is reflected in implicit measures. Along the same lines, Pacherie (2008) hypothesized that the characteristic feature of OCD is an abnormally low SoA that may be counteracted by compulsive behaviors. In this view, compulsive rituals serve to create an illusory SoA that helps to reinstate a desired feeling of control (see also Reuven-Magril et al., 2008). The fact that our findings regarding language are consistent with those obtained with implicit measures (Gentsch et al., 2012; Rossi et al., 2005) suggests that the assessment of SoA though language captures the more primary, diminished SoA in high OC individuals (it is worth noting that participants were not aware of the fact that the tasks in our study were designed to assess agency). An alternative explanation for our findings is that the omission of agency in the language of high OC individuals is motivated, rather than merely a reflection of a diminished SoA. Cognitive models of OCD have proposed that inflated responsibility is a key element of the disorder (e.g., Lopatka & Rachman, 1995; Salkovskis et al., 1999), hence one possible motivation to mitigate agency could be to lighten the burden of responsibility felt by high OC individuals. A related motivation may be to reduce guilt: Mancini and Gangemi (2004) have suggested that OC behavior is regulated by the fear of guilt that would result from behaving irresponsibly. According to empirical studies cited by the authors (e.g., Mancini & Gangemi, 2002, 2004), both responsibility and fear of guilt, but particularly the latter, increase obsessive-like behaviors. If the use of non-agentic language reflects the motivation of OC individuals to reduce feelings of responsibility and guilt, however, this should affect only or mostly their speech about themselves. This was not the case in our study: high OC participants avoided agents in their sentences both when the agent was the individual himnherself and when it was a different person or even an inanimate cause. We believe that this finding supports the hypothesis that the expression of the SoA through language reflects a generalized diminishment of the SoA in high OC individuals. Further support for this interpretation comes from the finding that there was no difference in high OC individuals’ tendency to avoid agents between negative and non-negative content: if the omission of agency in the language of high OC individuals was motivated, then aggressive actions should have triggered more motivation to avoid guilt, as compared to non-negative content. Nevertheless, more research is needed in order to empirically test these two possibilities. As results obtained in different paradigms and with explicit and implicit measures of the SoA in the same paradigm tend to be poorly correlated (Dewey, Pacherie, & Knoblich, 2014; Gentsch et al., 2012), future research should use multiple methods and measures to assess the SoA in OCD. Two implicit measures of the SoA, which were the focus of previous studies, are the Intentional Binding effect and the Sensory Attenuation effect (e.g., Moore & Obhi, 2012; Weiss, Herwig, & SchützBosbach, 2011). These two effects are based on the idea that sensory processing of one’s own actions tends to differ from that of externally triggered stimuli. Theories of the SoA postulate a ‘‘forward model” in the motor system that predicts the sensory consequences of executing an action. If the predicted sensory events occur, they are linked to the prediction of the forward model, which contributes to the SoA (Wolpert & Ghahramani, 2000). Studies showed that higher congruency between predicted and actual sensory consequences increases the SoA (e.g., Sato & Yasuda, 2005). In future studies, it would be important to explore whether or not these known implicit measures of the SoA converge with indices of agency in language, both in relation to OCD and in general. In our study, in addition to the implicit measure, we also used an explicit measure of the SoA. Previous studies found that unlike the implicit measures of SoA, explicit measures often show elevated SoA in OCD (e.g., Gentsch et al., 2012; ReuvenMagril et al., 2008). The explicit measures used in those studies assessed the SoA in the scope of the specific experimental tasks, and were intentionally vague in order to allow for the perception of elevated SoA, namely illusory control. Our direct measure, on the other hand, was aimed to capture the more basic concept of agency, hence reducing the potential of an illusory SoA. Our rationale was that by assessing the general SoA of OC individuals, as distinct from the specific experimental task, we might succeed in revealing the core diminished SoA and bypass the hypothesized compensation mechanism that results in illusory SoA. This was done using a novel self-report scale, which includes general statements related to participants’ perceptions regarding their SoA in their every-day lives. Indeed, we found that high OC participants reported less agency in their daily lives in comparison to low OC participants, in accord with their use of agency in spoken language. 132 E. Oren et al. / Consciousness and Cognition 42 (2016) 125–134 Our study focused on the linguistic strategies used to omit agency in Hebrew. It would be interesting to use the preference task and the question-about-a-picture task in other languages as well, thus allowing the examination of the use of agent-avoidance strategies that are less accessible in Hebrew, such as the passive voice (which is not frequently used in Hebrew but is very common in English and Italian, for instance). Previous work has shown that languages differ in their preference for agentic vs. non-agentic frames (e.g., Fausey & Boroditsky, 2011; Fausey, Long, Inamori, & Boroditsky, 2010). Conducting a comparison between languages with regards to the method used in this study could expand the current body of knowledge in the field of cross-linguistic differences in cognition. In addition, our findings are based on a non-clinical, highly functioning, largely female student sample, so their generalization to OCD requires replication with a clinical sample (preferably including a larger age range than the one used in our study). As previous studies showed that OC-like behaviors can be enhanced or reduced by manipulating certain situational aspects (e.g., Dek, van den Hout, Engelhard, Giele, & Cath, 2015; Lazarov, Cohen, Liberman, & Dar, 2015; Radomsky & Alcolado, 2010), it is possible that the tendency to use of nonagentic speech is also affected by contextual features. Hence, future studies might explore the short- and the long-term stability of OC tendencies or behaviors. Although the linguistic perspective of agency has received some attention in previous studies (e.g., Fausey & Boroditsky, 2010; Yamamoto, 2006), its connection to individual differences and to psychopathology has been largely unexplored. To our knowledge, this study is the first attempt to link OCD and the use of language and the first to examine language as an index of the SoA. As such, it also provides evidence to the relevance of linguistics in the field of clinical psychology research. Specifically, using language as an indicator of SoA has the potential to be relevant not only to OCD but to other conditions in which one might expect a distortion in the SoA, such as depression and psychosis. Therefore, our results do not imply that the use of non-agentic speech may be a diagnostic to obsessive–compulsive tendencies. However, linguistic research may have important implications in the field of psychotherapy. For example, Acceptance and Commitment Therapy (ACT), a major part of the third wave of behavioral and cognitive therapies, focuses on the ways in which clients understand and perpetuate their difficulties through language (Hayes, Strosahl, & Wilson, 2011). ACT is based on the Relational Frame Theory (RFT, Hayes, Barnes-Holmes, & Roche, 2001), a basic experimental analysis of human language and cognition. In its application, however, ACT employs mostly paradox, metaphors, stories, etc. (Hayes, 2004), rather than addressing the core, syntactic level of clients’ spoken language. Our results suggest that clinicians might enhance their understanding of their clients by attending to the omission (or in other cases inflation) of agency in their speech. More speculatively, drawing clients’ attention to their avoidance of agency, and even practicing using more agentic language, might enhance their sense of agency and control. Authors declaration The authors declare no actual or potential conflict of interest in relation to this study. This paper was not supported by any funding. We confirm that the manuscript has been read and approved by all named authors and that there are no other persons who satisfied the criteria for authorship but are not listed. We further confirm that the order of authors listed in the manuscript has been approved by all of us. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.concog. 2016.03.012. References Abramowitz, J. S., Fabricant, L. E., Taylor, S., Deacon, B. J., McKay, D., & Storch, E. A. (2014). The relevance of analogue studies for understanding obsessions and compulsions. Clinical Psychology Review, 34(3), 206–217. Alexiadou, A., & Schäfer, F. (2006). Instrument subjects are agents or causers. In D. Baumer, D. Montero, & M. Scanlon (Eds.). Proceedings of the 25th west coast conference on formal linguistics (Vol. 25, pp. 40–48). Somerville, MA: Cascadilla Proceedings Project. Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory.Englewood Cliffs, NJ: Prentice-Hall. Bandura, A. (2001). Social cognitive theory: An agentic perspective. Annual Review of Psychology, 52, 1–26. Belayachi, S., & van der Linden, M. (2010). Feeling of doing in obsessive–compulsive checking. Consciousness and Cognition, 19(2), 534–546. Belletti, A. (2008). Notes on passive object relatives. In P. Svenonius (Ed.), Functional structure from top to toe. Oxford: Oxford University Press. Collins, C. (2005). A smuggling approach to the passive in English. Syntax, 8(2), 81–120. Costa, J., Lobo, M., Carmona, J., & Silva, C. (2008). Clitic omission in European Portuguese: Correlation with null objects? In A. Gavarrò & M. João Freitas (Eds.), Language acquisition and development. Proceedings of generative approaches of language acquisition 2007 [GALA 2007] (pp. 133–143). Newcastle-uponTyne: Cambridge Scholars Publishing. Dar, R., Kahn, D. T., & Carmeli, R. (2012). The relationship between sensory processing, childhood rituals and obsessive–compulsive symptoms. Journal of Behavior Therapy and Experimental Psychiatry, 43(1), 679–684. Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-determination in human behavior.New York: Plenum Press. de Haan, S., Rietveld, E., & Denys, D. (2013). Being free by losing control: What obsessive–compulsive disorder can tell us about free will. In W. Glannon (Ed.), Free will and the brain: Neuroscientific, philosophical and legal perspectives. Cambridge, UK: Cambridge University Press. Dek, E. C., van den Hout, M. A., Engelhard, I. M., Giele, C. L., & Cath, D. C. (2015). Perseveration causes automatization of checking behavior in obsessive– compulsive disorder. Behaviour Research and Therapy, 71, 1–9. Dewey, J. A., Pacherie, E., & Knoblich, G. (2014). The phenomenology of controlling a moving object with another person. Cognition, 132(3), 383–397. Duranti, A. (Ed.). (2004). A companion to linguistic anthropology. Malden, MA: Blackwell Publishing Ltd. E. Oren et al. / Consciousness and Cognition 42 (2016) 125–134 133 Everaert, M., Marelj, M., & Siloni, T. (2012). The theta system: An introduction. In M. Everaert, M. Marelj, & T. Siloni (Eds.). The theta system: Argument structure at the interface. Oxford studies in theoretical linguistics (Vol. 37, pp. 1–19). Oxford, UK: Oxford University Press. http://dx.doi.org/10.1093/acprof: oso/9780199602513.003.0001. Fairclough, N. (1992). Discourse and social change.Cambridge, UK: Polity. Fausey, C. M., & Boroditsky, L. (2010). Subtle linguistic cues Influence perceived blame and financial liability. Psychonomic Bulletin & Review, 17(5), 644–650. Fausey, C. M., Long, B. L., Inamori, A., & Boroditsky, L. (2010). Constructing agency: The role of language. Frontiers in Psychology, 1(162), 1–11. http://dx.doi. org/10.3389/fpsyg.2010.00162. Fausey, C. M., & Boroditsky, L. (2011). Who dunnit? Cross-linguistic differences in eye-witness memory. Psychonomic Bulletin & Review, 18(1), 150–157. Foa, E. B., Huppert, J. D., Leiberg, S., Langner, R., Kichic, R., Hajcak, G., & Salkovskis, P. (2002). The obsessive–compulsive inventory: Development and validation of a short version. Psychological Assessment, 14, 485–496. Foa, E. B., Kozak, M., Salkovskis, P. M., Coles, M. E., & Amir, N. (1998). The validation of a new obsessive–compulsive disorder scale: The obsessive– compulsive inventory. Psychological Assessment, 10, 206–214. Friedmann, N., & Novogrodsky, R. (2004). The acquisition of relative clause comprehension in Hebrew: A study of SLI and normal development. Journal of Child Language, 31(03), 661–681. Friedmann, N., & Novogrodsky, R. (2007). Is the movement deficit in syntactic SLI related to traces or to thematic role transfer? Brain and Language, 101(1), 50–63. Friedmann, N., & Szterman, R. (2006). Syntactic movement in orally trained children with hearing impairment. Journal of Deaf Studies and Deaf Education, 11 (1), 56–75. Friedmann, N., Yachini, M., & Szterman, R. (2015). Relatively easy relatives: Children with syntactic SLI avoid intervention. In E. Di Domenico, C. Hamann, & S. Matteini (Eds.), Structures, strategies and beyond. Studies in honour of Adriana Belletti. Amsterdam, The Netherlands: John Benjamins, Linguistik Aktuell Series. Gentsch, A., Schütz-Bosbach, S., Endrass, T., & Kathmann, N. (2012). Dysfunctional forward model mechanisms and aberrant sense of agency in obsessive– compulsive disorder. Biological Psychiatry, 71(7), 652–659. Gentsch, A., & Synofzik, M. (2014). Affective coding: The emotional dimension of agency. Frontiers in Human Neuroscience, 8(608), 1–7. Gwilliam, P., Wells, A., & Cartwright-Hatton, S. (2004). Dose meta-cognition or responsibility predict obsessive–compulsive symptoms: A test of the metacognitive model. Clinical Psychology & Psychotherapy, 11(2), 137–144. Haggard, P. (2005). Conscious intention and motor cognition. Trends in Cognitive Sciences, 9, 290–295. Hajack, G., Huppert, J. D., Simons, R. F., & Foa, E. B. (2004). Psychometric properties of the OCI-R in a college sample. Behaviour Research and Therapy, 42, 115–123. Hayes, S. C. (2004). Acceptance and commitment therapy, relational frame theory, and the third wave of behavioral and cognitive therapies. Behavior Therapy, 35(4), 639–665. Hayes, S. C., Barnes-Holmes, D., & Roche, B. (Eds.). (2001). Relational frame theory: A post-skinnerian account of human language and cognition. New York: Plenum Press. Hayes, S. C., Strosahl, K. D., & Wilson, K. G. (2011). Acceptance and commitment therapy: The process and practice of mindful change.New York: Guilford Press. Jeannerod, M. (2003). The mechanism of self-recognition in humans. Behavioural Brain Research, 142, 1–15. Kuha, M. (2007). Acceptance and avoidance of responsibility in world leaders’ statements about climate change. Language and Ecology, 2(3), 1–26. Lazarov, A., Dar, R., Liberman, N., & Oded, Y. (2012a). Obsessive–compulsive tendencies and undermined confidence are related to reliance on proxies for internal states in a false feedback paradigm. Journal of Behavior Therapy and Experimental Psychiatry, 43(1), 556–564. Lazarov, A., Dar, R., Liberman, N., & Oded, Y. (2012b). Obsessive–compulsive tendencies may be associated with attenuated access to internal states: Evidence from a biofeedback-aided muscle tensing task. Consciousness and Cognition, 21(3), 1401–1409. Lazarov, A., Dar, R., Oded, Y., & Liberman, N. (2010). Are obsessive–compulsive tendencies related to reliance on external proxies for internal states? Evidence from biofeedback-aided relaxation studies. Behaviour Research and Therapy, 48(6), 516–523. Lazarov, A., Liberman, N., Hermesh, H., & Dar, R. (2014). Seeking proxies for internal states in obsessive–compulsive disorder. Journal of Abnormal Psychology, 123(4), 695–704. Lazarov, A., Cohen, T., Liberman, N., & Dar, R. (2015). Can doubt attenuate access to internal states? Implications for obsessive–compulsive disorder. Journal of Behavior Therapy and Experimental Psychiatry, 49, 150–156. Lewin, A. B., Wu, M. S., Murphy, T. K., & Storch, E. A. (2014). Sensory over-responsivity in pediatric obsessive compulsive disorder. Journal of Psychopathology and Behavioral Assessment, 37(1), 134–143. Liberman, N., & Dar, R. (2009). Normal and pathological consequences of encountering difficulties in monitoring progress toward goals. In G. B. Moskowitz & H. Grant (Eds.), The psychology of goals (pp. 277–303). New York, NY: Guilford. Lochmann, T., & Deneve, S. (2011). Neural processing as causal inference. Current Opinion in Neurobiology, 21(5), 774–781. Lopatka, C., & Rachman, S. (1995). Perceived responsibility and compulsive checking: An experimental analysis. Behaviour Research and Therapy, 33, 673–684. Mancini, F., D’Olimpio, F., & Cieri, L. (2004). Manipulation of responsibility in non-clinical subjects: Does expectation of failure exacerbate obsessive– compulsive behaviors? Behaviour Research and Therapy, 42(4), 449–457. Mancini, F., & Gangemi, A. (2002). Role of responsibility in conditional reasoning. Psychological Reports, 91(1), 275–288. Mancini, F., & Gangemi, A. (2004). The influence of responsibility and guilt on naive hypothesis-testing. Thinking and Reasoning, 10(3), 289–320. Moulding, R., & Kyrios, M. (2007). Desire for control, sense of control and obsessive–compulsive symptoms. Cognitive Therapy and Research, 31(6), 759–772. Moore, J. W., & Obhi, S. S. (2012). Intentional binding and the sense of agency: A review. Consciousness and Cognition, 21(1), 546–561. Novogrodsky, R., & Friedmann, N. (2006). The production of relative clauses in SLI: A window to the nature of the impairment. Advances in Speech-Language Pathology, 84, 364–375. Pacherie, E. (2008). The phenomenology of action: A conceptual framework. Cognition, 107, 179–217. Parise, C. V., Spence, C., & Ernst, M. O. (2012). When correlation implies causation in multisensory integration. Current Biology, 22(1), 46–49. Rachman, S. (1993). Obsessions, responsibility and guilt. Behaviour Research and Therapy, 31, 149–154. Radomsky, A. S., & Alcolado, G. M. (2010). Don’t even think about checking: Mental checking causes memory distrust. Journal of Behavior Therapy and Experimental Psychiatry, 41, 345–351. Reinhart, T. (2003). The theta system – An overview. Theoretical Linguistics, 28(3), 229–290. http://dx.doi.org/10.1515/thli.28.3.229. Reuven-Magril, O., Dar, R., & Liberman, N. (2008). Illusion of control and behavioral control attempts in obsessive–compulsive disorder. Journal of Abnormal Psychology, 117, 334–341. Rossi, S., Bartalini, S., Ulivelli, M., Mantovani, A., Di Muro, A., Goracci, A., & Passero, S. (2005). Hypofunctioning of sensory gating mechanisms in patients with obsessive–compulsive disorder. Biological Psychiatry, 57(1), 16–20. Salkovskis, P., Shafran, R., Rachman, S., & Freeston, M. H. (1999). Multiple pathways to inflated responsibility beliefs in obsessional problems: Possible origins and implications for therapy and research. Behaviour Research and Therapy, 37(11), 1055–1072. Sato, A., & Yasuda, A. (2005). Illusion of sense of self-agency: Discrepancy between the predicted and actual sensory consequences of actions modulates the sense of self-agency, but not the sense of self-ownership. Cognition, 94, 241–255. Sekaquaptewa, D., Vargas, P., & Von Hippel, W. (2010). A practical guide to paper-and-pencil implicit measures of attitudes. In B. Gawronski & B. Payne (Eds.), Handbook of implicit social cognition: Measurement, theory, and applications (pp. 140–155). New York, NY: Guilford Press. Semin, G. R., & Fiedler, K. (1988). The cognitive functions of linguistic categories in describing persons: Social cognition and language. Journal of Personality and Social Psychology, 54(4), 558–568. 134 E. Oren et al. / Consciousness and Cognition 42 (2016) 125–134 Shafran, R. (1997). The manipulation of responsibility in obsessive–compulsive disorder. British Journal of Clinical Psychology, 36, 397–407. Shafran, R., Thordarson, D. S., & Rachman, S. (1996). Thought-action fusion in obsessive compulsive disorder. Journal of Anxiety Disorders, 10(5), 379–391. Shapiro, D. (1965). Neurotic styles.New York: Basic Books. Soref, A., Dar, R., Argov, G., & Meiran, N. (2008). Obsessive–compulsive tendencies are related to focused information processing strategy in the flanker task. Behaviour Research and Therapy, 46, 1295–1299. Synofzik, M., Vosgerau, G., & Lindner, A. (2009). Me or not me – An optimal integration of agency cues? Consciousness and Cognition, 18(4), 1065–1068. Synofzik, M., Vosgerau, G., & Newen, A. (2008). Beyond the comparator model: A multifactorial two-step account of agency. Consciousness and Cognition, 17 (1), 219–239. Synofzik, M., Vosgerau, G., & Voss, M. (2013). The experience of agency: An interplay between prediction and postdiction. Frontiers in Psychology, 4(127), 1–8. http://dx.doi.org/10.3389/fpsyg.2013.00127. Ťápal, A., Oren, E., Dar, R., & Eitam, B. (2016). The Sense of Agency Scale (SoAS): A Measure of One’s Perceived Control over Mind, Body, and the Immediate Environment. (in preparation). Varlokosta, S., Belletti, A., Costa, J., Friedmann, N., Gavarró, A., Grohmann, K. K., ... Argemí, N. (2015). A cross-linguistic study of the acquisition of clitic and pronoun production. Language Acquisition. http://dx.doi.org/10.1080/10489223.2015.1028628. Weiss, C., Herwig, A., & Schütz-Bosbach, S. (2011). The self in action effects: Selective attenuation of self-generated sounds. Cognition, 121(2), 207–218. Wolpert, D. M., & Ghahramani, Z. (2000). Computational principles of movement neuroscience. Nature Neuroscience, 3(Suppl.), 1212–1217. Yamamoto, M. (2006). Agency and impersonality: Their linguistic and cultural manifestations.Amsterdam: Benjamins. Further reading Lavy, E., van Oppen, P., & van den Hout, M. (1994). Selective processing of emotional information in obsessive compulsive disorder. Behaviour Research and Therapy, 32(2), 243–246.

Consciousness and Cognition Consciousness and Cognition 14 (2005) 645–646 www.elsevier.com/locate/concog Papers to Appear in Forthcoming Issues Todd E. Feinberg, Julian Paul Keenan. Where in the brain is the self? Antonio Zadra, Sophie Desjardins, Éric Marcotte. Evolutionary function of dreams: A test of the threat simulation theory in recurrent dreams Benjamin Libet. The timing of brain events: Reply to the ‘‘Special Section’’ in this journal of September 2004, edited by Susan Pockett Shaun Gallagher, Jesper Brøsted Sørensen. Experimenting with phenomenology Jonathan Smallwood, Leigh Riby, Derek Heim, John B. Davies. Encoding during the attentional lapse: Accuracy of encoding during the semantic sustained attention to response task Julia Fisher, Elliot Hirshman, Thomas Henthorn, Jason Arndt, Anthony Passannante. Midazolam amnesia and short-term/ working memory processes Susan Pockett. The great subjective back-referral debate: Do neural responses increase during a train of stimuli? Chris Frith. The self in action: Lessons from delusions of control Cristina Becchio, Mauro Adenzato, Bruno G. Bara. How the brain understands intention: Different neural circuits identify the componential features of motor and prior intentions Debra A. Gusnard. Being a self: Considerations from functional imaging Daniel A. Pollen. Brain stimulation and conscious experience: Electrical stimulation of the cortical surface at a threshold current evokes sustained neuronal activity only after a prolonged latency David A. Oakley, Patrick Haggard. The timing of brain events: AuthorsÕ response to LibetÕs ÔReplyÕ Uta Frith, Frederique de Vignemont. Egocentrism, allocentrism, and Asperger syndrome Frederick Toates. A model of the hierarchy of behaviour, cognition, and consciousness Ap Dijksterhuis, Teun Meurs. Where creativity resides: The generative power of unconscious thought K. Valli, A. Revonsuo. Recurrent dreams: Recurring threat simulations? Cordula Becker, Mark A. Elliott. Flicker-induced color and form: Interdependencies and relation to stimulation frequency and phase Stephen A. Dewhurst, Selina J. Holmes, Karen R. Brandt, Graham M. Dean. Measuring the speed of the conscious components of recognition memory: Remembering is faster than knowing Catherine Deeprose, Jackie Andrade. Is priming during anesthesia unconscious? Stephan G. Boehm, Ellen C. Klostermann, Werner Sommer, Ken A. Paller. Dissociating perceptual and representation-based contributions to priming of face recognition doi:10.1016/S1053-8100(05)00097-8 646 Papers to Appear in Forthcoming Issues / Consciousness and Cognition 14 (2005) 645–646 Kathy Pezdek, Shirley Lam. What research paradigms have cognitive psychologists used to study ‘‘False memory,’’ and what are the implications of these choices? Donald D. Hoffman. The scrambling theorem: A simple proof of the logical possibility of spectrum inversion Maurizio Tirassa, Francesca Marina Bosco, Livia Colle. Rethinking the ontogeny of mindreading Mark Blagrove, Sarah-Jayne Blakemore, Ben R.J. Thayer. The ability to self-tickle following Rapid Eye Movement sleep dreaming

Consciousness and Cognition Consciousness and Cognition 14 (2005) 426–427 www.elsevier.com/locate/concog Papers to Appear in Forthcoming Issues David Kahn, J. Allan Hobson. State-dependent thinking: A comparison of waking and dreaming thought Henk Aarts, Ruud Custers, Daniel M. Wegner. On the inference of personal authorship: Enhancing experienced agency by priming effect information Séverine Fay, Michel Isingrini, Viviane Pouthas. Does priming with awareness reflect explicit contamination? An approach with a response-time measure in word-stem completion Cristina Becchio, Cesare Bertone. The ontology of neglect Hilde Haider, Peter A. Frensch, Daniel Joram. Are strategy shifts caused by data-driven processes or by voluntary processes? Anna Abraham, Sabine Windmann, Irene Daum, Onur Güntürkün. Conceptual expansion and creative imagery as a function of psychoticism Todd E. Feinberg, Julian Paul Keenan. Where in the brain is the self? John D. Eastwood, Daniel Smilek. Functional consequences of perceiving facial expressions of emotion without awareness Jean-Marie Danion, Christine Cuervo, Pascale Piolino, Caroline Huron, Marielle Riutort, Charles Siegfried Peretti, Francis Eustache. Conscious recollection in autobiographical memory: An investigation in schizophrenia Elke Geraerts, Elke Smeets, Marko Jelicic, Jaap van Heerden, Harald Merckelbach. Fantasy proneness, but not self-reported trauma is related to DRM performance of women reporting recovered memories of childhood sexual abuse Anne Giersch, Serge Caparos. Focused attention is not enough to activate discontinuities in lines, but scrutiny is Lucina Q. Uddin, Jan Rayman, Eran Zaidel. Split-brain reveals separate but equal self-recognition in the two cerebral hemispheres Anna Stone, Tim Valentine. Strength of visual percept generated by famous faces perceived without awareness: Effects of affective valence, response latency, and visual field Fuminori Ono, Jun-ichiro Kawahara. The effect of unconscious priming on temporal production Richard L. Abrams. Unconscious processing of multiple nonadjacent letters in visually masked words Antonio Zadra, Sophie Desjardins, Éric Marcotte. Evolutionary function of dreams: A test of the threat simulation theory in recurrent dreams Benjamin Libet. The timing of brain events: Reply to the ‘‘Special Section’’ in this journal of September 2004, edited by Susan Pockett Shaun Gallagher, Jesper Brøsted Sørensen. Experimenting with phenomenology Jonathan Smallwood, Leigh Riby, Derek Heim, John B. Davies. Encoding during the attentional lapse: Accuracy of encoding during the semantic sustained attention to response task doi:10.1016/S1053-8100(05)00073-5 Papers to Appear in Forthcoming Issues / Consciousness and Cognition 14 (2005) 426–427 427 Julia Fisher, Elliot Hirshman, Thomas Henthorn, Jason Arndt, Anthony Passannante. Midazolam amnesia and short-term/ working memory processes D. Ben Shalom. Autism and the experience of a perceptual object Susan Pockett. The great subjective back-referral debate: Do neural responses increase during a train of stimuli? Chris Frith. The self in action: Lessons from delusions of control Cristina Becchio, Mauro Adenzato, Bruno G. Bara. How the brain understands intention: Different neural circuits identify the componential features of motor and prior intentions Debra A. Gusnard. Being a self: Considerations from functional imaging Daniel A. Pollen. Brain stimulation and conscious experience: Electrical stimulation of the cortical surface at a threshold current evokes sustained neuronal activity only after a prolonged latency David A. Oakley, Patrick Haggard. The timing of brain events: AuthorsÕ response to LibetÕs ÔReplyÕ Uta Frith, Frederique de Vignemont. Egocentrism, allocentrism, and Asperger syndrome Frederick Toates. A model of the hierarchy of behaviour, cognition, and consciousness Ap Dijksterhuis, Teun Meurs. Where creativity resides: The generative power of unconscious thought

Consciousness and Cognition 43 (2016) 75–88 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Review article Mysticism and schizophrenia: A phenomenological exploration of the structure of consciousness in the schizophrenia spectrum disorders Josef Parnas, Mads Gram Henriksen ⇑ Center for Subjectivity Research, University of Copenhagen, Denmark Psychiatric Center Glostrup/Hvidovre, Copenhagen University Hospital, Denmark a r t i c l e i n f o Article history: Received 25 November 2015 Revised 18 May 2016 Accepted 20 May 2016 Keywords: Schizophrenia Mysticism Self-disorders Ipseity Disturbance Model Psychosis Delusion Hallucination Psychopathology Phenomenology Consciousness a b s t r a c t Mysticism and schizophrenia are different categories of human existence and experience. Nonetheless, they exhibit important phenomenological affinities, which, however, remain largely unaddressed. In this study, we explore structural analogies between key features of mysticism and major clinical-phenomenological aspects of the schizophrenia spectrum disorders—i.e. attitudes, the nature of experience, and the ‘other’, mystical or psychotic reality. Not only do these features gravitate around the issue of the basic dimensions of consciousness, they crucially seem to implicate and presuppose a specific alteration of the very structure of consciousness. This finding has bearings for the understanding of consciousness and its psychopathological distortions. Ó 2016 The Authors. Published by Elsevier Inc. This is an open access article under the CC BYNC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Contents 1. 2. 3. 4. 5. 6. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 Attitudes in mysticism and schizophrenia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 The nature of mystical and schizophrenic experience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 The other reality in mysticism and schizophrenia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 4.1. Ultimate reality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 4.2. The world of psychosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 Minimal self and the de-structuration of immanence in schizophrenia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 Conclusion and implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 Conflict of interest statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 ⇑ Corresponding author at: Center for Subjectivity Research, University of Copenhagen, Njalsgade 140-142, Building 25, 5th Floor, DK-2300 Copenhagen S, Denmark. E-mail address: mgh@hum.ku.dk (M.G. Henriksen). http://dx.doi.org/10.1016/j.concog.2016.05.010 1053-8100/Ó 2016 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). 76 J. Parnas, M.G. Henriksen / Consciousness and Cognition 43 (2016) 75–88 1. Introduction Mysticism and mental illness are two very different categories, located in distinct conceptual spaces. Yet, there are some superficial connections, e.g., both categories concern consciousness (subjectivity) and unusual experiences. Psychiatric literature has to some extent explored similarities between mystical states (often designated as ‘altered states of consciousness’), hallucinogenic drug-induced states, and global alterations of consciousness, typically ecstatic-confusional or dreamlike conditions encountered in some cases of acute-onset schizophrenia (Buckley, 1981; Deikman, 1971; Gouzoulis-Mayfrank et al., 1998; Nelson & Sass, 2008; Oxman, Rosenberg, Schnurr, Tucker, & Gala, 1988; Podvoll, 1979; Škodlar & Ciglenečki, 2015). By contrast, the literature on mysticism is keen to distinguish between genuine mystical states of ‘unio mystica’ (i.e. an experience of ineffable, boundless sense oneness with the Absolute) and a host of other, often pathological mental states such as euphorias, illusions, delusions, hallucinations, visions, raptures, and trances (Stace, 1960). In this article, we strive to go beyond the superficial and occasional similarities and explore the structural analogies between the features of unio mystica and major clinical-phenomenological aspects of schizophrenia, notably the disorders of the self, a long lasting focus of our research (Parnas & Henriksen, 2014). Our phenomenological approach is eclectic and not bound to any specific authority or school of thought. Rather, we seek to exploit and integrate different, mutually compatible approaches to achieve a cohesive psychopathological grasp of subjectivity in schizophrenia. Initially, we wish to emphasize that we do not entertain the absurd position that mystics suffer from schizophrenia (or psychosis) or vice versa. However, mystical states entail profound and complex alterations of waking consciousness that may offer a possibility of a comparative enlightenment, enabling a better grasp of the phenomenological vicissitudes of conscious life in schizophrenia. Such psychopathological understanding may not only have an intrinsic theoretical value but also pragmatic ends such as improving therapeutic efforts and indicating new areas of pathogenetic research. In this study, we focus exclusively on experiences of unio mystica, leaving aside other mysticism-related phenomena. For the sake of clarity, we address separately three characteristics of mysticism and schizophrenia, i.e. attitudes, nature of experience, and the ‘other’, mystical or psychotic reality, although these characteristics are, in fact, highly interdependent. As we shall see, the issue of self or consciousness is crucially at stake in the discussion of the three characteristics and therefore, in the final section of this essay, we explicitly address the issue of self and structure of consciousness in schizophrenia and mysticism. 2. Attitudes in mysticism and schizophrenia On the basis of mystics’ self-reports, Steinbock (2007) identified certain typical behavioral and mental attitudes that are adopted by mystics to facilitate the emergence of the mystical experience across history and religious traditions. These attitudes include a distancing from and disinterest towards reality and practical life, a suspension of ordinary ontological assumptions (i.e., phenomenologically speaking, to suspend the ‘natural attitude’ by effectuating the epoché [cf. Overgaard, 2015]), spiritual solitude, and a weakening of one’s sense of self (dés-istement de soimême [Depraz, 2001]). In a seminal work, Stace (1960) also pointed to the cross-cultural and temporal invariance of such attitudes. Most importantly, Stace (1960) argued that for the mystical experience to emerge, it is usually necessary to reach a specific state of mind or, using an old Chinese term, a state of ‘no-mind’ (wu-nien), i.e. sort of meditative tranquility in which all sensuous content, imagery, emotions, thoughts, etc. are obliterated from the mind, which thus is empty. Consistently, the Christian mystic Eckhart argued that only by the way of ‘pure detachment’ (2009, pp. 566–575), i.e. by fully freeing oneself from all needs, affections, and interests toward oneself, others, and the world, can the mind become completely empty and receptive of God. Eckhart quotes in this context St. Augustine: ‘The soul has a secret entrance to the divine nature, when all things become nothing for it’, and for Eckhart, ‘this entrance is nothing but pure detachment’ (2009, p. 573).1 The motif of self-annihilation is a recurring idea in many religions—to obtain an immediate experience of God or to reach salvation or liberation (nirvana) as it were, the self must be destroyed. This motif is aptly articulated in Kierkegaard’s religious philosophy: ‘all religion (. . .) aims at a person’s total transformation and wants, through renunciation and selfdenial, to wrest away from him all that, precisely that, to which he immediately clings, in which he immediately has his life’ (Kierkegaard, 1998, p. 248) or, in a more condensed form, ‘self-annihilation is the essential form for the relationship with God’ (Kierkegaard, 1992, p. 984). Remarkably similar considerations can be found in recent philosophical studies of religious experience of the Absolute (Bagger, 2007; Lacoste, 2004; Morgan, 2013). For example, Lacoste described a series of ‘liturgical’ steps involved in facilitating the desired contact with the Absolute. Lacoste based his analysis on Heidegger’s ‘topological’ notion of ‘place’, as inherence in a non-spatial locus of intersecting vital and existential coordinates. The liturgical procedure is a ‘transgression’ because it ‘subverts the dynamics of place’ or, as Lacoste also puts it, ‘‘Being-there’ [Dasein] is bracketed and transformed as being-towards (eschaton) (. . .) Liturgical experience no longer enables us to identify in ourselves a carnal dimension. . . confined within this world’ (Lacoste, 2004, p. 39). In brief, emergence of contact with the Absolute (God) requires also here a series of attitudes, particular ways of disposing of ‘self’ and ‘place’ (e.g., isolation, denial of place [dépaysement], prayer). 1 Eckhart’s mystical path does not preclude suffering for, as Schürmann (2001) puts it, ‘the logic of detachment somehow reflects the logic of the way of the cross’. J. Parnas, M.G. Henriksen / Consciousness and Cognition 43 (2016) 75–88 77 Turning to the schizophrenia spectrum disorders, we may note that many patients are exposed to and suffer from attitudes that in various ways resemble those preparatory steps taken by the mystics. We are here specifically concerned with persistent, trait-like, pre-onset (i.e. premorbid and prodromal) features, also shared by schizotypal disorder. These features are manifest as changed modes of existence and experience, i.e., not only observable from a third-person perspective as peculiar behaviors or conspicuous forms of existential orientation, but also lived through as profound changes of subjective life. Concerning the ‘objective’ features, epidemiological studies show that pre-schizophrenic and schizotypal individuals often appear eccentric, detached, disinterested, isolated from and sometimes rejected by their peers; they are rated by their schoolteachers as being withdrawn and passive (Olin et al., 1998), and rated by clinicians as exhibiting a diminished interpersonal emotional rapport and formal slips in language (e.g., Parnas & Jørgensen, 1989; Parnas, Schulsinger, Schulsinger, Mednick, & Teasdale, 1982; Tyrka et al., 1995). Phenomenologically informed empirical studies testify to an altered subjective life, reflected in the presence of certain trait-like (enduring rather than episodic), non-psychotic, anomalous self-experiences (viz. self-disorders). In the following, we present from a clinical-phenomenological perspective a picture of the altered subjectivity in schizophrenia, an alteration that often emerges many years prior to the onset of frank illness. Frequently and typically, the patients complain of feeling ephemeral, not fully self-present, lacking an inner ‘core’ or ‘nucleus’. They may suffer from a failing sense of being one with oneself (automatic, pre-reflective self-coincidence or ‘radical self-recognition’ [Parnas, 2007]), occasionally necessitating a reflective self-identification. Usually, they report feeling radically different from others (Anderssein) and often this feeling has persisted since early childhood. This particular feeling is frequently verbalized as being ‘wrong’ and it may resist further predication (Parnas & Henriksen, 2014). It is a feeling of a difference so fundamental that it entails a sense of being ontologically different from others and it is often associated with a profound solitude. Such ontological dislocation makes comparisons with others in terms of mundane characteristics difficult (e.g., being too fat or more clever than one’s peers). Regularly, this experience is expressed through complaints such as feeling ‘not really human’ (Saks, 2007, p. 193f.) or ‘I’m a psychomachine’ (Kimura, 2001). The following complaint from a patient illustrates these experiences: ‘‘I looked just like every other child, but inside I was different. It is as if I am another creature that somehow ended up inside a human body” (Henriksen & Nordgaard, 2014, p. 436). Often, these experiences are thematized by fantasies of being a time-traveller, extra-terrestrial or a secrete adoptee. A very typical case of self-disorders in schizophrenia is expressed in the following vignette: I feel like I’m not a natural human being or a proper human being or something like that. I have always tried so much to be a real human being, but I have the feeling that I’m not (. . .) I don’t feel like I have a core or a substance. . . Increasingly, I began to feel that I sort of fused with the surroundings. . . And I had a hard time recognizing myself from hour to hour, day to day (. . .) I had this idea that I didn’t look human and that everybody was sort of being nice to me and played along. Seriously, when I passed others in the street, they were polite and didn’t stare too much, but they would think ‘what was that?’ That is, I really thought that I wasn’t recognizable as a human being (. . .) I must have been 4 or 5 years old. I was starting dance class and I was looking in the mirror. I was standing next to the other kids and I remember that I looked alien. I felt like I sort of stuck out from that large wall mirror. As if I wasn’t a real child. This feeling has been very persistent from very early on (Henriksen & Nordgaard, 2016). Moreover, patients may report a diminished sense of embodied presence in the shared-social world, which often involves a decreased ability to be affected, drawn or stimulated by others, objects or situations—e.g., ‘‘I live in a sort of bubble, where the world does not matter. I lack synchrony with the people around me” (Henriksen & Nordgaard, 2014, p. 437). This diminished sense of immersion may also be linked to problems with ‘common sense’ (Blankenburg, 1971), indicative of disruptions at the level of immediate pre-conceptual resonance or attunement with the world. Typically, such problems manifest as difficulties with an immediate grasp of what is contextually relevant and appropriate, a failing sense of what others consider self-evident (e.g., the natural evidences or the tacit axioms of social interaction). These problems are exemplified in the following vignette: I have always struggled to understand why people didn’t take life more seriously. I mean, ‘‘How can you just walk around, be named ‘Angie,’ buy butter, and take riding lessons?” Every morning, when I wake up, I realize like for the first time that this is the real reality, that we are all going to die, that we don’t know why we are here, that nothing makes sense. . . This is one of the reasons why I feel different from others. They walk around and talk on their phone, plan what they want to do. . . It puzzles me that I haven’t gotten used to it (. . .) It hurts me that it is so easy and natural for the rest of the world. They don’t even think about it (Henriksen & Nordgaard, 2016). Problems with ‘common sense’ are often linked to a tendency to hyper-reflect, which sometimes takes the form of an ongoing (pre-reflective) self-monitoring or ‘simultaneous introspection’ (Nagai, 1991) or ‘hyper-reflexivity’ (Sass, 1992). Thus, the patient may complain about observing himself constantly in his engagement with the world and others, making a true immersion almost impossible (Nagai, 1991). Many patients report that certain thoughts appear intrusive, somehow alien, ‘as if’ they were not generated by themselves; also, uncontrollable, parallel trains of thoughts, occurring simultaneously with a loss of meaning, and experiences of sudden, complete emptiness of thoughts are occasionally described. Furthermore, the boundaries between self/other and self/world may be permeable or unstable in different ways (i.e. transitivism). Very often the reason for avoiding eye contact is the feeling that the others can penetrate into one’s immanent sphere. From this transformed existential position, the patient regularly finds himself unanchored from the world and alienated from others and from aspects of his own existence 78 J. Parnas, M.G. Henriksen / Consciousness and Cognition 43 (2016) 75–88 (e.g., thoughts, feelings, body), and, in his ontological solitude, he may have various quasi-solipsistic experiences, e.g., transient experiences of being the center of the universe with everything somehow referring to him, or having a privileged access to hidden and deeper layers of reality.2 During the last two decades, empirical studies have shown i.a. that such self-disorders hyperaggregate in schizophrenia spectrum disorders but not in other mental disorders (Haug et al., 2012; Nordgaard & Parnas, 2014; Parnas, Handest, Jansson, & Saebye, 2005; Parnas, Handest, Saebye, & Jansson, 2003; Raballo & Parnas, 2012) and that self-disorders predict schizophrenia spectrum diagnosis (Nelson, Thompson, & Yung, 2012; Parnas, Carter, & Nordgaard, 2016; Parnas et al., 2011—for a review, see Parnas & Henriksen, 2014). The generative disorder, underlying the various self-disorders and other domains of schizophrenic psychopathology, has been proposed to be a specific disturbance of ‘ipseity’ or ‘minimal self’ (Cermolacce, Naudin, & Parnas, 2007; Nelson, Parnas, & Sass, 2014; Sass & Parnas, 2003), destabilizing the very sense of self-presence and of being a self-coinciding subject of experience and action. We will now briefly summarize the ‘attitudinal’ analogies between mysticism and schizophrenia. The mystics’ preparatory moves share two interrelated goals: (i) detaching the mystic from the spontaneous immersion or absorption in the world, and (ii) subverting the sense of self, through ‘emptying the mind’ by means of a progressive elimination of all concerns, needs, affections, etc. This radical ‘no-self’ condition creates a state of receptive passivity in which the desired mystical experience finally may articulate itself. Yet, the mystic does not lose his sense of self completely; he retains his first-person perspective because otherwise the experience could not be experienced or reported. The self-disorders in schizophrenia spectrum conditions interlock in various ways on the shared goals of the mystics’ attitudes: the self-disorders radically unanchor the patient from the shared-social world and they entail a profoundly diminished sense of embodied, vital selfpresence. In contrast to the mystic’s attitudes, however, which generally are willingly strived for and at least partially controlled, the self-disorders are typically uncontrollable and involuntary, causing immense suffering (Møller & Husby, 2000). However, like the mystic, the patient does not cease to be a subject of awareness and action and in this regard the firstperson perspective remains intact. We return to these issues in the final section. 3. The nature of mystical and schizophrenic experience What is the nature of the mystical experience and how is it comparable to the nature of primary delusional and emergent hallucinatory experiences? To come closer to answering these questions, it is essential to realize that these different experiences broadly share a coincidence or unity of appearing and signification, i.e. the how (structure) and the what (content) of manifestation are but one. In our view, this is the source or root of the following distinctive features of these experiences. The mystical experiences have often been regarded as having the experiential quality of an epiphany (e.g., James, 2013; Stace, 1960; Steinbock, 2007). An epiphany does not have a ‘feel’ of inferential reasoning or of a gradual piecing together of parts to make a whole but rather of a frequently sudden and always profound realization, manifesting what it discloses as undoubtedly true, i.e. the experience is given in such way that it reveals it to be so. An epiphany is essentially a passive and pathic (affective) experience of striking and deep in-sight—light dawns instantly over the whole. In the clinical, psychopathological literature, it is well established that primary delusional and the formation of hallucinatory experiences in schizophrenia have the experiential form of an epiphany (Conrad, 2002; Ey, 1973; Jaspers, 1997; Škodlar & Ciglenečki, 2015). For example, Conrad argues that the core of schizophrenic experiencing is marked by a twin alteration of the structure of experience, which he articulates with the notions of ‘apophany’ (the experience has the form of a revelation) and ‘anastrophe’ (the experience is centered around the patient in a peculiar way; cf. self-reference without reason [Gruhle]) (Conrad, 2002, p. 269). According to Conrad, ‘delusional perception [a variant of primary delusion] (. . .) is manifestly identical to apophantic [revelatory] experience, which (. . .) is not ‘comprehensible’ but lies entirely outside the categories of comprehensibility’ (2002, p. 110; our translation; italics added). These features of schizophrenic experiencing have been well summarized by Gennart: ‘(1) The meaning of apophantic delusional experience articulates itself through its very manifestation; the meaning is not grasped through the subject’s efforts of interpretation; rather the meaning is simply revealed to the subject by the things themselves. (2) The revelation of meaning imposes itself on the subject, short-cutting his liberty or initiative of understanding’ (2011, p. 324; our translation). Second, paradigmatic revelatory experiences in mysticism do not have an ‘object’ in any ordinary or phenomenological sense of the term, i.e. mystical experiences do not fit the formula of an object presenting itself before an experiencing subject, e.g., like the tree is the object of my visual perception or the poem is the object of my evaluative judgment. Phenomenologically speaking, mystical experiences do not conform to the intentional noetic-noematic structure that otherwise characterizes our experiences (perception, imagination, thinking, etc.). To articulate the specific structure of the mystical experience, Steinbock (2007) employs the term of ‘verticality’ as opposed to ‘horizontality’ that designates the ordinary noetic-noematic or subject-object intentional structure. Steinbock argues that vertical experiences of Absolutes are revealed to us, whereas objects of horizontal experiences are presented or manifested to us (cf. Levinas, 1979; Marion, 1991). It seems to us that the notion of ‘verticality’ not only functions to distinguish the mystical givenness from the ‘horizontal’ duality of the 2 For further details and rich clinical descriptions of self-disorders, see Parnas and Handest (2003), Parnas et al. (2005), Henriksen and Parnas (2012), and Henriksen and Nordgaard (2014, 2016). J. Parnas, M.G. Henriksen / Consciousness and Cognition 43 (2016) 75–88 79 subject-objet intentional relation but also alludes to a spiritual existential vector that is at play, i.e. a vertical striving for a union with divinity (e.g., Steinbock claims that ‘[what] is given vertically incites awe’ [2007, p. 14]). At first glance, psychotic experiences may seem to have a clearly demarcated object (e.g., a hallucinatory ‘voice’ or a specific delusional content or theme), thus conforming to the ordinary subject-object intentional structure. This is certainly the case in the advanced illness stages when these symptoms have become well elaborated. However, phenomenological explorations of the emergence of primary delusional and hallucinatory experiences suggest that the psychotic ‘object’ only materializes and takes its clearly noematic form on the basis of an initially global, diffuse, and sometimes perplexed state of atmospheric pathic tension that has no clear and distinctive noetic-noematic intentional structure. The nature of ‘atmospheres’ does not lend itself to a clear description in terms of a subject-object relation (Griffoni, 2014). Conrad’s phenomenological study of the onset of acute schizophrenia in a sample of 107 patients points to certain typical phases: the revelatory articulation of delusion is often preceded by an increase of basic affective tone (erhöchte Bodenaffektivität), followed by an atmosphere of apprehension, free-floating anxiety or insecurity (occasionally of elation or ecstasy), perhaps of something impending, ‘something in the air’ (Wahnstimmung). The patient increasingly feels as if he is at the very center of what is happening (incipient ‘anastrophic’ experience). However, in this state of uncertainty, the sense of uncertainty is itself ‘absolutely certain’ (Müller-Suur, 1950, p. 45; our translation and italics). The patient may be uncertain about what is going on but he has absolutely no doubt that something is happening. According to Müller-Suur (1950), this is a distinctive feature of schizophrenia—i.e. non-schizophrenic delusional patients experience neither the sense of absolute uncertainty (Ungewißheits bewußtsein) in pre-psychotic stages nor the sense of absolute certainty (Gewissheitsbewusstsein) in psychosis; typically their delusions are not revealed to them but arise in a more gradual, progressive and inferential fashion. In schizophrenia, the fluidity of the experience of delusional mood condenses into a sense of an insidious or more sudden alien presence (copresence), a sentiment of anonymous, intrusive, otherness or alterity in the midst of the patient’s self-intimacy or ‘sphere of ownness’ (l’experience d’alterité) (Ey, 1973). This sense of alterity serves as a source of the noematic elaborations of psychotic projections. In the face of the disturbing uncertainty, the patient searches for solutions or answers. Ey (1973) called the cognitive efforts in this search for meaning and explanation for the ‘psychotic work’ (le travail psychotique). Through the ‘psychotic work’, the apophantic revelation is eventually cognitively elaborated into specific delusional contents, acquiring now a clear form of a noematic ‘object’. Conrad’s analysis lends support to the notion of ‘noetic’ or ‘egological’ delusion introduced in psychopathological literature (e.g., Giudicelli, 1990; Tatossian, 1978, 2014). In the same vein, Straus’ (1935) emphasized the ‘pathic’ (affective) rather than ‘gnostic’ (cognitive) element of the psychotic experience. The notion of ‘noetic’ delusion seems to emphasize both an experiential proximity to the constituting layers of immanence (Section 5) and an initial revelatory coincidence of appearing and signification, jointly underscoring a prominence of the noetic moments of experience with an absence or extreme poverty of clearly demarcated noematic elements. In brief, the psychotic ‘object’ is sensu stricto secondary to or derived from the primary psychotic experience (Conrad, 2002; Ey, 1973; Gennart, 2011; Jaspers, 1997). Third, we will now take a closer look at the mystical revelation. In short, the experience of absolute, undifferentiated unity (unio mystica) in which all distinctions and multiplicity are obliterated is the essential core of the mystical experience in all major mystical traditions (Burch, 1960; Carter, 2013; Forman, 1990; James, 2013; Stace, 1960; Steinbock, 2007). It is important to note that the mystic experiences this unity directly and does not simply interpret or infer it as such.3 Stace (1960) distinguishes between two kinds of mystical experience of union, i.e. two variants with identical phenomenological nucleus. First, an ‘outward-looking’, sensuous, extrovertive kind (usually considered as ‘inferior’ to the second kind), in which the perceptual experience is transfigured so that the unity shines through all diversity and multiplicity of the visual field (e.g., ‘in this light my spirit saw through all things and into all creatures and I recognized God in grass and plants’ [Böhme] or ‘all blades of grass, wood, and stone, all things are One. This is the deepest truth’ [Eckhart]; quoted in Stace, 1960, p. 69, p. 63). In the extrovertive mystical experience, something, say, the blades of grass is simultaneously perceived as ‘grass’ and ‘not grass’ (i.e. as an instantiation of the One). In Zen Buddhism, the mystical experience (satori) of ‘identity in difference’ is described with the notion of ‘soku hi’, i.e. something that both ‘is’ and ‘is not’. The notion of ‘soku hi’ aptly illustrates the experience’s fundamental violation of the principle of non-contradiction. Second, an ‘inward-looking’, non-sensuous, introvertive kind in which all mental content, differentiation, and individuation are obliterated and all that there remains is a pure void, i.e. an empty, unitary consciousness (e.g., ‘when you thus cease to be finite you become one with the Infinite’ [Plotinus], ‘it [the soul] is sunk and lost in the desert where its identity is destroyed’ [Eckhart] or ‘as a lump of salt thrown into water melts away. . .even so, O Maitreyi, the individual soul, dissolved, is the Eternal—pure consciousness, infinite, transcendent’ [The Upanishads]; quoted in Stace, 1960, p. 112, p. 98, p. 118). The dissolution of finite individuality into infinity, i.e. the breaking down of the walls of the self and its dissolution into the vast sea of Being, is the very experience of unio mystica in introvertive mysticism—an experience of absolute emptiness and fullness at once. In Sufism, there is even a technical term for this crucial aspect of the experience, viz. ‘fana’, 3 This may explain why the mystics’ descriptions of unio mystica, reported in various languages, across different historical eras and cultures, and within different religious or atheist frameworks are conspicuously similar. The interpretation, however, depends on their historical, cultural, and religious backgrounds. Thus, the ancient Hindu mystic, as described in the Upanishads, interprets his experience as ‘identity with’ Brahman or the Universal Self. Plotinus speaks of the unity as the ‘One’. The Christian mystics interpret their experiences as ‘union with God’ or, if we follow Eckhart, as union between the ‘ground of the soul’ and ‘the ground of God’ (i.e. the Godhead [divinitas]). In Mahayana Buddhism, the mystical union is between the deepest foundation of the soul and ‘sunyata’ (‘emptiness’ or ‘the Void’). There have of course also existed atheistic mystics, who did not use a religious language to describe their experiences (see, e.g., Stace, 1960). A contemporary secular mysticism has grown into a sort of ‘atheist religion’, with important implications for the emergence of the so-called ‘transpersonal’, spiritual psychology (Hunt, 2003). 80 J. Parnas, M.G. Henriksen / Consciousness and Cognition 43 (2016) 75–88 which literally means a ‘passing away’ (of the self into God). Finally, both the sensuous experience of identity in difference and the non-sensuous experience of absolute unity and dissolution of self have been described with the poetic notion of ‘oceanic feeling’ (Saarinen, 2014).4 Interestingly, Stace’s division between extrovertive and introvertive mystical experiences reminds of two general types of primary delusional experience (Conrad, 2002; Jaspers, 1997; Schneider, 1950). One is world-directed as a ‘delusional perception’ (Wahnwahrnehmung) in which a mainly undistorted perceptual content serves as a vehicle of delusional revelation (e.g., ‘at the steps of a catholic convent, a dog was waiting for me in upright position, watching me seriously. As I approached, it lifted its paw. By chance, another man was walking a meter from me. I quickly caught up with him and asked if the dog had also introduced itself to him. An astonished ‘no’ made me certain that I was here dealing with a plain revelation’ [Schneider, 1950, 106; our translation, italics added]). In another, ‘introspective’ kind (Wahneinfall), delusional significance is revealed in immanent content such as thought, remembrance or image (e.g., ‘[it] suddenly occurred to me one night, quite naturally, self-evidently but insistently, that Miss L. was probably the cause of all the terrible things through which I have had to go these last few years (telepathic influences, etc.) (. . .) I felt as if scales had fallen from my eyes and I saw why life had been precisely as it was through these last years [Jaspers, 1997, p. 103]).5 The revelation, articulated as delusional or hallucinatory experiences in schizophrenia, is typically lived as an apodictic felt insight into another dimension of reality—a dimension that is normally hidden for other people. A prominent German scholar on schizophrenia, Kurt Schneider, described this experience in the following way: ‘The significance [of the experience] is of a special kind; it nearly always carries a great import, is urgent and somehow personal, as a sign or message from another world. It is as if the perception expresses ‘‘a higher reality”’ (Schneider, 1950, p. 106; our translation, italics added). Of course, it is only a minority of patients that conceptualise their psychotic experience in philosophical terms; rather this sense of extraordinary contact with another dimension or layer of reality is manifest through the peculiar incorrigibility of delusions and the attitude of ‘double bookkeeping’ (Section 4.2). It is also important to note that this sentiment of contact with another dimension of reality is quite often present, in a more subtle, inchoate form (e.g., self-disorders), already long before the onset of a fully articulated illness. Sometimes, these pre-onset experiences may exhibit a quality of transient mystical states. Fourth, we will briefly address the issue of ineffability of the mystical experience, repeatedly emphasized by the mystics. It is important to realize that the ineffability of which they nonetheless speak arises not primarily from the quality of the content as such but from the very structure of the entire experience; it has to do with the nature of the mystical experience itself, which, as we shall see later, seems to reveal an ontological realm prior to all distinctions, conceptual carvings and fixations, without any texture, quality or form. Here, there is no difference between subject and object, inside and outside, knower and what is known—they are completely united. According to Stace, this is the logical reason behind the mystical experience’s ineffability; to describe or explain something, we must in some way be detached from it, while remaining framed by the mundane ontological context. Mystical experiences, however, do not fulfill these conditions (Stace, 1960, p. 105; Steinbock, 2007). The intrinsic nature of unio mystica belongs to a sphere of experience or an ontological domain over which our understanding, linguistic categories, and rules of logic have no power—they simply do not apply. Ineffability is also a crucial feature of schizophrenic experience and we have already touched upon several of its sources. In sum, (i) the revelatory givenness of primary delusional and hallucinatory experiences bypasses critical reflection and precludes any Cartesian-style doubt concerning the reality or validity of these experiences; (ii) the psychotic experiences usually arise on the basis of an elusive, diffuse ‘delusional mood’, defying description and understanding; (iii) the delusional mood is usually antedated by self-disorders, which, as we have seen, reflect subtle, yet profound and persistent disruptions at the level of pre-reflective self-experience—experiences that can barely be propositionally expressed (e.g., Anderssein); and (iv) the patient’s sense of being in touch with another ontological dimension makes it nearly impossible to articulate his experiences in the terms of the mundane ontological context. Finally, we will turn to the issue of the effects or traces that the mystical and psychotic experience leaves on the individual. The mystical experience, although usually only very short-lived, can be so powerful and profound that it may revolutionize the entire existence or existential perspective of the individual (Stace, 1960, p. 60f.). It is noteworthy, however, that there are cases where the mystical consciousness seems to have become permanent, ‘running concurrently with, and in some way fused and integrated with, the normal or common consciousness’ (Stace, 1960, p. 61).6 Sometimes, the experience of unio mystica may linger on with an attenuated intensity after the termination of the mystical climax, e.g., as a sort of feeling of enlightenment and clarity, an attitude of care, compassion, and tolerance. Typically, mystical experiences are transformative in a positive sense, i.e. enriching, enhancing existential openness, novel perspectives, and ethical demand. From a phenomenological perspective, an important point to make is that in the mystical states, the synchronic unity of experience and the diachronic unity of the stream of consciousness seem typically to be preserved, i.e. we are not dealing with dissociative phenomena or a disunity of the self. Turning to schizophrenia, we may note that the primary delusional or hallucinatory experience typically exerts a profound impact on the individual as well, but usually in much more disturbing way. This is not so only due to the specific nature of the psychotic experience but also because of the general consequences of falling ill (e.g., social and occupational). Most 4 Originally, the notion was introduced by the mystic and Nobel Prize wining novelist Rolland in a personal letter to Freud, see Parsons (1999, pp. 170–179). In the current diagnostic classifications (ICD-10, DSM-V), only delusional perception is listed due to reliability concerns. 6 This may be a key to understanding the essence of double bookkeeping in schizophrenia; a position where the patient simultaneously lives in two different realities (Section 4.2; Henriksen & Parnas, 2014). 5 J. Parnas, M.G. Henriksen / Consciousness and Cognition 43 (2016) 75–88 81 psychiatrists are familiar with descriptions of psychotic experiences that almost like a lightning strike significantly and sometimes even permanently alter the life of their patients (see, e.g., Müller-Suur, 1962, p. 81f.). Whereas the mystical experience usually has an enriching effect and the mystic typically is able to integrate his sense of extraordinary insights into a mundane-spiritual existence (e.g., by embracing an unsentimental accept of all that is), the life of many patients with schizophrenia are almost brought to a standstill, i.e. they can neither integrate their delusional insights into their mundane existence nor leave these insights behind—metaphorically speaking, the temporality of existence freezes and the patients are perpetually stuck in the now, a kind of eternal present (kairos), with a fixed, immutable past and an absence of openness and projects for the future (Blankenburg, 1965; Bovet & Parnas, 1993; Minkowski, 1933). 4. The other reality in mysticism and schizophrenia A mystic and some patients with schizophrenia claim to have attained a profound and penetrating insight into another dimension of reality. As we have seen, these particular experiences reside outside the realm of ordinary experience, language, and rules of logic—as Conrad puts it, ‘outside the categories of comprehensibility’ (Conrad, 2002, p. 110). But what kind of reality are the mystic and the patient with schizophrenia claiming to be in touch with? Is it at all the same kind of reality or dimension of reality? A way of dismissing the entire issue would be simply to say that in both cases we are dealing with nothing else than individual psychological states, reducible in a future neuroscience to certain functional configurations of the ‘connectome’ (i.e. a total pattern of interacting neural circuits). Here, we suspend, at least temporarily, the plain reductive-naturalist assumption, and raise instead the question of the ‘objective’ or ontological status of the other reality. We first present some major philosophical accounts of this other reality in mysticism before turning to the psychotic world in schizophrenia. 4.1. Ultimate reality For the mystics, there is absolutely no doubt that the reality or truth, revealed in the enlightenment experience, is not merely a psychological state, a ‘subjective’ phantasm, but entails a form of objectivity (e.g., ‘I experienced a complete certainty that at that moment I saw things as they really were’ [N.M.]; ‘I saw no new thing but I saw all the usual things in a miraculous new light—in what I believe is their true light. . . I have looked into the heart of reality; I have witnessed the truth’ [Montague]; quoted in Stace, 1960, p. 72, p. 83f.). Stace (1960) suggests here that the reality, disclosed in mystical experiences, is not ‘objective’ (in the sense of lending itself to a third-person definition and description) but rather, as he puts it, ‘transsubjective’ (1960, p. 148). A remarkably similar claim can be found in Nishida (1990), who argues that the enlightenment experience (‘pure experience’) is neither subjective nor objective but ‘trans-individual’. For Stace, ‘transsubjectivity’ is far from an arbitrary feature: ‘the fact that self-transcendence is a part of the experience itself is the reason why the mystic is absolutely certain of its truth beyond all possibility of arguing him out of it. A significance and interpretation of any experience can be doubted, but the experience itself is indubitable’ (1960, p. 154).7 The notions of a ‘trans-subjective’ or ‘trans-individual’ experience suggest a sort of ‘quasi-objective’ ontological status of the reality that is being revealed. According to this line of thought, the kind of reality that manifests itself in the mystical experience is not just a psychological accomplishment of an individual consciousness; rather it appears to somehow exist in itself, independently of the subject, though its manifestation only may be instantiated by an individual consciousness. In other words, the mystics accede to an at least potentially shared, but normally hidden dimension of reality, which usually is so deeply buried within us that we are entirely unaware of it, perhaps only, if ever, glimpsing it in rarified forms in certain aesthetic experiences (Henry, 2009) or meditative states. Several philosophers have addressed this metaphysical dimension of reality. James (2013) suggests that the apparent unanimity of mystical experiences may be indicative of the existence of a level of ultimate reality, a level of ‘pure experience’. According to Fink (1995), the ultimate goal to which a rigorous performance of the phenomenological reduction leads is a discovery of a meta-ontic layer of the Absolute, a layer of transcendental ‘openness-to the world’. It is a layer preceding the phenomenological accounting for intentionality, noetic-noematic correlations, and consciousness as such—as Bertolini puts it, ‘[it] is an originative, constituting movement, beyond the field of beings, beyond the ontic field as such. The relation of absolute origination (. . .) is between Being and Nothingness’ (2014, p. 39). In his later work, Heidegger (1957) also struggles to think Being (Sein) in itself (Being as Being), i.e. Being detached from beings (entities [Seiendes]). He argues that Being grounds beings but Being itself is without ground, i.e., Being is ‘a ground without ground’, which he also calls an ‘abyss’ (Abgrund). This conception echoes the mystical notions of ‘Ungrund’ (Böhme) and the ‘wayless abyss’ or ‘abysmal’ (Ruysbroeck), which describe the experience of unio mystica. Caputo (1990) has aptly illuminated a certain kinship between Heidegger’s analysis of Being in itself and Eckhart’s notion of the Godhead. For both Heidegger and Eckhart, access to Being (or God) requires a radical openness, which involves divestment of all forms of thinking, imagery, and judgment, detachment from beings (Abschied vom Seienden), and letting Being be (Gelassenheit). 7 In schizophrenia, we find strikingly similar features with regard to primary delusional experiences, viz. certainty, incorrigibility, and the unquestionable truth or validity of the experience—as one of Jaspers’ patients puts it, ‘everything is so dead certain that no amount of seeing to the contrary will make me doubtful’ (1997, p. 100). 82 J. Parnas, M.G. Henriksen / Consciousness and Cognition 43 (2016) 75–88 Possibly the most substantive account of the other dimension of reality is found in the works of the famous Japanese philosopher, Nishida, gravitating around his core notion of ‘pure experience’ (borrowed from James [2013]). According to Krueger, ‘pure experience for Nishida is both the primordial foundation of consciousness and the ultimate ground of all reality’ (2006, p. 12). Thus, for Nishida, as for the mystics in general, the ‘other’ reality is not somehow on a par with the reality we experience in our everyday life but instead the ultimate ground of this reality.8 The ultimate ground is not outside us but within us—to see it we must look with a ‘reversed eye’ (Nishida adopts here Böhme’s expression [Nishida, 1990, p. 81]). Pure experience is the intuition of oneness in all (‘identity in difference’) and this oneness is quintessentially the primal flow of reality as an original, creative, unifying activity, which is the final ground or force of life, perhaps comparable to Bergson’s ‘élan vital’ (1908), Henry’s metaphysical notion of ‘Life’ (C’est moi la verité) (2008) or to Levinas’ notion of God as a radical alterity, ‘beyond Being’ (au de la d’étre) (1991). These philosophical accounts of ultimate reality converge, despite their various formulations and specific differences, to Nishida’s claim that pure experience is in fact ‘trans-individual’. In pure experience, we intuit the unifying activity that is the final ground of both consciousness and reality, the meta-ontic transcendental level proposed by Fink (1995). As we have seen, emptying and stillness of mind, which is the desired aim of most schools of meditation (Fasching, 2008), is generally considered a precondition for the emergence of unio mystica. If we accept Nishida’s claim that ultimate reality, intuited in pure experience, is the foundation of both reality and consciousness, then the following question becomes pertinent: what, more precisely, do we experience when the mind is brought to a halt? The answer is that we experience no thing and yet we do not experience nothing. In other words, when we let go of all objects or contents of consciousness, we become aware of consciousness itself, viz. ‘pure consciousness’. As Fasching argues, ‘consciousness is not a phenomenon among phenomena but the taking place of the phenomenality of phenomena’ or, as he also puts it, ‘the taking place of presence’ (2008, p. 467, p. 466). Usually, we are aware of objects or contents that are present to us (e.g., the object of my perception or the thought I entertain) but we are unaware of presence itself. As Fasching puts it, ‘[the] presence of presence is not another presence in addition to the presence of the object, but simply this very presence itself’ (2008, p. 475). In other words, pure consciousness is not some sort of object located at the periphery of our experience, usually eluding our attention, but is to be found at the very heart of our experience (but not as an object)—it is the appearing (presence) of what appears (presents itself) in our experience. Notably, there is no distance between appearing and what appears. As the taking place of presence, pure consciousness is present to us in each and everything that is present (Fasching, 2008, p. 476). This self-presence or selfmanifestation is, as Henry famously phrases it, ‘the essence of manifestation’ (1973, p. 143). All manifestation is thus conditioned upon an ‘invisible’, generative layer of self-revelation, a transcendental dimension that also makes possible contact with the Absolute. This meta-ontic dimension is inherent in all forms of manifestation, independently of whether or not these manifestations are ‘subjective’ (e.g., thoughts, feelings, etc.) or ‘objective’ (e.g., perception) in kind. Against this backdrop, we may reconsider the mystics’ ‘detachment’ from the shared-social world. It is not so much a withdrawal to another, inner or private world, but a move into the normally invisible world, i.e., into the very presence of the world, or, to put it differently, it is a move from an object-awareness into an awareness of pure consciousness, i.e. into the invisible, generative layer that makes any appearing possible. This meta-ontic or meta-noetic level is ‘trans-subjective’ and potentially available to us all. 4.2. The world of psychosis We will now reconsider some phenomenological and ontological aspects of the schizophrenic psychosis in the light of our discussion of the ‘objective’ or trans-subjective status of the mystical experience. Extrapolating from clinical experience, a claim can be made that many patients with schizophrenia experience being in touch with another, hidden, ontological dimension of reality—an experience that is not entirely unlike the mystics’ breakthrough into ultimate reality or experience of pure consciousness. The experience of unique ontological access or insight, in schizophrenia or elsewhere, is of course not by itself an index of mental disorder. Nonetheless, the articulation of the schizophrenic psychosis is often related to this experience of insight, which is perceptible in the nature of psychotic symptoms in schizophrenia (primary delusions, hallucinations, and phenomena of external influence). Most importantly, these symptoms have a peculiar ‘subjective’ (perhaps best expressed as ‘solipsistic’) character. Viewed in their noematic projective content (e.g., concrete delusions), they appear always short of a complete articulation of a properly transcendent status. They are, so to say, ‘unfinished’ or ‘insufficiently objective’. They never become fully inscribed into the texture of the intersubjective world.9 Rather, they retain residua of quasi-phenomenal fragments of immanence or traces of a fragmented self (Parnas, 2004)—as Sass aptly puts it, ‘[the] world of many schizophrenic-type patients is not, then, a flesh-and-blood world of shared action and risk but a mind’s-eye world’ (1994, p. 46).10 It is crucial to realize that expressions such as the ‘mind’s eye’ (Schreber) or ‘reversed eye’ (Nishida) do not refer to an act of introspection or imagination but convey a felt union of subjectivity and the other ontological dimension. 8 Hunt (2006) has argued for the affinity of mystical states with experiential aspects of the quantum level of reality. By contrast, a delusional system in non-schizophrenic psychosis is well integrated in the external world, its logic, persecutory mechanisms, and actors. 10 Sass employs here Schreber’s (2000) notion of seeing with the ‘mind’s eye’ (geistigen Auge), which he distinguishes from seeing with the ‘bodily eye’ (körperlichen Auge): ‘I use here the expression ‘seeing with the mind’s eye’ (. . .) because I cannot find a more suitable one in our human language. We are used to thinking all impressions we receive from the outer world are mediated through the five senses, particularly that all light and sound sensations are mediated through eye and ear. This may be correct in normal circumstances’ (Schreber, 2000, p. 120). Schreber coined the term to articulate his experience of continual communication, through ‘nerve-contact’ and ‘rays’, with God. Schreber’s notion of the ‘mind’s-eye’ seems to echo Böhme’s and Nishida’s notion of the ‘reversed eye’, reflecting a breakthrough into the ultimate ground of reality, i.e. the experience of the primal flow of reality as an original, creative meta-noetic activity or, following Faching, as the experience of the sources of presence of the world. 9 J. Parnas, M.G. Henriksen / Consciousness and Cognition 43 (2016) 75–88 83 Primary delusions and hallucinations in schizophrenia are not of this world (Merleau-Ponty, 2002, p. 395). They presuppose and entail a changed ontological position, free of the natural certitudes and axioms of space, time, causality, and noncontradiction, or, as phrased by Ratcliffe (2012), they occur in another ‘modal space’. For example, hallucinatory ‘voices’ are in general not woven into the fabric of the intersubjective world. Rather, they are experienced as intrinsically private, i.e. patients rarely believe that others also hear their ‘voices’ (Aggernæs, 1972). Moreover, the ‘voices’ are often ubiquitous and violate the physical constraints of the sensorial space (Henriksen, Raballo, & Parnas, 2015). In other words, ‘voices’ are not in the world but rather ‘superimposed’ (Merleau-Ponty, 2002, p. 395) on it and they are felt to be hyper-proximate (le sentiment de sur-proximité [Charbonneau, 2004]) to the patient’s innermost recesses, precluding taking a protective distance, shelter (le sentiment de désabritement) or flight—as one of our patients puts it, ‘I cannot shut her [the voice] out. She is always there’. In this regard, ‘voices’ resemble the character of haunting. It also merits attention that patients rarely (typically only in the acute exacerbations) conflate their psychotic convictions with their ordinary beliefs about worldly matters. Renee described it in the following way: [Little] by little I brought myself to confine to my friends that the world was about to be destroyed, that planes were coming to bomb and annihilate us [i.e. her delusion]. Although I often offered these confidences jestingly I firmly believed them (. . .) Nonetheless, I did not believe the world would be destroyed as I believed in real facts (Sechehaye, 1951, p. 14f.; our italics). Schreber made a similar remark, ‘I could even say with Jesus Christ: ‘My Kingdom is not of this world’; my so-called delusions are concerned solely with God and the beyond’ (Schreber, 2000, p. 371; our italics). These paradigmatic cases suggest that schizophrenic delusions, though linguistically resembling epistemic propositions, are not really epistemic claims about worldly matters (Cermolacce, Sass, & Parnas, 2010; Parnas, 2004) but rather attempts to frame and verbalize anomalous experiences of an already altered subjectivity (Škodlar, Henriksen, Sass, Nelson, & Parnas, 2013). The ‘world of psychosis’ in schizophrenia, epitomized by primary delusions, hallucinations, and phenomena of external influence, and apprehended by the ‘mind’s eye’, does not exist alongside the shared-social world as if they were somehow two different, yet somehow commensurable, distinct realities but is revealed to the patient as insights into the very ground of the shared-social world. A patient of ours offers the following illuminating account of his experience: There are two worlds. There is the unreal world, which is the world I am in and we are in. And then there is the real world. The only thing that is real in the unreal world is my own self. Everything else - buildings, trees, houses - is unreal. All other humans are extras. My body is part of the charade. There is a real world somewhere and from there someone or something is trying to control me by putting thoughts into my head or by creating (. . .) screaming voices inside my head. Many, if not a majority of patients with schizophrenia, appear to simultaneously live in two different worlds or ontological dimensions. Bleuler (1950) described crucial aspects of this phenomenon with the notion of ‘double book-keeping’ (Henriksen & Parnas, 2014). Due to the two worlds’ different (incommensurable) ontological status, patients typically seem to experience them as not conflicting, thereby allowing them to coexist and only occasionally to collide (Henriksen & Parnas, 2014). The balance and interpenetration of these two ontological perspectives are variously negotiated by the individual patients and as a function of different illness stages. Crucially, double bookkeeping is never associated with a disunity (dissociation) of phenomenal consciousness. The dual ontological nature of the psychotic consciousness was already noticed and emphasized by the first alienists-psychiatrists such as Pinel (1806). Around that time, Hegel also observed, in such coexistence of ontological frameworks, a mark of madness, an ‘inner perversion of self-consciousness’ (Hegel, 1998, §374–375). Possibly the most lucid first-person description of double bookkeeping is offered by Prof. Saks: It was at this point, I think, that my life truly began to operate as though it were being lived on two trains, their tracks side by side. On one track, the train held the things of the ‘real world’—my academic schedule and responsibilities, my books, my connection to my family (. . .) On the other track: the increasingly confusing and even frightening inner workings of my mind. The struggle was to keep the trains parallel on their tracks, and not have them suddenly and violently collide with each other (2007, p. 64f.). 5. Minimal self and the de-structuration of immanence in schizophrenia The phenomenal nature of self-disorders in schizophrenia (Section 2) suggests that we are dealing with a disorder of firstperson perspective (ipseity) that is far more fundamental than the ‘self-related’ complaints and behaviors or characterological traits encountered in the disorders outside the schizophrenia spectrum, e.g., affective psychosis, other mood, anxiety, and otherwise ‘neurotic’ or personality disorders. In these latter disorders, the ‘self-related’ complaints or behavioral dispositions indicate problems in the domain of psychological self-image or self-esteem—i.e. self-representational problems located at the level of personal, narrative or ‘extended’ selfhood.11 At this level, the issue of ipseity, i.e. of being a selfcoinciding subject of experience and action is never at stake. We have elsewhere proposed that a generative trait feature of schizophrenia is a disorder of ipseity or minimal (core) self (Cermolacce et al., 2007; Nelson, Parnas, et al., 2014; Sass & 11 The patient with schizophrenia may also exhibit self-representational problems, largely consequential to his basic ipseity disorder. 84 J. Parnas, M.G. Henriksen / Consciousness and Cognition 43 (2016) 75–88 Parnas, 2003 [viz. the Ipseity Disturbance Model]). In the following, we articulate what is implied by the notion of minimal self or ipseity, and how it is disordered in schizophrenia. We then further clarify the analogies and differences between schizophrenia and mysticism. The concept of minimal self is a ‘thin’ phenomenological notion, referring to the first-personal manifestation or givenness of experience (Zahavi, 2005, 2014)—a structure that assures the subjectivity of experience, often designated as ‘mineness’, ‘myness’, ‘for-me-ness’ (Hart, 2009; Henry, 1973; Klawonn, 1991; Zahavi, 2005). We will here propose two mutually implicative aspects or moments that constitute the minimal self, viz. a formal and an affective aspect. The first-person perspective may be considered as a formal feature. All experience is given in a structural configuration as ‘mine’, i.e. in my (first-person) perspective. This formal aspect of ‘perspective’ is often illustrated by an analogy to visual perception: a spatial cone of vision radiates out to the world from my embodied, absolute here, my elusive experiential pole or source of my perspective. However, this spatial model does not fit other modalities of intentionality such as thinking, remembering or feeling. In these latter cases, the first-person perspective manifests itself as an interpenetration of the intentional act (e.g., thinking or perceiving) and the pre-reflective sense of ‘I-me-myself’ implied in the act, i.e. a sense of self-presence, tacitly imbuing all mental activity. The sense of self-presence is an affective aspect of the minimal self. Thus, the minimal self is not just a pure form; as a pure form it could never be given in conscious experience. In other words, the minimal self invariably involves a pre-reflective, affective sense of self-presence, self-familiarity or self-intimacy that persists across time and changing modalities of consciousness, permeating any particular intentional act (Hart, 2009). This pre-reflective sense of ‘I-me-myself’ is experientially manifest through its structuring effect on the flow of experience; it is not an independent mental entity that can be phenomenally accessed in itself, apart from experience. Phenomenologists (and other philosophers [e.g., Janzen, 2008]) agree that what we here call the ‘minimal self’ is given non-relationally, non-inferentially or non-observationally (e.g., Husserl, Sartre, Merleau-Ponty, Henry, and Levinas). Some aspects of Henry’s account (1973, 1975) appear to be helpful in explicating the affective aspect of the minimal self.12 On Henry’s view, the minimal self is a radically passive self-revelation (un-instigated, non-willed, non-dyadic, and not an outcome of a relation between two distinct relata), yet dynamic, self-affecting, pathic pulse of immanence.13 This self-revelation acquires its inchoate singularity or proto-individuation, because the affectivity of auto-affection never surpasses itself or alters in nature despite its changing experiential manifestations (e.g., changing moods and feelings). Affectivity, irrespective of its particular occurrent manifestation, remains the same ‘se sentir soi-même’ (self-sensing of self) (Henry, 1973, p. 465). The life of consciousness entails an incessant immanent self-affection, imbuing the first-person perspective with its affective dimension, an experiential feel of self-presence or self-familiarity, and an inchoate sense of singularity. As Hart notes, the self-familiarity of the ‘I-me-myself’ is paradoxical: on the one hand, it is ‘propertyless’ because it resists property-mediated description. Yet, on the other hand, it is a sense of selfness that is foundational of our identity (Hart, 2009). It functions as a necessary nucleus around which more sophisticated, complex or ‘extended’ feelings and self-representation of identity are formed throughout life. The notion of minimal self must not be reified but rather understood as a grid, a template or a structure of the immanent life. In this sense, the minimal self and immanence are not distinct or separable entities in a relation of mutual exteriority. The minimal self constitutes the experiential life as a singular, unified field of awareness. Phrased differently, the minimal self imposes limits and limitations on the field of immanence, making its vicissitudes conform to the basic structures of consciousness. In this pre-reflective milieu, the very alterity or the ‘me-not me’ (self-other) distinction articulates itself as an edge for the formation of transcendent objects. It is also at this level of passivity that the primordial, tacit threads of operative intentionality, anchoring us in the world, are formed as the so-called ‘passive syntheses’ (Husserl, 1973), ‘common sense’ or ‘natural self-evidence’ (Blankenburg, 1971, 2001) or pre-reflective immersion in the world (‘Being-in’; Heidegger, 2007). In schizophrenia spectrum disorders (and in the experience of unio mystica as well), the formal aspect of the minimal self appears to be preserved, i.e. the patient continues to be the subject of his complaints and continues to employ the firstpersonal pronoun ‘I’, although he may be tempted to express himself, using the third-personal pronoun, e.g., ‘one is’ or ‘it thinks’ (Minkowski, 1927). Only rarely, we encounter self-reports that may be interpreted as being indicative of a quasicomplete obliteration of the formal aspect of the minimal self, where the very sense of being a perspectival pole is affected (see, e.g., Saks, 2007, p. 12f.). By contrast, the affective aspect of the minimal self, i.e. the automatic, pre-reflective sense of self-presence or self-familiarity is unstable and threatened in schizophrenia spectrum disorders, causing an incomplete saturation of the immanent life and resulting in a variety of experiential anomalies (incl. self-disorders). This failing of ‘self-sensing of self’ manifests itself as an unspecifiable ‘lack’ (sentiment d’íncompletude [cf. Janet, 1929]) or wavering of the normally unproblematic, tacit sense of ‘I-me-myself’. A crucial aspect of the diminished self-presence is ‘operative hyper-reflexivity’, i.e. normally tacit and lived moments of the pre-reflective flow of consciousness may now pop-up into awareness as quasi-autonomous, alien or intrusive entities (Parnas & Sass, 2011, p. 537). Thus the ‘incomplete’ minimal self enables normally silent or anonymous regions to emerge with alien prominence within the very intimacy of one’s own subjectivity—it is, as Ey puts it, ‘a modification within the self’ (1973, p. 417). The unreliability of the very sense of being a self-coinciding subject of awareness and action is closely associated, and most likely interdependent, with the disruptions of pre-reflective world-immersion or ‘common sense’. 12 13 A similar emphasis on affectivity as grounding ipseity is also present in Levinas (1979). Henry (2008) understands immanence in the sense of a ‘material’ or hyletic substance (matiére phenomenologique). J. Parnas, M.G. Henriksen / Consciousness and Cognition 43 (2016) 75–88 85 In our view, the immanence becomes de-structured in schizophrenia, (Ey, 1978), affecting its very limits and limitations, including the ‘me-not-me’ distinction. Normally, these limits or structures of the field of immanence are assured by the minimal self. The de-structuration of immanence distorts a smooth deployment of experience, the most conspicuous feature of which is perhaps the permeability of the self-other or self-world boundary manifest in experiences of transitivism (e.g., the patient may feel somehow transparent, without any barriers, ‘radically exposed’ [Henriksen, Škodlar, Sass, & Parnas, 2010] or ‘as if’ fusing with others or the surroundings). Most importantly in the present context, the failing of ‘self-sensing of self’ and the de-structuration of immanence seem to facilitate an articulation of another presence or radical alterity in the midst of the patient’s sphere of ownness (Henriksen & Parnas, 2014). In our view, this disturbing openness to another presence within the very intimacy of one’s own subjectivity is the phenomenological core of primary psychotic experience in schizophrenia and a crucial source of double bookkeeping. This alien presence often takes form of a projective ‘Other’ and a sense of a breakthrough into a hidden ontological domain. Through the patient’s ‘psychotic work’, the sense of another presence may eventually materialize into a persecuting, influencing or hallucinatory Other,14 which continues to be felt as hyper-proximate, because it originates and remains linked to the de-structured immanence—a predicament that Rogozinski called ‘a carnal tear, a crisis of the chiasm’ (2010, p. 208). It seems to us that the de-structured immanence is in a certain sense a variant of what the mystic strives to achieve through a series of willfully adopted behavioral and mental attitudes that gravitate around the efforts of self-effacement (effacement of the sense of self-presence) and detachment from the shared-social world. The unio mystica is, as we have seen, a state of structureless, boundless immanence, ‘pure experience’ or ‘taking place of phenomenality’. The unio mystica entails some form of a sense of harmonious fusion or co-presence with the Absolute, and after the termination of the mystical state, the mystic recovers his mundane intentional consciousness with its structures and limitations. In schizophrenia, however, this experiential process of de-structuration happens involuntarily and in a fragmented way, because both its origin and course are influenced by the enduring disorder of minimal self—this is also key to understanding the different effects and traces such experiences leave on the individual in mysticism and schizophrenia, respectively. Often, the de-structuration of immanence brings the patient experientially proximate to the noetic layer of pure experience or the taking place of presence—a layer that is normally hidden in ordinary conscious experience. Several authors (e.g., Tatossian, Kimura, and Nagai) have drawn attention to the fact that certain complaints of these patients indicate some degree of awareness of the noetic, constituting activity of consciousness. 6. Conclusion and implications Our comparison of mystical states and certain central features of schizophrenia makes it very clear that we are dealing with very distinct conditions, which nonetheless exhibit some informative phenomenological analogies. Apart from our emphasis on the de-structuration of immanence as a precondition for the emergence of unio mystica, our comparison contributes only little to the study of mysticism. However, we believe that the explored analogies have important bearings on the issues of consciousness and schizophrenia. Overall, our study demonstrates that schizophrenic experience is measurable within the scope of human experience, which is to say that human beings have an intrinsic potential for such experiences and, moreover, that such experiences actually reveal essential structural features of human consciousness. This is not an attempt to romanticize a serious and often very painful, debilitating mental disorder. However, keeping this humane dimension in mind may assist the efforts to destigmatize the illness. The nature of schizophrenia has always been a hotly disputed topic (Urfer-Parnas, Mortensen, & Parnas, 2010). The recent surge in phenomenological psychopathology re-emphasized the spectrum idea, i.e. a quasi-dimensional distribution of the illness phenotype, varying in intensity and gross clinical picture but sharing what classic psychiatrists called ‘fundamental symptoms’, e.g., schizophrenic autism (Bleuler, 1950; Henriksen et al., 2010; Parnas, 2011, 2012; Parnas & Jansson, 2015). Phenomenological studies of the last decades have rekindled the interest in the fundamental features of the schizophrenia spectrum disorders. These disorders affect the basic structures of consciousness such as self-hood (self-disorders), intentionality, and intersubjectivity (crisis of common sense) (Parnas & Bovet, 1991; Parnas, Bovet, & Zahavi, 2002; Salice & Henriksen, 2015; Sass & Parnas, 2003). Our comparative study allows for a certain theoretical integration of the basic disorders of schizophrenia, a grasp of these features in conjunction and mutual dependencies. The study highlights the fact that self-effacement and world-epoché (as diminished self-presence and presence in the world) seem to operate as a structural alteration that makes the immanent life de-structured and vulnerable to the emergence of an intrusive radical alterity within the minimal self. The self-alienation may articulate itself as a sense of access to another, hidden ontological reality or as revelatory, primarily pathic-noetic delusional or hallucinatory experiences. For the patients, these experiences, lived in their spheres of ownness, present an indubitable impact of truth. From a therapeutic perspective, it is therefore a mistake to dismiss the patient’s delusions as simple cognitive errors (‘false beliefs’) and his ‘lack of insight’ as simply a deficit of critical self-reflection (metacognition) or denial of illness. Moreover, on the basis of our study, an emerging therapeutic target consists in aiding the patient in balancing or negotiating an existence exposed to a double ontological orientation (double bookkeeping). 14 These psychotic symptoms can be seen to reflect radical experiences of being watched or listened to (delusions of being filmed or bugged), touched (delusions of external influence) or spoken to (auditory verbal hallucinations) in the innermost recesses of one’s self. 86 J. Parnas, M.G. Henriksen / Consciousness and Cognition 43 (2016) 75–88 Finally, it seems to us that the generative disorder of schizophrenia is to be sought in the altered organization of the basic structures of pre-reflective consciousness, which we have sought to elicit in this study. If this proposal were to be translated into an empirical research program, one would need to address the fundamental philosophical and neuroscientific questions concerning the nature of consciousness, its ontogenetic, developmental trajectory, and its pathological distortions in schizophrenia. Although a review of neuroscientific studies concerning the self and mystical states is beyond the scope of this paper (but see Nelson, Whitford, Lavoie, & Sass [2014a, 2014b]), we will like to emphasize a relatively recent approach in the neurosciences, i.e. neuro-phenomenology (Petitot, Varela, Pachoud, & Roy, 2000). In brief, it is a scientific approach that attempts to investigate correlations between subjective experience and patterns of brain activity. For instance, a recent review of neuroimaging research in meditative states has pointed to a great variety of findings and methodologies across the studies. Although identifying important methodological shortcomings, this review seems to demonstrate correlations between mediation and brain structure, region, and function, respectively (Tang, Hölzel, & Posner, 2015). A recent meta-analysis of 78 functional neuroimaging studies of meditation found reliably dissociable patterns of brain activation and deactivation (Fox et al., 2016). In the spirit of neuro-phenomenology, Northoff has advocated ‘a shift from a content- or function-based concept of self to a process-based view of the self’ (Northoff, Qin, & Feinberg, 2011, p. 55). More specifically, ‘The process that establishes a relation between the organism and a stimulus is called self-related processing. It is distinguished from its cognitive counterpart, self-referential processing, that takes the contents be they bodily, mental or autobiographical as given (and preexisting)’ (Northoff et al., 2011, p. 55). Thus, self-related processing is the most basic activity from which object and subject articulate themselves. Obviously, studying this level of brain activity is most adequately addressed during the brain resting state and therefore may offer a promising neural correlate to self-disorders: ‘In the same way that the basic disturbance of the self is present everywhere and affects all its various functions, the resting state, metaphorically speaking, ‘‘has its hands” in all kinds of neural processing related to different stimuli, tasks, and their respectively associated functions. In short, schizophrenia may be characterized by an overall presence of the ‘‘basic disturbance of the self”’ (Northoff, 2014, p. 395). Conflict of interest statement The authors declare no conflicts of interest. Acknowledgement MGH was funded in part by the Carlsberg Foundation (grant no. 2012010195). References Aggernæs, A. (1972). The experienced reality of hallucinations and other psychological phenomena. An empirical analysis. Acta Psychiatrica Scandinavica, 48, 220–238. Bagger, M. (2007). The uses of paradox. Religion, self-transformation, and the absurd. New York: Columbia University Press. Bergson, H. (1908). Essai sur les données immédiates de la conscience. Paris: Librairies Félix Alcan et Guillaumin. Bertolini, S. (2014). The forces of the cosmos before genesis and before life: Some remarks on Eugen Fink’s philosophy of the world. Analsecta Husserliana, 116, 37–46. Blankenburg, W. (1965). Zur Differentialphänomenologie der Wahnwarhrehmung. Nervenarzt, 36, 285–298. Blankenburg, W. (1971). Der Verlust der natürlichen Selbstverständlichkeit. Ein Beitrag zur Psychopathologie symptomarmer Schizophrenien. Stuttgart: Enke. Blankenburg, W. (2001). First steps toward a psychopathology of ‘common sense’ (A. L. Mishara, Trans.). Philosophy, Psychiatry, and Psychology, 8, 303–315. Bleuler, E. (1950). Dementia praecox or the group of schizophrenias (J. Zinkin & N. D. C. Lewis, Trans.). New York: International University Press. Bovet, P., & Parnas, J. (1993). Schizophrenic delusions: A phenomenological approach. Schizophrenia Bulletin, 19, 579–597. Buckley, P. (1981). Mystical experience and schizophrenia. Schizophrenia Bulletin, 7, 516–521. Burch, G. B. (1960). Respect for things. Aryan Path, 31, 484–487. Caputo, J. D. (1990). The mystical element in Heidegger’s thought. New York: Fordham University Press. Carter, R. E. (2013). The Kyoto school. An introduction. Albany, NY: SUNY Press. Cermolacce, M., Naudin, J., & Parnas, J. (2007). The ‘minimal self’ in psychopathology: Re-examining the self-disorders in the schizophrenia spectrum. Consciousness and Cognition, 16, 703–714. Cermolacce, M., Sass, L. A., & Parnas, J. (2010). What is bizarre in bizarre delusion? A critical review. Schizophrenia Bulletin, 36, 667–697. Charbonneau, G. (2004). Introduction à la phénoménologie des hallucinations. In G. Charbonneau (Ed.), Introduction à la phénoménologie des hallucinations (pp. 17–42). Paris: Circle Hermeneutique. Conrad, K. (2002). Die beginnende Schizophrenie. Versuch einer Gestaltanalyse des Wahns. Bonn: Edition Das Narrenschiff im Psychiatrie-Verlag. Deikman, A. J. (1971). Bimodal consciousness. Archives of General Psychiatry, 25, 481–489. Depraz, N. (2001). En quête d’une métaphysique phénoménologique: La référence henryenne à Maître Eckhart. In A. David & J. Greisch (Eds.), Michel Henry, L’Epreuve de la Vie (pp. 255–280). Paris, France: Les Éditions Cerf. Eckhart, M. (2009). The complete mystical works of Meister Eckhart (M. O’C. Walshe, Trans.). New York: Crossroad Publishing. Ey, H. (1973). Traite des hallucinations, Tome I et II. Paris: Masson. Ey, H. (1978). Consciousness: A phenomenological study of being conscious and becoming conscious (J. H. Flodstrom, Trans.). Bloomington; London: Indiana University Press. Fasching, W. (2008). Consciousness, self-consciousness, and meditation. Phenomenology and the Cognitive Sciences, 7, 463–483. Fink, E. (1995). Sixth Cartesian meditation. Bloomington: Indiana University Press. Forman, R. K. C. (Ed.). (1990). The problem of pure consciousness: Mysticism and schizophrenia. New York: Oxford University Press. Fox, K. C. R., Dixon, M. L., Nijeboer, S., Girn, M., Floman, J. L., Lifshitz, M., ... Christoff, K. (2016). Functional neuroanatomy of meditation: A review and metaanalysis of 78 functional neuroimaging investigations. Neuroscience and Biobehavioral Reviews, 65, 208–228. Gennart, M. (2011). Corporéité et présence. Jalons pour une approche du corps dans la psychose. Argenteuil: Le Cercle Hermeneutique. J. Parnas, M.G. Henriksen / Consciousness and Cognition 43 (2016) 75–88 87 Giudicelli, S. (1990). La question de la subjectivité dans le champ psychiatrique. In S. Giudicelli & G. Lanteri-Lausa (Eds.), Sujet et Subjectivité. Questions philosophiques, Questions psychopathologiques (pp. 9–24). Toulouse: Éditions Érès. Gouzoulis-Mayfrank, E., Habermeyer, E., Hermle, L., Steinmeyer, A., Kunert, H., & Sass, H. (1998). Hallucinogenic drug induced states resemble acute endogenous psychoses: Results of an empirical study. European Psychiatry, 13, 399–406. Griffoni, T. (2014). Atmospheres: Aesthetics of emotional spaces (S. Sanctis, Trans.). Farnham: Ashgate. Hart, J. G. (2009). Who one is. Book 1. Meontology of the ‘I’: A transcendental phenomenology. Berlin: Springer. Haug, E., Lien, L., Raballo, A., Bratlien, U., Øie, M., Andreassen, O. A., ... Møller, P. (2012). Selective aggregation of self-disorders in first-treatment DSM-IV schizophrenia spectrum disorders. Journal of Nervous and Mental Disease, 200, 632–636. Hegel, G. W. F. (1998). Phenomenology of spirit (A. V. Miller, Trans.). Delhi: Motilal Banarsidass Publishers. Heidegger, M. (1957). Der Satz vom Grund. Pfullingen: Verlag Günter Neske. Heidegger, M. (2007). Being and time (J. Macquarrie & E. Robinson, Trans.). Oxford: Blackwell. Henriksen, M. G., & Nordgaard, J. (2014). Schizophrenia as a disorder of the self. Journal of Psychopathology, 20, 435–441. Henriksen, M. G., & Nordgaard, J. (2016). Self-disorders in Schizophrenia. In G. Stanghellini & M. Aragona (Eds.), An experiential approach to psychopathology. What is it like to suffer from mental disorders. New York: Springer. Henriksen, M. G., & Parnas, J. (2012). Clinical manifestations of self-disorders and the Gestalt of schizophrenia. Schizophrenia Bulletin, 38, 657–660. Henriksen, M. G., & Parnas, J. (2014). Self-disorders and schizophrenia: A phenomenological reappraisal of poor insight and noncompliance. Schizophrenia Bulletin, 40, 542–547. Henriksen, M. G., Raballo, A., & Parnas, J. (2015). The pathogenesis of auditory verbal hallucinations in schizophrenia: A clinical-phenomenological account. Philosophy, Psychiatry, & Psychology, 22, 165–181. Henriksen, M. G., Škodlar, B., Sass, L. A., & Parnas, J. (2010). Autism and perplexity: A qualitative and theoretical study of basic subjective experiences in schizophrenia. Psychopathology, 43, 357–368. Henry, M. (1973). The essence of manifestation (G. Etzkorn, Trans.). The Hague: Martinus Nijhoff. Henry, M. (1975). Philosophy and phenomenology of the body (G. Etzkorn, Trans.). The Hague: Martinus Nijhoff. Henry, M. (2008). Material phenomenology (S. Davidson, Trans.). New York: Fordham University Press. Henry, M. (2009). Seeing the invisible (S. Davidson, Trans.). London: Continuum. Hunt, H. T. (2003). Lives in spirit. Precursors and dilemmas of a secular western mysticism. New York: State University of New York Press. Hunt, H. T. (2006). The truth value of mystical experience. Journal of Consciousness Studies, 13, 5–43. Husserl, E. (1973). Experience and judgment (J. S. Churchill & K. Ameriks, Trans.). London: Routledge. James, W. (2013). Varieties of religious experience. Newburyport: Dover Publications. Janet, P. (1929). De l’angoisse à l’extase. Paris: Felix Alcan. Janzen, G. (2008). The reflexive nature of consciousness. Amsterdam: John Benjamins. Jaspers, K. (1997). General psychopathology (J. Hoenig & M. W. Hamilton, Trans.). London: Johns Hopkins University Press. Kierkegaard, S. (1992). Concluding unscientific postscript to philosophical fragments (H. V. Hong & E. H. Hong, Trans.). New Jersey: Princeton University Press. Kierkegaard, S. (1998). The moment and late writings (H. V. Hong & E. H. Hong, Trans.). New Jersey: Princeton University Press. Kimura, B. (2001). Cogito and I: A bio-logical approach. Philosophy, Psychiatry, & Psychology, 8, 331–336. Klawonn, E. (1991). Jeg’ets Ontologi. En Afhandling om Subjektivitet, Bevidsthed og Personlig Identitet. Odense: Odense Universitetsforlag. Krueger, J. W. (2006). The varieties of pure experience: William James and Kitaro Nishida on consciousness and embodiment. William James Studies, 1, 1–37. Lacoste, J.-Y. (2004). Experience and the absolute. Disputed questions on the humanity of man (M. Raftery-Skehan, Trans.). New York: Fordham University Press. Levinas, E. (1979). Totality and infinity (A. Lingis, Trans.). Pittsburgh: Martinus Nijhoff Publishers and Duquesne University Press. Levinas, E. (1991). Otherwise than being or beyond essence (A. Lingis, Trans.). Dordrecht: Kluwer Academic Publishers. Marion, J.-L. (1991). Dieu sans l’être. Paris: PUF. Merleau-Ponty, M. (2002). Phenomenology of perception. London: Routledge. Minkowski, E. (1927). La Schizophrénie. Psychopathologie des Schizoïdes et des Schizophrènes. Paris: Payot. Minkowski, E. (1933). Le temps vécu, etudes phénoménologiques et psychopathologiques. Paris: Collection de l’évolution psychiatrique. Morgan, B. (2013). On becoming god. Late medieval mysticism and the modern western self. New York: Fordham University Press. Müller-Suur, H. (1950). Das Gewissheitsbewusstsein beim schizophrenen und beim paranoischen Wahnerleben. Fortschritte der Neurologie, Psychiatrie, und ihrer Grenzgebiete, 18, 44–51. Müller-Suur, H. (1962). Das Schizophrene als Ereignis. In H. Kranz (Ed.), Psychopathologie Heute (pp. 81–93). Stuttgart: Georg Thieme Verlag. Møller, P., & Husby, R. (2000). The initial prodrome in schizophrenia: Searching for naturalistic core dimensions of experience and behavior. Schizophrenia Bulletin, 26, 217–232. Nagai, M. (1991). Naisei no kôzô: Seishin byôrigakuteki kôsatsu. Iwanami Shoten. Nelson, B., Parnas, J., & Sass, L. A. (2014). Disturbance of minimal self (ipseity) in schizophrenia: Clarification and current status. Schizophrenia Bulletin, 40, 479–482. Nelson, B., & Sass, L. A. (2008). The phenomenology of the psychotic break and Huxley’s trip: Substance use and the onset of psychosis. Psychopathology, 41, 346–355. Nelson, B., Thompson, A., & Yung, A. R. (2012). Basic self-disturbance predicts psychosis onset in the ultra high risk for psychosis ‘prodromal’ population. Schizophrenia Bulletin, 38, 1277–1287. Nelson, B., Whitford, T. J., Lavoie, S., & Sass, L. A. (2014a). What are the neurocognitive correlates of basic self-disturbance in schizophrenia? Integrating phenomenology and neurocognition. Part 1 (Source monitoring deficits). Schizophrenia Research, 152, 12–19. Nelson, B., Whitford, T. J., Lavoie, S., & Sass, L. A. (2014b). What are the neurocognitive correlates of basic self-disturbance in schizophrenia? Integrating phenomenology and neurocognition: Part 2 (aberrant salience). Schizophrenia Research, 152, 20–27. Nishida, K. (1990). An inquiry into the good (A. Masao & C. Ives, Trans.). New Haven: Yale University Press. Nordgaard, J., & Parnas, J. (2014). Self-disorders and schizophrenia spectrum: A study of 100 first hospital admissions. Schizophrenia Bulletin, 40, 1300–1307. Northoff, G. (2014). Unlocking the brain. Consciousness (Vol. 2). New York: Oxford University Press. Northoff, G., Qin, P., & Feinberg, T. E. (2011). Brain imaging of the self – Conceptual, anatomical and methodological issues. Consciousness and Cognition, 20, 52–63. Olin, S. C., Mednick, S. A., Cannon, T., Jacobsen, B., Parnas, J., Schulsinger, F., & Schulsinger, H. (1998). School teacher ratings predictive of psychiatric outcome 25 years later. British Journal of Psychiatry Supplement, 172, 7–13. Overgaard, S. (2015). How to do things with brackets: The epoché explained. Continental Philosophy Review, 48, 179–195. Oxman, T. E., Rosenberg, S. D., Schnurr, P. P., Tucker, G. J., & Gala, G. (1988). The language of altered states. Journal of Nervous Mental Disease, 176, 401–408. Parnas, J. (2004). Belief and pathology of self-awareness: A phenomenological contribution to the classification of delusions. Journal of Consciousness Studies, 11, 148–161. Parnas, J. (2007). Subjectivity in schizophrenia: The minimal self is too small. In A. Grøn, I. Damgaard, & S. Overgaard (Eds.), Subjectivity and transcendence (pp. 55–70). Tübingen: Mohr Siebeck. Parnas, J. (2011). A disappearing heritage: The clinical core of schizophrenia. Schizophrenia Bulletin, 37, 1121–1130. Parnas, J. (2012). The core Gestalt of schizophrenia. World Psychiatry, 11, 67–69. Parnas, J., & Bovet, P. (1991). Autism in schizophrenia revisited. Comprehensive Psychiatry, 32, 7–21. Parnas, J., Bovet, P., & Zahavi, D. (2002). Schizophrenic autism: Clinical phenomenology and pathogenetic implications. World Psychiatry, 1, 131–136. Parnas, J., Carter, J., & Nordgaard, J. (2016). Premorbid self-disorders and lifetime diagnosis in the schizophrenia spectrum: A prospective high-risk study. Early Intervention in Psychiatry, 10, 45–53. 88 J. Parnas, M.G. Henriksen / Consciousness and Cognition 43 (2016) 75–88 Parnas, J., & Handest, P. (2003). Phenomenology of anomalous experiences in early schizophrenia. Comprehensive Psychiatry, 44, 121–134. Parnas, J., Handest, P., Jansson, L., & Saebye, D. (2005). Anomalous subjective experience among first-admitted schizophrenia spectrum patients: Empirical investigation. Psychopathology, 38, 259–267. Parnas, J., Handest, P., Saebye, D., & Jansson, L. (2003). Anomalies of subjective experience in schizophrenia and psychotic bipolar illness. Acta Psychiatrica Scandinavica, 108, 126–133. Parnas, J., & Jansson, L. B. (2015). Self-disorders: Clinical and conceptual implications for the diagnostic concept of schizophrenia. Psychopathology, 48, 332–338. Parnas, J., & Jørgensen, A. (1989). Premorbid psychopathology in schizophrenia spectrum. British Journal of Psychiatry, 155, 623–627. Parnas, J., & Henriksen, M. G. (2014). Disordered self in the schizophrenia spectrum: A clinical and research perspective. Harvard Review of Psychiatry, 22, 251–265. Parnas, J., Møller, P., Kircher, T., Thalbitzer, J., Jansson, L., Handest, P., & Zahavi, D. (2005). EASE: Examination of anomalous self-experience. Psychopathology, 38, 236–258. Parnas, J., Raballo, A., Handest, P., Jansson, L., Vollmer-Larsen, A., & Saebye, D. (2011). Self-experience in the early phases of schizophrenia: 5-year Follow-up of the Copenhagen Prodromal Study. World Psychiatry, 10, 200–204. Parnas, J., & Sass, L. A. (2011). The structure of self-consciousness in schizophrenia. In S. Gallagher (Ed.), The Oxford handbook of the self (pp. 521–546). Oxford: Oxford University Press. Parnas, J., Schulsinger, F., Schulsinger, H., Mednick, S. A., & Teasdale, T. W. (1982). Behavioral precursors of schizophrenia spectrum: A prospective study. Archives of General Psychiatry, 396, 658–664. Parsons, W. B. (1999). The enigma of the oceanic feeling. Revisioning the psychoanalytic theory of mysticism. New York: Oxford University Press. Petitot, J., Varela, F. J., Pachoud, B., & Roy, J.-M. (2000). Naturalizing phenomenology. Issues in contemporary phenomenology and cognitive science. Stanford, Calif.: Stanford University Press. Pinel, P. (1806). A treatise on insanity (D. D. Davids, Trans.). Sheffield: W. Todd. Podvoll, E. M. (1979). Psychosis and the mystic path. Psychoanalytic Review, 66, 571–590. Raballo, A., & Parnas, J. (2012). Examination of anomalous self-experience: Initial study of the structure of self-disorders in schizophrenia spectrum. Journal of Nervous and Mental Disease, 200, 577–583. Ratcliffe, M. (2012). Phenomenology as a form of empathy. Inquiry, 55, 473–495. Rogozinski, J. (2010). The ego and the flesh. An introduction to egoanalysis (R. Vallier, Trans.). Stanford, California: Stanford University Press. Saarinen, J. A. (2014). The oceanic feeling. A case study in existential feeling. Journal of Consciousness Studies, 21, 196–217. Saks, E. R. (2007). The center cannot hold. New York: Hyperion. Salice, A., & Henriksen, M. G. (2015). The disrupted ‘we’: Schizophrenia and collective intentionality. Journal of Consciousness Studies, 22, 145–171. Sass, L. A. (1992). Madness and modernism. Insanity in the light of modern art, literature, and thought. Harvard: Harvard University Press. Sass, L. A. (1994). Paradoxes of delusion: Wittgenstein, Schreber, and the schizophrenic mind. Ithaca, NY: Cornell. Sass, L. A., & Parnas, J. (2003). Schizophrenia, consciousness, and the self. Schizophrenia Bulletin, 29, 427–444. Schneider, K. (1950). Klinische psychopathologie. Stuttgart: Georg Thieme Verlag. Schreber, D. P. (2000). Memoirs of my nervous illness (I. Macalpine & R. A. Hunter, Trans.). New York: New York Review of Books. Schürmann, R. (2001). Wandering joy. Meister Eckhart’s mystical philosophy. Great Barrington, MA: Lindisfarne Books. Sechehaye, M. (1951). Autobiography of a schizophrenic girl (G. Rubin-Rabson, Trans.). New York: Grune & Stratton. Škodlar, B., & Ciglenečki, J. (2015). Psychose als mißglücktes Abenteuer. Mystische Erfahrungen und ihr psychotherapeutisches Potential. In S. Grätzel & J. E. Schlimme (Eds.), Psycho-logik 10. Abenteuer und Selbstsorge (pp. 154–169). Freiburg/München: Verlag Karl Albe. Škodlar, B., Henriksen, M. G., Sass, L. A., Nelson, B., & Parnas, J. (2013). Cognitive-behavioral therapy for schizophrenia: A critical evaluation of its theoretical framework from a clinical-phenomenological perspective. Psychopathology, 46, 249–265. Stace, W. T. (1960). Mysticism and philosophy. London: MacMillan. Steinbock, A. J. (2007). Phenomenology and mysticism. Bloomington, IN: Indiana University Press. Straus, E. (1935). Vom Sinn der Sinne. Ein Beitrag zur Grundlegung der Psychologie. Berlin: Julius Springer. Tang, Y.-Y., Hölzel, B. K., & Posner, M. I. (2015). The neuroscience of mindfulness meditation. Nature Reviews Neuroscience, 16, 213–225. Tatossian, A. (1978). La phénoménologie des psychoses. Paris: Masson. Tatossian, A. (2014). Psychiatrie phénoménologique. Paris: MJW Fédition. Tyrka, A. R., Cannon, T. D., Haslam, N., Mednick, S. A., Schulsinger, F., Schulsinger, H., & Parnas, J. (1995). The latent structure of schizotypy: I. Premorbid indicators of a taxon of individuals at risk for schizophrenia-spectrum disorders. Journal of Abnormal Psychology, 104, 173–183. Urfer-Parnas, A., Mortensen, E. L., & Parnas, J. (2010). Core of schizophrenia: Estrangement, dementia or neurocognitive disorder? Psychopathology, 43, 300–311. Zahavi, D. (2005). Subjectivity and selfhood. Investigating the first-person perspective. Cambridge, MA: MIT Press. Zahavi, D. (2014). Self and other. Exploring subjectivity, empathy, and shame. Oxford: Oxford University Press.

Consciousness and Cognition 19 (2010) 802–815 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Retro- and prospection for mental time travel: Emergence of episodic remembering and mental rotation in 5- to 8-year old children q Josef Perner *, Daniela Kloo, Michael Rohwer Department of Psychology, University of Salzburg, Austria a r t i c l e i n f o Article history: Available online 22 July 2010 Keywords: Episodic memory Mental rotation Development Prospection Theory of mind Preschool period Mental time travel a b s t r a c t We investigate the common development of children’s ability to ‘‘look back in time” (retrospection, episodic remembering) and to ‘‘look into the future” (prospection). Experiment 1 with 59 children 5 to 8.5 years old showed mental rotation, as a measure of prospection, explaining specific variance of free recall, as a measure of episodic remembering (retrospection) when controlled for cued recall. Experiment 2 with 31 children from 5 to 6.5 years measured episodic remembering with recall of visually experienced events (seeing which picture was placed inside a box) when controlling for recall of indirectly conveyed events (being informed about the pictures placed inside the box by showing the pictures on a monitor). Quite unexpectedly rotators were markedly worse on indirect items than non-rotators. We speculate that with the ability to rotate children switch from knowledge retrieval to episodic remembering, which maintains success for experienced events but has detrimental effects for indirect information. Ó 2010 Elsevier Inc. All rights reserved. 1. Introduction We naturally speak of ‘‘looking back on past events” or ‘‘looking forward to the future.” Clearly, we cannot literally look back or forward in time (Martin, 2001). We cannot see the future or the past in the same way as we see an event unfolding in front of us or behind us. The best we can do to capture past or future experiences of an event is to re-experience the event (retrospection), or imagine experiencing a future event (prospection).1 The ability to retrospect in this sense has become a central feature for episodic memory with Tulving’s (1985) introduction of ‘‘autonoetic consciousness”, namely the awareness that remembering consists of ‘‘calling back into consciousness a seemingly lost state that is then ‘immediately recognized as something formerly experienced’ (Ebbinghaus, 1885, p. 1).” Philosophers spoke of ‘‘experiential memory,” at least since Locke (Owens, 1996). Although the different terms all capture the phenomenon adequately, and ‘‘retrospection” does so in nice juxtaposition to ‘‘prospection,” we prefer ‘‘episodic remembering” as our standard term. The choice of ‘‘remembering”, rather than ‘‘memory” is to emphasise that it is more than just q This article is part of a special issue of this journal on Self, Other and Memory. * Corresponding author. E-mail address: josef.perner@sbg.ac.at (J. Perner). 1 In response to an anonymous reviewer we need to point out that the precise form of how the future is addressed in different prospection studies varies. Sometimes it is a concern about a specific future event (e.g., on the way to the puzzle room I have to think taking the needed implement with me; Suddendorf & Nielsen, 2009) but sometimes a timeless assumption suffices that one can treat as placed in the future (e.g., When(-ever) I go to a hibernal resort (tomorrow) I will take my coat and not the swimsuit; Atance & Meltzoff, 2005). In both cases the events one imagines experiencing (into which one projects oneself) are imagined possibilities (i.e., although one intends to do the puzzle, it is not a fact that one will) in contrast to retrospection, where one re-experiences (projects oneself into) a past event that has actually taken place. By precedent and due to this commonality we keep using the term ‘‘prospection” for experiencing imagined events in contrast to retrospection for re-experiencing actual events. 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.06.022 J. Perner et al. / Consciousness and Cognition 19 (2010) 802–815 803 retrieval of knowledge about a past episode. It is a re-experiencing of that episode. This distinction is also brought out by the notion of ‘‘mental time travel” (MTT: Suddendorf & Corballis, 1997; Wheeler, Stuss, & Tulving, 1997): one has to not just retrieve information about the past or think about the likely future. It requires projecting oneself as an experiencing agent into the past or future. This distinction is also akin to the difference between having a theory of mind (Churchland, 1984; Gopnik & Astington, 1988) as opposed to simulating one’s own (or other people’s) mental processes (Goldman, 2006; Gordon, 1986; Heal, 1986). There is the strong and widespread, but not uncontested, intuition that these three abilities, theory of mind, episodic remembering, and prospection belong together. They may be uniquely human abilities (Gilbert & Wilson, 2007; Roberts, 2002; Roberts et al., 2008; Suddendorf & Busby, 2003; Suddendorf & Corballis, 1997, 2007; Wheeler et al., 1997). There is now also fast growing empirical evidence from different quarters that episodic remembering (often investigated under the label or as part of autobiographical memory), prospection, and theory of mind share a developmental schedule, and common neural substrate as shown in coactivation patterns and in common deficits in clinical cases. 1.1. Brain imaging Spreng, Mar, and Kim (2008) meta-analyzed, among other areas, all available brain imaging studies on theory of mind (n = 30, as a random selection of 50), autobiographical memory (n = 19), and prospection (n = 6). Direct overlap was observed in the medial temporal lobe (left parahypocampal gyrus: BA 36), medial parietal regions (precuneus, posterior cingulate, bilaterally: BA 31), left temporo-parietal junction (BA 39, touching on BA 19), medial prefrontal cortex (frontal pole: BA 10). Convergence within same Brodmann areas were also observed in right TPJ, left ventrolateral prefrontal cortex (BA 47), medial prefrontal cortex, and rostral anterior cingulate (BA 32), and lateral temporal lobe (BA 21, 22) especially left. All these regions tend to also be activated by navigation problems and default processing (areas that tend to be more strongly activated in the absence than in the presence of external stimulation). Two theories have been proposed as to the common denominator underlying these common activations. Hassabis and Maguire (2007) suggested that all these tasks require scene-construction and Buckner and Carroll (2007) that they all require projection of self into different time points or spatial locations. 1.2. Brain injury Patients without autonoetic consciousness in Tulving’s sense, i.e., amnesics with a special impairment of episodic remembering and autobiographical memory, have been reported to also have severe deficits of prospection, Patients K.C. (Tulving, 1985), R. (Stuss, 1991), M.L. (Levine et al., 1998), and D.B. (Klein, Loftus, & Kihlstrom, 2002). Loss of autonoetic consciousness does, however, not lead inevitably to an impairment in theory of mind (patients K.C. and M.L.: Rosenbaum, Stuss, Levine, & Tulving, 2007). This finding does not preclude theory of mind being necessary for autonoetic consciousness and episodic remembering, in particular, it may be crucial as a developmental requirement or linkage. In fact, there is growing evidence that theory of mind development around 4 and 5 years is linked to both, episodic remembering as well as prospection. 1.3. Development There are studies showing a specific relationship between advances in theory of mind and free recall as a measure of episodic remembering in relation to cued recall (Tulving, 1985). Perner and Ruffman (1995) were able to show that between 3 and 5 years, children’s improvement on free recall correlates significantly with their understanding of how knowledge depends on experience. Even when cued recall and verbal intelligence were partialled out, correlations stayed above .30. Tasks used included children’s ability to explain why they know the contents of a box (How-do-you-know test: Wimmer, Hogrefe, & Perner, 1988a; Wimmer, Hogrefe, & Sodian, 1988b), to distinguish a lucky guess from proper knowledge (Miscione, Marvin, O’Brien, & Greenberg, 1978), and to understand which sense modality to use to find out about colour or weight of an object (O’Neill et al., 1992). These results were largely replicated by Naito (2003) on a Japanese sample. She also found a relationship between free recall and children’s ability to understand when they had learned a fact (Taylor, Esbensen, & Bennett, 1994). Perner, Kloo, and Gornik (2007) used a different measure of episodic remembering. They contrasted recall of experienced events with recall of indirectly conveyed events. In the experience condition children put cards with drawings of simple objects into a box. In the indirect-information condition they were blindfolded and so could not see which cards they put inside. They were afterwards shown on a monitor the pictures that were on these cards (information about individual cards was thus indirectly conveyed). The reasoning was that children can have an episodic memory of putting a particular card inside the box only when they experienced putting that card into the box. Only with this experience can they later re-experience their action. When blindfolded, they cannot experience which card they had put inside. When later shown what was on these cards, they can only infer that they put that card inside the box. Hence there is no experience of putting that card inside, which they could re-experience. Free recall of experienced events correlated with performance on various theory of mind tasks (How-do-you-know, When-did-you-learn, and the modality-specificity test). Sprung (2008) and Sprung and Harris (2009) found that theory of mind abilities (especially introspective understanding) modulates children’s ability to report 804 J. Perner et al. / Consciousness and Cognition 19 (2010) 802–815 on their intrusive thoughts, which in case of traumatic experiences (e.g., Hurricane Katrina victims) provides a basis for outgrowing the trauma. Rapidly growing evidence suggests that children’s prospective abilities develop at about the same age. Moore, Barresi, and Thompson (1998, Experiment 1) reported some correlations between 3- and 4-year-old children’s understanding of desire and belief and of the benefits of increased but delayed reward. Bischof-Köhler (2000) investigated several abilities in a group of 3- to 4.5-year-olds with a large battery of tasks including understanding of duration, theory of mind (false belief and deception), planning (shopping, need for preparation), delay of gratification, and motivational conflict. These abilities all underwent a noticeable improvement from 3 to 4.5 years and correlated significantly even when children’s age was partialled out. Atance and O’Neill (2005) investigated 3- to 4-year-olds’ ability to think about what to pack for a trip. About half of these children went beyond packing typical (scripted) or attractive things. They also thought of providing for possible eventualities (e.g., pack telephone to contact someone in case of an emergency). In another task children were given the beginning of a drawing, e.g., a straight line, and then were asked what they wanted to draw on the basis of this first element. There was a strong correlation r = .65 between children’s ability to anticipate events in the trip task and their choice of a drawing that was feasible with the given element (e.g., a sun but not a ladder starting from a circle) even when language ability was controlled. Atance and Meltzoff (2005) tested for mental time travel and found that 3 year olds above chance chose a suitable item (e.g., winter coat) for going to a snowy place (as shown in a photograph) and they were able to explain their choice by reference to their future need created by that environment. By 4 and 5 years all children could do this. It remains unclear whether this shows children’s ability to project themselves into the future state of going to a wintery landscape. They might just know about the need for winter coats in a cold environment. Atance and Meltzoff (2006) overly satiated children on salty pretzels, which made them change their preference from pretzels to water. Even 5 year old children showed little sign of understanding that the next day their preference would return to pretzels. They predicted a lasting preference for water. Busby and Suddendorf (2005) simply asked 3- to 5-year-old children what they had done yesterday and what they were likely to do tomorrow. Discounting general, scripted answers (e.g., ‘‘I played”) there was a marked improvement in children’s ability to report past and likely future episodes. Suddendorf and Nielsen (2009) tested 3- and 4-year-olds on a simple problem (e.g., use a triangular key to open a box). When the key was not available children went to another room and played for 15 min. Before returning to the first room they were given the choice among three different keys (one triangular) and asked which one they would like to bring with them to the room. Three-year-olds chose at random, while 4-year-olds chose the triangular key above chance (with plenty of room for improvement). Atance and Jackson (2009) assessed different aspects of future thinking in 3- to 5-year-old preschoolers, including mental time travel (Atance & Meltzoff, 2005; Busby & Suddendorf, 2005), delay of gratification, planning (simple version of Tower of Hanoi, Carlson, Moses, & Claxton, 2004), and prospective memory (remembering to do something after finishing something else: Kvavilashvili, Messer, & Ebdon, 2001). Children improved on all these tasks and performance between all tasks was correlated but correlations were dependent on age and receptive vocabulary. Ford, Shum, and Driscoll (2009) found that prospective memory performance in 4- to 6-year-olds was related to inhibitory ability and understanding false beliefs. Finally, Russell, Alexis, and Clayton (2010) let 3–5 year old children play blow-football for which one needed to bring a straw for blowing and as player on the blue (as opposed to the red) side a box to stand on as that side of the table was otherwise too high for children at this age. Children were asked questions framed in the past, present and future about what they themselves or another child needed to play on the blue side. The variable was whether they thought of both required items or not. In case children gave different answers for themselves than for their peer, the authors reasoned, this is a likely sign that children bring to bear their own perspective when imagining how they had played, or were going to play, i.e., project themselves into the role of player. Systematic comparison of self and other was only carried out for present (Experiment 4) and future (Experiments 2 and 3). In particular, 4-year-olds gave similar answers for self and other in the present and for other in the future, but ignored their own needs for the future. This was interpreted as evidence that between 3 and 5 years children become able to project themselves into future situations. Supportive evidence for this age trend comes from Prencipe and Zelazo (2005), who found that 3 year olds make wise choices of delaying an immediate small reward for a larger one later when choosing for another person but not for themselves. This discrepancy ceases around 4 years. 1.4. Autism The spectrum of autism consists of a developmental disorder that affects the very triad of theory of mind, retrospection, and prospection. Children with autism spectrum disorder tend to have problems with basic theory of mind tasks (Baron-Cohen, Leslie, & Frith, 1985; Perner, Frith, Leslie, & Leekam, 1989) or the least impaired cases (Asperger’s syndrome) have deficits appreciating subtle mental interactions like sarcasm, jokes, etc. (Happé, 1994). They also have executive deficits (Ozonoff, Pennington, & Rogers, 1991; for a review, see, Hill, 2004), which tend to be more pronounced for planning (prospection) than for inhibition. When tested for memory their free recall seems to be specifically impaired over cued recall and recognition (Boucher & Lewis, 1989; Boucher & Warrington, 1976; Bowler, Gardiner, & Grice, 2000; Tager-Flusberg, 1991) and even the highest functioning individuals with Asperger syndrome give fewer ‘‘remember” judgments in recognition tasks (Bowler, Gardiner, & Gaigg, 2007) than unimpaired people. J. Perner et al. / Consciousness and Cognition 19 (2010) 802–815 805 1.5. Sharpening the issues The neurophysiological and developmental evidence leaves little doubt that there is some common ground for theory of mind, episodic remembering and prospective abilities. Whether the co-emergence of these different abilities and their sharing of common cerebral structures is good evidence for mental time travel, i.e., for the ability to re- and pre-experience past or future events as the common basis remains an open question. The evidence for that claim is weak in two respects: few of the reviewed studies attempt to isolate specific cases of episodic remembering (retrospection: re-experiencing a past situation by projecting oneself into the past) as opposed to merely retrieving knowledge about the past, and of projecting oneself into a future experience (prospection) as opposed to knowing what is likely to happen in the future or knowing what is needed in an imagined future situation. Consequently, the developmental synchrony could easily be explained by quite general features, e.g., thinking about the past and future. The contrast between free and cued recall does provide some measure for teasing out the retrospective aspects of memory. Its usefulness is based on Tulving’s (1985) argument that free recall is more dependent on the availability of episodic traces than is cued recall. Hence, if free recall correlates specifically with progress in theory of mind when controlling for cued recall, then that correlation ought to be due to changes in episodic remembering (Perner, 1990) and not some more general ability to think about or retrieve information about the past. The link between free recall and episodic remembering, one has to admit, is rather weak. Free recall does not depend completely on episodic remembering (even in free recall a few items can come to mind automatically). Moreover, free recall is not exclusively helped by episodic remembering, which can also enhance cued recall. The contrast between memory of experienced events and indirectly conveyed events (Perner et al., 2007) is somewhat sharper in this respect. It takes advantage of the fact that only experienced events can possibly be reexperienced. Indirectly conveyed events should, therefore, not profit at all from the developing ability to remember episodically. The self-other contrast used by Russell et al. (2010) helps isolate the prospective aspects of anticipating future needs. It also allows for projection into the past, unfortunately, not for remembering any past event but only for figuring out what was needed for a past action as can be seen from their example test question: ‘‘. . . point to the two things you think the little girl had to have to play blow-football on the blue side?” (Section 2.1.3). Even for prospection this method depends on the risky background assumption that the difference between predicting one’s own future needs and predicting someone else’s future needs reflects an inability of projection, which is primarily applicable to oneself and not to others. For believers in simulation theory projection is also applied in the case of others. More generally, the difference can be due to factors that tend to interfere more strongly in the case of self than other. In sum, the claim for a specific developmental relationship between the ability to retrospect (episodic remembering) and prospection rests on evidence that can be given a more general interpretation (e.g., thinking about different times) or fails to provide the needed correlations: Perner and Ruffman (1995), Naito (2003) and Perner et al. (2007) only provide evidence for retrospection and ToM but not prospection. Russell et al. only provide evidence for prospection. Our prime objective is to provide evidence for a specific developmental relationship between retrospection (episodic remembering) and prospection. To capture retrospection we use the free-cued recall contrast in our first experiment and the experienced-indirectly conveyed contrast in our second experiment. For measuring children’s prospective abilities (the study by Russell et al. was published well after our research was conceived and conducted) we relied on children’s ability to engage in mental rotation. We used the child appropriate simplification by Estes (1998) of the original tasks by Shepard and Metzler (1971). Children were shown pairs of teddy bears (monkeys in Estes’ original study) with one arm raised. Children had to judge whether the monkeys had raised the same side arm (e.g., both their left arm) or an arm on different sides (one left the other right). This is easy if the creatures stand side by side but increasingly difficult if one of them is rotated sideways (see Fig. 1). An easy solution is to take the tilted picture and rotate it back to upright. However, children are not allowed to do this physically. In that case one can rotate it mentally by forming an image of what one sees and rotate the image and then compare the rotated image of the tilted figure with the perception of the other figure. This procedure is an instance of prospection: one pre-experiences an actual rotation of the tilted figure. And the reaction times tell us whether a child used this method because mental rotation results in a linear increase of reaction time with the degree of rotation (Shepard & Metzler, 1971). Estes (1998) reported that children start using this method between 4 and 6 years. Fig. 1. Examples of stimuli used for the mental rotation task in both experiments. 806 J. Perner et al. / Consciousness and Cognition 19 (2010) 802–815 2. Experiment 1 2.1. Method 2.1.1. Participants Fifty-nine children (25 girls and 34 boys) participated in the study. Children came from a nursery school in Upper Austria and two after-school care clubs, one in the city of Salzburg and one in Upper Austria. Most participants came from a middle-class background. Children’s ages ranged from 4, 11 (years, months) to 8, 7 (M = 6, 8, SD = 13.67 months). To analyze and display age trends, we divided the children into three approximately same sized age-groups: Twenty children from 4, 11 to 6, 0 (M = 5, 5, SD = 3.99 months), 17 children ranging in age from 6, 1 to 7, 4 (M = 6, 8, SD = 5.51 months), and 22 children ranging in age from 7, 5 to 8, 7 (M = 7, 11 SD = 4.4 months). 2.1.2. Design Each child was tested individually in a quiet room of the nursery school or after-school care club. Children were given two memory tasks (one with free and one with cued recall). In addition, children received a computerized mental rotation task based on Estes (1998) and an age-appropriate measure of verbal intelligence (verbal subtest of KISTE, Häuser, Kasielke, & Scheidereiter, 1994; or HAWIK-III, Tewes, Rossmann, & Schallberger, 1999). All tasks were administered in two sessions a few days apart. Each session started with presentation of items for the memory task and ended with recall of memory items. Between presentation and recall of memory items, half of the children were given the mental rotation task in the first session and the verbal intelligence measure in the second session; the other half received these tasks in the opposite order. Half of the children received the free recall task in the first session and the cued recall task in the second session; the other half started with the cued recall task. 2.1.3. Procedure and materials 2.1.3.1. Mental rotation task (Estes, 1998). In a warm-up phase, children were presented with an odd one out game. They were shown three bears in upright position (two of them both raising the same arm and one of them raising a different arm). Children were asked to select the bear that was different from the other two. Then, children were asked to play a computerized mental rotation task presented on a notebook computer using the software package Presentation (Neurobehavioral Systems Inc., http://www.neuro-bs.com). Children were told to press the button with the smiling face if two bears both raising the same arm appeared on the screen and to press the button with the sad face if two bears each raising a different arm appeared. After experimenter and child had jointly completed six training trials, children were asked to play on their own. The test phase comprised 56 trials. Each trial consisted of a pair of bears. The bear on the left was always upright, whereas the bear on the right appeared in seven different orientations. He was either upright (i.e., 0° rotation) or rotated clockwise in 30° increments up to 180°. In addition, each bear raised one of his arms. For half of the trials, both bears raised the same arm (i.e. both right or both left). For the other half of trials, the bears raised different arms (i.e. the left bear raised its right arm and the right bear raised its left arm, or the reverse). In total, there were 28 different stimulus pairs: two ‘‘same” and two ‘‘different” pairs for each of the seven different orientations. Each child was given these 28 different stimulus pairs twice in a fixed random order. After the first 28 trials, there was a short 2 min break. Children had to press as fast as possible the button marked with a smiling face if the two bears had raised their same side arm and the button marked with a sad face if the bears had raised different side arms. Children had to keep their index fingers positioned on the response buttons throughout the trial block. Children were asked if they were ready before the experimenter initiated the next trial. Correct responses were followed by a brief tune, errors by silence. No other feedback was given. After Trials 5, 28, and 56, children were asked how they could tell if the two bears had raised their same arms or not. Sample stimuli are shown in Fig. 1. 2.1.3.2. Memory tasks. For the memory tasks two sets of 20 coloured pictures (21  29.7 cm) of familiar objects or animals were created (similar to Perner & Ruffman, 1995, Exp. 1). Each set comprised five categories, with four items in each category (see Appendix A for list of items). Pictures were presented in a fixed random order. Each child received both sets. Half of the children were given Set 1 for the free recall task and Set 2 for the cued recall task. For the other half the two sets were exchanged. Children were told to look carefully at each picture they are going to see, because they will be asked about them later. Each picture was shown for 5 s and the child had to name it. Slight misidentifications (e.g., calling the farm a barn) were accepted. Completely wrong answers were corrected. After the intervening task (mental rotation or verbal intelligence measure), children were reminded: ‘‘I’ve shown you some pictures earlier on. Can you remember?” In the free recall condition, children were then simply asked ‘‘What was in these pictures?” In the cued recall condition, they were asked for each category, e.g.: ‘‘There were some animals, what were they?” 807 J. Perner et al. / Consciousness and Cognition 19 (2010) 802–815 The number of correctly recalled items and of false alarms (items that had not been on the learning list) were recorded. The difference between the number of correct recalls and the number of false alarms was used as a measure of recall accuracy. 2.2. Results and discussion First, children’s performance on the memory tasks and on the mental rotation task was analyzed separately, and then we looked at the relationship between these tasks. 2.2.1. Memory tasks Recall accuracy (hits minus false alarms) on the two memory tasks was analyzed by an analysis of variance with agegroup (younger, middle, older) as between participants factor and recall (free, cued) as a within participants factor. There was a significant main effect of recall, F(1, 56) = 49.61, p < .001, partial g2 = .47. No other effect was significant p > .50. Correlation between free and cued recall accuracy was r = .17, p > .19. 2.2.2. Mental rotation We first looked at children’s number of correct answers. There was a clear bimodal distribution. One group of 24 children had values from 24 to 38 normally distributed around the mean of 30.21 correct, which is just two items more than the guessing level of 28 (of 56 items): t(23) = 3.35, p < .003. Presumably these are the children who did not resort to rotation, without which they could only judge about two items consistently correctly. There were no children with values 39–41 and the remaining 35 children attained values from 42 to 55 with a mean of 49.89 correct. To see how children managed to give correct answers they were classified as rotators or non-rotators according to the two criteria used by Estes (1998). For each child a regression analysis was computed for the median reaction times for each angle as dependent and angle of rotation as independent variable. When the regression coefficient (slope) was significantly different from zero the child was classified as a ‘‘rotator according to RT”. Children were also asked at the end how they had approached the task. If they indicated that they had rotated the stimuli in their mind (e.g., ‘‘I rotated the bear in my head”) they were classified as ‘‘rotators according to explanation”. The other children gave no insightful answer (e.g., ‘‘I just know it”). Correlation of these measures with the number correct solutions showed r = .64 for explanation, r = .66 for reaction time, and r = .67 for rotation by reaction time and explanation. We use this binary classification as ‘‘rotators” (rotators by RT and by explanation) vs. ‘‘non-rotators” for all further computations (we checked that the interpretations stay the same if anyone of the other two is used). This way of classifying children into rotators and non-rotators had a marginally significant relation with age (r = .25, p = .052) but a clear relationship with verbal intelligence (r = .40, p = .002). 2.2.3. Interrelations between retrospection and prospection The argument about the relationship between type of recall and episodic remembering is that both kinds are (can be) helped by the availability of episodic traces. This can be equally strong, so that no interactive effect of the availability of episodic traces on free vs. cued recall can be expected. However, free recall is more dependent on episodic traces than cued recall (which profits to a large degree from the availability of the cues). Availability of episodic traces should, therefore, correlate more strongly with free than with cued recall, since more alternative factors influence cued than free recall. The theoretical claim is that the ability for episodic remembering requires the ability to project oneself into the past and that the same projective ability is required in the mental rotation task by projecting oneself as an observer of a future or hypothetical action (rotating one of the items to be compared). On those grounds we expect that mental rotation should show a significant correlation with free recall even when any correlation with cued recall has been partialled out. Table 1 shows the correlations between relevant variables in the upper right corner above the main diagonal. Partial correlations with age and verbal intelligence taken into account are shown for the remaining variables in italics below the main diagonal. The correlation between rotation and free recall remains at least marginally significant but the correlation between rotation and cued recall reaches only about half its size. More importantly, when cued recall is partialled out the partial Table 1 Correlations and partial correlations with age and verbal intelligence controlled. 1. Age 2. Verbal intelligence 3. Mental rotation 4. Free recall: accuracy 5. Cued recall: accuracy 1 2 3 4 5 – .17 – .25+ .40** – .26+ .14 .10 .29* .35** – .06 .18 .39** .29* .17 – Note: Numbers in italics – partial correlations after partialling out age and verbal intelligence. + p < .06. * p < .05. ** p < .01. 808 J. Perner et al. / Consciousness and Cognition 19 (2010) 802–815 correlation of rotation with free recall remains clearly significant (pr = .31, p = .016). This is as predicted when retrospection (as measured by free recall beyond cued recall) is developmentally linked with prospection (as measured by mental rotation). In contrast, when free recall is partialled out the partial correlation of rotation with cued recall is weaker but does stay marginally significant (pr = .25, p = .06), which attests to the fact, pointed out earlier, that both free and cued recall profit from the ability of episodic remembering to project oneself back into the past. This highlights the need for a more discerning measure of episodic remembering in the next experiment. 3. Experiment 2 Experiment 1 provided some evidence that the development of prospection, assessed by the ability to engage in mental rotation, is specifically linked to retrospection (free recall) when controlling for general ability to retrieve information about the past (cued recall). The differential performance on free over cued recall is a very weak measure of episodic remembering (retrospection) since both types of recall can benefit to equal amounts from the availability of episodic traces. The only difference is that the overall variance of free recall should be more strongly dominated by episodic abilities than the variance of cued recall. This weakness is also reflected in the controversy over Tulving’s (1985) claim that relatively more items recalled in free recall should be judged as ‘‘remembered” than in cued recall (Jones & Roediger, 1995; Roediger & McDermott, 1995; vs. Hamilton & Rajaram, 2003). In this second experiment we try to get stronger evidence for a specific developmental link by employing a contrast that depends more directly on the availability of episodic traces. A suitable contrast is between (directly) experienced events and events about which one was indirectly informed (e.g., through verbal or pictorial media). Only experienced events can be reexperienced. Though imagined experiences can be mistaken for re-experiences (false memories) this should be relatively rare if children are not instructed to vividly imagine the event about which they are indirectly informed. Under these premises we can predict that development of retrospective abilities (episodic remembering) should uniquely improve recall of experienced events but not at all (except for the occasional false memory) recall of indirectly conveyed events. Following the method used by Perner et al. (2007) we had children put cards with different pictures into a box. In the direct-experience condition the children saw the picture on each card as they put it into the box. In the indirect-information condition they put the cards into the box when being blindfold and afterwards they were shown on a monitor the pictures on those cards. We expect that children’s ability to engage in mental rotation should coincide with improved recall of directly experienced items but not of indirectly conveyed items. 3.1. Method 3.1.1. Participants Thirty-one children (16 girls and 15 boys) from two nursery schools in towns near the city of Salzburg participated in the study. Most participants came from a middle-class background. Children’s ages ranged from 5, 0 (years, months) to 6, 4 (M = 5, 10; SD = 4.72 months). For later analysis children were divided into a younger (n = 13, M = 5, 5; ranging from 5, 0 to 5, 10) and an older group (n = 18, M = 6, 1; ranging from 5, 11 to 6, 4). 3.1.2. Design Each child was tested individually in a quiet room of the nursery school. Children were given two memory tasks (one with direct experience and one with indirect information). In addition, children received the mental rotation task based on Estes (1998) described in Experiment 1. All tasks were administered in three sessions a few days apart. In the first session, children received the mental rotation task and a modality-specificity task in counterbalanced order.2 In the second and third session, one of the two memory tasks was given. In each session, presentation and recall of memory items were separated by a 10-min delay. Half of the children started with the direct-experience condition and were then given the indirect-information condition. For the other half, the order was reversed. 3.1.3. Procedure and materials The mental rotation task (Estes, 1998) was administered as described in Experiment 1. 3.1.3.1. Memory tasks. For the memory tasks, four sets of 12 coloured pictures (21  29.7 cm) of familiar objects, animals, or human beings were used. Set 1 and Set 2 were administered in the first session. Set 3 and Set 4 were given in the second session. In each session, half of the children were given a particular set as test items and the other set as distractor items. For the other half, the two sets were exchanged, so that each set was equally often used as test or distractor set in the direct experience and in the indirect-information memory task. The experimenter presented the 24 pictures (test and distractor items), one after the other, alternating between items from each set. As each picture was produced, the child was asked to name it. If necessary (which was rarely the case), the experimenter provided the correct label. 2 This task was included to make good use of the retention time. It was part of a project of a student who helped testing the children. This task will not be further analysed. J. Perner et al. / Consciousness and Cognition 19 (2010) 802–815 809 In the direct-experience memory task, children were then asked to place the 12 test items into a (26  40  4 cm) box. They were allowed to look at each picture for 2 s and were instructed to keep these pictures in mind. In the indirect-information memory task, children were also asked to place the 12 test items into a box but they were blindfolded so that they could not see the pictures. After having placed the 12 items into the box, they were shown ‘‘what these 12 pictures had depicted” by means of a computerized presentation. Each picture was shown for two seconds, and children were instructed to keep these pictures in mind. After a 10-min delay, children were asked in both memory tasks, ‘‘Do you remember the pictures you put into the box?” If children answered with an item that was not put inside the box it was scored as a false alarm. On both memory tasks, the number of correct recalls (hits) and the number of false alarms (if children mentioned items that were not put inside the box) were recorded. Mostly, false alarms comprised items from the distractor set used in the familiarization phase. The distractor set was introduced to get a more precise measure of remembering the pictures being put into the box as opposed to mere familiarity with the pictures. The contrast between recall of items placed inside the box with false alarms (mostly items that children were familiarized with but did not put inside) sharpens the detection of episodic memories in contrast to mere familiarity answers. For instance, a child who recalls two correct items and no distractors is comparable with one who recalls 12 targets and 10 distractors. Both children vastly differ in recall of familiar items (pictures) but are similar in terms of memory for items placed inside the box. This is critical, because our manipulation of direct experience and indirect knowledge pertains to the placing of cards into the box, not to familiarity with the pictures (they are directly experienced in both conditions). For this reason it is essential to rely not only on the quantity of pictures recalled (number of placed pictures recalled in relation to all pictures placed into the box) but also check for accuracy (number of placed pictures recalled in relation to all pictures recalled; as this terminology is used by Koriat and Goldsmith (1996, p. 177)). A measure that captures both these aspects common in signal detection theory is d0 which is based on the difference between hits and false alarms. This is typical for forced choice recognition tasks and rarely used with recall (Koriat & Goldsmith, 1996, p. 183). Following this approach, we use the difference between number correct items recalled minus number false alarms as our critical indicator of episodic recall and refer to it, for want of a shorter label and in line with signal detection theory, as: recall accuracy. 3.2. Results First, children’s performance on the memory tasks and on the mental rotation task was analyzed separately, and then we looked at the relationship between these tasks. 3.2.1. Memory tasks Recall accuracy (hits minus false alarms) was subjected to an analysis of variance with age-group as between participants factor and experience (direct, indirect) as a within participants factor. There was a significant interaction between age-group and experience, F(1, 29) = 4.31, p < .047, partial g2 = .13. Whereas the recall of directly experienced items increased with age from 2.69 to 3.56 items, recall of indirect items declined with age from 3.46 to 2.50 items. This unexpected decline will be discussed below when looking at the relationship with mental rotation (prospection). No other effect was significant p > .70. Correlation between direct and indirect recall accuracy was r = .42, p = .019. 3.2.2. Mental rotation The results mirror those of Experiment 1. The numbers of correct answers range from 28 to 54. There was again a strong correlation (r = .57, p = .001) between number correct solutions and rotation by explanation (r = .38, p = .033), rotation by reaction time (r = .46, p = .01), and rotation by reaction and explanation (r = .58, p = .001). As in Experiment 1 we use rotation by reaction time and explanation for our further analyses. This classification had a marginal correlation with age in Experiment 1 but this time it had no correlation with age at all (r = .00), which is probably due to the much narrower age range in this experiment. 3.2.3. The relation between retrospection and prospection The contrast between directly experienced versus indirectly conveyed events works differently than the contrast between free and cued recall in Experiment 1. While free and cued recall can profit from episodic remembering, only directly experienced events can do so, because only experienced events can be re-experienced. This means that the emerging ability for episodic remembering should enhance recall of directly experienced events but not recall of indirectly conveyed events. Moreover, re-experiencing past events (episodic remembering) and pre-experiencing imagined events for mental rotation require the same ability of projecting oneself as an observer. Hence, we expect that children who can use mental rotation will have better recall of directly experienced events than of indirectly conveyed events. Children, who fail to use rotation, will show no, or at least a much reduced, difference in recall. An analysis of variance of recall accuracy (number of hits minus false alarms) with rotation (rotation by RT and explanation: rotators vs. non-rotators) as a between subjects factor and experience (experienced vs. indirectly conveyed events) showed significant main effects of rotation (F(1, 29) = 5.16, p = .031, partial g2 = .15) and experience (F(1, 29) = 4.51, p = .042, partial g2 = .13), and a highly significant rotation  experience interaction (F(1, 29) = 7.83, p = .009, partial g2 = .21). This interaction was predicted. However, the form of the interaction as displayed in the left panel of Fig. 2 harbours 810 J. Perner et al. / Consciousness and Cognition 19 (2010) 802–815 Fig. 2. Results for rotators and non-rotators on three different memory measures in Experiment 2. a surprise. The ability to mentally rotate does not lead to enhanced memory for experienced events but has a detrimental effect on trying to remember indirectly conveyed events. This decline in accuracy for indirectly conveyed events persists for rotators-by-RT but not for rotators-by-explanation (p = .016), although the pattern of results is still very similar. This sharp decline (by 3.2 items) in accuracy (hits minus false alarms) is, as the centre panel in Fig. 2 shows, less pronounced (1.6 items) for hits (items correctly recalled) but still significant (interaction: (F(1, 29) = 4.81, p = .037, partial g2 = .14)). At the same time the number of false alarms increases more sharply for indirectly conveyed items than experienced items (F(1, 29) = 4.45, p = .045, partial g2 = .13) as the right panel in Fig. 2 shows. 3.3. Discussion The ability to engage in mental rotation had a strong differential effect on children’s ability to recall directly experienced events and to recall indirectly conveyed events. The finding, however, contained a great surprise: the differential effect did not consist in rotators’ better recall of experienced events but was due to their worse recall of indirectly conveyed events. This unexpected decline in recall of indirectly conveyed items does, nevertheless, confirm a trend observed in two experiments by Perner et al. (2007), where recall of indirectly conveyed items declined with increased theory of mind competence. What seemed initially a curious and haphazard finding now looks increasingly robust: As children improve their theory of mind and develop mental rotation skills their recall of indirectly conveyed information declines. It is difficult to find an obvious explanation for this decline. One could suspect that the reason lies in the unusual procedure of the indirect condition. But why should the more sophisticated rotators become confused by this procedure, when adults find it unusual but clear and perform as well as in the more normal experience condition (Stöttinger, 2006). The only plausible but highly speculative explanation has been briefly alluded to by Perner et al. (2007, p. 480). The younger children use their non-episodic recall (knowledge retrieval) for both conditions more or less successfully. Then when their theory of mind competence increases and become able to project themselves as past perceivers of experienced events they switch to using episodic recall. This switch may (Perner et al., 2007) or may not (current Experiment 2) help them immediately to recall more items than with non-episodic recall. In any case, this switch does not work for indirectly conveyed items and children’s performance declines. Presumably some time after discovering episodic recall children will realize that it does not work in every case and will readjust their strategy for indirect information. Eventually they may also discover the possibility of indirect episodic recall for indirect information. That is, when told about an event they should not try to re-experience (episodically remember) the event (e.g., putting the crocodile into the box) but to remember the information event (i.e., remember the crocodile appearing on the monitor and the instructions that pictures on the monitor were on the cards put inside the box). It is left to future studies to explore this development. 4. General discussion We have presented some data showing that prospection (as measured by children’s ability to engage in mental rotation) is related to retrospection (episodic remembering: as measured by variance shared with free recall when controlling for cued recall and by the difference between recall of experienced events (which can be re-experienced) versus recalling indirectly conveyed events (which cannot be re-experienced). Our expectation was that recall of indirect information would be unaffected while recall of experienced events, which can be re-experienced, would increase with children’s ability to mentally rotate. Contrary to expectations recall of experienced events was unaffected while children’s recall of indirect information dropped drastically with their ability to engage in mental rotation. Although, retrospectively (sic!), this trend was already apparent in investigations of children’s theory of mind competence (Perner et al., 2007), the sharpness of this trend with J. Perner et al. / Consciousness and Cognition 19 (2010) 802–815 811 mental rotation came as a surprise. The only plausible explanation we found was that acquisition of mental rotation marks a change in recall strategy. Instead of retrieving knowledge as before, children expect to find answers by trying to re-experience the relevant events. This is a successful replacement for the old strategy in the case of experienced events but fails miserably for indirect information about events. Despite these unexpected results we were able to demonstrate a developmental link between retro- and prospection that is Difficult to reduce to a more general connection, e.g., understanding time, which could account for existing data as reviewed in the Introduction. The specific connection we identified is children’s ability to project themselves mentally as experiencing events. This characterisation is commensurable with the neurocognitive suggestions of self-projection (Buckner & Carroll, 2007) and scene-construction (Hassabis & Maguire, 2007). In fact, on our view these two abilities are necessary complements: To project oneself as an observer of events at different times one has to be able to project one self into different times AND be able to reconstruct the scenes that one is to observe. Our results are less compatible with the ages observed in earlier developmental studies. Results from many of the studies reviewed earlier that looked at prospection suggest that the main development takes place between 3 and 5 years, though not from all studies. For instance, Atance and Meltzoff’s (2006) report that 5-year olds still could not imagine/anticipate that their current desire for water induced by overfeeding on salty pretzels will have reverted next day to their usual desire for pretzels. The fact that prospection as measured with mental rotation points to later development might mean that the studies that see the development completed by 5 years did not assess children’s ability to project themselves as observers (experiencers) of events at different times but assessed their knowledge about events in the past or future. Research on children’s retrospection (episodic remembering) also finds development in this age bracket but clearly also points to improvements beyond the age of 5 years. The original studies by Perner and Ruffman (1995) and also Naito (2003) included children up to 7 years and their more difficult theory of mind tasks tended to show the stronger correlations with measures of episodic remembering. These findings are, therefore, more in line with the developments documented in the present studies. We want to end with more general considerations about the methodological difficulties in assessing young children’s and non-verbal creatures’ explanations and subjective phenomenology. The autonoetic awareness of re-experiencing a formerly experienced event is something very subjective, in the sense that evidence for it comes exclusively from introspective verbal reports. With adult participants in memory experiments the relevant experience has been assessed by the ‘‘R–K” test. People have to judge their subjective experience of recall or recognition as one of merely knowing (K) what had happened (presentation of an item) or remembering (R) this event. This test is less than straight forward with adults, its value highly controversial (Donaldson, 1996; Dunn, 2004; Gardiner & Richardson-Klavehn, 2000; Stöttinger, Aigner, Hanstein, & Perner, 2009; Stöttinger, Kaiser, & Perner, 2009) and its use for young children beyond contemplation. The only solution seems to be an objectively measurable indicator of these subjective phenomena. This enterprise requires a fine instrument in order to distinguish between knowing what happened and re-experiencing (remembering) the event retrospectively, and between knowing what will happen and pre-living the likely event in prospection. The instrument has to be precise because both kinds of mental processes can, in principle, produce the required information about the world. They are ‘‘computationally equivalent”: both can give the correct answer to the questions posed, e.g.: What happened? Which item was presented? The difference must primarily lie in how the system arrives at the answer. Cognitive Psychology can, supposedly, provide the relevant instruments. So, no wonder that one of its most impressive findings, mental rotation, has such appeal in this context. The critical point is not that mental rotation can provide the correct same–different judgments—there may be infinitely many computationally equivalent ways. It is the pattern of reaction times that tells us that the answer is arrived at by mental rotation as we subjectively experience it. However, even the reaction time pattern is not a fail safe indicator of prospection. Several species of animals can give correct responses, e.g., pigeons, but do not show the same linear pattern of reaction times with angle of rotation (Hollard & Delius, 1982). Their visual system directly provides rotation invariance and they do not have to mentally simulate perceiving a hypothetical rotation. However, the reaction times of baboons (Vauclair, Fagot, & Hopkins, 1993), a sea lion (Stich, Dehnhardt, & Mauck, 2003) and a lion-tailed macacque (Burmann, Dehnhardt, & Mauck, 2005) also show some signs of the linear relationship with angle of rotation, though their times deviate from the human pattern in ways adaptive for their respective perceptual environment. For argument’s sake let us assume their reaction times showed exactly the same pattern. We could still not conclude that these animals project themselves as perceivers of hypothetical rotation events (mental time travel). Their visual system might provide rotation invariance but processing time happens to depend on angle of rotation. Thus, the ultimate justification for interpreting the linear reaction time pattern in humans as a sign of mental time travel hinges on the fact that this linearity goes hand in hand with the subjective experience of mental rotation in adults. This relationship between observable answer and subjective experience in adults is not so tight for other measures. For instance, when asked what to pack for a trip to a place shown in a hibernal scene (Atance & Meltzoff, 2005) my answer ‘‘winter coat” could be given by projecting myself travelling to that place and noticing the need for a winter coat. However, it could equally well be based on just knowing that in places like this you need a winter coat. Mental rotation (provided the linear reaction time pattern does indicate it) is a much less vulnerable indicator of prospection. For investigating episodic remembering (retrospection) in animals and young children several proposals have been made. The technically most sophisticated has been developed by Yonelinas (1999) for distinguishing between recall based on famil- 812 J. Perner et al. / Consciousness and Cognition 19 (2010) 802–815 iarity versus recollection, which can be profitably used even with rats (Sauvage, Fortin, Owens, Yonelinas, & Eichenbaum, 2008). Unfortunately the distinction between familiarity of items and recollection of associative information is not quite the same as between knowing and remembering. Associative information can be known or remembered. We need methods that drive a wedge between knowing the past and remembering the past. Perner and Ruffman (1995) followed Tulving’s criterion that free recall produces relatively more experiences of remembering than cued recall. Unfortunately the link to the one direct method of assessment of the remember–know paradigm remains tenuous (Hamilton & Rajaram, 2003; Jones & Roediger, 1995; Roediger & McDermott, 1995). Perner et al. (2007); our Experiment 2) relied on an objective prerequisite for episodic remembering, i.e., the direct experience constraint (Stöttinger, Aigner, et al. 2009; Stöttinger, Kaiser, et al., 2009): only an experienced event can be re-experienced. Neither of these methods is apt to produce positive demonstrations of a child enjoying a re-experience of a past episode. These methods only allow an objective test of developmental hypotheses: if some other ability (theory of mind, mental rotation) can be claimed to index the onset of re-experiencing then the method allows for a testable prediction. The development of the index ability should have a differential effect on recall conditions (free vs. cued; recall of directly experienced vs. indirectly conveyed events). Although the method falls short of giving a direct existence proof of episodic remembering it does provide an empirical test of hypotheses about episodic remembering that could also be used on animals. As it has been applied here it requires some linguistic proficiency for the indirect-information condition. It would not work without being able to tell children that the items shown on the monitor are the same as on the cards they could not see. Indirect knowledge through verbal communication is difficult to use with animals but there are dogs (Miklósi, Polgárdi, Topál, & Csányi, 2000) and goats (Kaminski, Riedel, Call, & Tomasello, 2005), who understand pointing gestures, and there are chimpanzees who can infer from seeing that one of two containers is empty that the object must be in the other container (Call, 2004). So, if we can find a theory of which individuals (ontogenetic development, enculturation, evolution) are able of episodic remembering, then we can provide testable predictions of how these individuals differ in recalling directly observed locations vs. gestured or inferred locations. Clayton and Russell (2009) made a new suggestion of how to differentiate knowledge of the past from remembering the past empirically without use of language. Their minimal requirement for remembering is evidence during re-experience of the subjective perspective of the original experience. This is clearly a highly relevant aspect—but an important limitation has to be pointed out. Like the direct experience constraint, this criterion cuts only one way: features of subjective perspective at recall do not licence claims about re-experiencing. For, perspective features could result from how knowledge of the past is encoded. To use an example from Davies, Russell, and Russell (2009), children observed how to fix a toy boat by performing on a left or right lever either a pumping or levering motion. By 3 years most children imitated after some delay a series of two consecutive actions in the correct spatio-temporal order. The authors suggested that this shows earlier episodic remembering than commonly expected, because children reproduced the temporal perspective of their original experience. The problem with this conclusion is that the study confounds the subjective order of experiences with the objective knowledge of the order in which the actions were performed. What one might be able to do is the following. Let us assume children below 3 years can remember sequences of events to some degree. One would then have to show that at the purported age of 3 years children show a marked change in the order in which they recall witnessed events but do not show any change in the order in which they report the events having happened. This method put together with the direct experience vs. indirect information contrast (Perner et al., 2007) may provide a more powerful combination. Reports of events need not follow the order of the events. So, particularly strong evidence would be forthcoming if at the critical age children began recalling events in the order they had been reported rather than the order in which they happened, e.g., when told ‘‘he ate after his nap” the younger ones recall, ‘‘he slept”–‘‘he ate,” while the older ones recall, ‘‘he ate”–‘‘he slept”. This kind of reversal would provide an interesting analogy to our finding that rotators started to have problems with indirectly conveyed events because—our suggestion—they switched from knowledge retrieval to episodic remembering. There has been a certain resentment among animal and infancy researchers against insisting on Tulving’s (1985) and Wheeler et al. (1997)) criteria for episodic memory because they seem to be impossible to meet for non- or low verbal creatures. With our research and this final discussion we want to make the case that to capture the phenomenon of episodic remembering we need to insist on these criteria but that there is a way of testing theories of when episodic remembering develops or evolves. Acknowledgments The authors acknowledge the financial support from the Austrian Science Fund (FWF Grant P16215-G04 ‘‘Episodic Memory and Conscious Experience”). The data of Experiment 1 were collected by Michael Rohwer for his Diploma thesis (2006) ‘‘Zusammenhang zwischen episodischem Gedächtnis und mentaler Rotation.” We thank the Head and staff of the Kindergarten Seekirchen, Kindergarten Thalgau, Kindergarten Thalheim, Hort Taxham, and Hort Thalheim for their willing participation and Elisabeth Stöttinger for her guidance through the adult memory literature. J. Perner et al. / Consciousness and Cognition 19 (2010) 802–815 813 Appendix A Memory items used in Experiment 1, ordered by category. Original German words (English translation) Set 1 Set 2 Tiere (animals) Pflanzen (plants) – Schaf (sheep) – Pferd (horse) – Schwein (pig) – Kuh (cow) Gemüse (vegetables) – Karotte (carrot) – Erbse (pea) – Zwiebel (onion) – Paprika (red pepper) Möbel (furniture) – Bett (bed) – Tisch (table) – Stuhl (chair) – Schrank (wardrobe) Gebäude (buildings) – Kirche (church) – Burg (castle) – Haus (house) – Scheune (barn) Werkzeuge (tools) – Säge (saw) – Schaufel/Spaten (spade) – Zange (pliers) – Hammer (hammer) – Baum (tree) – Sonnenblume (sun flower) – Rose (rose) – Getreide/Weizen (wheat) Obst (fruit) – Birne (pear) – Weintrauben (grapes) – Kirschen (cherries) – Apfel (apple) Kleidung (clothes) – Pullover (sweater) – Schuhe (shoes) – Socken (socks) – Hose/Jeans (trousers/Jeans) Fahrzeuge (vehicles) – Bus (bus) – Auto (car) – Lastwagen (lorry/ truck) – Motorrad (motorbike) Musikinstrumente (musical instruments) – Trompete (trumpet) – Klavier/Piano (piano) – Geige/Violine (violin) – Trommel (drum) References Atance, C. M., & Jackson, L. K. (2009). The development and coherence of future-oriented behaviors during the preschool years. Journal of Experimental Child Psychology, 102(4), 379–391. Atance, C. M., & Meltzoff, A. N. (2005). My future self: Young children’s ability to anticipate and explain future states. Cognitive Development, 20(3), 341–361. Atance, C. M., & Meltzoff, A. N. (2006). Preschoolers’ current desires warp their choices for the future. Psychological Science, 17(7), 583–587. Atance, C. M., & O’Neill, D. K. (2005). The emergence of episodic future thinking in humans. Learning and Motivation, 36(2), 126–144. Baron-Cohen, S., Leslie, A. M., & Frith, U. (1985). Does the autistic child have a ‘‘theory of mind”? Cognition, 21, 37–46. Bischof-Köhler, D. (2000). Kinder auf Zeitreise: Theory of mind, Zeitverständnis und Handlungsorganisation. Bern ua: Hans Huber. Boucher, J., & Lewis, V. (1989). Memory impairments and communication in relatively able autistic children. Journal of Child Psychology and Psychiatry and Allied Disciplines, 30(1), 99–122. Boucher, J., & Warrington, E. (1976). Memory deficits in early infantile autism: Some similarities to the amnesic syndrome. British Journal of Psychology, 67, 73–87. Bowler, D. M., Gardiner, J. M., & Gaigg, S. B. (2007). Factors affecting conscious awareness in the recollective experience of adults with Asperger’s syndrome. Consciousness and Cognition, 16(1), 124–143. Bowler, D. M., Gardiner, J. M., & Grice, S. (2000). Episodic memory and remembering in adults with Asperger syndrome. Journal of Autism and Developmental Disorders, 30, 305–316. Buckner, R. L., & Carroll, D. C. (2007). Self-projection and the brain. Trends in Cognitive Sciences, 11(2), 49–57. Burmann, B., Dehnhardt, G., & Mauck, B. (2005). Visual information processing in the lion-tailed macaque (Macaca silenus): Mental rotation or rotational invariance? Brain, Behavior and Evolution, 65(3), 168–176. Busby, J., & Suddendorf, T. (2005). Recalling yesterday and predicting tomorrow. Cognitive Development, 20(3), 362–372. Call, J. (2004). Inferences about the location of food in the great apes (Pan paniscus, Pan troglodytes, Gorilla gorilla, and Pongo pygmaeus). Journal of Comparative Psychology Copyright, 118, 232–241. Carlson, S. M., Moses, L. J., & Claxton, L. J. (2004). Individual differences in executive functioning and theory of mind: An investigation of inhibitory control and planning ability. Journal of Experimental Child Psychology, 87, 299–319. Churchland, P. M. (1984). Matter and consciousness: A contemporary introduction to the philosophy of mind. Cambridge, MA: MIT Press. A Bradford book. Clayton, N. S., & Russell, J. (2009). Looking for episodic memory in animals and young children: Prospects for a new minimalism. Neuropsychologia, 47(11), 2330–2340. Davies, J., Russell, C., & Russell, J. (2009). Making deferred imitation a measure of episodic memory. Unpublished manuscript, Department of Experimental Psychology, Cambridge University. Donaldson, W. (1996). The role of decision processes in remembering and knowing. Memory and Cognition, 24, 523–533. Dunn, J. C. (2004). Remember–know: A matter of confidence. Psychological Review, 111(2), 524–542. Ebbinghaus, H. (1885). Über das Gedächtnis. Leipzig: Duncker und Humblot. Estes, D. (1998). Young children’s awareness of their mental activity: The case of mental rotation. Child Development, 69(5), 1345–1360. 814 J. Perner et al. / Consciousness and Cognition 19 (2010) 802–815 Ford, R., Shum, D., & Driscoll, T. (2009). Event-based prospective memory in young children: An individual difference analysis. Paper presented as part of the symposium on prospective memory at the BPS Cognitive Section 26th annual meeting, University of Hertfordshire, 1–3 September 2009. Gardiner, J. M., & Richardson-Klavehn, A. (2000). Remembering and knowing. In E. Tulving & F. I. M. Craik (Eds.), The Oxford handbook of memory (pp. 229–244). Oxford, New York: Oxford University Press. Gilbert, D. T., & Wilson, T. D. (2007). Prospection: Experiencing the future. Science (New York, NY), 317(5843), 1351–1354. Goldman, A. I. (2006). Simulating minds: The philosophy, psychology, and neuroscience of mindreading. Oxford: Oxford University Press. Gopnik, A., & Astington, J. W. (1988). Children’s understanding of representational change and its relation to the understanding of false belief and the appearance-reality distinction. Child Development, 59, 26–37. Gordon, R. M. (1986). Folk psychology as simulation. Mind and Language, 1, 158–171. Hamilton, M., & Rajaram, S. (2003). States of awareness across multiple memory tasks: Obtaining a ‘‘pure” measure of conscious recollection. Acta Psychologica, 112, 43–69. Happé, F. (1994). An advanced test of theory of mind: Understanding of story characters’ thoughts and feelings by able autistic, mentally handicapped, and normal children and adults. Journal of Autism and Developmental Disorders, 24, 129–154. Hassabis, D., & Maguire, E. A. (2007). Deconstructing episodic memory with construction. Trends in Cognitive Sciences, 11(7), 299–306. Häuser, D., Kasielke, E., & Scheidereiter, U. (1994). Kindersprachtest für das Vorschulalter (KISTE). Bern, Göttingen: Hogrefe. Heal, J. (1986). Replication and functionalism. In J. Butterfield (Ed.), Language, mind, and logic (pp. 135–150). Cambridge: Cambridge University Press. Hill, E. L. (2004). Executive dysfunction in autism. Trends in Cognitive Sciences, 8, 26–32. Hollard, V., & Delius, J. (1982). Rotational invariance in visual pattern recognition by pigeons and humans. Science, 218(4574), 804–806. Jones, T. C., & Roediger, H. L. I. (1995). The experiential basis of serial position effects. European Journal of Cognitive Psychology, 7, 65–80. Kaminski, J., Riedel, J., Call, J., & Tomasello, M. (2005). Domestic goats, Capra hircus, follow gaze direction and use social cues in an object choice task. Animal Behaviour, 69, 11–18. Klein, S. B., Loftus, J., & Kihlstrom, J. F. (2002). Memory and temporal experience. The effects of episodic memory loss on an amnesic patient’s ability to remember the past and imagine the future. Social Cognition, 20, 353–379. Koriat, A., & Goldsmith, M. (1996). Memory metaphors and the real-life/laboratory controversy: Correspondence versus storehouse conceptions of memory. Behavioral and Brain Sciences, 19, 167–228. Kvavilashvili, L., Messer, D. J., & Ebdon, P. (2001). Prospective memory in children: The effects of age and task interruption. Developmental Psychology, 37, 418–430. Levine, B., Black, S. E., Cabeza, R., Sinden, M., Mcintosh, A. R., Toth, J. P., et al (1998). Episodic memory and the self in a case of isolated retrograde amnesia. Brain, 121(10), 1951–1973. Martin, M. G. F. (2001). Out of the past: Episodic recall as retained acquaintance. In C. Hoerl & T. McCormack (Eds.), Time and memory: Issues in philosophy and psychology (pp. 257–284). Oxford: Clarendon Press. Miklósi, A., Polgárdi, R., Topál, J., & Csányi, V. (2000). Intentional behaviour in dog-human communication: An experimental analysis of ‘‘showing” behaviour in the dog. Animal Cognition, 3, 159–166. Miscione, J. L., Marvin, R. S., O’Brien, R. G., & Greenberg, M. T. (1978). A developmental study of preschool children’s understanding of the words”know” and ‘‘guess”. Child Development, 49, 1107–1113. Moore, C., Barresi, J., & Thompson, C. (1998). The cognitive basis of future-oriented prosocial behavior. Social Development, 7, 198–218. Naito, M. (2003). The relationship between theory of mind and episodic memory: Evidence for the development of autonoetic consciousness. Journal of Experimental Child Psychology, 85, 312–336. O’Neill, D. K., Astington, J. W., & Flavell, J. H. (1992). Young children’s understanding of the role that sensory experiences play in knowledge acquisition. Child Development, 63, 474–490. Owens, D. (1996). A Lockean theory of memory experience. Philosophy and Phenomenological Research, LVI, 319–332. Ozonoff, S., Pennington, B. F., & Rogers, S. J. (1991). Executive function deficits in high-functioning autistic children: Relationship to theory of mind. Journal of Child Psychology and Psychiatry, 32, 1081–1105. Perner, J. (1990). Experiential awareness and children’s episodic memory. In W. Schneider & F. E. Weinert (Eds.), Interactions among aptitudes, strategies, and knowledge in cognitive performance (pp. 3–11). New York, Berlin, Heidelberg, London, Paris, Tokyo, Hong Kong: Springer-Verlag. Perner, J., Frith, U., Leslie, A. M., & Leekam, S. R. (1989). Exploration of the autistic child’s theory of mind: Knowledge, belief and communication. Child Development, 60, 689–700. Perner, J., Kloo, D., & Gornik, E. (2007). Episodic memory development: Theory of mind is part of re-experiencing experienced events. Infant and Child Development, 16, 471–490. Perner, J., & Ruffman, T. (1995). Episodic memory an autonoetic consciousness: Developmental evidence and a theory of childhood amnesia. Journal of Experimental Child Psychology, 59(3), 516–548 (special issue: early memory). Prencipe, A., & Zelazo, P. D. (2005). Development of affective decision making for self and other: Evidence for the integration of first- and third-person perspectives. Psychological Science, 16(7), 501–505. Roberts, W., Feeney, M., MacPherson, K., Petter, M., McMillan, N., & Musolino, E. (2008). Episodic-like memory in rats: Is it based on when or how long ago? Science, 320(5872), 113–115. Roberts, W. A. (2002). Are animals stuck in time? Psychological Bulletin, 128(3), 473–489. Roediger, H. L. I., & McDermott, K. B. (1995). Creating false memories: Remembering words not presented in lists. Journal of Experimental Psychology: Learning, Memory and Cognition, 21, 803–814. Rosenbaum, R. S., Stuss, D. T., Levine, B., & Tulving, E. (2007). Theory of mind is independent of episodic memory. Science, 318, 1257. Russell, J., Alexis, D., & Clayton, N. (2010). Episodic future thinking in 3- to 5-year-old children: The ability to think of what will be needed from a different point of view. Cognition, 114(1), 56–71. Sauvage, M. M., Fortin, N. J., Owens, C. B., Yonelinas, A. P., & Eichenbaum, H. (2008). Recognition memory: Opposite effects of hippocampal damage on recollection and familiarity. Nature Neuroscience, 11(1), 16–18. Shepard, R. N., & Metzler, J. (1971). Mental rotation of three-dimensional objects. Science, 171, 701–703. Spreng, R. N., Mar, R. A., & Kim, A. S. (2008). The common neural basis of autobiographical memory, prospection, navigation, theory of mind, and the default mode: A quantitative meta-analysis. Journal of Cognitive Neuroscience, 21(3), 489–510. Sprung, M. (2008). Unwanted intrusive thoughts and cognitive functioning in kindergarten and young elementary school-age children following Hurricane Katrina. Journal of Clinical Child and Adolescent Psychology, 37(3), 575–587. Sprung, M., & Harris, P. (2009). Intrusive thoughts and young children’s knowledge about thinking following a natural disaster. Journal of Child Psychology and Psychiatry. Stich, K. P., Dehnhardt, G., & Mauck, B. (2003). Mental rotation of perspective stimuli in a California sea lion (Zalophus californianus). Brain, Behavior and Evolution, 61(2), 102–112. Stöttinger, E. (2006). Episodisches Gedächtnis – Erinnern als wirkliches Wiedererleben. Unpublished thesis, University of Salzburg. Stöttinger, E., Aigner, S., Hanstein, K., & Perner, J. (2009). Grasping the diagonal: Controlling attention to illusory stimuli for action and perception. Consciousness and Cognition, 18(1), 223–228. Stöttinger, E., Kaiser, W., & Perner, J. (2009). ‘‘Remember” judgments and the constraint of direct experience. Psychological Research – Psychologische Forschung, 73, 623–632. Stuss, D. T. (1991). Disturbance of self-awareness after frontal system damage. In G. P. Prigatano & D. L. Schacter (Eds.), Awareness of deficit after brain injury (pp. 63–83). Oxford: Oxford University Press. J. Perner et al. / Consciousness and Cognition 19 (2010) 802–815 815 Suddendorf, T., & Busby, J. (2003). Like it or not? The mental time travel debate: Reply to Clayton et al.. Trends in Cognitive Sciences, 7, 437–438. Suddendorf, T., & Corballis, M. C. (1997). Mental time travel and the evolution of the human mind. Genetic, Social, and General Psychology Monographs, 123, 133–167. Suddendorf, T., & Corballis, M. C. (2007). The evolution of foresight: What is mental time travel, and is it unique to humans? The Behavioral and Brain Sciences, 30(3), 299–313 (discussion 313-51). Suddendorf, T., Nielsen, M. (2009). The development of foresight: Children’s capacity to remember a novel problem and to secure its future solution. Unpublished manuscript, University of Queensland. Tager-Flusberg, H. (1991). Semantic processing in the free recall of autistic children: Further evidence for a cognitive deficit. British Journal of Developmental Psychology, 9, 417–430. Taylor, M., Esbensen, B., & Bennett, R. T. (1994). Children’s understanding of knowledge acquisition: The tendency for children to report they have always known what they have just learned. Child Development, 65, 1581–1604. Tewes, U., Rossmann, P., & Schallberger, U. (1999). Hamburg-Wechsler-Intelligenztest für Kinder III (HAWIK-III). Handbuch und Testanweisung (3. überarbeitete und ergänzte Auflage). Bern: Verlag Hans Huber. Tulving, E. (1985). Memory and consciousness. Canadian Psychology, 26, 1–12. Vauclair, J., Fagot, J., & Hopkins, W. D. (1993). Rotation of mental images in baboons when the visual input is directed to the left cerebral hemisphere. Psychological Science, 4(2), 99–103. Wheeler, M. A., Stuss, D. T., & Tulving, E. (1997). Toward a theory of episodic memory: The frontal lobes and autonoetic consciousness. Psychological Bulletin, 121(3), 331–354. Wimmer, H., Hogrefe, J., & Perner, J. (1988a). Children’s understanding of informational access as source of knowledge. Child Development, 59, 386–396. Wimmer, H., Hogrefe, J., & Sodian, B. (1988b). A second stage in children’s conception of mental life: Understanding sources of information. In J. W. Astington, P. L. Harris, & D. R. Olson (Eds.), Developing theories of mind (pp. 173–192). New York: Cambridge University Press. Yonelinas, A. P. (1999). The contribution of recollection and familiarity to recognition and source-memory judgments: A formal dual-process model and an analysis of receiver operating characteristics. Journal of Experimental Psychology: Learning, Memory and Cognition, 25(6), 1415–1434.

Consciousness and Cognition Consciousness and Cognition 14 (2005) 257–277 www.elsevier.com/locate/concog Unconscious word-stem completion priming in a mirror-masking paradigmq Walter J. Perrig*, Doris Eckstein University of Bern, Switzerland Received 13 February 2004 Available online 8 October 2004 Abstract The aim of this study was to investigate unconscious priming by the use of a spatial mirror-masking paradigm. Words and nonwords with no under-length letters are mirrored at their horizontal axis. The results are figures of geometric-like forms that contain letters in their upper part. In the three experiments reported in this study, a priming procedure used such mirrored words and nonwords as primes. Participants were ignorant of the nature of the construction of the stimuli. Perceptual reports of the participants revealed that they did not realize that words were hidden in the primes. Nevertheless, they showed priming in all three experiments. Priming effects were replicated with prime–target SOAs of between 1 and 3 s. Functional dissociations were found between ignorant and informed participants. Informed groups showed perceptual and semantic priming, while ignorant groups showed only perceptual priming.
2004 Elsevier Inc. All rights reserved. Keywords: Unconscious priming; Word-stem completion; Mirror masking q This research was supported by the Swiss National Science Foundation (Grant No. 1114-67180.01 to Walter Perrig). We thank Josephine Cock and William Banks for valuable comments on an earlier version of this paper, and Daniel Bhend for assistance in experimental design and data collection for Experiment 1. * Corresponding author. E-mail address: walter.perrig@psy.unibe.ch (W.J. Perrig). 1053-8100/$ - see front matter
2004 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2004.07.008 258 W.J. Perrig, D. Eckstein / Consciousness and Cognition 14 (2005) 257–277 1. Introduction Research on unconscious cognition is a fascinating topic but also a thorny one. Why does experimental psychology in this domain move so slowly and clumsily, when at the same time theories and speculations in the field of psychoanalysis about the unconscious roots of motives, feelings, and behavior have exerted such an influence on society for the last 100 years? In therapy, healing, and academic thinking, reasoning based on so-called unconscious psycho-dynamics has easily survived the centennial, while experimental findings and interpretations in unconscious cognition have struggled and barely managed to get out of the laboratory, let alone reach readers of specialized journals. It is certainly not the case that such findings would not be interesting or not important enough. For example, experimental proof that meaning can be unconsciously perceived and that it can influence feelings or decisions (Kunst-Wilson & Zajonc, 1980; Marcel, 1983) would appear to be of primal relevance to the better understanding of human cognition, from both a theoretical and a practical perspective. In contrasting the struggle of getting unconscious perception recognized as a phenomenon with the fairly easy and widespread acceptance of the Freudian unconscious one has to consider the fact that acceptance and rejection in the two cases are by different groups of people. Those who are predisposed to accept Freudian assumptions generally ignore findings of experimental research; those who are skeptical about the experimental findings, and maybe experimental scientists in general reject the Freudian assumptions. Understandably, scientific skepticism is often associated with experimental proofs in unconscious perception. For many scientists there is simply not enough evidence for the theoretical elaboration of unconscious effects to proceed any further at the present time. As explained below, there are two main reasons. The first major problem for experiments devoted to investigating unconscious perception is related to the difficulty of measuring the exact subjective phenomenal status of participants during perception. There has always been, and will always remain, the critical question as to whether some participants have at least a modicum of conscious perception of parts of the hidden stimulus information (e.g., Shanks & St. John, 1994). The second and, perhaps more important, problem concerns the difficulty of replicating findings that are considered to be the results of unconscious mechanisms. Reviews on unconscious information processing (e.g., Kihlstrom, 1987; Perrig, Wippich, & Perrig-Chiello, 1993) usually present various experimental approaches that—under the best of possible controls—seem to demonstrate unconscious effects influencing behavior. However, until recently there were no research programs with well-established paradigms that could easily facilitate the replication of unconscious priming effects and thereby enable the systematic study of unconscious perception. However, in the related area of implicit cognition such research programs do exist, as for instance in implicit learning and implicit memory (cf. Curran & Schacter, 2001; Perrig, 2001), automatic processes (e.g., Hasher & Zacks, 1979; Jacoby, 1991; Posner & Snyder, 1975; Schneider & Shiffrin, 1977), and procedural knowledge (Cohen, 1984). In those areas, the notion of unconscious processing is less radically defended than for instance in unconscious perception (Reber & Perrig, 2001). Fore-grounded is the demonstration of differences in behavioral effects between declarative knowledge and experience that is not the object of conscious thought. In implicit memory research, for instance, there are reliable functional dissociations between direct and indirect memory tests (Roediger & McDermott, 1993). Implicit cognition first of all is established on the operational level. As such, reliable differences produced W.J. Perrig, D. Eckstein / Consciousness and Cognition 14 (2005) 257–277 259 by direct and indirect measures of memory in the behavioral data provide a stable basis for argument and theoretical challenge. The question of the exact phenomenological status of participants carrying out indirect tests (Perruchet, Gallego, & Pacteau, 1992), together with that of the reliability of implicit measures (Buchner & Wippich, 2000; Meier & Perrig, 2000) remain under debate in this research domain as well. Greenwald and his co-workers (Greenwald, Klinger, & Schuh, 1995; Greenwald, Draine, & Abrams, 1996) developed a response window paradigm that allowed the reliable measurement of unconscious priming, something that might be considered a breakthrough in the experimental investigation of priming. In this paradigm, the prime is preceded and followed by a mask after which the participant must make a rapid response to a target (in less than 500 ms). Results show strong prime–target congruency effects. For example, if in an evaluative task the emotional valence of target words has to be judged, then the error rate increases when the prime and the target valences differ (positive versus negative). The failure of participants to respond correctly to the prime words is taken as evidence of the unconscious nature of their performance. To date, the available data suggest that learned, perceptual stimulus–response associations form the basis of prime–target congruency effects in the response window paradigm. This conclusion is based on the fact that congruency effects can only be demonstrated under the following constraints. First, the critical dimension in the experimental material (e.g., word versus nonword; positive versus negative word) must also be the criteria on which the participants make their responses to the targets throughout the experiment (Klinger, Burton, & Pitts, 2000). Second, the primes, or perceptual elements of the primes, need to be shown to participants beforehand, i.e., as visible targets (Abrams & Greenwald, 2000; Damian, 2001). This state of the art does not, of course, impair the valuable contribution of the response window technique. On the contrary, its empirical and theoretical output is challenging and motivates the development of alternative paradigms that might also allow the investigation of unconscious information processing. In this study, we report the use of a spatial mirror-masking paradigm that prevents the conscious reading of words. In the mirror-masking paradigm, words or letter strings with no under-length letters are mirrored at their horizontal axis. The result is a figure, comprising geometric-like forms, that unbeknown to participants may contain a word in its upper part. An example is presented in Fig. 1. The mirrored word adds spatial information that hides the word from the uninformed participants. In a priming procedure, these mirrored words or letter strings are used as primes, and participants are not informed about the nature of the stimuli, thereby producing a kind of inattentional or cognitive blindness for the prime words themselves. It is well known that people often fail to see information that they do not expect to be there. Simons and Chabris (1999) published striking demonstrations of sustained inattentional blindness even for dynamic events. Our mirror-masking paradigm can be considered as a method for studying implicit memory in participants in a mental state of inattentional or cognitive blindness. Fig. 1. Example of the mirror masked word awareness. 260 W.J. Perrig, D. Eckstein / Consciousness and Cognition 14 (2005) 257–277 Three experiments are reported here that all demonstrate priming effects in a word-stem completion task. In contrast to the work with the response window technique, stimuli are never repeated. Thus, priming represents one-trial memory effects. Our results so far suggest that the priming effects are of perceptual rather than semantic origin. 2. Experiment 1 The first experiment tested for priming effects of mirror-masked primes using a word-stem completion (WSC) test. In a WSC task, participants see the first letters of a word, which they must complete as a whole word. It has been shown that previously shown prime words influence the way in which targets are completed, even if primes are pattern masked (despite some of them being visible, e.g., Forster, Booker, Schacter, & Davis, 1990; Meier, Morger, & Graf, 2003; Merikle, Joordens, & Stolz, 1995). Primes in our first experiment were mirrored words and mirrored nonwords (referred to henceforth as mirror primes) presented for 750 ms and then replaced by a three-letter target word stem that had to be completed with two letters from the first suitable noun to come to mind. Three types of prime–target relations were used in the experiment: (a) mirrored prime words followed by their word stems (same prime condition), (b) mirrored prime words followed by word stems from other words (different prime condition), and (c) mirrored prime nonwords followed by word stems as controls. We expected a higher completion frequency and faster response times for completion of word stems in the same prime condition compared to the different prime condition and the nonword prime condition. Because the subjects were not informed about the hidden word, priming effects can be considered to be the result of unconscious information processing. Thus, if participants do not—and indeed cannot—read the words embedded in the visual (and visible) stimuli, then the mirror-masking paradigm should induce a situation of inattentional blindness. In order to succeed at our goal, of course, it is essential that participants remain unaware of the full nature of the primes. To shift the focus of attention away from the primes, participants were informed that ‘‘position marks’’ (these were the mirror primes themselves) would indicate where, out of four possible locations on the screen, the word stems would subsequently appear. In order to know whether participants actually identified the mirrored words consciously, they were questioned about the ‘‘position marks’’ after the experiment. 2.1. Method 2.1.1. Participants and design The sample consisted of 29 undergraduate students participating for credit, 17 women and 12 men. Three levels of prime–target condition were varied within subjects: (a) the same prime condition, where the target word stem was the same as the prime word stem, (b) the different prime condition, where the target word stem and the preceding prime word stem differed, and (c) the nonword condition, where the prime was a nonword. A balanced design was obtained by using equal numbers of nonword and word prime trials. W.J. Perrig, D. Eckstein / Consciousness and Cognition 14 (2005) 257–277 261 2.1.2. Materials A set of 48 standardized five-letter target words and their respective three-letter target word stems was selected from German word-stem completion norms (Meier & Eckstein, 1998). Each word stem could be completed to form at least three appropriate nouns. For the purpose of stimulus rotation across conditions, these words were divided into eight balanced groups of six words. Within each group, the probability of completing a stem to form the target word averaged p = .22. From a set of nonstandardized words, 12 additional five-letter prime words, whose word stems differed from those of the targets, were also selected. Twenty-four nonwords were constructed. The materials are shown in Appendix A. 2.1.3. Procedure Participants performed the test individually on an IBM compatible personal computer. The program was written in MEL 2.0 (Schneider, 1988). The first three trials were practice trials, in which stems with two possible completions were shown after nonword primes. The following test consisted of 12 same prime trials, 12 different prime trials, and 24 nonword prime trials. Order of trials was determined randomly before testing. The eight target stimulus groups were rotated between conditions and participants. Owing to the fact that mirror primes were used as fixation marks, every trial was randomly shown in one of four positions on the screen, the first two being 5 cm to the left and, respectively, to the right of the screen center, the second two being 3 cm to the top and, respectively, to the bottom of the screen center. Each trial started with a black screen that lasted for 1 s. Then, the mirror prime was shown in one of the four possible positions. After 750 ms, the stem replaced the prime at approximately the same screen location, and participants had to type their answer on the computer keyboard. Editing of the response was permissible, by using the delete and backspace keys. The subsequent trial started 1 s after the participant pressed the ÔenterÕ key. All stimuli were shown in small-size white lettering on a black background screen. The mirror primes were presented in mirrored ÔBoxie17Õ font, the stem and response letters being written in ÔChicago23Õ font. This difference in font prevented a line-up of the perceptual trace of the prime with the word stem (see Fig. 2). The instruction was to fixate the position mark (which was in fact the mirror prime), which indicated where the word stem would appear. The word stem had then to be completed as soon as possible with exactly two letters from the first appropriate German noun that came to mind. Participants were informed that the task was a difficult one, and that the next stem would be presented immediately, that is 1 s after their input. Fig. 2. Example of a stimulus sequence. The prime was shown for 750 ms (left) and then replaced by the stem (right). The visual overlap of the primeÕs perceptual trace with the subsequent stem as it appeared on the screen is shown in the middle. 262 W.J. Perrig, D. Eckstein / Consciousness and Cognition 14 (2005) 257–277 At the end of the test, participants were asked three consecutive questions: (1) whether they had noticed anything unusual in the patterns of the position marks, (2) whether they had seen single letters ‘‘hidden’’ in them, and (3) whether they had seen whole words in them. 2.2. Results Analysis of responses to the post-task questions revealed that only one person reported having seen words in the ‘‘position marks.’’ The data of this person were removed from subsequent analysis. Data from one other participant had to be excluded because he had completed word stems to form nonsense strings instead of real words. The data are presented in Table 1. Analysis was performed on target completion rates and response times separately. Target completion rates were computed as the number of target completions divided by the number of trials per condition; response latencies were defined as the latency between the stem presentation onset and the moment of pressing the ÔenterÕ key upon completion of each word. An a level of .05 was used for statistical tests. As can be seen in Table 1, highest completion rates and fastest response times were found in the same prime condition. Contrast tests showed that completion rates were significantly higher in the same prime condition compared to the nonword condition and the different prime condition, with F (1, 26) = 7.85, p = .009, and F (1, 26) = 12.64, p = .001, respectively. There was no significant difference between the different prime condition and the nonword condition, F (1, 26) = 1.22, p = .28. In the same way, response times in the same prime condition were significantly shorter than in the nonword condition, as well as in the different prime condition, with F (1, 26) = 4.47, p = .044, and F (1, 26) = 4.31, p = .048, respectively. There was no significant difference between the nonword condition and the different prime condition, F (1, 26) = 0.09, p = .77. 2.3. Discussion First of all, Experiment 1 demonstrates that the mirror-masking paradigm can successfully establish a situation of inattentional blindness in participants. Out of the 29 subjects only one reported noticing ‘‘something special.’’ From this subjective failure to see any letters or words hidden in the stimuli, we can fairly safely conclude that participants did not deliberately apply any kind of explicit knowledge about the mirror primes that were ostensibly presented as ‘‘position markers.’’ Indeed, not only did they not apply that knowledge, they did not appear to have acquired it. The data do, however, demonstrate clear priming effects in the same prime condition, Table 1 Proportional target completion frequencies (hit rate) and response times for completing word stems in Experiment 1 (N = 27) Prime condition Hit rate Response times (s) Same prime Different prime Nonword prime 0.31 (0.13) 6.42 (2.33) 0.19 (0.10) 7.37 (3.02) 0.22 (0.09) 7.28 (2.98) Note. Standard deviations are given in brackets. W.J. Perrig, D. Eckstein / Consciousness and Cognition 14 (2005) 257–277 263 where the word stem of the mirror-masked word has to be completed. The measured 31% target completion rate was well above the 22% completion rate from the norms, and it differed significantly from the nonword prime control condition. The demonstration of same word priming is as such already important. The data of this first experiment suggest that mirror masking might be a practical way of proceeding systematically in the investigation of unconscious information processing. The really interesting question, of course, is what might have caused the priming effect found in our mirror-masking paradigm. At the present time, there are several different explanations, all equally possible. For example, the prime could have activated lexical units, i.e., the words. The priming effect showing up in the same prime condition, where the word stem of the presented word has to be completed, could be either a semantic or a perceptual effect. However, it is also conceivable that no lexical entity is activated at all. In this case, the priming effect would be based solely on the repeated word stem, which improves access to words in the subsequent word-stem completion task. If we assume that completion of a word stem is facilitated by previous activation of meaning, we might expect that activation of a different meaning in the different prime condition could impair performance relative to the nonword prime condition. Such a pattern of results is not evident in our data. Therefore perceptual information is the more likely source of the priming found in Experiment 1. However, these priming effects cannot be interpreted as being the result of the lining up of a perceptual trace of prime and target, because the use of different fonts and slightly different positions prevented an exact perceptual line up of the two stimuli. In order to achieve a replication and to investigate further the source of priming, a second experiment was conducted that manipulated semantic relatedness of the prime words. Two treatment groups were incorporated in order to compare effects of conscious and unconscious perception. 3. Experiment 2 Two major changes were introduced in Experiment 2 compared to Experiment 1. First, a fourth prime condition (related prime) was added to the three prime types of Experiment 1 (same prime, different prime, and nonword). In the related prime condition, prime words were semantically related to the target word to be completed. This condition allows testing for semantic priming with respect to the unrelated prime condition. Second, two more groups were tested in addition to the ignorant group. Participants in these groups were informed about the content of the mirror primes. One informed group was merely informed about the fact that words could be hidden in the prime patterns, and a second, trained group was trained to recognize words in the prime patterns before proceeding to the experimental task. This manipulation was provided to compare the performance of those who were expected to show unconscious priming with that of participants who had conscious access to the prime information. From the data in Experiment 1, we assumed that perceptual information might be the primal source for the priming effect found in the same prime condition. On this premise, we would expect to find dissociations in performance between the ignorant group, the informed and the trained group, especially in the related prime condition. In particular, we would expect the informed and the trained group to score very well in the same prime and the related prime conditions, 264 W.J. Perrig, D. Eckstein / Consciousness and Cognition 14 (2005) 257–277 because the same prime or associated word meaning would be activated when participants (knowingly) read the word primes. The ignorant group, however, would be expected to show priming only in the same prime condition. Such a pattern of results, if found, would clearly dissociate conscious from unconscious processing in objective performance data (see Merikle & Daneman, 1998). 3.1. Method 3.1.1. Participants and design The ignorant sample consisted of 29 psychology students participating for credit, and 14 students participating in a linguistics course (34 women and 9 men). The informed sample consisted of 25 psychology students (19 women and 6 men) participating for credit. The trained group consisted of 18 psychology students (13 women and 5 men) participating for credit. The four prime conditions were varied within subjects. These were: (a) same prime condition, (b) related prime condition, whereby the prime word was semantically related to the target word, (c) different prime condition, and (d) nonword prime condition. Each participant saw 13 test stems in each of the three word prime conditions, and 39 test stems in the nonword prime condition. 3.1.2. Materials In order to have a collection of words related to the targets, we first chose 72 words out of the standardized set of Meier and Eckstein (1998). In a separate session, semantic word-stem completion priming effects of semantically related and unrelated four- to eight-letter word primes were measured, first in a paper-and-pencil test and then computerized with 30 participants in each test. None of these participants took part in the actual experiment. Based on the test results, 52 target words were selected out of this set with their corresponding semantically related and unrelated words. The 52 word triplets (word-related word-unrelated word) were then divided into four lists of 13 triplets, each being equal with respect to completion frequency. Within each group, the baseline target completion probability averaged p = .18. Twenty-six distractor stems were selected from a set of word stems, each with two possible word completions (nouns). These 26 stems were used as target stems in combination with nonword primes to allow for a balanced overall word to nonword prime proportion at test. Finally, 39 nonwords were constructed. These equalled the prime words in length (see Appendix B for the word lists). 3.1.3. Procedure Subjects in the three different groups received different instructions before proceeding to the experimental word-stem completion task. Subjects in the ignorant group were not informed about the presence of words in the masks. Subjects in the informed group were informed about the construction of the prime stimuli and were shown an example of the mirror primes. Subjects in the trained group were first informed about the mirror-masking technique, and then they received a practice block (with response feedback) in which they had to identify words and nonwords in the mirror primes. Subjects practised until they were able to recognize most word primes. None of these stimuli were part of the following experimental stimulus set. The instruction for the wordstem completion task was exactly the same for all three groups. W.J. Perrig, D. Eckstein / Consciousness and Cognition 14 (2005) 257–277 265 The procedure at test was the same as in Experiment 1, with three exceptions: (a) each participant saw 13 word primes in every experimental condition, and 39 nonword primes, (b) the time to complete the word was limited to 12 s; the next trial then started immediately, and (c) in 26 nonword trials, nonstandardized word stems with only two possible noun completions were presented. Completions of these latter word stems were not analyzed. The four stimulus lists were rotated across conditions and subjects. At the end of the experiment, participants were questioned about their retrospective awareness of having seen any letters or words in the prime patterns, and participants in the informed group were asked to estimate the proportion of words they had seen in the mirror masks. 3.2. Results Out of the 43 participants in the ignorant group, six reported having seen letters or words in the masked patterns. The following analyses were therefore performed only on the data of the 37 participants who did not detect any letters or words in the prime patterns. Registered response times of one person were lost due to a technical problem, reducing the sample to 36 for latency analyses. In the informed group, five subjects reported having seen more than 50% of the prime words, 13 subjects reported having seen at least one word, and seven subjects reported they did not notice any words at all. Analysis was performed on hit rates and response times separately. Hit rates were computed as the number of target completions divided by the number of trials per condition; response latencies were defined as the latency between the stem presentation onset and the moment of pressing the ÔenterÕ key upon completion of each word. An a level of .05 was used for statistical tests. For tests of priming, the different prime condition served as baseline. The data are presented in Table 2 (raw data) and Fig. 3 (A and B; priming data). Priming was computed as the difference between the same prime resp. the related prime conditions and the different prime condition. Table 2 Proportional target completion frequencies (hit rates) and response times for completing words in Experiment 2 Prime condition Same prime Related prime Different prime Nonword prime Ignorant group (N = 37) Hit rate Response times (s) 0.24 (0.13) 4.8 (1.1) 0.16 (0.11) 5.1 (1.2) 0.18 (0.11) 5.0 (1.5) 0.18 (0.10) 5.1 (1.2) Informed group (N = 25) Hit rate Response times (s) 0.35 (0.11) 3.5 (0.8) 0.25 (0.12) 4.3 (1.0) 0.21 (0.10) 4.4 (1.1) 0.23 (0.11) 4.5 (1.3) Trained group (N = 18) Hit rate Response times (s) 0.41 (0.18) 4.1 (1.1) 0.26 (0.14) 4.9 (1.2) 0.17 (0.07) 4.9 (1.3) 0.19 (0.09) 5.3 (1.9) Note. Standard deviations are given in brackets. Due to missing response time registration of one person in the ignorant group, response times are based on 36 subjects. 266 W.J. Perrig, D. Eckstein / Consciousness and Cognition 14 (2005) 257–277 Fig. 3. Hit rate (A) and latency priming (B) in Experiment 2. Significant priming is indicated by an asterisk. The bars indicate standard errors. First, equality of hit rates in the nonword prime and different prime control conditions was verified with a 2 (prime: different vs. nonword) · 3 (group: ignorant, informed, and trained) analysis of variance on hit rates. Neither the effect of prime type, F (1, 77) = .95, p = .33 nor the effect of group, F (2, 77) = 2.16, p = .12, nor the interaction were significant, F (2, 77) = .28, p = .75. In consequence, the nonword prime condition was not considered any more in the following analyses. Next, prime condition and group effects were analyzed by a 3 (prime condition: same, related, and different) · 3 (group: ignorant, informed, and trained) analysis of variance. Hit rates differed W.J. Perrig, D. Eckstein / Consciousness and Cognition 14 (2005) 257–277 267 significantly between prime conditions, F (2, 154) = 32.32, p < .001 and groups, F (2, 77) = 11.36, p < .001, and there was a significant interaction of prime condition and group, F (4, 154) = 3.83, p = .005. Repeated contrast tests on the prime condition showed that hit rates were larger in the same prime than in the related prime condition, F (1, 77) = 28.64, p < .001, and larger in the related prime than in the different prime condition, F (1, 77) = 4.70, p = .033. Contrasts on the group factor showed that ignorant subjects completed fewer target words than informed subjects, with an estimated difference of M = 0.076 (SE = 0.020), p < .001, whereas there was no evident difference between informed and trained subjects, M = 0.012 (SE = 0.023), p = .63. Repeated contrasts on the prime condition · group interaction revealed that the group factor interacted with the different prime vs. same prime condition, F (2, 77) = 7.81, p = .001, and different prime vs. related prime condition, F (2, 77) = 3.30, p = .042. However, there was no evident interaction with the same prime vs. related prime condition, F (2, 77) = .94, p = .40. As can be seen in Table 2, same word priming was significantly larger for nonignorant than ignorant groups, with a difference in priming of M = .08, t (60) = 2.04, p = .046 and M = 0.17, t (53) = 3.83, p < .001 for the informed and trained group, respectively. Conversely, related word priming was only significant in the trained group, t (17) = 2.25, p = .016, whereas there was a tendential effect in the informed group, t (24) = 1.49, p = .077 and no significant effect in the ignorant group, t (36) = 0.78, p = .22 (one-tailed tests). Response times were analyzed with multivariate analyses, as the sphericity assumption for univariate analysis had to be rejected. Degrees of freedom were corrected on WilkÕs Lambda criterion. Response latencies were equal in nonword and different word conditions, as there was no significant effect of prime condition, F (1, 76) = 1.07, p = .30 and no interaction with group, F (2, 76) = .33, p = .72. Response times differed significantly between conditions, F (2, 75) = 15.51, p < .001 and groups, F (2, 76) = 7.00, p = .002, but with no significant interaction, F (4, 150) = 1.45, p = .22. The effect of prime condition was due to the faster reactions for same primes compared to related primes, F (1, 76) = 21.53, p < .001, as there was no difference between the latencies in the related and the different prime condition, F (1, 76) = 0.0, p = .98. Further tests revealed that the same resp. related word priming effects appeared only in the informed group, t (24) = 4.60, p < .001 resp. t (24) = 4.35, p < .001 and the trained group, t (17) = 2.25, p = .038 resp. t (17) = 2.48, p = .024, whereas there was no latency priming in the ignorant group, F (1, 36) < 1.0. Over all conditions, trained subjects gave faster responses than informed subjects with an estimated difference of M = 57 ms (SE = 28 ms), p = .041, and informed subjects gave faster responses than ignorant subjects, M = 87 ms (SE = 23 ms), p < .001. In summary, priming in the different conditions depended on the group factor. Whereas there was a clear same prime effect in all groups on hit rates and response latencies, a related prime effect was only evident in hit rates of informed and trained subjects. Informed and trained subjects completed prime words more frequently than ignorant subjects. Furthermore, an effect of information was evident in response latencies. The fastest responses were given by trained subjects, and the slowest responses were given by ignorant subjects. 3.3. Discussion Experiment 2 nicely replicates our findings from Experiment 1 in showing significant priming effects in the same prime condition for subjects who did not (consciously) identify the prime stim- 268 W.J. Perrig, D. Eckstein / Consciousness and Cognition 14 (2005) 257–277 uli. Thus again, the mirror-masking technique proved to be a reliable paradigm for investigating unconscious information processing. In addition to this replication, Experiment 2 reveals valuable insight into the question of the source of the unconscious priming effect, and its possible differences to conscious processing of the prime stimuli. To test whether semantic information alone could be responsible for priming effects a related prime condition was introduced in Experiment 2. The prime patterns in this condition consisted of mirror prime words that were semantically related to the target word stem to be completed. As the results show, participants who were effectively blind to the hidden word information in the priming pattern, showed no significant priming effect in the related prime condition. This stands in sharp contrast to the performance of the two other groups introduced in Experiment 2: one group being informed about the mirror-masking construction of the prime patterns, and the other being trained to identify the information hidden in the prime patterns. These two groups showed overall significantly higher priming effects than the ignorant group. Most remarkable of all, the knowledgeable groups showed significant semantic priming in the related prime condition, whereas the ignorant group was obviously not able to extract semantically related information to help them with word-stem completion. In summary, Experiment 2 replicates the unconscious priming effects in the same prime condition found in Experiment 1. Furthermore, the priming effects dissociate conscious from unconscious processing in the different experimental conditions. In line with the dissociation logic of Merikle and Daneman (1998), this lends further support to the functionality of our mirror-masking paradigm. The missing priming effect in the related prime condition of the ignorant group corroborates our conclusion from Experiment 1 that the primal source for the unconscious priming effect is probably perceptual in nature. To investigate further the characteristics of the hidden prime information, a third experiment was conducted to test for the durability of the memory traces left by the mirror-masked prime words. 4. Experiment 3 Experiment 1 and Experiment 2 delivered support for the mirror-masking paradigm as a successful tool in producing a situation of inattentional blindness (i.e., where subjects do not/cannot consciously identify word information contained in prime patterns). The subsequent act of wordstem completion is, nevertheless, influenced by this prime information. Completion rate of targets increases when word stems of the words mirrored in the prime patterns have to be completed. The available data at this time suggest that some kind of a perceptual memory trace influences this performance. The question now arises of how durable this trace might be. Priming effects in implicit memory research have proved to be long lasting. Concerning the effects of masked primes, however, we do not know of any study investigating the time course of unconscious priming effects. Consequently, there is no reference data to elaborate on the expected outcome of Experiment 3, which was designed to investigate the duration of the memory trace of a mirror-masked prime influencing word-stem completion. For this end, target stems were shown after an interval of 1, 2 or 3 s after prime onset. In order to avoid unnaturally long trials, 3 s would be the longest sensible delay after a ‘‘position mark’’ (mirror prime pattern) has indicated the location where the word stem will shortly appear. Besides the manipulation of the SOA between prime pattern and word stem, W.J. Perrig, D. Eckstein / Consciousness and Cognition 14 (2005) 257–277 269 the following differences to Experiment 2 were made: only the same prime condition (word stem of the mirrored word-prime), and a nonword prime condition were used. Moreover, no group was trained to see the words in the prime patterns. It was hypothesized that Experiment 3 should produce reliable priming effects, at least at the shortest SOA condition, and with an established lack of awareness of the hidden word information in the prime patterns. 4.1. Method 4.1.1. Participants and design The ignorant sample consisted of 22 students (13 women and 9 men) participating for credit. The informed sample consisted of 16 women and 6 men, all students participating for credit. Word and nonword prime conditions were varied within subjects. In the three word prime conditions, target stems were shown 1, 2 or 3 s after the mirrored word-prime pattern, whereas in the nonword prime condition, the target stem was shown 1 s after the nonword prime. 4.1.2. Materials Material and word groups were the same as in Experiment 2. However, only the same prime condition was used: target stems consisted of the three first letters of word primes. 4.1.3. Procedure Each participant saw 13 target stems in the 1, 2, 3 s SOA conditions, and 39 target stems in the nonword prime condition. Order of trials was randomly determined before testing, and the four balanced word stimulus groups were rotated between conditions and participants. Participants in the two groups differed only in the information they received before being tested. Participants in the informed group were shown an example of the construction of the mirror primes, and they were told that such primes would be presented in the test. The other group was left ignorant concerning the construction of the prime patterns. After that, the usual instruction was displayed on the screen. Both groups were asked to do the task exactly as asked in the displayed instructions. As in the previous two experiments, primes were presented for 750 ms at one of four positions on the screen. Subjects were given 15 s to enter their completions according to the first noun that came to mind. At the end of the experiment, subjects were questioned about their retrospective awareness of having seen any letters or words in the prime patterns. Participants in the informed group were asked to estimate the proportion of words they had seen in the masks. 4.2. Results Out of the 22 participants in the ignorant group, only two reported having seen letters or words in the masked patterns. The following analyses were therefore performed on the data of the 20 participants who did not detect any letters in the patterns. On average, the informed group reported seeing ca. 50% of the words in the primes (six subjects reported seeing more than 50%, seven subjects 50%, and nine subjects less than 50%). Analysis was performed on hit rates and response times separately. Hit rates were computed as the number of target completions divided by the number of trials per condition; response latencies were defined as the latency between the stem presentation onset and the moment of pressing the Ôen- 270 W.J. Perrig, D. Eckstein / Consciousness and Cognition 14 (2005) 257–277 Table 3 Proportional target completion frequencies (hit rates) and response times for completing words in Experiment 3 SOA of word prime and target Nonword primes One second Two seconds Three seconds Ignorant group (N = 20) Hit rate Response times (s) 0.29 (0.16) 4.8 (1.0) 0.27 (0.16) 5.0 (1.1) 0.27 (0.14) 4.9 (1.1) 0.20 (0.10) 5.7 (1.7) Informed group (N = 22) Hit rate Response times (s) 0.43 (0.21) 3.6 (1.3) 0.45 (0.16) 4.2 (1.5) 0.44 (0.16) 4.1 (1.3) 0.06 (0.04) 4.3 (0.7) Note. Standard deviations are given in brackets. terÕ key upon completion of each word. An a level of .05 was used for statistical tests. Completion rates and response times for completing a word are presented in Table 3. Because variances and covariances differed between conditions over all groups due to the control condition, analysis was performed separately for the groups on the completion rates and latencies. Next, the priming effect was computed, and stability of responses over different SOAs was computed for both groups. In a one-factorial ANOVA with four levels (prime condition: same-word prime with 1 s SOA, 2 s SOA and 3 s SOA, and nonword prime with 1 s SOA) on completion rates of the ignorant group priming condition was significant, F (3, 57) = 3.17, p = .031. Hits were always more frequent in experimental compared to the control condition as tested with contrasts, F (1, 19) = 7.64, p = .012, F (1, 19) = 4.70, p = .043, and F (1, 19) = 4.68, p = .044 for the 1, 2, and 3 s SOA conditions, respectively. As can be seen in Table 3, priming was even stronger for the informed group, F (1, 21) > 55.0, p < .001. The same analysis on reaction latencies revealed a significant main effect of prime for the ignorant group, F (3, 17) = 4.69, p = .015 (multivariate tests were computed because of a lack of sphericity). Contrasts showed tendentially significant differences in experimental vs. the control condition, F (1, 19) = 3.56, p = .074, F (1, 19) = 4.66, p = .044, and F (1, 19) = 11.13, p = .003 for the 1, 2, and 3 s SOA conditions, respectively. Priming was also evident in the latencies of the informed group, F (3, 19) = 4.69, p = .013. However, contrasts revealed that only the 1 s SOA latencies were significantly lower than those of the control condition, with F (1, 21) = 10.25, p = .004, F (1, 21) = .04, p = .84, and F (1, 21) = .57, p = .46 for the 1, 2, and 3 s SOA conditions, respectively. In a 3 (SOA: 1, 2, and 3 s) · 2 (group: ignorant vs. informed) analysis of variance on completion priming, the only significant finding was that of a group effect reflecting greater priming in the informed group, F (1, 40) = 47.44, p < .001. Other F-values of main effects, interaction, and contrast tests were below 1.0. Group differences were not significant in the same analysis over latencies, F (1, 40) = 2.16, p = .15, neither was the main effect of SOA, F (2, 80) = 1.66, p = .20, nor the interaction of SOA and group, F (2, 80) = 0.931, p = .40. Consequently, there were no signs of change of priming over time. 4.3. Discussion In Experiment 3, influence of prime–target SOA of 1, 2, and 3 s on priming was investigated. ParticipantsÕ subjective reports show that the mirror-masking paradigm can be used W.J. Perrig, D. Eckstein / Consciousness and Cognition 14 (2005) 257–277 271 successfully with quite long prime–target SOAs as demonstrated here, up to 3 s at least. More importantly, we again find very robust priming effects in word-stem completion and reaction times in the ignorant group. This means that, even in situations where the participant has quite a long time to prepare for the appearance of the word stem, there is no hint that participants might be able to gain some insight into the construction principle of the mirrored prime patterns. Participants in Experiment 3 showed priming effects to the same degree as those in Experiments 1 and 2. In Experiment 3, the informed group showed much higher completion rates in all three SOA word-prime conditions, although their answers were not much faster compared to the completion after nonword primes. In fact, their very low completion rates after nonword primes are particularly striking. This suggests that participants in the informed group tended to rely very much on the prime information and were therefore inhibited in their ‘‘brainstorming’’ when the prime was a nonword. Thus again, there are behavioral differences between conscious and the unconscious use of prime information in Experiment 3 in the data pattern. Experiment 3 demonstrates the robustness of the unconscious priming in so far as it can last for at least 3 s with no indication of decay over the investigated prime–target intervals. 5. General discussion The aim of this study was to investigate unconscious priming by the use of a spatial mirrormasking paradigm. The method presented here can be considered as an alternative to temporal masking techniques in the investigation of unconscious priming effects. Words and nonwords with no under-length letters are mirrored at their horizontal axis. The results are figures of geometriclike forms that contain potentially legible words in the upper part. The mirrored stimuli contain spatial (visual) information (in the lower part) that should prevent detection of the words. In the three experiments reported in this study, such mirrored words and mirrored nonwords were used as primes in a priming procedure. Groups of naive participants were not informed about the nature of the construction of the stimuli. Instead, their attention was drawn away from the mirrormasked stimuli by informing them that ‘‘position markers’’ (the mirror prime patterns) would indicate where (out of four possible locations on the screen) the next word stem to be completed would appear. The first important finding shows that subjects did not detect the hidden words in the prime patterns. This was revealed by three retrospective questions that were given after the experiment. Participants were asked directly whether they had seen any letters or words in the ‘‘position markers’’ (in the mirror prime stimuli themselves). Overall, out of 94 subjects only nine reported having noticed words or letters in the masks. Thus, we succeeded in producing a kind of inattentional, or cognitive, blindness for words. It is argued that participants did not notice the information simply because they did not expect it to be there. Nevertheless, these ignorant or cognitively blind groups show reliable priming in all three experiments. This is the most striking finding of the present study, because, as a new experimental paradigm, it seems to open new possibilities in the investigation of unconscious priming. In the present study, unconscious priming occurs in the same prime condition of word-stem completion, 272 W.J. Perrig, D. Eckstein / Consciousness and Cognition 14 (2005) 257–277 where stems taken from words in the prime patterns have to be completed. Priming was evident in completion rates (between 6 and 9%) as well as in response times (between 12 and 23%). It was also shown that the unconscious priming effect was not affected by an increase in prime–target SOAs of up to 3 s (Experiment 3). Neither different words in the prime patterns in Experiment 1, nor semantically related prime words in Experiment 2 affected completion rates or response times of ignorant participants. This means that there is no evidence of unconscious semantic lexical priming in our study. This conclusion clearly stands in sharp contrast to the findings of Marcel (1983), that demonstrated semantic association to be the strongest factor in forced choice decisions in a task requiring selection of the one word (out of a word pair) that was similar to a previously masked prime word. Unfortunately, the paradigm from MarcelÕs study never evolved into a proper, extensive research programme in unconscious priming. The perceptual nature of the priming found in our study seems rather more congruent with findings presented by Klinger et al. (2000), Abrams and Greenwald (2000), and Greenwald and Abrams (2002). These three studies, although demonstrating reliable unconscious priming, do not lend support to semantic influences in unconscious priming. They rely on the well-established, ‘‘sandwich-masking,’’ response-window technique of Greenwald and his collaborators (see Draine & Greenwald, 1998). Klinger et al. (2000) present strong evidence that unconscious priming found in the response-window paradigm is not the result of spreading activation in a meaning associated net, but rather the result of response competition in a two-alternative categorical decision. Abrams and Greenwald (2000) resp. Greenwald and Abrams (2002) show the priming effect to be mediated by perceptual, sub-lexical word parts, and not by meaning. Reference to these studies asks for a comparison of our mirror-masking technique with the response-window paradigm. A number of differences immediately pop out. The first obvious one relates to the task itself. In the response-window paradigm, participants are trained to identify and to differentiate stimuli (e.g., lexical decision), or to evaluate features of stimuli (emotional valence) under time pressure. The hit rate is the only indicator for eventual priming. In our wordstem completion task subjects need not be trained for fast responses, and response times are, in addition to completion rates, possible indicators for priming. Another important difference is that in our experiments priming is shown in conditions where primes and targets were only ever presented once, while in the Greenwald paradigm words are repeated as primes and target several times throughout the experiment. As Abrams and Greenwald (2000) and Damian (2001) show, the previous presentation of the primes as targets seems to be a necessary condition for priming effects to appear. In the response-window technique, manipulated features of the stimulus material (e.g., emotion valence) only affect performance in the response task if these features are to be evaluated (Klinger et al., 2000). In contrast with this finding, the priming effect of our experiments does not rely on such restrictions. These differences could suggest that the response-window technique is especially apt to study automatized stimulus–response mappings in the tradition of prime categorization studies (Damian, 2001; Kunde et al., 2003), while the mirror-masking priming reflects one-trial memory on the perceptual level of letter representation or lexical representation. This assumption follows from the fact that we used different fonts for the prime and the target stimuli. But of course, this is but a coarse grained differentiation of the poten- W.J. Perrig, D. Eckstein / Consciousness and Cognition 14 (2005) 257–277 273 tially relevant fields to situate explanations for the priming effects found in the two different experimental paradigms. To summarize, the mirror-masking paradigm presented here can effectively be used to mask words in priming research, and to demonstrate replicable unconscious perceptual priming. The advantages of the method are technical simplicity, threshold-independence, and unlimited prime duration. Its practicability has been demonstrated here in a word completion task, but the general method could be used in many different types of implicit memory tests. Of course, as in all research on implicit cognition, it is difficult to infer from post-experimental reports the exact state of phenomenological awareness of participants during the experiment itself. Here, we wish to defend the position that participantsÕ (admittedly subjective) perceptual reports can have, a priori, the same scientific status as behavioral data, and consequently they should be treated with the same care. Fortunately we have striking demonstrations of sustained inattentional blindness even for dynamic events (Simons & Chabris, 1999). We would argue that our mirror-masking paradigm can be considered as a method for studying the role of implicit memory in such mental states. Besides retrospective reports about perceptions, we have a second control in our study by comparing performance of ignorant groups and informed groups. The informed groups show overall larger priming effects than the ignorant groups. Furthermore, the informed groups very clearly show priming when semantically related primes are presented, while the ignorant group show no priming at all in that condition. This dissociation in performance corroborates further the usefulness of the mirror-masking technique in the study of unconscious priming effects. Appendix A. Words and nonwords in Experiment 1 Stems Same primes Different primes Nonwords bus ess fro kre tra sto zwe tre stu rei mil leh fra kan lin dra sei wei war tan busch essen front krebs trank stoff zweck trend sturm reise milbe lehne frack kanal linse drall seife weide warte tanne abend boden elend flaum hitze konto meile notar orden rauch dreck fluch niete rinde trahe botme hhcle hilem ehsch hrhho hefns ronue ritth hmifu nrmhe cinrt nimie udsfh mfnit canvh nmneb nurts hciak lnaet (continued on next page) 274 W.J. Perrig, D. Eckstein / Consciousness and Cognition 14 (2005) 257–277 Appendix A (continued) Stems Same primes rad kur kar bra for kla lau str due sal tro sch sti kas kri hau bon fal kom sta tot kam dat mot fei man hal lot radau kurde karre braue forum klaue laune stroh duett salbe troll schal stier kasse krimi hauch bonus falte kombi stadt totem kamin datum motor feind manie halde lotus Different primes Nonwords wrtzi voahh ovnez zltri rbuid kdsfl hdrui anrlh Appendix B. Words and word stems in Experiments 2 and 3 Stems 1 Stems 2 Same primes Related primes Nonrelated primes alt blu dau fah fer fla flo gil haf kat ker kun leb kle bon fra kan kla kre man min rad sei sta str tot tro bra bonze frack kante klaue krebs manko minus radau seife stadt stroh totem trost brand reichtum musiker ecke huf tumor fehler verlust aufruhr waschen dorf halm indianer affekt funke haut stunde idee lexikon schwein zeit brust wurst deutsch bissen farbe anwalt blau ehre W.J. Perrig, D. Eckstein / Consciousness and Cognition 14 (2005) 257–277 275 Appendix B (continued) Stems 1 Stems 2 Same primes Related primes Nonrelated primes mas nie ord pau rie rud seg sum ten ton tru wol bus ess fal kas kna kom kri lin stu the tra wan due fei hal kar kur lau leh mil lot rei sal tan wei dra fro hau kam kra lun mot sch sti sto tre tri zwe busch essen falte kaste knall komet krimi linse sturm theke trank wanze duett feind halde karma kurde lauch lehne milbe lotus reife salbe tanne weide draht frost hauch kamin krach lunch motor schal stier stoff treue trick zweck dickicht mahlzeit riss indien schuss leere detektiv brille orkan bardame durst insekt trio hass schutt hindu tuerke zwiebel sofa biene asien samen arznei baum strauch achse winter atem feuer streit imbiss antrieb mantel torero materie liebe zauber nutzen kirche schrift schein film frucht lacher dasein made inhalt dusche bahnhof nuss fassade artikel brief bosheit kolibri himmel stich heer hehler koffer fazit kaiser ruhm leute tasche wert hilfe sinn wolle abend ballast werk einsicht charme schneise alkohol References Abrams, R. L., & Greenwald, A. G. (2000). Parts outweigh the whole (word) in unconscious analysis of meaning. Psychological Science, 11(2), 118–124. Buchner, A., & Wippich, W. (2000). On the reliability of implicit and explicit memory measures. Cognitive Psychology, 40(3), 227–259. Cohen, N. J. (1984). Amnesia and the distinction between procedural and declarative knowledge. In L. R. Squire & N. Butters (Eds.), The neuropsychology of memory (pp. 83–103). New York: Guilford Press. 276 W.J. Perrig, D. Eckstein / Consciousness and Cognition 14 (2005) 257–277 Curran, T., & Schacter, D. L. (2001). Implicit learning and memory, psychological and neural aspects. In N. J. Smelser & P. B. Baltes (Eds.), International encyclopedia of the social and behavioral sciences (pp. 103–125). Kidlington: Elsevier Science. Damian, M. F. (2001). Congruity effects evoked by subliminally presented primes: Automaticity rather than semantic processing. Journal of Experimental Psychology: Human Perception and Performance, 27(1), 154–165. Draine, S. C., & Greenwald, A. G. (1998). Replicable unconscious semantic priming. Journal of Experimental Psychology: General, 127, 286–303. Forster, K., Booker, J., Schacter, D. L., & Davis, C. (1990). Masked repetition priming: Lexical activation or novel memory trace?. Bulletin of the Psychonomic Society, 28(4), 341–345. Greenwald, A. G., & Abrams, R. L., 2002. Visual masking reveals two qualitatively different levels of unconscious cognition. Paper presented at 43rd Annual Meeting of the Psychonomic Society, Kansas City, MO, November, 2002. Greenwald, A. G., Draine, S. C., & Abrams, R. H. (1996). Three cognitive markers of unconscious semantic priming. Science, 273, 1699–1702. Greenwald, A. G., Klinger, M. R., & Schuh, E. S. (1995). Activation by marginally perceptible (‘‘subliminal’’) stimuli: Dissociation of unconscious from conscious cognition. Journal of Experimental Psychology: General, 124, 22–42. Hasher, L., & Zacks, R. T. (1979). Automatic and effortful processes in memory. Journal of Experimental Psychology: General, 108, 356–388. Jacoby, L. L. (1991). A process dissociation framework: Separating automatic from intentional uses of memory. Journal of Memory and Language, 30, 513–541. Kihlstrom, J. F. (1987). The cognitive unconscious. Science, 237, 1445–1452. Klinger, M. R., Burton, P. C., & Pitts, G. S. (2000). Mechanisms of unconscious priming: Response competition, not spreading activation. Journal of Experimental Psychology: Learning, Memory, and Cognition, 26(2), 441–455. Kunde, W., Kiesel, A., & Hoffmann, J. (2003). Conscious control over the content of unconscious cognition. Cognition, 88, 223–242. Kunst-Wilson, W. R., & Zajonc, R. B. (1980). Affective discrimination of stimuli that are not recognized. Science, 207, 557–558. Marcel, A. (1983). Conscious and unconscious perception: Experiments on visual masking and word recognition. Cognitive Psychology, 15, 197–237. Meier, B., & Eckstein, D. J. (1998). Experimentiermaterial für die Untersuchung impliziter und expliziter Gedächtniseffekte mit der Wortanfangs-Ergänzungsaufgabe. Sprache und Kognition, 17(1), 89–105. Meier, B., Morger, V., & Graf, P. (2003). Competition between automatic and controlled processes. Consciousness and Cognition, 12(2), 309–319. Meier, B., & Perrig, W. J. (2000). Low reliability of perceptual priming: Consequences for the interpretation of functional dissociations between explicit and implicit memory. Quarterly Journal of Experimental Psychology, 53, 211–233. Merikle, P. M., & Daneman, M. (1998). Psychological investigations of unconscious perception. Journal of Consciousness Studies, 5, 5–18. Merikle, P. M., Joordens, S., & Stolz, J. A. (1995). Measuring the relative magnitude of unconscious influences. Consciousness and Cognition, 4(4), 422–439. Perrig, W. J. (2001). Implicit memory. In N. J. Smelser & P. B. Baltes (Eds.), International encyclopedia of the social and behavioral sciences (pp. 103–125). Kidlington: Elsevier Science. Perrig, W. J., Wippich, W., & Perrig-Chiello, P. (1993). Unbewusste Informationsverarbeitung. Huber: Bern. Perruchet, P., Gallego, J., & Pacteau, C. (1992). A reinterpretation of some earlier evidence for abstractiveness of implicitly acquired knowledge. Quarterly Journal of Experimental Psychology A, 44, 193–210. Posner, M., & Snyder, C. (1975). Attention and cognitive control. In R. L. Solso (Ed.), Information processing and cognition: The Loyola Symposium. Hillsdale, NJ: Erlbaum. Reber, R., & Perrig, W. J. (2001). The psychology of perception without awareness. In N. J. Smelser & P. B. Baltes (Eds.), International encyclopedia of the social and behavioral sciences. Kidlington: Elsevier Science. Roediger, H. L., & McDermott, K. B. (1993). Implicit memory in normal human subjects. In H. Spinnler & F. Boller (Eds.), Handbook of neuropsychology (pp. 63–131). Amsterdam: Elsevier. W.J. Perrig, D. Eckstein / Consciousness and Cognition 14 (2005) 257–277 277 Schneider, W. (1988). Micro Experimental Laboratory: An integrated system for IBM-PC compatibles. Behavioral Research Methods, Instruments, and Computers, 20, 206–217. Schneider, W., & Shiffrin, R. M. (1977). Controlled and automatic human information processing: I. Detection, search, and attention. Psychological Review, 84, 1–66. Shanks, D. R., & St. John, M. M. F. (1994). Characteristics of dissociable human learning systems. Behavioral and Brain Sciences, 17, 367–447. Simons, D. J., & Chabris, C. F. (1999). Gorillas in our midst: Sustained inattentional blindness for dynamic events. Perception, 28(9), 1059–1074.

Consciousness and Cognition 22 (2013) 965–974 Contents lists available at SciVerse ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog The influence of cognitive and emotional suppression on overgeneral autobiographical memory retrieval Sang Quang Phung, Richard A. Bryant ⇑ School of Psychology, University of New South Wales, Australia a r t i c l e i n f o Article history: Received 1 August 2012 Available online 19 July 2013 Keywords: Autobiographical memory Thought suppression Emotional inhibition Avoidance a b s t r a c t Over-general autobiographical memory (OAM) retrieval is characterized by retrieval of categoric autobiographical memories. According to the CarFAX model, this tendency may result from avoidance which functions to protect the person against recalling details of upsetting memories. This study tested whether avoidance strategies impact on the ability to retrieve specific autobiographical memories. Healthy participants (N = 51) watched a negative video clip and were instructed to either suppress any thought (thought suppression), suppress any feeling (emotional inhibition), or think and feel naturally (controls) in response to the video. Participants then completed the Autobiographical Memory Test. Participants engaging in either thought suppression or emotional inhibition retrieved fewer categoric autobiographical memories than controls. These findings challenge the affect regulation component of the CarFAX model insofar as they suggest that regulatory strategies that aim to reduce awareness of adverse emotional memories do not necessarily lead to increased recall of categoric autobiographical memories. Ó 2013 Elsevier Inc. All rights reserved. 1. Introduction There is strong convergent evidence that psychological disorders are characterized by overgeneral autobiographical memory (OAM). This retrieval style involves the tendency to retrieve categoric personal memories from one’s past (e.g. ‘all the times that I’ve failed my exams’, ‘my holidays in summer were all great as a child’). This pattern has been observed in those suffering from suicidality (William & Broadbent, 1986), complicated grief (Golden, Dalgleish, & Mackintosh, 2007), major depressive disorder (Wilhelm, McNally, Baer, & Florin, 1997), posttraumatic stress disorder (Harvey, Bryant, & Dang, 1998), eating disorder (Dalgleish et al., 2003), and borderline personality disorder (Jones et al., 1999). Further, this overgeneral retrieval style has been associated with other maladaptive functioning, such as deficits in problem-solving (Evans, Williams, O’Loughlin, & Howells, 1992), reduced ability to plan for the future (Williams et al., 1996), and engagement in maladaptive coping strategies (Jones et al., 1999). Williams and his colleagues (Williams et al., 2007) have proposed the CarFAX model to explain OAM retrieval. This model posits that the retrieval search is impaired prior to locating specific memories by three possible mechanisms: (a) ruminating on why adverse events occur, (b) avoiding reminders of these negative events to regulate more positive affective responses, or (c) depletion of executive cognitive resources available to regulate the retrieval of specific memories. Whereas many studies have examined the impact that rumination (e.g. Sutherland & Bryant, 2007; Watkins & Teasdale, 2001) and the depletion of executive cognitive resources (e.g. Dalgleish et al., 2007) have on the manifestation of OAM, relatively few studies have examined the role of affect regulation or avoidance in the retrieval of OAMs. CarFAX posits that one may abort the retrieval ⇑ Corresponding author. Address: School of Psychology, University of New South Wales, NSW 2052, Australia. Fax: +61 2 9385 3641. E-mail address: r.bryant@unsw.edu.au (R.A. Bryant). 1053-8100/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.concog.2013.06.008 966 S.Q. Phung, R.A. Bryant / Consciousness and Cognition 22 (2013) 965–974 search for a personal memory prematurely as a means of regulating unwanted affective states linked to the memories. Some support for the role of avoidance or affect-regulation in OAM comes from correlational findings that individuals who are less specific in their memory recall score higher on a range of avoidance assessments, including behavioral avoidance, experiential avoidance, and thought suppression (Hermans, Defranc, Raes, Williams, & Eelen, 2005). Other support comes from evidence that less specific autobiographical memories of negative events are associated with less affective response (Raes, Hermans, de Decker, Eelen, & Williams, 2003). 2. Thought suppression One common form of cognitive avoidance involves thought suppression, which requires deliberate inhibition of a specific thought. There is convergent evidence that attempted suppression leads to the paradoxical effect of increased occurrence of the suppressed thought, either during the period of attempted suppression or after suppression has terminated (Wenzlaff & Wegner, 2000). Whereas some studies have reported the initial enhancement effect (e.g. Wegner, Schneider, Carter, & White, 1987), others have noted the subsequent rebound effect (e.g. Davies & Clark, 1998; Harvey & Bryant, 1998), others have found neither of the two effects (e.g. Kelly & Kahn, 1994; Muris et al., 1993), and others have found instead a decrease in the to-be-suppressed stimuli (e.g. Roemer & Borkovec, 1994). One meta-analysis of 28 studies concluded that thought suppression is associated with a small to moderate rebound effect (Abramowitz, Tolin, & Street, 2001). 2.1. Thought suppression and the ironic process theory The prevailing model of thought suppression is the ironic process theory, which posits that thought suppression involves two processes (Wenzlaff & Wegner, 2000). The operating process is a conscious, effortful process that searches for mental contents consistent with the desired state. This process searches for other mental materials as substitutes to engage the person’s thinking away from the to-be-suppressed thoughts. The other is the monitoring process, which is less effortful and occurs simultaneously to inspect consciousness for traces of the to-be-suppressed stimulus so that the operating process can be re-engaged to maintain suppression. Given that the operating process is effortful, it can be truncated in conditions where the person’s cognitive resources are depleted by other competing cognitive tasks. Accordingly, ironic control theory holds that the primary reason attempted suppression is unsuccessful is because under conditions of cognitive load or cessation of suppression, the monitoring process heightens sensitivity to the unwanted thought. Consistent with this proposal, increased cognitive load interferes with suppression (Wegner, Erber, & Zanakos, 1993). 2.2. Thought suppression and OAM retrieval Several studies have examined the effect of attempted thought suppression on the recall of autobiographical memories. Dalgleish and Yiend (2006) asked dysphoric adults to recall a specific negative past event which they were then either asked or not asked to suppress. They found that those suppressing the target memory showed a faster recall of memories of other negative past experiences. It should be noted, however, that Dalgleish and Yiend (2006) examined the effect of suppression on the accessibility, and not the specificity, of autobiographical memories. Extending on this finding, Neufeind, Dritschel, Astell, and MacLeod (2009) focused on a non-clinical sample and examined how suppressing memories related to a distressing video clip affected recall of other autobiographical memories. This study found that greater engagement in thought suppression significantly correlated with (in Study 1), and directly engaging in the suppression of thoughts of a distressing video led to (in Study 2), faster recall of other memories of negative experiences. Both of these studies concluded that the paradoxical effects of thought suppression are not limited to the recall of target memories but also the recall of other autobiographical memories. If suppression is conceptualized as a form of affect regulation, it appears that both the Dalgleish and Yiend (2006) and the Neufeind et al. (2009) studies do not support the Car FAX prediction regarding the affect regulation component of the model. However, neither of these studies specifically compared the predictions of ironic control and CarFAX model in relation to the impact of thought suppression on specific autobiographical memories retrieval. It is worth noting that the effort associated with avoidance is purportedly different between the CarFAX and Ironic Control theories. Whereas the ironic control theory posits an explicitly active avoidance strategy that is cognitive demanding, CarFAX suggests that is a more passive process that can become habitual over time such that it more effortlessly limits the extent to which the retrieval search is completed. The distinction between these two approaches appears blurred, however, because Williams, Watts, MacLeod, and Mathews (1997) also state that although ‘‘it is possible that this functional memory strategy is governed by consciously controlled processes, we do not exclude the possibility that it is a response that is shaped without the involvement of conscious monitoring (p. 134).’’ That is, it is uncertain the extent to which the avoidance functions within the CarFAX model are passive or active. In either case, the CarFAX and Ironic Control theories appear to propose distinct hypotheses regarding avoidance insofar as CarFAX predict that avoidance should lead to the recall of fewer specific autobiographical memories. In contrast, ironic control theory predicts that avoidance in the form of suppression should paradoxically lead to greater awareness of related autobiographical memories. Although previous studies have only demonstrated faster recall of memories following suppression, rather than more specific recall (Dalgleish & Yiend, 2006; Neufeind et al., 2009), we propose that the act of S.Q. Phung, R.A. Bryant / Consciousness and Cognition 22 (2013) 965–974 967 suppressing may activate specific memories because the monitoring process seeks occurrence of the specific unwanted memories, and in this way the event specific knowledge level of the memory system is primed. 3. Emotional inhibition In contrast to thought suppression, emotional inhibition can be defined as ‘‘the lack of emotional responsivity to stimuli that would normally elicit emotional responses’’ (Bryant & Kapur, 2006, p. 281). Whereas thought suppression is a strategy that requires constant monitoring of the suppressed material, emotional inhibition is a generic strategy that attempts to blunt all responses and does not require focus on any specific emotion or memory. 3.1. Emotion inhibition and memory Richards and Gross (2000) demonstrated that participants who inhibited emotions whilst watching an emotional film clip subsequently showed poorer memory for details of that film clip than did participants who were allowed to express their emotions (see also Richards & Gross, 1999). There is also evidence to suggest that emotional regulation compromises performance on concurrent memory functioning. For example, Richards and Gross (1999) found that women who suppressed their emotions to a series of emotion-eliciting slides showed reduced memory for the information that was orally paired with these slides than did women who did not suppress their emotions. This concurs with other findings that participants required to inhibit the expression of emotions in response to an emotional film had shorter endurance in a subsequent hand grip task, as well as solved fewer problems in an anagram task (Baumeister, Bratslavsky, Muraven, & Tice, 1998), suggesting that emotion suppression is cognitively taxing. Gross (1998) suggests a difference between antecedent-focused emotion regulation (which refers to regulating affect as a means of preparing oneself from experiencing undesirable emotions) and response-focused emotion regulation (which refers to regulating affect after the undesirable emotions have been triggered). These two emotion-regulatory processes purportedly differ in their associated cognitive costs (Richards & Gross, 2000); whereas expressive suppression (response-focused) requires the person to continually monitor their own emotions and to correct any discrepancy between what is felt and what is to be expressed, it is cognitively more demanding and effortful than reappraisal (antecedent-focus) which is exercised before the emotional event to alter its emotional significance after which further cognitive work is not necessary. 4. Affect-regulation strategies and OAM Although the CarFAX model specifies that avoidance will result in aborted retrieval searches as a means to regulate adverse emotions, there is a dearth of evidence regarding the differential impact of avoidance strategies on retrieval. Accordingly, the goal of this study was to specify the role that different avoidant strategies play in the retrieval of autobiographical memories by comparing the effects of thought suppression and emotional suppression on the retrieval of autobiographical memories. On the basis of ironic process theory, we hypothesized that thought suppression would result in retrieval of fewer categoric and more specific autobiographical memories; in contrast, the CarFAX model would predict that suppression would lead to more categoric and reduced specific memories. To determine the influence of an alternate avoidance strategy, we also studied the effect of emotion inhibition on autobiographic retrieval. 5. Method 5.1. Participants Participants were 51 (26 females, 25 males) English-speaking psychology undergraduate students of mean age 19.25 years (SD = 4.31) who were studying at the Community College of City University (Hong Kong). Participants were randomized to either the Thought Suppression (n = 17), Emotional Inhibition (n = 16), or Control (n = 18) conditions. 5.2. Materials 5.2.1. Beck Depression Inventory – II (BDI-II; Beck, Steer, & Brown, 1996) The BDI-II is a 21-item self-report questionnaire that measures the severity of affective, somatic, cognitive, and behavioral symptoms of depression in the preceding 2 weeks. Each item is indexed on a 4-point scale, and has excellent reliability (a = .92) and test–retest reliability (r = .90) (Beck et al., 1996). 5.2.2. Beck Anxiety Inventory (BAI; Beck & Steer, 1990) This BAI is a 21-item self-report questionnaire that measures the presence and frequency of physiological, affective, and cognitive symptoms of anxiety. The BAI has good internal consistency (r = .92), test–retest reliability (r = .75), and convergent and discriminant validities (Beck & Steer, 1990). 968 S.Q. Phung, R.A. Bryant / Consciousness and Cognition 22 (2013) 965–974 5.2.3. White Bear Suppression Inventory (WBSI; Wegner & Zanakos, 1994) The WBSI is a 15-statement inventory that indexes the tendency to suppress unwanted thoughts. Participants rate each statement on a 5-point scale on the tendency to engage in various suppressive strategies. The WBSI has good internal consistency (a = .87–.89) and test–retest reliability (r = .69–.92) (Wegner & Zanakos, 1994). 5.2.4. Thought Control Questionnaire (TCQ; Wells & Davies, 1994) The TCQ indexes the frequency of use of five strategies of thought control: distraction, punishment, reappraisal, social control, and worry. The TCQ is a 30-item questionnaire rated on a four-point Likert-type scale (1 = ‘‘never’’, 4 = ‘‘almost always’’). The five subscales possess moderately high internal consistency reliabilities (a = 0.64–0.83). 5.2.5. Video stimulus The target stimulus for the thought suppression and emotional inhibition instructions was a 4-min video clip from the Texas Chainsaw Massacre, depicting a psychopathic killer skinning a victim alive while in the secret view of the victim’s girlfriend. Pilot testing of 16 (8 females, 8 males) undergraduate students rated the video on a 100-point Likert scale (1 = extremely negative, 100 = extremely positive) as very negative (M = 11.44, SD = 17.96). 5.2.6. Visual analogue scales Visual analogue scales (VAS) in the form of a 0–100 point (0 = not at all distressed, 100 = very distressed) were used to assess participants’ mood. Comparable scales were used as manipulation checks to index the extent to which participants complied with thought suppression, emotion inhibition, or control instructions: prior to the instructions and following the AMT, participants were asked to indicate on a 100-point scale (0 = not at all, 100 = very much) how much they tried to not be aware of the video (How much did you try to not be aware of the video?). 5.2.7. Autobiographical Memory Test (AMT; Williams & Broadbent, 1986) Developed by Williams and Broadbent (1986) to assess the specificity of autobiographical memories, the AMT requires participants to recall a specific personal memory in response to a cue word. This study adopted 5 positive (e.g. enjoy, luck), 5 neutral (e.g. boat, actor), 5 negative (e.g. help, frighten), and 2 practice (egg and chocolate) cue words, which had been selected from the 1000 most frequently used English words (Carroll et al., 1971) matched on frequency of usage. Each cue word was displayed individually on A4-sized white paper with order of presentation randomized across participants. For each cue word, participants were given 30 s to describe a specific personal memory which was voice-recorded for coding later. Participants were instructed to provide a different memory for each cue word. If no memory was described within the 30 s, it was scored as an ‘omission’. 5.2.7.1. Coding of autobiographical memories retrieved. A rater blind to participants’ assigned conditions coded recorded memories. A memory was coded as extended if it referred to an event lasting more than a day (‘the holiday I had in December was a bit boring’), categoric if it referred to repeated activities or categories of memories of people or places (‘each time I go to the beach I have fun’), and specific if it referred to an event that occurred at a specific time or place lasting for a day or less (‘on Christmas Day my boyfriend gave me a special necklace as a gift’). A second independent rater, also blind to the aims of the study, coded a random sample of 20% of the taped memories. The mean kappa reliability coefficient was 0.82. 6. Procedure Following written informed consent, participants completed the BDI-II, BAI, WBSI, and TCQ to determine potential baseline differences in tendencies for depression, anxiety, or suppression. Participants then rated their mood on the first VAS (Mood Rating 1). Participants were then presented with the video, after which a second mood rating was obtained (Mood Rating 2). Participants were then randomly assigned to one of the three conditions, and were instructed to follow the instruction until it was terminated (i.e., a 5-min period). Participants in the thought-suppression condition were instructed to ‘‘try their hardest to suppress any thought related to the video they had watched until the end of the experiment’’. Participants in the emotional inhibition condition were instructed to ‘‘try their hardest to suppress any feeling related to the video they had watched until the end of the experiment’’. Participants in the control condition were instructed to ‘‘think and feel normally about the video they had watched’’. The number of times the word ‘video’ was mentioned was matched across instructional sets. During the 5-min instruction period, participants were also asked to record the frequency at which they experienced thoughts related to the video by placing a tick on a sheet of paper every time they experienced these intrusions. At the end of the 5 min, another mood rating was obtained (Mood Rating 3), and manipulation checks were obtained. Participants then completed the AMT whilst still engaging in suppression, emotional inhibition, or normal responding. The manipulation checks were re-administered, and a final mood rating was obtained (Mood Rating 4). The experiment then ended with the participants debriefed and thanked. 969 S.Q. Phung, R.A. Bryant / Consciousness and Cognition 22 (2013) 965–974 7. Results 7.1. Participant characteristics Table 1 presents the participants’ characteristics. Participants in the three conditions did not differ in terms of age [F(2, 48) = 0.28, p = .76] or scores on the BDI-II [F(2, 48) = 0.31, p = .74], BAI [F(2, 48) = 0.28, p = .76], TCQ Distraction [F(2, 48) = .58, p = .56, TCQ Social [F(2, 48) = 1.01 p = .37], TCQ Worry [F(2, 48) = 1.55, p = .22], TCQ Punishment [F(2, 48) = .03, p = .97], TCQ Reappraisal [F(2, 48) = .59, p = .56], or WBI [F(2, 48) = 2.84, p = .09]. This suggests that participants in the three conditions did not differ in terms of depression, anxiety, or tendency to suppress. 7.2. Preliminary analyses To assess the effects of viewing the video on participants’ mood, a 3 (Instructional Set)  4 (Mood Rating) repeated measures analyses of variance (ANOVA) was conducted on the four mood ratings (see Table 2). This indicated significant main effects for Mood Rating, F(3, 46) = 32.11, p < .001, and Instructional Set, F(2, 48) = 3.17, p < .05. Participants rated more distress immediately after seeing the movie relative to baseline (p < .001), after the induction (p < .001), and at the completion of the study (p < .001). Overall, participants in the Control condition reported more distress than those in the other two conditions. Additionally, participants rated less distress at the completion of the study than at baseline (p < .001) and after the induction (p < .001). There was no Instructional Set  Mood Rating interaction [F(6, 94) = 0.84, p = .54], indicating that the induction instructions did not differentially affect mood ratings. To index the extent to which participants complied with the experimental instructions pertaining to thought suppression, emotion inhibition, and control instructions, a 3 (Instructional Set)  2 (Manipulation Rating) repeated measures ANOVA was conducted on the two manipulation ratings (see Table 2). This indicated a significant main effect for Manipulation Rating, F(1, 48) = 50.77, p < .001. Specifically, participants across conditions rated their compliance with the initial instructions weaker following the AMT than immediately following the instruction period. 7.2.1. Effect of regulation strategy on autobiographical memory retrieval Four separate 3 (Instructional Set)  3 (Cue Valence) repeated measures ANOVAs were conducted on specific, categoric, extended, and omitted memories, respectively (see Table 3). In terms of specific memories, there was a marginal main effect for Instructional Set [F(2, 48) = 2.83, p = .08, g2 = .10] but not Cue Valence or the interaction. Specifically, there was a trend for participants in the Control condition to retrieve fewer specific memories than those in the other two conditions. In terms of categoric memories, there were main effects for Cue Valence [F(2, 47) = 32.10, p < .001, g2 = .58] and Instructional Set [F(2, 48) = 4.97, p < .01, g2 = .17] but not the interaction. Categoric memories were retrieved less frequently to positive cues than to neutral (p < .05) and negative (p < .05) cues. Categoric memories were retrieved more frequently in the control condition than in the emotional suppression (p < .01) and thought suppression (p < .05) conditions. In terms of extended memories, there were main effects for Cue Valence [F(2, 47) = 14.45, p < .001, g2 = .38] but not Induction Conduction or the interaction. Extended memories were retrieved less frequently to negative cues than to neutral (p < .01) and positive (p < .01) cues. In terms of omitted memories, there was a main effect for Cue Valence [F(2, 47) = 10.76, p < .001, g2 = .31] but not Instructional Set. There were fewer omissions to negative cues than to neutral (p < .01) and positive (p < .01) cues. To index the contribution of the instructional set relative to other known major predictors of memory accessibility, we conducted separate linear regressions to predict categoric retrieval, respectively. BDI-II scores were entered at Step 1 because of evidence that depression is associated with categoric retrieval (Williams et al., 2007), WBSI score was entered at Step 2 because of the relationship between tendency to suppress and accessibility of suppressed thoughts (Wenzlaff & Wegner, 2000), gender was entered at Step 3 because of evidence that women can recall emotional memories more strongly than men (Andreano & Cahill, 2009), and instructional set was entered at Step 4. Table 4 presents the summaries of the regression Table 1 Mean scores for participant characteristics. Age (years) BDI-II BAI WBSI TCQ Distraction TCQ Social TCQ Worry TCQ Punishment TCQ Reappraisal Thought suppression Emotion inhibition Control 19.78 (0.98) 13.65 (5.74) 13.94 (8.55) 49.18 (4.79) 15.06 (2.48) 16.82 (3.81) 11.76 (2.39) 10.24 (2.33) 14.82 (2.88) 19.33 (5.84) 14.31 (5.86) 12.31 (6.04) 44.19 (7.73) 15.50 (2.20) 15.63 (3.44) 10.63 (2.47) 10.06 (1.88) 14.56 (3.52) 18.68 (4.79) 12.72 (6.13) 12.33 (7.04) 49.56 (8.52) 15.89 (2.11) 15.28 (2.76) 12.17 (2.94) 10.06 (2.71) 15.61 (2.48) Note: Standard deviations appear in parentheses. BDI-II = Beck Depression Inventory – Second Edition. BAI = Beck Anxiety Inventory. WBSI = White Bear Suppression Inventory. TCQ = Thought Control Questionnaire. 970 S.Q. Phung, R.A. Bryant / Consciousness and Cognition 22 (2013) 965–974 Table 2 Mean mood ratings, manipulation checks, and intrusions. Thought suppression Emotion inhibition Control Mood ratings Baseline Post-Film Post-Induction Study Completion 42.89 (21.05) 67.12 (19.80) 42.59 (15.40) 25.00 (15.71) 45.75 (24.13) 66.25 (22.84) 51.56 (20.80) 33.75 (23.91) 39.06 (24.70) 55.00 (23.39) 39.72 (20.40) 27.22 (16.47) Manipulation rating Post-Instruction Post-AMT Intrusions 57.65 (21.58) 16.47 (27.60) 6.00 (3.68) 46.88 (31.19) 22.19 (28.34) 4.12 (4.34) 53.89 (30.17) 25.72 (25.83) 6.68 (6.31) Note: Standard deviations appear in parentheses. Table 3 Mean Number of specific, categoric, extended, and omitted autobiographical memories. Thought suppression Emotion suppression Control Specific Positive Neutral Negative 1.49 (.88) 1.47 (.81) 1.28 (1.20) 1.69 (1.06) 1.45 (1.15) 1.06 (1.00) .80 (.73) 1.33 (.97) .67 (.75) Categoric Positive Neutral Negative .00 (.00) .06 (.24) 2.24 (1.71) .19 (.75) .25 (.77) 2.50 (1.90) 1.28 (1.77) 1.06 (1.47) 2.94 (1.70) Extended Positive Neutral Negative 2.61 (1.18) 2.63 (1.66) 1.45 (1.33) 2.31 (1.15) 2.37 (1.60) .94 (1.29) 1.79 (1.33) 1.89 (1.08) .94 (.97) Omitted Positive Neutral Negative .90 (1.24) .75 (.87) .03 (.00) .81 (1.31) .93 (1.37) .50 (1.32) 1.13 (1.31) .72 (.83) .45 (.85) Note: Standard deviations appear in parentheses. Table 4 Summary of hierarchical regression model for retrieval of specific and categoric memories. B SEB p b Specific memories Step 1 Step 2 Step 3 Step 4 BDI-II WBSI Gender Instructional set .02 .06 .19 1.03 .05 .04 .52 .34 .07 .24 .05 .44 .63 .10 .72 .004 Categoric memories Step 1 Step 2 Step 3 Step 4 BDI-II WBSI Gender Instructional set .03 .03 .99 1.20 .08 .06 .89 .57 .05 .06 .16 .31 .36 .69 .27 .04 Note: Specific memories: Step 1 R2 = .00, DR2 = .00. Step 2 R2 = .00, DR2 = .00. Step 3 R2 = .00, DR2 = .00. Step 4 R2 = .14, DR2 = .13. Categoric memories: Step 1 R2 = .00, DR2 = .00. Step 2 R2 = .00, DR2 = .00. Step 3 R2 = .04, DR2 = .03. Step 4 R2 = .12, DR2 = .09. BDI-II = Beck Depression Inventory. WBSI = White Bear Suppression Inventory. models. The only significant predictor of both specific and categoric memories was the instructional set (accounting for 13% and 9% of the variance, respectively); that is, receiving an instruction to suppress thoughts or inhibit emotions resulted in more specific and fewer categoric memories. Regarding reported intrusive memories of the initial stimulus (see Table 2), a oneway ANOVA indicated no difference between intrusions between conditions, (see Table 2), F(2, 48) = 1.51, p = .23. S.Q. Phung, R.A. Bryant / Consciousness and Cognition 22 (2013) 965–974 971 8. Discussion 8.1. Thought suppression and autobiographical memory retrieval The major finding of this study is that both thought suppression and emotional inhibition reduced categoric memories relative to the control condition. This finding may seem at odds with the CarFAX model’s (Williams et al., 2007) proposal that cognitive avoidance impedes retrieval of specific retrieval, and may contribute to categoric recall. We also note that our finding is inconsistent with one previous study that did not find that suppression influenced the specificity of autobiographical memories (Geraerts, Hauer, & Wessel, 2010). The interpretation of these findings requires the acknowledgment that avoidance is a multifaceted construct that entails numerous different strategies, and that there is a need to deconstruct the avoidance construct into specific strategies, and under controlled conditions, to determine how regulatory processes aimed at minimizing distress may influence the specificity of retrieving both positive and negative memories. As alluded to by Williams et al. (1997), functional avoidance in the CarFAX model may involve controlled or more habitual strategies, and it is possible that different functional avoidance processes result in distinct outcomes. Reinforcing this possibility is recent evidence from two studies that used a Think/No-Think paradigm for cues associated with autobiographical memories and found that suppression diminished specific retrieval of autobiographical memories (Noreen & MacLeod, 2013; Stephens, Braid, & Hertel, 2013). It is likely that training participants to forget cues associated with personal memories is not identical to the direct instruction in this experiment to fully suppress any awareness of the memory itself. The effect of suppression on increasing specific and reducing categoric retrieval can be understood in terms of the paradoxical nature of suppression. According to ironic process theory (Wegner, 1994), when participants were suppressing thoughts or emotions associated with the stimulus video whilst performing the AMT, the monitoring process may have scanned for potential triggers of the video, and in so doing increased sensitivity of these memories, thereby making them less categoric (Rassin et al., 1997). This interpretation is consistent with Dalgleish and Yiend’s (2006) and Neufeind et al.’s (2009) findings that suppression led to speedier access to other negative memories, prompting the suggestion that suppression of a negative experience may cause any negative memory to be more accessible (Neufeind et al., 2009). This explanation is complicated, however, by the finding that attempted suppression of thoughts or emotions resulted in fewer categoric memories in response to neutral as well as negative cues. Several explanations may account for this. First, participants’ ratings indicated that watching the video caused participants to feel distressed, which may have resulted in a tendency to retrieve negative memories to all cues; this mood-congruent has been reported repeatedly in previous studies (Eich, 1995; Williams et al., 1997). That is, positive and neutral cues may have triggered negative associations because of the distressed mood at the time of retrieval. This tendency to retrieve negative memories to all cues may explain the decreased categoric retrieval to all cues in both suppression and inhibition conditions because both involved sensitivity to the suppressed negative stimuli. Second, as there were many stimuli in the complex scenes of the video, it is possible that suppressing memories or emotions of these scenes involved suppression of a very wide range of stimuli that may have involved associations of both the neutral and positive cue words. Third, suppressing or emotionally inhibiting the video content during the retrieval of autobiographical memories may have increased specificity of retrieval because the detailed nature of the information in the film may have primed specific-level processing, as well as providing a distraction for the participants attempting to suppress the film-related material; these factors may have contributed to fewer categoric memories because of the salience of the film content. In this sense it is worth noting that ironic control processes imply activation of specific levels of memory processing because the monitoring process is constantly seeking the occurrence of the suppressed memories, and this can lead to more awareness of specific, rather than categoric, memories. Finally, we note that suppression did not lead to different intrusion frequency than emotion inhibition or control instructions, which is contrary to the prediction of ironic control theory. It is possible that the suppression instructions did not have the desired effect in terms of contributing to subsequent intrusions (Wenzlaff & Wegner, 2000). 8.2. Emotional inhibition and categoric memory retrieval Interestingly, emotional inhibition was also associated with the retrieval fewer categoric autobiographical memories. This finding needs to be considered in relation to Richards and Gross’ (2000) finding that inhibiting emotional expression impeded memory for events. These authors posited that because this strategy is a response-focused technique that is employed after registration of the emotional material, it is cognitively demanding and results in reduced cognitive resources to consolidate associated memories. However, several differences exist between the current study and Richards and Gross’ (2000) paradigm. First, whereas the current study required participants to inhibit emotional experience, Richards and Gross (2000) instructed participants to suppress emotional expression. Suppressing emotional expression and experience may involve different processes, varying degrees of encoding of the emotional event, and variable effort (Hayes, Wilson, Gifford, Follette, & Strosahl, 1996). It is possible that inhibiting emotional experience, rather than expression, resulted in a more antecedent strategy than suppressing expression, and hence fewer cognitive resources were required. Because we did not index cognitive demands in this experiment, future studies could employ dual-task paradigms to assess the extent to which different regulation strategies are resource demanding. Second, whereas we studied the impact of emotional inhibition on the retrieval of autobiographical memories, Richards and Gross (2000) studied memory for events encoded at the time of 972 S.Q. Phung, R.A. Bryant / Consciousness and Cognition 22 (2013) 965–974 suppression. The impact of suppressing expression on encoding concurrent information may be more marked than its impact on retrieving previously encoded autobiographical memories. 8.3. Cue valence and autobiographical memory retrieval Across conditions, participants retrieved fewer categoric and extended memories to positive than other cue words. This finding is consistent with some (e.g. Williams & Dritschel, 1988) but not other (e.g. Brittlebank, Scott, Williams, & Ferrier, 1993) studies which have varied in methodologies and populations. This valence effect may be attributed to mood-congruent recall. Because the video scenario elicited distress in participants, and their ratings indicated that this distress persisted to the AMT, this could have led to a relative deficit in retrieving personal memories to positive cues than to negative cues. We also note that neutral words may have been easier to retrieve because in contrast to the abstract nature of the positive (e.g. peace) and negative (e.g. urgency) cues, neutral cues were objects (e.g. boat). Indeed, cues that are more imageable are more readily retrieved than abstract cues (Hauer, Wessel, Geraerts, Merckelbach, & Dalgleish, 2008), which may account for the differential retrieval to positive cues relative to negative and neutral cues. Relative to the greater focus on the rumination and executive control components of the CarFAX model, this is one of the few studies to address the affect regulation aspect of the model in an experimental framework. This study employed a novel regulation manipulation that extended beyond a single avoidance strategy. However, we also note a number of methodological restrictions. First, this study has a relatively small sample size which might have restricted its power. Second, our instructions to inhibit or suppress may not mimic avoidance strategies used in more naturalistic settings. According to Wenzlaff and Wegner (2000), in the real world suppression is typically self-initiated. This discrepancy restricts the generalizability of the effects found in the laboratory. Accordingly, they recommended the adoption of a paradigm that would encourage spontaneous suppression such as priming participants to death-related thoughts that would naturally evoke suppression (Arndt, Greenberg, Solomon, Pyszczynski, & Simon, 1997). This issue is related to the extent to which avoidance is passively or actively implemented in the two theories tested here, and we recognize that it is possible that the active nature of the suppression/inhibition paradigms may not be the optimal means to test functional avoidance in the CarFAX model. Third, this study focused on emotional inhibition and thought suppression at the exclusion of other strategies with demonstrated efficacy. As such, future studies could examine the effects of other regulation strategies, such as reappraisal, on the retrieval of specific autobiographical memories (Richards & Gross, 2000). Fourth, the AMT may not be the optimal means to assess the accessibility of autobiographical memories in non-clinical populations, remembering that it was developed primarily to assess memory patterns in depressed and suicidal groups (Raes, Hermans, Williams, & Eelen, 2007). The AMT leads to low levels of categoric retrieval in non-clinical populations (Raes et al., 2003), and the reported associations between categoric memories and depressive and ruminative tendencies that are often reported in clinical groups (Watkins & Teasdale, 2001; Watkins & Teasdale, 2004) have not been found in some non-clinical studies (see Griffith et al., 2012). It has been suggested that the instructional set of the AMT, along with the practice trials, results in a test that is not sensitive to indexing categoric retrieval in non-clinical participants (Raes et al., 2007). Conducting a more implicit test of memory accessibility, such as a sentence-completion task (Raes et al., 2007) may have led to a different pattern of findings. This approach may also overcome a potential problem of participants remembering to focus on both tasks simultaneously, which we did not index in this study. The finding that compliance with the suppression/inhibition instructions diminished during the AMT suggests that participants may not have been focused on both tasks. Providing a variant of the AMT that contains minimal instructions that do not explicitly require specific memories can provide a more sensitive measure of retrieval (Debeer, Hermans, & Raes, 2009). We also note a distinctive trend in this study in which specific memories were close to floor effects. Compared to other studies of non-clinical participants (Sutherland & Bryant, 2007), the participants in this study reported much fewer specific memories than would normally be expected. This finding cannot be attributed to distinctive instructions, timing allowed for provision of memories, or nature of cue words, which have been offered as methodological reasons for variations in findings across studies (Griffith et al., 2009). One possible explanation for this tendency is the use of a student population in Hong Kong, for whom English is a second language; the study was conducted in English, and it is possible that either the instructions were not sufficiently understood or the linguistic capabilities of participants hindered their capacity to articulate specific memories. In either case, it suggests that this finding requires replication with participants who complete the task in their first language. 8.4. CarFAX and affect regulation Collectively, the current findings questions the proposal of the CarFAX model that attempted avoidance or inhibition results in increased categoric retrieval, presumably because one is aborting the retrieval search prior to locating memories that are aversive (Williams et al., 2007). The finding that both forms of suppression did not affect autobiographical memory specificity, and more importantly decreased categoric retrieval, questions the direct effect of avoidance. We note several caveats to this conclusion. First, this study did not focus on participants with clinical levels of anxiety or depression. It is possible that their motivation to avoid personal memories may be greater than what could be achieved experimentally in this study, and hence avoidance may play a differential role in individuals who are retrieving very distressing memories. Second, we note that avoidance may function in concert with rumination and depleted cognitive resources. As we did not index these other S.Q. Phung, R.A. Bryant / Consciousness and Cognition 22 (2013) 965–974 973 factors, we are unable to infer the extent to which the interaction of these factors contributes to retrieval patterns. Taken together, this finding underscores the need for more detailed research on the affect regulation component of the CarFAX model. References Abramowitz, J. A., Tolin, D. F., & Street, G. P. (2001). Paradoxical effects of thought suppression: A meta-analysis of controlled studies. Clinical Psychology Review, 21, 683–703. Andreano, J. M., & Cahill, L. (2009). Sex influences on the neurobiology of learning and memory. Learning and Memory, 16, 248–266. Arndt, J., Greenberg, J., Solomon, S., Pyszczynski, T., & Simon, L. (1997). Suppression accessibility of death-related thoughts, and cultural worldview defense: Exploring the psychodynamics of terror management. Journal of Personality and Social Psychology, 73(1), 5–18. Baumeister, R. F., Bratslavsky, E., Muraven, M., & Tice, D. M. (1998). Ego depletion: Is the self a limited resource? Journal of Personality and Social Psychology, 74, 1252–1265. Beck, A. T., & Steer, R. A. (1990). Manual for the beck anxiety inventory. San Antonio, TX: The Psychological Corporation. Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Beck depression inventory manual (2nd ed.). San Antonio, TX: The Psychological Corporation. Brittlebank, A. D., Scott, J., Williams, J. M., & Ferrier, I. N. (1993). Autobiographical memory in depression: State or trait marker? British Journal of Psychiatry, 162, 118–121. Bryant, R. A., & Kapur, A. (2006). Hypnotically induced emotional numbing: The roles of hypnosis and hypnotizability. International Journal of Clinical and Experimental Hypnosis, 54(3), 281–291. Carroll, J. B., Davies, P., & Richman, B. (1971). Word frequency book. New York: American Heritage. Dalgleish, T., Tchanturia, K., Serpell, L., Hems, S., Yiend, J., de Silva, P., et al (2003). Self-reported parental abuse relates to autobiographical memory style in patients with eating disorders. Emotion, 3(3), 211–222. Dalgleish, T., Williams, J. M. G., Golden, A. J., Perkins, N., Barrett, L. F., Barnard, P. J., et al (2007). Reduced specificity of autobiographical memory and depression: The role of executive control. Journal of Experimental Psychology: General, 136(1), 23–42. Dalgleish, T., & Yiend, J. (2006). The effects of suppressing a negative autobiographical memory on concurrent intrusions and subsequent autobiographical recall in dysphoria. Journal of Abnormal Psychology, 115, 467–473. Davies, M. I., & Clark, D. M. (1998). Thought suppression produces a rebound effect with analogue post-traumatic intrusions. Behaviour Research and Therapy, 36(6), 571–582. Debeer, E., Hermans, D., & Raes, F. (2009). Associations between components of rumination and autobiographical memory specificity as measured by a Minimal Instructions Autobiographical Memory Test. Memory, 17, 892–903. Eich, E. (1995). Searching for mood-dependent memory. Psychological Science, 6, 67–75. Evans, J., Williams, J. M. G., O’Loughlin, S., & Howells, K. (1992). Autobiographical memory and problem-solving strategies of parasuicide patients. Psychological Medicine, 22, 399–405. Geraerts, E., Hauer, B. J. A., & Wessel, I. (2010). Effects of suppressing negative memories on intrusions and autobiographical memory specificity. Applied Cognitive Psychology, 24, 387–398. Golden, A., Dalgleish, T., & Mackintosh, B. (2007). Levels of specificity of autobiographical memories and of biographical memories of the deceased in bereaved individuals with and without complicated grief. Journal of Abnormal Psychology, 116(4), 786–795. Griffith, J. W., Sumner, J. A., Debeer, E., Raes, F., Hermans, D., Mineka, S., Zinbarg, R. E., & Craske, M. G. (2009). An item response theory/confirmatory factor analysis of the Autobiographical Memory Test. Memory, 17(6), 609–623. Griffith, J. W., Sumner, J. A., Raes, F., Barnhofer, T., Debeer, E., & Hermans, D. (2012). Current psychometric and methodological issues in the measurement of overgeneral autobiographical memory. Journal of Behavior Therapy and Experimental Psychiatry, 43(Suppl 1), S21–31. Gross, J. J. (1998). The emerging field of emotion regulation: An integrative review. Review of General Psychology, 2, 271–299. Harvey, A. G., & Bryant, R. A. (1998). The effect of attempted thought suppression in acute stress disorder. Behaviour Research and Therapy, 36(6), 583–590. Harvey, A. G., Bryant, R. A., & Dang, S. T. (1998). Autobiographical memory in acute stress disorder. Journal of Consulting and Clinical Psychology, 66, 500–506. Hauer, B. J. A., Wessel, I., Geraerts, E., Merckelbach, H., & Dalgleish, T. (2008). Autobiographical memory specificity after manipulating retrieval cues in adults reporting childhood sexual abuse. Journal of Abnormal Psychology, 117, 444–453. Hayes, S. C., Wilson, K. G., Gifford, E. V., Follette, V. M., & Strosahl, K. (1996). Experiential avoidance and behavioral disorders: A functional dimensional approach to diagnosis and treatment. Journal of Consulting and Clinical Psychology, 64, 1152–1168. Hermans, D., Defranc, A., Raes, F., Williams, J. M. G., & Eelen, P. (2005). Reduced autobiographical memory specificity as an avoidant coping style. British Journal of Clinical Psychology, 44, 583–589. Jones, B., Heard, H., Startup, M., Swales, M., Williams, J. M. G., & Jones, R. S. P. (1999). Autobiographical memory and dissociation in borderline personality disorder. Psychological Medicine, 29, 1397–1404. Kelly, A. E., & Kahn, J. H. (1994). Effects of suppression of personal intrusive thoughts. Journal of Personality and Social Psychology, 66, 998–1006. Muris, P., Merckelbach, H., & de Jong, P. (1993). Verbalization and environmental cuing in thought suppression. Behaviour Research and Theray, 31(6), 609–612. Neufeind, J., Dritschel, B., Astell, A. J., & MacLeod, M. D. (2009). The effects of thought suppression on autobiographical memory recall. Behaviour Research and Therapy, 47, 275–284. Noreen, S., & MacLeod, M. D. (2013). It’s all in the detail: Intentional forgetting of autobiographical memories using the autobiographical think/no-think task. Journal of Experimental Psychology: Learning, Memory, and Cognition, 39, 375–393. Raes, F., Hermans, D., de Decker, A., Eelen, P., & Williams, J. M. (2003). Autobiographical memory specificity and affect regulation: An experimental approach. Emotion, 3(2), 201–206. Raes, F., Hermans, D., Williams, J. M., & Eelen, P. (2007). A sentence completion procedure as an alternative to the Autobiographical Memory Test for assessing overgeneral memory in non-clinical populations. Memory, 15(5), 495–507. Rassin, E., Merckelbach, H., & Muris, P. (1997). Effects of thought suppression on episodic memory. Behaviour Research and Therapy, 35(11), 1035–1038. Richards, J. M., & Gross, J. J. (1999). Composure at any cost? The cognitive consequences of emotion suppression. Personality and Social Psychology Bulletin, 25(8), 1033–1044. Richards, J. M., & Gross, J. J. (2000). Emotion regulation and memory: The cognitive costs of keeping one’s cool. Journal of Personality and Social Psychology, 79(3), 410–424. Roemer, L., & Borkovec, T. D. (1994). Effects of suppressing thoughts about emotional material. Journal of Abnormal Psychology, 103(3), 467–474. Stephens, E., Braid, A., & Hertel, P. T. (2013). Suppression-induced reduction in specificity of autobiographical memories. Clinical Psychological Science, 1, 163–169. Sutherland, K., & Bryant, R. A. (2007). Rumination and overgeneral autobiographical memory. Behaviour Research and Therapy, 45(10), 2407–2416. Watkins, E., & Teasdale, J. D. (2001). Rumination and overgeneral memory in depression: Effects of self-focus and analytic thinking. Journal of Abnormal Psychology, 110, 353–357. Watkins, E., & Teasdale, J. D. (2004). Adaptive and maladaptive self-focus in depression. Journal of Affective Disorders, 82(1), 1–8. Wegner, D. (1994). Ironic process of mental control. Psychological Review, 101, 34–52. Wegner, D. M., Erber, R., & Zanakos, S. (1993). Ironic processes in the mental control of mood and mood-related thought. Journal of Personality and Social Psychology, 65(6), 1093–1104. 974 S.Q. Phung, R.A. Bryant / Consciousness and Cognition 22 (2013) 965–974 Wegner, D. M., Schneider, D. J., Carter, S. R., & White, T. L. (1987). Paradoxical effects of thought suppression. Journal of Personality and Social Psychology, 53(1), 5–13. Wegner, D. M., & Zanakos, S. (1994). Chronic thought suppression. Journal of Personality, 62, 615–640. Wells, A., & Davies, M. (1994). The tought control questionnaire: A measure of individual differences in the control of unwanted thoughts. Behaviour Research and Therapy, 32, 871–878. Wenzlaff, R. M., & Wegner, D. M. (2000). Thought suppression. Annual Review of Psychology, 51, 59–91. Wilhelm, S., McNally, R. J., Baer, L., & Florin, I. (1997). Autobiographical memory in obsessive–compulsive disorder. British Journal of Clinical Psychology, 36, 21–31. Williams, J. M. G., Barnhofer, T., Crane, C., Herman, D., Raes, F., Watkins, E., et al (2007). Autobiographical memory specificity and emotional disorder. Psychological Bulletin, 133(1), 122–148. Williams, J. M., & Broadbent, K. (1986). Autobiographical memory in attempted suicide patients. Journal of Abnormal Psychology, 95, 144–149. Williams, J. M. G., & Dritschel, B. H. (1988). Emotional disturbance and the specificity of autobiographical memory. Cognition & Emotion, 2, 221–234. Williams, J. M. G., Ellis, N. C., Tyers, C., Healy, H., Rose, G., & MacLeod, A. K. (1996). The specificity of autobiographical memory and imageability of the future. Memory & Cognition, 24, 116–125. Williams, J. M. G., Watts, F. N., MacLeod, C., & Mathews, A. (1997). Cognitive psychology and emotional disorders (2nd ed.). Chichester, England: Wiley.

Consciousness and Cognition 22 (2013) 1105–1113 Contents lists available at SciVerse ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Self in the mirror Wolfgang Prinz ⇑ Max Planck Institute for Human Cognitive and Brain Sciences, Department of Psychology, Stephanstrasse 1 a, 04103 Leipzig, Germany a r t i c l e i n f o Article history: Available online 12 February 2013 Keywords: Mirror systems Mental self Social mirroring Subjectivity a b s t r a c t What are mirror systems good for? Several suggestions have been made in response to this question, addressing the putative functions of mirror systems in minds and brains. This paper examines possible contributions of mirror systems to the emergence of subjectivity. At the heart of the discussion is the notion of social mirroring, which has a long tradition in social philosophy and social anthropology. Taking the existence of mirror devices in minds and brains for granted, I argue that social mirroring is a prerequisite for the constitution of mental selves, and, hence, the emergence of subjectivity. However, the fact that self and subjectivity are socially created should not be taken to indicate that they are illusory. They are as real as natural facts are. Ó 2013 Elsevier Inc. All rights reserved. 1. The looking glass self Mirrors have two faces. In one sense, they are innocent physical things—nothing but polished surfaces reflecting light according to simple geometrical rules. Yet, in another sense, when used by human observers, they are powerful cultural things. This is because mirrors help people to extend the reach of what they can see. Mirrors help people to look around corners and see what is happening behind their backs—as well as allowing them to look at themselves and check their outer appearance. Probably by virtue of this remarkable capacity, in many cultures mirrors have made their career as both technical devices for exploring the outer world and symbolic devices for exploring the inner self. Not only do they function as technical tools for checking one’s outer appearance, but also as symbolic instruments for deeper ways of reflecting one’s inner self. Such symbolic use of mirrors is widespread in Western art and literature. For instance, in the act of portraying oneself, mirrors are often thought to reflect aspects of the artist’s inner self through his or her outer appearance. Occasionally, the act of self-mirroring may even be included in the self-portrait as a symbolic indication of the reflective and reflexive intentions entailed in that act. For instance, the Uffizi Gallery harbors a beautiful painting by the Austrian painter Johannes Gumpp, which shows the painter’s image in a mirror, as well as the painting that he is painting of that image (Gumpp, 1646). In a similar vein, the act of detecting and recognizing oneself in the mirror has become a classical literary topic. Perhaps one of the most dramatic incidences of mirror self-recognition is reported in the myth of Narcissus. According to the testimony of Ovid’s Metamorphoses, Narcissus was bound to die at the very moment at which he discovered that the beautiful face in the mirror that he had fallen in love with did not belong to some beloved other but, rather, to himself (Ovid., 1977, p. 157). In more recent times, the French psychoanalyst Jaques Lacan revitalized the myth of mirror self-recognition. Lacan believed that the very first incidence of mirror self-recognition is a dramatic ‘‘Aha! experience’’ for the young infant—an act of utmost importance in the formation of the human mind and self-consciousness (Amsterdam, 1972; Lacan, 1949/1977). Yet, in the meantime, we have learned that mirror self-recognition cannot be considered a specific signature of human mentality. ⇑ Fax: +49 (0)341 9940 2330. E-mail address: prinz@cbs.mpg.de 1053-8100/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.concog.2013.01.007 1106 W. Prinz / Consciousness and Cognition 22 (2013) 1105–1113 Chimps do it, elephants do it, crows do it, and so do many other species (e.g., Gallup, 1970). Taken together, these observations seem to suggest that mirror self-recognition has perhaps often more to say about these animals’ understanding of mirrors than of themselves. In the context of self-recognition and self-reflection, one also encounters a metaphorical use that applies to social rather than physical mirrors. In social mirroring other individuals serve as mirrors for the self. This notion has, more or less independently, been promoted by a number of classical authorities from social philosophy and social science like, e.g., Cooley (1902), Hegel (1807), Mead (1934), Smith (1759/1976), and Whitehead (2001). The crucial idea here is that individuals come to understand themselves through mirroring themselves in others—that is, by learning to understand how their conduct is perceived, received and understood by others. To address this notion, Cooley (1902) has coined the term ‘‘looking glass self’’. What this suggests is that, for individuals, social mirrors can play a role similar to that of physical mirrors: Both help them to ‘‘see’’ how others perceive them. One may wonder what lies behind the widespread use of mirrors as symbols and metaphors for reflecting the self and reflecting upon the self. What can looking at our outer appearance in the mirror possibly add to what we already know from our inner experience? Do not we have direct, unmediated access to that experience? Obviously, if this were true, mirrors would provide redundant information: they would just replicate knowledge that we already have. In that case, it would be hard to understand how mirrors can have made such a fantastic career as cultural symbols for seeing one’s mind through one’s body. What that career seems to reflect, instead, is that perceiving oneself from the outside may often deliver a clearer and more adequate picture of our true feelings and actions than the inside perspective can provide. This paper examines the looking glass self from a cognitive science perspective. While the classical discussion of this idea in social philosophy and social science is mainly grounded in language as a symbolic mirror between others and the self, a novel view has recently emerged from several branches of cognitive science. This view claims that social mirroring may, likewise, be grounded in actions that serve as embodied mirrors between the self and others. In fact, one may even claim that social mirroring first relies on embodied acting (i.e., the ways in which individuals interact with each other), and only later shifts to symbolic talking (i.e., the ways in which individuals communicate about their acting); e.g., Prinz (2008, 2012, 2013). Adopting this view, we still need to ask how social mirroring works and what it does to individuals. Is there anything serious behind the widespread symbolic and metaphorical use of mirrors for self-recognition and self-reflection? What does it mean for individuals to perceive themselves in ways they otherwise cannot and what can they make of it? Here, I argue that mirrors may indeed play an important role in the formation of our mental selves—provided that the mirrors outside are met by mirrors inside. By mirrors outside I refer to social and physical mirrors that individuals encounter in their environments. By mirrors inside I refer to mirror-like representational devices operating inside their minds. I propose that these two kinds of mirrors interact with each other in ways that give rise to the formation of our mental selves. Over the past decades literature on social mirroring has been scarce, at least as far as solid empirical research is concerned. Accordingly, the framework I am sketching here is not meant to cover and integrate extant empirical evidence. It is rather meant to provide a heuristic basis for guiding future experimental research into mechanisms and practices of social mirroring. 2. Social mirroring To understand social mirroring we need to consider two perspectives, that of the target individual whose acting is being mirrored and that of the mirror individual who is mirroring the target’s acting. For the target individual, the mirror individual provides a living mirror in his or her environment. In what ways can the target individual find his or her own action mirrored through the mirror individual’s action? To answer this question it may be useful to draw two distinctions, one between two basic modes of mirroring and another one between two basic modes of communication. 2.1. Modes of mirroring We may discern two basic modes of mirroring: reciprocal and complementary. In the most fundamental form of social mirroring, reciprocal mirroring, target individual ‘‘T’’ sees her own action imitated, or replicated by mirror individual ‘‘M’’. In a setting like this, M acts as a mirror for T in a more or less literal sense. Social mirrors are, of course, fundamentally different from physical mirrors. Even if M attempts to provide as-perfect-as-possible copies of T’s acting, those copies will always be delayed in time and their kinematics will never be as perfectly correlated with T’s acting as specular images are. We can speak of reciprocal mirroring as long as T is in a position to recognize and understand M’s acting as a delayed copy of her own preceding action. Hence, the constitutive feature of reciprocal mirroring is T’s understanding of M’s action as a copy of T’s own foregoing action. A slightly different form of social mirroring, complementary mirroring, arises when T sees her own action continued and carried on by M rather than simply being replicated. This is, of course, entirely different from what physical mirrors do. Still, what complementary mirroring has in common with reciprocal mirroring is that (1) M’s action is strongly contingent upon T’s preceding action and (2) that contingency needs to be perceived and understood by T. In this case, complementary mirroring requires that T is in a position to assess M’s doing as a continuation of his or her own doing. W. Prinz / Consciousness and Cognition 22 (2013) 1105–1113 1107 2.2. Modes of communication The scenario considered so far draws on what we may call mirroring through embodied communication. It starts with T acting in a particular way; then M, upon perceiving T’s acting, starts replicating or continuing that action, and eventually that replication or complement is perceived and understood by T. Here, communication is embodied in the sense that it relies on T’s and M’s competence for both production of own action and perception of foreign action. Such embodied mirroring does not require a language system in which the two communicate. It does not even require explicit intentions to communicate something to someone else on either side. The sole requirement is that competent perceivers/actors meet, act and watch each other. Routines for embodied mirroring play an important role in interactions between young infants and their caretakers. Babies and their mothers will often find themselves involved in what has been called ‘‘protoconversational interactions’’ (Trevarthen, 1993, 1998; Trevarthen, Kokkinaki, & Fiamenghi, 1999). They imitate or continue each other’s action and take turns in this funny game from time to time. Though these interactions have been extensively studied, most studies focus on the baby’s production of imitative action but not on their perception of their mother’s imitative action. In other words, most of work in the field views the baby in the role of individual M (who mirrors mother’s actions), but not in the role of individual T (who perceives herself as being mirrored by mother). This, however, is the perspective that we need to adopt if we want to understand how social mirroring contributes to the formation of the self (cf., e.g., Meltzoff, 1990; Nadel, 2002; Paukner, Anderson, Borelli, Visalberghi, & Ferrari, 2005; Trevarthen, 1993; Zukow-Goldring, 2006). Perhaps more familiar to us as adults is action mirroring through symbolic communication. T acts in a particular way, and M, upon perceiving T’s acting, starts talking about T’s acting, and that verbal account is finally perceived and understood by T as referring to her preceding acting. In a setting like this, M’s verbal account of T’s acting can vary not only along the dimension replication/continuation but also along the dimension description/explanation/evaluation. In any case, such symbolic mirroring is dependent on the two individuals’ competences for the production and perception of spoken language. M communicates to T a verbal message concerning T’s action, and that message is then decoded and understood by T. However, competences for production and perception of spoken language are not enough for symbolic mirroring to work. More importantly, the two individuals need to share a conceptual framework for action. They need to draw on a shared action ontology, which specifies for them what actions are, how they can be parsed and individuated, and how physical action can be explained through foregoing mental action. This is precisely what their folk psychology provides—a common-sense framework for action description and explanation to which they resort when they reflect and communicate about what people are doing and why they do what they do (cf., e.g., Bogdan, 1991; Greenwood, 1991; Kusch, 1999; Malle, 2004; Malle, Moses, & Baldwin, 2001). 3. Mirrors inside Let us now see what kinds of representational structures and processing mechanisms target individual T requires in order to be in a position to capitalize on M’s mirroring, to build up a representation of self. Evidently, the mere fact of being mirrored from the outside will not do the job by itself. Pet owners, for instance, will often entertain mirror conversations with their cats and dogs all day long—without any obvious consequences for their pets’ mental architectures. Human babies seem to be different in that respect. They do exploit social mirrors for shaping and, in fact, for making their minds. What then, do humans have that cats and dogs do not? Humans, I submit, utilize mirrors inside. Mirrors inside are representational devices that help people to exploit what mirrors outside afford, henceforth referred to as mirror devices. Basically, these mirror devices serve to couple perception and action. But they do so in a special way, allowing for the operation of similarity between perception and action. 3.1. Design principles How do mirror devices work and how do they interact with mirrors outside? Before we can begin to answer this question, there is a functional problem to be solved: Consider our target individual T, watching what M is doing. Suppose that M will occasionally mirror T, but that, for most of the time, M will be doing something else. This raises the problem of how T can tell mirroring from non-mirroring in M’s actions. As long as this problem is unsolved, T has no way of capitalizing on what the social mirror facing him or her affords. Mirror devices solve this problem by virtue of two basic design principles, common coding and distal reference. The notion of common coding posits a shared representational domain for perception and action. Common coding demands that the same representational resources are used for both, planning and control of own action and perception of foreign action. In other words, tokens of own action will get their entries in that domain on exactly the same dimensions as tokens of foreign action. Common coding makes it possible both to perceive and produce similarity between own action and foreign action. With reference to production, M’s mirroring of T’s acting will rely on production of own action that resembles perceived foreign action (Hommel, Müsseler, Aschersleben, & Prinz, 2001; Prinz, 1990, 1997, 2002, 2005). Conversely, concerning perception, T’s understanding of the mirror nature of M’s action will rely on the perception of foreign 1108 W. Prinz / Consciousness and Cognition 22 (2013) 1105–1113 action as being similar to previously produced own action. Common coding is thus a prerequisite for the mirror game between the two to work. How can representations of own and foreign action be commensurate? The key feature here is distal reference (cf. Prinz, Aschersleben, & Koch, 2009). Distal reference is fairly obvious on the perceptual side (Brunswik, 1944, 1952, 1955). What we see and what we hear are neither patterns of sensory stimulation nor patterns of brain activation. Instead, we perceive objects and events in the environment—distal events rather than proximal stimuli or even central activations. No less obvious is distal reference on the action side. For instance, when we use a hammer to knock a nail into the wall, we do not plan that action in terms of muscle contractions or activations in our motor cortex. Instead, what we plan is the action itself and its intended outcome in the environment (cf. James, 1890, II, p. 520; Wulf & Prinz, 2001). Distal reference has two important implications: efficiency and publicity. Distal representations are efficient in the sense that they represent environmental events in a way that allows the choosing of actions, which satisfy both current goals and current environmental conditions. Distal representations are public in the sense that they represent events in a way that satisfies the needs for successful communication. This is because they refer to public (i.e., shared) events in the environment, not to private (i.e., unshared) patterns of sensory, motor or brain activations. These two design principles underpin mirror devices inside. These mirrors go either way—to produce own action resembling perceived foreign action and to perceive foreign action resembling own action. Their operation is based on priming through similarity. Perceived foreign action will prime corresponding own action, and own action will prime the perception of corresponding foreign action. Accordingly, they provide information about the self through others. They help us to see and understand our own behavior by way of watching others mirroring our behavior. 3.2. Embodied mirrors So much for the design principles. Let us now consider how they are instantiated in the human mind. This brings us back to the two basic modes of mirroring: embodied and symbolic. Embodied devices operate on implicit procedural knowledge for perception and control of bodies and actions. Much of this knowledge is likely to be contained in representational structures that build on innate resources. Conversely, symbolic devices operate on explicit declarative knowledge about bodies and actions. Most of that knowledge will be contained in representational structures that build on acquired, language-based resources. Here, I focus on devices for embodied mirroring, concentrating on two major kinds of such devices: body schemes and action schemes. I posit that these schemes are, from the outset, shared between others and the self—perhaps even having first developed for knowing and understanding others and only then projected back for knowing and understanding the self (Prinz, 2012). Obviously, a view like this poses a challenge to the widely accepted notion that knowledge of one’s own mind is the natural fundament from which knowledge of other minds may derive. Instead, it suggests either the parallel emergence of understanding the self and others or even the reverse order: that knowledge of others is the fundament from which knowledge of the self derives. The notion of body scheme was initially introduced to account for the tacit representational basis of posture and movement (Gallagher, 1995; Gallagher & Meltzoff, 1996; Head, 1920; Schilder, 1935). What kinds of information must body schemes contain and how do they instantiate the design principles for mirrors inside? The natural answer to the first part of the question seems to be that the body scheme must be created from information provided by sense organs in the body, particularly in muscles, tendons and joints. Yet, if we want the body to be spatially and temporarily attuned to the environment we need to represent both, body and environment, in the same format of distal events. Hence, in order to be in a position to control the body from the inside, the body scheme needs to ‘‘know’’ the body from the outside as well. This conclusion is supported by clinical and experimental observations on so-called ‘‘autoscopic’’ phenomena. This term refers to a set of often bizarre experiences that may either be provoked in certain kinds of virtual-reality based experimental settings (Arzy, Seeck, Ortigue, Spinelli, & Blanke, 2006; Blanke, Ortigue, Landis, & Seeck, 2002), or else occur in the context of a variety of neurological and psychiatric disorders (e.g., Brugger, 2002, 2005). The key feature of autoscopic phenomena is the illusory reduplication of the body. Patients sometimes feel, or see, either their own body, or some body who comes very close to their own. For instance, in autoscopic hallucinations they see their own face in front of them as if in a mirror, or they see a person facing them and repeating their own actions in a mirror-like fashion. The body scheme acts like a mirror here. Patients see themselves as others and like others, and they look at themselves like others do. These observations suggest that the creation of equivalence between own and foreign body is a genuine function that the body scheme has adopted beyond its role for posture and movement. In these pathologies the own body appears like another one. Experimental evidence suggests that the reverse may hold as well, i.e., that foreign bodies or body parts may be equivalent to one’s own (e.g., van den Bos & Jeannerod, 2002). A striking demonstration of the tight coupling between the inside and the outside perspective on body parts comes from the rubber hand illusion (Armel & Ramachandran, 2003; Botvinick & Cohen, 1998; Tsakiris & Haggard, 2005). When observers watch a rubber hand in front of them while their own hand is hidden behind an occluder they may come to perceive the rubber hand as their own. For instance, when the rubber hand and their own hand are touched simultaneously, they tend to localize that touch in the rubber hand they are watching, not in their own hidden hand. Equivalence of perceiving what happens to others and to the self is also substantiated by brain imaging. Recent studies have shown that, when people observe others being touched or pricked at certain locations on their body surface, this will activate the same brain sites that become active W. Prinz / Consciousness and Cognition 22 (2013) 1105–1113 1109 when the observer is touched or pricked at these locations (Keysers, Wicker, Gazzola, Anton, & Fogassi, 2004; Singer & Lamm, 2009; Singer et al., 2004). Last but not least, there is another line of evidence suggesting that part of a mirror-like body scheme is already functional at birth. This evidence comes from studies of facial imitation in newborns (Meltzoff, 2002, 2005; Meltzoff & Moore, 1977, 1983, 1989; Nagy et al., 2005). When newborn babies, in the very first hours of their lives, watch a human face performing gestures like opening the mouth or protruding the tongue, they tend to produce similar gestures by themselves. Similarity here concerns both, body parts and action patterns. Babies will respond to tongue protrusion with tongue protrusion, not lip protrusion, and they will respond to lip protrusion with lip protrusion, not lip opening. At least as far as facial parts and gestures are concerned, newborns seem to come with mirrors inside. These help them to relate others’ faces to their own face. The notion of action scheme refers to representational devices for matching own action to foreign action and vice versa. Such devices have been extensively studied over the past two decades, often with explicit reference to the mirror metaphor, using terms like mirror neurons or mirror systems. What kind of information do action schemes contain and how do they instantiate the design principles for embodied mirrors? In a nutshell, the notion of action schemes implies that representational resources subserving the production of (own) action will also subserve the perception of (foreign) action. This notion has recently gained strong support from both behavioral and brain studies. Here, I will focus on the behavioral side. The rationale for behavioral studies is simple and straightforward: If action production and action perception share representational resources, then some kind of interference should result when both perception and production draw on these resources at the same time. Perception of foreign action should then modulate production of own concurrent action, and likewise production of own action should modulate perception of foreign concurrent action. In both cases, the degree of mutual modulation should depend on the representational overlap between perceived and produced action. One line of evidence for this comes from studies showing that action perception may modulate concurrent action production. For instance, it has been shown that the initiation and selection of particular gestures may be modulated by concurrent perception of related gestures (Brass, Bekkering, & Prinz, 2001; Brass, Bekkering, Wohlschläger, & Prinz, 2000; Jacobs & Shiffrar, 2005). The same interference effect may be obtained for the perception of static postures as dynamic gestures. Interference is particularly pronounced for postures reflecting the goal states of the gestures to be produced—suggesting that representations of goal states play a crucial role for the mechanisms underlying gesture selection (Stürmer, Aschersleben, & Prinz, 2000). A prominent role for action goals is also supported by studies on imitation. Here, it has been shown that movement errors (i.e., incorrect movements to correct goals) are much more frequent than goal errors (i.e., correct movements to incorrect goals). This suggests that the underlying representational resources must contain more information than just the kinematics of perceived and to-be-produced movements. They seem to contain information about full-fledged goal-directed actions, with goals taking the lead over movements (Bekkering & Prinz, 2002; Bekkering & Wohlschläger, 2002). Further evidence comes from studies showing that action production may modulate the perception of actions and events. In a variety of experimental paradigms, it has been shown that ongoing action may modulate the concurrent perception of related objects, events and actions. For instance, the perception of the direction of an arrow, the orientation of a line or the rotation direction of an ambiguous apparent motion can be modulated through concurrent action (Craighero, Bello, Fadiga, & Rizzolatti, 2002; Craighero, Fadiga, Rizzolatti, & Umiltà, 1998; Craighero, Fadiga, Rizzolatti, & Umiltà, 1999; Koch & Prinz, 2005; Müsseler & Hommel, 1997a,b; Schubö, Aschersleben, & Prinz, 1998, 2001; Wohlschläger, 2000). Remarkably, such modulation is often not only obtained while the action is actually being performed, but also while it is being planned and prepared. Furthermore, it may take either form: facilitation or inhibition. For instance, the planning or execution of rotary hand movements has been shown to facilitate the visual perception of apparent rotary motions in the same direction. Conversely, the pressing of keys on the left vs. right hand side has been shown to impede the perception of arrows pointing in the same direction. Even more evidence comes from a third type of study showing that perceptual performance may depend on action-related knowledge and skills. A famous example is the impact of procedural action knowledge on the perception of motion velocity. On the action side, it has long been known that the velocity of drawing movements is lawfully related to the radius of the trajectory (Viviani, 2002; Viviani & Terzuolo, 1982). More recent studies have shown that the same lawful relationship is also effective in perception. For instance, the velocity of a moving dot is perceived to be constant if (and only if) its motion actually follows the law governing movement production, i.e., if it accelerates and decelerates, depending on local curvature. Conversely, if physical velocity is kept constant, perceived velocity is inversely related to curvature (Viviani, 2002; Viviani & Stucchi, 1992). Again, these findings show that procedural knowledge for action production is also involved in perceptual processing. In summary, the behavioral evidence suggests that action perception and production do, in fact, share common representational resources. Action schemes act like mirrors inside, providing embodied procedures for matching one’s own action to others’. The very same conclusion is, of course, suggested by the literature on mirror neurons and mirror systems, which are involved in action perception and production. In that literature, the evidence comes from both electrophysiological studies of the monkey brain and imaging and interference studies of the human brain (for overviews cf., e.g., Rizzolatti & Craighero, 2004; Rizzolatti & Sinigalia, 2007). 1110 W. Prinz / Consciousness and Cognition 22 (2013) 1105–1113 4. Mirror games and policies The mere existence of mirror devices does not guarantee that they can be put to use. For instance, for individuals like Robinson Crusoe who live in isolation, devices like the body or action schemes cannot fulfill their mirror function. For mirror devices to work, two basic conditions must be met. One is that other individuals need to be around. This is what Friday’s advent affords for Robinson Crusoe: mirrors inside need to be complemented by mirrors outside. The other is that the two individuals need to interact in particular ways. This is what their reciprocal acting and talking affords: they need to engage in mirror games. Mirror games are, in other words, social practices designed to confront mirrors inside with mirrors outside. Importantly, humans are not unique in this respect. Many songbirds, for instance, also combine mirror-like representational devices with behavioral practices of interactive mirroring (e.g., Marler & Slabbekoorn, 2004; Zeigler & Marler, 2008). It is far from clear what songbird and human mechanisms and practices have in common and in what ways they differ. A crucial difference may pertain to the kind of operations that are supported by these practices. The interactive practices that are prevalent in songbirds seem to be designed to support active mirroring rather than perception of being mirrored. Accordingly, they seem to be tailored to the needs of skill acquisition rather than self-perception and constitution. Here we focus on the human case, without implying that songbirds, like humans, use their devices for creating subjectivity. 4.1. Mirror games Turning back to human practices, we may discern two basic kinds of mirror games: symbolic and embodied. Whereas symbolic games rely on reciprocal talking about action, embodied games, on which I focus here, rely on reciprocal acting. In early infancy embodied mirroring is the only game in town. For caretakers, the practice of reciprocating or complementing the baby’s doings is common and widespread—perhaps a human universal. For babies, these games seem to be of crucial importance for tuning in with and becoming attached to others, as well as laying the ground for perceiving and understanding themselves like others. However, embodied mirror games are not restricted to interactions with preverbal infants. Mirroring habits also apply to interactions among adults. For instance, an individual may place his hands behind his head while facing another individual doing the same (reciprocation). Likewise, an individual may accompany another individual’s acting through pertinent facial and bodily gestures, thereby commenting on that acting in a non-verbal format. As a rule, such mirroring is not explicitly cultivated as a social practice. Individuals will often have no explicit intention to communicate anything to others and they may not even be aware of what they are doing. Their mirroring reflects automatic habits rather than controlled and cultivated practices (Bargh, Chen, & Burrows, 1996; Chartrand & Bargh, 1999), and sometimes these habits may even be considered impropriate conduct that ought to be suppressed. Still, from the viewpoint of the others, these implicit habits have exactly the same functions and consequences as explicit practices: they let people perceive and receive their own doings through the mirroring of others. In a nutshell, this is what mirror games afford. They provide self-related information through the other. By engaging in mirror games, people make capital out of their capacity to understand agency in others in order to construe mentality and agency in themselves. In a way, these games exploit others to build selves. 4.2. Mirror policies This term refers to tactics and strategies governing individuals’ proneness to engage and become engaged in mirror games. Here, we may discern two basic dimensions on which mirror policies vary. One concerns conditions under which individuals may be prone to imitate others and/or become imitated by others. Recent evidence suggests (Nagy, 2006; Nagy & Molnar, 2004; Trevarthen, 2003) that even though newborns may, at times, be prepared to not only imitate certain gestures but also provoke imitated responses by their caretakers, at other times, they may be less prepared to do so. Mirroring and becoming mirrored is thus already for them controlled by the proneness to engage in the game. The other dimension of mirror policies concerns selectivity. Individuals may, in fact, be quite selective in playing mirror games. For instance, they may mirror some kinds of behaviors, but not others. They may engage in mirror games under some kinds of circumstances, but not under others. Most importantly, they may be selective with respect to the target individuals whom they mirror. They may be prone to mirror certain individuals, but refuse to mirror certain others. For instance, they may mirror their kids, their folks and perhaps their peers, but perhaps not, or to a much lesser degree, strangers, disabled individuals or old people. We can, therefore, think of each mirror individual as entertaining an implicit list of target individuals with whom he or she is prone to engage in mirror games, and of each target individual as being included in some individuals’ personal mirror lists, but excluded from other individuals’ lists. Mirror policies can, thus, act to induce both, social assimilation and dissimilation—and eventually discrimination. Assimilation is based on the dialectics of mirroring and perceiving being mirrored. Likewise, dissimilation and discrimination are based on the dialectics of refusing to mirror and perceiving being refused. Policies for embodied mirroring may, thus, add to the various kinds of language-based games and policies through which social assimilation and discrimination are established and maintained. W. Prinz / Consciousness and Cognition 22 (2013) 1105–1113 1111 5. Real selves If it is true that the mental self is created in mirror games, then subjectivity is an artifact made by humans. It is not a natural fact; rather, it is created in attribution discourses and mirror practices, and in the dialectic back and forth of attribution and appropriation. Yet, when one looks at this more closely, attribution discourses and mirror practices actually only explain how individuals come to understand themselves as subjects, building up a representation of their own subjectivity. Representation of subjectivity is not, however, the same as subjectivity itself. Could it, therefore, be that the personal experience of subjectivity and selfhood is nothing more than a collectively produced and shared illusion that has nothing to do with the reality of the subpersonal machinery of cognition and volition? I maintain that this concern is unfounded. Basically, there are two arguments against it: a general and a specific one. The general argument pertains to the ontological status of social artifacts. In the realm of social artifacts, which are created by social exchange, reality is always constructed by representation. Therefore, it would be misleading to conclude that only the subpersonal machinery is real and efficacious, while representations of its workings at the personal level are epiphenomenal and inefficacious. Social artifacts are, on the contrary, real and efficacious by virtue of the beliefs and representations pertaining to them. Popes, presidents and prime ministers are popes, presidents and prime ministers because we consider them to be this. In exactly the same way, subjects are real and efficacious because we (and they themselves) believe them to be. Added to this—and this is the special argument—is the fact that subjects are auto-artifacts, that is, systems which observe themselves and develop representations of their own activity. As we know from systems theory, self-observation can become operative in such auto-referential systems (e.g., Luhmann, 1984/1995). That means that the representations that systems develop of their own activity may acquire the power to control and modulate this activity. In the case at hand, this principle is realized by the fact that personal interpretations of subpersonal processes must, likewise, naturally be borne by subpersonal processes—by processes that are cut from the same cloth as the subpersonal processes which represent the outer world, bring about decisions and control actions. Both kinds of processes are acted out in the same functional medium and can have a direct influence on each other. Are selves real? People have a self in the same sense as they have, for example, money, courts of law, or governments. Money, courts of law, and governments are social institutions that people create and recognize. The same applies to mental selves as auto-artifacts. Once established, these artifacts are, for the person, in no way fictitious or even illusory; rather, they are real facts which determine and constrain their range of action in exactly the same way as the facts of the natural surroundings in which they live. References Amsterdam, B. (1972). Mirror self-image reactions before age two. Developmental Psychobiology, 5, 297–305. Armel, K. C., & Ramachandran, V. S. (2003). Projecting sensations to external objects: Evidence from skin conductance response. Proceedings of the Royal Society B: Biological Sciences, 270, 1499–1506. Arzy, S., Seeck, M., Ortigue, S., Spinelli, L., & Blanke, O. (2006). Induction of an illusory shadow person. Nature, 443, 287. Bargh, J. A., Chen, M., & Burrows, L. (1996). Automaticity of social behavior: Direct effects of trait construct an stereotype activation on action. Journal of Personality and Social Psychology, 71, 230–244. Bekkering, H., & Prinz, W. (2002). Goal representations in imitative actions. In K. Dautenhahn & C. L. Nehaniv (Eds.), Imitation in animals and artifacts (pp. 555–572). Cambridge, MA: MIT Press. Bekkering, H., & Wohlschläger, A. (2002). Action perception and imitation: A tutorial. In W. Prinz & B. Hommel (Eds.). Common mechanisms in perception and action: Attention and performance (Vol. XIX, pp. 294–314). Oxford: Oxford University Press. Blanke, O., Ortigue, S., Landis, T., & Seeck, M. (2002). Stimulating illusory own-body perceptions. Nature, 419, 269–270. Bogdan, R. J. (Ed.). (1991). Mind and common sense: Philosophical essays on commonsense psychology. Cambridge, UK: Cambridge University Press. Botvinick, M., & Cohen, J. (1998). Rubber hands ‘‘feel’’ touch that eyes see. Nature, 391, 756. Brass, M., Bekkering, H., & Prinz, W. (2001). Movement observation affects movement execution in a simple response task. Acta Psychologica, 106, 3–22. Brass, M., Bekkering, H., Wohlschläger, A., & Prinz, W. (2000). Compatibility between observed and executed finger movements: Comparing symbolic, spatial, and imitative cues. Brain and Cognition, 44, 124–143. Brugger, P. (2002). Reflective mirrors: Perspective-taking in autoscopic phenomena. Cognitive Neuropsychiatry, 7, 179–194. Brugger, P. (2005). From phantom limb to phantom body: Varieties of extracorporal awareness. In G. Knoblich, I. M. Thornton, M. Grosjean, & M. Shiffrar (Eds.), Human body perception from the inside out (pp. 171–209). New York: Oxford University Press. Brunswik, E. (1944). Distal focussing of perception: Size constancy in a representative sample of situations. Psychological Monographs, 254. Brunswik, E. (1952). Conceptual framework of psychology (Vol. 1). Chicago: University of Chicago Press. Brunswik, E. (1955). Representative design and probabilistic theory in a functional psychology. Psychological Review, 62, 193–217. Chartrand, T. L., & Bargh, J. A. (1999). The chameleon effect: The perception-behavior link and social interaction. Journal of Personality and Social Psychology, 76, 893–910. Cooley, C. H. (1902). Human nature and the social order. New York: Scribner. Craighero, L., Bello, A., Fadiga, L., & Rizzolatti, G. (2002). Hand action preparation influences the processing of hand pictures. Neuropsychologia, 29, 517–537. Craighero, L., Fadiga, L., Rizzolatti, G., & Umiltà, C. (1998). Visuomotor priming. Visual Cognition, 5, 109–125. Craighero, L., Fadiga, L., Rizzolatti, G., & Umiltà, C. (1999). Action for perception: A motor-visual attentional effect. Journal of Experimental Psychology: Human Perception and Performance, 25, 1673–1692. Gallagher, S. (1995). Body schema and intentionality. In J. Bermúdez, A. Marcel, & N. Eilan (Eds.), The body and the self (pp. 225–244). Cambridge, MA: MIT Press. Gallagher, S., & Meltzoff, A. W. (1996). The earliest sense of self and others: Merleau-Ponty and recent developmental studies. Philosophical Psychology, 9, 213–236. Gallup, G. J. (1970). Chimpanzees: Self-recognition. Science, 167, 86–87. Greenwood, J. D. (Ed.). (1991). The future of folk psychology: Intentionality and cognitive science. Cambridge, UK: Cambridge University Press. Gumpp, J. (1646). Self portrait. <http://commons.wikimedia.org/wiki/File:Johannes_gumpp.jpg> Retrieved. Head, H. (1920). Studies in neurology. London: Frowde & Hodder. Hegel, G. W. F (1807). Phänomenologie des Geistes (Vol. 1). Bamberg: Joseph. 1112 W. Prinz / Consciousness and Cognition 22 (2013) 1105–1113 Hommel, B., Müsseler, J., Aschersleben, G., & Prinz, W. (2001). The theory of event coding (TEC): A framework for perception and action planning. Behavioral and Brain Sciences, 24, 849–878. Jacobs, A., & Shiffrar, M. (2005). Walking perception by walking observers. Journal of Experimental Psychology: Human Perception & Performance, 31, 157–169. James, W. (1890). The principles of psychology. New York: Holt. Keysers, C., Wicker, B., Gazzola, V., Anton, J., & Fogassi, L. (2004). A touching sight: SII/PV activation during the observation and experience of touch. Neuron, 42, 335–346. Koch, I., & Prinz, W. (2005). Response preparation and code overlap in dual tasks. Memory & Cognition, 33, 1085–1095. Kusch, M. (1999). Psychological knowledge: A social history and philosophy. London: Routledge. Lacan, J. (1949/1977). The mirror stage as formative of the function of the I as revealed in psychoanalytic experience (A. Sheridan, Trans.) Écrits: A selection. New York: Norton. Luhmann, N. (1984/1995). Social systems (J. Bednarz & D. Baecker. Trans.). Stanford: Stanford University Press. Malle, B. F. (2004). How the mind explains behavior. Folk explanations, meaning, and social interaction. Cambridge, MA: MIT Press. Malle, B. F., Moses, L. J., & Baldwin, D. A. (2001). Intentions and intentionality. Foundations of social cognition. Cambridge, MA: MIT Press. Marler, P. R., & Slabbekoorn, H. (2004). Nature’s music: The science of birdsong. San Diego, CA: Academic Press. Mead, G. H. (1934). Mind, self, & society: From the standpoint of a social behaviorist. Chicago: University of Chicago Press. Meltzoff, A. N. (2005). Imitation and other minds: The ‘‘like me’’ hypothesis. In S. Hurley & N. Chater (Eds.), Perspectives on imitation: From neuroscience to social science. Imitation, human development, and culture (Vol. 2, pp. 55–77). Cambridge, MA: MIT Press. Meltzoff, A. N. (1990). Foundations for developing a concept of self: The role of imitation in relating self to other and the value of social mirroring, social modeling, and self-practice in infancy. In D. Cicchetti & M. Beeghly (Eds.), The self in transition: Infancy to childhood (pp. 139–164). Chicago: University of Chicago Press. Meltzoff, A. N. (2002). Elements of a developmental theory of imitation. In A. N. Meltzoff & W. Prinz (Eds.), The imitative mind (pp. 19–41). Cambridge, UK: Cambridge University Press. Meltzoff, A. N., & Moore, M. K. (1977). Imitation of facial and manual gestures by human neonates. Science, 198, 75–78. Meltzoff, A. N., & Moore, M. K. (1983). Newborn infants imitate adult facial gestures. Child Development, 54, 702–709. Meltzoff, A. N., & Moore, M. K. (1989). Imitation in newborn infants: Exploring the range of gestures imitated and the underlying mechanisms. Developmental Psychology, 25, 954–962. Müsseler, J., & Hommel, B. (1997a). Blindness to response-compatible stimuli. Journal of Experimental Psychology: Human Perception and Performance, 23, 861–872. Müsseler, J., & Hommel, B. (1997b). Detecting and identifying response-compatible stimuli. Psychonomic Bulletin & Review, 4, 125–129. Nadel, J. (2002). Imitation and imitation recognition: Functional use in preverbal infants and nonverbal children with autism. In A. N. Meltzoff & W. Prinz (Eds.), The imitative mind: Development, evolution, and brain bases (pp. 42–62). Cambridge, UK: Cambridge University Press. Nagy, E. (2006). From imitation to conversation: The first dialogues with human neonates. Infant and Development, 15, 223–232. Nagy, E., Compagne, H., Orvos, H., Pal, A. P., Molnar, P., & Janszky, I. (2005). Index finger movement imitation by human neonates: Motivation, learning, and left-hand preference. Pediatric Research, 58, 749–753. Nagy, E., & Molnar, P. (2004). Homo imitans or homo provocans? Human imprinting model of neonatal imitation. Infant Behavior and Development, 27, 54–63. Ovid (1977). Metamorphoses. Books I–VIII. Cambridge, MA: Harvard University Press. Paukner, A., Anderson, J. R., Borelli, E., Visalberghi, E., & Ferrari, P. F. (2005). Macaques (Macaca nemestrina) recognize when they are being imitated. Biological Letters, 1, 219–222. Prinz, W. (1997). Perception and action planning. European Journal of Cognitive Psychology, 9, 129–154. Prinz, W. (2005). An ideomotor approach to imitation. In S. Hurley & N. Chater (Eds.), Perspectives on imitation: From neuroscience to social science. Mechanisms of imitation and imitation in animals (Vol. 1, pp. 141–156). Cambridge, MA: MIT Press. Prinz, W. (2012). Open minds: The social making of agency and intentionality. Cambridge, MA: MIT Press. Prinz, W. (2013). Real artifacts: The social creation of subjectivity. In B. L. Kaldis (Ed.), Mind and society: Cognitive science meets the social sciences. Springer. Prinz, W., Aschersleben, G., & Koch, I. (2009). Cognition and action. In E. Morsella, J. A. Bargh, & P. M. Gollwitzer (Eds.). Oxford handbook of human action (Vol. 2). Oxford: Oxford University Press. Prinz, W. (2002). Experimental approaches to imitation. In A. N. Meltzoff & W. Prinz (Eds.), The imitative mind: Development, evolution, and brain bases (pp. 143–162). Cambridge, UK: Cambridge University Press. Prinz, W. (1990). A common coding approach to perception and action. In O. Neumann & W. Prinz (Eds.), Relationships between perception and action: Current approaches (pp. 167–201). Berlin: Springer. Prinz, W. (2008). Mirrors for embodied communication. In I. Wachsmuth, M. Lenzen, & G. Knoblich (Eds.), Embodied communication in humans and machines (pp. 111–128). Oxford: Oxford University Press. Rizzolatti, G., & Craighero, L. (2004). The mirror-neuron system. Annual Review of Neuroscience, 27, 169–192. Rizzolatti, G., & Sinigalia, C. (2007). Mirrors in the brain: How our minds share actions, emotions, and expertise. New York: Oxford University Press. Schilder, P. M. (1935). The image and appearance of the human body (Vol. 4). London: Kegan Paul, Trench, Trubner & Co. Schubö, A., Aschersleben, G., & Prinz, W. (2001). Interactions between perception and action in a reaction task with overlapping S-R assignments. Psychological Research, 65, 145–157. Schubö, A., Aschersleben, G., & Prinz, W. (1998). Interferenz bei der Ausführung von Bewegungssequenzen. In H. Lachnit, A. Jacobs, & F. Rösler (Eds.), Experimentelle Psychologie (pp. 319–320). Lengerich: Pabst. Singer, T., & Lamm, C. (2009). The social neuroscience of empathy. The Year in Cognitive Neuroscience 2009. Annals of the New York Academy of Sciences, 1156, 81–96. Singer, T., Seymour, B., O’Doherty, J., Kaube, H., Dolan, R. J., & Frith, C. D. (2004). Empathy for pain involves the affective but not sensory components of pain. Science, 303, 1157–1162. Smith, A. (1759/1976). The theory of moral sentiments (Vol. 1). Oxford: Clarendon Press. Stürmer, B., Aschersleben, G., & Prinz, W. (2000). Correspondence effects with manual gestures and postures: A study of imitation. Journal of Experimental Psychology: Human Perception and Performance, 26, 1746–1759. Trevarthen, C. (1993). The self born in intersubjectivity: The psychology of an infant communicating. In U. Neisser (Ed.), The perceived self. Ecological and interpersonal sources of self-knowledge (pp. 121–173). Cambridge, UK: Cambridge University Press. Trevarthen, C. (2003). Stepping away from the mirror: Pride and shame in adventures of companionship. Reflections on the nature and emotional needs of infant intersubjectivity. In C. S. Carter, L. Ahnert, K. E. Grossmann, S. B. Hrdy, M. E. Lamb, S. W. Porges & N. Sachser (Eds.), Attachment and bonding (pp. 55–84). Cambridge, MA: MIT Press. Trevarthen, C. (1998). The concept and foundations of infant intersubjectivity. In S. Bråten (Ed.), Intersubjective communication and emotion in early ontogeny (pp. 15–46). Cambridge, UK: Cambridge University Press. Trevarthen, C., Kokkinaki, T., & Fiamenghi, G. A. J. (1999). What infants’ imitations communicate: with mothers, with fathers and with peers. In J. Nadel & G. Butterworth (Eds.), Imitation in infancy (pp. 127–185). Cambridge, UK: Cambridge University Press. Tsakiris, M., & Haggard, P. (2005). The rubber hand illusion revisited: Visuotactile integration and self-attribution. Journal of Experimental Psychology: Human Perception & Performance, 31, 80–91. van den Bos, E., & Jeannerod, M. (2002). Sense of body and sense of action both contribute to self-recognition. Cognition, 85, 177–187. W. Prinz / Consciousness and Cognition 22 (2013) 1105–1113 1113 Viviani, P. (2002). Motor competence in the perception of dynamic events: A tutorial. In W. Prinz & B. Hommel (Eds.). Common mechanisms in perception and action: Attention and Performance (Vol. XIX, pp. 406–442). Oxford: Oxford University Press. Viviani, P., & Stucchi, N. (1992). Biological movements look uniform: Evidences of motor-perceptual interactions. Journal of Experimental Psychology: Human Perception & Performance, 18, 603–623. Viviani, P., & Terzuolo, C. (1982). Trajectory determines movement dynamics. Neuroscience, 7, 431–437. Whitehead, C. (2001). Social mirrors and shared experiential worlds. Journal of Consciousness Studies, 8, 3–36. Wohlschläger, A. (2000). Visual motion priming by invisible actions. Vision Research, 40, 925–930. Wulf, G., & Prinz, W. (2001). Directing attention to movement effects enhances learning: A review. Psychonomic Bulletin & Review, 8, 648–660. Zeigler, H. P., & Marler, P. R. (Eds.). (2008). Neuroscience of birdsong. Cambridge, UK: Cambridge University Press. Zukow-Goldring, P. (2006). Assisted imitation: Affordances, effectivities, and the mirror system in early language development. In M. A. Arbib (Ed.), Action to language via the mirror neuron system (Vol. 469–500). Cambridge, UK: Cambridge University Press.

Consciousness and Cognition 42 (2016) 113–124 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Imagining the future: A cross-cultural perspective on possible selves q Clare J. Rathbone a,⇑, Sinué Salgado b, Melisa Akan c, Jelena Havelka d, Dorthe Berntsen b a Oxford Brookes University, Department of Psychology, Social Work and Public Health, Oxford, United Kingdom Aarhus University, Department of Psychology, Center on Autobiographical Memory Research, Aarhus, Denmark c Bogaziçi University, Department of Psychology, Istanbul, Turkey d University of Leeds, School of Psychology, Leeds, United Kingdom b a r t i c l e i n f o Article history: Received 31 July 2015 Revised 24 February 2016 Accepted 6 March 2016 Keywords: Possible self Future self Self-image Identity Future cognitions Imagining Life scripts a b s t r a c t This study examined the impact of culture on the qualitative and quantitative features of possible selves. Young adults from Turkey (n = 55), Serbia (n = 64), and the United Kingdom (n = 73) generated images of eight possible selves (e.g. I will be a doctor) which were dated and rated for vividness, positivity, imagery perspective, rehearsal, and according to whether or not they involved other people. All possible selves were coded according to categories (e.g. job, parenthood, self-improvement). There were cross-cultural differences in the types of possible selves generated and in the ratings for vividness, positivity, and rehearsal. Across all three cultures, specific possible selves were more frequently generated than abstract possible selves. Specific possible selves were rated as significantly more vivid and were dated as emerging later than abstract possible selves. Results are discussed with reference to cultural life scripts and the effects of culture on future cognitions. Ó 2016 Elsevier Inc. All rights reserved. 1. Introduction To what extent does culture affect the way people think about their future? It is commonly accepted that culture impacts on self-construal (Markus & Kitayama, 1991), influencing how we define ourselves (Rhee, Uleman, Lee, & Roman, 1995; Wang, 2001, 2004) and how we remember our earliest (Wang, 2006) and most self-defining memories (Jobson & O’Kearney, 2008). In the field of autobiographical memory, it has been suggested that cultural life scripts organize the retrieval of memories across the lifespan, influencing the way people construct both their past (Berntsen & Rubin, 2004), and future (Berntsen & Bohn, 2010). Thus, culture is argued to play a central role in the construction of our identities and in how we recall the past and imagine the future. One key method of examining people’s expectations for the future is to ask them to generate possible selves, that is, identities that people anticipate becoming in the future (Markus & Nurius, 1986). This study bridges the fields of possible selves and cultural life scripts, by examining the commonalities and differences in the ways young adults from three nations (the United Kingdom, Serbia, and Turkey) imagine who they will become in the future. By the use of a new coding scheme it also provides novel findings on cross-cultural differences in the contents of possible selves. q Clare J. Rathbone was supported by the Economic and Social Research Council (ES/K000918/1). We also thank the Danish National Research Foundation (DNRF89) for funding and Chris Moulin for his support and suggestions during the development of this project. ⇑ Corresponding author at: Department of Psychology, Oxford Brookes University, Oxford OX3 0BP, United Kingdom. E-mail address: crathbone@brookes.ac.uk (C.J. Rathbone). http://dx.doi.org/10.1016/j.concog.2016.03.008 1053-8100/Ó 2016 Elsevier Inc. All rights reserved. 114 C.J. Rathbone et al. / Consciousness and Cognition 42 (2016) 113–124 1.1. Interdependent and independent cultures Research suggests that culture can affect the way people process information, impacting on emotion, motivation, and cognition (Markus & Kitayama, 1991). In the domain of memory research, for example, cross-cultural differences have been found in the content of autobiographical memories (Conway, Wang, Hanyu, & Haque, 2005; Wang & Conway, 2004), selfdefining memories (Jobson & O’Kearney, 2008), earliest memories (Wang, 2006) and the centrality of memories of positive and negative life events (Zaragoza Scherman, Salgado, Shao, & Berntsen, 2014). These cross-cultural effects typically have been interpreted to reflect the use of relatedness (i.e. referring to a wider social group) or autonomous (i.e. referring to the self) focus, depending on whether the participant is from an independent/individualist or interdependent/collectivist culture (Markus & Kitayama, 1991). For example, Jobson and O’Kearney (2008) found that Australian participants (independent culture) provided more elaborate autonomous memories, whereas Asian participants (interdependent culture) generated more elaborate relatedness memories. Rhee et al. (1995) examined self-descriptions (i.e., ‘I am’ statements) from participants in individualistic and collectivistic cultures. They found that participants who strongly identified as Asian Americans tended to generate a higher proportion of social, and lower proportion of autonomous, self-descriptions compared to European Americans. Similarly, Wang (2001) examined the self-descriptions generated by American and Chinese college students. The American students tended to describe themselves using autonomous traits (such as being studious) more frequently than the Chinese students, who generated more collective, social descriptions (such as being a sister). Conway et al. (2005) compared the distribution of autobiographical memories from participants in Japan, China, Bangladesh, England, and the United States, and analysed the content of these memories in the Chinese and American samples. They found that the temporal distribution of memories was relatively stable across cultures, with all five groups showing similar lifespan retrieval curves, characterised by childhood amnesia during the first five years of life and increased retrieval during the reminiscence bump period of ages ten to thirty (e.g. Rubin, Wetzler, & Nebes, 1986). In contrast, there were cross-cultural differences in the content of Chinese and American participants’ memories. The Chinese group’s memories contained more events that involved interdependent (e.g., social) self-focus, whereas the American group recalled more events associated with an autonomous self-focus. Related findings were more recently reported by Zaragoza Scherman, Salgado, Shao, and Berntsen (2015). Although studies such as these suggest important differences in the ways members of different cultures define themselves and recall the past, we know less about the effect of culture on future cognitions. Here we begin to fill this gap by examining possible selves across cultures. 1.2. Possible selves Possible selves are ideas about who a person might become in the future. They are thought to be highly goal-related, incentivizing behaviour by acting as an outcome to be achieved or avoided (Markus & Nurius, 1986). For example, a feared possible self of being someone who fails school exams might motivate a student to revise. Alternatively, a desired possible self of being able to drive to visit friends and family might prompt someone to book driving lessons. Possible selves provide a valuable framework for studying cultural differences in identity as previous research has established that possible selves can influence behaviour (e.g. Oyserman, Bybee, Terry, & Hart-Johnson, 2004). For example, Oyserman, Bybee, and Terry (2006) showed that possession of academic possible selves, linked with plausible strategies for their attainment, was related to improved school attendance and academic performance. Hoppmann, Gerstorf, Smith, and Klumb (2007) studied the relationship between possible selves and behaviour in older adults. They found that having hoped-for possible selves relating to health and social relations was associated with a higher probability of engaging in activities within these domains. Importantly, those who engaged in hope-related daily activities had a higher probability of survival over a 10 year period. As reviewed by Lee et al. (2015) possible selves have been measured in various ways, including the content of a person’s most important possible self (Hooker & Kaus, 1992), the presence of a single target possible self such as being a ‘‘problem drinker” (Corte & Szalacha, 2010) or the number of feared (Oyserman & Markus, 1990), or expected possible selves (Aloise-Young, Hennigan, & Leong, 2001). For example, Aloise-Young et al. (2001) found that possessing a lower number of positive possible selves was related to adolescent alcohol use and cigarette smoking. Together, these studies highlight the important role of possible selves in shaping behaviour. Recent theoretical developments suggest that possible selves may impact on behaviour through their role in self-regulatory processes that influence both motivation and behaviour (Hoyle & Sherrill, 2006; vanDellen & Hoyle, 2008). This work on the goal-directed function of possible selves, although predominantly from the field of social psychology, reflects cognitive models of the self, such as the Self Memory System (Conway & Pleydell-Pearce, 2000), which conceptualizes the self as a goal hierarchy. As the work reviewed above demonstrates (e.g. Aloise-Young et al., 2001; Hoppmann et al., 2007; Oyserman et al., 2006), the way we think about our future goals has implications for the way we live our lives. In spite of the large body of cross-cultural work comparing self-construals (Markus & Kitayama, 1991; Wang, 2001, 2004), autobiographical memories (Wang, 2006; Wang & Conway, 2004) and life scripts (e.g., Ottsen & Berntsen, 2014; Rubin, Berntsen, & Hutson, 2009) to our knowledge, no work has directly compared the possible selves of people living in different countries. Previous research has focused on the possible selves of participants from a range of specific cultures including aboriginals (Senior & Chenhall, 2012) and Latinos (Yowell, 2000). Studies that have directly compared possible selves of participants from different cultures have been based on participants living in one country. For example, work by Oyserman and C.J. Rathbone et al. / Consciousness and Cognition 42 (2016) 113–124 115 colleagues compared the possible selves of high school students from a range of races and ethnicities within the United States (Oyserman & Fryberg, 2006; Oyserman, Gant, & Ager, 1995), and Waid and Frazier (2003) compared the possible selves of older adult native English speakers and native Spanish speakers living in the United States. Thus, the present study is novel in its examination of possible selves from participants living in different countries. The present study also aimed to extend understanding of cross-cultural differences in possible selves by using the cultural life script framework established by Berntsen and Rubin (2004), Rubin and Berntsen (2003). Cultural life scripts are culturally shared representations of the timing of major transitional life events (Berntsen & Rubin, 2004; Rubin & Berntsen, 2003). Thus cultural life scripts refer to the normative life events one would expect to experience in a given cultural group and the order in which they are expected to occur. These events typically include positive social landmarks such as graduating from school, getting a job, getting married, and becoming a parent (Berntsen & Rubin, 2004; Erdogan, Baran, Avlar, Tas, & Tekcan, 2008). A number of studies have compared the life scripts of different cultures (e.g., Ottsen & Berntsen, 2014; Rubin et al., 2009; Zaragoza Scherman, 2013), however, to date, there have been no cross-cultural investigations that bridge the fields of possible selves and cultural life scripts. As well as providing a framework for autobiographical retrieval, cultural life scripts also play a central role in the way people imagine important events in the future. Berntsen and Bohn (2010) asked young adult participants to generate memories and future events. When the cue was simply to think of an important future event, 71% of these future events were life-script related. Cultural life scripts provide a useful framework for exploring cross-cultural differences in possible selves as they allow analysis to go beyond a simple, dichotomous coding of how individualistic versus collectivistic a given group’s possible selves are. Cultural life script analysis generates an extensive set of categories that emerge from the data itself, enabling a more fine-grained examination of subtle differences in the ways participants from different cultures imagine the future. 1.3. The key contributions of the present study The present study adds to the existing literature in a number of important ways. First, the present study examined the content of cross-cultural possible selves using the cultural life script framework established by Berntsen and Rubin (2004), Rubin and Berntsen (2003). We analysed the frequency with which life script categories featured in the possible selves of participants from different cultures. It was predicted that participants would generate possible selves that reflect cultural life script events, such as marriage, occupational transitions, and parenthood (events that feature in the life scripts of participants from a range of cultures, e.g., Erdogan et al., 2008; Janssen, Uemiya, & Naka, 2014; Rubin et al., 2009). This proposal is supported by the results of Rathbone, Conway, and Moulin (2011), who examined the temporal distributions of British young adults’ possible and current selves. Although detailed coding of possible selves was not conducted, 55% of all future selves generated were related to either marriage, becoming a parent, or getting a job – all key features of the cultural life script. In the present study, we also examined whether the individualistic–collectivistic distinction would be demonstrated in possible selves by coding possible selves according to whether they were autonomous or social (e.g. Rhee et al., 1995). We chose to compare results from Turkey, the United Kingdom and Serbia as these groups differ on the collectivis tic–individualistic continuum, with Turkey and Serbia both considered collectivistic, whilst the United Kingdom is classed as individualistic (Erdogan et al., 2008; Hofstede, 1980). Second, also adding to previous work, we here compared the proportion of abstract and specific possible selves generated by participants in different cultures, with abstract self-images denoting traits and specific self-images referring to relationships, jobs, and other social roles (e.g. being retired) (Rhee et al., 1995). Broadly speaking, we expected abstract possible selves to be less associated with life script categories. In contrast, specific possible selves could map onto either life script or non-life-script categories, depending on the specific possible self in question. For example, ‘I will be a mother’ is both specific and life-script-related, whilst ‘I will be someone who reads more books’ is specific but not life-script-related. As such, this study investigated life-script-relatedness and the specific/abstract dichotomy as independent constructs. Third, previous cross-cultural work on possible selves has tended to focus on the content of identities generated (e.g. Oyserman & Fryberg, 2006; Waid & Frazier, 2003). Here we extend the analysis to also include the temporal distribution (based on participant-generated dates) and ratings of possible selves, following the protocols for cognitive examination of the self (Rathbone et al., 2011) and life scripts (Berntsen & Rubin, 2004). For example, would all participants picture themselves at a similar point in the future, or would there be cultural differences in the timeframes of possible selves? This was an empirical question. Previous work has found that young adults in the UK date possible selves as emerging at a mean of only 6.35 (Chessell, Rathbone, Souchay, Charlesworth, & Moulin, 2014) and 7.39 (Rathbone et al., 2011) years into the future, despite having almost the whole lifespan to sample from. We aimed to investigate whether this short-term temporal focus would be replicated in the UK sample, and extend to participants from other cultures. 1.4. Aims The present study had three broad aims. The first aim was to examine the content of possible selves across cultures. Based on previous research on the self-concept in interdependent compared to independent cultures (e.g. Markus & Kitayama, 1991; Rhee et al., 1995; Wang, 2001, 2004) it was predicted that the British participants would generate possible selves that were more autonomous (e.g. referring to personal traits, possessions, or physical descriptions of themselves) rather than social (such as occupation, and family relationships and marriage), compared to the Turkish and Serbian participants. In 116 C.J. Rathbone et al. / Consciousness and Cognition 42 (2016) 113–124 order to undertake this analysis, we developed a new coding scheme, which enabled us to examine the prevalence of social categories (e.g. marriage) compared to more autonomous possible selves (e.g. I will be content; I will be rich), and to carry out cross-cultural comparisons of the content of possible selves. To undertake this analysis all possible selves were coded according to a set of categories, which then formed the basis for the social-autonomous coding. The development of this extended category scheme (see Appendix A) was intended to provide a set of norms for future cross-cultural possible selves research. The second aim was to investigate the phenomenological features of possible selves generated across three cultures, by comparing ratings of vividness, positivity, rehearsal, imagery perspective, and whether mental images of possible selves featured the self alone or with others. We did not have specific predictions about these variables, other than that it was expected that the Serbian and Turkish participants would generate more possible selves involving others than the UK participants, reflecting previous work on the self-concept in interdependent and independent cultures (e.g., Rhee et al., 1995). We were also interested in the temporal distribution of possible selves, and predicted that participants from all cultures would date possible selves as emerging a mean of six to eight years into the future (e.g., Chessell et al., 2014; Rathbone et al., 2011). Our third aim was to compare the phenomenological features of specific and abstract possible selves. This is of theoretical relevance, in part because specific, concrete possible selves may be more likely to motivate behaviour (Hoppmann et al., 2007; Markus & Nurius, 1986; Oyserman et al., 2006). Although no previous studies have compared abstract and specific possible selves in this way, some support for this hypothesis is found in Oyserman et al.’s (2006) findings that possession of academic possible selves that were associated with specific, plausible strategies for their attainment were more likely to have a positive impact on behaviour at school. Whilst it was not the aim of this study to explore the impact of possible selves on behaviour, we aimed to better understand the phenomenological features of specific compared to abstract possible selves. By definition, specific possible selves involve a more concrete approach to considering the future, and we were interested in the phenomenological characteristics associated with these types of future cognitions. These subjective ratings of event characteristics derive from the literature on episodic future thinking and remembering. Thus, including them also helps to connect the present research to more widely used approaches to future thinking (see Szpunar, Spreng, & Schacter, 2014, for a review of different forms of future thoughts). We predicted that specific possible selves, by virtue of their specificity, would be more vivid and more rehearsed than abstract possible selves. Additionally, in line with construal level theory (Trope & Liberman, 2010) we expected specific possible selves to be dated nearer to the present than abstract possible selves. 2. Method 2.1. Participants Participants were recruited from undergraduate psychology degree programmes at universities in the UK, Turkey and Serbia. All received course credits for participating. Seven participants were excluded from the Serbian sample (five reported non-Serbian nationality and two did not provide their nationality) and seven were excluded from the British sample (six reported non-British nationalities and one did not provide their nationality). One participant was excluded from the Turkish sample as his responses suggested he had misunderstood the instructions. The data presented do not include these 15 participants. The Turkish sample (n = 55; 38 females, 17 males) had a mean age of 19.76 (SD = 1.48; Range = 18–26); the British sample (n = 73; 61 females, 12 males) had a mean age of 19.59 (SD = 1.77; Range = 18–30); and the Serbian sample (n = 64; 61 females, 3 males) had a mean age of 19.75 (SD = 1.04; Range = 19–23). 2.2. Materials and procedure All participants completed the questionnaire online in their native language. Participants gave their age, gender, and nationality, and were then asked to generate up to eight ‘‘I will be. . .” statements that might describe their identity in the future, but did not describe them at present. Specific instructions were as follows: We are interested in how you imagine yourself being in the future. Please give up to eight ‘‘I will be. . .” statements that might describe your identity in the future, and do not describe how you are at present. These ‘future identities’ or future personal characteristics might refer to personality traits, careers, hobbies, family roles or anything else that you feel might define your identity in the future. After participants had generated their set of statements, they were re-presented with each statement (one at a time) and asked to think about themselves in the future, acting in accordance with the statement provided. As an example, they were told that if they had said ‘I will be healthier’, they should imagine themselves in the future being healthy in some way. They were instructed to hold each future image in mind whilst they completed a series of rating scales. Each future image was rated on a scale of 1–10 (1 minimum; 10 maximum) for vividness, positivity, and rehearsal. For rating vividness, participants were instructed: On a scale of 1–10 (with 10 being very vivid, and 1 being not vivid at all), how vivid and clear is the image of you being this particular future identity? For example, if you can imagine the event happening very clearly, with details like sounds and smells, this would be rated highly for vividness. For positivity they were instructed: With 10 being very positive and 1 being very negative, how positive is the image of you being this particular future identity? For rating C.J. Rathbone et al. / Consciousness and Cognition 42 (2016) 113–124 117 rehearsal participants were asked: Is this a future image you have thought about a lot, or is this the first time you have imagined it? On a scale of 1–10 (10 being very regularly, and 1 being never) how often have you thought of this future identity before now? Participants also used a dichotomous rating to show whether they saw the future image through their own eyes (field perspective) or as though they were watching themselves (observer perspective), and provided the age they thought they would be when the imagined image took place (they were instructed to provide a specific age in years, rather than a range of ages, for each possible self). After generating these ratings for all possible selves, participants stated whether the images associated with each possible self featured only themselves, or whether they featured other people (and if so, whether this was one, two, three, or at least four other people). These final items about the presence of other people were placed last so as not to influence the images generated during the section in which they were rated for features such as positivity and vividness. The questionnaire was originally prepared in English and was then translated by authors fluent in both English and Serbian (JH) or Turkish (MA). Written responses were translated back into English by these same translators prior to coding. 2.2.1. Coding scheme A coding scheme was developed to analyse the ‘I will be’ statements (see Appendix A). This scheme was in part based on categories in the life script norms for Danish (Berntsen & Rubin, 2004) and Turkish (Erdogan et al., 2008) samples, but also emerged from the initial coding process. Because we asked for possible selves not events – in contrast to the task used in life script studies – the categories sampled here often did not correspond to distinct event categories and thus required the development of new categories, such as being ‘happy’. The coders were instructed to include a category in the scheme if there were two or more instances of that category across the whole sample. If more than one category was generated within one statement (e.g. I will be a husband and father) they were instructed to only code the first statement. All statements were independently coded by two experts (authors CJR and SS), both blind to participant nationality (the dataset used for coding did not contain details about participant group and all statements were listed alphabetically, not organised by participant). The measured Cohen’s Kappa was 0.76, indicating a substantial agreement. Following this, all discrepancies were discussed and raters reached 100% agreement. Two further forms of coding (orthogonal to the categories above) were carried out based on Rhee et al. (1995). First, all statements were dichotomously coded as either abstract (lacking specific details and typically containing references to traits or emotional states) or specific (associated with social roles or qualified by specific details, e.g., I will be a mother, I will work with animals). Second, according to criteria for analysing differences between collectivistic and individualistic cultures (e.g. Markus & Kitayama, 1991; Rhee et al., 1995) all statements were coded as either autonomous (reflecting independent selves; e.g. traits, having possessions, being rich or successful, physical descriptions) or social (reflecting interdependent selves; e.g. family or occupational roles, friendships, religion, falling in love). The criteria used for this coding were the same as those developed by Rhee et al. (1995). Where categories in the coding scheme did not map onto a simple social-autonomous dichotomy (i.e., skill development, other, and activities) each possible self that had been judged as belonging to such a category was additionally coded according to whether the participant rated it as featuring the self alone (autonomous) or with others (social). This applied to a total of just 105 cases. The coding of these cases was carried out independently by both expert raters (CJR and SS) with a high level of agreement (Kappa = 0.99). The discrepancies were discussed until 100% agreement was reached. 3. Results Participants generated a maximum of eight and a minimum of three possible selves. The Serbian sample generated a mean of 7.16 possible selves (SD = 1.48), the Turkish mean was 7.95 (SD = .41) and the British mean was 7.14 (SD = 1.54). There was a significant main effect of culture (F[2, 189] = 7.46, p = .001). Post-hoc Bonferroni-corrected comparisons revealed that the Turkish sample generated significantly more possible selves than the British and Serbian samples (p < .05, corrected), and that there were no differences between the number of possible selves generated by the British and the Serbian samples (p > .05, corrected). As the male:female ratio was not matched across samples, all ANOVAs were repeated including gender as a covariate. Gender had no significant effect on the pattern of results reported. 3.1. Content of possible selves by nationality To compare potential qualitative differences in the categories generated by participants of each nationality, all possible selves were coded by category. Table 1 shows the frequencies (and percentages) of each category within participants from each nationality. Table 1 shows that ‘self-improvement’ (e.g. abstract plans about being better, cleverer or kinder in the future) was the most frequently generated category across all cultures. In addition, a number of established life-script related categories featured heavily in the possible selves of participants from all cultures (e.g. marriage, jobs, parenthood). In order to compare the frequencies of specific categories across cultures, we took the ten most frequently generated categories (i.e., those most frequently mentioned across all three nationalities) and, for each participant, calculated a category score between 0 and 8 for each of these 10 categories (for example, if a participant generated 3 possible selves that were 118 C.J. Rathbone et al. / Consciousness and Cognition 42 (2016) 113–124 Table 1 Content of possible selves by nationality. Category Self-improvement Job-specific Parenthood Marriage Happy Job-general Finance self-improvement Successful Activities Other Health Family Skill development Move Friendship Physical appearance Trait-general Fall in love Acquiring property Education-university Learn languages Aging Possessions Relationship Busy Grandchildren Military Turkey Serbia United Kingdom Count % Count % Count % 141 79 32 20 20 14 19 12 8 18 9 9 9 4 9 4 9 2 1 3 7 0 4 1 0 1 2 32.27 18.08 7.32 4.58 4.58 3.20 4.35 2.75 1.83 4.12 2.06 2.06 2.06 0.92 2.06 0.92 2.06 0.46 0.23 0.69 1.60 0.00 0.92 0.23 0.00 0.23 0.46 133 82 45 31 21 24 12 20 16 11 10 10 7 3 5 9 3 6 0 3 2 1 2 0 0 2 0 29.04 17.90 9.83 6.77 4.59 5.24 2.62 4.37 3.49 2.40 2.18 2.18 1.53 0.66 1.09 1.97 0.66 1.31 0.00 0.66 0.44 0.22 0.44 0.00 0.00 0.44 0.00 129 42 48 46 44 34 26 22 23 5 15 14 8 15 5 4 4 4 11 5 2 6 0 4 4 1 0 24.76 8.06 9.21 8.83 8.45 6.53 4.99 4.22 4.41 0.96 2.88 2.69 1.54 2.88 0.96 0.77 0.77 0.77 2.11 0.96 0.38 1.15 0.00 0.77 0.77 0.19 0.00 coded as self-improvement, such as becoming a better person, clever, and academic, then they would have a selfimprovement score of 3). These scores were analysed with one-way between subjects ANOVAs to examine the effect of nationality on category score for each of the 10 top categories (see Table 2 for results). The results in Table 2 indicate several significant cross-cultural differences in the types of possible selves generated. These were examined using post-hoc Bonferroni-corrected pairwise comparisons. In general, and as predicted, the British sample stood out. The British sample generated significantly fewer possible selves that concerned ‘self-improvement’ compared to Turkey (p < .02, corrected); the British sample generated significantly fewer possible selves that were ‘job-specific’ compared to the Turkish (p < .001, corrected) and Serbian samples (p < .001 corrected); the British sample generated significantly more marriage-related possible selves than the Turkish sample (p < .01, corrected); and the British sample generated significantly more possible selves associated with being ‘happy’ than the Turkish (p < .03, corrected) or Serbian samples (p < .005, corrected). We examined whether participants from traditionally more collectivistic cultures (Turkey and Serbia) generated more social possible selves compared to participants from the more individualistic UK by examining the frequencies of social compared to autonomous possible selves across cultures. The Turkish sample generated 234 autonomous and 203 social possible selves, the British sample 277 autonomous and 244 social possible selves, and the Serbian sample 224 autonomous and 234 social possible selves. Within all three cultures there was no significant difference between the number of social compared to autonomous possible selves generated (v2 < 2.20, df = 1, p > .138) and there was no significant effect of nationality (v2 = 2.45, df = 2, p = .293). 3.2. Ratings of future images by nationality In order to compare the ratings of possible selves across cultures, mean rating scores were calculated for four of the possible self measures: vividness, positivity, rehearsal and distance from present (each mean score was calculated from all possible selves generated by each participant). Distance from present was calculated as date of possible self minus participant age, and was used instead of date of possible self to account for small differences in participants’ ages. Proportional scores were calculated for the dichotomous self measures: field/observer and alone/with others (proportional scores were calculated as the proportion of each participant’s possible selves that were field, and only featured themselves, respectively). See Table 3 for comparison of mean ratings and dates of possible selves by nationality. There were significant cross-cultural differences on ratings of vividness, positivity, and rehearsal. Bonferroni-corrected post-hoc comparisons showed that the British sample’s possible selves were significantly less vivid than those from the Turkish (p < . 002, corrected) and Serbian (p < .001, corrected) samples. The British sample’s possible selves were significantly less positive than the Serbian sample (p < .002, corrected), and significantly less rehearsed than the Turkish sample (p < .01, 119 C.J. Rathbone et al. / Consciousness and Cognition 42 (2016) 113–124 Table 2 Mean category prevalence score (for top 10 categories) by nationality. Category Self-improvement Job-specific Parenthood Marriage Happy Job-general Finance self-improvement Successful Activities Other Turkey Serbia United Kingdom Mean SD Mean SD Mean SD 2.55 1.44 0.58 0.36 0.36 0.25 0.35 0.22 0.15 0.35 1.63 1.07 0.50 0.49 0.49 0.44 0.48 0.42 0.40 0.58 2.06 1.28 0.70 0.48 0.33 0.38 0.19 0.31 0.25 0.25 1.54 1.16 0.46 0.50 0.51 0.52 0.39 0.47 0.44 0.56 1.74 0.58 0.66 0.63 0.60 0.47 0.36 0.30 0.32 0.23 1.42 0.71 0.48 0.49 0.49 0.58 0.48 0.46 0.57 0.49 F p g2 4.38 14.55 0.96 4.69 6.23 2.56 2.80 0.76 1.92 0.75 .01 <.001 .38 .01 <.001 .08 .06 .47 .15 .47 .04 .13 .01 .05 .06 .03 .03 .01 .02 .01 F p g2 10.53 7.77 5.03 2.76 0.36 2.04 <.001 <.001 .01 .07 .70 .13 .10 .08 .05 .03 .00 .02 Table 3 Mean phenomenological ratings and dates of possible selves by nationality. Scale Vividness Positivity Rehearsal Proportion field Proportion alone Distance from present Turkey Serbia United Kingdom Mean SD Mean SD Mean SD 7.90 8.27 7.24 0.48 0.34 8.13 1.17 1.00 1.59 0.24 0.16 3.48 7.96 8.90 7.01 0.45 0.36 7.21 1.07 1.03 1.33 0.29 0.22 3.21 7.04 8.07 6.42 0.57 0.33 6.92 1.59 1.59 1.63 0.31 0.19 3.59 corrected). Finally, the Serbian sample’s possible selves were significantly more positive than the Turkish sample (p < .03, corrected). There were no significant cross-cultural differences on the number of images featuring the self alone, or viewed from a field perspective, and no effects of culture on the dated ages of possible selves (i.e. distance from present). 3.3. Abstract and specific possible selves across nationalities In addition to the category coding, all statements were also coded as abstract or specific. The Turkish sample generated 284 specific and 153 abstract statements, the British sample 343 specific and 178 abstract statements, and the Serbian sample 308 specific and 150 abstract statements. Participants of all nationalities generated more specific than abstract statements (all v2 > 39.27, df = 1, p < .001), and there were no significant effects of nationality (v2 = 0.524, df = 2, p = .77). We were interested in whether the ratings of possible selves would differ depending on whether the possible self generated was abstract (e.g. often associated with traits) or specific (e.g. associated with clear roles and social groups, such as being married, employed or a parent). In order to examine more clearly the effects of self-type (specific/abstract) on self rating scales, the data were analysed across all participants (collapsed across nationality).1 Table 4 shows the mean rating scores for all 1416 possible selves generated (of which 935 were coded as specific and 481 were abstract). Specific possible selves were significantly more vivid, less likely to feature the self alone, and dated further from the present, compared to abstract possible selves. There was a non-significant trend for specific possible selves to be rated more positively than abstract possible selves (p = .07). We were particularly interested in the idea that the temporal distribution of specific possible selves versus abstract possible selves might differ. To explore the data more closely, we plotted the temporal distributions of specific compared with abstract possible selves for each of the three nationalities and for the whole sample (see Fig. 1). Fig. 1 demonstrates a robust tendency for specific possible selves to be dated further from the present than abstract possible selves. Further evidence for this idea is provided by a breakdown of the percentage of possible selves that are abstract compared with specific in the first five year period from the present (e.g. years 0–4). Across the whole sample, 25.1% of specific possible selves were dated within the first five years, compared to 44.94% of the abstract possible selves. Thus, the five years following the present contains almost half of the entire set of abstract possible selves generated, but only a quarter of the specific possible selves. This pattern was replicated within each country (Turkey: 23.6% specific, 42.7% abstract; Serbia: 19.9% specific, 41.3% abstract; UK: 31.1% specific, 50% abstract). 1 These data were also analysed by nationality, through a series of 2 (self-type: abstract/specific)  3 (nationality: British, Serbian, Turkish) ANOVAs calculated for each of the possible self rating scales. These ANOVAs showed that the only significant interaction was between nationality and self-type on proportion alone (F[2, 1383] = 3.35, p = .005, partial g2 = .005). Thus, for clarity, results are presented on the whole dataset, rather than by nationality. 120 C.J. Rathbone et al. / Consciousness and Cognition 42 (2016) 113–124 Table 4 Mean phenomenological ratings and dates for specific and abstract possible selves. Scale Specific Mean SD 7.74 8.45 6.88 .49 .31 7.85 2.10 1.82 2.49 .50 .46 5.66 7.33 8.26 6.81 .52 .45 6.70 2.12 1.97 2.27 .50 .50 8.06 All nationalities 18 16 Specific Abstract 14 12 10 8 6 4 2 0 0 5 10 15 20 25 30 35 t SD Percentage of future images Percentage of future images Vividness Positivity Rehearsal Proportion field Proportion alone Distance from present Abstract Mean 40 45 50 55 60 65 70 75 80 Turkey Specific Abstract 14 12 10 8 6 4 2 0 5 10 15 Specific Abstract 14 12 10 8 6 4 2 5 10 15 20 25 30 35 40 45 50 55 60 65 70 Years between present and possible self 75 80 85 Percentage of future images Percentage of future images Serbia 0 20 25 30 35 40 45 50 55 60 65 70 75 80 85 Years between present and possible self 16 0 .001 .066 .598 .341 <.001 .006 16 Years between present and possible self 18 3.44 1.84 0.53 0.95 5.15 2.78 18 0 85 p United Kingdom 18 Specific Abstract 16 14 12 10 8 6 4 2 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 Years between present and possible self Fig. 1. Distribution of the distance of possible selves from present according to specific and abstract categories in the whole sample and within each cultural group. 3.4. Life script qualities of specific possible selves Finally, we were interested in the potential overlap between the concept of life script categories and specific possible selves (which by nature often refer to sociocultural transitions, such as becoming married, a parent or employed, see Rhee et al., 1995). To assess this, we examined the frequencies of different coding categories (e.g. parenthood, marriage etc.) according to whether they were coded as specific or abstract. Results showed that 93% of the abstract possible selves were categorized as ‘self-improvement, ‘happy’, or ‘successful’. In contrast, 53% of the specific possible selves were related to the three core life-script-related categories of marriage, occupation and parenthood. If we sum the findings for all specific selves, then we find that specific possible selves is the largest category for all cultures (e.g., 65% of the Turkish, 67% of the Serbian, and 66% of the British sample’s possible selves were specific). 4. Discussion This study explored the qualitative and quantitative features of possible future selves generated by young adults in Serbia, Turkey, and the United Kingdom. The first aim of this research was to compare the content of possible selves across cultures. We found several cross-cultural commonalties: the most commonly-generated category of possible self was selfimprovement, and participants from all three nationalities frequently generated possible selves relating to becoming married, parents and employed. However, there were also a number of cross-cultural differences. For example, the British sample generated significantly fewer possible selves that concerned self-improvement (compared with Turkish participants) and fewer specific occupations (compared to Turkish and Serbian participants). The British sample also generated significantly more marriage-related possible selves (compared with participants from Turkey) and significantly more possible selves associated with being happy (compared to the Turkish and Serbian participants). C.J. Rathbone et al. / Consciousness and Cognition 42 (2016) 113–124 121 We had predicted that the British participants would generate possible selves that were more autonomous and less focused on social roles, reflecting the individualism associated with the UK, compared to Serbia and Turkey (e.g. Hofstede, 1980). This prediction was not supported by results, which showed no significant differences between the proportions of social compared with autonomous possible selves across cultures. There was some support for this hypothesis in the finding that being ‘happy’ featured more frequently in the possible selves of British participants. However, this effect may simply reflect cross-cultural differences in the perceived importance of being happy, as previous work suggests that Western cultures may prioritize being happy to a greater extent than other cultures (e.g. Joshanloo & Weijers, 2014). Counter to predictions, we found that marriage featured more prominently in the possible selves of British than Turkish participants. This was somewhat surprising, as the UK is generally considered a less traditional and more individualistic culture compared to Turkey (e.g. Erdogan et al., 2008; Hofstede, 1980). There are three explanations for the lack of support for a traditional individualistic-collectivistic distinction in our results. The first possibility is that the participants sampled did not differ in terms of their collectivism-individualism. This may reflect the nature of our samples. In all three countries, participants were young and highly educated individuals studying at university. As such, they may have comprised a more homogeneous subset of each culture (e.g. Oyserman, Coon, & Kemmelmeier, 2002). In support of this view, Mishra (1994) found that Indian men who were younger, urban, and more educated tended to be less collectivistic than older men with less education living in rural areas. Further, the present findings may reflect the developmental stage of the participants in this study. Young adulthood is a period that is cross-culturally associated with identity formation (e.g. Erikson, 1950; Fitzgerald, 1988), and so there may be a developmental explanation for some of the cross-cultural similarities, such as the preponderance of possible selves focused on self improvement. The second possible explanation for the lack of individualistic-collectivistic cross-cultural differences is that there may have been genuine cross-cultural differences between groups but that these were overshadowed by more universal tendencies to view the future in terms of broader life goals concerning marriage, children, occupation, and financial security. One way of distinguishing between these two explanations would have been to use measures of collectivism and individualism alongside the possible selves task, in order to explore whether the predicted cross-cultural differences were present. However, many researchers have questioned the appropriateness of framing research around a reductionist view of individualis tic–collectivistic cultures. In a large-scale meta-analysis, Oyserman et al. (2002) reviewed evidence for the effects of individualism-collectivism on well-being, self-concept, cognition, and relationality. Although the authors found support for reliable cultural differences, these effects were far smaller and less systematic than commonly assumed. Other reviews have found little evidence for the individualistic–collectivistic distinction. For example, Takano and Osaka (1999) found that 14 out of 15 studies reviewed did not support the theory that the US is a more individualistic and less collectivistic culture than Japan. As concluded by Voronov and Singer (2002) perhaps a reductionist view of cultures pertaining to either a collectivist or individualist framework is simply inadequate. A third possibility is that our operationalization of individualistic versus collectivistic characteristics (such as the distinction between individual versus social selves) was not sufficiently sensitive to capture actual cultural differences. This possibility is supported by the fact that the UK participants did deviate from the Turkish and Serbian groups on a number of dimensions, as shown in Table 1. Related to this, it is possible that Hofstede’s (1980) classifications of Turkey and Serbia as collectivist cultures are not reflected in our samples. Following Hofstede’s seminal study, researchers interested in the individualistic/collectivistic orientations in different cultures arrived at findings irreconcilable with the original conceptualization of individualism/collectivism as a uni-dimensional, bi-polar construct (e.g. Imamoğlu, 1998, 2003; Oyserman et al., 2002). For instance, Imamoğlu (1998, 2003) showed that university students who predominantly come from better-educated, upper-middle class in Turkey showed an orientation towards individuation without an accompanying decrease in interrelatedness. Thus, she argued that individualism and collectivism are not opposite poles of a uni-dimensional continuum; rather, they are distinct, yet complementary, attributes. The Turkish sample recruited for the present study came from one of the most prestigious and competitive universities in Turkey. In addition, in terms of its education system and student culture, this particular Turkish university endorses the American system. Thus, it is possible that our Turkish sample endorsed Western values and lifestyles, were more inclined towards individuation, and thus were more focused on self-improvement in comparison with other segments of the Turkish society. Related to this idea, a fourth explanation is that the individualism/collectivism orientation no longer works as proposed 36 years ago in 1980. As several generations in Turkey and Serbia have grown up heavily exposed to Western culture, it is to be expected that these young adults would be more similar to individuals from a typical Western country than to young adults from more traditional cultures. Advances in communication technologies that have become widely available over the last 20 years have enabled young, educated, urban populations to be exposed to a vast range of information and influences. This factor may explain the cross-cultural similarities in our study, compared to what might have been expected 36 years ago. The second aim was to compare the phenomenological features of possible selves across cultures. In contrast to predictions, there was no difference across cultures between the proportion of possible selves that featured the self alone (i.e. less social images of the self in the future). However, there were significant cross-cultural differences in the ratings for vividness, positivity, and rehearsal. For example, although all participants rated their possible selves in a broadly positive light (at least 8/10 for positivity) demonstrating an optimistic view of their own future (e.g. Sedikides & Gregg, 2008) the Serbian sample’s possible selves were rated as more positive compared to the British and Turkish samples. The Serbian participants also rated their possible selves as more vivid (compared to the British participants), and the Turkish participants rated their possible 122 C.J. Rathbone et al. / Consciousness and Cognition 42 (2016) 113–124 selves as more frequently rehearsed (compared to the British participants). In support of our prediction, all participants dated their possible selves as emerging at a mean of six to eight years from the present. This replicates previous studies by Chessell et al. (2014) and Rathbone et al. (2011) and suggests that young adults may only focus on the relatively near future when considering the type of person they are likely to become in the future. This finding broadly echoes results from Conway et al. (2005), who showed that whilst the reminiscence bump occurs at approximately the same age across cultures, the content of the memories within the reminiscence bump is subject to cross-cultural variation. In a similar way, although we found a number of broad similarities between cultures (such as proportion of social compared to specific statements and the age at which possible selves are dated), we also found a number of fine-grained cross-cultural differences (e.g. differences in the frequencies of particular categories of possible selves, such as ‘happiness’, and ratings of the positivity and vividness of possible selves). Our third aim was to compare the phenomenological features of specific and abstract possible selves. We were particularly motivated to examine the features of these two types of future cognition as specific possible selves are considered to be more goal-related and, consequently, potentially more likely to influence behaviour (e.g. Hoppmann et al., 2007; Hoyle & Sherrill, 2006; Oyserman et al., 2006). We predicted that specific possible selves would be more vivid, more rehearsed, and dated closer to the present, compared to abstract possible selves. These predictions were partially supported, with specific possible selves rated as more vivid than abstract possible selves. This was unsurprising, as specific possible selves (associated with roles such as parenthood and occupation, or containing specific details such as ‘working with animals’) are likely to be associated with clearer and more vivid images than abstract possible selves (which, by definition, are not associated with specific details, but instead with emotional states or traits). There were no differences in rehearsal ratings. In contrast to predictions, specific possible selves were dated further from the present, compared to abstract possible selves. Although this result appears counter to construal level theory, which posits that events closer to the present will be more specific than those that are distant (Trope & Liberman, 2010), this finding may be explained by the content of specific and abstract possible selves. Specific possible selves tended to map onto culturally normative events, such as marriage and parenthood – events that are highly goal-relevant. Demblon and D’Argembeau (2014) found that thoughts about the distant future tended to be more organised around personal goals than thoughts about the near future. This is in keeping with our finding that specific possible selves (predominantly reflecting long-term personal goals about relationships and occupations) are projected further into the future than abstract (traitlinked) possible selves. Another possibility is that trait-linked, abstract possible selves comprised more stable – and potentially already present – aspects of identity. Although participants were instructed not to generate any possible selves that described themselves at present, it is possible that there was some degree of overlap between how participants imagined themselves in the future and how they perceived themselves in the present. Such stable or overlapping selves would be more likely to be dated as emerging in the near future. To examine this possibility we suggest that future studies incorporate a rating scale of overlap between future self and current self. An alternative explanation for the nearness of abstract compared to specific possible selves may be found in temporal self-appraisal theory (e.g. Ross & Wilson, 2002). Temporal self-appraisal theory posits that events that cast the self in a positive light tend to be dated closer to the present, in contrast to events that involve more negative self-appraisal, and that this bias exists to enhance a positive view of the self. As 93% of the abstract possible selves concerned self-improvement, it is possible that motivation to enhance the present self (as predicted by temporal self-appraisal theory) caused these overly positive identities to be dated closer to the present (although note that there was no significant difference in the positivity ratings for abstract compared to specific possible selves). Finally, it is possible that the coding of specific and abstract possible selves does not map well onto the idea of specific and abstract construals, as operationalised in construal level theory. One of the more specific aims of this research was to generate a possible selves coding scheme that could be used in future studies to explore commonalities and differences in the ways people think about themselves in the future. As discussed above, cases where there were similarities across cultures may have reflected the homogenous nature of young, university educated participants, as well as a blurring of collectivist/individualist boundaries across the cultures sampled. These results may also reflect the stage in life our participants were at. Thus, it would be fruitful to compare the norms generated in the present study with the possible selves of older adults in each culture. For example, it is possible that ‘self-improvement’ was a central goal for young adults in this study as they were all engaged with university study, presumably with the aims of becoming better educated and more employable. Furthermore, it would be interesting to compare these possible selves norms with the possible selves of young and old adults in more traditionally collectivist cultures such as China. This may reveal a wider range of cross-cultural differences than found in the present study. It is important to note that the present study used categories derived from the coding process to analyse the data in this study. As there are psychometric limitations in taking this approach, we acknowledge that the full benefits of this coding scheme will be recognized when applying it to new data sets in future studies, such as those outlined above. Finally, future work could explore the robustness of the distribution of possible selves found in the present study and in previous work (e.g. Chessell et al., 2014; Rathbone et al., 2011). Whilst young adults have a tendency to date possible selves as developing 6–8 years in the future, Chessell et al. (2014) found that older adults dated possible selves closer to the present (a mean of 2.6 years in the future). We know little about the mechanism behind the dating of possible selves; it may reflect differences in cultural life scripts for different age groups or it may be a function of age-related changes in future time perspective (Spreng & Levine, 2006). Future work with adults of different ages will elucidate how the temporal distribution of possible selves might change with age. C.J. Rathbone et al. / Consciousness and Cognition 42 (2016) 113–124 123 In conclusion, this study explored the way young adults from different cultures think about who they might become in the future. Although all participants tended to picture themselves in the future using a similar time frame and through reference to specific, positive, culturally normative categories, we found a small number of differences in the content and phenomenological features of possible selves between cultures. Appendix A. Possible self coding scheme Category Explanation & examples Acquiring property Activities Home-owner etc. Gardener, traveller, charity volunteer (activities that do not require ‘learning’ – these would be classed as skill development) Relating to getting older Relating to university, degrees, graduating Aging Educationuniversity Fall in love Family Finance selfimprovement Friendship Grandchildren Happy Health Busy Possessions Move Job-general Job-specific Learn languages Marriage Military Other Parenthood Physical appearance Relationship Religion Self-improvement Skill development Successful Trait-general In love, fall in love Family used for being a daughter or son, aunt or uncle, or about family in general Relating to being wealthy, rich, well-paid, anything associated with money Relating to being friends Grandmother, grandfather, grandparent Has own category (even though fits into trait – self-improvement – should be coded separately) To do with fitness, illness (mental and physical), health in general Has own category as P2 occurrences Anything owned that isn’t property (which is coded as ‘acquiring property’) or job-related (e.g. ‘the owner of a cafe’ would be coded as job-specific) Relating to moving house or country If job is discussed in broad terms (e.g. do a job I enjoy, be good at my job) If specific details about job/occupation are given (e.g. be a psychologist, writer, doctor) Although skill development this is a separate category to allow cross-cultural comparisons Partner, husband, wife, married Relates to military service Response does not fit into any other category Father, mother, parent Relating to looks, weight, attractiveness Relationship used for boyfriend or girlfriend, if not explicitly about marriage (partner is classed as marriage) Any mention of religion Abstract statements relating to being better in the future (e.g. clever, a better person, academic) Skill development if the future self involves learning (e.g. driving a car, cooking, dancing), but ‘activity’ if not (e.g. giving to charity, going travelling) Has own category (even though fits into trait – self-improvement – should be coded separately) If not associated with self-improvement (e.g. negative or neutral traits, such as lazy, stressed, different) References Aloise-Young, P. A., Hennigan, K. M., & Leong, C. W. (2001). Possible selves and negative health behaviors during early adolescence. The Journal of Early Adolescence, 21(2), 158–181. Berntsen, D., & Bohn, A. (2010). Remembering and forecasting: The relation between autobiographical memory and episodic future thinking. Memory & Cognition, 38(3), 265–278. Berntsen, D., & Rubin, D. C. (2004). Cultural life scripts structure recall from autobiographical memory. Memory & Cognition, 32(3), 427–442. Chessell, Z., Rathbone, C. J., Souchay, C., Charlesworth, L., & Moulin, C. J. A. (2014). Autobiographical memory, past and future events and self-images in younger and older adults. Self and Identity, 13(4), 380–397. Conway, M. A., & Pleydell-Pearce, C. W. (2000). The construction of autobiographical memories in the self-memory system. Psychological Review, 107(2), 261–288. Conway, M. A., Wang, Q., Hanyu, K., & Haque, S. (2005). A cross-cultural investigation of autobiographical memory: On the universality and cultural variation of the reminiscence bump. Journal of Cross-Cultural Psychology, 36(6), 739–749. Corte, C., & Szalacha, L. (2010). Self-cognitions, risk factors for alcohol problems, and drinking in preadolescent urban youths. Journal of Child & Adolescent Substance Abuse, 19(5), 406–423. Demblon, J., & D’Argembeau, A. (2014). The organization of prospective thinking: Evidence of event clusters in freely generated future thoughts. Consciousness & Cognition, 24, 75–83. 124 C.J. Rathbone et al. / Consciousness and Cognition 42 (2016) 113–124 Erdogan, A., Baran, B., Avlar, B., Tas, A. C., & Tekcan, A. (2008). On the persistence of positive events in life scripts. Applied Cognitive Psychology, 22, 95–111. Erikson, E. H. (1950). Childhood and society.New York: W.W. Norton & Company. Fitzgerald, J. M. (1988). Vivid memories and the reminiscence phenomenon: The role of a self narrative. Human Development, 31, 261–273. Hofstede, G. H. (1980). Culture’s consequences: International differences in work-related values. London: Sage Publications. Hooker, K., & Kaus, C. R. (1992). Possible selves and health behaviors in later life. Journal of Aging and Health, 4(3), 390–411. Hoppmann, C. A., Gerstorf, D., Smith, J., & Klumb, P. L. (2007). Linking possible selves and behavior: Do domain-specific hopes and fears translate into daily activities in very old age? The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 62(2), 104–111. Hoyle, R. H., & Sherrill, M. R. (2006). Future orientation in the self-system: Possible selves, self-regulation, and behavior. Journal of Personality, 74, 1673–1696. Imamoğlu, E. O. (1998). Individualism and collectivism in a model and scale of balanced differentiation and integration. The Journal of Psychology, 132(1), 95–105. Imamoğlu, E. O. (2003). Individuation and relatedness: Not opposing but distinct and complementary. Genetic, Social, and General Psychology Monographs, 129(4), 367–402. Janssen, S., Uemiya, A., & Naka, M. (2014). Age and gender effects in the cultural life script of Japanese adults. Journal of Cognitive Psychology, 26, 307–321. Jobson, L., & O’Kearney, R. (2008). Cultural differences in retrieval of self-defining memories. Journal of Cross-Cultural Psychology, 39(1), 75–80. Joshanloo, M., & Weijers, D. (2014). Aversion to happiness across cultures: A review of where and why people are averse to happiness. Journal of Happiness Studies, 15(3), 717–735. Lee, C. K., Corte, C., Stein, K. F., Park, C. G., Finnegan, L., & McCreary, L. L. (2015). Prospective effects of possible selves on alcohol consumption in adolescents. Research in Nursing & Health, 38(1), 71–81. Markus, H. R., & Kitayama, S. (1991). Culture and the self – Implications for cognition, emotion, and motivation. Psychological Review, 98(2), 224–253. Markus, H. R., & Nurius, P. (1986). Possible selves. American Psychologist, 41(9), 954–969. Mishra, R. C. (1994). Individualist and collectivist orientations across generations. In U. Kim, H. C. Triandis, C. Kagitcibasi, S.-C. Choi, & G. Yoon (Eds.), Individualism and collectivism: Theory, method, and applications (pp. 225–238). Thousand Oaks, CA: Sage. Ottsen, C. L., & Berntsen, D. (2014). The cultural life script of Qatar and across cultures: Effects of gender and religion. Memory, 22, 390–407. Oyserman, D., Bybee, D., & Terry, K. (2006). Possible selves and academic outcomes: How and when possible selves impel action. Journal of Personality and Social Psychology, 91(1), 188–204. Oyserman, D., Bybee, D., Terry, K., & Hart-Johnson, T. (2004). Possible selves as roadmaps. Journal of Research in Personality, 38(2), 130–149. Oyserman, D., Coon, H. M., & Kemmelmeier, M. (2002). Rethinking individualism and collectivism: Evaluation of theoretical assumptions and metaanalyses. Psychological Bulletin, 128(1), 3–72. Oyserman, D., & Fryberg, S. A. (2006). The possible selves of diverse adolescents: Content and function across gender, race, and national origin. In C. Dunkel & J. Kerpelman (Eds.), Possible selves: Theory, research, and application (pp. 17–39). Huntington, NY: Nova. Oyserman, D., Gant, L., & Ager, J. (1995). A socially contextualized model of African American identity: Possible selves and school persistence. Journal of Personality and Social Psychology, 69(6), 1216–1232. Oyserman, D., & Markus, H. (1990). Possible selves and delinquency. Journal of Personality and Social Psychology, 59(1), 112–125. Rathbone, C. J., Conway, M. A., & Moulin, C. J. A. (2011). Remembering and imagining: The role of the self. Consciousness & Cognition, 20(4), 1175–1182. Rhee, E., Uleman, J. S., Lee, H. K., & Roman, R. J. (1995). Spontaneous self descriptions and ethnic identities in individualist and collective cultures. Journal of Personality and Social Psychology, 69, 142–152. Ross, M., & Wilson, A. E. (2002). It feels like yesterday: Self-esteem, valence of personal past experiences, and judgments of subjective distance. Journal of Personality and Social Psychology, 82(5), 792–803. Rubin, D. C., & Berntsen, D. (2003). Life scripts help to maintain autobiographical memories of highly positive, but not highly negative, events. Memory and Cognition, 31(1), 1–14. Rubin, D. C., Berntsen, D., & Hutson, M. (2009). The normative and the personal life: Individual differences in life scripts and life story events among USA and Danish undergraduates. Memory, 17, 54–68. Rubin, D. C., Wetzler, S. E., & Nebes, R. D. (1986). Autobiographical memory across the adult lifespan. In D. C. Rubin (Ed.), Autobiographical memory (pp. 202–221). Cambridge: Cambridge University Press. Sedikides, C., & Gregg, A. P. (2008). Self-enhancement: Food for thought. Perspectives on Psychological Science, 3(2), 102–116. Senior, K. A., & Chenhall, R. D. (2012). Boyfriends, babies and basketball: Present lives and future aspirations of young women in a remote Australian Aboriginal community. Journal of Youth Studies, 15(3), 369–388. Spreng, R. N., & Levine, B. (2006). The temporal distribution of past and future autobiographical events across the lifespan. Memory & Cognition, 34(8), 1644–1651. Szpunar, K. K., Spreng, R. N., & Schacter, D. L. (2014). A taxonomy of prospection: Introducing an organizational framework for future-oriented cognition. Proceedings of the National Academy of Sciences of the United States of America, 111, 18414–18421. Takano, Y., & Osaka, E. (1999). An unsupported common view: Comparing Japan and the US on individualism–collectivism. Asian Journal of Social Psychology, 2, 311–341. Trope, Y., & Liberman, N. (2010). Construal-level theory of psychological distance. Psychological Review, 117, 440–463. vanDellen, M. R., & Hoyle, R. H. (2008). Possible selves as behavioral standards in self-regulation. Self and Identity, 7(3), 295–304. Voronov, M., & Singer, J. A. (2002). The myth of individualism-collectivism: A critical review. Journal of Social Psychology, 142(4), 461–480. Waid, L. D., & Frazier, L. D. (2003). Cultural differences in possible selves during later life. Journal of Aging Studies, 17(3), 251–268. Wang, Q. (2001). Culture effects on adults’ earliest childhood recollection and self-description: Implications for the relation between memory and the self. Journal of Personality and Social Psychology, 81(2), 220–233. Wang, Q. (2004). The emergence of cultural self-constructs: Autobiographical memory and self-description in European American and Chinese children. Developmental Psychology, 40(1), 3–15. Wang, Q. (2006). Earliest recollections of self and others in European American and Taiwanese young adults. Psychological Science, 17(8), 708–714. Wang, Q., & Conway, M. A. (2004). The stories we keep: Autobiographical memory in American and Chinese middle-aged adults. Journal of Personality, 72(5), 911–938. Yowell, C. M. (2000). Possible selves and future orientation: Exploring hopes and fears of latino boys and girls. The Journal of Early Adolescence, 20(3), 245–280. Zaragoza Scherman, A. (2013). Cultural life script theory and the reminiscence bump: A reanalysis of seven studies across cultures. Nordic Psychology, 65(2), 103–119. Zaragoza Scherman, A., Salgado, S., Shao, Z., & Berntsen, D. (2014). Event centrality of positive and negative autobiographical memories to identity and life story across cultures. Memory. http://dx.doi.org/10.1080/09658211.2014.962997. Advance online publication. Zaragoza Scherman, A., Salgado, S., Shao, Z., & Berntsen, D. (2015). Life span distribution and content of positive and negative autobiographical memories across cultures. Psychology of Consciousness: Theory, Research, and Practice, 2(4), 475–489.

Consciousness and Cognition 43 (2016) 48–56 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Not sensitive, yet less biased: A signal detection theory perspective on mindfulness, attention, and recognition memory Eyal Rosenstreich a,⇑, Lital Ruderman b a b Department of Behavioral Sciences, Peres Academic Center, Rehovot, Israel Section of Comparative Medicine, Yale School of Medicine, New Haven, CT 06510, USA a r t i c l e i n f o Article history: Received 22 April 2015 Revised 10 April 2016 Accepted 16 May 2016 Keywords: Mindfulness Recognition memory Divided attention Sensitivity Response bias a b s t r a c t The practice of mindfulness has been argued to increase attention control and improve memory performance. However, it was recently suggested that the effect of mindfulness on memory may be due to a shift in response-bias, rather than to an increase in memory-sensitivity. The present study examined the mindfulness-attention-memory triad. Participants filled in the five-facets of mindfulness questionnaire, and completed two recognition blocks; in the first attention was full, whereas in the second attention was divided during the encoding of information. It was found that the facet of non-judging (NJ) moderated the impact of attention on memory, such that responses of high NJ participants were less biased and remained constant even when attention was divided. Facets of mindfulness were not associated with memory sensitivity. These findings suggest that mindfulness may affect memory through decision making processes, rather than through directing attentional resources to the encoding of information. Ó 2016 Elsevier Inc. All rights reserved. 1. Introduction Mindfulness is an ongoing awareness of the present moment, emotions, and state of mind (Kabat-Zinn, 2003). Mindfulness practice was theorized to promote de-automatization, that is, to reduce the extent to which behavior is based upon ‘‘non-thinking” (Kang, Gruber, & Gray, 2013). Specifically, Kang et al. suggested that mindfulness affects behavior through four factors: Attention (an increase in attentional control and cognitive flexibility), awareness (which reduces automatic processing), being at the present moment (which promotes de-centering), and non-judgmental acceptance (embracing thoughts and feelings as they are, without attempting to suppress them). In line with Kang et al.’s (2013) model, the practice of mindfulness has been demonstrated to improve the management of attentional resources (e.g., Morrison, Goolsarran, Rogers, & Jha, 2013; Tang et al., 2007). For example, Jensen, Vangkilde, Frokjaer, and Hasselbalch (2012) compared a mindfulness practice group to control groups which received a nonmindfulness relaxation practice or a financial incentive in various tasks of attention (e.g., to determine whether a target was presented in gray or white, or to find and cross a given target letter within a large set of letters). The results indicated that mindfulness practice was associated with a more stable performance in selective attention tasks, manifested in low reaction time variability and stable error rate across trials (see also Galla, Hale, Shrestha, Loo, & Smalley, 2012; Ruocco & Direkoglu, 2013). ⇑ Corresponding author at: Department of Behavioral Sciences, Peres Academic Center, 10 Peres St., Rehovot, Israel. E-mail address: eyal@pac.ac.il (E. Rosenstreich). http://dx.doi.org/10.1016/j.concog.2016.05.007 1053-8100/Ó 2016 Elsevier Inc. All rights reserved. E. Rosenstreich, L. Ruderman / Consciousness and Cognition 43 (2016) 48–56 49 It seems, therefore, that mindfulness may improve the use of attentional resources. The current study further investigated the connection between mindfulness and attentional demands through the scope of memory performance. Memory performance has been argued to be very sensitive to deficit of attentional resources. In particular, dividing participants’ attention during the encoding of words reduced the amount of words recollected in a later memory test (e.g., Knott & Dewhurst, 2007; Naveh-Benjamin, Craik, Gavrilescu, & Anderson, 2000). It has therefore been suggested recently that memory performance may benefit from mindfulness practice (Rosenstreich, 2014). That is, mindfulness may improve memory performance by increasing the availability of attentional resources during encoding. Indeed, studies which examined the effects of mindfulness on memory typically showed that mindfulness was associated with increased memory performance (for a review, see Rosenstreich, 2014). One study, conducted by Alberts and Thewissen (2011), presented mindfulness trainees and controls with a to-be-remembered list of 30 positive, negative, and neutral words (ten words of each type). After a retention interval, participants were asked to recall as many studied words as they could. It was found that although mindfulness practitioners did not recall more words in general when compared to controls, they recalled significantly fewer negative words (see also van Vugt, Hitchcock, Shahar, & Britton, 2012). In another study, Lykins, Baer, and Gottlob (2012), presented the California Verbal Learning Test (CVLT) to matched groups of meditators and non-meditators. The CVLT is a common diagnostic memory test, consisting of two study lists with 32 neutral words (16 in each). Each of the lists is presented five times during the study stage. Immediate and delayed memory tests revealed that mindfulness meditators recalled more words than non-meditators. However, this effect diminished when retrieval was cued (i.e., cued-recall test). Because retrieval in free recall tests is typically more difficult than in cued recall test (where the retrieval cues may make memory more accessible) (cf. Carpenter, Pashler, & Vul, 2006), this finding may indicate that mindfulness improves memory accessibility when a retrieval cue is absent. Nevertheless, the findings described above show that in order to fully understand the connection between mindfulness and memory, it is crucial to employ different memory tests. Along with this notion, the present study will focus on a scarcely investigated memory test in the mindfulness literature, recognition memory. The potential contribution of recognition tests in elucidating the mindfulness-memory connection has been demonstrated recently (Rosenstreich, 2015). A short mindfulness practice increased the correct recognition of studied words, but at the same time also increased the rate of incorrect recognition. That is, although mindfulness practitioners correctly remembered more words than randomized control participants who engaged in a mind-wandering workshop, they were also more prone to memory distortions. Moreover, the application of a recognition test enables the assessment of two informative measures of memory performance: memory sensitivity and response bias. Derived from Signal Detection Theory (SDT), sensitivity represents the ability to discriminate between studied (hits) and unstudied items (false alarms), whereas response bias represents a participant’s tendency to respond ‘‘studied” or ‘‘unstudied” during a recognition test, regardless of his level of memory performance (for further details, see Macmillan & Creelman, 2005; Rotello, Masson, & Verde, 2008). Incorporating the assessment of sensitivity and response bias within a study of mindfulness may deepen our understanding of the mindfulness-memory association. In particular, an improvement in memory performance could be either due to an increase in sensitivity (i.e., improved ability to discriminate between targets and foils), or due to a change in response patterns (i.e., increased tendency to judge target and foils as targets). A recent study employing these two measures revealed that the increased hit rate observed after mindfulness practice was due to response bias rather than increased sensitivity (Rosenstreich, 2015). That is, whereas control participants tended to favor the ‘‘unstudied” response, mindfulness practitioners were less biased and tended to favor both ‘‘studied” and ‘‘unstudied” responses in a similar proportion. Furthermore, both groups did not differ in their sensitivity. Nevertheless, two questions remain open following that study (Rosenstreich, 2015) regarding the connection between mindfulness and memory sensitivity and response bias. First, the study was aimed to investigate the impact of mindfulness practice on false memories, that is, to experimentally provoke memory distortions. It is therefore not clear whether the effect observed on response bias and the null effect observed on sensitivity were a result of the specific experimental design, or rather represent the underlying mechanisms of mindfulness. It remains to be seen what happens in the absence of such a memory distortion. The second question arises from the way mindfulness was operationalized. As described earlier, mindfulness is theorized as an ongoing awareness of the present moment. This theoretical construct could be either promoted by an intervention (for a review of mindfulness-based interventions, see Shapiro & Carlson, 2009), or measured as a predisposed trait. Employing a mindfulness intervention, as in Rosenstreich (2015), enables conclusions on causal connections between mindfulness and various variables, while measuring mindfulness as a trait may provide insights regarding the underlying factors of mindfulness. Specifically, the construct of mindfulness was argued to consist of five facets: (1) Non-Reacting: the ability to withhold reaction; (2) Observing: the ability to observe and direct attention; (3) Awareness: the ability to act with awareness; (4) Describing: the ability to describe thoughts and feelings; and (5) Non-Judgment: the ability to act without judging the self and others (Baer, Smith, Hopkins, Krietemeyer, & Toney, 2006). It was recently suggested that the facets of awareness and acceptance (non-judgment) play a significant role in attention and memory, respectively (Ruocco & Direkoglu, 2013). Specifically, Ruocco & Direkoglu showed that the facet of awareness was associated with improved performance in sustained attention tasks, whereas acceptance was associated with improved performance in working memory tasks, Hence, in order to better understand the connection between mindfulness, memory sensitivity, and response bias, mindfulness should not only be induced; rather, facets of mindfulness should first be assessed. 50 E. Rosenstreich, L. Ruderman / Consciousness and Cognition 43 (2016) 48–56 The present study was designed to further investigate the connection between mindfulness and recognition memory by addressing the questions which arose from Rosenstreich (2015). Specifically, it was focused on analyzing memory performance in terms of hits, false alarms, sensitivity, and response bias, and their connection to the assessment of the five facets of mindfulness (Baer et al., 2006). Furthermore, because mindfulness is intimately related to the availability and management of attentional resources (e.g., Tang et al., 2007), and because memory was demonstrated to be impaired by the reduction of attentional resources availability (Knott & Dewhurst, 2007; Maddox, Naveh-Benjamin, Old, & Kilb, 2012; Naveh-Benjamin et al., 2000). This study is mainly aimed at the investigation of the connection between mindfulness and memory in the context of attentional resource deficiency. To this end, two recognition tasks were carried out: the first was conducted while participants were fully attentive to the recognition task. The second recognition task was conducted while participants were also engaged in an auditory categorization task (i.e., secondary task), in which they were asked to classify auditory tones into three categories based on their pitch (low, medium, or high). This tone classification task was has been demonstrated to be highly demanding of attentional resources, and to impair the encoding of information (Craik, Naveh-Benjamin, Ishaik, & Anderson, 2000; Gardiner & Parkin, 1990; Naveh-Benjamin, Guez, Kilb, & Reedy, 2004). Participants’ rates of hits, false alarms, sensitivity, and response-bias were assessed. It was hypothesized that mindfulness would moderate the impact of attention on memory hits; in particular, it was predicted that mindfulness scores would be positively correlated with hit rate, with a stronger correlation under divided than under full attention. No specific predictions were made for false alarms, although it was demonstrated recently that recognition false alarms may be reduced after mindfulness practice (Lloyd, Szani, Rubenstein, Colgary, & Pereira-Pasarin, 2016). More important, following Rosenstreich (2015), it was hypothesized that facets of mindfulness would be associated with response bias, but not with sensitivity. Specifically, under full attention, mindful participants were predicted to be less biased in their recognition judgments than less-mindful participants, whereas both groups were not predicted to differ in their ability to discriminate between studied and unstudied words. Finally, as in the first hypothesis, response bias was predicted to be less affected by divided attention among participant with high, as compared to low, mindfulness scores. Due to the pioneer nature of this study, no prediction was made regarding the association between memory and any specific facet of mindfulness. 2. Method 2.1. Participants Sixty-eight first year psychology students at Tel-Aviv University participated in this study in exchange for course credit (mean age = 23.55, SD = 3.72; 50 females). Data from two participants were excluded from analyses due to their expressed difficulty in understanding a large portion of the mindfulness questionnaire. Thus, the study consisted of a sample size of 66 participants. The study was approved by the university ethics committee, and participants gave their written consent of willingness to participate in the study. 2.2. Design The study was correlative, with self report of mindfulness levels and attention (full, divided at encoding) as independent variables. Attention was manipulated within-subjects. The dependent variable was recognition memory performance. 2.3. Materials 2.3.1. Mindfulness Dispositional mindfulness was assessed using a shortened form of the ‘‘five facets of mindfulness questionnaire” (FFMQ-SF; Bohlmeijer, Peter, Fledderus, Veehof, & Baer, 2011). The FFMQ-SF is aimed at assessing the five facets of dispositional mindfulness (Baer et al., 2006), which are as follows: non-reactivity (the ability to withhold thoughts, emotional expressions, and physical actions), ability to observe (the direction of attention to the surroundings), acting with awareness (being aware of inner and outer sensations and stimuli), ability to describe (the ability to verbalize thoughts and feelings), and being non-judgmental (accepting the self and the others as they are). Its 24 statements (e.g., I notice the smells and aromas of things) were rated on a 6-point Likert scale (1 - never or very rarely true, 6 - very often or always true). Items of two facets, awareness and being non-judgmental, were presented in reverse scales and were transformed before the analysis of the data. The FFMQ-SF was translated into Hebrew and back into English by the author, and a professional language editor then compared the original and translated English versions. In the current study, the Hebrew version of the FFMQ-SF was found reliable, with Cronbach’s alphas of 0.67, 0.75, 0.81, 0.85, and 0.73 in the non-reactivity, observe, awareness, describe, and non-judgmental factors, respectively. E. Rosenstreich, L. Ruderman / Consciousness and Cognition 43 (2016) 48–56 51 2.3.2. Attention The secondary task in the study was designed to divide participants’ attention using a tone classification task (Gardiner & Parkin, 1990). This task was comprised of three auditory tones which differed in their frequencies: 220 Hz (low), 440 Hz (medium), and 880 Hz (high). These tones were generated using AudacityÒ audio software (http://www.audacityteam. org), and were 550 ms long. In this task, participants were requested to classify the tones as ‘‘Low”, ‘‘Medium”, and ‘‘High”, by pressing designated keys on a USB numeric keypad connected to a computer. The tones were presented randomly, with a 300 ms inter-stimulus interval (ISI). If a given tone was not classified within the 850 ms interval (tone onset plus ISI), a new tone was presented. Therefore, the secondary task was rapid, and demanded the participants to continuously focus their attention, thus limiting the attention allotted for the primary task. 2.3.3. Memory The primary task in the study was designed to assess memory performance. Specifically, memory was assessed using 80 Hebrew nouns which served as target words, with four additional words serving as buffers. All words were presented in black Arial font (bold), size 60. The 80 words were randomly assigned to two study and test blocks of 40 words. In each block, the study list consisted of 20 words which were presented randomly at a rate of 1000 ms per word. Each word was preceded with a ‘+’ sign presented for 250 ms, followed by a blank screen for 100 ms. The offset of the target word was followed by a blank screen for 100 ms. The test list consisted of 40 randomly presented words, half of which were studied, while the other half were new and unstudied. Participants were instructed to determine whether each test word was studied or not studied. These recognition judgments were conducted by pressing designated buttons on a computer keyboard, marked with red (the L key: ‘‘not studied”) and blue (the A key: ‘‘studied”) stickers. If a word was judged as studied, then participants were prompted to provide a ‘‘Remember-Know-Guess” judgment (Gardiner, Ramponi, & Richardson-Klavehn, 1998). The Remember-KnowGuess paradigm was employed as part of a different research program, which is reported elsewhere (Rosenstreich & Goshen-Gottstein, 2015); thus, these judgments are not within the scope of this article. The two 40-word blocks were counterbalanced across the two attention conditions, so that each block appeared an equal number of times under the full and divided-attention conditions. 2.4. Procedure Participants were first introduced to the structure of the experiment. They were informed that response accuracy was very important for both the memory and tone classification tasks, and that speed was also essential during the tone classification task. After a general practice during which the recognition and tone classification tasks were introduced, the experiment began (see Fig. 1 for a schematic overview). The experiment consisted of two separate blocks. In the first block, participants studied the list of words and performed a recognition test under full attention. In the second block, participants studied the list of words while performing the tone classification task. That is, here words were encoded under divided attention but retrieved under full attention. Each block began with a reminder of the instructions for the forthcoming task, as well with a short practice that simulated the task. In addition, in both blocks, the tone classification task was performed for 30 s after the study stage, but before the test, thus serving as a filler intended to prevent participants from memorizing the study list. Finally, at the completion of the second memory test, participants were asked to complete the mindfulness questionnaire using an online survey platform. Each participant performed the study alone on a computer connected to a 1700 CRT monitor. A research assistant was sitting nearby in order to operate the experiment, and provided assistance and clarification if needed. Fig. 1. A schematic overview of the design. 52 E. Rosenstreich, L. Ruderman / Consciousness and Cognition 43 (2016) 48–56 3. Results 3.1. Preliminary data preparation 3.1.1. Mindfulness Total scores were calculated for the five facets of mindfulness. These scores were calculated as the sum of the ratings in each of the non-reactivity (NR), observe, awareness, describe, and non-judgmental (NJ) facets. 3.1.2. Memory (primary task) performance Overall memory proportions were calculated for each participant. Specifically, the following four scores were calculated: proportion of correct responses (hits) under full attention, proportion of false positive responses (false alarms; FAs) under full attention, proportion of hits under divided attention, and proportion of FAs under divided attention. The proportions of hits were calculated by dividing the number of studied words judged as ‘‘studied” by 20 (the total number of studied words in each block). The proportions of FAs were calculated by dividing the number of unstudied words judged as ‘‘studied” by 20 (the total number of unstudied words in each block). In addition, memory performance was also assessed using two measures derived from signal detection theory (SDT), namely, sensitivity and response bias. Sensitivity is calculated as the distance between the hit and the FA distributions, whereas response bias represents the location of the participants’ decision criterion (for rationale and other measures of sensitivity, see Rotello et al., 2008). Specifically, sensitivity was calculated as follows: Sensitiv ity ¼ zðhitsÞ  zðFAsÞ ð1Þ In this equation, sensitivity scores larger than zero indicated that the participant successfully discriminated between studied and unstudied words. Next, response bias was calculated as follows: Response bias ¼ zðhitsÞ þ zðFAsÞ 2 ð2Þ Here, positive scores represented the adoption of a relaxed decision criterion, that is, an elevated tendency to respond ‘‘studied” regardless of the item’s actual status. Negative scores represented a more conservative decision criterion, which manifests in a decreased tendency to respond ‘‘studied”. A score of zero represented a balanced criterion. Sensitivity and response bias are assumed to be independent (Rotello et al., 2008), and indeed were not correlated in this study, r = 0.047, p = 0.709. 3.1.3. Secondary task performance Finally, four scores were calculated for each participant in the tone classification task: accuracy and mean reaction time (RT) for correct tone judgments under full attention, and accuracy and mean RT under divided attention. Accuracy and mean RT were calculated after extreme RTs (±2SD) were excluded (mean percent of extremes = 3.13%, SD = 1.50). Therefore, accuracy was calculated as the number of non-extreme correct tone identifications divided by the total number of non-extreme trials. Due to technical problems, data from two participants was not registered. Hence, analyses of the secondary task data were conducted with N = 64. Accuracy and mean RT were not correlated in this study, r = 0.001, p = 0.994, thus indicating that performance in the tone classification task was not due to speed-accuracy tradeoff. Descriptive statistics and Pearson correlations between the facets of mindfulness and the memory and tone classification tasks are presented in Table 1. First, in order to confirm that our data correspond with previous studies, the correlation between mindfulness and secondary task performance was examined. Specifically, Table 1 reveals that the facet of awareness was positively correlated with accuracy in the tone classification task, but only when attention was divided. That is, when both tasks were engaged, participants with high awareness scores performed more accurately in the secondary task than participants with low awareness scores. This effect corresponds with previous findings regarding the association between mindfulness and attention (e.g., Ruocco & Direkoglu, 2013). In addition, paired-samples t-tests revealed that, as compared to performing under full Table 1 Means, SDs, and Pearson correlation coefficients between the five facets of mindfulness and the performance in the memory and tone classification tasks. Memory (main) task (N = 66) Non-reactivity Observe Awareness Describe Non-judgmental Mean (SD) 16.29 (3.84) 16.11 (3.97) 21.67 (4.58) 22.64 (4.68) 20.15 (4.38) Note. FAs – false alarms. * p < 0.05. Tone classification (secondary) task (N = 64) Full attention Divided attention Full attention Hits FAs Hits FAs RT Accuracy RT Divided attention Accuracy 0.88 (0.13) 0.032 0.099 0.158 0.071 0.202 0.22 (0.17) 0.167 0.042 0.046 0.009 0.259* 0.65 (0.16) 0.124 0.168 0.134 0.132 0.161 0.28 (0.15) 0.129 0.130 0.062 0.010 0.064 187.40 (61.69) 0.109 0.066 0.045 0.125 0.029 71.74 (16.17) 0.015 0.110 0.055 0.079 0.064 234.34 (66.31) 0.145 0.195 0.048 0.160 0.049 63.39 (16.01) 0.088 0.006 0.260* 0.080 0.132 53 E. Rosenstreich, L. Ruderman / Consciousness and Cognition 43 (2016) 48–56 attention, divided attention significantly reduced memory hits, t(65) = 12.33, p < 0.001, RTs in the tone classification task, t (63) = 5.35, p < 0.001, and tone classification accuracy, t(63) = 3.19, p = 0.002. These findings confirm that the attention manipulation used in this study replicated previously reported effects of attention on memory (e.g., Naveh-Benjamin et al., 2000). After establishing that our mindfulness and attention data was in-line with previous studies, we turned to examine our research hypotheses. Table 1 reveals that the facets of mindfulness were generally not correlated with memory performance, with the exception of the facet of being non-judgmental (NJ) being negatively correlated with FAs when attention was full. That is, participants with high NJ scores were less likely to falsely recognize unstudied test words, as compared to participants with low NJ scores. Because our first hypothesis argued that facets of mindfulness will be associated with a higher hit rate, and that a stronger association will be found in the divided than in the full attention condition, these findings do not correspond with our hypothesis. 3.2. The joint association between mindfulness and attention on memory Next, it was examined whether facets of mindfulness moderated the impact of attention on memory performance. To this end, repeated-measures analyses of covariance (ANCOVAs) were conducted, with attention (full, divided) as the independent variable and sensitivity and response bias as the dependent variables. In order to assess the moderation effect of mindfulness, each facet of mindfulness was treated as a covariate. Because multiple comparisons were conducted for each dependent variable, a Family-wise Error Rate (FWER; Benjamini & Hochberg, 1995) approach was adopted in order to reduce the risk of alpha inflation. According to this approach, the p-value of each comparison is compared to a differential alpha value (i.e., alpha per comparison: apc). Specifically, the FWER approach derived from the Bonferroni correction for multiple comparisons, in which the original alpha value (0.05) is divided by the number of comparisons. However, this correction dramatically reduces statistical power, and increases the probability of type II error. In the FWER, apc is the product of multiplying Bonferroni’s corrected alpha by the comparison’s p-value location on an ordinal scale. Hence, a specific apc is calculated for each comparison, thus promoting an overall consideration of the significance levels in the model. Therefore, throughout this section, the FWER approach was employed for each family of comparisons. The stricter Bonferroni’s correction was employed when post hoc analyses were carried. ANCOVAs and their FWER alpha corrections are presented in Table 2. For sensitivity, Table 2 reveals that memory sensitivity was higher under full attention (M = 2.48, SD = 1.11) than under divided attention (M = 1.10, SD = 0.52). However, none of the facets of mindfulness moderated the impact of attention on sensitivity. For response bias, Table 2 reveals that participants’ responses were positively biased when attention was full (M = 0.27, SD = 0.52), but negatively biased when attention was divided (M = 0.13, SD = 0.46). That is, when attention was full, participants were more likely to judge test items as ‘‘studied”, than when attention was divided. Moreover, Table 2 also reveals that the facet of NJ moderated the effect of attention on response bias. In order to examine the source of this moderation effect, NJ was recoded into a dichotomous variable, based on its median (Md = 20). Next, six t-tests were conducted, each with a Bonferroni’s corrected alpha value of 0.008. The first test revealed that when attention was full, participants with low NJ were more biased (M = 0.43, SD = 0.48) than participants with high NJ (M = 0.09, SD = 0.51), t(64) = 2.85, p = 0.006. The second test revealed that there was no significant difference in response bias when attention was divided among participants with low (M = 0.12, SD = 0.42) and high (M = 0.13, SD = 0.50) NJ scores, t(64) = 0.12, p = 0.90. The third test revealed a significant shift of the response criterion as a function of attention among participants with low NJ scores, t(33) = 7.10, p < 0.001. Among participants with high NJ scores, however, the fourth test revealed that response criterion remained fairly stable, t(31) = 2.20, p = 0.04. Finally, two single-sample t-tests confirmed that participants with high NJ scores were not biased (that is, their response bias score did not differ from zero) when attention was full, t(31) = 0.95, p = 0.35, as well as when attention was divided, t(31) = 1.52, p = 0.14. These findings confirm the second hypothesis, that mindfulness is associated with memory performance through response bias rather than through sensitivity. Table 2 ANCOVAs results with Attention as independent variable, five facets of mindfulness as moderators, and with memory sensitivity and response bias as dependent variables. Memory sensitivity (d0 ) F(1, 64) Att. Att.  NR Att.  Observe Att.  Awareness Att.  Describe Att.  NJ * 112.01 0.77 0.28 0.39 0.07 0.05 p-value (apc) <0.001 (0.008) 0.38 (0.017) 0.60 (0.030) 0.54 (0.025) 0.79 (0.042) 0.83 (0.050) Response bias (c) g2 0.63 0.01 0.00 0.01 0.00 0.00 F(1, 64) * 35.57 2.30 0.76 1.97 0.01 6.39* p-value (apc) g2 <0.001 (0.008) 0.13 (0.025) 0.39 (0.042) 0.17 (0.030) 0.91 (0.050) 0.01 (0.017) 0.35 0.04 0.01 0.03 0.00 0.09 Note. Att. - Attention; NR - Non-reactivity; NJ - Non-judgmental; apc - alpha per comparison, following the family-wise error rate approach. * Significant result. 54 E. Rosenstreich, L. Ruderman / Consciousness and Cognition 43 (2016) 48–56 4. Discussion This study aimed to examine the connection between facets of mindfulness and the impact of attention on recognition memory. In particular, this research was guided by the question of whether mindfulness will moderate the effects of divided attention on memory performance, such that high levels of mindfulness will reduce the impact of attention on memory. Memory performance was assessed by the proportions of hits, false alarms, memory sensitivity, and response bias. It was hypothesized that facets of mindfulness will be positively associated with hit rates under both full and divided attention conditions, with a stronger association under full attention. It was also predicted that facets of mindfulness will be associated with response bias, but not with sensitivity. Finally, it was predicted that facets of mindfulness will moderate the effects of divided attention on response bias. It was found that in general, the five facets of mindfulness were not associated with memory performance. Specifically, facets of mindfulness were not correlated with hit rate, yet the facet of NJ was negatively correlated with false alarm rate when attention was full. These findings are not in line with previous findings (Rosenstreich, 2015), in which hit rates were increased and false alarms were not affected as a function of mindfulness. The discrepancy between the two studies could be accounted for by differences in their designs. First, the two studies varied in the length of the filler task presented between the study and the test: a five-minute filler was employed in Rosenstreich (2015), whereas a 30-s filler was employed in the current study. Filler length may moderate the effects of different manipulations on memory performance (cf., Dennis, Lee, & Kinnell, 2008). In the current study, the short filler might have resulted in an overall high recognition rate, which may have led to the null association between facets of mindfulness and hit rate. As for the association found between mindfulness and false alarms, although it was absent in studies which were designed to provoke memory distortions (Rosenstreich, 2015; Wilson, Mickes, Stolarz-Fantino, Evrard, & Fantino, 2015), it was observed when ‘‘traditional” recognition tests were carried (Lloyd et al., 2016). Hence, the negative correlation between NJ and false alarms observed in this study correspond with recent literature. Finally, in Rosenstreich (2015), mindfulness was practiced and promoted, whereas in the current study predisposed levels of mindfulness were measured. Thus, it is possible that the brief mindfulness practice employed in Rosenstreich (2015) was strong enough to affect memory; along with this notion, questionnaires of dispositional mindfulness may not be sensitive enough to the individual differences in mindfulness associated with recognition performance (cf., Grossman & Van Dam, 2011). The scope of this study was the investigation of the association between facets of mindfulness and signal detection measures of memory (i.e., memory sensitivity and response bias). As predicted, it was found that facets of mindfulness were not associated with sensitivity, and did not moderate the impact of attention on this measure. However, the facet of NJ was positively associated with response bias, such that participants with high NJ scores were less biased than participants with low NJ scores. That is, this finding shows that high NJs characterized with an even distribution of ‘‘studied” and ‘‘unstudied” responses, whereas low NJs characterized with a tendency to favor the ‘‘studied” response. These findings further support the patterns observed in Rosenstreich (2015), in which mindfulness practice did not affect sensitivity, yet promoted less biased response distribution. The notion that mindfulness does not improve or is not associated with the ability to discriminate between studied (target) and unstudied (foil) items, may bring to light some of the underlying cognitive mechanisms of mindfulness. In particular, it seems that mindfulness does not improve the extent to which information is encoded into memory; such an improvement would have been manifested in increased sensitivity (cf. Rosenstreich & Goshen-Gottstein, 2015). Furthermore, it seems that mindfulness is not associated with better allocation of attention to the processing of to-be-remembered information, because more attentional resources are typically associated with increased sensitivity (Naveh-Benjamin et al., 2000; Rosenstreich & Goshen-Gottstein, 2015). The association between response bias and mindfulness—with the facet of NJ in particular—may suggest that mindfulness has more to do with retrieval rather than encoding processes. That is, recognition retrieval processes are sometimes characterized in terms of decision making processes (e.g., Dunn, 2004; Wixted & Stretch, 2004; Yonelinas, Aly, Wang, & Koen, 2010), in which a decision criterion (or criteria) is set at a certain level of memory strength; thus, if an item’s strength exceeds this criterion, it is judged as ‘‘studied”. Response bias is a shift of this criterion, either to a relaxed (more ‘‘studied” responses) or a conservative (more ‘‘unstudied” responses) position. It seems that mindfulness in general and NJ in particular are associated with a balanced decision criterion. Moreover, whereas among low NJs the location of the decision criterion shifted as a function of attention, high NJs’ criterion was resilient to attentional deficit. The notion that mindfulness is connected with less biased decision making has been recently demonstrated by Hafenbrack, Kinias, and Barsade (2013). They showed that mindfulness practice reduced the sunk cost bias, a bias in which a cannot-be-recovered investment affects ones’ action. Specifically, participants who underwent mindfulness practice were more likely to ignore a past financial investment and to reach for a new financial opportunity, as compared to control participants (see also, Chong, Kee, & Chaturvedi, 2015). Furthermore, it was found that participants who scored high in the Mindfulness Awareness and Attention Scale (MAAS; Brown & Ryan, 2003) were more likely to overcome the sunk cost bias; it should be noted, however, that whereas the MAAS is comparable with Baer et al.’s facet of awareness, other facets of mindfulness were not assessed. Hence, it is not clear whether the facet of NJ plays a role in the sunk cost bias. Finally, the findings reported in this study support previous findings regarding the association between mindfulness and attention. In particular, mindfulness was argued to promote better allocation of attentional resources (Chambers, Lo, & Allen, 2008; Jha, Krompinger, & Baime, 2007; cf. Zeidan, Grant, Brown, McHaffie, & Coghill, 2012). Similarly, the facet of awareness E. Rosenstreich, L. Ruderman / Consciousness and Cognition 43 (2016) 48–56 55 was argued to be associated with performance in a continuous performance test (CPT) of attention (Ruocco & Direkoglu, 2013); participants with high awareness scores typically performed better in the attention task. In the current study as well, the facet of awareness was associated with the tone classification task—an auditory CPT (Naveh-Benjamin et al., 2000): CPT performance under divided attention was scarcely affected among participants with high, as compared to low, awareness scores. 5. Conclusions, limitations, and future directions Taken together, the findings reported here lead to the conclusion that mindfulness may be associated with memory through decision making processes, rather than through better allocation of attention during the encoding of information. In addition, the study provided further support to the notion that different facets of mindfulness are associated with memory and attention (Ruocco & Direkoglu, 2013): whereas response bias in recognition memory was associated with the facet of NJ, performance in the attention task was associated with the facet of awareness. Nevertheless, it should be noted that Ruocco and Direkoglu employed the Philadelphia Mindfulness Scale (PHLMS) and the acceptance scale in order to assess awareness and acceptance, respectively. In the current study, however, the FFMQ was employed in order to assess the five facets of mindfulness. Although these different measures may draw from common mindfulness factors, they often are subject to different interpretations. For example, the act with awareness factor in the PHLMS may correspond with the observation factor—rather than with the act with awareness factor—in the FFMQ (Bergomi, Tschacher, & Kupper, 2013). In addition, it has been demonstrated that measures of mindfulness may draw from theoretical constructs that vary from mindfulness (Grossman & Van Dam, 2011). This evidence therefore limits the ability to draw conclusions about the association between mindfulness as a trait and memory performance. Although similar findings have been also demonstrated when mindfulness was practiced (Rosenstreich, 2015), further study is needed in order to better understand the mindfulness-attention-memory triad. Such studies should examine memory performance when mindfulness is both practiced and assessed. This may crystallize the pathway through which mindfulness affects memory performance. Another limitation lies in the nature of the memory task itself. Specifically, the recognition task employed in this study was designed to address the lack of research-oriented memory tasks, rather than diagnostic memory tests, in mindfulness studies. Although the recognition task employed in this study deepened our understanding of the connection between mindfulness and memory, it also restricted the findings to the characteristics of its materials. That is, neutral words served as stimuli in the recognition task; thus, one cannot generalize the conclusions regarding mindfulness and memory performance to other stimuli (such as emotional words or faces). Considering the attention-demanding nature of emotional stimuli (e.g., Pessoa, 2009), it would be interesting to examine how mindfulness is connected with memory sensitivity and the response bias of such stimuli. References Alberts, H. J., & Thewissen, R. (2011). The effect of a brief mindfulness practice on memory for positively and negatively valenced stimuli. Mindfulness, 2, 73–77. Baer, R. A., Smith, G. T., Hopkins, J., Krietemeyer, J., & Toney, L. (2006). Using self-report assessment methods to explore facets of mindfulness. Assessment, 13, 27–45. Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society. Series B (Methodological), 289–300. Bergomi, C., Tschacher, W., & Kupper, Z. (2013). Measuring mindfulness: First steps towards the development of a comprehensive mindfulness scale. Mindfulness, 4, 18–32. Bohlmeijer, E., Peter, M., Fledderus, M., Veehof, M., & Baer, R. (2011). Psychometric properties of the five facet mindfulness questionnaire in depressed adults and development of a short form. Assessment, 18, 308–320. Brown, K. W., & Ryan, R. M. (2003). The benefits of being present: Mindfulness and its role in psychological well-being. Journal of personality and social psychology, 84(4), 822–848. Carpenter, S. K., Pashler, H., & Vul, E. (2006). What types of learning are enhanced by a cued recall test? Psychonomic Bulletin & Review, 13, 826–830. Chambers, R., Lo, B. C. Y., & Allen, N. B. (2008). The impact of intensive mindfulness training on attentional control, cognitive style, and affect. Cognitive Therapy and Research, 32, 303–322. Chong, Y. W., Kee, Y. H., & Chaturvedi, I. (2015). Effects of brief mindfulness induction on weakening habits: Evidence from a computer mouse control task. Mindfulness, 6, 582–588. Craik, F. I., Naveh-Benjamin, M., Ishaik, G., & Anderson, N. D. (2000). Divided attention during encoding and retrieval: Differential control effects? Journal of Experimental Psychology: Learning, Memory, and Cognition, 26(6), 1744. Dennis, S., Lee, M. D., & Kinnell, A. (2008). Bayesian analysis of recognition memory: The case of the list-length effect. Journal of Memory and Language, 59(3), 361–376. Dunn, J. C. (2004). Remember-know: A matter of confidence. Psychological Review, 111, 524–542. Galla, B. M., Hale, T. S., Shrestha, A., Loo, S. K., & Smalley, S. L. (2012). The disciplined mind: Associations between the Kentucky inventory of mindfulness skills and attention control. Mindfulness, 3, 95–103. Gardiner, J. M., Ramponi, C., & Richardson-Klavehn, A. (1998). Experiences of remembering knowing and guessing. Consciousness and Cognition, 7, 1–26. Gardiner, J. M., & Parkin, A. J. (1990). Attention and recollective experience in recognition memory. Memory and Cognition, 18, 579–583. Grossman, P., & Van Dam, N. T. (2011). Mindfulness, by any other name. . .: Trials and tribulations of sati in western psychology and science. Contemporary Buddhism, 12, 219–239. Hafenbrack, A. C., Kinias, Z., & Barsade, S. G. (2013). Debiasing the mind through meditation mindfulness and the sunk-cost bias. Psychological Science, 25, 369–376. Jensen, C. G., Vangkilde, S., Frokjaer, V., & Hasselbalch, S. G. (2012). Mindfulness training affects attention—Or is it attentional effort? Journal of Experimental Psychology: General, 141, 106–123. 56 E. Rosenstreich, L. Ruderman / Consciousness and Cognition 43 (2016) 48–56 Jha, A. P., Krompinger, J., & Baime, M. J. (2007). Mindfulness training modifies subsystems of attention. Cognitive, Affective, & Behavioral Neuroscience, 7, 109–119. Kabat-Zinn, J. (2003). Mindfulness-based practices in context: Past, present, and future. Clinical Psychology: Science and Practice, 10, 144–158. Kang, Y., Gruber, J., & Gray, J. R. (2013). Mindfulness and de-automatization. Emotion Review, 5, 192–201. Knott, L. M., & Dewhurst, S. A. (2007). The effects of divided attention at study and test on false recognition: A comparison of DRM and categorized lists. Memory and Cognition, 35, 1954–1965. Lloyd, M., Szani, A., Rubenstein, K., Colgary, C., & Pereira-Pasarin, L. (2016). A brief mindfulness exercise before retrieval reduces recognition memory false alarms. Mindfulness, 1–8. Lykins, E. L., Baer, R. A., & Gottlob, L. R. (2012). Performance-based tests of attention and memory in long-term mindfulness meditators and demographically matched nonmeditators. Cognitive Therapy and Research, 36, 103–114. Macmillan, N. A., & Creelman, C. D. (2005). Detection theory: A user’s guide (2nd ed.)Mahwah, NJ: Erlbaum. Maddox, G. B., Naveh-Benjamin, M., Old, S., & Kilb, A. (2012). The role of attention in the associative binding of emotionally arousing words. Psychonomic Bulletin & Review, 19, 1128–1134. Morrison, A. B., Goolsarran, M., Rogers, S. L., & Jha, A. P. (2013). Taming a wandering attention: Short-form mindfulness training in student cohorts. Frontiers in Human Neuroscience, 7, 897. Naveh-Benjamin, M., Craik, F. I., Gavrilescu, D., & Anderson, N. D. (2000). Asymmetry between encoding and retrieval processes: Evidence from divided attention and a calibration analysis. Memory and Cognition, 28, 965–976. Naveh-Benjamin, M., Guez, J., Kilb, A., & Reedy, S. (2004). The associative memory deficit of older adults: Further support using face-name associations. Psychology and Aging, 19(3), 541. Pessoa, L. (2009). How do emotion and motivation direct executive control? Trends in Cognitive Sciences, 13, 160–166. Rosenstreich, E. (2014). Mindfulness and memory. In N. N. Singh (Ed.), The psychology of meditation. New York: Nova. Rosenstreich, E. (2015). Mindfulness and false-memories: The impact of mindfulness practice on the DRM paradigm. The Journal of Psychology, 1–17 (aheadof-print). Rosenstreich, E., & Goshen-Gottstein, Y. (2015). Recollection-based retrieval is influenced by contextual variation at encoding but not at retrieval. PLoS ONE, 10(7), e0130403. http://dx.doi.org/10.1371/journal.pone.0130403. Rotello, C. M., Masson, M. E., & Verde, M. F. (2008). Type I error rates and power analyses for single-point sensitivity measures. Perception & Psychophysics, 70, 389–401. Ruocco, A. C., & Direkoglu, E. (2013). Delineating the contributions of sustained attention and working memory to individual differences in mindfulness. Personality and Individual Differences, 54, 226–230. Shapiro, S. L., & Carlson, L. E. (2009). The art and science of mindfulness: Integrating mindfulness into psychology and the helping professions. New York: APA Publications. Tang, Y., Ma, Y., Wang, J., Fan, Y., Feng, S., Lu, Q., ... Posner, M. I. (2007). Short-term meditation training improves attention and self-regulation. Proceedings of the National Academy of Sciences, 104, 17152–17156. van Vugt, M. K., Hitchcock, P., Shahar, B., & Britton, W. (2012). The effects of mindfulness-based cognitive therapy on affective memory recall dynamics in depression: A mechanistic model of rumination. Frontiers in Human Neuroscience, 6, 257. http://dx.doi.org/10.3389/fnhum.2012.00257. Wilson, B. M., Mickes, L., Stolarz-Fantino, S., Evrard, M., & Fantino, E. (2015). Increased false-memory susceptibility after mindfulness meditation. Psychological science. http://dx.doi.org/10.1177/0956797615593705. Wixted, J. T., & Stretch, V. (2004). In defense of the signal detection interpretation of remember/know judgments. Psychonomic Bulletin & Review, 11, 616–641. Yonelinas, A. P., Aly, M., Wang, W. C., & Koen, J. D. (2010). Recollection and familiarity: Examining controversial assumptions and new directions. Hippocampus, 20, 1178–1194. Zeidan, F., Grant, J. A., Brown, C. A., McHaffie, J. G., & Coghill, R. C. (2012). Mindfulness meditation-related pain relief: Evidence for unique brain mechanisms in the regulation of pain. Neuroscience Letters, 520, 165–173.

Consciousness and Cognition 22 (2013) 1047–1060 Contents lists available at SciVerse ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Coloured Letters and Numbers (CLaN): A reliable factor-analysis based synaesthesia questionnaire Nicolas Rothen a,⇑, Elias Tsakanikos b,c, Beat Meier d, Jamie Ward a a School of Psychology and Sackler Centre for Consciousness Science, University of Sussex, Brighton, United Kingdom Institute of Psychiatry, King’s College London, United Kingdom c Centre of Research in Individual Differences (CRID), Department of Psychology, University of Roehampton, United Kingdom d Institute of Psychology and Center for Cognition, Learning and Memory, University of Bern, Switzerland b a r t i c l e i n f o Article history: Received 23 April 2013 Available online 9 August 2013 Keywords: Synaesthesia Colour Questionnaire Factor analysis Consistency Stroop a b s t r a c t Synaesthesia is a heterogeneous phenomenon, even when considering one particular sub-type. The purpose of this study was to design a reliable and valid questionnaire for grapheme-colour synaesthesia that captures this heterogeneity. By the means of a large sample of 628 synaesthetes and a factor analysis, we created the Coloured Letters and Numbers (CLaN) questionnaire with 16 items loading on 4 different factors (i.e., localisation, automaticity/attention, deliberate use, and longitudinal changes). These factors were externally validated with tests which are widely used in the field of synaesthesia research. The questionnaire showed good test–retest reliability and construct validity (i.e., internally and externally). Our findings are discussed in the light of current theories and new ideas in synaesthesia research. More generally, the questionnaire is a useful tool which can be widely used in synaesthesia research to reveal the influence of individual differences on various performance measures and will be useful in generating new hypotheses. Ó 2013 Elsevier Inc. All rights reserved. 1. Introduction Synaesthesia can be described as extraordinary perceptual experiences associated with normal sensory or cognitive process. For instance, in grapheme-colour synaesthesia a letter printed in black triggers a highly specific and consistent colour experience. So far, many different forms of synaesthesia have been identified in scientific work, including spatial associations with days and months (time–space synaesthesia; Simner, Mayo, & Spiller, 2009; Smilek, Callejas, Dixon, & Merikle, 2007), tactile experiences elicited from observing touch (mirror-touch synaesthesia; Banissy & Ward, 2007; but see Rothen & Meier, 2013), and even colours evoked by swimming styles (swimming-style colour synaesthesia; Nikolić, Jürgens, Rothen, Meier, & Mroczko, 2011; Rothen, Nikolić, et al., 2013). Not only do many different forms exist, but also the heterogeneity within a particular form is large. For instance, synaesthetic experiences may be perceived in the mind’s eye (often referred to as associator synaesthetes) or as if projected onto the synaesthesia inducing stimulus (often referred to as projector synaesthetes) (Dixon, Smilek, & Merikle, 2004; Ward, Li, Salih, & Sagiv, 2007). Yet, by means of conventional statistical analyses (e.g., factor analysis) no psychometrically sound questionnaire exists to assess how different aspects of the heterogeneity of synaesthetic experiences influence performance in various tasks and its underlying neural activation. It is our aim to provide the field of synaesthesia research with such an instrument using standard statistical techniques. ⇑ Corresponding author. Address: School of Psychology and Sackler Centre for Consciousness Science, University of Sussex, Falmer, Brighton BN1 9QH, United Kingdom. E-mail address: nicolas.rothen@gmail.com (N. Rothen). 1053-8100/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.concog.2013.07.005 1048 N. Rothen et al. / Consciousness and Cognition 22 (2013) 1047–1060 Questionnaires (and/or interviews) in synaesthesia research have been generally applied for three reasons. Firstly, to assess that the reported experiences are indeed likely to be synaesthesia rather than some other phenomena (e.g. for diagnostic purposes). For instance, questions relating to synaesthesia as a life-long trait, that is consistent and automatic have been used for this purpose (Baron-Cohen, Harrison, Goldstein, & Wyke, 1993; Rothen & Meier, 2010a; Simner, Harrold, Creed, Monro, & Foulkes, 2009; Simner et al., 2006). Secondly, questionnaires have also been applied to reveal the relationship between synaesthesia and other phenomenological experiences or associated personality traits. For instance, to assess the link between synaesthesia and eidetic imagery (Glicksohn, Steinbach, & Elimalach-Malmilyan, 1999) or visual imagery (Barnett & Newell, 2008), to estimate the prevalence of synaesthesia in meditators (Walsh, 2005) or fine art students (Domino, 1989). Moreover and most important in the context of this article, questionnaires have been applied to capture (and quantify) individual differences of the synaesthetic phenomenology. Questionnaires concerned with individual differences in grapheme-colour synaesthesia almost exclusively focus on the associator–projector type distinction (Rouw & Scholte, 2007; Skelton, Ludwig, & Mohr, 2009). Phenomenological differences between associator and projector type synaesthetes, identified by using the PA questionnaire, are associated with functional and structural differences in the brain (Rouw & Scholte, 2007, 2010; Van Leeuwen, den Ouden, & Hagoort, 2011) and, as such, this measure has some external validity. However, test–retest reliability and factor structure of the questionnaire have never been reported. The analyses performed using the questionnaire assume a single factor structure (along the projector–associator dimension), but this has not been established using techniques such as factor analysis and others have argued that further fractionations are possible (Ward et al., 2007). An illustrated projector–associator questionnaire (i.e., Illustrated Synaesthesia Questionnaire – ISEQ) has been created and shown to produce reliable test–retest results on the basis of a rather small sample of 12 grapheme-colour synaesthetes (Skelton et al., 2009). The projector–associator distinction has also been linked to various behavioural differences linked to the processing of graphemes and/or colours (e.g., Dixon et al., 2004; Gebuis, Nijboer, & van der Smagt, 2009; Hubbard, Arman, Ramachandran, & Boynton, 2005; Ward et al., 2007). In a synaesthetic version of the Stroop-test, participants must name the real colour of a grapheme and ignore their synaesthesia or, alternatively, name their synaesthetic colour and ignore the real colour (Dixon et al., 2004; Ward et al., 2007). Performance on these tasks has been found to relate to the projector–associator distinction (measured by self-report rather than questionnaire scales). Projector synaesthetes were found to exhibit stronger Stroop interference (i.e., RT difference between incongruent and congruent trials) when the veridical colour of the synaesthetic inducers on the screen had to be named in comparison to when the colour of the synaesthetic experiences had to be named. In contrast, associator synaesthetes were found to exhibit weaker Stroop interference when the veridical colour of the synaesthetic inducers on the screen had to be named in comparison to when the colour of the synaesthetic experiences had to be named (Dixon et al., 2004). Furthermore, it was found that projector synaesthetes were generally faster in synaesthetic colour naming as compared to veridical colour naming and associator synaesthetes were generally slower in synaesthetic colour naming as compared to veridical colour naming (Dixon et al., 2004; Ward et al., 2007). To date Stroop-type tasks are probably the most often applied tasks in synaesthesia research and very often used to validate genuine synaesthetic experiences (Banissy & Ward, 2007; Nikolić et al., 2011). However, it is important to note that ‘synaesthetic’ Stroop effects may also be found in non-synaesthetes who have been trained to have similar associations (Meier & Rothen, 2009; Rothen, Nikolić, et al., 2013; Rothen, Wantz, & Meier, 2011). In the present study, we link performance on this task with particular phenomenology and, hence, show that Stroop-type interference is linked to synaesthetic experiences rather than the presence of grapheme-colour associations per se (such as those that may be learned). Besides the projector–associator distinction, other sources of individual differences exist in grapheme-colour synaesthesia which may also provide valuable information to cognitive scientists, but were only very rarely considered (e.g., Barnett et al., 2008; Rich, Bradshaw, & Mattingley, 2005). For instance, some synaesthetes report vivid synaesthetic experiences whereas others perceive them rather faded. Also synaesthetes differ in the way they use their synaesthetic experiences in everyday life (e.g., as mnemonics). Moreover, whilst synaesthesia is normally defined as being automatic, several of our participants report that, for them, they need to think about the colour before it appears. It is unclear whether this maps onto the projector–associator distinction or is orthogonal to it. Therefore, it is important to assess other sources of individual differences in order to correctly classify a synaesthete and when we try to link performance in a specific task with individual differences in grapheme-colour synaesthesia. 2. Questionnaire 2.1. Method 2.1.1. Participants We recruited 628 self-referred grapheme-colour synaesthetes – via the synaesthesia research websites of the University of Sussex and the University of Bern, and via different internet platforms relevant to synaesthesia – to complete an online questionnaire and to do an internet based test of consistency to assess the genuineness of their synaesthesia (Eagleman, Kagan, Nelson, Sagaram, & Sarma, 2007). In total 249 participants completed both the internet based test of consistency and the questionnaire. The remaining 379 participants completed the questionnaire, but not the test of consistency. Of the initial 628 participants, 405 participants completed the English version of the questionnaire and 223 participants N. Rothen et al. / Consciousness and Cognition 22 (2013) 1047–1060 1049 completed the German version of the questionnaire (cf., Section 2.1.2). Fifty-three participants of the 628 participants completed the questionnaire a second time on another occasion (enabling us to measure test–retest reliability). Hence, the questionnaire has been completed a total of 681 times. This test does not ask for demographic details. 2.1.2. Materials The original questionnaire consisted of 30 items (Appendix A). All items were related to grapheme-colour synaesthesia. Each item was rated on 5-point Likert-scale (1 strongly disagree, 2 moderately disagree, 3 neither agree nor disagree, 4 moderately agree, 5 strongly agree). We created an English version (Appendix A) and a German version (Supplementary material) of the questionnaire in order to be able to access more participants. The questionnaire consisted of two internet pages. On the first page, there were 21 items specifically related to synaesthetic experiences for letters and numbers. For the purpose of these questions, participants were asked to think about their phenomenology with respect to some illustrative examples of letters (Aa Bb Cc Dd Ee Ff Gg) and numbers (0 1 2 3 4 5 6 7 8 9). The last 9 items were presented on the subsequent internet page and referred to grapheme-colour experiences in general (e.g. outside of the current testing scenario). The specific items were related to the following aspects of synaesthetic experiences: the location of the colour experiences (conceptually related to the projector–associator distinction); the extent to which the colours appear automatically or require effort/attention; the extent to which the participants use their associations in daily life; the extent to which the associations have changed in intensity over time (longitudinal changes) or can be changed at will (intensity/flexibility); and, finally, the extent to which colours also trigger thoughts about graphemes (bidirectionality). Note that the questionnaire is only concerned with aspects of the synaesthetic experience but not with aspects of the synaesthetic inducer. As test of consistency, we used the standardised grapheme-colour consistency test, which is accessible via the internet (www.synesthete.org). That is, the material consisted of letters from A to Z and numbers from 0 to 9 as synaesthetic inducers and possible synaesthetic colour concurrents in HSV space (Eagleman et al., 2007). 2.1.3. Procedure After clicking on the link to the questionnaire on one of our recruitment pages on the internet, participants were presented with an introductory screen which informed about the nature of the questionnaire. By continuing, participants were presented with a data protection statement on the subsequent page. Thereafter, participants were presented with 21 items followed by nine additional items as specified in Section 2.1.2. Note, it was not possible to go back to the first set of items again, after proceeding to the second set. Next, participants were asked to indicate their first name, last name, and e-mail address which would allow us to contact them and associate the data from the questionnaire with the data of the consistency test. Once participants submitted their responses to the items, they were presented with a thank you and goodbye screen providing further contact details of our research group for any questions and comments they may have wanted to submit. Next, those participants who had not yet completed the test of consistency before (e.g., as part of another project), were invited by email to do the test by following a link. During the assessment of grapheme-colour consistency, participants were presented with the graphemes A–Z and 0–9 in random order three times each. Participants had to pick one colour represented in HSV (Hue, Saturation, and Value) colour space for each grapheme according to the elicited synaesthetic colour experience. There was also a no-colour option for the cases where the presented grapheme did not elicit a synaesthetic colour experience. 2.1.4. Analysis We conducted a factor analysis on the data of the questionnaire. We were particularly interested in items that load highly on one factor but not on others. The data were analysed in SPSS 19 (Statistical Package for the Social Sciences, version 19) as follows. Given the design of the questionnaire we performed a maximum likelihood factor analysis with a fixed number of five factors (i.e., corresponding to localisation, automaticity/attention, deliberate use, longitudinal changes, and intensity/flexibility) and varimax rotation. Note that the varimax rotation was justified by the fact that the unrotated factors were not significantly correlated with each other. The factor analysis was not conducted with a fixed number of six factors because the aspect of bidirectionality was only covered by two items and hence, cannot result in a meaningful factor based on its own items. Therefore, the items related to bidirectional aspects of synaesthesia were not taken into consideration for the factor analysis. Colours selected in the test of consistency were saved as representations in RGB (Red, Green, Blue) colour space. To calculate consistency of syneasthetic colour experiences, we adapted the method of Rothen, Seth, et al. (2013). That is, we converted RGB values into CIELUV values. First, we linearised RGB values by applying inverted gamma functions (gamma compression), and converted these linear RGBs into tristimulus values (XYZ; Brainard, Pelli, & Robson, 2002). Based on these XYZ values, we calculated CIELUV (Hunt & Pointer, 2011, p. 55). We used standard RGB (sRGB; Stokes, Anderson, & Chandrasekar, 1996) to obtain monitor specifications as these are most representative of a range of random monitors. Next, we calculated Euclidean distances in CIELUV space between the two instances of each inducer on an individual basis. For each subject the grand mean of these Euclidean distances was calculated as a representative value for overall consistency. Hence, a lower value implies greater consistency. A cut-off value of 135 has been suggested as resulting in maximal sensitivity (i.e., true positives) and specificity (i.e., true negatives), but can be adjusted depending on whether someone is concerned with sensitivity or specificity (Rothen, Seth, et al., 2013). Because this method is based on a perceptual colour model, it provides a more accurate measure for synaesthetic consistency as compared to synaesthetic consistency based on non-perceptual colour spaces, such as for example RGB and HSV colour spaces (Rothen, Seth, et al., 2013). 1050 N. Rothen et al. / Consciousness and Cognition 22 (2013) 1047–1060 The alpha level was set to .05 for all statistical analyses and t-tests were two-tailed. Exceptions are indicated at the specific location. 2.2. Results Average consistency was 63 (SD = 31) for the 249 participants who completed the grapheme-colour consistency test. Participants whose consistency value differed by more than 3 SD from the average consistency were regarded as outliers (N = 2) and excluded from the sample. The consistency values of the remaining 247 participants ranged from 7 to 156 with five participants exceeding the previously described optimal cut-off value of 135. If not further specified, the following analyses are based on this sample. Table 1 presents the results of the factor analysis. The Kaiser–Meyer Olkin measure of sampling adequacy suggested that our sample was factorable (KMO = .733). Factor loadings smaller than .4 were suppressed in the output and excluded, resulting in a factor-solution of 5 factors with a total of 17 items. However, only one item with a factor loading greater than .4 loaded on the last factor. Since it is not feasible for a single item to build a factor on its own, the item/factor was excluded. The mean score on the remaining 16 items of the questionnaire was 47.17 (SD = 9.146) and the value of Cronbach’s alpha was .755. The mean values of each item and factor and Cronbach’s alpha for each of the factors are included in Table 1. Based on the result of this analysis, the reduced version of the questionnaire consists of 16 items loading on 4 different factors. Following this finding, we tested whether those 16 items of the final CLaN (that loaded on the first 4 factors in the previous analysis, see Appendix B) lead to the same factor structure for the English and the German version, respectively. That is, we repeated the analysis including the specified 16 items seeking a 4 factor solution. The analysis was conducted twice, once for data from the English version and once for the data from the German version. Factor loadings smaller than .3 were suppressed in the output. We obtained similar results for both analyses (Table 2). Next, we repeated this procedure with all 628 self-referred synaesthetes (irrespective of synaesthetic consistency) in order to test for the generalisation of the factor structure. Again, we obtained similar results for both versions of the questionnaire (Supplementary Table 1). Finally, we assessed test–retest reliability of the final 16 items of the CLaN for those 53 participants who completed the questionnaire twice. The average time between the two occasions of questionnaire completion was 119.45 days (SE = 17.21). Fig. 1 presents the descriptive results and correlations between test and retest for the final 16 items of the CLaN which have been identified to lead to a reliable 4 factor solution. All correlations were significant at the .01 level, but one which was significant at the .05 level, indicating generally good test–retest reliability. 2.3. Discussion Towards our aim to create a reliable questionnaire which covers different sources of individual differences in synaesthesia, we were able to identify 4 factors in a sample of verified synaesthetes. These 4 factors consisted of a total of 16 items which showed good internal consistency. Moreover, the internal consistency for each factor separately was good. Table 1 Rotated factor matrix of factor loadings for the 28 initial questionnaire items. Note: The percentage of cumulative total variance is 36.147. Factor loadings smaller .4 were suppressed and excluded and for the reason of simplicity of inspection the respective items with lower factor loadings are not shown. Mean scores and Cronbach’s alpha of items with negative factor loadings are based on reversed scales. Item numbers correspond to items in Appendix A. N. Rothen et al. / Consciousness and Cognition 22 (2013) 1047–1060 1051 Using these 16 items, we were able to replicate the 4 factors when assessing the data of the English and the German version of the questionnaire separately. This was also the case when we assessed the two versions including all participants who filled in the online questionnaire irrespective of the verification of their synaesthesia via a test of consistency. However, we acknowledge that order of the different factors was different for the English version of the questionnaire when including all participants. Moreover, we also acknowledge that the items of factor 4 changed the direction of their factor loadings between the different analyses. However, we would like to point out that it was always the same items that were opposed to each other. That is, if item A loaded negatively on the factor and item B positively in one analysis, they were always opposed to each other in that when A loaded positively on the factor in another analysis B would then load negatively on that factor in for the same analysis. Nevertheless, in all of the three analyses the same items grouped together to factors indicating a relatively stable factor structure of the questionnaire especially for the first three factors. Furthermore, we were able to show that the 16 items (of the 4 factors) generally exhibit very good test–retest reliability. With respect internal consistency and test–retest reliability, our results indicate that we succeeded in creating a valid instrument to measure individual differences in the experience synaesthesia. Table 2 Rotated factor matrix of factor loadings for the 16 final items of the CLaN. Note: The percentage of cumulative total variance is 49.207 for the English version and 54.466 for the German version of the questionnaire. Factor loading smaller than .3 were suppressed. In the interest of comparability across the different factor analyses in this article, number of items, mean item score, and Cronbach’s alpha are always based on the greyshaded items of the respective factor (even though item 8 in the English version of the questionnaire has a greater factor loading on factor two than one). Also in the interest of comparability, mean item score and Cronbach’s alpha for factor 4 in the German version of the questionnaire were calculated by reversing the scale of item 24 instead of items 28 and 30. Item numbers correspond to items in Appendix A. 1052 N. Rothen et al. / Consciousness and Cognition 22 (2013) 1047–1060 Fig. 1. Mean (N = 53), SD, and test–retest correlation for the 16 final items of the CLaN. Mean scores of items 13, 5, 7, 24 were calculated by reversing their scale to account for their negative factor loadings and comparability with other descriptive measures in this article.  significant at the .01 level and  significant at the .05 level. Item numbers correspond to items in Appendix A. Before we assess the external validity of the 4 factors, it is important to interpret the 4 factors. Factor 1 is based on items related to the localisation of the synaesthetic experience and bears resemblance to the projector–associator dimension. Factor 2 is based on items related to aspects of automaticity and attention of the synaesthetic experience. Higher scores on this factor indicate that the synaesthetic experiences tend to be elicited with greater automaticity and require less attention. Factor 3 is based on items related to the deliberate use of synaesthetic experiences in everyday life. Higher scores on this indicate a tendency to stronger usage of synaesthetic experiences in everyday life. Factor 4 is based on items related to longitudinal change of synaesthetic experiences. Higher scores on this factor indicate that the synaesthetic colours changed intensity over time and were either weaker or stronger in the past. 3. External validity Next, we assessed the external validity of the final 4 factors of the CLaN. Our first goal was to identify factors that are related to synaesthetic consistency and the bandwidth of synaesthetic inducer (i.e., number of items [letters/numbers] which induce synaesthetic experiences). Our second goal was to identify factors correlated with two variants of the synaes- N. Rothen et al. / Consciousness and Cognition 22 (2013) 1047–1060 1053 thetic Stroop test. The latter goal was based on the fact that synaesthetic Stroop tests are probably the most often applied tests in the field of synaesthesia research. The materials and procedure for the Stroop tests were adopted from a previous study (Ward et al., 2007; see also, Dixon et al., 2004). 3.1. Methods 3.1.1. Participants We used the data of those 247 participants who completed both the test of consistency, the questionnaire, and who were not regarded as outliers. In addition, a subset of 18 participants completed two versions of a Stroop test in the laboratory; mean age = 36 years (SD = 11), 6 male and 12 female, all fluent English speakers. 3.1.2. Materials We described the stimuli for the test of consistency already in Section 2 of this article. The stimuli for the Stroop test consisted of 8 different graphemes (letters and digits) and the corresponding synaesthetic colours which varied between individual synaesthetes. The colours were selected on the basis of each individual’s data in the test of consistency. All graphemes existed in their congruent version (i.e., coloured in the corresponding synaesthetic colour) and in different incongruent versions (i.e., reassigned colours that corresponded to one of the other graphemes). 3.1.3. Procedure We described the procedure for the test of consistency already in Section 2 of this article. During the synaesthetic Stroop test, participants were presented with 64 trials. For half of the trials, the grapheme was presented in a colour congruent with their synaesthetic experience and for the other half of the trials the grapheme was presented in a colour incongruent with the synaesthetic experience. The order of stimulus presentation was randomized. Each trial started with a fixation cross presented at the centre of the screen for 1000 ms followed by the grapheme. All stimuli were presented against a grey background. The stimuli remained on the screen until the participant made a response into a microphone. There were two different versions of the Stroop test which differed in the task demands as follows. Participants were required to name either their synaesthetic colour and ignore the veridical colour, or to name the veridical colour and ignore their synaesthetic colour. Each participant did both versions of the Stroop test. The order of the different versions of the Stroop test was counterbalanced across participants. 3.2. Analysis and results We first consider the sample of those 247 participants who completed the questionnaire and the test of consistency. In order to identify factors that are related to synaesthetic consistency and the bandwidth of the synaesthetic inducer, we correlated each of the four factor scores with the synaesthetic consistency value (i.e., distance in CIELUV colour space) and the number of synaesthetic inducers. The average consistency value was 61 (SD = 25) and average number of synaesthetic inducers graphemes was 30 (SD = 9). Notably, because consistency is expressed as distance in colour space lower values denote greater consistency. We identified a significant negative correlation between synaesthetic consistency and the factor deliberate use (r = .146, p < .05). That is, participants with higher scores on the deliberate use factor tend to show greater synaesthetic consistency. Synaesthetic bandwidth was correlated with the factors deliberate use (r = .250, p < .001), localisation (r = .230, p < .001), and automaticity/attention (r = .203, p < .01). That is, higher scores on these factors were linked with greater sets of synaesthetic inducers. No other factor was significantly correlated with synaesthetic consistency and/or bandwidth, respectively. For the Stroop test (i.e., subsample N = 18), incorrect response trials and trials in which the microphone was inappropriately triggered (e.g., because the microphone was accidentally triggered or the response was not detected) were excluded from the reaction time analysis. The mean reaction times and error rates of the Stroop task are depicted in Fig. 2. We conducted a two-factorial Analysis of Variance (ANOVA) with the within subject factors task version (veridical vs. synaesthetic) and congruency (congruent vs. incongruent) to assess synaesthetic Stroop effects in both versions of the task. The ANOVA revealed a main effect congruency F(1,17) = 29.897, p < .001. There were no other significant effects or interactions, all Fs < 1.156, all ps > .297. We explored whether Stroop interference (i.e., incongruent minus congruent) is related to any of the four factors of the CLaN for both versions of the Stroop test separately. For the task of naming synaesthetic colours (and ignoring real colours) there was a significant negative correlation between synaesthetic colour naming Stroop interference and the factor automaticity/attention (r = .551, p < .05). This is shown in Fig. 3. Thus, those synaesthetes who claim that the experience is less automatic show more interference from the incongruent veridical colour. There were no other significant correlations. Considering next the absolute naming times in the two tests (i.e. collapsing across congruency), the factor localisation was significantly correlated with absolute naming times of veridical (r = .529, p < .05) and synaesthetic colours (r = .701, p < .01). That is, a tendency to see colours located on the text is associated with shorter RTs for both, naming synaesthetic colours and real colours. This is shown in Fig. 4. The factor of deliberate use was related to a relative difference in colour naming times across the two tasks (r = .469, p < .05). That is, those participants who claim to use their synaesthesia less in 1054 N. Rothen et al. / Consciousness and Cognition 22 (2013) 1047–1060 everyday life are relatively slower (relative to naming the veridical colours of graphemes) at naming the synaesthetic colours of graphemes. This is shown in Fig. 5. There were no other significant correlations. Although not entirely independent, we were able to confirm the previous finding of a significant correlation between the factor deliberate use of synaesthetic experiences and synaesthetic consistency in CIELUV colour space in this subsample (r = .601, p < .01, Fig. 6). 3.3. Discussion The assessment of the external validity of the CLaN was based on two samples, one including all participants who filled in the questionnaire and passed the test of consistency and the other on a subsample which additionally participated in a Stroop test experiment in the laboratory. The analyses based on the first sample indicated that greater synaesthetic consistency is linked to a greater tendency to use synaesthetic experiences in everyday life (e.g., to remember PIN codes, etc.) and was confirmed in the subsample of the synaesthetes who participated in the laboratory based Stroop experiment. Moreover, having a greater bandwidth in grapheme-colour synaesthesia was related to higher scores on the factors deliberate use, localisation, and automaticity/attention in the sample including all synaesthetes who passed the test of consistency. It is important to note that correlations are not able to indicate the direction of causation. That is, synaesthetes who show greater consistency may use their experiences more in everyday life because they have more consistent experiences and hence, the experiences are more useful. However, it may also be the other way round in that the synaesthetes who deliberately use their experiences more in everyday life are more consistent as a result of this increased use of their experiences. The same applies to the number of synaesthetic inducers and the relationship to the factors deliberate use, localisation, and automaticity/ attention. In a subsample of 18 synaesthetes who participated in a laboratory Stroop test, we replicated previous findings that synaesthetes are slower in naming real and synaesthetic colours that are incongruent with the colour of the eliciting stimulus in comparison to naming real and synaesthetic colours that are congruent with the colour of the eliciting stimulus (Dixon et al., 2004; Ward et al., 2007). Moreover, three of the four factors of the CLaN were meaningfully related to different measures of the Stroop test. That is, (1) those who use their synaesthesia less in everyday life tend to be slower at naming their synaesthetic colours (relative to veridical colours) whereas those who use their synaesthesia more show the opposite profile; (2) a tendency to see colours that have a specific location is associated with shorter RTs for both naming veridical colours and synaesthetic colours in Stroop tests; and (3) a lower tendency to experience synaesthesia automatically is linked to more interference from an incongruent veridical colour when required to name the synaesthetic colour. 4. General discussion By the means of a factor analysis, we created a synaesthesia questionnaire (CLaN) consisting of 16 items which load on 4 different factors: (F1) is based on the location of the synaesthetic experience with higher scores denoting a tendency to see colours that are located, (F2) is based on aspects of automaticity and attention of the synaesthetic experience with higher scores indicating greater automaticity and less attention to the inducing stimulus for the synaesthetic experience to be elicited, (F3) is based on deliberate use of synaesthetic experiences in everyday life with higher scores indicating increased usage of synaesthetic experiences in everyday life, and (F4) is based on longitudinal changes of synaesthetic experiences with higher scores indicating that the synaesthetic colours changed intensity over time. The CLaN exists in two different languages: English and German. Our findings indicate that both versions reliably lead to the same factor structure, are internally consistent, Fig. 2. Mean RTs and mean proportion of error rates of veridical and synaesthetic colour naming Stroop tests. Error bars represent standard errors of RTs. N. Rothen et al. / Consciousness and Cognition 22 (2013) 1047–1060 1055 Fig. 3. A greater tendency to experience synaesthesia automatically without attention is associated with less Stroop interference in the synaesthetic colour naming task. Fig. 4. A tendency to see colours that have a specific location is associated with shorter RTs for both, naming veridical colours and synaesthetic colours in Stroop tests. and exhibit generally good test–retest reliability. Moreover, meaningful correlations between the first three factors of the CLaN and different measures of performance in synaesthesia – including measures of synaesthetic consistency, bandwidth (i.e. number of graphemic inducers that produce colour), and Stroop performance – generally suggested good external validity. The factor that we termed localisation (F1) resembles the projector–associator dimension proposed by others, although it may not be identical to it (for localizer/non-localizer distinction see also, Cytowic & Eagleman, 2011). A high score on this factor would be indicative of a tendency to experience the synaesthetic colours as externally localised to the inducing grapheme itself (i.e., ‘projector’). However, it is less clear what the phenomenology of a low score on this dimension would consist of (except a tendency to not see it localised on the inducer). For instance, claims to ‘‘know but not see’’ synaesthetic colours (Q10 Appendix A) are not significantly loaded onto this factor (as might be expected if ‘associator’ was at the opposite extreme of this dimension), and neither are claims to experience the colours inside the body (Q2 Appendix A) or on an inner screen (Q14 Appendix A). It is possible that these could constitute additional factors that were not revealed because too few questions were included that loaded on them. Recent neuroimaging evidence suggests that such finer cuts might exist. Specifically, those claiming to see colours on a mental screen show greater influence of the parietal lobes in synaesthesia than those claiming to project colours onto the inducers (van Leeuwen et al., 2011). The former may be associated with a slower time course, which could also explain why – in our study – those scoring higher on the localisation factor tended to be faster at naming colours on the Stroop tests. Another interesting finding of the current study was that higher scores on F2 deliberate use (i.e., the most stable factor) of synaesthetic experiences in everyday life (e.g., as mnemonic) were associated with greater synaesthetic consistency measures as smaller Euclidean distances in CIELUV colour space. Since correlational results do not indicate a causal relationship between two variables this effect is somewhat difficult to interpret. Three interpretations seem possible. Although they 1056 N. Rothen et al. / Consciousness and Cognition 22 (2013) 1047–1060 Fig. 5. A greater tendency to use synaesthetic experiences in everyday life is associated with longer RTs in veridical colour naming as compared to synaesthetic colour naming, and vice versa a smaller tendency to use synaesthetic experiences in everyday life is associated with longer RTs in synaesthetic colour naming as compared to veridical colour naming. Fig. 6. Participants with higher scores on the deliberate use factor show greater synaesthetic consistency (as indicated by smaller Euclidean distances in CIELUV colour space). are related to different aspects of synaesthesia they are not mutually exclusive. The first interpretation we will discuss is related to enhanced memory performance in synaesthesia. High consistency is crucial in order to use synaesthetic colour information as memory retrieval cues. Hence, greater consistency in synaesthetic experiences would lead to better memory or, at least, to more reliable possibilities to use synaesthetic experiences as a memory aid in everyday life. Related to the notion that synaesthesia may indeed enhance memory performance (e.g., Pritchard, Rothen, Coolbear, & Ward, 2013; Rothen & Meier, 2010b; for a review see, Rothen, Meier, & Ward, 2012; see also, Meier & Rothen, in press), it is important to mention that the most recent theory has linked enhanced memory with enhanced perception in synaesthesia and not with a benefit from additional memory cues (Rothen et al., 2012). A second possibility to interpret the correlation between deliberate use of synaesthetic experiences in everyday life and synaesthetic consistency is related to the development of synaesthesia. Synaesthetic consistency may simply evolve from deliberate use of synaesthesia. That is the repeated use of colour associations in response to graphemes will consolidate the specific associations entirely in the sense of neurons that fire together wire together (Hebb, 1949). In line with this notion it is no surprise that there is a higher prevalence of grapheme-colour synaesthetes in fine-art students, since specific characteristics of artistic techniques may support the evolution of synaesthetic associations (Rothen & Meier, 2010a). Moreover, a recent study provided evidence that specific synaesthetic associations may indeed be learnt from childhood toys (Witthoft & Winawer, 2013). However, training synaesthetic associations has yet not been successful in acquiring synaesthesia (e.g., Colizoli, Murre, & Rouw, 2012; Kusnir & Thut, 2012; Meier & Rothen, 2009; Rothen, Nikolić, et al., 2013; Rothen et al., 2011). N. Rothen et al. / Consciousness and Cognition 22 (2013) 1047–1060 1057 A third possibility to interpret the correlation between deliberate use of synaesthetic experiences in everyday life and synaesthetic consistency is related to consciousness and cross-modal correspondences. Recently it has been hypothesised that synaesthesia may be strong and conscious form of binding which also exists in non-synaesthetes (e.g., non-random pitch-colour associations) (Rothen & Terhune, 2012), but not to a degree to be accessible by the means of conscious awareness nor to achieve the level of consistency observed in synaesthetes (Cohen Kadosh & Henik, 2007; but see Deroy & Spence, 2013). In line with this notion the degree of conscious awareness of grapheme-colour associations might be a covert moderator variable between the deliberate use of synaesthetic experiences and consistency. Higher scores on the automaticity/attention factor were found to be associated with smaller Stroop interference in the synaesthetic colour naming task. This finding is not surprising and suggests that, for less automatic synaesthetic experiences, more effort is required in the production of the synaesthetic colour name when the real colour on the screen is incongruent. Moreover, the result is meaningful in that the automaticity/attention factor is associated with a task which is said to measure the automaticity of associations (e.g., Rothen, Nikolić, et al., 2013). The longitudinal change factor of synaesthetic experiences was not related to any of our performance measures (i.e., consistency and Stroop). This may be understood in the light that changes in the synaesthetic experience over a prolonged period of time may be related to the age of the individual. However, since age was not asked for in our questionnaire or the other tests used (Eagleman et al., 2007), we may have missed interesting insights and it is clearly a challenge for future research to uncover the relationship between the factor longitudinal change of the CLaN and age and task performance. Hence, future research will need to show if this factor is externally valid or not. Overall, our findings from the Stroop task are consistent with previous studies using this task as either diagnostic marker for synaesthesia or as a paradigm to assess individual differences in synaesthesia. Regarding the diagnostic aspects of this task, we replicated previous findings that synaesthetes are slower in naming real and synaesthetic colours that are incongruent with the colour of the eliciting stimulus in comparison to naming real and synaesthetic colours that are congruent with the colour of the eliciting stimulus (e.g., Nikolić et al., 2011). Regarding the assessment of individual differences in grapheme-colour synaesthesia, two studies have to be mentioned explicitly. It was found that projector synaesthetes were generally faster in synaesthetic colour naming as compared to veridical colour naming and associator synaesthetes were generally slower in synaesthetic colour naming as compared to veridical colour naming (Dixon et al., 2004; Ward et al., 2007). Furthermore, the overall naming times were shorter for projector synaesthetes as compared to associator synaesthetes (Dixon et al., 2004; and numerically in Ward et al. (2007)). Our results are broadly consistent with these findings in that a tendency to see colours localised on the graphemes is associated with shorter naming times for both, veridical colours and synaesthetic colours in the Stroop tests. The finding that higher scores on the localisation factor are associated with quicker naming times for veridical colours may be particularly interesting. Recent studies suggest generally enhanced colour processing in grapheme-colour synaesthetes (Banissy, Walsh, & Ward, 2009; Terhune, Wudarczyk, Kochuparampil, & Cohen Kadosh, in press; Yaro & Ward, 2007). Hence, it would not be surprising if higher scores on the localisation factor would be associated with better colour discrimination. In line with this, it would be interesting to see how each factor relates to individual differences in synaesthetic colours based on the frequency of inducer graphemes (Beeli, Esslen, & Jäncke, 2007; Cohen Kadosh, Henik, & Walsh, 2007; Simner & Ward, 2008; Simner et al., 2005; Smilek, Carriere, Dixon, & Merikle, 2007) or the shape of inducer graphemes (Albertazzi et al., 2013; Brang, Rouw, Ramachandran, & Coulson, 2011; Watson, Akins, & Enns, 2012). Moreover, as already mentioned in the introduction, ‘synaesthetic’ Stroop effects may also be found in non-synaesthetes who trained similar associations (Meier & Rothen, 2009; Rothen, Nikolić, et al., 2013; Rothen et al., 2011). Future research will have to show how phenomenological reports for trained colour associations are related to various measures deriving from the Stroop task. To conclude, the obtained results suggest that the CLaN has good test–retest reliability and that the subscales have good construct validity. As discussed, the CLaN provides interesting results which not only corroborate existing ideas in synaesthesia research but also provide grounds for new ideas that can be tested in future research. Moreover, although the questionnaire was developed to be used as an instrument to measure individual differences in grapheme-colour synaesthesia, it may be adapted for the use with other forms of synaesthesia which may engage similar cognitive mechanisms. Acknowledgments N.R. is supported by the Swiss National Science Foundation and the Holcim Foundation for the Advancement of Scientific Research, Switzerland, with additional support from the Dr. Mortimer and Theresa Sackler Foundation. Appendix A Initial 30 items in the order of presentation (final 16 items highlighted). 1058 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. N. Rothen et al. / Consciousness and Cognition 22 (2013) 1047–1060 I can choose to alter the intensity of the synaesthetic colour I see the synaesthetic colours literally inside my body (e.g., behind my forehead) I experience the synaesthetic colours even if I do not attend to them specifically (e.g., while reading a book) I see the synaesthetic colours on the computer screen (or very close to it) It feels like I have to go and fetch the colours, rather than the colours coming to me I experience the synaesthetic colours in several locations at the same time (for instance, both on the screen and literally inside my head or some other combination) I only experience the synaesthetic colours of letters/numbers if I think about them as having a colour When I am looking quickly at a page of a book the synaesthetic colours appear before I am aware of what the letters/words are I know exactly what colour goes with a particular letter/number My synaesthetic experience of colours feels more like knowing than seeing The synaesthetic colours appear weak and faded I can choose to alter the location of the synaesthetic colours (e.g., by projecting them onto the wall) I do not ‘‘see’’ colours when I look at the letters/numbers I see my synaesthetic colours on an imagined screen that has no physical location that I can point to The synaesthetic colours appear vivid It seems that the colour is on the screen where the letter/number is printed The synaesthetic colours appear automatically without any effort on my part If I am asked to imagine letters or numbers as having different colours from my own (e.g., A = white, B = red, C = pink) then it is easy for me to do so The colours have the same shape as the letters/numbers I see the synaesthetic colours in the space outside my body (e.g., 10 inches in front of my tummy, or above my shoulder) but not on the computer screen I can point to the location of the synaesthetic colours When I am introduced to a new person, I think about the colour of their name When I see a colour I automatically experience the letter/number My synaesthetic colours did not change their intensity over the years I use my synaesthetic colours deliberately for remembering sequences of numbers (e.g., PINs, telephone numbers) I deliberately try to use my synaesthetic colours in my everyday life I use my synaesthetic colours to remember dates and plan appointments (e.g., 28.02.2010) My synaesthetic colours were weaker in the past (i.e., years ago) When I see a colour I do not automatically experience the letter/number My synaesthetic colours were stronger in the past (i.e., years ago) Appendix B Final 16 items of the CLaN 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. I experience the synaesthetic colours even if I do not attend to them specifically (e.g., while reading a book) I see the synaesthetic colours on the computer screen (or very close to it) It feels like I have to go and fetch the colours, rather than the colours coming to me I experience the synaesthetic colours in several locations at the same time (for instance, both on the screen and literally inside my head or some other combination) I only experience the synaesthetic colours of letters/numbers if I think about them as having a colour When I am looking quickly at a page of a book the synaesthetic colours appear before I am aware of what the letters/words are I do not ‘‘see’’ colours when I look at the letters/numbers It seems that the colour is on the screen where the letter/number is printed The synaesthetic colours appear automatically without any effort on my part I can point to the location of the synaesthetic colours My synaesthetic colours did not change their intensity over the years I use my synaesthetic colours deliberately for remembering sequences of numbers (e.g., PINs, telephone numbers) I deliberately try to use my synaesthetic colours in my everyday life I use my synaesthetic colours to remember dates and plan appointments (e.g., 28.02.2010) My synaesthetic colours were weaker in the past (i.e., years ago) My synaesthetic colours were stronger in the past (i.e., years ago) N. Rothen et al. / Consciousness and Cognition 22 (2013) 1047–1060 1059 Note that the items were re-enumerated and hence do not have the same number as in Appendix A. Important, questions are scored on a 5-point Likert-scale from 1 to 5 (i.e., strongly disagree, moderately disagree, neither agree nor disagree, moderately agree, strongly agree). Item number 3, 5, 7, and 11 need to be reverse coded to be scored (i.e., 5–1 applies to strongly disagree, moderately disagree, neither agree nor disagree, moderately agree, strongly agree). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.concog.2013.07.005. References Albertazzi, L., Da Pos, O., Canal, L., Micciolo, R., Malfatti, M., & Vescovi, M. (2013). The hue of shapes. Journal of Experimental Psychology: Human Perception and Performance, 39(1), 37–47. http://dx.doi.org/10.1037/a0028816. Banissy, M. J., Walsh, V., & Ward, J. (2009). Enhanced sensory perception in synaesthesia. Experimental Brain Research, 196(4), 565–571. http://dx.doi.org/ 10.1007/s00221-009-1888-0. Banissy, M. J., & Ward, J. (2007). Mirror-touch synesthesia is linked with empathy. Nature Neuroscience, 10(7), 815–816. http://dx.doi.org/10.1038/nn1926. Barnett, K. J., Finucane, C., Asher, J. E., Bargary, G., Corvin, A. P., Newell, F. N., et al (2008). Familial patterns and the origins of individual differences in synaesthesia. Cognition, 106(2), 871–893. http://dx.doi.org/10.1016/j.cognition.2007.05.003. Barnett, K. J., & Newell, F. N. (2008). Synaesthesia is associated with enhanced, self-rated visual imagery. Consciousness and Cognition, 17(3), 1032–1039. http://dx.doi.org/10.1016/j.concog.2007.05.011. Baron-Cohen, S., Harrison, J., Goldstein, L. H., & Wyke, M. (1993). Coloured speech perception: Is synaesthesia what happens when modularity breaks down? Perception, 22(4), 419–426. Beeli, G., Esslen, M., & Jäncke, L. (2007). Frequency correlates in grapheme-color synaesthesia. Psychological Science, 18(9), 788–792. http://dx.doi.org/ 10.1111/j.1467-9280.2007.01980.x. Brainard, D. H., Pelli, D. G., & Robson, T. (2002). Encyclopedia of imaging science and technology. In J. Hornak (Ed.), Display characterization (pp. 172–188). New York: Wiley. Brang, D., Rouw, R., Ramachandran, V. S., & Coulson, S. (2011). Similarly shaped letters evoke similar colors in grapheme-color synesthesia. Neuropsychologia, 49(5), 1355–1358. http://dx.doi.org/10.1016/j.neuropsychologia.2011.01.002. Cohen Kadosh, R., & Henik, A. (2007). Can synaesthesia research inform cognitive science? Trends in Cognitive Sciences, 11(4), 177–184. http://dx.doi.org/ 10.1016/j.tics.2007.01.003. Cohen Kadosh, R., Henik, A., & Walsh, V. (2007). Small is bright and big is dark in synaesthesia. Current Biology, 17(19), R834–R835. http://dx.doi.org/ 10.1016/j.cub.2007.07.048. Colizoli, O., Murre, J. M. J., & Rouw, R. (2012). Pseudo-synesthesia through reading books with colored letters. PLoS ONE, 7(6), e39799. http://dx.doi.org/ 10.1371/journal.pone.0039799. Cytowic, R. E., & Eagleman, D. M. (2011). Wednesday is indigo blue: Discovering the brain of synesthesia. The MIT Press. Deroy, O., & Spence, C. (2013). Why we are not all synesthetes (not even weakly so). Psychonomic Bulletin & Review. http://link.springer.com/article/10.3758/ s13423-013-0387-2/fulltext.html. Dixon, M. J., Smilek, D., & Merikle, P. M. (2004). Not all synaesthetes are created equal: Projector versus associator synaesthetes. Cognitive, Affective, & Behavioral Neuroscience, 4(3), 335–343. Domino, G. (1989). Synaesthesia and creativity in fine arts students: An empirical look. Creativity Research Journal, 2, 17–29. Eagleman, D. M., Kagan, A. D., Nelson, S. S., Sagaram, D., & Sarma, A. K. (2007). A standardized test battery for the study of synesthesia. Journal of Neuroscience Methods, 159(1), 139–145. http://dx.doi.org/10.1016/j.jneumeth.2006.07.012. Gebuis, T., Nijboer, T. C. W., & van der Smagt, M. J. (2009). Multiple dimensions in bi-directional synesthesia. European Journal of Neuroscience, 29(8), 1703–1710. http://dx.doi.org/10.1111/j.1460-9568.2009.06699.x. Glicksohn, J., Steinbach, I., & Elimalach-Malmilyan, S. (1999). Cognitive dedifferentiation in eidetics and synaesthesia: Hunting for the ghost once more. Perception, 28(1), 109–120. http://dx.doi.org/10.1068/p2755. Hebb, D. O. (1949). The organization of behavior. New York: Wiley. Hubbard, E. M., Arman, A. C., Ramachandran, V. S., & Boynton, G. M. (2005). Individual differences among grapheme-color synesthetes: Brain-behavior correlations. Neuron, 45(6), 975–985. http://dx.doi.org/10.1016/j.neuron.2005.02.008. Hunt, R. W. G., & Pointer, M. R. (2011). Measuring colour (4th ed.). Chichester, UK: John Wiley & Sons, Ltd.. <http://onlinelibrary.wiley.com/doi/10.1002/ 9781119975595.index/summary>. Kusnir, F., & Thut, G. (2012). Formation of automatic letter–colour associations in non-synaesthetes through likelihood manipulation of letter–colour pairings. Neuropsychologia, 50(14), 3641–3652. http://dx.doi.org/10.1016/j.neuropsychologia.2012.09.032. Meier, B., & Rothen, N. (in press). Synaesthesia and memory. In J. Simner & E. M. Hubbard (Eds.), Oxford handbook of synaesthesia. Oxford, UK: Oxford University Press. Meier, B., & Rothen, N. (2009). Training grapheme-colour associations produces a synaesthetic Stroop effect, but not a conditioned synaesthetic response. Neuropsychologia, 47(4), 1208–1211. http://dx.doi.org/10.1016/j.neuropsychologia.2009.01.009. Nikolić, D., Jürgens, U. M., Rothen, N., Meier, B., & Mroczko, A. (2011). Swimming-style synesthesia. Cortex, 47(7), 874–879. doi: 16/j.cortex.2011.02.008. Pritchard, J., Rothen, N., Coolbear, D., & Ward, J. (2013). Enhanced associative memory for colour (but not shape or location) in synaesthesia. Cognition, 127(2), 230–234. http://dx.doi.org/10.1016/j.cognition.2012.12.012. Rich, A. N., Bradshaw, J. L., & Mattingley, J. B. (2005). A systematic, large-scale study of synaesthesia: Implications for the role of early experience in lexicalcolour associations. Cognition, 98(1), 53–84. http://dx.doi.org/10.1016/j.cognition.2004.11.003. Rothen, N., & Meier, B. (2010a). Higher prevalence of synaesthesia in art students. Perception, 39(5), 718–720. Rothen, N., & Meier, B. (2010b). Grapheme-colour synaesthesia yields an ordinary rather than extraordinary memory advantage: Evidence from a group study. Memory, 18(3), 258–264. http://dx.doi.org/10.1080/09658210903527308. Rothen, N., & Meier, B. (2013). Why vicarious experience is not an instance of synesthesia. Frontiers in Human Neuroscience, 128. http://dx.doi.org/10.3389/ fnhum.2013.00128. Rothen, N., Meier, B., & Ward, J. (2012). Enhanced memory ability: Insights from synaesthesia. Neuroscience & Biobehavioral Reviews, 36(8), 1952–1963. http://dx.doi.org/10.1016/j.neubiorev.2012.05.004. Rothen, N., Nikolić, D., Jürgens, U. M., Mroczko-Wa˛sowicz, A., Cock, J., & Meier, B. (2013a). Psychophysiological evidence for the genuineness of swimmingstyle colour synaesthesia. Consciousness and Cognition, 22(1), 35–46. http://dx.doi.org/10.1016/j.concog.2012.11.005. Rothen, N., Seth, A. K., Witzel, C., & Ward, J. (2013b). Diagnosing synaesthesia with online colour pickers: Maximising sensitivity and specificity. Journal of Neuroscience Methods, 215(1), 156–160. http://dx.doi.org/10.1016/j.jneumeth.2013.02.009. Rothen, N., & Terhune, D. B. (2012). Increased resting state network connectivity in synesthesia: Evidence for a neural basis of synesthetic consistency. The Journal of Neuroscience, 32(40), 13641–13643. http://dx.doi.org/10.1523/JNEUROSCI.3577-12.2012. 1060 N. Rothen et al. / Consciousness and Cognition 22 (2013) 1047–1060 Rothen, N., Wantz, A.-L., & Meier, B. (2011). Training synaesthesia. Perception, 40(10), 1248–1250. Rouw, R., & Scholte, H. S. (2007). Increased structural connectivity in grapheme-color synesthesia. Nature Neuroscience, 10(6), 792–797. http://dx.doi.org/ 10.1038/nn1906. Rouw, R., & Scholte, H. S. (2010). Neural basis of individual differences in synesthetic experiences. The Journal of Neuroscience, 30(18), 6205–6213. http:// dx.doi.org/10.1523/JNEUROSCI.3444-09.2010. Simner, J., Harrold, J., Creed, H., Monro, L., & Foulkes, L. (2009a). Early detection of markers for synaesthesia in childhood populations. Brain, 132(1), 57–64. http://dx.doi.org/10.1093/brain/awn292. Simner, J., Mayo, N., & Spiller, M.-J. (2009b). A foundation for savantism? Visuo-spatial synaesthetes present with cognitive benefits. Cortex, 45(10), 1246–1260. http://dx.doi.org/10.1016/j.cortex.2009.07.007. Simner, J., Mulvenna, C., Sagiv, N., Tsakanikos, E., Witherby, S. A., Fraser, C., et al (2006). Synaesthesia: The prevalence of atypical cross-modal experiences. Perception, 35(8), 1024–1033. Simner, J., & Ward, J. (2008). Synaesthesia, color terms, and color space: Color claims came from color names in Beeli, Esslen, and Jäncke (2007). Psychological Science, 19(4), 412–414. http://dx.doi.org/10.1111/j.1467-9280.2008.02101.x. Simner, J., Ward, J., Lanz, M., Jansari, A., Noonan, K., Glover, L., et al (2005). Non-random associations of graphemes to colours in synaesthetic and nonsynaesthetic populations. Cognitive Neuropsychology, 22, 1069–1085. http://dx.doi.org/10.1080/02643290500200122. Skelton, R., Ludwig, C., & Mohr, C. (2009). A novel, illustrated questionnaire to distinguish projector and associator synaesthetes. Cortex, 45(6), 721–729. http://dx.doi.org/10.1016/j.cortex.2008.02.006. Smilek, D., Callejas, A., Dixon, M. J., & Merikle, P. M. (2007a). Ovals of time: Time–space associations in synaesthesia. Consciousness and Cognition, 16(2), 507–519. http://dx.doi.org/10.1016/j.concog.2006.06.013. Smilek, D., Carriere, J. S. A., Dixon, M. J., & Merikle, P. M. (2007b). Grapheme frequency and color luminance in grapheme-color synaesthesia. Psychological Science, 18(9), 793–795. http://dx.doi.org/10.1111/j.1467-9280.2007.01981.x. Stokes, M., Anderson, M., & Chandrasekar, S. (1996). A Standard Default Color Space for the Internet – sRGB. <http://www.w3.org/Graphics/Color/ sRGB.html> Retrieved 09.07.13. Terhune, D. B., Wudarczyk, O. A., Kochuparampil, P., & Cohen Kadosh, R. (in press). Enhanced dimension-specific visual working memory in grapheme-color synesthesia. Cognition. Van Leeuwen, T. M., den Ouden, H. E. M., & Hagoort, P. (2011). Effective connectivity determines the nature of subjective experience in grapheme-color synesthesia. The Journal of Neuroscience, 31(27), 9879–9884. http://dx.doi.org/10.1523/JNEUROSCI.0569-11.2011. Walsh, R. (2005). Can synaesthesia be cultivated?: Indications from surveys of meditators. Journal of Consciousness Studies, 12, 5–17. Ward, J., Li, R., Salih, S., & Sagiv, N. (2007). Varieties of grapheme-colour synaesthesia: A new theory of phenomenological and behavioural differences. Consciousness and Cognition, 16(4), 913–931. http://dx.doi.org/10.1016/j.concog.2006.09.012. Watson, M., Akins, K., & Enns, J. (2012). Second-order mappings in grapheme-color synesthesia. Psychonomic Bulletin & Review, 19(2), 211–217. http:// dx.doi.org/10.3758/s13423-011-0208-4. Witthoft, N., & Winawer, J. (2013). Learning, memory, and synesthesia. Psychological Science. http://dx.doi.org/10.1177/0956797612452573. Yaro, C., & Ward, J. (2007). Searching for Shereshevskii: What is superior about the memory of synaesthetes? The Quarterly Journal of Experimental Psychology, 60(5), 681–695. http://dx.doi.org/10.1080/17470210600785208.

Consciousness and Cognition Consciousness and Cognition 14 (2005) 296–303 www.elsevier.com/locate/concog Hypnotic susceptibility, baseline attentional functioning, and the Stroop task Sandro Rubichia,*, Federico Riccib, Roberto Padovanic, Lorenzo Scagliettib a Dipartimento di Scienze Sociali, Cognitive e Quantitative, Reggio Emilia, Italy b Scuola Europea di Psicoterapia Ipnotica, Milan, Italy c Dipartimento di Scienze Biomediche, Università di Modena e Reggio Emilia, Italy Received 23 February 2004 Available online 11 September 2004 Abstract According to the theoretical framework relating hypnosis to attention, baseline attentional functioning in highly hypnotizable individuals should be more efficient than in low hypnotizable individuals. However, previous studies did not find differences in Stroop-like tasks in which the measure indicative of the Stroop interference effect was based on response latencies. This study was designed to determine whether subjects with different levels of hypnotic susceptibility show differences in baseline attentional functioning. To assess this hypothesis, high, medium, and low hypnotizable subjects performed a Stroop task designed to evaluate accuracy performance, before being subjected to hypnotic induction. Results showed that the Stroop interference effect was smaller in high hypnotizable subjects than in low hypnotizable subjects, whereas it was not different between high, and medium hypnotizable subjects. This outcome supports the notion that baseline attentional functioning is related to hypnotic susceptibility. Ó 2004 Elsevier Inc. All rights reserved. Keywords: Hypnosis; Attentional functioning; Stroop task; Hypnotic susceptibility * Corresponding author. E-mail address: rubichi@unimore.it (S. Rubichi). 1053-8100/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2004.08.003 S. Rubichi et al. / Consciousness and Cognition 14 (2005) 296–303 297 1. Introduction The recent research on theoretical and experimental hypnosis underlines the study of cognitive and neurophysiological mechanisms related to hypnotic susceptibility as one of its major interests (see e.g., Raz & Shapiro, 2002; for a review). There is evidence indicating that the attentional system is involved in determining the responsiveness to the hypnotic induction. From an experimental point of view, this issue is studied by comparing the performance in attentional functioning between individuals who show different responsiveness to hypnosis, namely high, and low susceptible subjects. In this framework, the neuropsychophysiological model (e.g., Crawford, 1994; Crawford & Gruzelier, 1992; Gruzelier, 1988, 1998) emphasizes the stages in which the traditional hypnotic relaxation induction occurs. In the first stage, focused, and selective attention processes would be engaged to focus subjectÕs attention on hypnotistÕs voice. A general decrease of attentional functioning is assumed to underlie the second and third stages of induction. This would lead to the suspension of critical evaluation and to the deep relaxation state that are peculiar to the hypnotic state. Thus, the neuropsychophysiological model posits that high hypnotizable (HH) subjects, compared to low hypnotizable (LH) subjects, should manifest (a) a higher performance in attentional functioning in the baseline state, and (b) a decrease of attentional functioning after the hypnotic induction. In other words, attentional processing by HH subjects compared to LH subjects should be more efficient in the baseline state and less efficient in the hypnotic state. There are both neurophysiological and behavioral data supporting the decrease of attentional functioning after hypnotic induction. For example, a strong inhibition of the activity of the frontal cerebral regions has been shown (Kaiser, Barker, Haenschel, Baldeweg, & Gruzelier, 1997). This has been associated with the gradual loss by the subject of both the critical evaluation of reality and the awareness of motor volition typical of the hypnotic state. Later, two contemporary shifts of the general cerebral activation from left to right regions and from anterior to posterior areas take place, as supported by neuropsychological (Crawford & Allen, 1983; Gruzelier & Warren, 1993) and electrophysiological (Mészáros, Crawford, Szabó, Nagy-Kovács, & Résvész, 1989) evidence regarding measures of laterality (but see, for example, Jasiukaitis, Nouriani, Hugdahl, & Spiegel, 1997; for dissenting results). Given that frontal lobes are extensively involved in attentional networks (e.g., Posner & Raichle, 1997), these data strongly support the relevance of the attentional system for hypnotic susceptibility. In addition, HH subjects during hypnosis score lower than LH subjects on a task measuring verbal fluency to letter-designated categories, whereas there is no difference in verbal fluency to semantically designated categories (Gruzelier & Warren, 1993; Kallio et al., 2001). The dissociation between letter fluency and semantic fluency is also seen in patients with lesions in the left dorsal lateral or medial frontal lobe. This suggests that an alteration of activity in these areas might affect processing when HH subjects are under hypnosis. The clearest support to the notion that the attentional system of HH subjects functions less efficiently than that of LH subjects after hypnotic induction is provided by the performance on the Stroop task (Stroop, 1935; see MacLeod, 1991; for a review). In it, subjects are required to name the ink color of a colored word. In the case of incongruent color word (e.g., the word GREEN displayed in red ink), subjects usually make more errors and are slower with respect to both a congruent word (e.g., the word RED diplayed in red ink) and a control colored string (e.g., the string 298 S. Rubichi et al. / Consciousness and Cognition 14 (2005) 296–303 XXXX displayed in red ink). Attentional researchers refer to this phenomenon as the Stroop Interference Effect (SIE). The accounts of the SIE are based on the assumption that reading words is automatic and mandatory and that the processing of the word accesses the wordÕs meaning even if subjects are instructed to ignore it and attend only the wordÕs ink (but see Raz, Shapiro, Fan, & Posner, 2002a, 2002b; Raz et al., 2003; for the modulation of the SIE with hypnotic suggestion). Basically, both the relevant (ink color) and irrelevant (meaning) information would be processed and would compete for the activation and emission of the response. In the case of an incongruent word, two competing responses are activated, thus generating errors and slowing of the performance. The Stroop task is considered a measure of the cognitive abilities that are fundamental for the selection of relevant information and the suppression of irrelevant information. Compared to LH subjects, HH subjects under hypnosis show an increase in the SIE as indexed by the number of errors (Kaiser et al., 1997; Nordby et al., 1999) and by the time difference in response latencies (Blum & Graef, 1971; Sheenan, Donovan, & MacLeod, 1988). More controversial is whether at the baseline state, that is, out of hypnosis, HH subjects perform better than LH subjects in attentional tasks. Here, we will briefly review the works that investigated baseline attentional functioning, with specific attention to those studies, where Stroop-like tasks were employed. Kallio, Revonsuo, Hamalainen, Markela, and Gruzelier (2001) found no difference between HH and LH subjects in four neuropsychological tests assumed to assess frontal functions (Stroop task, verbal fluency tasks, reaction time tasks, and vigilance task). In the Stroop task that they developed, subjects were required to name aloud, on a handout, the color of control colored strings (Xes-strings) and incongruent words. The author measured the time required to name colors in the two conditions and found that HH and LH subjects performed very similarly in the baseline state. Similar results were also reported by Aikins and Ray (2001) that presented HH and LH subjects with different neuropsychological tests of executive functioning (word fluency to letter, Stroop test, Towers of Hanoi, Wisconsin Card Sorting test) in the baseline state. The authors found that HH subjects were better than LH subjects in the task-shifting ability only, as measured by the Wisconsing Card Sorting Task (Heaton, 1981). The Stroop task that they used was administered by a computer program that generated screens of 50 incongruent words each. The response time of five screens were recorded. No control or congruent conditions were provided, thus the performances referred to the incongruent condition latencies only. Again and consistently with Kallio et al.Õs study (2001), no difference was found between HH and LH subjects in the baseline state. Stroop-like tasks were also employed in two works by Dixon and collaborators (Dixon, Brunet, & Laurence, 1990; Dixon & Laurence, 1992). In their first work, the authors employed a Stroop task introduced by Cheesman and Merikle (1986) in which visual words were presented with (a) different congruent-trial probability (the ratio between congruent and incongruent stimuli was either 25 or 75%), and (b) perceptually degraded letters that delayed the time necessary to recognize words. The degraded words were presented in black ink color while a colored rectangular appeared around the stimuli. Participants were requested to name the ink of the rectangular border around the words. HH subject had a larger SIE as measured by the difference in the reaction time of congruent and incongruent stimuli than both LH subjects and medium hypnotizable (MH) subjects. Such a finding was described as supporting the notion that HH subjects process words more automatically than LH subjects and was confirmed in a later study (Dixon & Laurence, 1992). However, given that the SIE was more pronounced in HH than in LH individuals, S. Rubichi et al. / Consciousness and Cognition 14 (2005) 296–303 299 this finding seems to contradict the idea that baseline attentional functioning is more efficient in HH subjects. To sum up, while there is evidence that a transient decrease of attentional functioning is related to high hypnotizability, it is still not clear whether HH subjects have a more efficient functioning of the attentional system at the baseline state. Recall that the neuropsychophysiological model (e.g., Crawford, 1994; Crawford & Gruzelier, 1992; Gruzelier, 1988, 1998) predicts that HH subjects, compared to LH subjects, should manifest a higher performance in attentional functioning at baseline, which allows them to focus attention on instructions during hypnotic induction. We believe that among attentional mechanisms, the ability to select, and focus on the relevant information and to suppress irrelevant information is fundamental for complying with the instructions given during hypnotic induction. Therefore, it seems surprising that previous studies on the Stroop task did not find empirical support of the expected more efficient performance of HH subjects. A possible reason might rely on the kind of Stroop task used and on the measures that have been considered as indicative of the SIE. Even if very different Stroop-like tasks were so far used, all of them had focused on the time needed to perform the task, and there is evidence that other measures are, in some cases, more sensitive than reaction times for discriminating different populations. More specifically, accuracy performance as measured by the number of errors in the Stroop task have suggested, in some works, discriminating between frontal damaged patients, and normal control subjects, while reaction times have not (Swick & Jovanovic, 2002; Vendrell et al., 1995). The main purpose of the present work was to investigate whether high susceptibility to hypnosis is related to a more efficient performance on the Stroop task in the baseline state. We decided to design a Stroop task along the following lines. In our view, the Stroop task proposed by Kallio et al. (2001) is the most suitable for studying the selection/suppression ability during a given period of time. In it, handouts with neutral (control) or incongruent stimuli were given, and the time required to name colors in the two conditions was recorded. In this Stroop task, the selection/suppression ability must be maintained over time, until all the stimuli present in the handout are named. This situation strictly mimics the ability to focus on hypnotic instructions, which is supposed to be more efficient in HH subjects. For the reasons exposed above, instead of recording the time needed to name in the two conditions, we preferred to give a settled period of time for naming neutral, and incongruent handouts. The measures considered as indicative of the SIE were the number of correct namings in the two conditions and the errors (see Section 2 for a detailed description). 2. Methods 2.1. Subjects Participants were volunteers among 100 students of the University of Modena and Reggio Emilia who decided to attend, for extra credit, an introductory course on hypnosis. Thirty-four subjects (13 male and 21 female), who were unaware of the experimental aims, took part in the experimental session. Written informed consent was obtained from all participants to investigate the relation between hypnosis and cognitive performance. 300 S. Rubichi et al. / Consciousness and Cognition 14 (2005) 296–303 The Italian version of the Stanford Hypnotic Susceptibility Scales Form A (henceforth SHSSA; Weitzenhoffer & Hilgard, 1959) was used to categorize the hypnotic susceptibility of each subject as low (score of 4 or less out of a possible 12), medium (score between 5 and 7 out of a possible 12) or high (score of 8 or more out of a possible 12). One male participant interrupted the administration of SHSS-A after six of the 11 items scale because, as he said, he was afraid of the loss of control associated with the hypnotic induction. Therefore, his data were not considered. The remaining subjects were classified as LH subjects (7 participants), MH subjects (6 participants) and HH subjects (20 participants). 2.2. Materials and apparatus The Stroop task designed for the present study consisted of two conditions. In the Neutral Naming (NN) condition, subjects were asked to name filled circles that were colored in blue, green or red ink. In the Incongruent Naming (IN) condition subjects were required to name the ink color of the colored words BLUE, GREEN, or RED that were always colored with incongruent inks. The stimuli of the two conditions were printed on A3 handouts organized in six columns of ten items each. In each condition subjects were required to name as many color as they could in 20 s. Before the testing session, each participant performed a training for both NN and IN conditions with 60 stimuli. The SIE was calculated as follows: (NN IN)/(NN + IN), where NN and IN were the number of correct responses for the Neutral Naming and Incongruent Naming tasks, respectively. Therefore, small SIEs were indexed by small ratio values, whereas greater ratio values reflected larger SIEs. This manner of calculating the SIE considered both the difference in correct responses between the incongruent and neutral conditions and the total number of correct responses given by each participant. Thus, a good performance is indexed by a high number of correct responses on the whole and a small difference of correct responses between the two conditions. Both explicit naming errors and those responses with corrections were considered as incorrect responses. 2.3. Design and procedure Each subject was tested individually, performing the Stroop task before hypnotic induction. The Stroop task and the hypnotic induction were administered by different experimenters. After the Stroop task, participants underwent the hypnotic induction with the administration of the SHSS-A that allowed us to classify them as LH, MH, or HH subjects. 3. Results Table 1 presents the mean scores and the standard deviations for the SHSS-A and the Stroop task as a function of the 3 groups (i.e., LH, MH, and HH group of subjects). An analysis of variance (ANOVA) considering Group (high, medium, and low suggestible subjects) as a between-subjects factor and the SIE score as a dependent variable, was significant, F2,32 = 8.15, p = .0015. Planned comparisons revealed that the SIE value of LH subjects was S. Rubichi et al. / Consciousness and Cognition 14 (2005) 296–303 301 Table 1 Mean scores (M) and standard deviations (SD) of SHSS-A and SIE values as a function of the three groups LH SHSS-A SIE MH HH M SD M SD M SD 3.14 .32 .90 .07 6.17 .21 .75 .07 9.90 .19 1.33 .08 SHSS, Stanford Hypnotic Susceptibility Scale form A; SIE, Stroop Interference Effect. significantly lower than that of both MH (p = .033) and HH (p = .001) subjects, whose scores did not significantly differ. A second ANOVA with Group (high, medium, and low suggestible subjects) as a between-subjects factor and Error Type (neutral naming and incongruent naming) as within-subjects factor was computed on errors. Only the main effect of Error Type was significant, F1, 30 = 11.54, p = .002, indicating that errors in the incongruent naming condition (.91) were more frequent than errors in neutral naming conditions (.24). To further examine the relations between cognitive performance and hypnotic susceptibility, the Pearson correlation coefficient was calculated between the SHSS-A score and the SIE score. The SHSS-A showed a significant correlation with SIE, (r = .48, p = .004), indicating that hypnotic susceptibility increases along with the decrease of the SIE score. 4. Discussion The present work was designed to test the relationships between hypnotic susceptibility and the attentional system. A high level of hypnotic susceptibility is characterized by high concentration on hypnotic instructions and by a pronounced loss of awareness after the hypnotic induction. According to the neuropsychophysiological model (e.g., Crawford, 1994; Crawford & Gruzelier, 1992; Gruzelier, 1988, 1998), these characteristics, which determine the level of hypnotic susceptibility, depend on the efficiency of the attentional system that is mainly supported by the frontal, and prefrontal areas of the brain. More precisely, this model implies that HH subjects, compared to LH subjects, should have both a higher performance in attentional functioning at the baseline state, that is out of hypnosis, and a decrease of attentional functioning after the induction. While there is converging evidence that supports the stronger decrease of the attentional system activity in HH during hypnosis, more controversial is whether the attentional system is more efficient in HH than in LH subjects at baseline. We reasoned that, among the attentional mechanisms, the ability to focus on instructions while filtering out irrelevant information should be fundamental to comply with hypnotic instructions. The Stroop task (MacLeod, 1991; Stroop, 1935) allows us to test the ability to select the relevant information and to suppress the irrelevant information. If hypnotic susceptibility depends on this attentional ability, then HH subjects should perform better than LH subjects on the Stroop task, that is, they should have a smaller SIE. Previous studies did not show this predicted difference on Stroop performance as a function of different levels of hypnotic susceptibility. In the present study, we used a version of the Stroop task that was designed to measure the SIE by considering accuracy performance. 302 S. Rubichi et al. / Consciousness and Cognition 14 (2005) 296–303 As predicted, the SIE of HH subjects was smaller than that of LH subjects. The SIE of the MH subjects was equivalent to that obtained by HH subjects. Moreover, the SIE and the SHSS-A score correlated negatively, thus corroborating the idea that the selection/suppression attentional ability, and hypnotic susceptibility are closely interrelated. Thus, also the prediction of the neuropsychophysiological model stating that baseline attentional functioning is more efficient in HH than in LH subjects gathered empirical support. The fact that MH subjectsÕ performance is equivalent to that of HH subjects seems to indicate that there are likely other factors underlying high levels of hypnotizability. Future researches should be directed at investigating the nature of such factors. A final observation regards the tasks that are often used for comparing the performance of subjects with different degrees of hypnotic susceptibility. Given that cognitive tasks designed to detect differences between normal subjects and neuropsychological patients are often used, we believe that the reported lack of differences in attentional performance between HH and LH subjects may depend, at least in part, on the power of the adopted tools. In the study by Aikins and Ray (2001), for example, neuropsychological tests designed to assess frontal lobe dysfunctions were used. The results led the authors to conclude that they found ‘‘scant evidence for frontal differences between high and low hypnotizable across eight indices of functions posited to operate frontally’’ (p. 104). However, with these tasks, often non-brain damaged subjects reach the optimum performance, or nearly so. In other words, it is likely that no differences are found between HH and LH subjects because ceiling effects are reached. In the present work, also, participants made very few errors, and the ANOVA on error did not yield any significant effect. As Aikins and Ray (2001) suggested, the frontal lobe, and the attentional system may have a more subtle role in explaining hypnotic susceptibility and, more in general, hypnotic phenomena. Thus, the broad conclusion suggested here is to examine the relationships between the attentional system and hypnotic susceptibility by means of more specific hypotheses and fine-grained tools. Acknowledgments The study was supported by grants from MIUR. We thank two anonymous reviewers for their very helpful comments on an earlier draft of this work. References Aikins, D., & Ray, W. J. (2001). Frontal lobe contributions to hypnotic susceptibility: A neuropsychological screening of executive functioning. International Journal of Clinical and Experimental Hypnosis, 49, 320–339. Blum, G. S., & Graef, J. R. (1971). The detection over time of subjects simulating hypnosis. International Journal of Clinical and Experimental Hypnosis, 19, 211–224. Cheesman, J., & Merikle, P. M. (1986). Distinguishing conscious from unconscious perceptual processes. Canadian Journal of Psychology, 40, 343–367. Crawford, H. J. (1994). Brain dynamics and hypnosis: Attentional and disattentional processes. International Journal of Clinical and Experimental Hypnosis, 52, 204–232. Crawford, H. J., & Allen, S. N. (1983). Enhanced visual memory during hypnosis as mediated by hypnotic responsiveness and cognitive strategies. Journal of Experimental Psychology: General, 112, 662–685. S. Rubichi et al. / Consciousness and Cognition 14 (2005) 296–303 303 Crawford, H. J., & Gruzelier, J. H. (1992). A midstream view of the neuropsychophysiology of hypnosis: Recent research and future directions. In E. Fromm & M. R. Nash (Eds.), Contemporary hypnosis research (pp. 227–266). London: Guilford. Dixon, M., Brunet, A., & Laurence, J. R. (1990). Hypnotizability and automaticity: Toward a parallel distributed processing model of hypnotic responding. Journal of Abnormal Psychology, 99, 336–343. Dixon, M., & Laurence, J. R. (1992). Hypnotic susceptibility and verbal automaticity: Automatic and strategic processing differences in the Stroop color-naming task. Journal of Abnormal Psychology, 101, 344–347. Gruzelier, J. H. (1988). The neuropsychology of hypnosis. In M. Heap (Ed.), Hypnosis: current clinical, experimental and forensic practices (pp. 68–76). London: Croom Helm. Gruzelier, J. H. (1998). A working model of the neurophysiology of hypnosis: A review of evidence. Contemporary Hypnosis, 15, 3–21. Gruzelier, J. H., & Warren, K. (1993). Neuropsychological evidence of reductions of left frontal tests with hypnosis. Psychological Medicine, 23, 93–101. Heaton, R. K. (1981). Wisconsin card sorting test. Odessa, FL: Psychological Assessment Resources. Jasiukaitis, P., Nouriani, B., Hugdahl, K., & Spiegel, D. (1997). Relateralizing hypnosis: Or, have we been barking up the wrong hemisphere?. International Journal of Clinical and Experimental Hypnosis, 2, 158–177. Kaiser, J., Barker, R., Haenschel, C., Baldeweg, T., & Gruzelier, J. H. (1997). Hypnosis and event-related potential correlates of error processing in a Stroop-type paradigm: A test of the frontal hypothesis. International Journal of Psychophysiology, 27, 215–222. Kallio, S., Revonsuo, A., Hamalainen, H., Markela, J., & Gruzelier, J. H. (2001). Anterior brain functions and hypnosis: A test of the frontal hypothesis. International Journal of Clinical and Experimental Hypnosis, 49, 95–108. MacLeod, C. M. (1991). Half a century of research on the Stroop effect: An integrative review. Psychological Bulletin, 109, 163–203. Mészáros, I., Crawford, H. J., Szabó, C., Nagy-Kovács, A., & Résvész, M. A. (1989). Hypnotic susceptibility and cerebral hemisphere preponderance: Verbal-imaginal discrimination task. In V. Ghorghui, R. Netter, H. Eysenck, & R. Rosenthal (Eds.), Suggestion and suggestibility: Theory and research (pp. 191–204.). New York: Springer-Verlag. Nordby, H., Hugdahl, K., Jasiukaitis, P., & Spiegel, D. (1999). Effects of hypnotizability on performance of a Stroop task and event-related potentials. Perception and Motor Skills, 88, 819–830. Posner, M. I., & Raichle, M. E. (1997). Images of the mind. New York: Scientific Americam Library. Raz, A., & Shapiro, T. (2002). Hypnosis and neuroscience: A crosstalk between clinical and cognitive research. Archives of General Psychiatry, 59, 85–90. Raz, A., Landzberg, K. S., Schweizer, H. R., Zephrani, Z. R., Shapiro, T., Fan, J., & Posner, M. I. (2003). Posthypnotic suggestion and the modulation of Stroop interference under cyclopegia. Consciousness and Cognition, 12, 332–346. Raz, A., Shapiro, T., Fan, J., & Posner, M. I. (2002a). Hypnotic modulation of Stroop interference: Behavioral and neuroimaging accounts. Journal of Cognitive Neuroscience, A34 (Suppl). Raz, A., Shapiro, T., Fan, J., & Posner, M. I. (2002b). Hypnotic suggestion and the modulation of Stroop interference. Archives of General Psychiatry, 59, 1155–1161. Sheenan, P. W., Donovan, P., & MacLeod, C. M. (1988). Strategy manipulation and the Stroop effect in hypnosis. Journal of Abnormal Psychology, 94, 249–255. Stroop, J. R. (1935). Studies of interference in serial verbal reactions. Journal of Experimental Psychology, 18, 643–661. Swick, D., & Jovanovic, J. (2002). Anterior cingulated cortex and the Stroop task: Neuropsychological evidence for topographic specificity. Neuropsychologia, 40, 1240–1253. Vendrell, P., Junquè, C., Pujol, J., Jurado, M. A., Molet, J., & Grafman, J. (1995). The role of prefrontal regions in the Stroop task. Neuropsychologia, 33, 341–352. Weitzenhoffer, A. M., & Hilgard, E. R. (1959). Stanford hypnotic susceptibility scale. Palo Alto, California: Consulting Psychologists press.

Consciousness and Cognition 22 (2013) 931–943 Contents lists available at SciVerse ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog The role of metacognition in prospective memory: Anticipated task demands influence attention allocation strategies Jan Rummel ⇑, Thorsten Meiser University of Mannheim, School of Social Sciences, Mannheim, Germany a r t i c l e i n f o Article history: Received 22 February 2013 Available online 13 July 2013 Keywords: Prospective memory Metacognition Monitoring processes Attention allocation a b s t r a c t The present study investigates how individuals distribute their attentional resources between a prospective memory task and an ongoing task. Therefore, metacognitive expectations about the attentional demands of the prospective-memory task were manipulated while the factual demands were held constant. In Experiments 1a and 1b, we found attentional costs from a prospective-memory task with low factual demands to be significantly reduced when information about the low to-be-expected demands were provided, while prospective-memory performance remained largely unaffected. In Experiment 2, attentional monitoring in a more demanding prospective-memory task also varied with information about the to-be-expected demands (high vs. low) and again there were no equivalent changes in prospective-memory performance. These findings suggest that attention–allocation strategies of prospective memory rely on metacognitive expectations about prospective-memory task demands. Furthermore, the results suggest that attentional monitoring is only functional for prospective memory to the extent to which anticipated task demands reflect objective task demands. Ó 2013 Elsevier Inc. All rights reserved. 1. Introduction Event-based prospective memory (PM) denotes the ability to remember performing an intended action at the occurrence of a target cue (Kvavilashvili & Ellis, 1996). Typically, the PM cue occurs while one is engaged in other ongoing activities (Ellis & Kvavilashvili, 2000). For example, when remembering to give a colleague a message the next time one sees her, encountering the colleague can cue the intention. When the colleague passes one’s office, however, one is usually engaged in answering phone calls, writing emails, or talking to other colleagues and thus one is likely to miss the opportunity to pass the message. Therefore, it might be reasonable to devote some attention to monitor for the occurrence of the colleague to make sure that one will not miss the appropriate moment of intention fulfillment. How individuals distribute their attentional resources between a PM intention and other ongoing tasks, however, is not yet completely understood. Therefore, in the present research, we investigated how metacognitive expectations about the demands of a given PM task influence attention–allocation strategies of PM. Expectations about cognitive demands have been demonstrated to influence strategies in various memory tasks. For example, such metacognitive expectations have been shown to affect study-time allocation in recognition memory (Dunlosky & Ariel, 2011; Hines, Touron, & Hertzog, 2009; Mazzoni, Cornoldi, & Marchitelli, 1990; but see also Son & Metcalfe, 2000), the illusion of competence in cued-recall tasks (Castel, McCabe, & Roediger, 2007; Koriat & Bjork, 2005), or systematic guessing behavior in recognition and source memory (Förster & Strack, 1998; Meiser, Sattler, & von Hecker, 2007; Strack, Förster, & ⇑ Corresponding author. Address: Department of Psychology, University of Mannheim, D-68131 Mannheim, Germany. E-mail address: rummel@uni-mannheim.de (J. Rummel). 1053-8100/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.concog.2013.06.006 932 J. Rummel, T. Meiser / Consciousness and Cognition 22 (2013) 931–943 Werth, 2005). Two different approaches have been suggested to investigate metacognitive expectations about task-demands. The most prominent one is to ask participants to rate their own performance in a cognitive task either prior to or after performing the task. Such metacognitive judgments have been successfully used to assess individual expectancies about the demands of and the performances in different memory tasks (e.g., Koriat & Bjork, 2005; Nelson & Dunlosky, 1991; Touron, Oransky, Meier, & Hines, 2010). Another less frequently used approach is to manipulate metacognitive expectations experimentally via (bogus) information provided with memory-task instructions. Förster and Strack (1998), for example, investigated the role of task-demands expectations in a recognition task by inducing the beliefs that the music played at encoding would foster (hinder) learning of the words and found that the induced expectancies about the music’s impact affected the hit rate and false alarm rate in a subsequent recognition test. Building on this second approach in the present research, we experimentally manipulated the expectations about the cognitive demands of a PM task to investigate whether, among other factors, metacognitive expectations would influence the strategies of how attentional resources are distributed between a PM intention and an ongoing task. Current theories of PM assume that holding a PM intention as well as performing an ongoing task usually draw on limited attentional resources (McDaniel & Einstein, 2007; Smith, 2003). In line with this idea, adding a PM intention to an ongoing task has been found to slow down performance in the latter task (Marsh, Cook, & Hicks, 2006; Marsh, Hicks, & Cook, 2005, 2006; Marsh, Hicks, Cook, Hansen, & Pallos, 2003; Smith, 2003). The level of these PM-induced costs to the ongoing task, and thus supposedly the level of attentional cue-monitoring, seems to depend on the cognitive demands of the PM intention and the ongoing task. When there are fewer different PM cues (Cohen, Jaudas, & Gollwitzer, 2008), when the PM cues become more salient (Einstein, McDaniel, Manzi, Cochran, & Baker, 2000; Graf, Uttl, & Dixon, 2002), or when there is a strong overlap between the processes engaged to perform the ongoing task and to identify the PM cue (Einstein et al., 2005; McBride & Abney, 2012; Meier & Graf, 2000; Rummel, Boywitt, & Meiser, 2011; see also Einstein & McDaniel, 2005, for a review) PM-induced costs are significantly reduced, sometimes even to non-significant levels. These findings suggest that the level of cue-monitoring is adapted to the reduced PM-task demands. On the other hand, PM performance also varies with the demands of the ongoing task (Marsh, Hancock, & Hicks, 2002). Therefore, Hicks and colleagues (Hicks, Marsh, & Cook, 2005) argue that during intention formation individuals adopt a resource-allocation strategy of how they distribute their attention between the ongoing task and the PM task in accordance with the cognitive demands of both tasks. The engagement in attentional cue-monitoring, however, is not completely determined by objective task demands. For example, Einstein et al. (2005) showed that additionally stressing the importance of the PM task relative to the ongoing task results in higher PM-induced costs and improved PM performance (see also Marsh et al., 2005). Similarly, individuals become more cautious in their ongoing task responding when they think it is very likely that the PM cue will occur (Boywitt & Rummel, 2012). In line with findings that cue-monitoring can be willingly controlled, Einstein and McDaniel (2008) recently theorized that individuals are aware of the PM-task demands and calibrate their cue-monitoring accordingly. With non-habitual PM tasks, however, individuals supposedly have little prior knowledge about the factual PM-task demands while forming an intention, because they have not yet experienced the PM task (and their own performance in the task) at this point. Therefore, the question arises what a potential awareness of PM-task demands is based on. We argue that resource-allocation strategies of PM are not only calibrated to objective PM-task demands but also in accordance with idiosyncratic beliefs about these demands, that is, expectations about the cognitive effort necessary for fulfilling the intention. Importantly, such metacognitive expectations about task demands can differ from factual task demands. Research asking participants to predict their PM performance has found, for instance, that participants tend to underestimate their PM performance (Meeks, Hicks, & Marsh, 2007; Schnitzspahn, Zeintl, Jäger, & Kliegel, 2011), especially when objective PM-task demands are low (Rummel, Kuhlmann, & Touron, 2013). Although PM-prediction accuracy has to be interpreted with caution because predictions can reactively change PM performance (Meier, von Wartburg, Matter, Rothen, & Reber, 2011) as well as the engagement in attentional cue-monitoring (Rummel et al., 2013), these results suggest that knowledge about PM-task demands might be rather inaccurate, especially before task experience. If attention–allocation strategies build on biased expectations about PM-task demands, however, attention allocation will be biased as well. In particular, if the actual PM-performance is underestimated, individuals might allocate additional attention to the PM task exceeding the objective PM-task demands. This would result in additional PM-induced costs, which are not functional for PM. Importantly, the observed PM-induced costs, in terms of slowed ongoing-task performance, would contain both functional and non-functional cost-components. Evidence that PM-induced costs are not always functionally related with PM-performance comes from Meiser and Schult (2008), who found significant correlations between PM-performance and PM-induced costs only when objective PM-task demands were high and ongoing-task accuracy was additionally stressed. Furthermore, Scullin, McDaniel, and Einstein (2010) found increased PM-induced costs with a reminder of the PM task several trials prior to a PM cue, but this increase in costs was not accompanied by an increase in PM performance when factual PM-task demands were low. Thus, in these studies, the PM-induced costs were not always entirely functional for PM. 2. The current study The aims of the present study were twofold. The major goal was to show that attention–allocation strategies of PM are calibrated to metacognitive expectations about PM-task demands. Others have reasoned that metacognitive expectations about one’s own PM performance should play a critical role for the engagement in attentional cue-monitoring (Einstein & J. Rummel, T. Meiser / Consciousness and Cognition 22 (2013) 931–943 933 McDaniel, 2008), but to our best knowledge, this notion has not yet been tested empirically. To fill this gap we manipulated expectations about the cognitive demands imposed by a PM task during intention formation and independent of factual PMtask demands during intention realization. In doing so, we further aimed at testing, whether ill-calibrated expectations about PM-task demands can result in additional, non-functional costs to an ongoing task. The manipulation of anticipated PM-task demands in the present experiments was derived from the metamemory literature suggesting that knowledge of one’s own memory functioning can be based on two sources of information, namely on explicit declarative information about memory demands and, if available, on prior experience (cf. Koriat, 1997; Koriat, Bjork, Sheffer, & Bar, 2004; Strack et al., 2005). We expected individuals to utilize these two sources of information to develop expectancies about PM-task demands. In a first step, a pilot study was conducted to corroborate the assumption that providing declarative information about the to-be-expected PM-task demands with PM instructions would change expectancies regarding the PM-task’s difficulty and PM performance. To support the manipulation of anticipated PM-task demands, participants not only received metacognitive information but had also the opportunity to experience the specifics of the future PM cue before receiving PM instructions. In Experiments 1a and 1b, we then manipulated expectations about the demands of a PM task with salient PM cues using the manipulation that was pretested in the pilot study. Factual task demands, however, were held constant to investigate whether attention–allocation strategies vary with anticipated PM-task demands. As identifying salient PM cues should not require much attention (Graf et al., 2002), we expected a high PM-performance even with an anticipation-induced de-allocation of attentional resources from the PM task. To investigate the role of anticipated PM-task demands under conditions where PM-performance is expected to rely heavily on attentional cue-monitoring (McDaniel & Einstein, 2007), participants in Experiment 2 were presented with a common PM task with non-salient cues and anticipated PM-task demands were manipulated again via declarative information. 3. Pilot study In order to test whether expectations about PM-task demands and one’s own PM performance are sensitive to metacognitive information provided with PM instructions, we compared PM-predictions of one group of participants who received standard PM instructions with those of another group who received additional information about the to-be-expected PM demands. Therefore, all participants first received PM instructions, asking them to respond to red-letter words by pressing a specific key during a lexical-decision task consisting of black-letter stimuli, which they would perform later in the experiment. Participants in the metacognitive-information condition had been pre-exposed with an exemplar of the red PM-target word before receiving PM instructions. Thus, they had experienced the saliency of a red-colored word before they knew that this word would require a specific response later in the experiment. As part of the PM instructions, participants in this condition were further informed that the red words would leap to the eye so that correct PM responding would not require attention. After a short delay, participants in both conditions were then asked to rate the likelihood that they will detect the PM-target word and respond appropriately and to rate the difficulty of doing so. Higher PM-performance and lower PM-difficulty ratings in the metacognitive-information condition would corroborate the notion that expectancies about PM-task demands can be manipulated via metacognitive information. Furthermore, participants were asked to rate the perceived importance of the PM task to demonstrate that the suggested manipulation can be distinguished from importance manipulations (Kliegel, Martin, McDaniel, & Einstein, 2001). 3.1. Method 3.1.1. Participants Sixty-seven students participated for monetary compensation or course credit. Thirty-four of them were randomly assigned to the standard-PM condition and 33 were assigned to the metacognitive-information condition. 3.1.2. Materials and procedure Participants first received instructions for a lexical-decision task asking them to press the A key for words and the L key for nonwords, followed by 21 practice trials. Twenty-one words of medium frequency and length were chosen from a German word-norm database and nonwords were created by swopping two or three letters in ten of these words. Lexical-decision trials were separated by a blank screen for 1750 ms and words and nonwords were presented until participants entered a response. For the metacognitive-information condition, the word at the 16th practice trial was displayed in red letters. After the practice trials, participants in both conditions received the PM instruction to respond to red words by pressing the 7 key when performing the lexical-decision task for the next time. Participants in the metacognitive-information condition were further told that previous research had shown that correct responding to the red words does not require attention and that it was not necessary to think of the additional task continuously, because, as they may have realized by previously encountering the red word, the color would leap to the eye immediately. Participants in both conditions then solved figural reasoning problems for five minutes before they were asked to rate their own PM performance and their ongoing-task performance (in percent). Additionally, they were asked to rate the PM-task’s difficulty (0 – not at all difficult; 100 – extremely difficult) and the perceived importance of the PM task compared 934 J. Rummel, T. Meiser / Consciousness and Cognition 22 (2013) 931–943 to the ongoing task (0 – the ongoing task is much more important; 100 – the PM task is much more important). After answering these questions, participants were informed that they would not have to perform the lexical-decision task and were debriefed and dismissed. 3.2. Results and discussion Independent sample t-tests were conducted comparing performance, difficulty, and importance ratings between the standard-PM and the metacognitive-information condition. Expectancies concerning ongoing-task performances did not differ between the two conditions, t < 1. Participants in the metacognitive-information condition, however, expected the PM task to be significantly less difficult (M = 47.15; SE = 4.39) than participants in the standard-PM condition (M = 58.18; SE = 2.67), t(65) = 2.20, p = .031, d = .55. Furthermore, participants in the metacognitive-information condition expected their PM performance to be significantly higher (M = 75.15; SE = 4.10) than participants in the standard-PM condition (M = 61.18; SE = 3.00), t(65) = 2.77, p = .007, d = .69. The perceived importance of the PM task, however, was quite similar in the metacognitive-information (M = 46.67; SE = 3.35) and the standard-PM condition (M = 50.79; SE = 3.37), t < 1. This pattern of results demonstrates that the suggested manipulation of anticipated PM-task demands selectively affected participants’ expectations about their own PM performance and the perceived difficulty of the PM task while leaving expectations about ongoing task performance and the perceived importance of the PM task unaffected. Notably, these results not only provide evidence that the manipulation of anticipated task demands was successful but also that this manipulation can be distinguished from manipulations merely attenuating the perceived importance of the PM task (e.g., Harrison & Einstein, 2010). After corroborating the validity of the manipulation of anticipated PM-task demands, we aimed at testing whether variations in anticipated PM-task demands will translate into differences in attention–allocation strategies of PM in the Experiments 1a and 1b. To avoid a confound of the present manipulation with reactive effects of PM-performance predictions (Meier et al., 2011; Rummel et al., 2013), we did not assess PM-performance predictions in the following experiments. 4. Experiment 1a While there is converging evidence that the addition of a PM task causes performance costs in an ongoing task (e.g., Marsh et al., 2003), results are mixed when the PM cues are highly salient. Some authors did not find significant PM-induced costs with salient PM cues, like participants’ own names for instance (Harrison & Einstein, 2010). Other researchers, however, did find significant PM-induced costs under comparable conditions (Smith, Hunt, McVay, & McConnell, 2007). Based on these inconsistencies, Einstein and McDaniel (2010) argued that costs may occur that are not necessary to realize an intention and not functionally related to PM performance (see also Scullin et al., 2010). Similarly, we suggest that the inconsistent results can be explained by taking metacognitive expectations about PM-task demands into account. That is, when receiving PM instructions, participants may sometimes be unsure about the factual attentional demands of the task, and they may allocate more attention to the PM task than objectively necessary, hence, engaging cue-monitoring at the cost of the ongoing task that is partially not functional for PM. To test these assumptions, we realized three conditions in Experiment 1a in which participants had to perform an ongoing lexical-decision task. In two conditions, participants received the additional PM task to press a special key in response to a salient cue, that is, a red letter word in the context of black words and nonwords (cf. Smith et al., 2007, Experiment 1). In one of these conditions (i.e., the metacognitive-information condition), participants’ expectations about the attentional demands of the PM task were manipulated using the manipulation that was pretested in the pilot study while the objective PM-task demands during intention realization were held constant between the two conditions. In a third condition (i.e., the control condition), participants’ had to perform the ongoing task without PM demands to assess whether the addition of a PM intention resulted in attentional costs to the ongoing task. 4.1. Method 4.1.1. Participants Ninety-six students participated for monetary compensation and were randomly assigned to one of the three conditions. Data of one participant were excluded because their mean response time in the baseline block of the ongoing task (see below) was more than 3.5 standard deviations above the sample mean. The final sample consisted of 95 participants with 32 participants in the standard-PM condition, 30 in the metacognitive-information condition, and 33 in the control condition. 4.1.2. Materials and procedure In all conditions, the ongoing lexical-decision task consisted of 206 trials including 103 German words and 103 nonwords, which were created as in the pilot study. The first six trials (half words, half nonwords) were practice trials. The remaining 200 trials formed two blocks of 100 trials each, with the first block serving as a baseline block. All experimental manipulations (i.e., imposing PM demands and/or providing metacognitive information) were realized after the first block and thus the second block served as the experimental block. Both the baseline block and the experimental block were further divided into two sub-blocks and each sub-block consisted of 50 trials with equal numbers of word and nonword probes. Within each 935 J. Rummel, T. Meiser / Consciousness and Cognition 22 (2013) 931–943 Table 1 Ongoing task performance as a function of condition (Experiment 1a). Condition Accuracy Response times Standard PM Metacognitive information Control .96 (.006) 956 (24) .96 (.006) 892 (25) .96 (.006) 916 (23) Note. Accuracy (i.e., adjusted mean proportion of accurate responses in the experimental block of the ongoing task) and response times (i.e., adjusted mean response times on words in milliseconds in the experimental block of the ongoing task) as a function of condition in Experiment 1a. Standard errors of the means are displayed in parentheses. sub-block, the order of probes was randomized. In all conditions, a red-letter word was presented at the 51st trial of the experimental block while all other probes were black. In the condition with metacognitive-information condition, the 51st trial in the baseline block also contained a red-letter word. Experimental sessions started with instructions for the lexical decision task (see pilot study). After the six practice trials and the baseline block, instructions for the lexical decision task were repeated, and in the two PM conditions, they were expanded by the PM-task instructions. Then, participants in the metacognitive-information condition were informed that PMtask performance would require little attention. This information was identical to the one used in the pilot study. Participants in the control condition did not receive a PM task. After solving figural reasoning problems for five-minutes, participants performed the experimental block of the lexical decision task. Finally, participants were asked to recall when they had to press another key than usual and which key that was to ensure that they retrospectively remembered the PM task. 4.2. Results and discussion The level of statistical significance was set to .05 for all analyses. 4.2.1. Ongoing-task performance Mean response times (RTs) on words1 and the proportion of accurate responses in the experimental block were analyzed to assess PM-induced costs to the ongoing task. RT analyses were confined to accurate responses and responses slower than 300 ms but faster than two standard deviations above the individual mean to control for aberrant RTs (e.g., Einstein et al., 2005). Deviations from a normal distribution were tested using the Shapiro–Wilk test. The red-word trials as well as the trials directly succeeding a red word were excluded from the analyses, to avoid artifactual costs associated with these trials. For all analyses the performance in the baseline block was included as a covariate to adjust for inter-individual differences in response speed and to increase statistical power.2 Adjusted mean accuracy-rates and adjusted mean RTs for each condition are presented in Table 1. After excluding outliers, RT-distributions still deviated from a normal distribution in the baseline block [standard PM: p = .033; metacognitive information: p = .002; control: p = .003] as well as in the experimental block [standard PM: p = .006; metacognitive information: p = .133; control: p = .033]. Therefore, mean RTs were log-transformed for each participant and block (Ratcliff, 1993).3 After this transformation, the RT distributions did not significantly differ from a normal distribution, all ps > .12. We specified two orthogonal planned contrasts (Rosenthal & Rosnow, 1985) in an analysis of covariance of RTs with condition (standard PM, metacognitive information, control) as between-subjects factor and RTs in the baseline block as covariates.4 The contrasts directly tested the hypothesis that costs from a PM-task with salient cues can be avoided when metacognitive information about the low PM-task demands is provided and thus provide the most powerful test of the present hypothesis (cf. Rosenthal & Rosnow, 1985). The first contrast compared the standard-PM condition to both the metacognitiveinformation condition and the control condition to test the prediction that higher PM-induced costs (in terms of slower baseline-adjusted RTs) are observed in the standard-PM condition than in the metacognitive-information and the control condition. The second contrast compared the metacognitive-information condition with the control condition to test for residual costs in the metacognitive-information condition. The results revealed the predicted pattern. RT-costs in the standard-PM condition were significantly higher than in the other two conditions, t(91) = 2.08, p = .041, d = .44, suggesting that participants engaged in resource-demanding monitoring for the salient cues when no additional information about the low PM-task demands was provided (Smith et al., 2007). No RT-cost differences were found, however, between the metacognitive-information condition and the control condition, t(91) < 1, d = .18. The finding that PM-induced costs were significantly reduced when information 1 As the PM cue was pre-defined to be a word, RTs on words should better reflect processes engaged to identify the cue (cf. Marsh et al., 2003). Only the first 50 trials of the baseline block were analyzed because in the metacognitive-information condition the 51st trial in this block was a red word and this could have influence processing of the subsequent stimuli. Including all baseline-block trials would not have changed the pattern of results. 3 Using non-transformed RTs as dependent variables and covariates in the present experiments would have led to similar conclusions. 4 The ANCOVA approach allowed us to compare RTs in the experimental block between experimental conditions with baseline RT variations being controlled for. Therefore, the adjusted RTs reflect PM-induced costs to the ongoing task. While simple one-way ANOVAs did not show differences in RTs between experimental conditions for the baseline block (Mstandard = 816; SE = 40; Mmetacognitive information = 917; SE = 41; Mcontrol = 867; SE = 39), F(2, 92) = 1.94, p = .150, g2p ¼ :04, and for the second block (Mstandard = 909; SE = 44; Mmetacognitive information = 940; SE = 45; Mcontrol = 917; SE = 43), F < 1, the ANCOVA with planned contrasts enabled us to test the more specific predictions concerning PM-induced costs, that is, RT differences in the second block that cannot be accounted for by natural variation as assessed in the baseline block. 2 936 J. Rummel, T. Meiser / Consciousness and Cognition 22 (2013) 931–943 about the low task demands was provided with PM instructions suggests that attention–allocation strategies of PM are calibrated in accordance with anticipated PM-task demands. The same contrasts did not reveal a significant difference with ongoing task accuracy as the dependent variable, ts < 1, indicating that potential PM-induced costs were not reflected by ongoing task accuracy (cf. Marsh et al., 2003; Smith et al., 2007). 4.2.2. Memory for PM instructions and PM performance All participants recalled the PM task in the final questionnaire. The proportion of accurate PM responses was analyzed to compare PM performance in the two PM conditions. Pressing the 7 key as the first response to the cue as well as pressing the 7 key within the inter-stimulus interval after having pressed another key (i.e., either the A or L key for the lexical decision task) were counted as PM responses. PM performance did not differ between the standard-PM condition (M = .84, SE = .07) and the metacognitive-information condition (M = .80, SE = .07), t < 1, d = .10.5 In both PM conditions, PM performance was relatively high, which is in line with previous research using salient PM cues (Harrison & Einstein, 2010; Smith et al., 2007). However, the 95% confidence intervals (CIs) did not overlap with one, neither in the standard-PM condition, CI = [0.70; 0.98], nor in the metacognitive-information condition, CI = [0.66; 0.94], indicating that the results are unlikely to be due to ceiling performance.6 As we argued that PM-induced costs can be reduced by providing task-demands information without impeding PM performance, the statistical power to detect PM-performance differences between the two PM conditions is critical in the present analysis. Thus power analyses were conducted with the software G*Power (Faul, Erdfelder, Lang, & Buchner, 2007) revealing that the power to detect superior PM performance in the standard-PM condition relative to the metacognitive-information condition was moderate, 1  b = .62, to reveal at least a medium-sized effect (Cohen’s d P .50). At the same time, the empirical effect-size was negligibly small (d = .10), questioning the practical relevance of a potential PM-performance difference of this size. The issue of statistical power will be addressed in a joint analysis of Experiments 1a and 1b to rule out that a mere lack of statistical power can account for the present findings. 5. Experiment 1b Results of Experiment 1a provide evidence that participants allocate their attentional resources in line with metacognitive information about the to-be-expected PM-task demands. Attention–allocation strategies of PM, however, are likely to vary with task experience. For instance, Loft, Kearney, and Remington (2008) showed that attentional monitoring is reinitialized after the presentation of a PM cue. Because participants in Experiment 1a encountered the PM cue only once, the question arises whether individuals would re-calibrate their attention allocation with increasing PM-task experience. The metacognition literature suggest that metacognitive judgments are partly based on experiences made while performing a cognitive task (Koriat, 1997). In Experiment 1a, participants in the metacognitive-information condition had experienced the saliency of a red-colored word before they received PM instructions but they had little online experience with the PM task, that is, they experienced only once how easy/difficult it would be to respond to the salient cue with the intended action. To further investigate the role of task experience for metacognitive expectations about PM-task demands, the same setting as in Experiment 1a was realized in Experiment 1b, but participants were presented with three PM cues in the course of the experiment. This allowed for analyzing potential effects of online experience with processing salient PM cues. Additionally, realizing the same experimental conditions as in Experiment 1a allowed for a more powerful test of our hypothesis that some cost-components are non-functional for PM performance in a joint analysis of both experiments. 5.1. Method 5.1.1. Participants One hundred and nine students participated for monetary compensation and were randomly assigned to one of the three conditions. Data of one participant whose mean response time in the lexical decision task was more than 3.5 standard deviations above the sample mean and of two participants who did not follow the ongoing task instructions were excluded. The final sample consisted of 106 participants in total with 36 participants in the standard-PM condition, 35 in the metacognitive-information condition, and 35 in the control condition. 5.1.2. Materials and procedure The materials and procedure were identical to Experiment 1a with the following exceptions. The experimental block of the lexical decision task was extended to 163 trials to include two more PM trials. To avoid excessive numbers of trials, we did not include a baseline block but we extended the number of practice trials to 21. Imperceptible to the participants, the 5 A Chi-Square test also yielded no difference in PM performance between the two PM conditions, v2(1) = 0.15, p = .70, w = .10. Excluding all participants who failed to respond to the PM cue correctly did not change the RT results. Because PM performances in both PM conditions are perfect after excluding participants who missed the PM cue, this result suggests that the reduction of costs in the metacognitive-information condition remained stable after controlling for PM-performance differences between the two PM conditions. Therefore, it is unlikely that the present results were due to differential trade-offs between the PM task and the ongoing task. 6 937 J. Rummel, T. Meiser / Consciousness and Cognition 22 (2013) 931–943 Table 2 Ongoing task performance as a function of condition and segments (Experiment 1b). Condition Standard PM Metacognitive information Control Accuracy Overall Segment1 Segment2 Segment3 Segment4 .96 (.006) .96 (.006) .96 (.009) .96 (.008) .95 (.009) .96 (.006) .96 (.006) .97 (.006) .96 (.006) .96 (.007) .96 (.006) .97 (.006) .96 (.009) .96 (.008) .96 (.009) Response times Overall Segment1 Segment2 Segment3 Segment4 1015 (39) 1059 (46) 1017 (41) 1021 (41) 966 (36) 873 (39) 937 (47) 854 (41) 870 (42) 843 (37) 931 (39) 964 (47) 923 (42) 948 (42) 903 (37) Note. Accuracy (i.e., adjusted mean proportion of accurate responses in the experimental block of the ongoing task) and response times (i.e., adjusted mean response times on words in milliseconds in the experimental block of the ongoing task) as a function of condition in Experiment 1b. The segments 1–4 indicate subsequent blocks of 40 trials of the lexical decision task separated by PM trials. Bold values indicate overall performance (i.e., performance aggregated across segments 1-4). Standard errors of the means are displayed in parentheses. experimental block was further divided into four segments of 40 trials (half word, half nonword trials randomly intermixed) each, with the PM trials located between adjacent segments. Accordingly, the red-letter words, which served as PM cues in the PM conditions, appeared at Trials 41, 82, and 123. For the metacognitive-information condition, an additional red-letter word appeared at the 16th trial of the practice block, serving as an exemplar of the salient word. Metacognitive instructions in the metacognitive-information condition were equivalent to Experiment 1a. 5.2. Results and discussion 5.2.1. Ongoing-task performance RTs were trimmed using the same criteria as in Experiment 1a and RTs in the practice block were used as a covariate to adjust for inter-individual differences.7 Adjusted mean RTs and mean accuracy-rates for each condition are reported in Table 2. Again, the trimmed RTs were significantly non-normal, all ps 6 .003, and thus they were log-transformed. After transformation, RTs did not differ significantly from a normal distribution, all ps P .250. The same planned contrasts as in Experiment 1a were specified. The first contrast compared the standard-PM condition with the metacognitive-information condition and the control condition; the second contrast compared the metacognitive-information condition with the control condition. Replicating the results of Experiment 1a, the planned contrasts showed that participants in the standard-PM condition showed higher levels of RT-costs than participants in the other two conditions, t(102) = 2.89, p = .005, d = .57. Participants in the metacognitive-information condition were again numerically faster than participants in the control condition were, but this comparison was not significant, t(102) = 1.11, p = .267, d = .22. These findings further support the interpretation that PM-induced costs could be avoided when information about task demands was provided. These results are additional evidence that metacognitive information about task demands is used to calibrate attention-allocation strategies of PM. The same contrasts using ongoing task accuracy as the dependent variable and covariate did not yield significant effects, ts < 1. To explore the effects of PM-task experience, we tested whether there was a disproportionate RT decrease in the standard-PM condition during the course of the experiment as compared to the other two conditions. Such a disproportionate reduction of PM-induced costs in the standard-PM condition would suggest that attention–allocation strategies are re-calibrated based on increased task experience. RTs were analyzed as a function of condition (standard-PM, metacognitive-information, control) and segment (segments 1–4). As segments were separated by the PM trials, PM-task experience should increase over segments. Mean RTs as a function of segment and condition are reported in Table 2. The 3 (condition)  4 (segment) ANCOVA yielded a significant main effect of condition, F(2, 102) = 4.79, p = .010, g2p ¼ :09, mirroring the results from the contrast analysis reported above. Furthermore, there was a marginally significant main effect of segment, F(3, 306) = 2.52, p = .058, g2p ¼ :02, suggesting that there was an overall practice effect in the lexical-decision task. Although the statistical power to reveal a medium-sized interaction effect (g2p P :06) exceeded .99, the interaction was not significant, F < 1, g2p ¼ :01. Thus, in the present experiment, there was no evidence that the level of costs in the standard-PM condition converged to the level of PM-induced costs in the metacognitive-information condition over the course of the lexical-decision task with multiple PM cues. 7 Practice-block RTs (Mstandard = 935; SE = 56; Mmetacognitive information = 924; SE = 57; Mcontrol = 1040; SE = 57) did not differ significantly between conditions, F(2, 105) = 1.98, p = .143, g2p ¼ :04. Experimental-block RTs (Mstandard = 1001; SE = 46; Mmetacognitive information = 853; SE = 47; Mcontrol = 965; SE = 47) varied significantly with conditions, F(2, 105) = 3.26, p = .042, g2p ¼ :06. 938 J. Rummel, T. Meiser / Consciousness and Cognition 22 (2013) 931–943 5.2.2. Memory for PM instructions and PM performance All participants recalled the PM task in the final questionnaire. Again, the proportion of accurate responses in the PM task did not differ between the standard-PM condition (M = .89, SE = .04) and the metacognitive-information condition (M = .92, SE = .03), t < 1, d = .17. Descriptively, PM performance was even higher in the metacognitive-information than in the standard-PM condition, while PM performance was off ceiling in both the standard-PM, 95%-CI = [0.83; 0.95], and the metacognitive-information condition, 95%-CI = [0.86; 0.98].8 5.2.3. Joint analysis of Experiments 1a and 1b Realizing the same experimental conditions in Experiments 1a and 1b, allows us to combine the data from the two experiments in a joint ANCOVA with experiment (1a vs. 1b) as an additional between-subjects factor, which increased the statistical power to detect PM-performance differences and/or residual costs in the PM conditions with metacognitive information. While we demonstrated that anticipated task demands affect the level of PM-induced costs, a second goal of the present research was to demonstrate that some cost components are not functional for PM. To assure that costs were non-functional it is critical to demonstrate that PM performance was not impeded when costs were reduced and thus it is important to rule out that potential PM-performance differences remained undetected due to a lack of statistical power. The pattern of results in the planned-comparison analysis of ongoing-task RTs remained the same in the joint analysis as in the separate analyses of Experiments 1a and 1b. A significant PM-induced slowing was found in the standard-PM condition (Madjusted = 984; SE = 24) when compared with the other two conditions, t(195) = 3.14, p = .002, d = .45. At the same time, ongoing-task RTs in the metacognitive-information condition (Madjusted = 900; SE = 25) did not statistically differ and were even numerically faster than those in the control condition (Madjusted = 923; SE = 24), t < 1. Thus, there was no evidence for residual PM-induced costs in the metacognitive-information condition. The combination of Experiments 1a and 1b resulted in N = 133 participants in the two PM conditions. PM performance data of these participants were submitted to a 2  2 ANOVA with the between-subjects factors experiment (1a vs. 1b) and condition (standard PM, metacognitive information). For this analysis, the statistical power to find a PM performance decrement of medium effect size (g2p ¼ :06) in the metacognitive-information condition as compared to the standard-PM condition was good, 1  b = .82 (cf. Cohen, 1968). Despite the good power, the joint analysis did not show a significant difference in PM performance between the standard-PM (M = .87, SE = .04) and the metacognitive-information condition (M = .86, SE = .04), F = 1.4, p = .236, g2p < :001, corroborating the notion that the reduction of PM-induced costs after metacognitive information was not accompanied by a PM-performance decline. This is evidence that PM-induced costs can reflect processes that are not functional for PM, because costs from a PM task with salient cues were not related to PM performance across conditions and PM performance sustained in the absence of observable costs (cf. Scullin et al., 2010). 6. Experiment 2 Results of Experiments 1a and 1b suggest that providing explicit information about PM-task demands appears to be effective in order to avoid non-functional PM-induced costs when PM cues are salient. It remains unclear, however, whether and to which extent anticipated task demands contribute to PM-induced costs when PM cues are not overly salient and attentional monitoring is necessary for PM performance (Einstein & McDaniel, 2010; McDaniel & Einstein, 2007; Smith, 2010; Smith et al., 2007). One could expect different effects of metacognitive-information manipulations in these situations. On the one hand, one could expect that people will stop monitoring after being told that a PM task is not that demanding. Consequently, they would miss most PM cues or, at least, detect significantly fewer cues. On the other hand, based on findings that people underestimate their PM performance in demanding PM situations (Meeks et al., 2007), one could expect that a functional level of monitoring is maintained even after anticipated task-demands are lowered via metacognitive instructions. That is, analogously to Experiments 1a and 1b, an overestimation of PM-task demands could result in cost components that are not functional for PM over and above the extent of attentional monitoring that is necessary and functional. To address this issue, in Experiment 2 we realized two PM conditions with non-salient cues, that is, words belonging to the animal category in an ongoing lexical-decision task (e.g., Hicks et al., 2005; Rummel, 2010). While holding the objective task demands constant across PM conditions, we again manipulated the anticipated PM-task demands to investigate their impact on attention–allocation strategies in common PM situations. Moreover, we used a different manipulation of anticipated PM-task demands to ensure that the effects were not due to the specifics of the manipulation used in Experiments 1a and 1b. A pseudo-priming procedure was used in which participants were told that the presentation of primes prior to each probe stimulus would influence their PM performance, that is, either fostering or hindering the identification of the PM cues. Analogously to the metacognitive information about PM-task demands given in the previous experiments, this manipulation should result in different expectations about PM-task demands while objective PM-task demands remained identical. Furthermore, we analyzed RTs on ongoing-task trials directly preceding PM trials as a function of PM accuracy (cf. West, Krompinger, & Bowry, 2005). Slowed RTs on trials prior to the occurrence of a PM cue have been shown to be predictive of enhanced PM performance under demanding PM conditions (Scullin et al., 2010). Thus, slower RTs on trials preceding correct 8 Excluding all participants who did not achieve 100% PM performance from the RT analysis did not change the pattern of results, rendering it unlikely that trade-off differences can account for the differences between conditions. 939 J. Rummel, T. Meiser / Consciousness and Cognition 22 (2013) 931–943 Table 3 Ongoing task performance as a function of condition (Experiment 2). Condition Accuracy Response times Response times on the six trials prior to a correct PM response Response times on the six trials prior to an incorrect PM response Difficult PM expectations Easy PM expectations Control .94 (.005) 792 (14) 853 (39) 868 (49) .94 (.005) 748 (13) 727 (34) 678 (42) .94 (.005) 705 (12) – – Note. Accuracy (i.e., adjusted mean proportion of accurate responses in the experimental block of the ongoing task) and response times (i.e., adjusted mean response times on words in milliseconds in the experimental block of the ongoing task) as a function of condition in Experiment 2. Standard errors of the means are displayed in parentheses. as compared to on trials preceding incorrect PM responses is evidence that RTs reflect processes that are functional for PM (see also Loft & Yeo, 2007, for a similar rationale). 6.1. Method 6.1.1. Participants Eighty-nine students participated for course credit or monetary compensation and were randomly assigned to one of three conditions. Data of two participants were excluded because they could not recall the PM task in a final questionnaire. The remaining sample consisted of 87 participants with 29 participants in a PM condition with easy PM expectations, 28 in a PM condition with difficult PM expectations, and 30 in the control condition without PM task. 6.1.2. Materials and procedure Materials were similar to the ones used in Experiments 1a and 1b, but lexical-decision trials were slightly changed for the manipulation of anticipated PM-task demands. Each trial started with a fixation cross of variable duration (250–750 ms), followed by a mask (40 ms), a pseudo-prime (10 ms), and again a mask (40 ms). According to pretests, the short presentation duration of the pseudo-primes did not allow for processing the content of the pictures consciously but allowed noticing that something was presented. Then, a probe (word or nonword) appeared on the screen until a response was made. Experimental sessions started with instructions for the lexical decision task. After accomplishing six practice trials and a baseline block of 106 trials (half words, half nonwords), instructions of the lexical decision task were repeated and in the two PM conditions they were expanded by the PM-task instructions. In both PM conditions, participants were asked to press the 7 key in response to members of the animal category. Then, participants in the two PM conditions received additional metacognitive instructions. Participants in one PM condition were told that the pictures presented subliminally on each trial would show the specific animal prior to the presentation of an animal name and thus render the detection of the animal names quite easy (i.e., easy PM expectations condition). Participants in the other PM condition were told that the pictures presented subliminally on each trial would distract their attention from the PM intention and thus render the detection of the animal names rather difficult (i.e., difficult PM expectations condition). In fact, a subliminally presented pseudo-prime without special meaning was randomly chosen from a body of 32 deeply zoomed-in pictures of buildings and plants at each trial. Because the primes were identical in all conditions, if they exerted any effect at all, they should influence PM performance in both PM conditions to the same extent. Participants in the control condition did not receive any PM instructions. After solving figural reasoning problems for five minutes, participants performed the experimental block with 106 lexicaldecision trials with the embedded PM trials. The PM cues were either ‘‘fox,’’ ‘‘hamster,’’ ‘‘wasp,’’ ‘‘zebra,’’ ‘‘carp,’’ and ‘‘blackbird’’ or ‘‘badger,’’ ‘‘lizard,’’ ‘‘dragonfly,’’ ‘‘lama,’’ ‘‘pike,’’ and ‘‘finch’’ and were presented at the Trials 20, 30, 45, 75, 90, 104. The two PM-cue lists were counterbalanced across conditions and the order of PM-cue occurrences was randomized. Finally, participants in the two PM conditions were asked to recall the PM task. 6.2. Results and discussion 6.2.1. Ongoing-task performance RTs and ongoing-task accuracy-rates in the experimental block of the ongoing task were analyzed to assess PM-induced costs. The same criteria as in Experiment 1a were used to control for outliers and artifactual costs and performance in the baseline block was again included as a covariate.9 Adjusted mean RTs and accuracy-rates for each condition are reported in Table 3. After trimming, RT-distributions in the PM condition with difficult PM expectations and in the control conditions remained significantly non-normal, all ps 6 .04. RT-distributions in the PM condition with low anticipated task demands 9 Baseline RTs (MDifficult PM expectations = 754; SE = 24; MEasy PM expectations = 680; SE = 24; Mcontrol = 730; SE = 23) did not differ significantly between conditions, F(2, 84) = 2.74, p = .070, g2p ¼ :06, whereas experimental-block RTs (MDifficult PM expectations = 823; SE = 26; MEasy PM expectations = 710; SE = 26; Mcontrol = 714; SE = 26) varied significantly with conditions, F(2, 84) = 6.11, p = .003, g2p ¼ :127. 940 J. Rummel, T. Meiser / Consciousness and Cognition 22 (2013) 931–943 did not differ significantly from a normal distribution, ps P .11, but for consistency, RTs of all three conditions were logtransformed. After transformation, RT-distributions did not differ significantly from a normal distribution, all ps P .11. We specified planned orthogonal contrasts in an analysis of covariance of RTs in the experimental block with condition (difficult PM expectations vs. easy PM expectations vs. control) as between-subjects factor and RTs in the baseline block as covariate. The first contrast compared both the two PM conditions with the No-PM control condition, to test whether the addition of a demanding PM task resulted in general PM-induced slowing. The second contrast compared the two PM conditions, in order to test whether the manipulation of anticipated task demands changed the level of costs. The results revealed the predicted pattern. RT-costs were significantly higher in the two PM conditions than in the control condition, t(83) = 4.06, p < .001, d = .89, indicating that the addition of the demanding PM task resulted in the engagement of significant cue-monitoring in both PM conditions. This is in line with previous findings using comparable tasks (e.g., Marsh et al., 2003). In addition to this general cost effect, the level of costs varied as a function of anticipated PM-task demands, that is, RT-costs were significantly higher in the difficult than in the easy PM expectations condition, t(83) = 2.07, p = .041, d = .45. This finding suggests that the level of cue-monitoring was further influenced by the additional metacognitive information about PM-task demands over and above the factual task demands. The same contrast as for the RT-analysis applied to ongoing-task accuracy did not yield significance, both ts < 1. 6.2.2. Memory for PM instructions and PM performance Two participants could not recalled the PM task in the final questionnaire and had been excluded from all analyses of Experiment 2 (see above). Proportions of accurate PM responses was similar in the difficult PM expectations condition (M = .59, SE = .07) and in the easy PM expectations condition (M = .55, SE = .06), t < 1, d = .12, suggesting that significant changes in PM-induced costs were not accompanied by significant PM-performance differences. 6.2.3. PM performance as a function of costs To investigate whether and to which extent PM-induced costs were associated with better PM performance in the two PM conditions, we correlated PM accuracy and RTs to words in the experimental block with RTs to words in the baseline block being partialled out of the latter variable to achieve a (baseline-adjusted) RT-cost measure. In the easy PM expectation condition, the semi-partial correlation turned out to be significant, r(28) = .37, t = 2.16, p = .040. In the difficult PM expectation condition, however, the same semi-partial correlation was numerically lower and not significant, r(27) = .30, t = 1.55, p = .133. This correlational pattern is preliminary evidence that the PM-induced costs in the difficult PM expectation condition were probably not functional for PM performance to the same extent as in the easy PM expectation condition. To further investigate the functional relationship between PM-induced costs and PM performance, RTs on the six ongoing-task trials preceding correct responses to PM cues were compared with RTs on the six trials preceding incorrect responses to PM cues as a function of PM condition. RTs on trials prior to a PM trial allow for assessing cue-monitoring which is directly functional for PM performance, as RTs on these trials should reflect cue-monitoring at the time the cue actually occurred (Loft & Yeo, 2007; West et al., 2005). We confined the analysis to the six trials preceding a PM trial, because others have shown that performance in a demanding PM task was improved when cue-monitoring was initiated six trials prior to the occurrence of the PM cue (Scullin et al., 2010). Mean RTs on the trials preceding PM trials with correct and incorrect PM responses are presented in Table 3 for each PM condition. An ANOVA with PM condition (difficult PM expectations, easy PM expectations) as between-subjects factor and PM accuracy (correct, incorrect response) as within-subjects factor yielded a marginally significant main effect of PM accuracy, F(1, 35) = 3.75, p = .061, g2p ¼ :10, and a significant main effect of PM condition, F(1, 35) = 7.33, p = .010, g2p ¼ :17. 10 In particular, across correct and incorrect PM responses, RTs preceding a PM cue were significantly slower in the difficult PM expectations condition (M = 860, SE = 43) than in the easy PM expectations condition (M = 702, SE = 37). This effect was qualified by a significant interaction with PM accuracy, F(1, 35) = 4.27, p = .046, g2p ¼ :11. To test the interaction, paired-sample t-tests were computed comparing RTs preceding correct vs. incorrect PM responses for each PM condition separately. Results showed that in the easy PM expectations condition, RTs preceding correct PM responses were significantly slower than RTs preceding incorrect PM responses, t(20) = 4.18, p < .001, d = 1.87, suggesting that time-consuming cue-monitoring was engaged. In the difficult PM expectations condition, in which RTs preceding a PM cue were generally slower than in the PM condition with low anticipated PM demands, RTs preceding correct PM responses did not differ from RTs preceding incorrect PM responses, t < 1, d = .04. These findings suggests that the enlarged time costs in the difficult PM expectations condition as compared to the easy PM expectations condition were contaminated with non-functional processes that were not predictive for PM performance. 7. General discussion The major goal of the present research was to investigate the role of metacognitive expectations about PM-task demands for attention–allocation strategies in PM. It has been suggested that during intention formation individuals decide how to distribute their attentional resources between a PM task and an ongoing task (Hicks et al., 2005) and that these decisions follow metacognitive beliefs about the to-be-expected PM-task demands (Einstein & McDaniel, 2008). However, up to our 10 Participants with PM performance of zero or 1 are not included because these participants provided missing data for one cell of the design. J. Rummel, T. Meiser / Consciousness and Cognition 22 (2013) 931–943 941 knowledge, this notion has never been tested empirically. Therefore, in a series of experiments, we investigated whether attention–allocation strategies of PM reflect the expectations people hold about PM-task demands at intention formation. Results from Experiments 1a and 1b showed that PM-induced costs were significantly reduced when participants were informed about the low attentional demands imposed by a PM task with salient cues. Simultaneously, a high level of PM performance was maintained even in the absence of PM-induced costs. These results are in line with a multiprocess view of PM, assuming that in non-demanding PM situations performance can strongly rely on rather effortless spontaneous retrieval processes (McDaniel & Einstein, 2007). Building on these findings, in Experiment 2 we investigated to which extent anticipated task demands can affect the allocation of attention to a demanding PM task with non-salient PM cues. We found that independent of the anticipated PM-task demands, all participants engaged in a significant level of cue-monitoring as reflected by PM-induced costs. This finding is in line with most theories of PM because they agree that PM performance should rely heavily on cue-monitoring processes under such demanding PM-task conditions (cf. McDaniel & Einstein, 2007; Smith, 2010). Although factual PM-task demands were identical, however, the PM-induced costs to the ongoing task varied with anticipated demands, demonstrating that the level of costs is sensitive to expectations about PM-task demands even when the PM cues are not salient. Taken together, the present results provide empirical evidence that attention–allocation strategies in PM are calibrated to subjective expectations about PM-task demands in a given situation. Furthermore, the present research suggests that a combination of declarative information and prior experience effectively changes PM demands expectations (cf. Koriat et al., 2004). In the present Experiment 1b, PM-task experience alone did not change PM processing to the same extent as the combination of prior experience and declarative information about PM demands did. Because participants of Experiment 1b had only view opportunities to gain online experience with the PM task, however, it remains an open question whether a more extensive task experience might be as effective as providing declarative information for a the calibration of attention–allocation strategies. Another goal of the present research was to further investigate the nature of PM-induced costs. In particular, we were interested in the functional relationship between PM-induced costs and PM performance. Previous findings suggested that PM-induced costs are not always functionally related to PM performance. For instance, in a comparison of participants who showed substantial costs and participants who showed no costs in a non-demanding PM task, Einstein et al. (2005) did not find PM performance differences (but see also Smith, 2010, for potential limitations of these findings). The results from the joint analysis of the present Experiments 1a and 1b provided strong evidence that PM performance in a PM task with salient cues remained largely unaffected by changes of the cognitive processes causing PM-induced costs. These results can help to understand inconsistencies in previous research on salient PM cues (Harrison & Einstein, 2010; Smith et al., 2007). Individuals may not be very aware that a salient cue can be identified with low cognitive effort and engage in non-necessary cuemonitoring resulting in PM-induced costs. This notion is supported by the pilot study in which estimated PM performance in a standard-PM condition was quite low (i.e., 61%) which was lower than the PM performance assessed in Experiments 1a and 1b under comparable conditions. However, findings of Experiments 1a and 1b also imply that individuals can optimize their attention–allocation strategies based on explicit information about PM-task demands. This finding lends empirical support to Einstein and McDaniel’s (2010) recent theorizing that the mere existence of costs does not imply that these costs are necessary to realize an intention. Further evidence for the existence of non-functional cost components comes from the analysis of trials preceding PM trials in Experiment 2. While there was a general slowing on the trials directly preceding PM trials in the difficult PM expectations condition as compared to the easy PM expectations condition, only in the easy PM expectations condition was a slowing on these trials predictive for successful PM performance. This result suggests that some of the additional costs caused by the experimentally increased anticipated PM-task demands reflected processes that were not functional for PM performance or at least that PM improves not to the same extent to which PM-induced costs increase after a certain (optimal) cost threshold is reached. It remains to be tested which processes underlie these non-functional costs. It might be the case that they represent an (overly) cautious ongoing-task responding (Boywitt & Rummel, 2012) or a different (and inefficient) type of monitoring (cf. Albinski, Sedek, & Kliegel, 2012). The present results, however, already provide first insights into why non-functional costs can occur inasmuch as they result from an inefficient allocation of attention in favor of the PM task due to biased metacognitive expectations about PM-task demands. Notably, whenever PM task-demand expectations are biased, and there is evidence that they usually are (cf. Meeks et al., 2007; Rummel et al., 2013; Schnitzspahn et al., 2011), observed costs may only partially reflect functional cue-monitoring of PM. This finding has important implications for PM research in general, because such non-functional processes may often contribute to differences in observable PM-induced costs. Finally, the present results also suggest that providing declarative information about the cognitive demands of a PM-task can be an effective strategy to avoid costs that are not functional for PM. Given the importance of PM in daily life and the tremendous consequences which can result from PM failures (cf. Dismukes, 2008; McDaniel & Einstein, 2007), ways to optimize PM performance have found particular attention by researchers. These include non-cognitive strategies like the usage of external reminders (Einstein & McDaniel, 1990; Henry, Rendell, Phillips, Dunlop, & Kliegel, 2012) or enactment of the intended action (Pereira, Ellis, & Freeman, 2012), but also cognitive strategies, such as mentally simulating the intended action (Brewer, Knight, Meeks, & Marsh, 2011; Meiser & Rummel, 2012), intention rehearsal (Stone, Dismukes, & Remington, 2001), or forming specific if–then associations between the PM cue and the intended action (Gollwitzer, 1999; McDaniel & Scullin, 2010; Rummel, Einstein, & Rampey, 2012). The present results suggest that improving the metacognitive understanding of 942 J. Rummel, T. Meiser / Consciousness and Cognition 22 (2013) 931–943 PM-task demands via explicit information seems to be a another promising (meta-)cognitive strategy allowing to optimize attention–allocation strategies of PM beyond an optimization based on mere increased task experience. Acknowledgments The present research was supported by a Grant from the German Research Foundation (DFG) to the second author. We thank C. Dennis Boywitt for his helpful comments on a previous version of this manuscript. References Albinski, R., Sedek, G., & Kliegel, M. (2012). Differences in target monitoring in a prospective memory task. Journal of Cognitive Psychology, 24, 916–928. Boywitt, C. D., & Rummel, J. (2012). A diffusion model analysis of task interference effects in prospective memory. Memory & Cognition, 40, 19–27. Brewer, G. A., Knight, J., Meeks, J. T., & Marsh, R. L. (2011). On the role of imagery in event-based prospective memory. Consciousness & Cognition, 20, 901–907. Castel, A. D., McCabe, D. P., & Roediger, H. L. (2007). Illusions of competence and overestimation of associative memory for identical items: Evidence from judgments of learning. Psychonomic Bulletin & Review, 14, 107–111. Cohen, A.-L., Jaudas, A., & Gollwitzer, P. M. (2008). Number of cues influences the cost of remembering to remember. Memory & Cognition, 36, 149–156. Cohen, J. (1968). Statistical power analysis for the behavioral sciences. New Jersey: Lawrence Erlbaum Associates. Dismukes, R. K. (2008). Prospective memory in aviation and everyday settings. In M. Kliegel, M. A. McDaniel, & G. O. Einstein (Eds.), Prospective memory: Cognitive, neuroscience, developmental, and applied perspectives (pp. 411–428). New York: Lawrence Erlbaum. Dunlosky, J., & Ariel, R. (2011). Self-regulated learning and the allocation of study time. Psychology of Learning and Motivation, 54, 103–140. Einstein, G. O., & McDaniel, M. A. (1990). Normal aging and prospective memory. Journal of Experimental Psychology: Learning, Memory, and Cognition, 16, 717–726. Einstein, G. O., & McDaniel, M. A. (2005). Prospective memory: Multiple retrieval processes. Current Directions in Psychological Science, 14, 286–290. Einstein, G. O., & McDaniel, M. A. (2008). Prospective memory and metamemory: The skilled use of basic attentional and memory processes. In A. S. Benjamin & B. Ross (Eds.). The psychology of learning and motivation (Vol. 48, pp. 145–173). San Diego: Elsevier. Einstein, G. O., & McDaniel, M. A. (2010). Prospective memory and what costs do not reveal about retrieval processes: A commentary on Smith, Hunt, McVay, and McConnell (2007). Journal of Experimental Psychology: Learning Memory and Cognition, 36, 1082–1088. Einstein, G. O., McDaniel, M. A., Manzi, M., Cochran, B., & Baker, M. (2000). Prospective memory and aging: Forgetting intentions over short delays. Psychology and Aging, 15, 671–683. Einstein, G. O., McDaniel, M. A., Thomas, R., Mayfield, S., Shank, H., Morrisette, N., et al (2005). Multiple processes in prospective memory retrieval: Factors determining monitoring versus spontaneous retrieval. Journal of Experimental Psychology: General, 134, 327–342. Ellis, J. A., & Kvavilashvili, L. (2000). Prospective memory in 2000: Past, present, and future directions. Applied Cognitive Psychology, 14, S1–S9. Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G⁄Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39, 175–191. Förster, J., & Strack, F. (1998). Subjective theories about encoding may influence recognition: Judgmental regulation in human memory. Social Cognition, 16, 78–92. Gollwitzer, P. M. (1999). Implementation intentions: Strong effects of simple plans. American Psychologist, 54, 493–503. Graf, P., Uttl, B., & Dixon, R. (2002). Pro- and retrospective memory in adulthood. In P. Graf & N. Ohta (Eds.), Lifespan memory development (pp. 257–282). Cambridge, MA: MIT Press. Harrison, T., & Einstein, G. O. (2010). Prospective memory: Are preparatory attentional processes necessary for a single focal cue? Memory & Cognition, 38, 860–867. Henry, J. D., Rendell, P. G., Phillips, L. H., Dunlop, L., & Kliegel, M. (2012). Prospective memory reminders: A laboratory investigation of initiation source and age effects. The Quarterly Journal of Experimental Psychology, ifirst.. http://dx.doi.org/10.1080/17470218.17472011.17651091. Hicks, J. L., Marsh, R. L., & Cook, G. I. (2005). Task interference in time-based, event-based, and dual intention prospective memory conditions. Journal of Memory and Language, 53, 430–444. Hines, J. C., Touron, D. R., & Hertzog, C. (2009). Metacognitive influences on study time allocation in an associative recognition task: An analysis of adult age differences. Psychology and Aging, 24, 462–475. Kliegel, M., Martin, M., McDaniel, M. A., & Einstein, G. O. (2001). Varying the importance of a prospective memory task: Differential effects across time- and event-based prospective memory. Memory, 9, 1–11. Koriat, A. (1997). Monitoring one’s own knowledge during study: A cue-utilization approach to judgments of learning. Journal of Experimental Psychology: General, 126, 349–370. Koriat, A., & Bjork, R. A. (2005). Illusions of competence in monitoring one’s knowledge during study. Journal of Experimental Psychology: Learning Memory and Cognition, 31, 187–194. Koriat, A., Bjork, R. A., Sheffer, L., & Bar, S. K. (2004). Predicting one’s own forgetting: The role of experience-based and theory-based processes. Journal of Experimental Psychology: General, 133, 643–656. Kvavilashvili, L., & Ellis, J. (1996). Varieties of intentions: Some classifications and distinctions. In M. A. Brandimonte, G. O. Einstein, & M. A. McDaniel (Eds.), Prospective memory: Theory and applications (pp. 23–52). Mahwah, NJ: Lawrence Erlbaum. Loft, S., Kearney, R., & Remington, R. (2008). Is task interference in event-based prospective memory dependent on cue presentation? Memory & Cognition, 36, 139–148. Loft, S., & Yeo, G. (2007). An investigation into the resource requirements of event-based prospective memory. Memory & Cognition, 35, 263–274. Marsh, R. L., Cook, G. I., & Hicks, J. L. (2006). Task interference from event-based intentions can be material specific. Memory & Cognition, 34, 1636–1643. Marsh, R. L., Hancock, T. W., & Hicks, J. L. (2002). The demands of an ongoing activity influence the success of event-based prospective memory. Psychonomic Bulletin & Review, 9, 604–610. Marsh, R. L., Hicks, J. L., & Cook, G. I. (2005). On the relationship between effort toward an ongoing task and cue detection in event-based prospective memory. Journal of Experimental Psychology: Learning, Memory, and Cognition, 31, 68–75. Marsh, R. L., Hicks, J. L., & Cook, G. I. (2006). Task interference from prospective memories covaries with contextual associations of fulfilling them. Memory & Cognition, 34, 1037–1045. Marsh, R. L., Hicks, J. L., Cook, G. I., Hansen, J. S., & Pallos, A. L. (2003). Interference to ongoing activities covaries with the characteristics of an event-based intention. Journal of Experimental Psychology: Learning, Memory, and Cognition, 29, 861–870. Mazzoni, G., Cornoldi, C., & Marchitelli, G. (1990). Do memorability ratings affect study-time allocation. Memory & Cognition, 18, 196–204. McBride, D. M., & Abney, D. H. (2012). A comparison of transfer-appropriate and multi-process frameworks for prospective memory performance. Experimental Psychology, 59, 190–198. McDaniel, M. A., & Einstein, G. O. (2007). Prospective memory: An overview and synthesis of an emerging field. Los Angeles: Sage. McDaniel, M. A., & Scullin, M. K. (2010). Implementation intention encoding does not automatize prospective memory responding. Memory & Cognition, 38, 221–232. J. Rummel, T. Meiser / Consciousness and Cognition 22 (2013) 931–943 943 Meeks, J. T., Hicks, J. L., & Marsh, R. L. (2007). Metacognitive awareness of event-based prospective memory. Consciousness & Cognition, 16, 997–1004. Meier, B., & Graf, P. (2000). Transfer appropriate processing for prospective memory test. Applied Cognitive Psychology, 14, 11–27. Meier, B., von Wartburg, P., Matter, S., Rothen, N., & Reber, R. (2011). Performance predictions improve prospective memory and influence retrieval experience. Canadian Journal of Experimental Psychology, 65, 12–18. Meiser, T., & Rummel, J. (2012). False prospective memory responses as indications of automatic processes in the initiation of delayed intentions. Consciousness & Cognition, 21, 1509–1516. Meiser, T., Sattler, C., & von Hecker, U. (2007). Metacognitive inferences in source memory judgements: The role of perceived differences in item recognition. The Quarterly Journal of Experimental Psychology, 60, 1015–1040. Meiser, T., & Schult, J. C. (2008). On the automatic nature of the task-appropriate processing effect in event-based prospective memory. European Journal of Cognitive Psychology, 20, 290–311. Nelson, T. O., & Dunlosky, J. (1991). When people’s judgments of learning (Jols) are extremely accurate at predicting subsequent recall – The delayed-Jol effect. Psychological Science, 2, 267–270. Pereira, A., Ellis, J., & Freeman, J. (2012). Is prospective memory enhanced by cue-action semantic relatedness and enactment at encoding? Consciousness & Cognition, 21, 1257–1266. Ratcliff, R. (1993). Methods for dealing with reaction time outliers. Psychological Bulletin, 114, 510–532. Rosenthal, R., & Rosnow, R. (1985). Contrast analysis: Focused comparisons in the analysis of variance. Cambridge, UK: Cambridge University Press. Rummel, J. (2010). Psychological distance to a prospective memory cue influences the probability of fulfilling a delayed intention. Memory, 18, 284–292. Rummel, J., Boywitt, C. D., & Meiser, T. (2011). Assessing the validity of multinomial models using extraneous variables: An application to prospective memory. The Quarterly Journal of Experimental Psychology, 64, 2194–2210. Rummel, J., Einstein, G. O., & Rampey, H. (2012). Implementation intention encoding in a prospective memory task enhances spontaneous retrieval of intentions. Memory, 20, 803–817. Rummel, J., Kuhlmann, B. G., & Touron, D. R. (2013). Performance predictions affect attentional processes of event-based prospective memory. Consciousness & Cognition, 22, 729–741. Schnitzspahn, K. M., Zeintl, M., Jäger, T., & Kliegel, M. (2011). Metacognition in prospective memory: Are performance predictions accurate? Canadian Journal of Experimental Psychology, 65, 19–26. Scullin, M. K., McDaniel, M. A., & Einstein, G. O. (2010). Control of cost in prospective memory: Evidence for spontaneous retrieval processes. Journal of Experimental Psychology: Learning, Memory, and Cognition, 36, 190–203. Smith, R. E. (2003). The cost of remembering to remember in event-based prospective memory: Investigating the capacity demands of delayed intention performance. Journal of Experimental Psychology: Learning, Memory, and Cognition, 29, 347–361. Smith, R. E. (2010). What costs do reveal and moving beyond the cost debate: Reply to Einstein and McDaniel (2010). Journal of Experimental Psychology: Learning, Memory, and Cognition, 36, 1089–1095. Smith, R. E., Hunt, R. R., McVay, J. C., & McConnell, M. D. (2007). The cost of event-based prospective memory: Salient target events. Journal of Experimental Psychology: Learning, Memory, and Cognition, 33, 734–746. Son, L. K., & Metcalfe, J. (2000). Metacognitive and control strategies in study-time allocation. Journal of Experimental Psychology: Learning, Memory, and Cognition, 26, 204–221. Stone, M., Dismukes, K., & Remington, R. (2001). Prospective memory in dynamic environments: Effects of load, delay, and phonological rehearsal. Memory, 9, 165–176. Strack, F., Förster, J., & Werth, L. (2005). ‘‘Know thyself!’’ The role of idiosyncratic self-knowledge in recognition memory. Journal of Memory and Language, 52, 628–638. Touron, D. R., Oransky, N., Meier, M. E., & Hines, J. C. (2010). Metacognitive monitoring and strategic behaviour in working memory performance. The Quarterly Journal of Experimental Psychology, 63, 1533–1551. West, R., Krompinger, J., & Bowry, R. (2005). Disruptions of preparatory attention contribute to failures of prospective memory. Psychonomic Bulletin & Review, 12, 502–507.

Consciousness and Cognition 22 (2013) 806–809 Contents lists available at SciVerse ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Commentary Measuring and testing awareness of emotional face expressions Kristian Sandberg a,b,⇑, Bo Martin Bibby c, Morten Overgaard a,d a Cognitive Neuroscience Research Unit, Aarhus University Hospital, Nørrebrogade 44, Building 10G, 8000 Aarhus C, Denmark UCL Institute of Cognitive Neuroscience, University College London, 17 Queen Square, WC1N 3AR London, United Kingdom Department of Biostatistics, Aarhus University, Bartholins Allé 2, Building 1261, 8000 Aarhus C, Denmark d Cognitive Neuroscience Research Unit, Dept. of Communication and Psychology, Aalborg University, Kroghstræde 3, 9220 Aalborg Ø, Denmark b c a r t i c l e i n f o Article history: Received 26 February 2013 Available online 30 May 2013 Keywords: Consciousness Perceptual awareness scale (PAS) Confidence ratings (CR) Post-decision wagering (PDW) Emotion Awareness a b s t r a c t Comparison of behavioural measures of consciousness has attracted much attention recently. In a recent article, Szczepanowski et al. conclude that confidence ratings (CR) predict accuracy better than both the perceptual awareness scale (PAS) and post-decision wagering (PDW) when using stimuli with emotional content (fearful vs. neutral faces). Although we find the study interesting, we disagree with the conclusion that CR is superior to PAS because of two methodological issues. First, the conclusion is not based on a formal test. We performed this test and found no evidence that CR predicted accuracy better than PAS (p = .4). Second, Szczepanowski et al. used the present version of PAS in a manner somewhat different from how it was originally intended, and the participants may not have been adequately instructed. We end our commentary with a set of recommendations for future studies using PAS. Ó 2013 Elsevier Inc. All rights reserved. 1. Introduction Over the last few years, behavioural methods for assessing consciousness have become a topic of much scientific debate, and the main focus has been on comparing different measures (scales) in order to examine if one is superior to the others (Overgaard & Sandberg, 2012). A large part of this debate was motivated by the proposal of post-decision wagering (PDW) as a so-called ‘‘objective’’ measure of consciousness by Persaud, McLeod, and Cowey (2007). Here, the authors showed, among other findings, an imperfect correlation between task accuracy in a visual detection task and wagering behaviour of a blindsight patient, GY, thus indicating that GY was not fully aware of the targets that he was nevertheless able to report. However, PDW was quickly criticized for being an indirect measure of mental content rather than awareness (Seth, 2008), and it was pointed out that loss aversion (Schurger & Sher, 2008) and/or other factors could prevent participants from wagering high in the presence of sensory awareness (Clifford, Arabzadeh, & Harris, 2008). These claims were subsequently supported by empirical studies finding PDW to be inferior to simpler measures such as confidence ratings (CR) and ratings of the clarity of the visual experience (Dienes & Seth, 2010; Sandberg, Timmermans, Overgaard, & Cleeremans, 2010; Wierzchoń, Asanowicz, Paulewicz, & Cleeremans, 2012). Overlapping with, and slightly predating this debate, work in our group focused on examining which scale participants prefer to use for reporting conscious experiences if allowed to construct the scale themselves (Ramsøy & Overgaard, 2004). This resulted in the 4-point so-called perceptual awareness scale (PAS), ranging from ‘‘no experience’’ over ‘‘a brief glimpse’’ to ‘‘an almost clear experience’’ and ending at ‘‘a clear experience’’. PAS was subsequently compared to a dichotomous measure of awareness for normal participants (Overgaard, Rote, Mouridsen, & Ramsøy, 2006) and for a blindsight patient DOI of original article: http://dx.doi.org/10.1016/j.concog.2012.12.003 ⇑ Corresponding author at: Cognitive Neuroscience Research Unit, Aarhus University Hospital, Nørrebrogade 44, Building 10G, 8000 Aarhus C, Denmark. E-mail address: krissand@rm.dk (K. Sandberg). 1053-8100/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.concog.2013.04.015 K. Sandberg et al. / Consciousness and Cognition 22 (2013) 806–809 807 (Overgaard, Fehl, Mouridsen, Bergholt, & Cleeremans, 2008), as well as to 4-point CR and PDW scales (Sandberg et al., 2010). In all cases, PAS was found to be superior to the competing scales, most frequently by showing better accuracy-awareness correlation and/or less subliminal perception when participants reported not having seen the target. In other words, these studies suggested that the best results were obtained by asking participants to report their experiences directly and allowing them to do this on a scale created either by themselves or other participants presented with similar stimuli. In another recent article Szczepanowski, Traczyk, Wierzchoń, and Cleeremans (2013) compared participants’ use of CR, PAS, PDW during a visual identification task, now using faces with fearful and neutral expressions instead of the simple visual shapes used in Sandberg et al. (2010). Szczepanowski et al. estimated the relationship between accuracy and awareness for each scale using Pearson chi-square tests of independence as well as type 2 receiver operating characteristics (ROC) analysis. Based on the resulting values, the authors argued that CR produced the best relationship between accuracy and awareness, PAS the second best, and PDW the worst, thus indicating that CR was the most exhaustive subjective measure when examining stimuli with emotional content. Although we find the article very interesting and a relevant contribution to the field, we disagree with the authors’ conclusion that confidence ratings are more exhaustive than PAS in paradigms with emotional stimuli for two reasons. The first is the employed statistical method, and the second is the manner in which PAS was used. In the following, we will elaborate these points in detail. 2. Statistical model selection and tests of differences As mentioned above, Szczepanowski et al. used type 2 ROC analysis and Pearson chi-square tests to examine if task accuracy varied across awareness ratings within each scale and compare the obtained chi2 values numerically in order to rank the scales. The type 2 ROC analysis indicated that all scales were equally sensitive (mean sensitivity was between 0.58 and 0.59 for all scales) although no formal test was performed. The conclusions regarding the inter-scale differences were based on the chi-square test. We find that this method is inappropriate in two ways: First, it assumes that all observations are independent, which is not the case as each participant contributed with 80 observations for each scale. An optimal statistical model should thus take into account that the data originates from N clusters (where N is the number of participants). Second, even if it was reasonable to assume independence, then ranking of chi2 values is not a formal comparison of scales and does thus not allow us to conclude whether a difference between scales is significant or not. An optimal model should allow for such formal testing of the null hypotheses that the relationship between task accuracy and awareness rating is equally good for scale 1 and 2, scale 1 and 3, and scale 2 and 3. Both these goals can be achieved by use of logistic regression. Logistic regression furthermore allows for a second test of scale exhaustiveness – tests using the so-called ‘‘guessing criterion’’ (Dienes, Altmann, Kwan, & Goode, 1995). Using this, performance is compared between scales for those cases where participants claim to be guessing or to have no experience (that is, they use the lowest awareness rating). It has been argued that above-chance performance in these cases reflect unconscious processing. Yet if one scale finds unconscious processing but another does not, it is highly likely that the first scale is less exhaustive (unless the second scale is not exclusive, i.e. if it misclassifies unconscious processing as conscious processing) ((Sandberg et al., 2010), but see also Dienes and Seth (2010) and Timmermans, Sandberg, Cleeremans, and Overgaard (2010)). Szczepanowski et al. have kindly allowed us access to the original data to test for differences in: (a) how well each scale predicted task accuracy in general, and (b) how much subliminal perception was indicated at the lowest awareness ratings. We created a logistic regression mixed model with accuracy as the dependent variable and with scale and awareness rating as independent variables, and with participant as a random effect. First, we conducted pair-wise comparisons between all three scales in order to test if one scale predicted accuracy significantly better than the others. We found that both CR and PAS predicted accuracy significantly better than PDW (z = 4.07, p < .001 for CR vs. PDW, and z = 2.28, p = .023 for PAS vs. PDW), but importantly, no significant difference was observed between CR and PAS (z = 0.87, p = .39). Overall, this analysis thus found both CR and PAS to be superior to PDW, but there was no evidence for a significant difference between CR and PAS. Second, we calculated task accuracy at the lowest awareness rating for each scale and compared this to the chance-level of 50%. For CR, task accuracy was 42%, 95%-CI: (28;55)%, which was not significantly different from chance (z = 1.21, p = .23). For PAS, task accuracy was 46%, 95%-CI: (41;52)%, which was not significantly different from chance (z = 1.19, p = .24). For PDW, task accuracy was 56%, 95%-CI: (50;62)%. This was almost but not quite significantly different from chance (z = 1.87, p = .06). Furthermore, we found that accuracy for awareness rating 1 was significantly higher for PDW than for both CR (z = 2.75, p = .006) and PAS (z = 2.15, p = .032), but importantly, no significant difference was observed between CR and PAS (z = 0.77, p = .44). Overall, we thus failed to reject the null hypothesis that accuracy was at chance for all scales although particularly for PDW, this may have been a matter of statistical power. The accuracy was significantly higher for PDW than for CR and possibly also than for PAS. Again, it is important to note that no significant difference was found between CR and PAS. In summary, we found no evidence of a significant difference in the exhaustiveness of CR and PAS, but both scales were significantly more exhaustive than PDW. Compared to Sandberg et al. (2010), the main impact of Szczepanowski et al.’s emotional stimuli were thus a lack of a statistically significant difference between CR and PAS. The absence of a statistically significant difference between PAS and CR could, in principle, be a matter of statistical power, yet as mentioned above, it could also be related to the manner in which PAS was used. Below, we will discuss this last point in detail. 808 K. Sandberg et al. / Consciousness and Cognition 22 (2013) 806–809 3. The use of PAS Szczepanowski et al. report that the PAS ‘‘is a 4-point verbal scale that attempts to measure the quality of conscious experience directly. It asks participants to evaluate the visibility of the percept as subjective certainty of its presence (the metacognitive judgment of a percept’s accessibility’’ (p. 213). However, this interpretation is only partially true. In the publication originally introducing PAS, Ramsøy and Overgaard (2004) wrote: ‘‘In describing and reporting sensations in terms of clearness, it is important to make the distinction between degrees of clearness and degrees of certainty about one’s answer’’ (p. 10). Thus, we would like to emphasize that PAS should not be equated with ‘‘subjective certainty’’. Essentially, ‘‘subjective certainty’’ is exactly what CR measures, and, accordingly, if PAS was introduced to participants by Szczepanowski et al. using a ‘‘subjective certainty’’ terminology, the non-significant differences between CR and PAS conditions are not surprising. We acknowledge, of course, that some judgement is needed when rating on the PAS (and one may be more or less confident in the accuracy of the judgment), but our main point here is that this is related to a comparison of the (remembered) visual experience and the scale step description and not an assessment of the probability of a stimulus being present based on the (remembered) visual experience. Ramsøy and Overgaard (2004) report that PAS is constructed from the reasoning that methods for subjective reports of conscious experience should be developed in collaboration with the reporting subjects. Thus, they make no claim that 4point scales should be preferred a priori – on the contrary, they argue that one cannot prefer any particular scale construction a priori. In the study, participants generated the scale while reporting their experience of the colour, shape, and position of simple figures. The validity of the PAS in later studies may reasonably be assumed to depend on the similarity of the stimulus material in the original and later study as well the how well the participants were instructed in the interpretation of scale step descriptions. In most later studies performed by our group using PAS (e.g. Overgaard, Nielsen, & Fuglsang-Frederiksen, 2004; Overgaard et al., 2006; Overgaard et al., 2008), the exact meaning of the individual labels were discussed with the subjects and tested in several pilot trials where the experimenter made sure that there were no misunderstandings. One crucial aspect is the distinction between ‘‘brief glimpse’’ and an ‘‘almost clear experience’’. For brief glimpses, there is a conscious experience caused by the presentation of the relevant stimulus, yet the participant is nevertheless unable to report the content of this experience. Typically, this experience is a vague short-lived sensation that ‘‘something was there’’ or simply that ‘‘it was different from nothing at all’’ (see reviews in Overgaard and Sandberg (2012), and Overgaard (2012)). Few participants understand this distinction immediately if the procedure does not include thorough instruction, discussion and pilot trials. Indeed, recent (not yet published) findings indicate that in-depth instructions improve the accuracy-awareness correlation for PAS. The absence of in-depth instruction may thus have been a contributing factor to the results of Szczepanowski et al. Yet it should be noted that only written instructions were given on scale use in Sandberg et al. (2010) to ensure that participants were not given more in-depth instructions than participants using CR and PDW. For this reason, it is unlikely that verbal instructions alone are the cause of the difference in the findings. The other possibility is the nature of the stimuli used in the experiment. Most studies using PAS, have used geometric figures very similar to those used by Ramsøy and Overgaard (2004) and asked participants to report the shape (Overgaard et al., 2008; Sandberg, Bibby, Timmermans, Cleeremans, & Overgaard, 2011; Sandberg et al., 2010; Schwiedrzik, Singer, & Melloni, 2011) while one study used different, but still very simple stimuli, and asked participants to report the position (Overgaard et al., 2006). Common to these studies is that perceiving only part of a stimulus would typically allow the participant to answer correctly (i.e. seeing part of a circle allows the participant to infer that it cannot have been a triangle). The task used by Szczepanowski et al. is markedly different from the tasks used in these above-mentioned previous studies and the application of PAS is not straightforward. For instance, it is possible to perceive several features of a face (e.g. the nose, part of the hair and eye-brows) without improving classification accuracy of emotional content. Similarly, the perception of just a few features (e.g. the eyes or the mouth) may improve classification accuracy significantly. The stimuli also differ simply in terms of their overall complexity. Accordingly, it may be too hasty to apply PAS in its ‘‘original form’’ to their study, and having applied it, it is difficult whether the results are caused by the emotional content of the stimuli, the overall stimulus complexity, or participants reporting on overall visual clarity rather than clarity of the relevant features. Of course, we welcome very much the integration of subjective and objective measures and the use of PAS when relevant. We acknowledge that previous studies have not been entirely consistent in the use of PAS regarding stimulus material and instructions, and for this reason, we would like to make some recommendations for future experiments: 1. If necessary and when in doubt to re-do the entire calibration procedure and create a new scale as explained in Ramsøy and Overgaard (2004) rather than to just ‘‘import’’ the 4-point scale. 2. If one decides not to re-do the original procedure, always include (1) a full instruction explaining the meaning of each scale point in detail, (2) a pilot test with a good amount of trials (e.g. 30–50) in which the experimenter interrupts the subject frequently to ask about the use of the individual scale points (e.g. ‘‘I noticed you just reported ‘‘brief glimpse’’ – why did you do that/what did you mean with that/how would you define brief glimpse?’’). If one does not do 1 or 2, we would be very reluctant to use PAS to test the correlation between subjective experience and any other measure. K. Sandberg et al. / Consciousness and Cognition 22 (2013) 806–809 809 4. Conclusions Overall, we disagree with the claim of Szczepanowski et al. that CR is more exhaustive than PAS when using emotional stimuli. From a statistical perspective, we argue that the data needs to be appropriately modelled and formal tests are needed to support the conclusions. Using logistic regression, we fail to reject the null hypothesis that PAS and CR predict accuracy ratings to the same extent, and neither are we able to reject the null hypothesis that the two scales indicate the same amount of subliminal perception when participants report to be guessing or have no visual experience of the stimulus – in fact, we found no evidence for subliminal perception for either scale. In contrast, we find evidence to support that both scales are more exhaustive than PDW. We therefore see the study mainly as evidence that PDW performs poorly when using stimuli with emotional content. Based on this finding, we judge that there is now convincing evidence that PDW should only be used when participants are unable to use other, more direct measures such as CR or PAS (e.g. in studies using non-human animals), and when doing this, analyses should take loss aversion into account. Turning to the procedures of the study, we are concerned that PAS was used in a manner that is somewhat different from how it was originally intended. The idea behind PAS was that it should be a scale that reflects the way that participants prefer to report. Therefore, the current PAS should not be seen as a single scale to be used in all tasks and for all stimuli, but rather as something to be used in visual identification tasks using simple stimuli. In the present study, PAS was used with somewhat more complex stimuli (faces) with several features of which only a small part were relevant (those separating fearful and neutral expressions). We are also concerned that the instructions to participants may not have clearly distinguished subjective visibility and certainty and that the scale points may not have been adequately explained. We recommend that ideally (though not always practically possible), it should be tested how participants prefer to report their visual experience for the particular stimulus type used. If this is not possible, we recommend that the experimenter thoroughly informs the participants of the distinction between certainty of accuracy and clarity of visual experience as well as how participants of previous experiments have defined and explained the meaning of the scale steps. References Clifford, C., Arabzadeh, E., & Harris, J. A. (2008). Getting technical about awareness. Trends in Cognitive Sciences, 12(2), 54–58. http://dx.doi.org/10.1016/ j.tics.2007.11.009. Dienes, Z., Altmann, G. T. M., Kwan, L., & Goode, A. (1995). Unconscious knowledge of artificial grammars is applied strategically. Journal of Experimental Psychology: Learning, Memory, and Cognition, 21(5), 1322–1338. http://dx.doi.org/10.1037/0278-7393.21.5.1322. Dienes, Z., & Seth, A. K. (2010). Measuring any conscious content versus measuring the relevant conscious content: Comment on Sandberg et al.. Consciousness and Cognition, 19(4), 1079–1080. http://dx.doi.org/10.1016/j.concog.2010.03.009. Overgaard, M. (2012). Blindsight: Recent and historical controversies on the blindness of blindsight. Wiley Interdisciplinary Reviews: Cognitive Science, 3(6), 607–614. http://dx.doi.org/10.1002/wcs.1194. Overgaard, M., Fehl, K., Mouridsen, K., Bergholt, B., & Cleeremans, A. (2008). Seeing without seeing? Degraded conscious vision in a blindsight patient. PLoS ONE, 3(8), e3028. http://dx.doi.org/10.1371/journal.pone.0003028. Overgaard, M., Nielsen, J. F., & Fuglsang-Frederiksen, A. (2004). A TMS study of the ventral projections from V1 with implications for the finding of neural correlates of consciousness. Brain and Cognition, 54(1), 58–64. Overgaard, M., Rote, J., Mouridsen, K., & Ramsøy, T. Z. (2006). Is conscious perception gradual or dichotomous? A comparison of report methodologies during a visual task. Consciousness and Cognition, 15(4), 700–708. http://dx.doi.org/10.1016/j.concog.2006.04.002. Overgaard, M., & Sandberg, K. (2012). Kinds of access: Different methods for report reveal different kinds of metacognitive access. Philosophical transactions of the Royal Society of London. Series B, Biological sciences, 367(1594), 1287–1296. http://dx.doi.org/10.1098/rstb.2011.0425. Persaud, N., McLeod, P., & Cowey, A. (2007). Post-decision wagering objectively measures awareness. Nature Neuroscience, 10(2), 257–261. http://dx.doi.org/ 10.1038/nn1840. Ramsøy, T. Z., & Overgaard, M. (2004). Introspection and subliminal perception. Phenomenology and the Cognitive Sciences, 3(1), 1–23. http://dx.doi.org/ 10.1023/B:PHEN.0000041900.30172.e8. Sandberg, K., Bibby, B. M., Timmermans, B., Cleeremans, A., & Overgaard, M. (2011). Measuring consciousness: Task accuracy and awareness as sigmoid functions of stimulus duration. Consciousness and Cognition, 20(4), 1659–1675. http://dx.doi.org/10.1016/j.concog.2011.09.002. Sandberg, K., Timmermans, B., Overgaard, M., & Cleeremans, A. (2010). Measuring consciousness: Is one measure better than the other? Consciousness and Cognition, 19(4), 1069–1078. http://dx.doi.org/10.1016/j.concog.2009.12.013. Schurger, A., & Sher, S. (2008). Awareness, loss aversion, and post-decision wagering. Trends in Cognitive Sciences, 12(6), 209–210 (author reply 210. doi: http://dx.doi.org/10.1016/j.tics.2008.02.012). Schwiedrzik, C. M., Singer, W., & Melloni, L. (2011). Subjective and objective learning effects dissociate in space and in time. Proceedings of the National Academy of Sciences, 108(11), 4506–4511. http://dx.doi.org/10.1073/pnas.1009147108. Seth, A. K. (2008). Post-decision wagering measures metacognitive content, not sensory consciousness. Consciousness and Cognition, 17(3), 981–983. http:// dx.doi.org/10.1016/j.concog.2007.05.008. Szczepanowski, R., Traczyk, J., Wierzchoń, M., & Cleeremans, A. (2013). The perception of visual emotion: Comparing different measures of awareness. Consciousness and Cognition, 22(1), 212–220. http://dx.doi.org/10.1016/j.concog.2012.12.003. Timmermans, B., Sandberg, K., Cleeremans, A., & Overgaard, M. (2010). Partial awareness distinguishes between measuring conscious perception and conscious content: Reply to Dienes and Seth. Consciousness and Cognition, 19(4), 1081–1083. http://dx.doi.org/10.1016/j.concog.2010.05.006. Wierzchoń, M., Asanowicz, D., Paulewicz, B., & Cleeremans, A. (2012). Subjective measures of consciousness in artificial grammar learning task. Consciousness and Cognition, 21(3), 1141–1153. http://dx.doi.org/10.1016/j.concog.2012.05.012.

Consciousness and Cognition 19 (2010) 816–828 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog ROC in animals: Uncovering the neural substrates of recollection and familiarity in episodic recognition memory q Magdalena M. Sauvage * Functional Architecture of Memory unit (www.rub.de/fam), Mercator Research Group, Faculty of Medecine, Ruhr University Bochum, Universitätsstraße 150, 44 801 Bochum, Germany Center for Memory and Brain, Boston University, 2 Cummington St., Boston, MA 02215, USA a r t i c l e i n f o Article history: Available online 5 August 2010 Keywords: ROC Recollection Familiarity Episodic memory Hippocampus MEC a b s t r a c t It is a consensus that familiarity and recollection contribute to episodic recognition memory. However, it remains controversial whether familiarity and recollection are qualitatively distinct processes supported by different brain regions, or whether they reflect different strengths of the same process and share the same support. In this review, I discuss how adapting standard human recognition memory paradigms to rats, performing circumscribed brain lesions and using receiver operating characteristic (ROC) methods contributed to solve this controversy. First, I describe the validation of the animal ROC paradigms and report evidence that familiarity and recollection are distinct processes in intact rats. Second, I report results from rats with hippocampal dysfunction which confirm this finding and lead to the conclusion that the hippocampus supports recollection but not familiarity. Finally, I describe a recent study focusing on the medial entorhinal cortex (MEC) that investigates the contribution of areas upstream of the hippocampus to recollection and familiarity. Ó 2010 Elsevier Inc. All rights reserved. 1. Introduction The contribution of two memory types to recognition memory function has been discussed since the time of Aristotle. One of these processes is described as a vague feeling of familiarity or ‘reminiscence’; for example when you recognize somebody, but cannot identify this person by name. The second, the recollection process, involves additional dimensions of memory; for example the spatial or temporal context in which this person was encountered (for a review see Yonelinas (2002)). Interestingly, the recollection and the familiarity processes are differentially affected in aging and in patients with amnesia. Indeed, the recollection process is strongly impaired in aging and amnesia, while the familiarity process is relatively spared (Barbeau et al., 2005; Brandt, Gardiner, Vargha-Khadem, Baddeley, & Mishkin, 2008; Daselaar, Fleck, Dobbins, Madden, & Cabeza, 2006; Duverne, Habibi, & Rugg, 2008; Düzel, Vargha-Khadem, Heinze, & Mishkin, 2001; Howard, Bessette-Symons, Zhang, & Hoyer, 2006; Peters & Daum, 2008; Prull, Dawes, Martin, Rosenberg, & Light, 2006; Quamme, Yonelinas, Widaman, Kroll, & Sauve, 2004; Turriziani, Serra, Fadda, Caltagirone, & Carlesimo, 2008; Vann et al., 2009; Yonelinas, Kroll, Dobbins, Lazzara, & Knight, 1998; but see Knowlton & Squire, 1995; Schacter, Verfaellie, & Anes, 1997). Hence, uncovering the neural substrates of the recollection and the familiarity processes could contribute to the characterization of new targets to rescue at least part of these deficits. A standard method to analyze human recognition memory performance is to use receiver q This article is part of a special issue of this journal on Self, Other and Memory. * Functional Architecture of Memory unit (www.rub.de/fam), Mercator Research Group, Faculty of Medecine, Ruhr University Bochum, Universitätsstraße 150, 44 801 Bochum, Germany. Fax: +49 (0) 23432 14504. E-mail address: magdalena.sauvage@rub.de 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.06.023 817 M.M. Sauvage / Consciousness and Cognition 19 (2010) 816–828 operating characteristic (ROC) functions (for a review see Yonelinas and Parks (2007)). In humans, episodic recognition memory is usually assessed by presenting a study list of items to a subject (for example a list of words appearing on a screen one at a time), and after a delay, presenting a longer list of items composed of the same items intermixed with an equal number of new items, also appearing one at a time. The probability of correct recognition of a study list item (p‘hit’) is plotted as a function of the probability of incorrect recognition of a ‘new’ item (p‘false alarm’) across confidence or bias levels, and the best fitting curve is defined to generate an ROC function (Yonelinas, 1999; see Fig. 1A and C for idealized human ROC curves for item and associative recognition memory, respectively). B 1 Familiarity Index (F) 0.6 P(Hit) 0.8 l ve ce le an ch 0.4 F 0.2 Recollection Index (R) 0.6 F 0.4 R Probability Estimates P(Hit) 0.8 1 Probability Estimates A 0.2 0 0 0.2 0.4 0.6 0.8 1 0 0.2 D 1 0.6 0.8 1 1 0.8 0.6 0.6 P(Hit) 0.8 F 0.4 R Probability Estimates F Probability Estimates 0.4 0.2 0.4 P(Fa) P(Fa) P(Hit) 0.2 0 0 C R 0.4 0.2 R 0.4 0.2 0 0 0 0 0.2 0.4 0.6 0.8 1 P(Fa) E 0 0.2 0.4 0.6 0.8 1 P(Fa) 1 P(Hit) 0.8 0.6 F Probability Estimates 0.4 0.2 R 0.4 0.2 0 0 0 0.2 0.4 0.6 0.8 1 P(Fa) Fig. 1. ROC functions for humans and rats recognition memory. Recollection (R) and familiarity (F) indices are shown as bar diagrams (see Yonelinas & Parks, 2007 for detailed calculations). (A) Ideal human ROC curve for single item recognition memory: the ROC curve is asymmetrical and curvilinear, reflecting the contribution of recollection and familiarity to recognition memory. (B) Similar ROC function observed in rats for single odor recognition memory (graph from Fortin et al., 2004). (C) Ideal human ROC curve for associative recognition memory: the ROC is asymmetrical and linear suggesting that recognition performance is based essentially on recollection. (D) Comparable ROC function observed in rats (graph from Sauvage, Fortin, Owens, Yonelinas, & Eichenbaum, 2008). (E) ROC function observed in rats for single odor recognition with a speeded response deadline demand (data from Sauvage, Beer, & Eichenbaum, 2010). The ROC is symmetrical and curvilinear, suggesting recognition relies primarily on familiarity. 818 M.M. Sauvage / Consciousness and Cognition 19 (2010) 816–828 Multiple models of recognition memory are based on the contribution of the familiarity and the recollection processes to recognition memory function (for a review see Yonelinas (2002)). Among them, two have been used extensively to study episodic recognition memory in humans: the dual-process model and the one-process model (for reviews see Wixted (2007) and Yonelinas and Parks (2007) respectively). The dual-process model describes familiarity and recollection processes as qualitatively distinct processes. In light of this model, the familiarity process is described as a rapid and continuous signal detection process sensitive to perceptual manipulations, while the recollection process is described as a slower and conceptually driven threshold process (Atkinson & Juola, 1973, 1974; Mandler, 1980; see for a review Yonelinas (2002)). Studies using the dual-process model of recognition memory also report that familiarity and recollection have different neural substrates: the hippocampus for recollection, the parahippocampal region for familiarity (Aggleton et al., 2005; Bowles et al., 2007; Yonelinas et al., 2002, 2007; see for reviews Diana, Yonelinas, and Ranganath (2007) and Eichenbaum, Yonelinas, and Ranganath (2007)). According to this model, two distinct indices can be generated from the analysis of ROC functions. The recollection index (R), y-intercept of the ROC function, which reflects the contribution of the recollection process to recognition memory performance, and the familiarity index (F), reflected by the degree of curvilinearity of the function, which reveals the contribution of familiarity to recognition memory performance (see Yonelinas, 1994; Yonelinas & Parks, 2007 for calculation of the indices). Hence, within the frame of the dual-process model, ROC functions are asymmetrical and curvilinear when familiarity and recollection contribute to recognition memory performance (Fig. 1A), asymmetrical and linear when the recollection process is primarily involved (Fig. 1C), or symmetrical and curvilinear when recognition memory is essentially based on the familiarity process (Fig. 5; Aged group). A major alternative to this model is the one-process theory that describes familiarity and recollection as qualitatively similar processes differing only in the strength of memory they reflect (see for a review Wixted (2007)). Familiarity would reflect weak memory, while recollection would reflect stronger memory or memory involving more information. As a consequence, recollection and familiarity would have a unique neural substrate, the hippocampus and could not vary independently (Manns, Hopkins, Reed, Kitchener, & Squire, 2003; Stark & Squire, 2000; Wais, Wixted, Hopkins, & Squire, 2006; see for a review Squire, Stark, and Clark (2004)). According to this view, the degree of curvilinearity of the ROC function reflects the sum of the strengths of memory components, and its asymmetry reflects greater variability in strength for the ‘old’ than for ‘new’ items. Among human ROC studies, some report that damage restricted to the hippocampus impairs specifically the recollection process (Aggleton et al., 2005; Quamme et al., 2004; Yonelinas et al., 2002), while others report that both processes are affected in patients with damage thought to be circumscribed to the hippocampus (Manns et al., 2003; Reed & Squire, 1997; Stark & Squire, 2000; Wais et al., 2006). This discrepancy resides mainly in the fact that the hippocampus and the parahippocampal region are adjacent brain structures. Indeed, identifying the precise extent of brain damage in patients with amnesia, or the precise source of brain activity within adjacent brain regions is beyond the spatial and neuropsychological resolution of standard techniques used currently in humans (e.g. functional and structural MRI imaging, and psychological tests). Thus, it has been suggested that the familiarity impairments accompanying the recollection deficits reported in some studies in patients with damage to the hippocampus resulted from additional damage to areas adjacent to the hippocampus (e.g. the parahippocampal region), rather than from damage to the hippocampus per se. Given the medial temporal lobe is exceptionally conserved across species (Manns & Eichenbaum, 2006), one way to clearly define whether the hippocampus supports familiarity as well as recollection is to perform lesion restricted to the hippocampus in animals, and assess the effect of this circumscribed lesion with behavioral paradigms that allow for the generation of distinct recollection and familiarity indices. We discuss this approach in the present review. A second major issue in human recognition memory is whether the parahippocampal region is functionally segregated in terms of its contribution to the recollection and the familiarity processes. Indeed, recent human and animal studies suggest that specific areas of the parahippocampal region, which are adjacent and strongly interconnected, contribute to different aspects of memory function (see Eichenbaum et al. (2007) for a review). The perirhinal cortex (PRc) and the lateral entorhinal cortex (LEC) would process information about the familiarity of individual items. In contrast, the postrhinal cortex (POR; parahippocampal cortex in humans) and the medial entorhinal cortex (MEC) would support recollection by representing the spatial and temporal context (whether the items are new or old) in which items have been experienced. However, this hypothesis could not be thoroughly tested in humans principally because of two reasons. One: it is not possible to determine the precise source of brain activity during recognition memory tasks in humans when areas are adjacent. Two: because cases showing restricted lesions to a single area of the parahippocampal region are extremely rare. In this review I will show how we addressed these two controversial issues by developing behavioral animal ROC paradigms that allow for recognition memory performance to be evaluated in a similar manner to the way that it is in humans (e.g. translational paradigms). Moreover, given these tasks are performed with animals, they present a key advantage over human ROC studies in that they can be combined with stereotactic surgery, which allows for brain areas to be damaged in a very restricted manner, while preventing additional damage to the adjacent brain structures. Using this approach, we first aimed at investigating the contribution of the hippocampus to recollection and familiarity by performing lesions circumscribed to the hippocampus, and defined whether the hippocampus supports recollection and familiarity, or recollection only. Second, we investigated the contribution of areas upstream of the hippocampus, more specifically the contribution of the MEC, to recollection and familiarity by performing lesion circumscribed to the MEC. 819 M.M. Sauvage / Consciousness and Cognition 19 (2010) 816–828 In the first part of this review, I will discuss how animal ROC paradigms contributed to bridging human and animal recognition memory, and brought evidence that recollection and familiarity are qualitatively distinct processes using intact rats. In a second part, I will show that this finding was confirmed in rats with impaired hippocampal function, and will show that animal ROC paradigms brought compelling evidence that the hippocampus supports recollection but not familiarity. In the last part of the review, I will report the first step of a series of studies investigating the functional segregation of the parahippocampal region, which focuses on the characterization of the contribution of the MEC to recollection and familiarity. 2. A translational model of episodic recognition memory A prerequisite for animal ROC paradigms to be appropriate translational models is to yield results comparable to those observed in humans, e.g. the contribution of familiarity and recollection to recognition memory performance in rats would be expected to be comparable to that of humans under the same experimental conditions. Within the frame of the dual-process model, a typical standard human ROC function for single item recognition memory reflects the contribution of the recollection and the familiarity processes, as shown by the asymmetry of the ROC function A Study list: 10 odors presented one at a time 1 3 2 , coriander , sage 10 Test list: 20 odors (10 old, 10 new) 30 min delay , ... paprika lemon cumin 1 3 11 , paprika 'old' 'new' 20 , ... , coriander coffee 'old' 'new' B (2) The odor is *new*: dig and retrieve reward C Bias levels Cup sizes (1) The odor is *old*: repress (1) digging and retrieve reward in the back (2) Reward for 'new' odors (buried in the cup for a correct response) Reward for 'old' odors (at the back of the cage for a correct response) 5 4 3 2 1 Fig. 2. Example of ROC paradigm in rats. (A) Sequence of odor presentations on the sample and test phases (Fortin et al., 2004). (B) Non-matching-tosample rule. (C) Bias levels obtained by varying cup heights and payoff ratios of cereal rewards. 820 M.M. Sauvage / Consciousness and Cognition 19 (2010) 816–828 (y-intercept: R different from 0) and its curvilinearity (F different from 0) respectively (Fig. 1A; Yonelinas, 1994; see for a review Yonelinas and Parks (2007)). In 2004, Fortin and colleagues laid out the ground work for ROC studies in animals by readily adapting a standard human episodic recognition memory task to rats using their innate ability to discriminate odors and to forage (Fortin, Wright, & Eichenbaum, 2004). The animal behavioral ROC paradigm was designed to be as similar as possible to the human paradigm to minimize the potential effect of methodological differences on the interpretation of the data. In human and rat ROC paradigms, items are presented during a study phase, and after a delay the same items are presented again intermixed with new items. Correct recognition of the items presented during the study phase (a hit) and incorrect recognition of new items (a false alarm; fa) are assessed across five decision criteria, and the probability of a hit (phit) is plotted as a function of the probability of false alarm (pfa). Subsequently, the exact same analytical methods are used to evaluate the shape of human or rat ROC functions, and to generate recollection and familiarity indices (see Yonelinas, 1999 and Yonelinas & Parks, 2007 for details). Understandably though, stimulus modalities differed as they were defined to yield an optimal performance for each species. In humans, stimuli used are usually visual, while in rats stimuli used are olfactory. In addition, human subjects reported that they recognized a stimulus item as ‘old’ or ‘new’ verbally or by pressing appropriate keyboard keys, while rats were trained on a non-matching to sample rule to ‘show’ that they recognized the test stimuli as being ‘new’ or ‘old’. In more detail, the animal protocol was performed as follows: every day, a unique study list of 10 odors mixed in sand contained in cups would be presented to a given animal. Odors were common household spices (cumin, coriander etc.) chosen from a pool of 40 odors. Only 20 odors were used per session (10 ‘study’ odors, 10 ‘new’ odors; one session a day), and the study list changed every day because episodic recognition memory was investigated. Sampling of the odors during the study phase was ensured by placing a small piece of cereal buried in the cup. After a 30 min delay, recognition memory was tested as animals were presented with the same odors (‘old’ odors) intermixed with additional odors that had not been presented that day (‘new’ odors), but which the animal was highly familiar with (Fig. 2A). Animals were trained on a delay non-matching to sample rule, following which they had to dig in the cup to retrieve a food reward if the odor presented was ‘new’, or repress digging if the odors were ‘old’ and go to the back of the cage to collect a food reward (Fig. 2B). To prevent that rats solve the task by smelling the presence of the reward buried in the cups containing the ‘new’ odors, cups containing ‘old’ odors were baited with food rewards that were not accessible to the animals. In addition, spatial information could not be used to solve the task because all stimulus cups were presented at the same location (in the front of the cage), and the reward location was experienced only after each trial ended, e.g. after the recognition judgment was completed. Once animals were trained on the delayed non-matching to sample rule, recognition performance was assessed across five bias levels, ranging from conservative to liberal, by manipulating the amount of reward and the cup sizes (each bias type once a week in a pseudorandom order; Fig. 2C). Recognition memory performance was assessed by collecting the number of hits and the number of false alarms (fa) over 20 trials per session. Subsequently, the probability of hits and fas were calculated, and data averaged over four sessions for each bias level to plot and generate ROC functions for odor recognition memory (Fortin et al., 2004). In this study, Fortin and colleagues showed for the first time that the recollection and the familiarity processes contributed to single item recognition memory in rats, as it is the case in humans (Yonelinas, 1994). Indeed, performance in rats resulted in an asymmetrical ROC function, reflecting the contribution of the recollection process. In addition, the ROC function for odor recognition memory in rats was curvilinear, reflecting the contribution of the familiarity process (Fig. 1B). This finding brought the first concrete evidence that analyzing single item recognition memory performance with ROC methods yielded comparable results in humans and rats with familiarity and recollection contributing to single item recognition memory, thereby suggesting that the animal ROC paradigm for odor recognition was an appropriate translational tool to study single item recognition memory. With this in mind, we pursued the validation of the animal ROC paradigms by testing whether rat ROC functions varied in a similar manner to human ROC functions when subjected to the same experimental conditions. In addition, this manipulation allowed us to test the hypothesis that recollection and familiarity were qualitatively different processes. 3. Are recollection and familiarity distinct processes? evidence from intact rats The first controversial issue that we addressed using animal ROC paradigms was whether recollection and familiarity were qualitatively distinct processes, or whether they reflected different strengths of the same process. Compelling evidence that these processes are distinct would be that one process could support recognition memory performance without a significant contribution of the other. To test this hypothesis, we investigated recognition memory performance of intact rats under memory demands reported to favor the contribution of the recollection process or the contribution of the familiarity process to recognition memory. 3.1. Can recognition memory performance solely rely on the recollection process? In humans, episodic recognition memory for pairs relies primarily on the recollection process, as revealed by an ROC function which is asymmetrical, reflecting a strong contribution of the recollection process, and linear, which reflects the lack of contribution of the familiarity process to recognition performance (Fig. 1C; see Yonelinas, 2002 for a review). Although linear ROC functions have been reported in a number of independent studies focusing on associative recognition memory (Arndt & Reder, 2002; Kelley & Wixted, 2001; Rotello, Macmillan, & Van Tassel, 2000; Slotnick, Klein, Dodson, & Shimamura, 2000), M.M. Sauvage / Consciousness and Cognition 19 (2010) 816–828 821 the finding of linear ROCs remained controversial because they have been rarely reported within the frame of the one model process (Wais et al., 2006). To determine whether the main process contributing to associative recognition memory in rats is the recollection process, we examined recognition memory performance of rats in an associative ROC paradigm involving the recognition of odors paired with media (Sauvage et al., 2008). Given that rats can separately attend to odors and media as distinct dimension when presented as an odor-medium pair (Birrell & Brown, 2000), we adapted the ROC protocol described in the previous section (Fortin et al., 2004) by presenting odor-medium pairs instead of simple odors. For example, among the 10 study pairs, cumin was mixed with beads and thyme with cotton balls. During the recognition phase, the same study pairs (old pairs) would be presented intermixed with ‘new’ rearranged odor-medium pairs (for example: cumin with cotton balls, thyme with beads). Using model-independent parameters (polynomial and linear regression functions) to assess the shape of the ROC function, and subsequently analyzing the data within the frame of the dual-process model, we reported that the rat ROC function for associative recognition memory was asymmetrical and linear, reflecting a recollection-based performance with no significant contribution of the familiarity process, as observed in humans (Fig. 1D; Sauvage et al., 2008). This result clearly showed that recognition memory performance could be achieved without a significant contribution of the familiarity process, supporting the claim that recollection and familiarity are qualitatively distinct processes. Note that partisans of the one-process model suggested that the linearity of the ROC function for associative recognition memory stemmed from the use of differential reward payoffs to manipulate response biases, rather than from memory demands (Wixted & Squire, 2008). Indeed, Wixted and colleague suggested that using different reward payoffs led to a ‘differential outcomes effect’, which in turn, was responsible for the linearity of the ROC function. A differential outcomes effect occurs when animals learn faster (as measured by an increase in performance accuracy), after hundreds of repetitions, stimulus–response reward combinations that yield a large reward than those that yield a smaller one. There are multiple reasons why this effect is not relevant to our protocol (see Eichenbaum, Sauvage, Fortin, & Yonelinas, 2008 for details). First, a simple visual inspection of the rat ROC function for single odor recognition memory reveals that it is possible to obtain a curvilinear ROC function using the exact same differential reward payoffs but different memory demands (compare the ROC for item recognition: Fig. 1B to the ROC for associative recognition: Fig. 1D). This clearly suggests that the shape of ROC functions is tied to memory demands and not to the use of differential payoff rewards. Second, the ‘differential outcomes effect’ requires hundreds of stimulus–response reward pairings for the preferential learning to take place. However, in our paradigm, stimuli are showed only once and the reward is received only after the recognition judgment is completed. Hence, there is no possibility for the animal to predict the amount of reward that will be received, and this to affect its performance. Last but not least, opposite to the predictions of the ‘differential outcomes effect’, accuracy is actually the lowest for the bias which yields the largest difference between rewards for the ‘new’ stimulus and the ‘old’ stimuli (Bias 1: 1=4 compared to 3 froot loops), while it is the highest for the bias for which there is NO difference between rewards for ‘old’ and ‘new’ stimuli (Bias 5: 1/2 froot loop in both cases). In conclusion, the differential outcomes effect is clearly not at work in our study, and the linearity of the rat ROC function for associative recognition memory is tied to memory demands and not to the use of differential payoff rewards. In the present study, we have reported that the ROC function for associative recognition in rats is linear, reflecting a strong contribution of the recollection process to the memory for pairs, without significant contribution of the familiarity process, which suggests that the recollection and the familiarity processes are qualitatively distinct processes. Moreover, given that linear ROC functions for associative recognition memory have been reported in the human literature (Arndt & Reder, 2002; Rotello et al., 2000; Slotnick et al., 2000; Yonelinas, 1997), even by the principal detractor of the dual-process model (Kelley & Wixted, 2001; see ROC functions for associative recognition with rearranged pairs), obtaining comparable ROC functions for associative recognition memory in rats constitutes a second step in the validation of animal ROC paradigms as proper translational tools to investigate the contribution of recollection and familiarity to recognition memory. In the next section, we focused on the missing piece of the puzzle, and studied whether recognition memory performance in rats could be achieved on the basis of the familiarity process only. 3.2. Can recognition memory performance solely rely on the familiarity process? A counterpart to the previous experiment was to examine recognition memory performance under conditions that favor the familiarity process. Familiarity is usually described as a rapid process that is based on pattern matching, and is sensitive to perceptual modulations, whereas recollection is characterized as a slower and conceptually driven process (see Mandler (2008) and Yonelinas (2002) for reviews). Indeed, studies involving process dissociation procedures (Yonelinas & Jacoby, 1994), response deadlines (Hintzman, Caulton, & Levitin, 1998; Hintzman & Curran, 1994; McElree, Dolan, & Jacoby, 1999) and evoked response potentials or electroencephalograms (Curran, 2004; Duarte, Ranganath, Trujillo, & Knight, 2006; Duzel, Yonelinas, Mangun, Heinzem, & Tulving, 1997; Smith, 1993; Woodruff, Hayama, & Rugg, 2006) reveal a two-component temporal function that includes a rapid familiarity process and a slower recollective process. Consistent with this view, limiting appropriately the latency to respond in a recognition memory task should allow for the familiarity process to be fully completed but prevent a significant contribution of the recollection process to recognition memory performance. To test this hypothesis, we first assessed odor recognition memory performance in rats without a deadline, and obtained a typical ROC function for odor recognition memory: asymmetrical and curvilinear, reflecting the contribution of the recollection and the familiarity processes to recognition memory performance. However, when a deadline was subsequently applied, by giving the animals half the time to respond, the ROC function remained curvilinear, reflecting the contribution of the familiarity process to recognition memory, 822 M.M. Sauvage / Consciousness and Cognition 19 (2010) 816–828 but became symmetrical (y-intercept not different from zero) suggesting that the recollection process did not contribute to recognition memory performance in a significant manner (Fig. 1E; Sauvage et al., 2010). These results suggested that under speeded conditions recognition memory performance relies essentially on the familiarity process, while the contribution of recollection is negligible, giving further support to the claim that recollection and familiarity are qualitatively distinct processes. Furthermore, this last experiment completed the validation of animal ROC paradigms as appropriate tools to study the contribution of familiarity and recollection to recognition memory in animals, given the animal ROC function under speeded conditions mirrored results obtained in the human literature under comparable conditions. In conclusion, in this part of the review, we reported a double dissociation of the contribution of familiarity and recollection to episodic recognition memory in intact rats. This finding reveals that the contribution of each process depends on memory demands, and strongly supports the hypothesis that recollection and familiarity are qualitatively distinct processes. Moreover, we validated animal ROC paradigms as appropriate translational tools to investigate the contribution of recollection and familiarity to recognition memory, given that animal and human ROC studies yielded comparable results. In the following sections, we report findings that emerged from animal studies using ROC paradigms combined with lesions restricted to target areas to address issues that remained controversial in humans because of the difficulty of defining the extent of brain damage in patients with amnesia, or because of the impossibility of dissociating the precise source of activity in adjacent brain areas in healthy subjects. 4. Does the hippocampus support the recollection and the familiarity processes? evidence from rats with hippocampal dysfunction A major controversy in recognition memory is whether the hippocampus supports the recollection process only or whether it also supports the familiarity process. Previous studies report a predominant role for the hippocampus in the recollection process while the parahippocampal region is suggested to primarily contribute to the familiarity process (Brown & Aggleton, 2001; Eichenbaum, Otto, & Cohen, 1994). However, conflicting results emerged from the ROC literature in humans. Evidence from studies on amnesic patients with damage restricted to the hippocampus suggests that the hippocampus specifically contributes to the recollection process but not to the familiarity process. Indeed, the y-intercept of the ROC function of patients with amnesia is significantly reduced compared to healthy subjects, while the curvilinear shape of the ROC function is maintained (Aggleton et al., 2005; Turriziani et al., 2008; Yonelinas et al., 1998, 2002; see for a review Eichenbaum et al. (2007)). In addition, studies focusing on specific areas of the parahippocampal region, the perirhinal cortex (PRc) and the entorhinal cortex (EC), report a preponderant role of these regions in the familiarity process (Bowles et al., 2007; Haskins, Yonelinas, Quamme, & Ranganath, 2008; Yonelinas et al., 2007; for reviews see Diana, Yonelinas, and Ranganath (2010) and Diana et al. (2007)). In striking contrast, other human studies report that both the recollection and the familiarity processes are affected following damage thought to be restricted to the hippocampus (Manns et al., 2003; see for a review Wixted and Squire (2004)), reflected by an alteration in both the asymmetry and the curvilinear shape of the ROC function (Wais et al., 2006). In addition, studies from the same group suggest that activity in the hippocampus and the PRc is correlated to the memory strength of remembered items, rather than to the specific contribution of these brain areas to the recollection or the familiarity processes to recognition memory (Shrager, Kirwan, & Squire, 2008; for a review see Squire, Wixted, and Clark (2007)). These findings left the field of human recognition memory divided regarding the contribution of the hippocampus to the recollection and the familiarity processes. The controversy principally stems from the fact that using standard functional or structural MRI imaging techniques, it is impossible to determine with precision whether brain activity (or brain damage) is truly circumscribed to the hippocampus, or whether it also extends to areas adjacent to the hippocampus (the parahippocampal region). A clear advantage of animal studies over human studies is that brain damage can be generated in a very controlled and restricted manner using stereotactic surgery techniques. Hence, as a first step to study the contribution of the hippocampus to recollection and familiarity, we performed lesions unequivocally circumscribed to the hippocampus, and assessed the effect of this selective lesion on the contribution of the recollection and the familiarity processes to item and associative recognition memory using animal ROC paradigms. In addition, we report later animal ROC findings related to the ethological model of reduced hippocampal function that is aging. Of note, data from these experiments have been analyzed with model-independent parameters (linear and polynomial regressions) to evaluate the shape of the ROC functions, with the dual-process model and with the one-process model (see Fortin et al., 2004; Robitsek, Fortin, Koh, Gallagher, & Eichenbaum, 2008; Sauvage et al., 2008 for details). In each experiment, analysis with the one-process model confirmed the dual-process findings by revealing that reducing hippocampal function selectively eliminates one parameter of the ROC function, specifically the inequality of variances between ‘old’ and ‘new’ items. By contrast, no change was observed in the other parameter of the ROC function, d0 , which reflects the difference in memory strength between ‘old’ and ‘new’ items. In summary, analysis with the one-process model confirmed that hippocampal lesions and aging selectively eliminated one parameter supporting the identification of old items, which is comparable to altering the recollection process in the dual-process theory, and left intact another parameter comparable to familiarity. 4.1. Contribution of the hippocampus to recollection and familiarity in item recognition memory? In 2004, Fortin and colleagues performed lesions restricted to the hippocampus in rats and reported that the ROC function of hippocampal-damaged rats remained curvilinear, reflecting the contribution of the familiarity process to recognition 823 M.M. Sauvage / Consciousness and Cognition 19 (2010) 816–828 A B 1 Sham Hippocampus 0.6 F 0.2 0 0.2 0.4 F 0.2 0 S 0 R 0.4 Probability Estimates 0.4 Probability Estimates P (Hit) 0.8 0.6 P(Fa) H S 0.8 H 1 0 0.2 0.4 R 0.4 0.2 0 0.6 0.8 1 P(Fa) Fig. 3. Item recognition ROC in rats with circumscribed hippocampal lesion (graph from Fortin et al., 2004). (A) Hippocampal lesion eliminates recollectionbased performance while sparing the familiarity process. (B) ROC function in sham rats with extended memory delay. Extending the delay significantly affects familiarity not recollection. memory. In addition, the ROC function became fully symmetrical suggesting that the recollection process did not contribute significantly to recognition memory performance of hippocampal-lesioned rats (Fig. 3A, Hippocampus). This finding was the first evidence that damage unequivocally restricted to the hippocampus significantly impaired the recollection process without affecting the familiarity process, which suggested that the hippocampus supports recollection but not familiarity. Of note, opponents of the dual-process model argued that hippocampal lesions did not specifically affect the recollection process but rather reduced overall memory performance because the accuracy level of the hippocampus group was slightly but significantly reduced compared to that of sham rats. To test this hypothesis, Fortin and colleagues lowered the overall memory performance of sham rats by increasing the delay between study and recognition phases (Yonelinas et al., 2002), and studied the effect of this manipulation on the contribution of recollection and familiarity to recognition memory. Against the predictions of the one-process model, the ROC function of sham rats became linear and asymmetrical (Fig. 3B) instead of curvilinear and symmetrical as observed following hippocampal lesion (Fig. 3A, Hippocampus). This manipulation clearly stated that reducing overall recognition memory performance affected the familiarity process and not the recollection process, and therefore could not account for the recollection deficit observed after hippocampal lesion. In summary, this study showed for the first time that the hippocampus supports recollection and not familiarity, given that lesions were specifically restricted to the hippocampus and only recollection was affected. Moreover, these data confirmed the results obtained in intact rats suggesting that recollection and familiarity are qualitatively distinct processes since the recollection and the familiarity processes were not affected in a similar manner following hippocampal lesion. 4.2. Contribution of the hippocampus to recollection and familiarity in associative recognition memory? To study further the contribution of the hippocampus to the familiarity and the recollection processes, we studied recognition memory performance for pairs using the associative ROC paradigm described in Sauvage et al., 2008. Rats were presented with a study list of odors paired with media, and recognition memory performance was subsequently assessed by presenting the same pairs intermixed with rearranged odor-medium pairs. As previously described, the ROC function for associative recognition memory was asymmetrical and linear, reflecting a recollection-based memory performance with no significant contribution of the familiarity process (Fig. 4, Sham). In striking contrast, recollection was significantly impaired following hippocampal lesion as reflected by a significant drop of the y-intercept (index of recollection, R), suggesting that the hippocampus is critical for recollection in associative recognition memory. Even more interesting, the ROC function of hippocampal- lesioned rats was now curvilinear reflecting a significant contribution of the familiarity process to recognition memory following hippocampal damage (Fig. 4, Hippocampus). This finding was the first clear evidence in the human and animal literature that the hippocampus did not support the familiarity process since the contribution of the familiarity process to recognition memory increased following hippocampal damage. An alternative interpretation of these data, in line with the predictions of the one-process model, was recently voiced. It has been argued that if the hippocampus supports familiarity and recollection, and if recollection judgements are ‘more demanding’ than familiarity judgments, and the lesion of the hippocampus is only partial, it could be possible that hippocampal lesions significantly affect the recollection process, but that enough of the hippocampus remains to complete the ‘less demanding’ familiarity judgments. As much as this could theoretically be the case in humans, for example in mild cases of stroke (assuming that damage remains in the vicinity of blood vessels or brain arteries, and potentially leaves part of the hippocampus and its connections intact), it is highly improbable in our animal studies given the type of lesion we perform. Indeed, our surgery aims at lesioning the hippocampus 824 M.M. Sauvage / Consciousness and Cognition 19 (2010) 816–828 1 P(Hit) 0.8 0.6 Sham Hippocampus F Probability Estimates 0.4 0.2 R 0.4 0.2 0 S H S H 0 0 0.2 0.4 0.6 0.8 1 P(Fa) Fig. 4. Associative recognition ROC of rats with restricted hippocampal lesion (graph from Sauvage et al., 2008). Hippocampal lesion affects recollection and familiarity in an opposite manner. Hippocampal damage decreases the recollection index (R: y-intercept) while increasing the familiarity index (F). 1 Young Aged 0.6 F R 0.4 Probability Estimates P(Hit) 0.8 0.2 0.4 0.2 0 Y A Y A 0 0 0.2 0.4 0.6 0.8 1 P(Fa) Fig. 5. Item recognition ROC in aging (graph from Robitsek et al., 2008). Aging reduces the contribution of recollection to recognition performance while sparing familiarity. by applying numerous small lesions (24 total; 12 per hemisphere) spread along the rostro-caudal, dorso-ventral and mediolateral axis. Hence, even if the lesion of the hippocampus is not total, the remaining hippocampal ‘bits’ are definitively disconnected and unlikely to be functional. Second, if in agreement with the single-process model, the hippocampus supports familiarity and recollection, the prediction of this model would be that hippocampal damage should also impair familiarity (even if it affects much more the recollection process), which is opposite to the results obtained in the present study, since a significant increase of the contribution of familiarity to recognition memory was reported. Finally, it is also important to underline that the specific recollection impairment observed in the Hippocampus group cannot be explained by a difference in memory strength between the Hippocampus and the Sham groups, since overall recognition memory performance did not significantly differ between the Hippocampus group and the Sham rats (Sauvage et al., 2008). Thus, the ‘memory strength hypothesis’ could not apply in the present study, and our data suggested that rats without a functional hippocampus recognized pairs to a similar level as sham rats but used an alternate strategy to solve the task. Interestingly, damage to the hippocampus in rats was reported to increase the tendency to unitize stimulus elements of a pair into a single stimulus; for example, lemon and sand could be encoded as lemon-scented sand by rats with a hippocampal lesion (Eichenbaum et al., 1994), and such a strategy was also observed in patients with amnesia. Indeed, amnesic patients who performed very poorly on memory tasks that use unrelated items as stimulus pairs, were reported to perform better on associative recognition memory tasks when given the possibility to rely more heavily on familiarity. For example, when compound words such as fireman, hardware or sleepwalk were used as stimulus pairs instead of unrelated words, and patients could consider both elements of a pair as a single item (Giovanello, Keane, & Verfaellie, 2006; Quamme, Yonelinas, & Norman, 2007; Turriziani, Fadda, Caltragirone, & Carlesimo, 2004). In summary, this study provided the first clear evidence that the hippocampus did not support familiarity, because the contribution of familiarity to recognition memory performance was enhanced following damage restricted to the hippocampus. It also revealed that the hippocampus supports the recollection process in associative recognition memory, as it is the case in item recognition memory. Furthermore, it strongly suggests that familiarity and recollection are qualitatively distinct processes because they can vary in an opposite manner. M.M. Sauvage / Consciousness and Cognition 19 (2010) 816–828 825 4.3. Contribution of the hippocampus to recollection and familiarity in aging? Finally, the finding of a selective contribution of the hippocampus to recollection and not familiarity was further supported by results from a ROC study with aged rats, used as a model of reduced hippocampal function (Daselaar et al., 2006; Rosenzweig & Barnes, 2003; Wilson, Gallagher, Eichenbaum, & Tanila, 2006). In this study, the ROC function for item recognition of memory-impaired rats was found to be symmetrical and curvilinear, as it was the case following hippocampal lesion, suggesting that reducing hippocampal function limits the contribution of recollection to recognition memory while sparing the familiarity process (Fig. 5, Robitsek et al., 2008). Interestingly, as observed in associative recognition memory, the contribution of the familiarity process to recognition memory performance was increased in a subset of the memory-impaired rats, confirming in an ethologically relevant model that the hippocampus supports recollection and not familiarity, and that the two processes are qualitatively distinct since they can vary in an opposite manner. In conclusion, data from ROC studies in rats with reduced hippocampal function (lesion and aging studies) confirmed results in intact rats by showing that the recollection and the familiarity processes are qualitatively distinct, because they can vary independently when hippocampal function is compromised. In addition, these data brought convincing evidence that the hippocampus supports recollection but not familiarity, given the contribution of the recollection process to single item and associative recognition memory is significantly reduced following hippocampal lesion, whereas the contribution of familiarity is either unaltered or enhanced. In this section, I have discussed how animal ROC paradigms contributed to characterize the role of the hippocampus in recollection and familiarity in recognition memory by studying animals with hippocampal dysfunction. In the following section, I report how animal ROC paradigms were used to identify brain structures, upstream of the hippocampus, that provide information necessary to complete recollection-based-judgments to the hippocampus. 4.4. What areas provide information required for recollection-based-judgments to the hippocampus? Here, I will address a second issue that considerably fuels the current debate in recognition memory, which is the investigation of the functional segregation of the parahippocampal region in terms of recollection and familiarity. More precisely, I report here the first step of a series of studies which aim at characterizing the specific contribution of each area of the parahippocampal region to recollection and familiarity, starting with the medial entorhinal cortex (MEC), which has recently drawn a lot of attention for its dedicated role in spatial navigation and path integration (Fyhn, Molden, Witter, Moser, & Moser, 2004; Hafting, Fyhn, Molden, Moser, & Moser, 2005, see for reviews McNaughton, Battaglia, Jensen, Moser, and Moser (2006) and Moser and Moser (2008)). The parahippocampal region includes the perirhinal cortex (PRc), the parahippocampal cortex (PHc), the lateral entorhinal cortex (LEC) and the medial entorhinal cortex (MEC), which share intricate and bidirectional projections (see for a review Van Strien, Cappaert, and Witter (2009)). Despite this complex network of connections, data emerging from the human literature suggest that some of the areas of the parahippocampal region support the recollection process, while others contribute the familiarity process (for reviews see Diana et al. (2007, 2010) and Eichenbaum et al. (2007)). According to these studies, the PRc and the LEC would process information about the familiarity of individual items, whereas the PHc and the MEC would support recollection by representing spatial and temporal contexts (whether items are old or new) in which items have been experienced. Subsequently, both the item and contextual information would be combined within the hippocampus (see for a review Lipton and Eichenbaum (2008)). Thus, human fMRI studies report that activity in the PRc is correlated to familiarity-based judgments in recognition memory tasks (Davachi, Mitchell, & Wagner, 2003; Haskins et al., 2008; Henson, Cansino, Herron, Robb, & Rugg, 2003; Montaldi, Spencer, Roberts, & Mayes, 2006; Ranganath et al., 2004; Suchan, Gayk, Schmid, Köster, & Daum, 2008). In addition, ablation of the same structure, in a patient suffering from an intractable case of epilepsy, led to deficits specifically tied to the familiarity process (Bowles et al., 2007). In contrast, evidence from imaging studies points towards a preponderant role of the PHc and the hippocampus in the recollection of single items and associations (Aminoff, Gronau, & Bar, 2007; Bar, Aminoff, & Schacter, 2008; Cansino, Maquet, Dolan, & Rugg, 2002; Eldridge, Knowlton, Furmanski, Bookheimer, & Engel, 2000; Suchan et al., 2008; Woodruff, Johnson, Uncapher, & Rugg, 2005; Yonelinas et al., 2005). The contribution of the MEC and LEC to recollection and familiarity has been essentially extrapolated from their anatomical connections to the PHc and the PRc, because it is not possible with fMRI imaging techniques to dissociate the activity that occurs in the medial part of the EC during a memory task from that occurring in the lateral part of the EC. Moreover, no patients with damage restricted to the MEC or LEC have been reported to this date. Thus, as a part of the PHc–MEC complex, the MEC is suggested to play a preponderant role in the recollection process. Conversely, the LEC would contribute more specifically to the familiarity process, as a part of the PRc–LEC complex. However, direct evidence of these selective contributions is missing. To investigate the involvement of the MEC in the recollection process, we studied the effect of damage restricted to the MEC on the contribution of recollection and familiarity to odor recognition memory (Sauvage, Beer, Ekovich, Ho, & Eichenbaum, submitted for publication). Following MEC lesions, the ROC function remained curvilinear, reflecting a strong contribution of the familiarity process (Fig. 6, MEC). However, in sharp contrast to the sham rats, the ROC function of MEC rats became symmetrical (y-intercept not different from 0), suggesting that the recollection process no longer significantly contributed to the recognition performance. These results provided the first evidence that the MEC contributes specifically to recollection-based-judgments, whereas its contribution to the familiarity process is minimal. These findings are consistent 826 M.M. Sauvage / Consciousness and Cognition 19 (2010) 816–828 1 P(Hit) 0.8 Sham 0.6 MEC 0.4 Probability estimates 0.4 0.2 0.2 0 S 0 0 0.2 0.4 M 0.6 S 0.8 M 1 P(Fa) Fig. 6. Item recognition ROC of rats with restricted lesion to the MEC (graph from Sauvage et al., submitted for publication). MEC lesion reduces the contribution of recollection to recognition memory while familiarity remained unaffected. with a broader view of the MEC function in recognition memory than its dedicated role in spatial representation, since MEC damage significantly altered performance in our non-spatial (odor) recognition memory task. In agreement with this finding, a recent electrophysiological study reported that dorsocaudal MEC (dcMEC) neurons had distinct activity patterns depending on whether rats turned left or right on a T-maze in an alternation working memory task (Lipton, White, & Eichenbaum, 2007). This finding suggested that dcMEC cells signaled which of the two episodes was ongoing, and were encoding spatial and temporal contexts (Lipton & Eichenbaum, 2008). Similar findings regarding the representation of spatial and temporal contexts were reported for the PHc, which is adjacent and strongly connected to the MEC. Indeed, the PHc was commonly found to be activated during the viewing of spatial environments, of objects that evoke strong contextual associations and during the viewing of objects that have strong temporal associations (Aminoff et al., 2007; Bar et al., 2008; Epstein & Kanwisher, 1998). Thus, the PHc which is strongly connected with the MEC was also suggested to represent the spatial and temporal context of remembered objects. In light of previous studies focusing on the selective contribution of the PRc to the familiarity process, the present study brought further support to the hypothesis of a functional segregation within the parahippocampal region, by showing that MEC is not involved in familiarity for specific items but, is a part of the parahippocampal and medial entorhinal network that is essential to memory for the context in which events occur, a defining feature of episodic recollection (Eichenbaum et al., 2007). This perspective is consistent with the observations on the dcMEC in spatial representation (grid cells), but suggests a broader role in the representation of spatial and temporal contexts which is of importance to episodic memory. In summary, results from this study suggested once more that recollection and familiarity are qualitatively distinct processes, since damage restricted to the MEC affected recollection while sparing familiarity. This study also allowed us to firmly conclude that the MEC contributes selectively to the recollection process, possibly by providing information regarding the temporal context of the remembered items to the hippocampus (whether items are ‘old ‘or ‘new’). This study is the first of a series of ROC studies in animals focusing on the investigation of the functional segregation of the parahippocampal region in terms of recollection and familiarity, and calls for further studies to characterize the contribution of other areas of the parahippocampal region, for example the LEC, for which virtually no data are available. In conclusion, developing ROC paradigms in animals contributed to bridging recognition memory in humans and animals by assessing and analyzing recognition memory performance in a comparable manner. In addition, combining circumscribed brain damage in animals to the use of animal ROC paradigms allowed for major controversies in human recognition memory to be addressed, and led consistently to the conclusion that familiarity and recollection are qualitatively distinct processes, and that the hippocampus supports recollection but not familiarity. Finally, we started to apply this approach to the characterization of the specific contribution of the parahippocampal areas to recollection and familiarity in recognition memory and are confident that this translational approach, combined to more standard approaches, will contribute to significant progress in the elucidation of the neural substrates of recognition memory. Acknowledgments We thank Jarret Frank for his help in Graphic designs and Zachery Beer for proof-reading. These experiments were Funded by MH71702, MH51520, MH52090 and AG09973. References Aggleton, J. P., Vann, S. D., Denby, C., Dix, S., Mayes, A. R., Roberts, N., et al (2005). Sparing of the familiarity component of recognition memory in a patient with hippocampal pathology. Neuropsychologia, 43(12), 1810–1823. M.M. Sauvage / Consciousness and Cognition 19 (2010) 816–828 827 Aminoff, E., Gronau, N., & Bar, M. (2007). The parahippocampal cortex mediates spatial and nonspatial associations. Cerebral Cortex, 17(7), 1493–1503. Arndt, J., & Reder, L. M. (2002). Word frequency and receiver-operating characteristic curves in recognition memory: Evidence for a dual-process interpretation. Journal of Experimental Psychology: Learning, Memory, and Cognition, 28(5), 830–842. Atkinson, R. C., & Juola, J. F. (1973). Factors influencing speed and accuracy of word recognition. In S. Kornblum (Ed.). Attention and performance (Vol. 4, pp. 583–612). New York: Academic Press. Atkinson, R. C., & Juola, J. F. (1974). Search and decision processes in recognition memory. In D. H. Krantz, R. C. Atkinson, & R. D. Luce (Eds.), Contemporary developments in mathematical psychology (pp. 243–293). San Francisco: W.H. Freeman. Bar, M., Aminoff, E., & Schacter, D. L. (2008). Scenes unseen: The parahippocampal cortex intrinsically subserves contextual associations, not scenes or places per se. Journal of Neuroscience, 28(34), 8539–8544. Barbeau, E. J., Felician, O., Joubert, S., Sontheimer, A., Ceccaldi, M., & Poncet, M. (2005). Preserved visual recognition memory in an amnesic patient with hippocampal lesions. Hippocampus, 15(5), 587–596. Birrell, J. M., & Brown, V. J. (2000). Medial frontal cortex mediates perceptual attentional set shifting in the rat. Journal of Neuroscience, 20, 4320–4324. Bowles, B., Crupi, C., Mirsattari, S. M., Pigott, S. E., Parrent, A. G., Pruessner, J. C., et al (2007). Impaired familiarity with preserved recollection after anterior temporal-lobe resection that spares the hippocampus. Proceedings of the National Academy of Sciences of the United States of America, 104(41), 16382–16387. Brandt, K. R., Gardiner, J. M., Vargha-Khadem, F., Baddeley, A. D., & Mishkin, M. (2008). Impairment of recollection but not familiarity in a case of developmental amnesia. Neurocase, 15(1), 60–65. Brown, M. W., & Aggleton, J. P. (2001). Recognition memory: What are the roles of the perirhinal cortex and hippocampus? Nature Reviews Neuroscience, 2, 51–61. Cansino, S., Maquet, P., Dolan, R. J., & Rugg, M. D. (2002). Brain activity underlying encoding and retrieval of source memory. Cerebral Cortex, 12(10), 1048–1056. Curran, T. (2004). Effects of attention and confidence on the hypothesized ERP correlates of recollection and familiarity. Neuropsychologia, 42, 1088–1106. Daselaar, S. M., Fleck, M. S., Dobbins, I. G., Madden, D. J., & Cabeza, R. (2006). Effects of healthy aging on hippocampal and rhinal memory functions: An event related fMRI study. Cerebral Cortex, 16, 1771–1782. Davachi, L., Mitchell, J. P., & Wagner, A. D. (2003). Multiple routes to memory: Distinct medial temporal lobe processes build item and source memories. Proceedings of the National Academy of Sciences of the United States of America, 100(4), 2157–2162. Diana, R. A., Yonelinas, A. P., & Ranganath, C. (2010). Medial temporal lobe activity during source retrieval reflects information type, not memory strength. Journal of Cognitive Neuroscience, 22(8), 1808–1818. Diana, R. A., Yonelinas, A. P., & Ranganath, C. (2007). Imaging recollection and familiarity in the medial temporal lobe: A three-component model. Trends in Cognitive Sciences, 11(9), 379–386. Duarte, A., Ranganath, C., Trujillo, C., & Knight, R. T. (2006). Intact recollection memory in high performing older adults: ERP and behavioral evidence. Journal of Cognitive Neuroscience, 18(1), 33–47. Duverne, S., Habibi, A., & Rugg, M. D. (2008). Regional specificity of age effects on the neural correlates of episodic retrieval. Neurobiology of Aging, 29(12), 1902–1916. Düzel, E., Vargha-Khadem, F., Heinze, H. J., & Mishkin, M. (2001). Brain activity evidence for recognition without recollection after early hippocampal damage. Proceedings of the National Academy of Sciences of the United States of America, 98(14), 8101–8106. Duzel, E., Yonelinas, A. P., Mangun, G. R., Heinzem, H. J., & Tulving, E. (1997). Event-related brain potential correlates of two states of conscious awareness in memory. Proceedings of the National Academy of Sciences of the United States of America, 94(11), 5973–5978. Eichenbaum, H., Otto, T., & Cohen, N. J. (1994). Two functional components of the hippocampal memory system. Behavioral and Brain Sciences, 17(449–472), 472–518. Eichenbaum, H., Sauvage, M., Fortin, N., & Yonelinas, A. P. (2008). ROCs in rats? Response to Wixted and Squire. Learning and Memory, 15(9), 691–693. Eichenbaum, H., Yonelinas, A. R., & Ranganath, C. (2007). The medial temporal lobe and recognition memory. Annual Review of Neuroscience, 30, 123–152. Eldridge, L. L., Knowlton, B. J., Furmanski, C. S., Bookheimer, S. Y., & Engel, S. A. (2000). Remembering episodes: A selective role for the hippocampus during retrieval. Nature Neuroscience, 3(11), 1149–1152. Epstein, R., & Kanwisher, N. (1998). A cortical representation of the local visual environment. Nature, 392(6676), 598–601. Fortin, N. J., Wright, S. P., & Eichenbaum, H. (2004). Recollection-like memory retrieval in rats is dependent on the hippocampus. Nature, 431(7005), 188–191. Fyhn, M., Molden, S., Witter, M. P., Moser, E. I., & Moser, M. B. (2004). Spatial representation in the entorhinal cortex. Science, 305(5688), 1258–1264. Giovanello, K. S., Keane, M. M., & Verfaellie, M. (2006). The contribution of familiarity to associative memory in amnesia. Neuropsychologia, 44(10), 1859–1865. Hafting, T., Fyhn, M., Molden, S., Moser, M. B., & Moser, E. I. (2005). Microstructure of a spatial map in the entorhinal cortex. Nature, 436(7052), 801–806. Haskins, A. L., Yonelinas, A. P., Quamme, J. R., & Ranganath, C. (2008). Perirhinal cortex supports encoding and familiarity-based recognition of novel associations. Neuron, 59(4), 554–560. Henson, R. N., Cansino, S., Herron, J. E., Robb, W. G., & Rugg, M. D. (2003). A familiarity signal in human anterior medial temporal cortex? Hippocampus, 13(2), 301–304. Hintzman, D. L., Caulton, D. A., & Levitin, D. J. (1998). Retrieval dynamics in recognition and list discrimination: Further evidence of separate processes of familiarity and recall. Memory & Cognition, 26(3), 449–462. Hintzman, D. L., & Curran, T. (1994). Retrieval dynamics of recognition and frequency judgments: Evidence for separate processes of familiarity and recall. Journal of Memory and Language, 331, 1–18. Howard, M. W., Bessette-Symons, B., Zhang, Y., & Hoyer, W. J. (2006). Aging selectively impairs recollection in recognition memory for pictures: Evidence from modelling and receiver operating characteristic curves. Psychology and Aging, 21, 96–106. Kelley, R., & Wixted, J. T. (2001). On the nature of associative information in recognition memory. Journal of Experimental Psychology: Learning, Memory, and Cognition, 27(3), 701–722. Knowlton, B. J., & Squire, L. R. (1995). Remembering and knowing: Two different expressions of declarative memory. Journal of Experimental Psychology: Learning, Memory, and Cognition, 21(3), 699–710. Lipton, P. A., & Eichenbaum, H. (2008). Complementary roles of hippocampus and medial entorhinal cortex in episodic memory. Neural Plasticity, 2008, 258467. Lipton, P. A., White, J. A., & Eichenbaum, H. (2007). Disambiguation of overlapping experiences by neurons in the medial entorhinal cortex. Journal of Neuroscience, 27(21), 5787–5795. Mandler, G. (1980). Recognizing: The judgment of previous occurrence. Psychological Review, 87, 252–271. Mandler, G. (2008). Familiarity breeds attempts: A critical review of dual process theories of recognition. Perspectives in Psychological Science, 3(5), 392–401. Manns, J. R., & Eichenbaum, H. (2006). Evolution of declarative memory. Hippocampus, 16, 795–808. Manns, J. R., Hopkins, R. O., Reed, J. M., Kitchener, E. G., & Squire, L. R. (2003). Recognition memory and the human hippocampus. Neuron, 37, 171–180. McElree, B., Dolan, P. O., & Jacoby, L. L. (1999). Isolating the contributions of familiarity and source information to item recognition: A time course analysis. Journal of Experimental Psychology: Learning, Memory, and Cognition, 253, 563–582. McNaughton, B. L., Battaglia, F. P., Jensen, O., Moser, E. I., & Moser, M. B. (2006). Path integration and the neural basis of the ‘cognitive map’. Nature Reviews Neuroscience, 7(8), 663–678. Montaldi, D., Spencer, T. J., Roberts, N., & Mayes, A. R. (2006). The neural system that mediates familiarity memory. Hippocampus, 16(5), 504–520. Moser, E. I., & Moser, M. B. (2008). A metric for space. Hippocampus, 18(12), 1142–1156. 828 M.M. Sauvage / Consciousness and Cognition 19 (2010) 816–828 Peters, J., & Daum, I. (2008). Differential effects of normal aging on recollection of concrete and abstract words. Neuropsychology, 22(2), 255–261. Prull, M. W., Dawes, L. L., Martin, A. M., III, Rosenberg, H. F., & Light, L. L. (2006). Recollection and familiarity in recognition memory: Adult age differences and psychology. Journal of Experimental Psychology: Human Learning and Memory, 28(6), 1499–1517. Quamme, J. R., Yonelinas, A. P., & Norman, K. A. (2007). Effect of unitization on associative recognition in amnesia. Hippocampus, 17(3), 192–200. Quamme, J. R., Yonelinas, A. P., Widaman, K. F., Kroll, N. E. A., & Sauve, M. J. (2004). Recall and recognition in mild hypoxia: Using covariance structural modeling to test competing theories of explicit memory. Neuropsychologia, 42, 672–691. Ranganath, C., Yonelinas, A. P., Cohen, M. X., Dy, C. J., Tom, S. M., & D’Esposito, M. (2004). Dissociable correlates of recollection and familiarity within the medial temporal lobes. Neuropsychologia, 42(1), 2–13. Reed, J. M., & Squire, L. R. (1997). Impaired recognition memory in patients with lesions limited to the hippocampal formation. Behavioral Neuroscience, 111, 667–675. Robitsek, R. J., Fortin, N. J., Koh, M. T., Gallagher, M., & Eichenbaum, H. (2008). Cognitive aging: A common decline of episodic recollection and spatial memory in rats. Journal of Neuroscience, 28(36), 8945–8954. Rosenzweig, E. S., & Barnes, C. A. (2003). Impact of aging on hippocampal function: Plasticity, network dynamics, and cognition. Progress in Neurobiology, 69, 143–179. Rotello, C. M., Macmillan, N. A., & Van Tassel, G. (2000). Recall-to-reject in recognition: Evidence from ROC curves. Journal of Memory and Language, 43(1), 67–88. Sauvage, M., Beer, Z., & Eichenbaum, H. (2010). Recognition memory: Adding a response deadline eliminates recollection but spares familiarity”. Learning & Memory, 17(2), 104–108. Sauvage, M., Beer, Z., Ekovich, M., Ho, L., & Eichenbaum, H. (submitted for publication) The medial entorhinal cortex: A selective role in recollection-based memory. Journal of Neuroscience. Sauvage, M., Fortin, N. J., Owens, C. B., Yonelinas, A. P., & Eichenbaum, H. (2008). Recognition memory opposite effects of hippocampal damage on recollection and familiarity. Nature Neuroscience, 11(1), 16–18. Schacter, D. L., Verfaellie, M., & Anes, M. D. (1997). Illusory memories in amnesic patients: Conceptual and perceptual false recognition. Neuropsychology, 11(3), 331–342. Shrager, Y., Kirwan, C. B., & Squire, L. R. (2008). Activity in both hippocampus and perirhinal cortex predicts the memory strength of subsequently remembered information. Neuron, 59(4), 547–553. Slotnick, S. D., Klein, S. A., Dodson, C. S., & Shimamura, P. (2000). An analysis of signal detection and threshold models of source memory. Journal of Experimental Psychology: Human Learning and Memory., 28(6), 1499–1517. Smith, M. E. (1993). Neurophysiological manifestations of recollective experience during recognition memory judgments. Journal of Cognitive Neuroscience, 5(1), 1–13. Squire, L. R., Stark, C. E. L., & Clark, R. E. (2004). The medial temporal lobe. Annual Review of Neuroscience, 27, 279–306 (Review). Squire, L. R., Wixted, J. T., & Clark, R. E. (2007). Recognition memory and the medial temporal lobe: A new perspective. Nature Reviews Neuroscience, 8(11), 872–883. Stark, C. E., & Squire, L. R. (2000). Functional magnetic resonance imaging (fMRI) activity in the hippocampal region during recognition memory. Journal of Neuroscience, 20, 7776–7781. Suchan, B., Gayk, A. E., Schmid, G., Köster, O., & Daum, I. (2008). Hippocampal involvement in recollection but not familiarity across time: A prospective study. Hippocampus, 18(1), 92–98. Turriziani, P., Fadda, L., Caltragirone, C., & Carlesimo, G. A. (2004). Recognition memory for single items and for associations in amnesic patients. Neuropsychologia, 42, 426–433. Turriziani, P., Serra, L., Fadda, L., Caltagirone, C., & Carlesimo, G. A. (2008). Recollection and familiarity in hippocampal amnesia. Hippocampus, 18(5), 469–480. Van Strien, N. M., Cappaert, N. L., & Witter, M. P. (2009). The anatomy of memory: An interactive overview of the parahippocampal–hippocampal network. Nature Reviews Neuroscience, 10(4), 272–282 (Review). Vann, S. D., Tsivilis, D., Denby, C. E., Quamme, J. R., Yonelinas, A. P., Aggleton, J. P., et al (2009). Impaired recollection but spared familiarity in patients with extended hippocampal system damage revealed by 3 convergent methods. Proceedings of the National Academy of Sciences of the United States of America, 106(13), 5442–5447. Wais, P. E., Wixted, J. T., Hopkins, R. O., & Squire, L. R. (2006). The hippocampus supports both the recollection and the familiarity components of recognition memory. Neuron, 49(3), 459–466. Wilson, I. A., Gallagher, M., Eichenbaum, H., & Tanila, H. (2006). Neurocognitive aging: Prior memories hinder new hippocampal encoding. Trends in Neurosciences, 29, 662–670. Wixted, J. T. (2007). Dual-process theory and signal-detection theory of recognition memory. Psychological Review, 114(1), 152–176. Wixted, J. T., & Squire, L. R. (2004). Recall and recognition are equally impaired in patients with selective hippocampal damage. Cognitive, Affective & Behavioral Neuroscience, 4(1), 58–66 (Review). Wixted, J. T., & Squire, L. R. (2008). Constructing receiver operating characteristics (ROCs) with experimental animals: Cautionary notes. Learning & Memory, 15(9), 687–690. Woodruff, C. C., Hayama, H. R., & Rugg, M. D. (2006). Electrophysiological dissociation of the neural correlates of recollection and familiarity. Brain Research, 1100(1), 125–135. Woodruff, C. C., Johnson, J. D., Uncapher, M. R., & Rugg, M. D. (2005). Content-specificity of the neuralcorrelates of recollection. Neuropsychologia, 43, 1022–1032. Yonelinas, A. P. (1994). Receiver-operating characteristics in recognition memory: Evidence for a dual-process model. Journal of Experimental Psychology: Learning, Memory, and Cognition, 20(6), 1341–1354. Yonelinas, A. P. (1997). Recognition memory ROCs for item and associative information: The contribution of recollection and familiarity. Memory & Cognition, 25(6), 747–763. Yonelinas, A. P. (1999). The contribution of recollection and familiarity to recognition and source memory judgments: A formal dual-process model and an analysis of receiver operating characteristics. Journal of Experimental Psychology: Learning, Memory, and Cognition, 25, 1415–1434. Yonelinas, A. P. (2002). The nature of recollection and familiarity: A review of 30 years of research. Journal of Memory and Language, 46, 441–517. Yonelinas, A. P., & Jacoby, L. L. (1994). Dissociations of processes in recognition memory: Effects of interference and of response speed. Canadian Journal of Experimental Psychology, 48(4), 516–535. Yonelinas, A. P., Kroll, N. E., Dobbins, I., Lazzara, M., & Knight, R. T. (1998). Recollection and familiarity deficits in amnesia: Convergence of remember-know, process dissociation, and receiver operating characteristic data. Neuropsychology, 12, 323–339. Yonelinas, A. P., Kroll, N. E., Quamme, J. R., Lazzara, M. M., Sauvé, M. J., Widaman, K. F., et al (2002). Effects of extensive temporal lobe damage or mild hypoxia on recollection and familiarity. Nature Neuroscience, 5, 1236–1241. Yonelinas, A. P., & Parks, C. M. (2007). Receiver operating characteristics (ROCs) in recognition memory: A review. Psychological Bulletin, 133(5), 800–832. Yonelinas, A. P., Widaman, K., Mungas, D., Reed, B., Weiner, M. W., & Chui, H. C. (2007). Memory in the aging brain: Doubly dissociating the contribution of the hippocampus and entorhinal cortex. Hippocampus, 17(11), 1134–1140.

Consciousness and Cognition 22 (2013) 1082–1091 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Alcohol increases hypnotic susceptibility Rebecca Semmens-Wheeler ⇑, Zoltán Dienes, Theodora Duka Sackler Centre for Consciousness Science and the School of Psychology, University of Sussex, UK a r t i c l e i n f o Article history: Received 10 June 2012 Keywords: Cold control theory Hypnosis Higher order thought theory Dorsolateral prefrontal cortex Alcohol a b s t r a c t One approach to hypnosis suggests that for hypnotic experience to occur frontal lobe activity must be attenuated. For example, cold control theory posits that a lack of awareness of intentions is responsible for the experience of involuntariness and/or the subjective reality of hypnotic suggestions. The mid-dorso-lateral prefrontal cortex and the ACC are candidate regions for such awareness. Alcohol impairs frontal lobe executive function. This study examined whether alcohol affects hypnotisability. We administered 0.8 mg/kg of alcohol or a placebo to 32 medium susceptible participants. They were subsequently hypnotised and given hypnotic suggestions. All participants believed they had received some alcohol. Participants in the alcohol condition were more susceptible to hypnotic suggestions than participants in the placebo condition. Impaired frontal lobe activity facilitates hypnotic responding, which supports theories postulating that attenuation of executive function facilitates hypnotic response, and contradicts theories postulating that hypnotic response involves enhanced inhibitory, attentional or other executive function. Ó 2013 Published by Elsevier Inc. 1. Introduction Hypnotic suggestions give rise to a wide range of interesting experiences and behaviours. Typically these involve a sense of involuntariness, such as when one’s arm apparently rises by itself. Alternatively they may comprise the experience of an entirely convincing yet fabricated subjective reality, such as the experience of a mosquito on one’s hand. While there may be different underlying mechanisms involved in different types of hypnotic suggestions, and individuals may create the experience in different ways (e.g. see Terhune, Cardeña, & Lindgren, 2011; Woody & Barnier, 2008), a number of general theories have been developed in an attempt to explain hypnotic phenomena. Hypnosis can be construed either as a special state or as a way of responding to suggestions (Kirsch et al., 2011). In terms of the latter, hypnotic responding is a way of respondingin which the sense of volition or reality has been deliberately distorted (whether or not one is in a special state). In terms of the former, it is a state that may facilitate such responding. Here we investigate the effect of a drug state on hypnotic response in order to test different theories of hypnosis. Although several studies have examined the effects of drugs, including cannabis, psilocybin, diazepam and nitrous oxide on hypnotisability (Kelly, Fisher, & Kelly, 1978; Sjoberg & Hollister, 1965; Whalley & Brooks, 2009), surprisingly none has yet investigated the relationship of alcohol to hypnotic suggestibility. Yet, as we now describe, theories of hypnosis often postulate a role of the frontal lobes in hypnotic responding, and alcohol primarily disrupts frontal lobe functioning. A number of theories have emphasised the role of the frontal cortex and associated executive functions, such as attention. One broad approach posits that hypnotic phenomena arise from a state of hypofrontality (see Dietrich, 2003) and diminished executive functions such as attention. For example, Woody and Bowers (1994) postulate that hypnotic induction leads to impairment of executive functions, causing actions to be controlled by contention scheduling (i.e. habit). Woody and Sadler ⇑ Corresponding author. Address: School of Psychology, University of Sussex, Falmer, Brighton BN1 9QH, UK. Fax: +44 1273 678058. E-mail address: rebeccajsw@gmail.com (R. Semmens-Wheeler). 1053-8100/$ - see front matter Ó 2013 Published by Elsevier Inc. http://dx.doi.org/10.1016/j.concog.2013.07.001 R. Semmens-Wheeler et al. / Consciousness and Cognition 22 (2013) 1082–1091 1083 (2008) review a number of ways in which executive control mechanisms may be disrupted in order to produce hypnotic response. Similarly, Gruzelier (1998, 2006) has proposed that hypnosis results from a state of frontal lobe exhaustion and diminished attentional abilities resulting from extreme concentration during hypnotic induction. Gruzelier and Warren (1993), Kallio, Revonsuo, Hämäläinen, Markela, and Gruzelier (2001), and Farvolden and Woody (2004), found that hypnotic induction reduced letter fluency in high rather than low hypnotisables, although similar effects were not detected on other frontal tasks. Thus, responding hypnotically may involve a specific form of hypofrontality. If these theories are true then alcohol should increase hypnotic responding. Other theories would predict that the alcohol-induced frontal lobe impairment would reduce hypnotic responsiveness. The theories of both Spanos (e.g. 1986) and Hilgard (e.g. 1986) rely on the functioning of the frontal lobes for hypnotic response to be achieved. Spanos (e.g. Bertrand & Spanos, 1985; Spanos et al., 1982) has demonstrated that hypnotic behaviour can involve overcoming pre-potent responses, which necessarily involves executive functioning. Hilgard’s theory relies upon two intact but dissociated executive functions. In fact, Hilgard (1986) argued that maintaining the two dissociated streams itself took executive capacity, because the hypnotic rather than non-hypnotic performance of one of two simultaneous tasks involved more dual task interference (see also Tobis & Kihlstrom, 2010; Wyzenbeek & Bryant, 2012). Similarly Crawford, Knebel, and Vendemia (1998) argue that frontal lobe executive functions are required for hypnotic analgesia. Therefore, since alcohol impairs executive function, alcohol should decrease hypnotic susceptibility by these approaches. A more recent theory has highlighted the role of metacognition in hypnosis. The cold control theory of hypnosis (Dienes, 2012; Dienes and Perner, 2007; also see Barnier, Dienes, & Mitchell, 2009) explains hypnotic phenomena as the result of a strategic lack of awareness of the intention to perform a particular action. In other words, to respond hypnotically, the subject performs an action while thinking that they were not intending to perform that action: hypnosis essentially involves the lack of accurate higher order thoughts (HOTs) of intending. (Hence ‘cold control’: intentional control without HOTs.) Take, for example, the hypnotic suggestion that one’s arm is stiff and rigid as if splinted, so that it cannot bend. In order to perform the suggestion successfully, the subject might intend to contract the antagonistic muscles of the arm simultaneously to prevent it from bending (about 80% of participants do try to bend, Comey & Kirsch, 1999). The hypnotic aspect is the experience of involuntariness, and cold control posits that this occurs by way of avoiding HOTs of intending, which thus lead to the inaccurate HOT, ‘‘my arm has become stiff and rigid by itself and I cannot bend it.’’ Similarly, suggestions for analgesia or amnesia may involve distraction away from pain or the to-be-forgotten material. However, the hypnotic component is the ability to deceive oneself about having intended to do so; that is, by cold control theory this is done by avoiding accurate HOTs of intending. Note that on this theory hypnotic experience does not involve any alteration in first-order abilities (i.e. abilities with the function of dealing only with the world), but is achieved purely metacognitively. Thus, according to cold control, impairment of frontal function would enhance hypnotic response in virtue of the role of the frontal lobes in metacognition. Higher order thoughts of seeing have been linked to the dorsolateral prefrontal cortex (DLPFC). Lau and Passingham (2006) using fMRI found that the brain region that distinguished reports of ‘‘seeing’’ rather than of ‘‘guessing’’ for equivalent perceptual discrimination was the DLPFC; thus, the DLPFC was not linked to the first order mental state of seeing, but to awareness of seeing. In another study, subjects’ self-reported awareness of seeing was disrupted when theta burst TMS was applied to the area, even when first order perception was held constant with and without TMS (Rounis et al., 2010). That is, the disruption found was purely related to HOTs, and not first order perception. Fleming, Weil, Nagy, Dolan, and Rees (2010) also found the individual differences in the accuracy of higher order thoughts about perceiving correlated with grey and white matter volume in the same region. The neural substrate of accurate higher order thoughts may well extend beyond the DLPFC. The monitoring and cognitive control functions of the anterior cingulate cortex (the ACC) make it a likely co-candidate region for the production of HOTs. Indeed, Woody and Szechtman (2011) found in highly hypnotisable participants that there were greater levels of activation in the ACC during auditory hallucination compared to imagination of the same sounds. That is, the ACC may be involved in determining whether internally generated sensory representations are just that – imagination – or else misrepresented as perceptions. Alcohol impairs both the DLPFC (Wendt & Risberg, 2001) and the ACC (Ridderinkhof et al., 2002). Consistent with the claim these areas are involved in accurate metacognitive awareness, Sayette, Reichle, and Schooler (2009) found that alcohol compared to placebo decreased people’s awareness that they were mind wandering. The DLPFC and ACC not only have executive monitoring functions but also control functions. Thus, the effect of alcohol on these areas can also be shown by the effect of alcohol on tasks that test inhibition of pre-potent response (like the Stop Signal Task, SST; Fillimore & Weafer, 2004), or the ability to resist perseveration (like the letter fluency task; Peterson, Rothfleisch, Zelazo, & Pihl, 1990). For example, Marinkovic, Rickenbacher, Azma, and Artsy (2012) reported reduced activation in ACC bilaterally during the incongruent condition on the colour Stroop task. Similarly, Gundersen, Specht, Grüner, Ersland, and Hugdah (2008) observed decreased activation in ACC and cerebellum during a working memory task following alcohol consumption. With a different variant of a working memory task, Paulus, Tapert, Pulido, and Schuckit (2006) found less activation in the DLPFC in participants who consumed alcohol rather than placebo. In sum, alcohol impairs control and monitoring functions subserved by the DLPFC and ACC. According to theories postulating that hypnosis involves disruptions in executive control mechanisms (e.g. Woody and Sadler, 2008), alcohol should increase hypnotic responsiveness. According to cold control theory, as alcohol impairs the areas responsible for metacognitive monitoring, alcohol should make it harder to have accurate higher order thoughts of intending, and thereby facilitate hyp- 1084 R. Semmens-Wheeler et al. / Consciousness and Cognition 22 (2013) 1082–1091 notic response. On the other hand, theories that emphasize that hypnotic response involves extra executive capacity (e.g. Hilgard, 1986) predict that alcohol should impair hypnotic responding. It is well known that alcohol produces effects of an altered state of consciousness, i.e. drunkenness, and this may lead to increased expectations of hypnotic responding. Expectancy is a strong predictor of hypnotic response (Braffman & Kirsch, 1999), and indeed, according to response expectancy theory, is the final psychological mechanism by which hypnotic response is achieved (Kirsch, 1985). Theories that postulate some form of diminished frontal lobe functioning, such as cold control theory, predict that the effect of alcohol on hypnotic susceptibility will be observed above and beyond any effects of response expectancy. The main aim of the study was to determine the effect of alcohol on hypnotic suggestibility, specifically measuring the central element to successful responding: the sensation that actions and experiences ‘‘happen by themselves’’. We administered real or placebo alcohol to medium susceptible participants before they received a series of hypnotic suggestions. (Note: although suggestions typical of susceptibility scales were used, we did not assess an individual’s objective responsiveness with a pass/fail, but rather the rated feeling of automaticity, in order to increase sensitivity). Participants rated how strongly they experienced each suggestion. As a manipulation check, the effect of the alcohol on frontal function was determined by the letter fluency task and stop signal task. Before responding to hypnotic suggestions, participants also rated how strongly they expected to respond in order to control expectancy effects. 2. Methods 2.1. Participants Participants in this study were 32 undergraduate and postgraduate students aged between 18 and 39 years (M = 22, SD = 5.62) recruited from the University of Sussex hypnosis screening database. Participants scored in the medium range (4–8 suggestions passed out of 12) on the Waterloo-Stanford Group Scale of Hypnotic Susceptibility, Form C (WSGC; Bowers, 1998). Medium-susceptible participants were selected in order to allow for either an increase or a decrease in hypnotic suggestibility. If alcohol were to decrease hypnotic suggestibility then we might see a floor effect in lows, and if it were to increase it, then we might see a ceiling effect in highs. For study inclusion, minimum alcohol consumption as assessed with the Alcohol Use Questionnaire (AUQ; Mehrabian & Russell, 1978) of 10 units per week and maximum alcohol consumption of 40 units per week was defined. (One unit equals 8 g ethanol.) All participants were in good health, had a body mass index of between 18 and 28, and were not pregnant or breastfeeding. Participants included were not heavy smokers (>20 cigarettes per day) and were able to abstain from smoking for the duration of the test session. Volunteers with current symptoms of mental illness or neurological disease, a history of severe mental illness, drug or alcohol abuse or altered metabolism of alcohol (e.g. impaired liver function or gastroenteritis) were excluded from the study. Ethical approval was received from the University of Sussex ethical committee. Informed consent was obtained from each participant before commencing with the study. All 32 participants completed the study (12 males). Participants were remunerated with course credits or £5 per hour for their participation in the study. 2.2. Design Participants were randomly allocated to an alcohol or placebo condition according to a double blind between-subjects design and administered either an alcohol or placebo beverage. Participants were told that they were to receive a high or low dose of alcohol. The drinks were prepared by a laboratory assistant and administered to the participants by the researcher, who was blind to whether or not alcohol or placebo was being administered. 2.3. Materials 2.3.1. REY Auditory Verbal Learning Test (RAVLT) To assess whether the two groups differed in verbal learning and memory abilities, each participant completed the REY Auditory Verbal Learning Test (RAVLT; Rey, 1964), in which the experimenter read aloud 15 words at a rate of one per second. Participants were required to wait 2 min and then recall as many words from the list as possible.1 2.3.2. Letter fluency task Participants were asked to produce as many words as possible starting with the letters F and S (in a counterbalanced order across conditions) within 1 min. Proper nouns and variants of words already given counted as errors. This task was administered to assess the effects of alcohol on participants’ monitoring function, which is required in this task to avoid perseverations. It was scored by subtracting the score of the post-drink test from the score of the pre-drink test to establish that alcohol had impaired the frontal lobe functioning of participants in the alcohol compared to placebo condition. 1 The groups did not differ significantly on this task and thus this will not be discussed further. R. Semmens-Wheeler et al. / Consciousness and Cognition 22 (2013) 1082–1091 1085 2.3.3. Drunkenness scale and VASs for alcohol effects These scales were used to measure the participants’ general feelings of drunkenness and more specific experiences. The drunkenness scale requires participants to indicate how drunk they feel on a scale from 1 (I feel no effect of alcohol) to 9 (So drunk the room is spinning).2 The Subjective Effects Visual Analogue Scales (VASs; Loeber & Duka, 2009) required participants to indicate the degree to which they experienced each of light-headedness, contentment, stimulation, pleasant glow, irritability, alertness and relaxation, by marking a corresponding line labeled from 1 (I feel no effect of alcohol) to 10 (I feel a strong alcohol effect). 2.3.4. Hypnotic suggestions A total of nine hypnotic suggestions were made to participants following the second word fluency task, approximately 45 min after alcohol administration was finished. The suggestions covered both motor and cognitive types, as well as direct and challenge types (Barnier & Woody, 2008). The suggestions were (in order delivered): that they had a sour taste in their mouth; that their outstretched hands were attracted to each other, making them move together (magnetic hands); to feel that their outstretched right arm was weighed down by holding an imaginary heavy object that they could not keep it up (heavy arm); that a mosquito had landed on their hand and was tickling it; that their arm was so stiff and rigid that they could not bend it (rigid arm); to see two balls out of three placed in front of them (negative hallucination); that their arm was so heavy they could not lift it (arm immobilisation); to forget everything that had happened since they were hypnotised until told that they could remember (post-hypnotic amnesia) and to feel a strong urge to move seats when a clipboard was handed to them. 2.3.5. Expectancy ratings Explicit expectancy ratings were recorded using E-Prime 2.0. Before each hypnotic suggestion was made, participants were asked to report whether or not they expected to experience the suggestion. For example, ‘‘If you were given a hypnotic suggestion that your arm would feel so heavy that you would not be able to hold it up, do you expect that your arm would feel heavier than normal.’’ They responded by pressing ‘Y’ for yes and ‘N’ for no on a computer keyboard. They were then asked how confident they were about this expectancy (on a scale of 1–43). Yes/no responses were combined with confidence ratings to give a directional ‘‘explicit expectancy’’ scale, ranging from 4 (indicating a strong expectancy not to respond to a suggestion) to +4 (indicating a strong expectancy that one would respond to the suggestion). Additionally, reaction times for yes/no responses were recorded and used as a measure of unconscious expectancy (when explicit ratings are partialled out). 2.3.6. Subjective hypnotic response ratings Following each suggestion, participants were asked to rate how strongly they experienced the suggestion (on a scale of 0– 5). For example ‘‘On a scale from 0 to 5, how strongly did you feel your hand becoming heavy (where 0 means you felt your arm was no more heavy than normal and 5 means you felt your arm becoming heavy as though you had a heavy object in your hand, pulling it down.). 2.3.7. Stop Signal Task (SST) The SST was administered at the end of the session to check that participants were still generally influenced by alcohol, as response inhibition tends to be impaired under alcohol’s influence (Loeber & Duka, 2009; Fillimore & Weafer, 2004). The SST, from the CANTAB (Cambridge Cognition, Cambridge, UK; http://www.camcog.com), assesses response inhibition performance. In each trial, an arrow (go-stimulus) was presented on the screen and the participant was required to press the left or the right button of a two-choice response box as quickly as possible to indicate if the arrow was either right-facing or leftfacing. In 25% of the trials, an auditory stop signal (a beep sound) was presented at a variable delay after the go-stimulus. The subject was instructed to withhold their motor response on presentation of the stop signal. Five blocks of 64 trials were presented. The main variable was Stop Signal reaction time (SSRT) a measure of response inhibition (Robbins, 2007), which takes into account reaction time on go trials and is calculated from the length of time between the go stimulus and the stop stimulus when the participant is able to successfully inhibit his or her response in the latter 50% of trials. High SSRT indicates impaired inhibitory motor control. 2.4. Procedure Participants were instructed not to drink alcohol for at least 12 h before the start of the test session. On the day of testing zero blood alcohol concentration was ensured before the start of testing. Baseline breath alcohol concentration was measured using a breathalyser (Lion Alcolmeter SD-400, Lion Laboratories Ltd., UK) and participants completed the Subjective Effects Visual Analogue Scale (VAS) and the drunkenness scale. The REY Auditory Verbal Learning Test (RAVLT; Rey, 1964) and letter fluency task were administered to control for any pre-manipulation differences between the two groups. 2 1, Feel no effect of alcohol; 2, Feel the first effects of alcohol; 3, Slightly tipsy; 4, Feeling warm; 5, a bit disinhibited; 6, Very merry; 7, Beginning to feel uncoordinated; 8, Very drunk; hard to focus properly; 9, So drunk the room is spinning and I feel sick. 3 (1) I am completely guessing, I have no idea whether I would or wouldn’t. (2) I am more or less guessing, but I have some feeling I was right. (3) I am pretty sure I am right. (4) I am completely certain. 1086 R. Semmens-Wheeler et al. / Consciousness and Cognition 22 (2013) 1082–1091 Participants in the alcohol group were given an alcohol dose of 0.8 g/kg. For a 70 kg person this is equal to about 56 g of pure alcohol. This is equivalent to approximately 2.5 pints of lager or 5 glasses of wine (Weissenborn & Duka, 2003). The alcohol beverage consisted of 90% v/v alcohol diluted with tonic water (SchweppesÒ Indian Tonic Water) to make up a drink of 500 ml which was mixed with Angostura BitterÒ to mask the taste of the alcohol. The placebo beverage consisted of 500 ml tonic water and Angostura BitterÒ only. Drinks were divided into 10 portions, and participants consumed the ten portions at 3 min intervals in the presence of the experimenter. Participants were breathalysed 15 min after alcohol consumption and then completed another set of the Subjective Effects Visual Analogue Scales (VASs; Loeber and Duka (2009) and a 9 point drunkenness scale in order to obtain a subjective rating of how ‘drunk’ they felt. Participants were next given a brief hypnotic induction and nine suggestions. They were asked about their expectancy before each suggestion and subjective response after each suggestion. Following this they were breathalysed before completing the letter fluency task and finally the stop signal task. 3. Results 3.1. How blind were subjects to condition? Subjects were told at the end of the experiment that there had been two conditions: alcohol and placebo. Subjects were then asked if they thought they had received alcohol; 86% thought they had. 73% of those in the placebo condition and 100% of those in the alcohol condition, v2(1) = 5.8, p = .08). A t-test indicated that those who believed they had received alcohol rated their drunkenness higher (M = 3.4, SD = 2.1) than those who did not (M = 0.25, SD = 0.5), t(30) = 2.98, p = .006. Participants who believed they had received alcohol also rated their experience of ‘‘pleasant glow’’ higher than those who believe they had not (M = 5.7, SD = 1.4 and M = 3.9, SD = 2.8, respectively), t(30) = 2.14, p = .002. Further, when scores on the scales were averaged both before and after administration of alcohol or placebo, not only did alcohol change subjective feelings, from 2.82 (SD = 0.49) pre-administration to 5.04 (0.79), post administration, t(13) = 7.17, p < .001, d = 3.37, but so did the placebo, from 3.25 (0.75) to 3.81 (0.80), t(13) = 3.25, p = .006, d = 0.72. Nonetheless the change produced by alcohol (0.94, SD = 1.04) was detectably different from the change produced by placebo (0.29, SD = 0.73), t(26) = 2.94, p = .035, 1-tailed, d = 0.45. Participants who had alcohol reported feeling significantly more lightheaded, t(30) = 3.21, p = .003, d = 1.13 (see Table 1), and more intoxicated, t(30) = 13.23, p < .001, d = 1.84 (see Table 2) compared to those who had placebo. In sum, while there was a placebo effect, subjects potentially also had some knowledge about condition, and thus it is important to control expectancies in determining alcohol’s effect on hypnotic response. 3.2. Was enough alcohol administered to affect frontal lobe functioning? Blood alcohol levels (BAC) at 45 min ranged from 0.55 promille w/volume to 0.96 promille w/volume in the alcohol group. No participant of the placebo group had a detectable BAC (derived from the breath alcohol level; BrAC) or BrAC. As expected, alcohol impaired performance on tests of frontal lobe functioning. The alcohol group’s decline in performance on the word fluency task (M = 3.18, SD = 2.58) was greater than the placebo group’s (M = 0.33, SD = 1.45) Table 1 Mean subjective mood ratings (SEM) after administration of alcohol or placebo. * VAS Placebo Alcohol Drunkenness Light-headed Pleasant glow Irritable Relaxed Alert Stimulated Contented 1.34 (.19) 1.93 (.53) 4.73 (.34) 0.50 (.21) 5.73 (.41) 4.70 (.43) 4.73 (.27) 6.17 (.35) 4.25 (.23)* 4.5 (.61)* 6.12 (.49) 1.59 (.44) 6.15 (.29) 4.5 (.39) 4.91 (.46) 6.12 (.50) t(30) = 13.23, p < .001, d = 2.59 t(30) = 3.21, p = .003, d = 1.12 t(30) = 2.51, p = .018, d = 0.89 t(30) = 2.14, p = .040, d = 0.78 t(30) = 0.83, p = .411, d = 0.30 t(30) = 0.35, p = .733, d = 0.12 t(30) = 0.32, p = .749, d = 0.12 t(30) = .08, p = .938, d = 0.03 Significant after controlling for familywise error at the .05 level (Hochberg’s, 1988 sequential Bonferroni). Table 2 Mean scores on tests of frontal lobe function (SEM*). Letter fluency pre Letter fluency post SST post * Standard error of the mean. Placebo Alcohol 15.733 (1.08) 15.53 (0.71) 161.09 (15.88) 14.24 (1.01) 10.94 (0.65) 255.57 (29.82) t(30) = 1.01, p = .32 t(30) = 4.75, p < .001 t(30) = 2.69, p < .012 R. Semmens-Wheeler et al. / Consciousness and Cognition 22 (2013) 1082–1091 1087 Table 3 Mean subjective responses to individual hypnotic suggestions (SEM). Suggestion Placebo (SEM) Alcohol (SEM) Rigid arm Posthypnotic suggestion Negative hallucination Heavy arm Arm immobilisation Sour taste Magnetic hands Posthypnotic amnesia Mosquito hallucination 2.53 (.41) 0.73 (.25) 0.40 (.16) 3.73 (.37) 2.60 (.34) 1.67 (.35) 3.27 (.32) 1.67 (.39) 1.07 (.36) 3.82 (.31) 2.06 (.47) 1.71 (.50) 4.53 (.15) 3.41 (.24) 2.41 (.31) 4.00 (.24) 2.24 (.32) 1.12 (.27) t(30) = 2.53, p = .017 t(30) = 2.42, p = .022 t(30) = 2.36, p = .025 t(30) = 2.08, p = .047 t(30) = 1.99, p = .056 t(30) = 1.61, p = .119 t(30) = 1.87, p = .172 t(30) = 1.15, p = .259 t(30) = 0.12, p = .91 t(30) = 2.69, p = .014, d = 0.87. Kallio et al. (2004) found a hypnotic induction reduced letter fluency by 30%, similar to the 23% reduction alcohol produced in the current study (from 14.24 to 10.91): That is, we have administered sufficient alcohol to reduce frontal function by the order of magnitude relevant for effects on hypnotic response. The alcohol group’s reaction times (M = 255.66, SD = 122.96) on the stop signal task were also longer than those of participants in the placebo condition (M = 161.09, SD = 61.5, t(30) = 2.69, p = .012, d = .97, further indicating alcohol’s effect on frontal functioning. 3.3. Did alcohol affect hypnotic response? Table 3 shows the mean subjective response for each suggestion separately. When responses were averaged over suggestions, the alcohol group (M = 2.81, SD = 0.71) responded subjectively more to hypnotic suggestions than the placebo group placebo group (M = 1.96, SD = 0.75, t(30) = 3.27, p = .003, d = 1.16) (on a scale that went from 0 to 5). This is the key result of the study. An ANCOVA was performed on subjective hypnotic response between the alcohol group (adjusted M = 2.70, SEM = .19) and the placebo group (adjusted M = 1.96, SEM = 0.19) with the hypnotic suggestibility score (WGSC) as a covariate. The difference in hypnotic suggestibility remained significant, F(1, 27) = 7.81, p = .009. A significant positive correlation was found between expectancy and subjective response to suggestion overall, Pearson’s r = .55, p < .001. So was the effect of alcohol just based on expectancy? The expectancy ratings of the alcohol group (M = 0.56, SD = 0.91) and placebo group (M = 0.11, SD = 1.28), did not differ significantly t(30) = 1.16, p = .26, d = 0.40, (on an expectancy scale that went from 4 to +4).4 Crucially, when expectancy was put in as a covariate, the difference between the groups in hypnotic suggestibility remains (adjusted means: alcohol group M = 2.73, SEM = 1.7; placebo group M = 2.03, SE = 1.6), t(29) = 3.57, p = .002). A more formal mediation analysis looking at a range of possible mediators is conducted below. 3.4. Direct and indirect measures of expectation Direct and indirect measures of expectancy were taken. Before each hypnotic suggestion was made, participants were asked to report whether or not they expected to experience the suggestion and how confident they were about this expectancy (on a scale of 1–4) as a direct measure, and the reaction time for the yes/no response was taken as an indirect measure. We would expect that subjects with more confidence in ‘no’ responses would be less hypnotically suggestible, and vice versa for ‘yes’ responses. In order to examine the effect of these direct and indirect measures of expectancy on subjective ratings (hypnotic response), for each subject a multiple regression was run with expectation RT and confidence rating simultaneously predicting subjective response for yes and no responses separately. Regression weights (betas) for yes and no should be in opposite directions, as the more confident a subject is in a no response, the less hypnotically responsive they should be, so we reversed the sign for ‘no’ expectancy response regression weights and averaged across both (for both indirect and direct measures). The mean standardised regression weight for confidence was 0.11, t(24) = 1.00, p = .33, d = .51, indicating that conscious expectancy did not significantly predict subjective response controlling for RT differences. However, the mean standardised regression weight for RT was 0.29, significantly above chance, t(25) = 2.62, p = .015, d = .20, indicating that RT as an indirect measure of expectancy did predict subjective response. 4 In order to determine if this non-significant result was sensitive a Bayes Factor was used to compare the alternative hypothesis (that expectancy was higher for the alcohol rather than placebo group) to the null hypothesis. A Bayes Factor greater than 3 indicates strong evidence for the alternative over the null; less than a 1/3 indicates strong evidence for the null over the alternative; and anything in between indicates the data are insensitive (Dienes, 2011). First we need to specify what sizes of effect the alternative hypothesis predicts. The raw regression slope of hypnotic response on expectancy was 0.43 (t = 3.58, p = .001). The difference in hypnotic response between groups was 0.85 units. Thus the change in expectancy needed to produce the observed change in hypnotic response is 0.85/0.43 = 2.0 units. Thus, the predictions of the alternative hypothesis were modelled as a normal with a mean of 2.0 and standard deviation of 1.0 to indicate maximum vagueness in this estimate (see Dienes, 2011, Appendix, for this recommendation; and website for Dienes, 2008, for free online software for Bayes Factors). The Bayes Factor was 0.25; thus the non-significant difference in expectancy between groups is strong support of the null hypothesis. The ANCOVA also shows that expectation did not fully mediate the effect of group. The extent of mediation is more formally analysed below. 1088 R. Semmens-Wheeler et al. / Consciousness and Cognition 22 (2013) 1082–1091 Explicit Expectancy (M) b = .53* a = .21 Condition (X) c =.49* (.51)* Subjective response (Y) *p<.05 Fig. 1. Regression coefficients for the relationship of alcohol/placebo condition with subjective hypnotic response as mediated by explicit expectancy ratings. 3.5. Mediation analyses Measures of explicit and unconscious expectancy were entered into separate mediation analyses to investigate the effect they had on the relationship between alcohol/placebo condition and subjective hypnotic response. In order to measure unconscious expectancies differences between groups, Go trial RTs from the SST were partialled out of ‘yes’ and ‘no’ expectancy RTs as a baseline measure of reaction time so as to account for any between-groups differences, such as alcohol causing longer latencies. ‘Yes’ and ‘no’ expectancy RTs were regressed separately on SST go trial RTs and the residuals were used as the measure of unconscious expectancies. Fig. 1 displays mediation analysis for alcohol/placebo condition predicting hypnotic response with explicit expectancy as a mediator. The standardised regression weight predicting hypnotic response from condition partialling out expectancy is ‘‘c’’; the correlation between condition and expectation is ‘‘a’’; and the standardised regression weight predicting hypnotic response from expectancy, controlling condition is ‘‘b’’. If c > 0, then there is not full mediation; if ab = 0 there is no mediation; and if ab > 0, then there is some mediation (Woody, 2011). If there is full mediation by a variable, then c = 05 and if there is no mediation, then ab = 0 (see Woody, 2011). Conventional mediation analysis, being based on significance testing, does not provide a systematic method of establishing no meditation, as opposed to partial mediation by establishing if ab = 0. The amount of mediation depends on the size of ab. As c > 0 in all cases below, the only question is whether there is no mediation or partial mediation, and this depends on determining if ab = 0. Thus, we calculated a Bayes Factor, the only known method of determining degree of evidence for a point null hypothesis (Berger & Sellke, 1987). As the partial correlations were close to the original correlations in each case, we normalised a, b and c with Fisher’s Z, which has a known standard error. We can use the following formula for the standard error of ab in order to test if ab = 0 or if ab > 0: SEab ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 2 a2 SEb þ b SEa2 To represent the prediction of partial mediation for the purposes of calculating a Bayes factor, we used a uniform distribution from zero to an upper limit of the correlation between condition and hypnotic response – because the most the mediated effect can be (ab) is the full effect. If the Bayes Factor is <1/3, then it is strong evidence for no mediation. If it is >3, it is strong evidence for partial mediation, and if it is anywhere in-between, then there is insufficient evidence. (Note a limitation of this method is that the distribution of ab is not normal, see Mackinnon & Fairchild, 2009; however we use the method here as an approximation so as to use the online Bayes calculator.6 As can be seen in Fig. 1, c > 0, therefore there is not full mediation. The Bayes Factor for the test of ab > 0 as opposed to ab = 0 is 0.78. Therefore, the evidence is simply insensitive for indicating whether or not there was partial mediation of the effect of condition on subjective response by explicit expectancy. As can be seen in Fig. 2, c > 0, therefore there is not full mediation of ‘no’ expectancy RTs on the relationship between alcohol/placebo condition and subjective hypnotic response. The Bayes Factor for the test of ab > 0 as opposed to ab = 0 is 0.02. Therefore, there was evidence for no mediation by unconscious expectation not to respond to a suggestion. As can be seen in Fig. 3, c > 0, therefore there is not full mediation. The Bayes Factor for the test of ab > 0 as opposed to ab = 0 is 0.08. Therefore, there was evidence for no mediation by unconscious expectation to respond to a suggestion. As can be seen in Fig. 4, c > 0, therefore there is not full mediation. The Bayes Factor for the test of ab > 0 as opposed to ab = 0 is 0.13 Therefore, there was evidence for no mediation of condition on subjective response by SST no-go trial RTs (response inhibition), which were used as a measure of frontal lobe functioning. 5 Therefore, we need to first of all check to see if c is significant. If it is not, (and ab is non-zero) then there may be partial or full mediation. A Bayes Factor can be used to test c > 0 by using an estimate of the full effect as an upper limit of the uniform, in a similar way as for testing ab. 6 We checked the conclusions with a Bayes factor method that involves no violations of distributional assumptions, specifically by using the product of normals to represent the likelihoods, and all conclusions stand. This latter method will be described in a future publication. It is reassuring that they give similar answers. R. Semmens-Wheeler et al. / Consciousness and Cognition 22 (2013) 1082–1091 1089 No Expectancy RTs -.03 .06 Subjective response Condition .51* (.51)* *p<.05. Fig. 2. The effect of condition (alcohol or placebo) on subjective hypnotic response with ‘no’ expectancy RTs as a single mediator. Yes Expectancy RTs .09 -.25 Condition .50* (.51)* Subjective response *p<.05. Fig. 3. Regression coefficients for the relationship of alcohol/placebo condition with subjective hypnotic response as mediated by ‘yes’ expectancy RTs. SST No-go trial RTs -.11 .44* Condition .52* (.51)* Subjective response *p<.05. Fig. 4. Regression coefficients for the relationship of alcohol/placebo condition with subjective hypnotic response as mediated by SST no-go trial RTs. 4. Discussion In this study we examined the effect of alcohol versus placebo on how compelling the hypnotic experience is. The placebo was convincing as a possible alcoholic drink, although there was a significant difference between the ratings of drunkenness reported between the two groups. The key result was that alcohol consumption increased hypnotic responsiveness compared to placebo. The results conceptually replicate those of Dienes and Hutton (2013) who showed that applying rTMS to the left DLPFC increased subjective ratings of hypnotic experience, compared to stimulation of the vertex. While explicit expectancy strongly predicted performance, the effect of alcohol on hypnotic suggestibility remained after controlling for explicit expectancy. We measured unconscious as well as conscious expectancies for the first time, and showed that unconscious expectancy could predict hypnotic response above and beyond conscious expectancy. Yet the effect of alcohol could not be fully accounted for by either type of expectancy. Bayesian analysis indicated that while explicit expectancy might have partially mediated the relationship between alcohol condition and hypnotic responding, unconscious expectancies did not mediate at all. The results also confirmed that alcohol impaired frontal lobe functions, as demonstrated by the alcohol group’s decline in performance on the stop signal task and the letter fluency task, consistent with theories postulating a role for hypo- rather than hyper-frontality in hypnotic response. However, performance on the stop signal task did not mediate the relationship between alcohol condition and subjective hypnotic response, which is problematic for the original dissociated control theory (Woody & Bowers, 1994) and related neurophysiological approaches (Gruzelier, 1998; Gruzelier, 2006), which postulate a 1090 R. Semmens-Wheeler et al. / Consciousness and Cognition 22 (2013) 1082–1091 generalized hypofrontality producing hypnotic response. However, these findings support theories postulating that only some aspect of diminished frontal lobe functioning is related to hypnotic suggestibility, such as the cold control theory (e.g. Dienes, 2012) and some types of dissociation theory (e.g. second order dissociative control theory; Woody & Sadler, 2008). For example, according to cold control theory, it is only the effect of alcohol on higher order thoughts of intending that is relevant to its effect on hypnotic response to the suggestions used, and not the effect of alcohol on the inhibitory processes measured by the stop–go task. Similarly, on second order dissociated control theory, it is specifically disruption in the monitoring of control processes that leads to hypnotic response, and there is no reason why the stop–go task would measure that aspect of alcohol’s influence on the frontal system. These findings do not, however, rule out the arguments suggesting that sufficient frontal lobe impairment should reduce hypnotisability. For example, the theories of both Spanos (1986) and Hilgard (1977) rely on intact executive functioning. Similarly, cold control does not postulate a state of utter hypofrontality; on this theory executive control is still implemented in order to carry out the cognitive or motor action performed. Thus, if executive function is impaired to the degree that the action could not be performed by executive control, then hypnotic suggestibility would be impaired (Dienes and Perner, 2007; contrast Woody & Bowers, 1994). For moving the hands together, or even imagining a mosquito, it is obvious people have sufficient frontal function after the amount of alcohol we administered to still intend to voluntarily perform these actions and successfully perform them. What is crucial for hypnotic suggestibility to be facilitated is that diminished frontal lobe functioning reduces concomitant higher order thoughts about intentions, allowing for HOTs of not intending to arise (Barnier et al., 2009). Indeed, it may be that performance on suggestions that heavily involve executive functions would be impaired following alcohol consumption (e.g. consider the inhibition of pre-potent responses hypnotically suggested in Bertrand & Spanos, 1985). For example, forgetting the number 4, which involves overcoming habit, may become more difficult according to cold control (Dienes and Perner, 2007) but not dissociated control theories (Woody & Bowers, 1994). We found that alcohol increased responsiveness to a negative hallucination suggestion. This suggestion involves the hypnotic subject avoiding the perception of a clearly visible object, a mental task that one would naturally assume to be an inhibitory task. The increase in the alcohol group’s ability on this task despite a reduction in executive functioning suggests that responding to this suggestion does not involve above average inhibitory abilities (cf. Kirsch et al., 2011). It may be people attended away from the third ball without being aware of that intention (cold control) – the special ability of highs may not be in their ability to attend away, which may be normal (Dienes et al., 2009) but in their ability to not know that is what they were doing. We predict that once sufficient alcohol is administered to impair inhibition of prepotent responses under standard conditions, the corresponding response performed hypnotically will also be impaired. Frontal lobe impairment may be just one of a number of ways of creating a hypnotic experience. Highly hypnotisable subjects differ in the way they create hypnotic experiences. For example, Barber’s three-dimensional theory of hypnosis theory suggests that there are three types of hypnotisable subjects: those who are fantasy prone and spend much of their lives having ‘‘real-as-real’’ daydreams; amnesic subjects, who tend to forget life events and hypnotic experiences and subjects who are extremely motivated and have strong expectations about their ability to respond hypnotically (Barber, 1999). McConkey his colleagues (e.g. McConkey et al., 1989) have also identified two types of highly hypnotisables: those who actively construct hypnotic experience and those who are more passive, listening to the suggestions and waiting for the effects to happen to them. Similarly, Terhune et al. (2011; Terhune & Brugger, 2011) showed that highly hypnotisables can be separated into high and low dissociating groups which differ in their performance on executive tasks, and consequently in how they respond to suggestions. Future research could investigate the effect of alcohol on different types of highs. Although alcohol particularly disrupts the DLPFC and ACC, it also affects a large area of the prefrontal cortex and beyond (Kähkönen, Wilenius, Nikulin, Ollinkainen, & Ilmoniemi, 2003) and so we cannot definitively conclude that the increase in hypnotic suggestibility was specifically due a reduction in metacognition, nor even specifically executive function. In future, specific disruption of DLPFC function could be coupled with a measure of accuracy of higher order thoughts as well as hypnotic responsiveness. References Barber, T. X. (1999). Hypnosis: A mature view. Contemporary Hypnosis, 16(3), 123–127. Barnier, A., & Woody, E. Z. (2008). Hypnosis scales for the twenty-first century: What do we need and how should we use them? In M. Nash & A. Barnier (Eds.). The Oxford handbook of hypnosis: Theory, research, and practice (pp. 255–282). Oxford University Press. Barnier, A., Dienes, Z., & Mitchell, C. J. (2009). How hypnosis happens: New cognitive theories of hypnotic suggestibility. In M. Nash & A. Barnier (Eds.), The Oxford handbook of hypnosis: Theory, research, and practice (pp. 141–178). Oxford University Press. Berger, J. O., & Sellke, T. (1987). Testing a point null hypothesis: the irreconcilability of P-values and evidence. Journal of the American Statistical Association, 82, 112–122. Bertrand, L. D., & Spanos, N. P. (1985). The organization of recall during hypnotic suggestions for complete and selective amnesia. Imagination, Cognition, and Personality, 4, 249–261. Bowers, K. S. (1998). Waterloo-Stanford group scale of hypnotic susceptibility, form c: Manual and response booklet. International Journal of Clinical and Experimental Hypnosis, 46(3), 250–268. Braffman, W., & Kirsch, I. (1999). Imaginative suggestibility and hypnotizability: An empirical analysis. Journal of Personality and Social Psychology, 77(3), 578–587. Comey, G., & Kirsch, I. (1999). Intentional and spontaneous imagery in hypnosis: the phenomenology of hypnotic responding. International Journal of Clinical and Experimental Hypnosis, 47, 65–85. Crawford, H. J., Knebel, T., & Vendemia, J. M. C. (1998). The nature of hypnotic analgesia: Neurophysiological foundation and evidence. Contemporary Hypnosis, 15, 24–35. Dienes, Z. (2008). Understanding psychology as a science: An introduction to scientific and statistical inference. Palgrave Macmillan. R. Semmens-Wheeler et al. / Consciousness and Cognition 22 (2013) 1082–1091 1091 Dienes, Z. (2011). Bayesian versus Orthodox statistics: Which side are you on? Perspectives on Psychological Sciences, 6(3), 274–290. Dienes, Z. (2012). Is hypnotic responding the strategic relinquishment of metacognition? In M. Beran, J. L. Brandl, J. Perner, & J. Proust (Eds.), The foundations of metacogntion (pp. 267–278). Oxford University Press. Dienes, Z., Brown, E., Hutton, S., Kirsch, I., Mazzoni, G., & Wrightx, D. B. (2009). Hypnotic suggestibility, cognitive inhibition, and dissociation. Consciousness and Cognition, 18, 837–847. Dienes, Z., & Perner, J. (2007). The cold control theory of hypnosis. In G. Jamieson (Ed.), Hypnosis and conscious states: The cognitive neuroscience perspective (pp. 293–314). Oxford University Press. Dietrich, A. (2003). Introduction to consciousness. Hampshire: Palgrave Macmillan. Farvolden, P., & Woody, E. Z. (2004). Hypnosis, memory, and frontal executive functioning. International Journal of Clinical and Experimental Hypnosis, 52(1), 3–26. Fillimore, M., & Weafer, J. (2004). Alcohol impairment of behavior in men and women. Addiction, 99(10), 1237–1246. Fleming, S. M., Weil, R. S., Nagy, Z., Dolan, R. J., & Rees, G. (2010). Relating introspective accuracy to individual differences in brain structure. Science, 329, 1541–1543. Fromm, W., & Nash, M. (Eds.). (1992). Hypnosis: Research developments and perspectives (3rd ed.). New York: Guildford Press. Gruzelier, J. (1998). A working model of the neurophysiology of hypnosis: A review of evidence. Contemporary Hypnosis, 15, 3–21. Gruzelier, J. (2006). Frontal functions, connectivity and neural efficiency underpinning hypnosis and hypnotic susceptibility. Contemporary: Hypnosis, 23, 15–32. Gruzelier, J., & Warren, K. (1993). Neuropsychological evidence of reductions on left frontal tests with hypnosis. Psychological Medicine, 23, 93–101. Gundersen, H., Specht, K., Grüner, R., Ersland, L., & Hugdah, K. (2008). Separating the effects of alcohol and expectancy on brain activation: An fMRI working memory study. Neuroimage, 42(4), 1587–1596. Hilgard, E. R. (1977). Divided consciousness: Multiple controls in human thought and action. New York: Wiley-Interscience. Jamieson, G. (Ed.). (2007). Hypnosis and conscious states: The cognitive neuroscience perspective (pp. 293–314). Oxford University Press. Kähkönen, S., Wilenius, J., Nikulin, V., Ollinkainen, M., & Ilmoniemi, R. (2003). Alcohol reduces prefrontal cortical excitability in humans: A combined TMS and EEG study. Neuropsychopharmacology, 28, 747–754. Kallio, S., Revonsuo, A., Hämäläinen, H., Markela, J., & Gruzelier, J. (2001). Anterior brain functions and hypnosis: A test of the frontal hypothesis. International Journal of Clinical and Experimental Hypnosis, 49, 95–108. Kelly, S. F., Fisher, S., & Kelly, R. J. (1978). Effects of cannabis intoxication on primary suggestibility. Psychopharmacology, 56(2), 217–219. Kirsch, I. (1985). Response expectancy as a determinant of experience and behaviour. American Psychologist, 40, 1189–1202. Kirsch, I., Cardeña, E., Derbyshire, S., Dienes, Z., Heap, M., Kallio, S., et al (2011). Definitions of hypnosis and hypnotizability and their relation to suggestion and suggesitibility: A consensus statement. Contemporary Hypnosis and Integrative Therapy, 28(2), 107–111. Lau, H., & Passingham, R. (2006). Relative blindsight in normal observers and the neural correlate of visual consciousness. Proceedings of the National Academy of Science, 103, 18763–18768. Loeber, S., & Duka, T. (2009). Acute alcohol impairs conditioning of a behavioural reward-seeking response and inhibitory control processes–implications for addictive disorders. Addiction, 104(12), 2013–2022. Mackinnon, D. P., & Fairchild, A. J. (2009). Current directions in mediation analysis. Current Directions in Psychological Science, 18, 16–20. Marinkovic, K., Rickenbacher, E., Azma, S., & Artsy, E. (2012). Acute alcohol intoxication impairs top–down regulation of Stroop incongruity as revealed by blood oxygen level-dependent functional magnetic resonance imaging. Human Brain Mapping, 33, 319–333. Mehrabian, A., & Russell, J. A. (1978). A questionnaire measure of habitual alcohol use. Psychological Report, 43, 803–806. Peterson, J. B., Rothfleisch, J., Zelazo, P. D., & Pihl, R. O. (1990). Acute alcohol intoxication and cognitive functioning. Journal of Studies on Alcohol, 51, 114–122. Paulus, M., Tapert, S., Pulido, C., & Schuckit, M. (2006). Alcohol attenuates load-related activation during a working memory task: Relation to level of response to alcohol. Alcoholism, 30(8), 1363–1371. Rey, A. (1964). L’examen clinique en psychologie. Paris: Presses Universitaires de France. Ridderinkhof, K. R., de Vlugt, Y., Bramlage, A., Spaan, M., Elton, M., Snel, et al (2002). Alcohol consumption impairs detection of performance errors in mediofrontal cortex. Science, 298, 2209–2211. Robbins, T. W. (2007). Shifting and stopping: Fronto-striatal substrates, neurochemical modulation and clinical implications. Philosophical Transactions of the Royal Society: B Biological Sciences, 362, 917–932. Sayette, A. M., Reichle, E. D., & Schooler, J. W. (2009). Lost in the sauce: The effects of alcohol on mind wandering? Psychological Science, 20(6), 747–752. Sjoberg, B. M., & Hollister, L. E. (1965). The effects of psychotomimetic drugs on primary suggestibility. Psychopharmacology, 8, 251–262. Spanos, N. P., Radtke, H. L., & Dubreuil, D. L. (1982). Episodic and semantic memory in post-hypnotic amnesia: a re-evaluation. Journal of Personality and Social psychology, 43, 565–573. Spanos, N. (1986). Hypnotic behaviour: a social–psychological interpretation of amnesia, analgesia, and ‘trance logic’. Behavioural and Brain Sciences, 9, 449–502. Tobis, I., & Kihlstrom, J. (2010). Allocation of attentional resources in post hypnotic suggestion. International Journal of Clinical and Experimental Hypnosis, 58, 367–382. Terhune, D. B., & Brugger, P. (2011). Doing better by getting worse: Posthypnotic amnesia improves random number generation. PLoS ONE, 6(12), e29206. http://dx.doi.org/10.1371/journal.pone.0029206. Terhune, B., Cardeña, E., & Lindgren, M. (2011). Dissociative tendencies and individual differences in high hypnotic suggestibility. Cognitive Neuropsychiatry, 16(2), 113–135. Whalley, M., & Brooks, G. (2009). Enhancement of suggestibility and imaginative ability with nitrous oxide. Psychopharmacology, 203, 745–752. Weissenborn, R., & Duka, T. (2003). Acute alcohol effects on cognitive function in social drinkers: Their relationship to drinking habits. Psychopharmacology, 165(3), 306–312. Wendt, P., & Risberg, J. (2001). Ethanol reduces rCFB activation of left dorsolateral prefrontal cortex during a verbal fluency task. Brain and Language, 77, 197–215. Woody, E. Z. (2011). An SEM perspective on evaluating mediation: What every clinical researcher needs to know. Journal of Experimental Psychopathology, 2(2), 210–251. Woody, E., & Bowers, K. S. (1994). A frontal assault on dissociated control. In S. J. Lynn & J. W. Rhue (Eds.), Dissociation: Clinical and theoretical perspectives (pp. 52–79). New York: Guilford Publications. Woody, E., & Szechtman, H. (2011). Using hypnosis to develop and test models of psychopathology. Journal of Mind-Body Regulation, 1(1). Wyzenbeek, M., & Bryant, R. A. (2012). The cognitive demands of hypnotic suggestibility. International Journal of Clinical and Experimental Hypnosis, 60(67– 80), 2011.

Consciousness and Cognition Consciousness and Cognition 14 (2005) 641–644 www.elsevier.com/locate/concog Commentary Autism and the experience of a perceptual object D. Ben Shalom *,1 Zlotowski Center for Neuroscience, Ben Gurion University of the Negev, Israel Received 18 September 2004 Available online 27 June 2005 Abstract Sewards and Sewards (2002) argue that while computations necessary for object recognition occur throughout the ventral visual stream, object recognition awareness involves the anterior temporal lobe and the medial orbital prefrontal cortex. The present paper suggests, however, that the medial orbital prefrontal cortex has a unique contribution, namely that of producing a basic experience of a perceptual object. It is further argued that the mechanisms that produce this experience also result in making the object more important than its subparts and features. Finally, it is argued that a reduction in this importance may account for some perceptual difficulties in high-functioning autism. This view is consistent with evidence for early selective abnormalities in other systems involving the medial prefrontal cortex in autism. Ó 2005 Elsevier Inc. All rights reserved. Keywords: Autism; Object recognition; Medial prefrontal cortex 1. Introduction Object recognition and object recognition awareness have been the subject of much debate. In their 2002 review, Sewards and Sewards pinpointed object recognition awareness, or the conscious recognition of objects, to activity in the anterior temporal lobe and the medial orbital * Fax: +972 8 6472907 E-mail address: doritb@bgumail.bgu.ac.il. 1 Present adress: Department of Foreign Literatures and Linguistics, Ben Gurion University of the Negov, Beer Sheva 84105, Israel. 1053-8100/$ - see front matter Ó 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2005.03.005 642 D.B. Shalom / Consciousness and Cognition 14 (2005) 641–644 prefrontal cortex. The present commentary has two different purposes. One is to contribute to the discussion of the roles of the medial prefrontal cortex and the anterior temporal lobe in object recognition (Bar, 2004; Rolls, 2004; Sewards & Sewards, 2002). The other is to try to shed light on some peculiar perceptual difficulties in high-functioning autism, by connecting them with medial prefrontal function and dysfunction. These two purposes are probably related, in the sense that the unique contribution of the medial prefrontal cortex to object recognition is only visible in a population in which it is selectively impaired relative to other components of object recognition. 2. The medial prefrontal cortex and the anterior temporal lobe in object recognition Traditional, feedforward views of visual processing held that visual features are first extracted in lower-level cortical areas (Hubel & Weisel, 1962) and then gradually projected to higher-level cortical regions in which object recognition is achieved (e.g., Tanaka, 1996). One challenge to this view comes from the work of Ahissar and Hochstein (2000), who argue that explicit vision progresses in a reverse hierarchical direction, with conscious perception beginning at the highest cortical areas and gradually returning down the visual pathway as needed. This view is still generally consistent with the work of Sewards and Sewards (2002), who argued that the anterior temporal lobe and the medial orbital prefrontal cortex play similar roles in object recognition awareness. A more radical view, suggested by Bar (2004) and Rolls (2004) is that the prefrontal cortex and the anterior temporal lobe perform fundamentally different roles in object recognition. According to Bar (2004), low spatial frequencies in the image are extracted quickly from early visual areas to the prefrontal cortex. The global information conveyed by these frequencies is typically sufficient to limit the number of object representations that are considered. According to Rolls (2004), the orbitofrontal cortex decodes the reward value of primary reinforcers (such as taste) and enables sensory representations of objects (including visual objects) to be associated with these reinforcers. 3. Visual perceptual difficulties in autism A range of perceptual abnormalities in various modalities are described in firsthand accounts of high-functioning autism, and frequently portrayed as central to the autistic experience (for a review see OÕNeill & Jones, 1997). Some of these abnormalities may be simple hyper- or hyposensitivities, and could be related to amygdala dysfunction (e.g., Baron-Cohen et al., 2000). More intriguing are reports of what sounds like object fragmentation, e.g., ‘‘For instance, when I am confronted with a hammer I am initially not confronted with a hammer at all but solely with a number of unrelated parts’’ (van Dalen, 1994, in Brock, Brown, Boucher, & Rippon, 2002). or ‘‘When I was 7 I got a dollÕs house for Christmas. I saw a big red angular object with corrugated ridges. I picked it up with the top ridge in my mouth and played the ridges as a musical instrument quite happy with my found toy [. . .]. I later moved onto the series of flattened D.B. Shalom / Consciousness and Cognition 14 (2005) 641–644 643 white blocks, disassembled them from the structure and tapped at them as one does on a door before stacking them in piles according to size. No thinking in pictures. I saw no dollÕs house’’ (Williams, 2003). These and the clinically acknowledged intensive interest in parts of objects in low-functioning autism (American Psychiatric Association, 1994) are usually interpreted in terms of ‘‘weak coherence’’ (Frith, 1989), i.e., a bias toward local rather than global processing (at least when there is no explicit instruction to be aware of the global level, Plaisted, Swettenham, & Rees, 1999). 4. The medial prefrontal cortex in memory and emotion A general role for the medial prefrontal cortex in memory and emotion is well established (cf., Papez, 1937). A more specific role for this area, including the paracingulate and medial orbital prefrontal cortex can be proposed as the experiencing of basic cognitive objects, that are later the input for further interpretation. For example, in the emotion domain, the medial prefrontal cortex may produce a basic experience of an emotional state, which is later interpreted, for example, as a feeling of being angry or of being afraid (cf. LeDoux, 1996). In the memory domain, it may produce a basic experience of an episode or event, which is later interpreted as ÔrememberedÕ, ÔknownÕ or ÔnewÕ by such processes as familiarity or recollection (cf. Whittlesea, 1993). These basic experiences and the later interpretive processes have been argued to be atypical in high-functioning autism (Ben Shalom, 2000, 2003; Ben Shalom et al., in press). 5. Autism and the experience of a perceptual object What kind of basic experience can the medial prefrontal cortex produce in the case of perceptual processing? One option is that it is the basic experience of a perceptual _object_. Consider the case of an (empty) plate, with a white color and a round shape. When typically developing people look at such a plate, all three representations (plate, white and round) are probably activated in the anterior temporal lobe. BarÕs and RollsÕ prefrontal mechanism may have two types of consequences for these representations: one, originally suggested by Bar (2004) and Rolls (2004) is the facilitation of the selection of the correct object representations (Bar) or assignment of the correct reward value to the object representation (Rolls). Another is the establishment of an order of importance in the anterior temporal lobe. By selecting or assigning a value to ÔplateÕ rather than to ÔroundÕ or to Ôwhite,Õ these mechanisms may help establish that the object is more _important_than its subparts and features. In sum, this commentary agrees with Sewards and Sewards (2002) that object recognition awareness relies on the anterior temporal lobe and the medial orbital prefrontal cortex. It suggests, however, that the medial orbital prefrontal cortex has a unique contribution, namely that of producing a basic experience of a perceptual _object_. It is further argued that the mechanisms that produce this experience also result in making the object more _important_than its subparts and features. Finally, it is argued that a reduction in this importance may account for some of the perceptual difficulties in high-functioning autism. 644 D.B. Shalom / Consciousness and Cognition 14 (2005) 641–644 References Ahissar, M., & Hochstein, S. (2000). The spread of attention and learning in feature search: effects of target distribution and task difficulty. Vision Research, 40, 1349–1364. American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.), Washington, DC: Author. Bar, M. (2004). Visual objects in context. Nature Reviews Neuroscience, 5, 617–629. Baron-Cohen, S., Ring, H. A., Bullmore, E. T., Wheelwright, S., Ashwin, C., & Williams, S. C. (2000). The amygdala theory of autism. Neuroscience and Biobehavioral Reviews, 24, 355–364. Ben Shalom, D. (2000). Developmental depersonalization: the prefrontal cortex and self-functions in autism. Consciousness and Cognition, 9, 457–460. Ben Shalom, D. (2003). Memory in autism: Review and synthesis. Cortex, 39, 1129–1138. Ben Shalom, D., Mostofsky, S. H., Hazlett, R. L., Goldberg, M. C., Landa, R. J., & Faran, Y. et al. (in press). Normal physiological emotions but differences in expression of conscious feelings in children with high-functioning autism. Journal of Autism and Developmental Disorders. Brock, J., Brown, C. C., Boucher, J., & Rippon, G. (2002). The temporal binding deficit hypothesis of autism. Development and Psychopatholgy, 14, 209–224. Frith, U. (1989). Autism: Explaining the enigma. Oxford: Blackwell. Hubel, D. N., & Weisel, T. N. (1962). Receptive fields, binocular interaction and functional architecture in the catÕs visual cortex. Journal of Physiology, 160, 106–154. LeDoux, J. E. (1996). The emotional brain. New York: Simon and Shuster. OÕNeill, M., & Jones, R. S. (1997). Sensory-perceptual abnormalities in autism: a case for more research?. Journal of Autism and Developmental Disorders, 27, 283–293. Papez, J. W. (1937). A proposed mechanism of emotion. Archives of Neurology and Psychiatry, 38, 725–734. Plaisted, K., Swettenham, J., & Rees, L. (1999). Children with autism show local precedence in a divided attention task and global precedence in a selective attention task. Journal of Child Psychology and Psychiatry, 40, 733–742. Rolls, E. T. (2004). Convergence of sensory systems in the orbitofrontal cortex in primates and brain design for emotion. The Anatomical Record A, 281, 1212–1225. Sewards, T. V., & Sewards, M. A. (2002). On the neural correlates of object recognition awareness: relationship to computational activities and activities mediating perceptual awareness. Consciousness and Cognition, 11, 51–77. Tanaka, K. (1996). Inferotepmoral cortex and object vision. Annual Review of Neuroscience, 19, 109–139. van Dalen, J. G. T. (1994). Autism: Seen from the Inside. Unplublished manuscript. Whittlesea, B. W. A. (1993). Illusions of familiarityg, Memory, and Cognition. Journal of Experimental Psychology: Learning, Memory, and Cognition, 19, 1235–1253. Williams, D. (2003). Not thinking in pictures. Available from www.donnawilliams.net/Articles/Autism/ NotThinkingInPictures.php.

Consciousness and Cognition xxx (2013) xxx–xxx Contents lists available at SciVerse ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Does infant behaviour provide support for the mirror neuron theory of action understanding? q Victoria Southgate Centre for Brain and Cognitive Development, Birkbeck College, Malet Street, London WC1E 7HX, United Kingdom a r t i c l e i n f o Article history: Available online xxxx Keywords: Action understanding Mirror neurons Infants EEG a b s t r a c t The mirror neuron theory of action understanding makes predictions concerning how the limited motor repertoire of young infants should impact on their ability to interpret others’ actions. In line with this theory, an increasing body of research has identified a correlation between infants’ abilities to perform an action, and their ability to interpret that action as goal-directed when performed by others. In this paper, I will argue that the infant data does by no means unequivocally support the mirror neuron theory of action understanding and that alternative interpretations of the data should be considered. Furthermore, some of this data can be better interpreted in terms of an alternative view, which holds that the role of the motor system in action perception is more likely to be one of enabling the observer to predict, after a goal has been identified, how that goal will be attained. Ó 2013 The Authors. Published by Elsevier Inc. All rights reserved. 1. Introduction Human infants provide a compelling opportunity to explore the hypothesis that the mirror neuron system plays a functional role in action understanding. While the term ‘action understanding’ is used in different ways throughout the literature on mirror neurons, its most common usage seems to be synonymous with ‘goal understanding’. That is, mirror neurons are proposed to enable the observer to infer the immediate target, or goal, of an action (Rizzolatti & Sinigaglia, 2010). Since ‘goal understanding’ is an ability that young infants are believed to possess, a promising test of the mirror neuron theory of action understanding is to ask whether the limited, but developing, motor capabilities of infants, influence their capacity to interpret others’ actions as goal-directed. The process of identifying the goal of observed movements is proposed to happen via a mechanism that directly matches an observed movement onto a pre-existing motor representation of that action in the observer (Rizzolatti, Fogassi, & Gallese, 2001), or which codes the goal of the action in motor terms (Rizzolatti & Sinigaglia, 2010). An infant lacking a motor representation of the observed movement (because they have never performed that action) would thus presumably have no motor representation onto which they could map that movement. Therefore, the hypothesis that the mirror neuron system plays a role in action understanding via such a mechanism makes clear predictions about how the motor limitations experienced by young infants might impact on their ability to make sense of others’ actions. As both types of sensory-motor transformation would require that the observer accesses a corresponding motor representation (either of the movements or the goal), an infant without motor experience with an action would not have access to a corresponding motor representation, and so should be unable to make sense of the observed action. Independent of work on mirror neurons, a body of work had already emerged which suggested a relationship between infants’ experience with an action and their ability to understand that action when performed by others. This work has since q This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial-No Derivative Works License, which permits non-commercial use, distribution, and reproduction in any medium, provided the original author and source are credited. E-mail address: v.southgate@bbk.ac.uk 1053-8100/$ - see front matter Ó 2013 The Authors. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.concog.2013.04.008 Please cite this article in press as: Southgate, V. Does infant behaviour provide support for the mirror neuron theory of action understanding? Consciousness and Cognition (2013), http://dx.doi.org/10.1016/j.concog.2013.04.008 2 V. Southgate / Consciousness and Cognition xxx (2013) xxx–xxx been cited by advocates of the mirror neuron theory of action understanding as support for their theory (e.g. Casile, Caggiano, & Ferrari, 2011), and has motivated numerous other infant studies aimed at confirming this hypothesis for the function of mirror neurons. Indeed, many infant studies have produced new data consistent with the hypothesis that mirror neurons support goal understanding. However, in the haste to promote the data as support for the theory, little attention has been given either to alternative interpretations of the data, or to other data which fail to support the theory. For example, despite the reported relationship between infants’ action experience and their goal understanding, numerous studies have demonstrated goal understanding for actions for which infants could not possibly recruit a corresponding motor representation. Rather than an attempt to understand the existence of both supporting and non-supporting data within any alternative framework, data inconsistent with the theory has tended to be ignored1. Thus, the primary aim of this paper is to critically examine the claim that infant data provide strong evidence for the mirror neuron theory of action understanding, and to situate infant data within an alternative theory of the function of the motor system during action observation. 2. Evidence from infants The claim that infant behaviour provides evidence for the mirror neuron theory of action understanding is based on the apparent relationship between infant’s understanding of goals and their own motor abilities. The initial impetus for this claim came from pre-existing data demonstrating that infants could better attribute a goal to others’ object-directed actions if those actions were performed by a human hand behaving in a typical way than if they were performed by a mechanical claw or stick (Woodward, 1998), or even a human hand behaving in an atypical fashion (e.g. approaching the object with the back rather than the front of the hand) (Woodward, 1999). The importance of experience was directly tested in a study by Sommerville and colleagues in which 3-month-old infants, who ordinarily do not reach for objects and do not appear to interpret others’ actions as goal-directed, were given training with ‘grasping’ objects. The authors placed Velcro gloves on infant’s hands which, when they swiped their hands in the vicinity of objects, would result in them inadvertently ‘picking up’ those objects. After this training, infants’ goal-attribution abilities were tested, and only those infants who had this training experience could interpret the action of a gloved-hand as goal-directed (Sommerville, Woodward, & Needham, 2005). While not aimed at testing the mirror neuron theory of action understanding, these findings are certainly compatible. If mirror neurons indeed lead to goal understanding by recruiting a corresponding motor representation, the presumed effect of training in 3-month-olds was to provide them with a motor representation onto which they could match the observed action and understand its goal. Following this, several studies aimed to test the hypothesis that infant’s goal understanding was dependent on their being able to exploit a motor representation, acquired via first-person experience, of whatever action they were observing. As a measure of goal understanding, Falck-Ytter and colleagues asked whether infants would evidence prediction of the outcome of an action commensurate with their abilities to perform that action (Falck-Ytter, Gredeback, & von Hofsten, 2006). Goal understanding was operationalized as eye movements arriving at the location of the outcome of the action before the outcome was achieved. Infants observed either a human hand repeatedly placing balls in a bucket, or self-propelled balls following the same trajectory into the bucket. The authors found that older infants, who would likely be able to place objects into containers themselves, did generate predictive saccades towards the goal of the observed action. However, younger infants who were unlikely to be able to place objects in containers themselves did not show anticipatory saccades, suggesting that they were not able to infer the goal of this action. Furthermore, neither older nor younger infants generated anticipatory saccades when the movements observed were executed by self-propelled balls. Subsequent work has extended these findings by confirming that the ability to place objects into containers is indeed correlated with the ability to generate anticipatory saccades towards the outcome of the action when performed by others (Cannon, Woodward, Gredeback, von Hofsten, & Turek, 2011), and other work has shown that the maturity of infants’ reaching is correlated with their ability to interpret a reach as goal-directed in others (Kanakogi & Itakura, 2011). These findings are interpreted as support for the mirror neuron theory of action understanding and hypothesized to reflect the importance of being able to access a motor representation of the observed action in order to understand that action as goal-directed. In what follows, I will suggest a number of reasons why the data from infants is not unequivocal support for the mirror neuron theory of action understanding and suggest that such an interpretation is premature in the absence of further data. 3. Action specificity While it has been shown that an infant’s experience and competence with an action is related to their ability to interpret that action when performed by others, the specificity of this relationship has not been demonstrated. That is, we do not know whether it is specifically experience with the observed action that is crucial, or whether motor maturity more generally might facilitate action prediction. We do not know whether it is experience placing objects in containers that enables infants to predict the outcome of someone else placing objects in containers (Cannon et al., 2011; Falck-Ytter, Gredeback, & von Hofsten, 2006), or whether those infants who can place objects in containers are also more adept at other motor skills, and 1 This data is often interpreted as resulting from an alternative, less important mechanism which results in a goal understanding that is inferior to that derived from a mirror mechanism (e.g. Gallese, Keysers, & Rizzolatti, 2004). Please cite this article in press as: Southgate, V. Does infant behaviour provide support for the mirror neuron theory of action understanding? Consciousness and Cognition (2013), http://dx.doi.org/10.1016/j.concog.2013.04.008 V. Southgate / Consciousness and Cognition xxx (2013) xxx–xxx 3 whether general motor maturity might be correlated with superior action prediction. In order for the correlation between action skill and predictive saccades during action observation to be evidence that predictive saccades reflect recruitment of corresponding motor plans, one would need to demonstrate that it is specifically skill with the observed action that leads to this improvement. 4. What is special about self-produced actions? Infants can acquire experience with actions either from performing those actions themselves, or, through observing those actions performed by others. Consistent with the mirror neuron theory of action understanding, the experience acquired through action production appears to be superior to that acquired through action observation alone (Sommerville, Hildebrand, & Crane, 2008). If action understanding requires access to a motor representation of that action, then experience performing rather than observing that action would be crucial. However, this only provides incontrovertible evidence for the mirror theory if we can be sure that it is the addition of a motor representation generated from self-produced actions that is responsible for infants being better able to understand these actions when performed by others. What else, apart from a motor representation, could self-produced actions provide that other-produced actions could not? One possibility is that the experience obtained from self-produced actions provides a better learning opportunity than that obtained from other-produced actions. For example, children learn the labels for objects better when labelled objects are being manipulated by themselves, compared to when they are being manipulated by someone else (Yu & Smith, 2012), and information that is received by infants in response to their requests appears to be better retained than unsolicited information (Begus, Gliga, & Southgate, in preparation). One could hypothesize various reasons as to why self-produced information is better assimilated than other-produced information. For example, infants, like adults, may be more likely to learn information that they had a hand in eliciting (Kang et al., 2009), or self-produced effects may be especially arousing for infants (e.g. Lewis, Sullivan and Brooks-Gunn, 1985). Whatever the reason, the point is that there could be numerous explanations for why actions, with which the infant has first-person experience, are more readily understood when performed by others, only one of which is the availability of a motor representation onto which the observed action can be mapped. 5. How goal understanding is operationalized Perhaps the most important reason why the infant data does not provide incontrovertible evidence for the mirror neuron theory of action understanding is that there are a number of reasons to question the assumption that what is taken as evidence of action understanding or a lack thereof, is really that. The interpretation of infant data as support for this theory is heavily dependent on the assumption that a particular behaviour reflects a goal attribution, and an absence of that behaviour reflects a failure to attribute a goal. The evidence discussed so far comes from two paradigms, one using infant looking-time as evidence of goal attribution and the other using infant predictive gaze. In the following sections, I will examine whether behaviour on these paradigms really provides the kind of evidence that the mirror neuron theory of action understanding requires. 5.1. The evidence from looking-time In the first paradigm (Woodward, 1998), infants observe an agent repeatedly choosing one of two objects and then, after the objects have switched location, either continuing to choose the same object (in a new location) or switching her choice to the previously un-chosen object. When this action is a familiar reaching action, infants typically spend more time looking at the action towards the new object than the action towards the old object, even though, because of the location switch, the action towards the old object entails a change in path from what infants had seen during familiarization. The fact that infants look longer at the new object-directed action suggests that they had encoded the familiarization action in terms of the goal object, rather than other features like location of reaching. However, although the assumption that infants’ looking behaviour reflects a goal attribution is rarely questioned, it is important to establish its validity, given that the argument that infants’ behaviour provides support for the mirror neuron theory of action understanding relies on this assumption. In a recent paper, Hernik and Southgate (2012) provide data that question the assumption that this paradigm reflects a goal attribution that could be generated by mirror neuron activity. One exception to infants’ expectation that an agent who has previously approached a particular target will continue to approach that target is if that target was initially approached in isolation, rather than as one of a pair of objects (Luo & Baillargeon, 2005, 2007). The accepted interpretation of this caveat is that, while infants interpreted the approach towards the solitary object as goal-directed, they simply do not know which one the agent will approach now that there is a new object in the scene. In the classic two-choice paradigm, the infant has evidence that the agent does not choose the ignored object, but in the solitary object condition, they have no information as to the agent’s disposition towards any new object, and so do not know whether she might actually prefer this one when it is available. The problem with this interpretation is that, as a mechanism of goal attribution, it would surely fail. One can imagine many situations where a goal attribution, which should endure the appearance of new potential targets for which the observer might lack information concerning the agent’s disposition (Hernik & Southgate, 2012), would be abandoned if such information was necessary. For example, imagine an agent (A) chasing another (B) down a busy street. Under the Luo and Please cite this article in press as: Southgate, V. Does infant behaviour provide support for the mirror neuron theory of action understanding? Consciousness and Cognition (2013), http://dx.doi.org/10.1016/j.concog.2013.04.008 4 V. Southgate / Consciousness and Cognition xxx (2013) xxx–xxx Baillargeon explanation, the appearance of all these new targets that comprise the busy street should lead the observer to abandon their goal attribution, and no longer expect A to continue chasing B. Thus, it is difficult to see how a mechanism which requires up-to-date dispositional information could support goal attribution. In fact, it turns out not to be the case that infants require knowledge of the agent’s disposition towards all possible targets in order for them to generate an expectation concerning which target the agent should approach. Another paradigm which proposes to demonstrate goal understanding in infants is based on the interpretation of efficient action (Gergely & Csibra, 2003). In this paradigm, infants observe an agent repeatedly jumping over an obstacle to approach a target. In test trials, the obstacle is removed and infants see the agent either approaching the target with the same detour jumping action, or executing a straight path to the target. Despite the fact that the straight path is perceptually novel relative to the jumping action, infants appear to expect the agent to execute a straight path towards the goal, looking longer towards the jumping agent whose action is inefficiently related to the goal (Gergely, Nadasdy, Csibra, & Biro, 1995). Thus, infants can exploit action efficiency as a cue to goal-directedness. Now, if we make the approach to a solitary object an efficient action (by having the agent detour around a barrier on its approach), infants then continue to expect the agent to approach the same target object even in the presence of a new object towards which the infant has no knowledge of the agent’s disposition (Hernik & Southgate, 2012). The fact that the infant continues to expect the agent to act on the same target irrespective of the arrival of new targets is consistent with the infant having attributed a goal to the agent. Similar effects have been reported by Biro (Biro, Verschoor, & Coenen, 2011) and when a different kind of alternative cue to goal-directedness is added to a solitary approach event (Luo, 2011). What this data suggests is that it is the presence of ‘choice’ (i.e. selecting one of two available objects) that is enabling infants to attribute a goal on the standard two-target paradigm, rather than simply the reach towards the object (as, in isolation, this does not seem to lead to an enduring goal attribution). That infants require this ‘choice’ or preference information in order to attribute an enduring goal presents a problem for those who wish to use infant’s behaviour as evidence in support of the mirror neuron theory of action understanding. If a direct reach to the object does not result in a goal attribution, but rather other cues like a selective reach, or efficient action, are necessary, then it is difficult to see how mirror neurons could be playing a role in infant’s goal attribution. Mirror neurons are hypothesized to enable the direct understanding of the action through a process of simulation, but in this paradigm, goal understanding appears to be mediated by the agent’s ‘choice’ or preference. 5.2. The evidence from eye tracking Infant’s failure to attribute goals to a direct approach in the absence of additional cues such as choice or efficiency of the action has implications for other data interpreted as support for the mirror neuron theory of action understanding. While the success of 12-month-olds in predicting the goal of a hand action but their failure to predict the goal of a self-propelled ball is consistent with a mechanism which requires access to a motor representation for goal understanding, it is also possible that the condition with self-propelled balls lacked sufficient cues that the action was goal-directed. The direct movement of the balls into the bucket is devoid of any cues to efficiency of action in the same way as a direct path to a solitary object (Hernik & Southgate, 2012). While those who interpret infant’s predictive gaze as support for the mirror neuron theory of action understanding would likely argue that such cues are also lacking in the hand condition, it may be that, by 12 months of age, placing actions are assumed to be goal-directed based on experience of seeing hands acting in goal-directed ways (Biro & Leslie, 2007). For unfamiliar agents like self-propelled balls, further cues to goal-directedness may be required – such as evidence that the ball’s actions are efficiently related to the goal. In fact, a recent study has demonstrated that if a self-propelled ball detours over an obstacle en route to its goal, then 13-month-olds do evidence predictive gaze shifts to the goal (Biro, in press). These data suggest that it may not be the absence of a motor representation that precludes goal attribution in the self-propelled ball condition, but an absence of cues that indicate the movement is goal-directed. However, even if these actions did contain sufficient cues for infants to interpret them as goal-directed, there are still reasons to question the assumption that predictive eye movements reflect a goal prediction achieved by a motor process. While the presence of predictive saccades have been cited as evidence in themselves of a motor contribution to goal understanding (Flanagan & Johansson, 2003), they may equally be mediated by conceptual knowledge of actions and their likely outcomes. For example, a commonly used paradigm in adults involves observers watching a hand approaching a small and a large object (Ambrosini, Costantini, & Sinigaglia, 2011; Costantini, Ambrosini, Cardellicchio & Sinigaglia, in press). When the hand is pre-shaped so that it is in a whole-hand grasp, observers generate saccades towards the large object in advance of the hand arriving at that object, but when the hand is closed into a fist, observer’s eye movements do not show evidence for prediction of the goal. The authors argue that the grasping hand provides motor information which presumably can be matched to a motor representation associated with large objects whereas the closed hand does not provide any motor information that would lead to the prediction that one object is more likely to be the target than the other. However, it is also likely that we have accumulated experience that hands in whole-hand grasps are more likely to be reaching for bigger objects than smaller objects, but it is unlikely that we would have any associated outcome for a hand approaching objects in a fist shape. There is no inherent reason why predictive saccades should be interpreted as supporting one hypothesis over the other. Nevertheless, studies do show that if the motor system is unavailable, even adults fail to make predictive saccades towards a familiar action outcome. For example, studies using Transcranial Magnetic Stimulation (TMS) have shown that if the motor system is rendered temporarily unavailable, adult participant’s predictive saccades to goal outcomes are impaired (Constantini, Ambrosini, Cardellicchio, & Sinigaglia, in press; Elsner, D’Ausilio, Gredeback, Falck-Ytter, & Fadiga, 2013). These Please cite this article in press as: Southgate, V. Does infant behaviour provide support for the mirror neuron theory of action understanding? Consciousness and Cognition (2013), http://dx.doi.org/10.1016/j.concog.2013.04.008 V. Southgate / Consciousness and Cognition xxx (2013) xxx–xxx 5 data have been interpreted as evidence that a process of direct matching, implemented in the motor system, is required for generating a goal prediction. However, while these data clearly do suggest a role for the motor system in predictive tasks, what kind of prediction it is involved in generating is ambiguous. The ability to predict the goal of an action also permits one to predict a likely way in which that goal will be fulfilled; the path that the action will take en route to this goal and the kinematics involved (Csibra & Gergely, 2007). In an action such as a direct reach for an object, the location of saccades reflecting a goal prediction and those reflecting a prediction about how that goal will be achieved would be the same. While saccadic data cannot distinguish these two possibilities, only if they reflected a process of goal prediction would they be support for the mirror theory of action understanding. The same criticism applies to the infant studies using predictive saccades as evidence for motor involvement since in all studies the path direction and the goal location are the same. While the presence of predictive saccades towards an action outcome surely implies that the observer has generated a goal prediction regardless of whether they reflect the process of goal prediction, an absence of predictive saccades does not imply that no goal prediction has taken place. In order to generate a predictive saccade, an observer would not only need to generate a prediction (concerning either the goal or the action path), but they would also need to disengage from whatever they were currently attending to. While action familiarity might facilitate predictive saccades because generating predictive saccades may require the availability of a motor program, action familiarity might also facilitate attention disengagement. Since infants receive both motor and visual feedback from self-produced actions (Del Giudice, Manera, & Keysers, 2009), actions that the infant can produce would be more visually familiar than actions that they cannot produce. In studies measuring predictive saccades, the test of success is the speed with which the observer generates saccades away from a salient moving stimulus (e.g. a hand, mechanical claw or self-propelled ball) towards the action end point. The ability to visually disengage from an interesting stimulus, thought to reflect endogenous attention control, continues to improve across the first year of life, with infants at 7 months being slower to disengage than infants at 14 months (Elsabbagh et al., in press). Thus, an alternative explanation for the difference in predictive saccades in the familiar and unfamiliar action conditions of the studies discussed above is that they reflect infant’s developing capacity for attention disengagement. While the human action is likely highly familiar to infants, the mechanical claw (or any other non-familiar action), by virtue of its novelty, may be more difficult for infants to disengage from, and thus more difficult to generate predictive saccades away from. Furthermore, 12-month-olds might simply be better at shifting their attention from a moving stimulus to a static one than 6-month-olds. A possible test would be to occlude a portion of the movement in both conditions. Occluding the salient movement element may make it easier for infants to visually disengage from the action and saccade towards the outcome. If predictive saccades result from access to a corresponding motor representation then occlusion of the movement should have no effect and the results should be the same. In all of the studies using predictive saccades as a measure of goal understanding, the observed action has been something that the infants cannot produce themselves, and is unlikely to have ever seen before (i.e. self-propelled balls and mechanical claws). Clues to whether it is the lack of visual or motor experience that results in the lack of predictive saccades, could be found in looking time studies. When the measure is looking-time, 6month-old infants appear to more readily attribute a goal to a walking human adult (Kamewari, Kato, Kanda, Ishiguro, & Hiraki, 2005), than to an animate box, or a mechanical claw (Woodward, 1998), even though both of these actions are outside of their motor repertoire. But while they would likely have accumulated plenty of visual experience with walking adults, they would be unlikely to have any prior visual experience with self-propelled boxes or mechanical claws. Finally, despite not generating predictive saccades towards the target of a mechanical claw (Cannon & Woodward, 2012; Kanakogi & Itakura, 2011), a different measure of action prediction suggests that infants do generate predictions for such unfamiliar actions. Based on a documented correlation between motor activity and action prediction (Kilner, Vargas, Duval, Blakemore, & Sirigu, 2004; Southgate, Johnson, Osborne, & Csibra, 2009), we asked whether infants would exhibit motor activation when they could expect either a human hand or a mechanical claw to act (Southgate & Begus, in press). Infants were familiarized to a hand or claw consistently approaching and ‘grasping’ one of two objects. After familiarization, infants saw the two objects presented individually at the top of a screen with either the hand or the claw resting at the bottom of the screen. We hypothesized that if infants could predict the goal of an action, we should see motor activation during this static period. We used EEG to look for a reduction in the sensorimotor alpha rhythm, indicating motor activation (Hari & Salmelin, 1997; Hari et al., 1998), during this rest period and found that infants recruited their motor system when they saw the target object, irrespective of whether the effector they saw was a hand or a mechanical claw. This suggests that even if infants fail to make anticipatory saccades when they observe a mechanical claw approaching a target object, they can still interpret these actions as target related. Furthermore, the presence of motor activation when infants see the target object suggests that they are using their motor system for some part of this process. In the following section, I will discuss how data like this – suggesting that infants recruit their motor system during a phase in which action can be predicted, irrespective of the familiarity of that action – can be interpreted. 6. What is the role of the motor system in action observation? I have argued that the oft-made assumption that the superiority of familiar actions for goal understanding in infancy implies that the motor system plays a causal role in this goal understanding is unwarranted, and that there are various alternative possible interpretations that should be considered. Furthermore, the wealth of evidence suggesting that infants can attribute goals to numerous actions for which they could not recruit a motor representation suggests that access to corre- Please cite this article in press as: Southgate, V. Does infant behaviour provide support for the mirror neuron theory of action understanding? Consciousness and Cognition (2013), http://dx.doi.org/10.1016/j.concog.2013.04.008 6 V. Southgate / Consciousness and Cognition xxx (2013) xxx–xxx sponding motor representations is unnecessary for goal understanding. An alternative to the mirror neuron theory of action understanding is the cue-based view (e.g. Biro & Leslie, 2007). As discussed throughout the preceding sections, researchers have identified a number of different cues which infants appear able to exploit in order to identify an action as goal-directed. These include that the action can be interpreted as efficiently related to an outcome, the presence of a choice or selection on the part of the agent, action variability and salient action effects (for further discussion see Hernik & Southgate, 2012). Goals are identified and understood based on the presence of these cues rather than the familiarity of the action. However, action familiarity may still influence goal attribution. As Biro and Leslie (2007) have noted, familiar human actions (like reaching or grasping) may become associated with goal-directed behaviour to the point where they become cues to goal-attribution. Furthermore, the ease with which one can infer a goal is likely to be influenced by experience with an action; certain actions are likely to become associated with certain outcomes (e.g. a hand in a pincer grip is likely usually followed by the picking up of a small object rather than a large object). While proponents of the mirror neuron theory of action understanding have conceded that these cues could be helpful for actions that are outside of the motor repertoire of the observer, for actions for which the observer could access a motor representation, this would be the preferred route and would result in a superior understanding of the action (Gallese et al., 2004). However, data which demonstrate the recruitment of the motor system during the observation of non-executable actions (e.g. Southgate & Begus, in press) is difficult to reconcile with this view. The motivation for positing a secondary, nonmotor route to goal understanding derives from the logical assumption that a mirror mechanism that matches form and motion could not be recruited for making sense of actions for which the observer would lack a corresponding motor representation. The fact that the motor system is recruited irrespective of action familiarity suggests that it is doing something quite different from matching the observed action onto a corresponding motor representation. Thus, an alternative explanation for the presence of motor activation during action observation is that it reflects a process of action anticipation, or prediction. According to this view, goals are identified outside of the motor system but, once they are, the motor system is recruited in order to predict how that goal will be achieved (Csibra, 2007; Jacob, 2009; Prinz, 2006), including predicting the unfolding kinematics of the observed action (Kilner, 2011). Generating an action prediction is not only crucial for basic needs like fleeing a predator for whom you have identified yourself as its target, but for successful engagement in cooperative and collaborative activities (Sebanz, Bekkering, & Knoblich, 2006). This hypothesis accounts for the fact that goal understanding itself appears possible without any reliance on corresponding motor representations, while at the same time explaining the well-documented involvement of the observer’s motor system during predictive tasks. For example, Schubotz and colleagues have shown that areas of the observer’s motor system are activated when they generate a wide range of predictions, even those which have no relation to action such as what tone will come next in a sequence (Schubotz, 2007; Schubotz & von Cramon, 2002), and other findings also strongly implicate the observer’s motor system in predictive tasks (Cross, Stadler, Parkinson, Schütz-Bosbach, & Prinz, 2011; Kilner et al., 2004; Southgate et al., 2009). Furthermore, this hypothesis fits well with the fact that the motor system functions in a predictive fashion during action execution (Miall & Wolpert, 1996), and thus its predictive involvement in action observation is a natural product of its inherently predictive modus operandi. A further hypothesis concerning how the motor system facilitates action prediction is that it operates in an emulative fashion. Once the observer knows the goal of the agent, they can ask how they themselves would fulfil this goal; what motor commands they would execute to this end, and it is this emulative process that recruits the motor system (Csibra, 2007). Insofar as we generally attempt to achieve our goals in the most efficient way possible, when the actor and observer share the same motor capabilities, the observer will likely emulate the effect via a motor program that matches that which the actor will actually use. However, under the emulation hypothesis, when the observer lacks the motor program that the actor is using, they may nevertheless recruit an alternative motor program that could bring about the same outcome that the observer has identified as being the goal of the action. For example, it has been shown that adults who are born without hands recruit their foot motor region when observing hand actions (Gazzola et al., 2007), a finding that is consistent with the hypothesis that the motor system is recruited to emulate the goal of an action. Similarly, a 6-month-old infant who lacks a motor program for walking could identify a goal such as ‘approach target’ or ‘contact target’ based on available cues and then recruit whatever motor program they themselves could use to bring about this goal (e.g. a motor program for crawling or reaching). This view of the role of the motor system is also consistent with numerous findings, including that the mirror system appears sensitive not to the effectors used to achieve a goal, but to the goal itself (Umiltà et al., 2008). If the role of the motor system in action interpretation were one of predicting action unfolding (post goal identification), the ease with which the goal can be ascribed would be crucial to the subsequent involvement of the motor system in prediction. For example, motor system involvement in action perception is modulated by the presence of a visible goal. The fact that observation of pantomimed actions does not lead to motor activation (Muthukumaraswamy, Johnson, & McNair, 2004) would be expected because, in the absence of a known outcome, the observer has no basis on which to predict how the action will unfold. Similarly, the ostensibly conflicting finding that infants recruit their motor system while observing a mechanical claw reaching for an object, suggesting that motor activation is not dependent on motor experience (Southgate & Begus, in press) and findings in adults that expertise with an observed action such as ballet (Calvo-Merino, 2004; Cross, Hamilton, & Grafton, 2006; Orgs, Dombrowski, Heil, & Jansen-Osmann, 2008) or piano playing (Haslinger et al., 2005) results in greater motor activation can be reconciled under the prediction hypothesis. While the visible presence of a goal in the infant study and the inclusion of known visual cues to goal attribution (see Hernik & Southgate, 2012 for further discussion of cues) could have Please cite this article in press as: Southgate, V. Does infant behaviour provide support for the mirror neuron theory of action understanding? Consciousness and Cognition (2013), http://dx.doi.org/10.1016/j.concog.2013.04.008 V. Southgate / Consciousness and Cognition xxx (2013) xxx–xxx 7 enabled infants to emulate the goal via an alternative motor program, only experts with knowledge of ballet sequences or piano pieces would be able to predict how the sequence would unfold, and thus we would expect to see a relationship between expertise and motor involvement, especially during the prediction of intransitive actions. The ease with which a goal can be identified and its effect on motor activation was also demonstrated in a recent study with infants (Southgate, Johnson, Karoui, & Csibra, 2010). In this study, infants were shown a grasping hand or an unfamiliar back-of-hand disappearing behind an occluder. Crucially, infants had no prior knowledge of what, if anything was behind the occluder. Infants exhibited motor activation in the case of the grasping hand, but not when the hand was in a back-of-hand posture. A control condition confirmed that this was not simply because the infants could match the kinematics of the grasping hand but not of the back-of-hand. Instead, a possible explanation is that, while the familiarity of the grasping hand permitted infants to generate a hypothesis about the likely outcome of the action (perhaps based on an association that they had acquired during the first 9 months of their life, in which they likely often saw grasping hands approaching objects), the unfamiliar back-of-hand action did not. Armed with this goal knowledge in the grasping hand condition, infants could then recruit their motor system to generate a prediction concerning how that grasping action would unfold in pursuit of the goal. In the absence of any goal knowledge for the back-of-hand condition, infants would have had no basis on which to generate a prediction concerning how the action would unfold, and thus they did not recruit their motor system. 7. Conclusions Much has been made of the documented correlations between infant’s motor skills and their performance on tasks of action understanding, in part because they ostensibly provide good evidence for the mirror neuron theory of action understanding. This paper has attempted to expose the limitations in the existing infant data as support for the mirror neuron theory of action understanding. In an attempt to find supporting evidence for this theory, findings which are consistent with the theory have been hastily assimilated without sufficient consideration of alternative explanations. Advocates of the theory that actions are understood by motor matching accept that we can make sense of actions outside of our motor experience, arguing that there is likely to be both a motor and a non-motor route to action understanding, but that a motor route offers a ‘richer’ understanding of the action (Gallese et al., 2004; Rizzolatti & Sinigaglia, 2010). However, there is no evidence that a ‘richer’ understanding of an action is attained via a motor route, and, given that both infants (Southgate & Begus, in press) and adults (Cross et al., 2011; Schubotz, 2007) recruit their motor system for actions both within and outside of their motor repertoire, it is not clear on what basis we should postulate two distinct routes to action understanding. After all, this secondary route was only suggested in response to the assumption that the motor system would not be recruited for actions that are outside the observer’s own motor history. The fact that infants can make sense of actions outside of their motor repertoire, and that they recruit their motor system during the perception of non-executable actions, fits better with the alternative view of action understanding. Under this view, goals are identified on the availability of various cues, only one of which might be action familiarity. However, once a goal is attributed, the motor system may be recruited to help the observer predict how the goal will be attained. As with the mirror neuron theory of goal understanding, this is essentially a simulative process. However, rather than simulating the observed movements in order to identify a goal, the observer identifies the goal and then simulates a way in which this goal might be fulfilled. Acknowledgment This work was supported by a Wellcome Trust Research Career Development Fellowship (088427/Z/09/Z). References Ambrosini, E., Costantini, M., & Sinigaglia, C. (2011). Grasping with the eyes. Journal of Neurophysiology, 106(3), 1437–1442. Begus, K., Gliga, T., & Southgate, V. (in preparation). Infants learn what they want to learn. Biro, S. (in press). The role of the efficiency of novel actions in infants’ goal anticipation. Journal of Experimental Child Psychology. http://dx.doi.org/10.1016/ j.jecp.2012.09.011. Biro, S., & Leslie, A. M. (2007). Infants’ perception of goal-directed actions: Development through cue-based bootstrapping. Developmental Science, 10(3), 379–398. http://dx.doi.org/10.1111/j.1467-7687.2006.00544.x. Biro, S., Verschoor, S., & Coenen, L. (2011). Evidence for a unitary goal concept in 12-month-old infants. Developmental Science, 14(6), 1255–1260. Calvo-Merino, B. (2004). Action observation and acquired motor skills: An fMRI study with expert dancers. Cerebral Cortex, 15(8), 1243–1249. http:// dx.doi.org/10.1093/cercor/bhi007. Cannon, E. N., & Woodward, A. L. (2012). Infants generate goal-based action predictions. Developmental Science, 15(2), 292–298. http://dx.doi.org/10.1111/ j.1467-7687.2011.01127.x. Cannon, E. N., Woodward, A. L., Gredeback, G., von Hofsten, C., & Turek, C. (2011). Action production influences 12-month-old infants’ attention to others’ actions. Developmental Science, 15(1), 35–42. http://dx.doi.org/10.1111/j.1467-7687.2011.01095.x. Casile, A., Caggiano, V., & Ferrari, P. F. (2011). The mirror neuron system: A fresh view. The Neuroscientist, 17, 524–538. Constantini, M., Ambrosini, E., Cardellicchio, P., & Sinigaglia, C. (in press). How your hand drives my eyes. Social Cognitive and Affective Neuroscience. http:// dx.doi.org/10.1093/scan/nst037. Cross, E. S., Stadler, W., Parkinson, J., Schütz-Bosbach, S., & Prinz, W. (2011). The influence of visual training on predicting complex action sequences. Human Brain Mapping. http://dx.doi.org/10.1002/hbm.21450. Cross, E. S., Hamilton, A. F. deC., & Grafton, S. T. (2006). Building a motor simulation de novo: Observation of dance by dancers. NeuroImage, 31(3), 1257–1267. http://dx.doi.org/10.1016/j.neuroimage.2006.01.033. Please cite this article in press as: Southgate, V. Does infant behaviour provide support for the mirror neuron theory of action understanding? Consciousness and Cognition (2013), http://dx.doi.org/10.1016/j.concog.2013.04.008 8 V. Southgate / Consciousness and Cognition xxx (2013) xxx–xxx Csibra, G. (2007). Action mirroring and action understanding: An alternative account. Sensorimotor foundations of higher cognition, 22, 427–451. Csibra, G., & Gergely, G. (2007). ‘Obsessed with goals’: Functions and mechanisms of teleological interpretation of actions in humans. Acta Psychologica, 124, 60–78. Del Giudice, M. D., Manera, V., & Keysers, C. (2009). Programmed to learn? The ontogeny of mirror neurons. Developmental Science, 12(2), 350–363. http:// dx.doi.org/10.1111/j.1467-7687.2008.00783.x. Elsabbagh M., Fernandes J., Webb S.J., Dawson G., Charman T. and Johnson M. H. (in press). Disengagement of visual attention in infancy is associated with emerging autism in toddlerhood. Biological Psychiatry , ISSN 0006-3223. Elsner, C., D’Ausilio, A., Gredeback, G., Falck-Ytter, T., & Fadiga, L. (2013). The motor cortex is causally related to predictive eye movements during action observation. Neuropsychologia, 51, 488–492. Falck-Ytter, T., Gredeback, G., & von Hofsten, C. (2006). Infants predict other people’s action goals. Nature Neuroscience, 9(7), 878–879. http://dx.doi.org/ 10.1038/nn1729. Flanagan, J. R., & Johansson, R. S. (2003). Action plans used in action observation. Nature, 424, 769–771. Gallese, V., Keysers, C., & Rizzolatti, G. (2004). A unifying view of the basis of social cognition. Trends in Cognitive Sciences, 8(9), 396–403. http://dx.doi.org/ 10.1016/j.tics.2004.07.002. Gazzola, V., van der Worp, H., Mulder, T., Wicker, B., Rizzolatti, G., & Keysers, C. (2007). Aplasics born without hands mirror the goal of hand actions with their feet. Current Biology, 17(14), 1235–1240. Gergely, G., & Csibra, G. (2003). Teleological reasoning in infancy: The naïve theory of rational action. Trends in Cognitive Sciences, 7, 287–292. Gergely, G., Nadasdy, Z., Csibra, G., & Biro, S. (1995). Taking the intentional stance at 12 months of age. Cognition, 1–29. Hari, R., Forss, N., Avikainen, S., Kirveskari, E., Salenius, S., & Rizzolatti, G. (1998). Activation of human primary motor cortex during action observation: A neuromagnetic study. Proceedings of the National Academy of Sciences, 95(25), 15061–15065. Hari, R., & Salmelin, R. (1997). Human cortical oscillations: A neuromagnetic view through the skull. Trends in Neurosciences, 20(1), 44–49. Haslinger, B., Erhard, P., Altenmüller, E., Schroeder, U., Boecker, H., & Ceballos-Baumann, A. O. (2005). Transmodal sensorimotor networks during action observation in professional pianists. Journal of Cognitive Neuroscience, 17(2), 282–293. Hernik, M., & Southgate, V. (2012). Nine-months-old infants do not need to know what the agent prefers in order to reason about its goals: On the role of preference and persistence in infants’ goal-attribution. Developmental Science, 15(5), 714–722. http://dx.doi.org/10.1111/j.1467-7687.2012.01151.x. Jacob, P. (2009). What do mirror neurons contribute to human social cognition? Mind & Language, 23(2), 190–223. Kamewari, K., Kato, M., Kanda, T., Ishiguro, H., & Hiraki, K. (2005). Six-and-a-half-month-old children positively attribute goals to human action and to humanoid-robot motion. Cognitive Development, 20(2), 303–320. http://dx.doi.org/10.1016/j.cogdev.2005.04.004. Kanakogi, Y., & Itakura, S. (2011). Developmental correspondence between action prediction and motor ability in early infancy. Nature Communications, 2, 341. http://dx.doi.org/10.1038/ncomms1342. Kang, M. J., Hsu, M., Krajbich, I. M., Loewenstein, G., McClure, S. M., Wang, J., et al (2009). The wick in the candle of learning: Epistemic curiosity activates reward circuitry and enhances memory. Psychological Science, 20(8), 963–973. Kilner, J. M. (2011). More than one pathway to action understanding. Trends in Cognitive Sciences, 15(8), 352–357. http://dx.doi.org/10.1016/ j.tics.2011.06.005. Kilner, J. M., Vargas, C., Duval, S., Blakemore, S.-J., & Sirigu, A. (2004). Motor activation prior to observation of a predicted movement. Nature Neuroscience, 7(12), 1299–1301. http://dx.doi.org/10.1038/nn1355. Lewis, M., Sullivan, M. W., & Brooks-Gunn, J. (2003). Emotional behaviour during the learning of a contingency in early infancy. British Journal of Developmental Psychology, 3(3), 307–316. http://dx.doi.org/10.1111/j.2044-835X.1985.tb00982.x. Luo, Y. (2011). Three-month-old infants attribute goals to a non-human agent. Developmental Science, 14(2), 453–460. Luo, Y., & Baillargeon, R. (2005). Can a self-propelled box have a goal? Psychological Science, 16(8), 601–608. Luo, Y., & Baillargeon, R. (2007). Do 12.5-month-old infants consider what objects others can see when interpreting their actions? Cognition, 105, 489–512. Miall, R. C., & Wolpert, D. M. (1996). Forward models for physiological motor control. Neural Networks, 9(8), 1265–1279. Muthukumaraswamy, S. D., Johnson, B. W., & McNair, N. A. (2004). Mu rhythm modulation during observation of an object-directed grasp. Cognitive Brain Research, 19(2), 195–201. http://dx.doi.org/10.1016/j.cogbrainres.2003.12.001. Orgs, G., Dombrowski, J.-H., Heil, M., & Jansen-Osmann, P. (2008). Expertise in dance modulates alphabeta event-related desynchronization during action observation. European Journal of Neuroscience, 27(12), 3380–3384. http://dx.doi.org/10.1111/j.1460-9568.2008.06271.x. Prinz, W. (2006). What re-enactment earns us. Cerebral Cortex, 42, 515–517. Rizzolatti, G., Fogassi, L., & Gallese, V. (2001). Neurophysiological mechanisms underlying the understanding and imitation of action. Nature Reviews Neuroscience, 2(9), 661–669. Rizzolatti, G., & Sinigaglia, C. (2010). The functional role of the parieto-frontal mirror circuit: Interpretations and misinterpretations, 1–11. http://dx.doi.org/ 10.1038/nrn2805. Schubotz, R. I. (2007). Prediction of external events with our motor system: Towards a new framework. Trends in Cognitive Sciences, 11(5), 211–218. http:// dx.doi.org/10.1016/j.tics.2007.02.006. Schubotz, R. I., & von Cramon, D. Y. (2002). Predicting perceptual events activates corresponding motor schemes in lateral premotor cortex: An fMRI study. NeuroImage, 15(4), 787–796. http://dx.doi.org/10.1006/nimg.2001.1043. Sebanz, N., Bekkering, H., & Knoblich, G. (2006). Joint action: Bodies and minds moving together. Trends in Cognitive Sciences, 10(2), 70–76. Sommerville, J. A., Hildebrand, E. A., & Crane, C. C. (2008). Experience matters: The impact of doing versus watching on infants’ subsequent perception of tool-use events. Developmental Psychology, 44(5), 1249–1256. http://dx.doi.org/10.1037/a0012296. Sommerville, J. A., Woodward, A. L., & Needham, A. (2005). Action experience alters 3-month-old infants ‘‘perception of others’’ actions. Cognition, 96(1), B1–B11. http://dx.doi.org/10.1016/j.cognition.2004.07.004. Southgate, V., & Begus, K. (in press). Motor activation during the prediction of non-executable actions in infants. Perspectives on Psychological Science. Southgate, V., Johnson, M., Karoui, E. l., & Csibra, G. (2010). Motor system activation reveals infants‘‘ on-line prediction of others’’ goals. Psychological Science, 21(3), 355–359. Southgate, V., Johnson, M. H., Osborne, T., & Csibra, G. (2009). Predictive motor activation during action observation in human infants. Biology Letters, 5(6), 769–772. http://dx.doi.org/10.1098/rsbl.2009.0474. Umiltà, M. A., Intskirveli, I., Grammont, F., Rochat, M., Caruana, F., Jezzini, A., et al (2008). When pliers become fingers in the monkey motor system. Proceedings of the National Academy of Sciences, 105(6), 2209–2213. Woodward, A. (1998). Infants selectively encode the goal object of an actor’s reach. Cognition, 69, 1–34. Woodward, A. (1999). Infants’ ability to distinguish between purposeful and non-purposeful behaviors. Infant Behavior and Development, 22(2), 145–160. Yu, C., & Smith, L. B. (2012). Embodied attention and word learning by toddlers. Cognition, 125(2), 244–262. http://dx.doi.org/10.1016/ j.cognition.2012.06.016. Please cite this article in press as: Southgate, V. Does infant behaviour provide support for the mirror neuron theory of action understanding? Consciousness and Cognition (2013), http://dx.doi.org/10.1016/j.concog.2013.04.008

Consciousness and Cognition 22 (2013) 1013–1021 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog A differentiating empirical linguistic analysis of dreamer activity in reports of EEG-controlled REM-dreams and hypnagogic hallucinations Jana Speth a,⇑, Clemens Frenzel b, Ursula Voss c,d a Dpt. of Psychology, University of Bonn, Germany Dpt. of Neuroendocrinology, University of Luebeck, Germany c Dpt. of Psychology, University of Frankfurt, Germany d Vitos Waldkrankenhaus Köppern, Germany b a r t i c l e i n f o Article history: Received 29 December 2012 Available online 7 August 2013 Keywords: Hypnagogia REM-dream Dream reports Dream consciousness EEG Sleep stage Linguistic analysis Simulated activity level a b s t r a c t We present Activity Analysis as a new method for the quantification of subjective reports of altered states of consciousness with regard to the indicated level of simulated motor activity. Empirical linguistic activity analysis was conducted with dream reports conceived immediately after EEG-controlled periods of hypnagogic hallucinations and REM-sleep in the sleep laboratory. Reports of REM-dreams exhibited a significantly higher level of simulated physical dreamer activity, while hypnagogic hallucinations appear to be experienced mostly from the point of passive observer. This study lays the groundwork for clinical research on the level of simulated activity in pathologically altered states of subjective experience, for example in the REM-dreams of clinically depressed patients, or in intrusions and dreams of patients diagnosed with PTSD. Ó 2013 Elsevier Inc. All rights reserved. 1. Introduction An empirically founded description of the characteristic phenomenology of hallucinatory worlds experienced by dreamers in different states of conscious awareness forms a cornerstone of the comprehensive definition of dreaming – at which modern dream research is striving to arrive. The nomological network of dreaming consists of hypnagogic and hypnapompic hallucinations, of REM-sleep dreams, lucid REM-dreams, and NREM-sleep dreams. While it is nowadays fairly easy to document the physiological concomitants of different conscious states, the distinguishing features of the subjective dimension of dreaming still have to be determined. The study of phenomenological aspects of hybrid dream states like hypnagogia or lucid dreaming, with two levels of consciousness prevailing simultaneously, allows us to determine the relationship between different parts of the nomological network of dreaming. We can then seek to find pathological alterations in the hallucinations of patients with certain psychological disorders, and, ultimately, indications for treatment. Dream reports attest to mental events conceived in a distinct state of consciousness. If we want to investigate sleeping consciousness, we need to take special care in devising our tools. We want to introduce a new method to read out the black box of subjective dream consciousness in an objective, quantitative way: The empirical linguistic analysis of dream reports. ⇑ Corresponding author. Present address: Bruesseler Str. 23, 53117 Bonn, Germany. E-mail address: jana.speth@uni-bonn.de (J. Speth). 1053-8100/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.concog.2013.07.003 1014 J. Speth et al. / Consciousness and Cognition 22 (2013) 1013–1021 A large variety of linguistic theories and definitions describe how psychophenomenological characteristics of our mental processes – such as perspective(s) (Sanders & Spooren, 1997; Wiebe, 1994; Scheibman, 2002), time conception (Comrie, 1985), forms of reflective thinking (Fauconnier, 1994, 1997), or the (perceived) self-perception and self-efficacy that is agency (Reinhart, 2002) – become expressed in language in a rule-governed process, for example through definable grammatical constructions, or in semantic patterns. Such linguistic theories therefore contain an implicit or explicit cognitive theory. Linguistic dream reports, obtained after awakening, attest to mental events conceived in a different, distinct state of consciousness. Abnormalities in the frequency of specific linguistic constructions should thus refer to specific characteristics of this distinct state of consciousness. Here, we present first results of an exploratory study based on the systematic linguistic analysis of subjects’ reports describing hallucinatory experiences during hypnagogia and REM-sleep dreams. Our laboratory-based study focuses on the level of dreamer1 activity in hallucinations associated with hypnagogia and REM-dreams, such as simulated running, jumping, or moving objects. In order to measure the level of dreamer activity, we devised and employed a new tool based on linguistic theta theory. We hypothesized that reports of REM-dreams show a higher level of activity – indicated by a higher ratio of linguistic expressions of agency – than reports of hypnagogic hallucinations. In order to test this hypothesis, we collected and compared transcribed oral reports from healthy subjects after forced awakenings from sleep stage 1 and REM-sleep. 1.1. Dream consciousness We distinguish between different types of dreams based on three main interdependent criteria: The neurophysiologically distinct sleep or pre-sleep stage in which they occur (Iber, Ancoli-Israel, Chesson, & Quan, 2007; Rechtschaffen & Kales, 1968), the level of dreamer awareness (Voss, Holzmann, Tuin, & Hobson, 2009; J.A. Hobson & U. Voss, 2010; A. Hobson & U. Voss, 2010), and further formal phenomenological aspects such as bizarreness, thematic coherence or emotion (Stickgold, Rittenhouse, & Hobson, 1994; Sutton et al., 1994; J.A. Hobson & U. Voss, 2010; A. Hobson & U. Voss, 2010; Voss, SchermellehEngel, Windt, Frenzel, & Hobson, 2013). Physiologically speaking, hypnagogic hallucinations are experienced in a resting state, during the transition to sleep. Researchers agree that hypnagogic hallucinations are by definition a sleep stage 1 phenomenon (Vaitl et al., 2005): a state of a low voltage, mixed frequency EEG accompanied by slow eye movement usually terminated by sleep stage 2 (Rechtschaffen & Kales, 1968). Here, mostly visual, non-emotional hallucinations occur in conjunction with decreasing alpha activity (Kuhlo & Lehmann, 1964). In contrast, REM sleep more frequently occurs in the 2nd half of the night, concurring with further decrease in muscle tone, and rapid binocular movement (Iber et al., 2007). One reason why the phenomenon of hypnagogic hallucinations has recently begun to re-attract scientific interest (e.g. Kussé, Shaffil-Le Bourdiec, Schrouff, Matarazzo, & Maquet, 2012) is the fact that this hybrid state shows characteristics of both waking and dreaming consciousness. In a laboratory setting, the subject proves to be still aware of being partly awake, insofar as he or she is usually able to communicate with the researchers via pre-arranged and EEG-observable eye movements. With respect to reflective consciousness, hypnagogia appears, at first sight, similar to another dream phenomenon: lucid dreaming, a hybrid state with characteristics of both waking and dreaming consciousness (J.A. Hobson & U. Voss, 2010; J.A. Hobson & U. Voss, 2010). In lucid dreaming, the dreamer is aware of the fact that he or she is dreaming. With sufficient training, the dreamer is able to signal lucidity to the investigator through a distinct pattern of eye movement (Voss et al., 2009). Despite similarities in dream awareness however, we have reason to assume that lucid dreaming and hypnagogic hallucinations are not the same phenomenon: Whereas, per definitionem, lucid dreams arise from REM-sleep, and are a state in which the dreamer remains physiologically asleep (Voss et al., 2009), hypnagogic hallucinations occur prior to sleep-onset (Vaitl et al., 2005). This study neither aims at a physiological comparison, nor at a comparison of dream awareness to describe the differences of REM-dreams and hypnagogic hallucinations – but will take a third approach by focusing on the phenomenological differences of both states of consciousness. The phenomenology of REM-dream consciousness draws from the self-stimulation of the brain (Hobson, 2009). Mostly isolated from environmental stimuli (Rechtschaffen, 1978) by an input/output gating mechanism controlled by the brain stem (Pompeiano, 1967), the dreaming brain relies on endogenous stimulation for fictive sensual perception and fictive motor activity as the dreaming self interacts with its dream world (Hobson, 2009). Dreaming has hallucinatory qualities in so far as hallucinations are characterized by a vivid sense of reality without an adequate external source (American Psychiatric Association, 1994). Considering the different background of conscious awareness against which hypnagogic events and REM sleep dreams occur, common sense would predict that we find some crucial differences between the phenomenology of hypnagogic hallucinations and of REM-sleep. During REM hallucinations, the subject is asleep and non-responsive; he or she appears to be fully enclosed in the dream world. Hypnagogic hallucinations are also called ‘‘hypnagogic images’’, which indicates that they 1 In this text, the term ‘‘dreamer’’ refers to the experiencer of (REM-)dreams as well as to the experiencer of hypnagogia and other hallucinations. J. Speth et al. / Consciousness and Cognition 22 (2013) 1013–1021 1015 may be experienced from a more passive point of view (as suggested by Ardis & McKellar, 1956; McKellar & Simpson, 1954; Oswald, 1962). They occur while the subject is still physiologically awake and partly responsive. It appears that during hypnagogic hallucinations, the subject is not entirely lost in reverie, not fully enclosed in a dream world, and thus possibly not as active a participant in hallucinated dream scenarios. Surprisingly, previous studies indicate fundamental similarity of content between REM-dreams and hypnagogic hallucinations (Foulkes, Spear, & Symonds, 1966; Foulkes & Vogel, 1965). Foulkes & Vogel (1965) find reports of hypnagogia to be quite similar to REM-dream reports with respect to mental organization and mental functioning. In spite of their findings however, Foulkes et al. (1966) express their confidence that the content of hypnagogia may ultimately prove discriminable from dream content associated with REM sleep, preserving the ‘‘undisturbed [...] association of REM sleep with a unique class of mental fantasy’’ (Foulkes et al., 1966: 280). The current study employs a new method, dreamer activity analysis, which is derived from applied linguistic analysis. Systematic analyses of dream reports based on this method point towards phenomenological differences between REM-dreams and hypnagogia, and thus suggest that Foulkes’ foresight was indeed justified. 1.2. Linguistic analysis Applied linguistic analysis refers to the scientific study of a sample of natural language. Analysis can be carried out with a grammatical (morphological, syntactical), semantic, phonological, or pragmatic focus. It supplies us with information on how humans use language to convey information, to express what is on their minds. Verbalization of mental events can only happen during waking consciousness, but language can still tell us about the subjective experiences of dreamers. We can investigate in how far characteristics of dream phenomenology associated with different brain states – such as REM sleep and hypnagogic hallucinations – can be distinguished with the help of subjects’ verbal dream reports. With regard to a variety of dream research questions, empirical linguistic analysis may close the gap between existing approaches to the analysis of dream phenomenology: It combines the advantages of the high objectivity of standardized or even automatic counting systems (e.g. Amini, Sabourin, & De Koninck, 2011; Bulkeley, 2009; Domhoff & Schneider, 2008) with the flexibility of human rating and scoring (e.g. Fosse, Stickgold, & Hobson, 2001) – a flexibility that is necessary when we are working with language in the context of subjective verbal reports. The empirical linguistic analysis of dream reports wants to advance the tools of formal analysis by establishing clear rules for the linguistic testing of research questions which until now appeared to be approachable only with content analytical methods (e.g. Hall & Van de Castle, 1966). It speaks for the validity of the methods of empirical linguistic analysis of dream reports that they can be based on linguistic, or even cognitive-linguistic theories which contain implicit or even explicit assumptions on how certain mental events are systematically expressed in natural language. In the context of many dream studies, analysis systems are only developed in an ad hoc fashion and with regard to a specific research question (e.g. Bértolo et al., 2003; Hartmann & Brezler, 2008; Mulder, Hochstenbach, Dijkstra, & Geertzen, 2008). In contrast to this, empirical linguistic analysis can yield tools that are variably applicable for different investigations of characteristics of hallucinatory states. As these tools may be found useful for different studies, they could be continuously validated and refined. The empirical linguistic analysis of dream reports could offer several methodical advantages that may prove useful to modern dream research. First of all, the analysis of dream reports with regard to specific linguistic characteristics follows clear rules: Through the synergy of psychological or physiological theories with specific linguistic theories we can make informed predictions on how psychophenomenological characteristics inherent to dream consciousness find their expression in objectively definable linguistic constructs in the dream report. Raters of the reports can identify and quantify these linguistic constructs. We can formulate basic requirements to any linguistic method for the empirical analysis of psychophenomenological events: (a) Deduction: The hypotheses that are to be tested are derived from psychological/neurophysiological theories as well as from established linguistic models. (b) Rule-governed system: The linguistic method is based on an explicit assumption on how mental events become systematically expressed in natural language. (c) Empiricism and quantifiability: Hypotheses are empirical, and the respective linguistic constructions are countable. (d) Presence: The linguistic constructs that are to be tested occur in such frequency in natural language that they can be statistically captured in a limited data base. (e) Applicability: The respective rules can be outlined clearly and comprehensively for the raters of the studies. The rules can be applied easily and economically during the rating process. On many occasions, questionnaires may yield important insight on dream consciousness (Voss, Frenzel, Koppehele-Gossel, & Hobson, 2012; Voss et al., 2013) and dream phenomenology (Pace-Schott et al., 2001; Takeuchi, Ogilvie, Ferrelli, Murphy, & Belicki, 2001). However, with regard to various research questions, the linguistic analysis of the dream report may bring additional advantages at minimal cost. First, even complex details of dream phenomenology could thus be analyzed with relatively high qualitative and quantitative accuracy. When determining characteristics of dream phenomenology we do not have to rely solely on the – probably 1016 J. Speth et al. / Consciousness and Cognition 22 (2013) 1013–1021 differing – reflective phenomenological understandings of subjects just awoken from REM-sleep, and neither on their explicit numerical estimation of various dream details. Second, as we simply have to ask subjects one general question, to tell us about their dreams, we prevent different demand effects and systematic distortions of results, which is conducive to research objectivity. Third, dream reports that were collected upon this principle present a cornucopia of information on a variety of different psychophenomenological aspects of the dreaming experience, and therefore allow for research efficiency. In the study presented here, we applied a grammatical and semantic approach based on linguistic theta theory to the empirical analysis of reports from hypnagogia and from REM-sleep dreams, as verified by EEG recordings. We expected characteristics of the two brain states to express themselves in the ratio of distinct linguistic features. In applying the activity analysis method, our interest was the degree of simulated activity in the two states of consciousness. We compared the degree to which hypnagogic hallucinations and REM-sleep dreams are experienced from the perspective of passive observer, versus as active participant – who is actively involved in dream activity, i.e. carrying out (simulated) physical activity such as running, flying or moving objects within his or her dream world. In linguistic theta system theory, the agent, the initiator of an event, takes on a specific thematic (theta) role within a sentence or phrase. Other theta roles are: Agent, cause, experiencer, instrument, and theme (Field, 2004; Reinhart, 2002). The agent is usually described by a noun phrase, but the agent does not necessarily correlate with the grammatical subject of a clause: The agent is defined through his or her relationship to the activity (described by the predicate). The following example shows that the agent can be the subject (i), while it does not have to be (ii): (i) Jim opens the box. (ii) The box is opened by Jim. Jim is the agent in both (i) and (ii), but subject only in (i). If we want to identify instances of dreamer activity in reports of REM-dreams and of hypnagogic hallucinations, we need to get behind the surface structure of the phrases and describe the constituents of the sentences in terms of their theta roles. More than that, the quantitative analysis of agency rate in reports has to be carried out in connection with careful agency analysis: We need not only determine the agent, but also determine if the agent represents the dreamer, as opposed to another dream figure. This study focuses exclusively on such instances of agency in reports of hypnagogic hallucinations and of REM-dreams where the dreamer describes him- or herself as the agent, the initiator of observable, physical, dynamic action. 2. Materials and methods A total of 26 healthy subjects (15 female, 11 male) between 19 and 28 years of age were recruited from the student population of Bonn University. Recruitment was conducted via written announcements in the university building and online at the department’s website as well as via oral announcements during lectures. The reports were conceived in the sleep laboratory of the University of Bonn. Informed consent on methods, side effects and aims of the study was obtained prior to the investigation. We paid 50 € compensation. 2.1. EEG In accordance with standard procedures, EEG (SOMNOscreen™ plus) was recorded from 6 channels, placed at Fp1, Fp2, F3, F4, Pz, and O1. Electrodes were referenced to linked mastoids. Bandpass filter was set at 0.3–70 Hz; sampling rate was 265 Hz. EOG was taken from the outer canthi of both eyes and supraorbitally to the left eye. Submental EMG electrodes were fixed at the chin. Waking and REM sleep EEG was scored online according to the criteria of the AASM (2007). 2.2. Design As shown in Fig. 1, subjects reported to the laboratory at around 9 pm. Subjects were then introduced to the concept of hypnagogic hallucinations as fictive, possibly dream-like experiences. Ample time was allowed to ask questions. To signal Fig. 1. Schedule of data acquisition. 1017 J. Speth et al. / Consciousness and Cognition 22 (2013) 1013–1021 hypnagogic hallucinations, an eye movement feedback similar to the one used in lucid dreaming studies (Voss et al., 2009) was explained and trained while subjects were still fully awake. Reports of hypnagogic hallucinations were collected at the onset of sleep. The subject was asked to react with a distinct eye movement pattern (two times left–right–left) upon the appearance of the first hypnagogic hallucinations, and to repeat the pattern at as regular intervals as possible for as long as the hallucinations continued. As soon as the ocular feedback was paused for more than 2 min and/or the investigator detected sleep stage epiphenomena indicating the onset of stage 2 sleep, the investigator entered the room to interview the subject. For REM sleep dream report collection, participants were then allowed to sleep without interruptions until the second half of the night, when REM sleep is most likely to occur. REM sleep reports were obtained after forced awakenings from the first REM period after 3 am. In both conditions, the subject was then asked to orally report all impressions as detailed as possible. The oral report was audio-taped and later transcribed for further analysis. The subject then stayed the rest of the night in the laboratory without further interruptions of sleep. Reports of hypnagogic hallucinations and REM-dreams were collected from different subjects during different nights. 2.3. Data analysis The transcribed records were analyzed according to level of agency by two blind independent German native speakers. Instances of initiated observable, physical, dynamic (– static) agency were identified, counted, and classified. Instances were classified as dreamer agency (‘‘dreamer_agency’’) whenever described from a first person point of view. Instances were classified as agency of other dream characters when occurring in the grammatical second or third person (‘‘other_agency’’). Instances of grammatical second person occur frequently on occasions when subjects dream about the researchers in the sleep laboratory in which they are just being tested, and, in giving an oral report to the investigator, call him or her ‘‘you’’ – as in ‘‘And then I saw you coming up the stairs to the laboratory. . .’’). As the act of speaking is also, strictly speaking, partly observable physical activity (jaw movements, etc.), instances of talking were counted as agency whenever there was an identifiable agent (as opposed to mere disembodied voices), but counted separately (‘‘dreamer_talk’’, ‘‘other_talk’’). Additionally appearances of dream characters other than the dreamer were counted (‘‘other_single’’ for single individuals, ‘‘other_group’’ for groups e.g. crowds consisting of an undefined number of individuals). Reports that were shorter than 50 words of dream-related content were excluded because they typically consisted more of meta-description than of actual analyzable dream description, such as ‘‘Tonight I only had a very short dream, which I do not remember very well [...]’’. This length criterion is consistent with that of previous dream content studies (Hobson, PaceSchott, & Stickgold, 2000; Stickgold et al., 1994; Sutton, Rittenhouse, Pace-Schott, Stickgold, & Hobson, 1994; Voss, Tuin, Schermelleh-Engel, & Hobson, 2011). Upon pre-arrangement, any plural agency which included the dreamer (‘‘I am running through the forest with my friends. . . we arrive at a little house. . .’’) was counted as dreamer agency. Repetitions were counted separately (for example, ‘‘I am running and running and running’’ would be counted as 3 instances of dreamer agency). As a peculiarity of dreams and hallucinations, animals, fantastic figures, and clearly animated objects can also take the role of agent, and of speaking agent, and were classified as such in our study (see Table 1). 3. Results A total of 59 reports of both hypnagogic images and REM-dreams were collected. We eliminated 12 reports because they were lacking content and consisted only of meta-descriptions of the dream experience or general context-framing discourse not relevant to our study. Both independent raters agreed on which reports to eliminate. Those reports which were considered devoid of relevant dream content descriptions were entirely congruent with the cut-off criterion of 50 word of dream related content. In the event of 5 REM-periods, and after 1 hypnagogic stage, subjects claimed to have no recall at all. A number of 40 reports (26 from female and 14 from male subjects) remained to be analyzed (19 reports of hypnagogic hallucinations; 21 of REM-dreams). Those reports had a mean length of 197.5 words (SD = 152); reports of hypnagogic images had a Table 1 Agency classification with linguistic examples. Linguistic occurrence Type of agency Example 1st person action 2nd and 3rd person action 1st person speaking 2nd and 3rd person speaking Disembodied voice Dreamer and others take action together Repetitions dreamer_agency other_agency dreamer_talk other_talk Not counted dreamer_agency Counted separately I ran You ran; She ran; They ran I said: ‘‘Run!’’ You said: ‘‘Run!’’, He said ‘‘Run!’’ I heard: ‘‘Run!’’; A voice said [...] We ran I ran and ran and ran [. . .] 1018 J. Speth et al. / Consciousness and Cognition 22 (2013) 1013–1021 Fig. 2. Mean of other (non-dreamer) characters, groups of other characters, and types of agency per report. Not corrected for word length of report. mean length of 191.2 words (SD = 89.8), and reports of REM-dreams had a mean length of 203.2 words (SD = 194.2). The groups did not differ significantly with regard to word length (see Fig. 2). The ratings of both rates correlated highly (r = .813 for other_single, r = .733 for other_group, r = .934 for dreamer_agency, r = .802 for other_agency, r = .905 for dreamer_talk and r = .77 for other talk – all significant for p < .01). For further analysis, the mean of every pair of ratings was calculated. A one-way Analysis of Variance (ANOVA) for the factor condition (hypnagogic vs. REM) revealed a significant difference for dreamer agency (F = 16.775, df = 1, 38, p < .001, g2 = .306), other_agency (F = 10.866, df = 1, 38, p < .01, g2 = .222), dreamer_talk (F = 10.866, df = 1, 38, p < .05, g2 = .1) and other_talk (F = 6.84, df = 1, 38, p < .05, g2 = .152). There were more elements of dreamer agency, of agency of others, of talking by the dreamer, and of talking of others in REM-dream reports than in reports of hypnagogic images. The amount of single dream characters or groups of characters did not differ (Please see Table 2). Further, the numbers of elements were corrected for the word length of the reports and then analyzed as above. These results did not differ from the results for uncorrected values. An ANOVA for the factor condition revealed a significant difference for dreamer_agency (F = 23.608, df = 1, 38, p < .001, g2 = .383), other_agency (F = 9.797, df = 1, 38, p < .01, g2 = .204), dreamer_talk (F = 8.708, df = 1, 38, p < .01, g2 = .186), and other_talk (F = 4.295, df = 1, 38, p < .5, g2 = .102). Again, the amount of dream characters did not differ. To investigate the relationship between the different categories, the variables (corrected for word length) were Pearson cross correlated (Please see Table 3). Only other_single/other_talk (r = .341, p < .05), dreamer_agency/dreamer_talk (r = .324, p < .05) and other_agency/other_talk (r = .389, p < .05) correlated significantly. 4. Discussion Our exploratory study of a relatively small number of dream reports indicates that reports of hypnagogic hallucinations exhibit a significantly lesser degree of dreamer agency than reports of REM-dreams. The hypothesized difference between reports of hypnagogic hallucinations and REM-dreams can indeed be considered a difference specifically on the agency level, as there was neither a difference in word length, nor a difference in the general number of other dream characters to be observed between both kinds of report. Both raters also agreed with respect to instances of speaking in the reports: The dreamer seems to engage in imagined verbal interaction significantly more often in REM-dreams than during hypnagogic hallucinations. More instances of talking were found for both dreamer and other dream characters in REM-dreams, which had not been predicted. Since talking is physical, observable and dynamic action – and can therefore be regarded as a subclass of dreamer agency – this finding does not contradict the original hypothesis. Dream characters other than the one who is linguistically linked to the first person perspective – and therefore the identity – of the dreamer also show a higher level of general observable, physical dynamic agency, as well as of speaking agency, in reports of REM-dreams than in reports of hypnagogic hallucinations. An indirect possible explanation for this phenomenon is that action is often interaction: The concept of communication for example usually requires both sender and receiver, and the concept of fighting likewise involves twoway effects. A direct possible explanation is that, of course, dream characters other than the one assuming first person perspective also pose an extension of the dreamer’s self – coming, after all, from the dreamer’s brain – and are thus more likely to be likewise fictively-physically animated when the dreamer’s neuronal motor patterns are activated. 1019 J. Speth et al. / Consciousness and Cognition 22 (2013) 1013–1021 Table 2 Descriptive statistics of other (non-dreamer) characters, groups of other characters, and types of agency per report. Not corrected for word length of report. other_single other_group dreamer_agency other_agency dreamer_talk other_talk Hypnagogic Mean N SD 3.32 19 2.45 1.53 19 2.04 .79 19 1.36 1 19 1.05 .21 19 .71 .316 19 .75 REM Mean N SD 4.24 21 3.87 1.48 21 2.18 5.43 21 4.76 4.14 21 4.03 1.43 21 2.48 1.9 21 2.55 Total Mean N SD 3.8 40 3.27 1.5 40 2.09 3.26 40 4.24 2.65 40 3.37 .85 40 1.94 1.15 40 2.06 Table 3 Correlations between other (non-dreamer) characters, groups of other characters, and types of agency. Corrected for word length of report. other_single other_group dreamer_agency other_agency dreamer_talk other_talk * .244 .197 .272 .164 .341* other_group .041 .059 .128 .132 dreamer_agency .062 .324* .201 other_agency dreamer_talk .026 .389* .292 p < .05. We have several reasons to assume that these data are not merely the result of differences in the subjects’ competence or compliance: First, systematic differences between the conditions only occurred for the predicted agency constructions and for instances of talking; the amount of dream characters did not differ. Second, correcting for length of the dream reports did not change the results. Third, if we correct for length, only some measured variables are intercorrelated, namely other_single/other_talk (the more other characters appear, the more likely is the dreamer to talk to someone), dreamer_agency/dreamer_talk (the more dreamer activity, the more the dreamer is talking), and other_agency/other_talk (the more other characters are acting, the more other characters are talking). We find these correlations intuitively plausible. However there appears to be no general tendency that some subjects report all variables more often. 4.1. Implications The method of quantitative linguistic analysis of agency in reports of REM-dreams and related states of consciousness promises to be reliable, economical, and relatively easy to apply. Insofar, activity analysis may be a recommendable tool for the investigation of reports of subjective experience, for example in a developmental or a clinical setting. Activity analysis could help to shed some light on fictive dream movement in REM-dreams. Although movement is not reported in all REMdreams, and is not exclusively reported in REM-dreams, it has been associated with REM-dreaming: Virtual motor action as a feature especially of REM-dreaming (Porte & Hobson, 1996) has been linked to the activation of sensorimotor and motor patterns in the brain (Hobson, Pace-Schott, Stickgold, & Kahn, 1998) and is thus part of the activation axis of the AIM model(Hobson et al., 2000). Further research will show if reports of children’s REM-dreams exhibit a higher level of dreamer agency during crucial stages of motor skill and respective cognitive development (Diamond, 2000). It can also be investigated if pathologies connected to a lower degree of daytime activity, such as major depressive disorder, correlate with a lower agency level in patients’ REM-dreams, or if we can observe specific patterns of simulated dreamer activity in reports of patients diagnosed with PTSD. Linguistic dreamer activity analysis can thus contribute to a better understanding of the subjective dimension of dreaming and other hallucinatory states, possibly even in connection with specific clinical disorders. 4.2. Limitations We receive information on the subjective state of hypnagogia and REM-sleep consciousness solely from introspective reports, which are nearly always written from a first person perspective. The reports are ideally obtained in a laboratory setting, and immediately upon forced awakening toward the end of the specific sleep stage determined via EEG measuring. In the case of hypnagogic hallucinations, the beginning of sleep stage 2 marks the end of the hypnagogic period. Intentional recall of dream events can only happen in a different brain state, during waking consciousness. We must thus expect a certain loss of information: Recall is often poor and possibly inaccurate, due to quickly vanishing memory traces. The verbalization of dream events in a state of waking consciousness brings with it problems of its own: Subjects may, voluntarily or involuntarily, decide to leave out or add information in an attempt to establish coherence, or in order to 1020 J. Speth et al. / Consciousness and Cognition 22 (2013) 1013–1021 conceal intimate details. The use of certain illustrating linguistic expressions during the process of verbalization (e.g. figures of speech, especially metaphors and similes) can further cloud the picture. For research-economical reasons, there was no acclimation night for the participants. While we tried to keep our explanation of hypnagogic hallucinations as general and brief as possible when instructing our participants, we cannot completely exclude the possibility of demand effects arising from the instruction to describe sleep onset experiences. The 50 words of dream-related content-criterion could be a further limitation. By excluding 12 reports we might systematically overlook aspects of dream phenomenology that are specifically covered by these short reports with nearly no detectable reference to the actual dream experience. However we have reason to assume that the criterion is valid: First, the criterion is well established in report-based dream research, and thus enhances the comparability of our findings with other findings in the field. Second, in our database we found the category of <50 words dream reports to be completely congruent with the category of dream reports both raters agreed to be devoid of analyzable dream related content. 4.3. Conclusions Differences in the phenomenology of dream consciousness can be investigated by means of empirical linguistic analysis of reports of subjective experience. In our study we developed and employed linguistic Activity Analysis to show that reports of REM-sleep dreams differ significantly from reports of hypnagogic hallucinations with regard to the level of simulated dreamer activity. The tool of activity analysis proved to be reliable, easy to apply, and effective. It thus revealed differences in the phenomenology of hallucinations with different physiological characteristics. In contrast to those of hypnagogic hallucinations, reports of REM-dreams indicate a higher level of simulated dreamer activity. While our study focused on sleep-onset and REM-dreams exclusively, we also expect the analysis of reports of other dissociative states such as isolated sleep paralysis (Takeuchi, Miyasita, Sasaki, Inugami, & Fukuda, 1992) or lucid dreaming (Voss et al., 2009) to be rewarding. We are planning to apply Activity Analysis to further altered states of consciousness in order to shed more light onto the subjective dimensions of the nomological network of dreaming and hallucinations as altered states of consciousness. References Amini, R., Sabourin, C., & De Koninck, J. (2011). Word associations contribute to machine learning in automatic scoring of degree of emotional tones in dream reports. Consciousness and Cognition, 20, 1570–1576. American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC. Ardis, J. A., & McKellar, P. (1956). Hypnagogic imagery and mescaline. Journal of Mental Science, 102, 22–29. Bértolo, H., Paiva, T., Pessoa, L., Mestre, T., Marques, R., & Santos, R. (2003). Visual dream content, graphical representation and EEG alpha activity in congenitally blind subjects. Cognitive Brain Research, 15, 277–284. Bulkeley, K. (2009). Seeking patterns in dream content: A systematic approach to word searches. Consciousness and Cognition, 18, 905–916. Comrie, B. (1985). Tense. Cambridge University Press. Diamond, A. (2000). Close interrelation of motor development and cognitive development and of the derebellum and prefrontal cortex. Child Development, 71, 44–56. Domhoff, G. W., & Schneider, A. (2008). Studying dream content using the archive and search engine on DreamBank. Net, Consciousness and Cognition, 17, 1238–1247. Fauconnier, G. (1994). Mental spaces: Aspects of meaning construction in natural language. Cambridge University Press. Fauconnier, G. (1997). Mappings in thought and language. Cambridge University Press. Field, J. (2004). Psycholinguistics. London: Routledge. Fosse, R., Stickgold, R., & Hobson, J. A. (2001). The mind in REM sleep: Reports of emotional experience. Sleep, 24, 947–955. Foulkes, D., Spear, P. S., & Symonds, J. D. (1966). Individual differences in mental activity at sleep onset. Journal of Abnormal Psychology, 71, 280–286. Foulkes, D., & Vogel, G. (1965). Mental activity at sleep onset. Journal of Abnormal Psychology, 70, 231–243. Hall, C. S., & Van de Castle, R. L. (1966). The content analysis of dreams. New York: Appleton-Century-Crofts. Hartmann, E., & Brezler, T. (2008). A systematic change in dreams after 9/11/01. Sleep, 31, 213. Hobson, J. A. (2009). REM sleep and dreaming: Towards a theory of protoconsciousness. Nature Reviews Neuroscience, 10, 803–813. Hobson, J. A., Pace-Schott, E. F., & Stickgold, R. (2000). Dreaming and the brain: Toward a cognitive neuroscience of conscious states. Behavioral and Brain Sciences, 23, 793–842. Hobson, J. A., Pace-Schott, E. F., Stickgold, R., & Kahn, D. (1998). To dream or not to dream? Relevant data from new neuroimaging and electrophysiological studies. Current Opinion in Neurobiology, 8, 239–244. Hobson, A., & Voss, U. (2010). A mind to go out of: Reflections on primary and secondary consciousness. Consciousness and Cognition, 20, 993–997. Hobson, J. A., & Voss, U. (2010). Lucid dreaming and the Bimodality of Consciousness. In E. Perry, D. Collerton, F. E. N. LeBeau, & H. Ashton (Eds.), Towards new horizons in consciousness research from the boundaries of the brain (pp. 155–165). Amsterdam: John Benjamins Publishing. Iber, C., Ancoli-Israel, S., Chesson, A., & Quan, S. F. (2007). The AASM manual for the scoring of sleep and associated events: Rules, terminology and technical specifications. Westchester, IL: American Academy of Sleep Medicine. Kuhlo, W., & Lehmann, D. (1964). Das einschlaferleben und seine neurophysiologische korrelate. Archiv für Psychiatrie und Zeitschrift für die Gesamte Neurologie, 205, 687–716. Kussé, C., Shaffil-Le Bourdiec, A., Schrouff, J., Matarazzo, L., & Maquet, P. (2012). Experience-dependent induction of hypnagogic images during daytime naps: A combined behavioural and EEG study. Journal of Sleep Research, 21, 10–20. McKellar, P., & Simpson, L. (1954). Between wakefulness and sleep: Hypnagogic imagery. British Journal of Psychology, 45, 266–276. Mulder, T., Hochstenbach, J., Dijkstra, P. U., & Geertzen, J. H. B. (2008). Born to adapt, but not in your dreams. Consciousness and Cognition, 17, 1266–1271. Oswald, I. (1962). Induction of illusory and hallucinatory voices with considerations of behaviour therapy. Journal of Mental Science, 108, 196–212. Pace-Schott, E. F., Gersh, T., Silvestri, R., Stickgold, R., Salzman, C., & Hobson, J. A. (2001). SSRI treatment suppresses dream recall frequency but increases subjective dream intensity in normal subjects. Journal of Sleep Research, 10, 129–142. Pompeiano, O. (1967). The neurophysiological mechanisms of the postural and motor events during desynchronized sleep. Research Publications – Association for Researcher in Nervous and Mental Disease, 45, 351–423. Porte, H. S., & Hobson, J. A. (1996). Physical motion in dreams: one measure of three theories. Journal of Abnormal Psychology, 105, 329–335. Rechtschaffen, A. (1978). The single-mindedness and isolation of dreams. Sleep, 1, 97–109. J. Speth et al. / Consciousness and Cognition 22 (2013) 1013–1021 1021 Rechtschaffen, A., & Kales, A. (Eds.). (1968). A manual of standardized terminology, techniques and scoring system for sleep stages of human subjects. Bethesda: HEW Neurological Information Network. Reinhart, T. (2002). The theta system. Theoretical Linguistics, 28, 229–290. Sanders, J., & Spooren, W. (1997). Perspective, subjectivity, and modality from a cognitive linguistic point of view. In Discourse and perspective in cognitive linguistics. Amsterdam: John Benjamins Publishing. Scheibman, J. (2002). Point of view and grammar: Structural patterns of subjectivity. In American English conversation. Philadelphia: John Benjamins Publishing. Stickgold, R., Rittenhouse, C. D., & Hobson, J. A. (1994). Dream splicing: A new technique for assessing thematic coherence in subjective reports of mental activity. Consciousness and Cognition, 3, 114–128. Sutton, J. P., Rittenhouse, C. D., Pace-Schott, E., Stickgold, R., & Hobson, J. A. (1994). A new approach to dream bizarreness: Graphing continuity and discontinuity of visual attention in narrative reports. Consciousness and Cognition, 3, 61–88. Takeuchi, T., Miyasita, A., Sasaki, Y., Inugami, M., & Fukuda, K. (1992). Isolated sleep paralysis elicited by sleep interruption. Sleep, 15, 217–225. Takeuchi, T., Ogilvie, R. D., Ferrelli, A. V., Murphy, T. I., & Belicki, K. (2001). The dream property scale: An exploratory English version. Consciousness and Cognition, 10, 341–355. Vaitl, D., Birbaumer, N., Gruzelier, J., Jamieson, G. A., Kotchoubey, B., Kubler, A., et al (2005). Psychobiology of altered states of consciousness. Psychological Bulletin, 131, 98–127. Voss, U., Frenzel, C., Koppehele-Gossel, J., & Hobson, J. A. (2012). Lucid dreaming: An age dependent brain dissociation. Journal of Sleep Research, 21(6), 634–642. Voss, U., Holzmann, R., Tuin, I., & Hobson, J. A. (2009). Lucid dreaming: A state of consciousness with features of both waking and non-lucid dreaming. Sleep, 32, 1191–1200. Voss, U., Schermelleh-Engel, K., Windt, J., Frenzel, C., & Hobson, J. A. (2013). Measuring consciousness in dreams: The lucidity and consciousness in dreams scale. Consciousness and Cognition, 22, 8–21. Voss, U., Tuin, I., Schermelleh-Engel, K., & Hobson, J. A. (2011). Waking and dreaming: Related but structurally independent. Dream reports of congenitally paraplegic and deaf-mute persons. Consciousness and Cognition, 20, 673–687. Wiebe, J. M (1994). Tracking point of view in narrative computational linguistics, vol. 20. MIT Press, pp. 233–287.

Consciousness and Cognition 19 (2010) 778–801 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Psychogenic amnesia – A malady of the constricted self q Angelica Staniloiu a,*, Hans J. Markowitsch a,b, Matthias Brand c,d a Physiological Psychology, University of Bielefeld, Bielefeld, Germany Alfried Krupp Institute for Advanced Study, Greifswald, Greifswald, Germany General Psychology: Cognition, University of Duisburg-Essen, Germany d Erwin L. Hahn Institute for Magnetic Resonance Imaging, Essen, Germany b c a r t i c l e i n f o Article history: Available online 23 July 2010 Keywords: Dissociative amnesia Inferolateral prefrontal cortex Self-consciousness Emotion Autobiographical–episodic memory (AEM) a b s t r a c t Autobiographical–episodic memory is the conjunction of subjective time, autonoetic consciousness and the experiencing self. Understanding the neural correlates of autobiographical–episodic memory might therefore be essential for shedding light on the neurobiology underlying the experience of being an autonoetic self. In this contribution we illustrate the intimate relationship between autobiographical–episodic memory and self by reviewing the clinical and neuropsychological features and brain functional imaging correlates of psychogenic amnesia – a condition that is usually characterized by severely impaired retrograde memory functioning, in absence of structural brain damage as detected by standard imaging. We demonstrate that in this disorder the autobiographical–episodic memory deficits do not exist in isolation, but occur with impairments of the autonoetic self-consciousness, emotional processing, and theory of mind or executive functions. Furthermore functional and metabolic brain alterations involving regions that are agreed upon to exert crucial roles in memory processes were frequently found to accompany the psychogenic memory ‘‘loss”. Ó 2010 Elsevier Inc. All rights reserved. 1. Introduction ‘‘Memory connects innumerable single phenomena into a whole, and just as the body would be scattered like dust in countless atoms if the attraction of matter did not hold it together so consciousness – without the connecting power of memory – would fall apart in as many fragments as it contains moments.” (Ewald Hering, 1870/1895, p. 12). ‘‘Memory requires more than mere dating of a fact in the past. It must be dated in my past. In other words, I must think that I directly experienced its occurrence.” (William James, 1890/1950, p. 650). Autobiographical–episodic memory (AEM) has a close connection to self and self understanding, which is concisely captured by Tulving’s (2005) last definition of AEM. According to this, AEM could be viewed as a threefold cord that results from the uniting of subjective time, autonoetic consciousness and the experiencing self (Tulving, 2005). As William James’ quotation suggests, the appearance of AEM on the one hand, reflects and occurs in the context of the attainment of a new stage of self understanding and self awareness (Nelson, 2003). On the other hand, the emergence of AEM supports further self development and the capacity – that is crucial to at least highly individualized societies – for maintaining a consistent feeling of identity and a coherent awareness of self‘s continuity over time (Nelson, 2003, 2005; Newen & Schlicht, 2009; Vogeley & Kupke, 2007; Welzer & Markowitsch, 2005). q This article is part of a special issue of this journal on Self, Other and Memory. * Corresponding author. Address: Physiological Psychology, University of Bielefeld, P.O.B. 100131, D-33501 Bielefeld, Germany. E-mail address: astaniloiu@uni-bielefeld.de (A. Staniloiu). 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.06.024 A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 779 In the field of clinical psychiatry and psychoanalysis the role of AEM for one’s sense of self coherence and ‘‘trans-temporal sameness” (Metzinger, 2008), well-being or capacity to reconstruct oneself has been emphasized since Freud and ‘‘beyond”, despite differences in the working models of the self, which ranged from a vertically-organized self, with areas of conflict ‘‘buried” by repression to a horizontally aligned and sequenced view of self, with territories of incompatibility separated by dissociation (S.A. Mitchell & Black, 1995). Patients with psychogenic amnesia have been described for more than a century, providing evidence for the significant influence of social environment on AEM (cf. Markowitsch, 1992). In this review, after defining psychogenic amnesia, we outline the memory systems and the main brain structures that are involved in AEM and the processes necessary for building an episodic memory. We then illustrate the ways in and by which stress-related and psychological factors could lead to both a severe compromise of AEM and disturbances of several functions underlying the experience of being a self. 2. What and where (in the international classifications of diseases) is psychogenic amnesia? Borrowed from Greek (Pearce, 2007) the word amnesia describes the most severe form of memory impairment and refers to an inability to learn new information or recall previously learned information. The term is nowadays used to refer to a symptom of a disorder, a syndrome or a specific disease. In the latter case, its employment is congruent with the traditional view of amnesia as a memory impairment that occurs in an alert, responsive person in the absence of (or out of proportion in comparison to) other significant cognitive impairments, being subsequently restricted to specific disorders. These disorders are named amnesic (ICD-10, 1992) or amnestic (DSM-IV-TR, 2000) disorders. They have as a core feature a memory impairment that is not due to dementia, has functional relevance and represents a decline from a previously attained level of function. Based on their preponderant etiological link amnes(t)ic disorders are categorized as being due to a general medical condition (e.g. neurological event), direct effects of substances (e.g. alcohol) persisting beyond the period of intoxication or withdrawal or psychological factors. The term psychogenic amnesia has traditionally been used to describe episodes of retrograde and/or anterograde (AEM) memory loss, which are precipitated by psychological stresses and occur in the absence of identifiable brain damage. The memory impairment in psychogenic amnesia has classically been regarded to be reversible (Loewenstein, 1991), with older studies reporting a high percentage of recovery within a month from the onset of amnesia (Kanzer, 1939). Newer studies that employed extensive neuropsychological testing have however identified a prolonged course of the memory impairments in a substantial number of patients with psychogenic memory loss (Kritchevsky, Chang, & Squire, 2004; Markowitsch, Thiel, Kessler, von Stockhausen, & Heiss, 1997). Apart from psychogenic, other terminologies (such as hysterical, dissociative, functional or medically unexplained or mnestic block syndrome) have over the years been employed to capture the category of amnesic disorders without direct evidence of significant brain damage on conventional structural imaging techniques. In the current main official classifications of diseases (DSM-IV-TR, 2000; ICD-10, 1992) earlier diagnostic designations of hysterical or psychogenic amnesia are now preponderantly subsumed under the diagnostic categories of dissociative disorders in DSM-IV-TR and dissociative (conversion) disorders in ICD-10, but also under other diagnostic subcategories such as somatization disorder (in DSM-IV-TR only), post-traumatic stress disorder and acute stress disorder (DSM-IV-TR and ICD-10). Despite its gradual vanishing from the international nomenclatures of diseases, the terminology psychogenic amnesia has however survived in the scientific literature for a number of reasons. As opposed to the current ICD-10 and DSM-IV-TR diagnostic subcategory of dissociative amnesia, the term psychogenic amnesia is more comprehensive, being employed to encompass more than one diagnostic entity (see below). In addition, in contrast to the designation dissociative amnesia, which carries a specific theoretical load, the term psychogenic amnesia conveys the preponderant etiological link of the amnesic condition to psychological factors, without making a priori assumptions about the nature of the psychological mechanisms involved (McKay & Kopelman, 2009).This is of importance given that not only dissociation (defined by Janet [1907, p. 23] as ‘‘an inability of the personal self to bind together the various mental components in an integrated whole under its control”), but several other psychological mechanisms, such as hyper-suppression or cognitive avoidance (Fujiwara et al., 2008; Lemogne et al., 2009; Tramoni et al., 2009) might concurrently contribute to the psychogenic memory loss. While the term amnesia was used rather loosely until the 1980s, being equalized with ‘global amnesic syndrome’, studies of patients with amnesic disorders of neurological or psychogenic nature have provided evidence that the memory impairment is more likely to involve certain kinds of memory (in particular the memory for personal events). Therefore any employment of the term amnesia should nowadays be accompanied by a clear description of the type and severity of the memory problems involved, which presupposes an understanding of the current main classifications of the memory systems and memory processes. 3. Autobiographical–episodic memory system and its place within memory systems Memory is not unitary, but is divided along a chronological and content axis, respectively. Along the time axis, memory is classified in short-term and long-term memory. The term short-term memory has been employed to describe the online holding of information such as telephone numbers. It has a limited capacity of a few bits (4–7) (G.A. Cowan, 2001; Miller, 1956) and encompasses a time range of seconds to minutes. Any information that is not lost and exceeds the limited capacity 780 A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 of short-term and working memory is assigned to long-term memory stores. The time-related dichotomy of memory was further refined by the introduction of the term ‘working memory’ by Alan Baddeley (Baddeley, 2000; Baddeley & Hitch, 1974). As captured by its name, working memory refers to working with memory – this involves not only time-limited online holding of new information, but also retrieving portions of old, already stored information. Further time-related categorizations of memories involve distinctions between old and new and anterograde and retrograde memories, respectively (Fig. 1). The compromised ability to access information that happened before the memoryimpairing incident corresponds to retrograde memory impairment, while anterograde memory impairment refers to the compromised capacity to long-term acquire new information after the incident. Anterograde amnesia can occur with minimal or no retrograde amnesia. Though traditional accounts of amnesia had led to the assumption that retrograde amnesia should be always accompanied by anterograde amnesia, cases of isolated retrograde amnesia have been reported after either neurological or psychological incidents. In some cases of isolated retrograde amnesia following a neurological insult (e.g. severe traumatic brain injury), significant anterograde amnesia was present initially, but then resolved or became subtle (Levine, Svoboda, Turner, Mandic, & Mackey, 2009; Levine et al., 1998). The etiological mechanisms involved in isolated or disproportionate retrograde amnesia for personal events are still a source of debate, with some authors arguing that, even in the cases with clear evidence of brain pathology psychological factors might make substantial contribution to the presence of residual, disproportionate retrograde AEM loss (Kopelman, 2000, 2008). The categorization of memory along the content dimension has evolved based on data from patients with different types of memory impairments, corresponding to different types of brain lesions, neuroimaging studies of patients with memory impairments or normal subjects, animal memory research and human developmental studies. In the 1970s and 1980s a revival of early conceptualizations of memory subdivisions (Schneider, 1912, 1928; Semon, 1904) was proposed in the field of human memory research by Tulving (1972, 1983) and in the arena of animal research by Mortimer Mishkin (Mishkin & Petri, 1984). Mishkin distinguished between ‘habit’ and a ‘memory’ system; the habit system referred to procedures and routines, whereas the memory system was concerned with the acquisition of facts and relations between objects. Tulving initially differentiated between semantic and episodic memory, implying that semantic memory refers to general knowledge, while episodic memory refers to single episodes with a specific embedding in time and locus. Later, Tulving and other researchers expanded the categorization of memory systems, in particular by adding those systems involving automatic, implicit and subconscious levels of processing – such as procedural memory and the priming system. Procedural memory refers to highly automated sensory-motor skills, which include complex motor acts like driving a car, riding a bike, or skiing. Priming describes the higher probability of (‘‘automatically”) choosing a stimulus which was perceived earlier in the same or a similar manner. The ‘perceptual memory system’ was identified as a legitimate distinct system relatively recently. In contrast to the priming and procedural memory systems, this system acts ‘consciously’, but on a presemantic level and relies on familiarity judgments. An example is the conscious identification of an apple without hesitance, no matter what color it has or whether it is already half eaten or not. Patients with semantic dementia, who lose the capabilities for language and semantic memory, may still be able to distinguish for example an apple from a peach or pear without the need to access semantic information, by accessing perceptual representations of information via the perceptual memory system. Fig. 2 outlines these memory systems, accordingly to their assumed (Tulving, 2005) ontogenetic and phylogenetic hierarchy (starting with the simplest memory systems) and points to differences in their neuroanatomical substrates. The conceptualizations of episodic memory have undergone several revisions over the years (Tulving, 1972, 1983, 1985, 1995, 2000, 2002, 2005). While a few decades ago the term episodic could be applied to describe memory for laboratory stimuli with a specific embedding in time and place (Tulving, 1972), currently the episodic memory system is viewed as equivalent to the AEM system and AEM is defined as the conjunction of subjective time, autonoetic consciousness and the experiencing self (Tulving, 2005). One characteristic of AEM is mental time traveling through subjective time from Fig. 1. Relations between anterograde and retrograde amnesia. The flash symbol represents the time point of a brain infarct or of a major psychic traumatic event, leading to either anterograde or retrograde amnesia or to both. Note that for retrograde amnesia the frequently observed gradient – termed Ribot’s law (cf. Markowitsch, 2008), is indicated by stating that usually very old memories are preserved in retrograde amnesia, while those close to the point of the event are impaired. A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 781 Fig. 2. The five long-term memory systems and their assumed brain bases (for further description see the text). present to both past and future. Piolino, Desgranges, and Eustache (2009) suggest that mental time traveling is the last feature of AEM that becomes fully functional, but the first feature that is affected in most amnesic conditions. In comparison to semantic memory that is accompanied by noetic consciousness only, AEM requires a higher level of self-consciousness – autonoetic consciousness (Markowitsch, 2003). The terms autonoetic [‘‘self-conscious”, or ‘‘self-aware”], noetic [‘‘aware”] and anoetic [‘‘not-aware”] were introduced and elaborated on by Tulving (1995). Autonoetic consciousness was defined as ‘‘capacity that allows adult humans to mentally represent and to become aware of their protracted existence across subjective time” (Wheeler, Stuss, & Tulving, 1997, p. 335). Although the designations autobiographical and episodic are sometimes used interchangeably, not all autobiographical memories are, however episodic. A distinction should therefore be made between autobiographical–episodic and autobiographical–semantic memory. The latter refers to personal knowledge, such as one’s name or date of birth or trait self-knowledge, and might be preserved, relearned or updated despite blocked access to AEM and even in the presence of certain impairments of semantic knowledge of impersonal facts (Klein & Gangi, 2010; Markowitsch, Calabrese, Neufeld, Gehlen, & Durwen, 1999). This latter observation led to hypothesizing dissociations within the semantic memory system (Klein & Gangi, 2010). A significant amount of general evaluative self-knowledge is available from autobiographical–semantic memory, especially in adulthood (Pfeifer, Dapretto, & Lieberman, 2010), while during childhood and adolescence self-evaluations seem to have a stronger connection to AEM (Newen et al., 2009; Oddo et al., 2010; Pfeifer et al., 2010). This may explain why patients who suffer AEM impairments, but continue to have an intact autobiographical–semantic memory might be able to maintain some aspects of personal identity (Rathbone, Moulin, & Conway, 2009). Although both AEM and autobiographical–semantic memory are involved in sustaining personal identity, the maintenance of an enduring personal identity in a frequently changing and challenging environment most likely requires periodical and persuasive update and confirmation, which draw on emotionally-laden episodic memories of personal experience. Among memory systems, the AEM system develops the latest ontogenetically (Nelson & Fivush, 2004; Piolino et al., 2007; Tulving, 2005) and is arguably uniquely human (Clayton & Dickinson, 1998; Tulving, 2005). In comparison to other memory systems, AEM is more vulnerable to neuronal alterations (Markowitsch, 2008; Tulving, 2005). AEM system is also susceptible to distortions, misinformation and dynamic reshaping (Loftus, 2005). Already in 1896 Freud noted that ‘‘the material present in the form of memory traces [is] subjected from time to time to a rearrangement” (Masson, 1985, p. 207). Furthermore, Freud (1901) remarked that there is in general no guarantee for the correctness of our memory; nevertheless we much more frequently than is justified assume that we can trust its information. The recent research on false memory syndromes supports Freud’s early insight (Loftus, 2005). In a combined behavioral and neuroimaging study it was found that university students made roughly 45% errors when judging whether a shown scene had or had not been included in two short movies (Kühnel, Woermann, Mertens, & Markowitsch, 2008). These results were reflected in the same study by the patterns of functional brain activations (Fig. 3): It was found that correct identifications of previously perceived scenes resulted in medial prefrontal activations, while incorrect identifications led to bilateral activations in the visual association cortex and precuneus. It can be speculated that the medial prefrontal activations reflect processes of monitoring and confidence, while the posterior activations reflect processes of trying to match internally generated images or imaginations with those seen in the outer world. 782 A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 Fig. 3. Horizontal sections through the human brain showing voxel-based bilateral activations (white regions) in the regions of the visual association cortex and the precuneus (left section) for falsely recognized pictures and in the medial prefrontal and anterior cingulate cortex (right section) for correctly identified pictures in the study of Kühnel et al. (2008). The research on false memories has implications for psychogenic amnesia, for example with regards to the legal aspect of hypnosis recall. As we will elaborate below, some cases of psychogenic amnesia might occur in relationship to legal difficulties or a forensic background. Though hypnosis may facilitate access to dissociated memories in certain psychogenic amnesic conditions, there is some evidence suggesting a higher likelihood of confident errors accompanying hypnosis recall (Maldonado & Spiegel, 2008). 4. Autobiographical–episodic memory and the brain The processing of memory has been a source of debate since the beginning of brain research. A waxing and waning interest in theories that proposed a strict mosaic-like localization (Markowitsch, 1994) within specific brain structures alternated with a similarly fluctuating interest in theories which promoted a widespread, Gestalt-like representation within the brain (Markowitsch, 1992). Nowadays, a compromise has generated a theoretical approach to memory processing that integrates both views and reflects the neuroscientific stance that both specialization and integration characterize the human brain (Paus, 2010; Pessoa, 2008). It is assumed that information enters the brain via the sensory organs and is then further processed depending on the kind of information and the process selected or triggered. Somewhat simplified, this means that subconsciously processed information preponderantly engages unimodal neocortical structures (priming), or – for procedural learning – the basal ganglia, premotor and other motor-related areas. Consciously processed information recruits more widespread networks, which are still largely neocortical for perceptual learning, but include limbic regions for the other two memory systems – the semantic and AEM systems. These two memory systems require the activation of limbic structures where its biological and social relevance is extracted and where the information is compared with already existing, related memories and later bound to and integrated with these. This process is also named synchronization. Further consolidation occurs during sleep (Stickgold & Walker, 2005). The link between sleep and memory has possible relevance for psychogenic amnesia. Loewenstein (1991) for example remarked that dissociated memories in psychogenic amnesia can often reveal their presence indirectly in nightmares. Kritchevsky et al. (2004) also described the case of man with severe retrograde functional amnesia, who during recovery of AEM began to experience nightmares with content related to past personal traumatic events. Though a connection between REM sleep and consolidation of emotional memory has been put forth by several authors, it has not unequivocally been substantiated empirically (for a review see Peigneux, Schmitz, & Urbain, 2010). Storage of facts and events is largely a matter of the cerebral cortex, though it has to be emphasized that storage is never final, as new information or the retrieval of already existing one leads to re-consolidation and new storage in the context of the last re-consolidation (Fulton, 2006; Haist, 2001; C.A. Miller & Sweatt, 2006; Wood, 2003). Retrieving facts and events requires an engagement of at least three closely interacting networks, namely activating brainstem structures comprising portions of the reticular activating system, a neocortical network containing the main information of the respective fact or event, plus a limbic network providing the emotional tagging of events or episodes (cf. Fig. 4.29 in Markowitsch & Welzer, 2010). Importantly, while encoding is based on a hierarchical arrangement of memory systems from procedural to AEM, retrieval allows independence in that way that no matter how the information was encoded, it can be retrieved in any memory system. Tulving (1995) suggested this flexibility and termed it the SPI-model, SPI referring to SERIAL encoding, PARALLEL processing and INDEPENDENT retrieval. For example, a patient with retrograde amnesia no longer knew or remembered that he had possessed a precious collection of antique clocks, but was able to manipulate the fine and complicated mechanical components within them without hesitance. Consequently, while he did not consciously remember his skills (serially encoded), he could retrieve them in an automatic way (parallel retrieval of procedural memories) (Markowitsch, Calabrese, Haupts, et al., 1993). While the main approach of the last century memory research consisted of a careful collection of data from neurological patients in order to establish relations between brain regions and memory processing, this has recently been complemented A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 783 and in part even replaced by functional imaging investigations in control subjects. Nevertheless, there are numerous examples of painstaking analyses in single patients demonstrating that regions of the limbic system are of crucial importance for encoding autobiographical memories (Fig. 4). The limbic system is seen as an interface between neocortical and brain stem structures (Fig. 4A). It contains important circuits and regions implicated in emotion and cognition (Fig. 4B). Within it, the hippocampal formation has received by far the most attention, due to the many influential papers on the amnesic patient H.M. of the Canadian neuropsychologist Brenda Milner. H.M. was a young man who due to pharmacologically resistant epileptic attacks underwent in 1953 bilateral surgery of his medial temporal lobes (Fig. 5). After surgery he became severely anterogradely and in part retrogradely amnesic (cf. Fig. 1) and remained so over many decades until his death in 2008. Many scientists concluded that the region within the medial temporal lobes, responsible for the amnesic condition with respect to AEM, was the hippocampus proper (e.g. Vargha-Khadem et al., 1997). Two other regional complexes within the limbic system have been regarded to be important ‘‘bottleneck” structures (Brand & Markowitsch, 2003) for AEM encoding. One is the medial and anterior diencephalon, the second is the amygdaloid body. Bilateral damage to the medial – and to a somewhat lesser degree also to anterior – diencephalic structures (anterior and mediodorsal thalamic nuclei, paratenial and midline nuclei) regularly leads to severe anterograde amnesia (Markowitsch, 1988; Vann, 2009). Given that this region contains a number of fibers (mammillothalamic tract, internal medullary lamina), Fig. 4. (A) Schematic sagittal section through a mammalian brain, illustrating the limbic system in two sagittal sections. The top schematic section reflects the frequently expressed idea of a ‘triune brain’ in which the limbic system works as an interface between the neocortex (controlling intellectual functions) and the brainstem (controlling basic physiological functions). (B) Schematic sagittal section through the human cerebrum, showing the main structures of the expanded limbic system. 784 A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 Fig. 5. Sketch of the extent of surgical removal in patient H.M. Top: View of H.M.’s cerebrum from below. The extent of surgical removal of the temporal lobes is indicated for one hemisphere only. The marked levels a–d refer to the bottom coronal sections which illustrate the removed medial temporal regions. Again, for comparison, the opposite hemisphere is presented intact, while in fact it had the same extent of removal. some diencephalic amnesic conditions may reflect a ‘‘disconnection syndrome” due to disruption of interactions between distant brain structures. The case of a former medical professor with above-average intelligence, illustrates this configuration of damage (Markowitsch, von Cramon, & Schuri, 1993). After a diencephalic infarct, this patient became totally anterogradely amnesic while his retrograde memory was partially preserved. His anterograde amnesia, which was also accompanied by a significant impairment of the ability to reflect on his condition, was so severe that his memory for new events lasted for seconds only. The other region, the amygdaloid body, is rarely damaged exclusively and bilaterally. An exception is the Urbach–Wiethe disease, a genetic condition that can lead to a selective calcification of both amygdalae (Markowitsch et al., 1994). As a consequence of bilateral damage of amygdala-structures, which are involved in the appraisal of incoming stimuli according to their biological and social significance, patients with this kind of damage may suffer problems with the processing of emotionally-laden AEMs (Cahill, Babinsky, Markowitsch, & McGaugh, 1995; Koenigs et al., 2008; Siebert, Markowitsch, & Bartel, 2003). As outlined above, the structures of the limbic system are engaged in a complementary, but closely interwoven manner in the acquisition of AEM information, while neocortical and in part limbic areas represent the major storage places. The A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 785 combined activation of right-hemispheric fronto-temporal regions serves as trigger stations for retrieving stored AEM events (Brand & Markowitsch, 2008; Fink et al., 1996; Kroll, Markowitsch, Knight, & von Cramon, 1997; LaBar & Cabeza, 2006). The corresponding regional complex in the left hemisphere seems to trigger semantic old memories (Markowitsch, Calabrese, et al., 1999). The fiber system, interconnecting the fronto-temporal regions is the uncinate fascicle (UF). UF has a temporal, frontal and insular part (Ebeling & von Cramon, 1992) and was ascribed functions in memory and emotional processing (Markowitsch, 1995). The ventromedial portion of UF primarily connects the amygdala and uncus with the gyrus rectus and the subcallosal area (Ebeling & von Cramon, 1992). According to a histopathological study, the right UF contains 33% more fibers and is 27% larger than the left one (Highley, Walker, Esiri, Crow, & Harrison, 2002). The ventral portion of the right UF has been implicated in the retrieval of episodic–autobiographical memories, in particular in ecphorizing affect-laden personal events (Fink et al., 1996; Levine et al., 1998; Markowitsch, 1995). According to some authors, UF may be involved in the formation of memory as well (Sepulcre et al., 2008). Levine et al. (1998, 2009) reported a case of isolated dense retrograde autobiographical–episodic amnesia covering the entire life, which occurred after a severe traumatic brain injury and was associated with a focal lesion of the frontal portion of the right UF. Despite normal performance on standard anterograde memory tests, the patient reported a feeling of disconnection from the post-accident autobiographical events. Subsequent refined testing of his anterograde AEM revealed that he assigned significantly less ‘‘remember‘‘ – ratings to his post-injury autobiographical events in comparison to normal subjects. He also generated less event-specific details from anterograde AEM than normal subjects, but this was not statistically significant. The lack of first person autonoetic connection with past personal events was also described in some psychopathological conditions accompanied by AEM impairments (Lemogne et al., 2006). One of the advanced models for the amnesia in dissociative identity disorder (DID) hypothesizes that the autobiographical traumatic information can in fact be accessed, but is discarded, because it is not perceived as belonging to the person’s autobiographical experience (Dorahy & Huntjens, 2007). As opposed to other brain fiber connections (such as fornix or parts of corpus callosum), UF matures later and more slowly and may continue its development beyond 30 years (Lebel, Walker, Leemans, Phillips, & Beaulieu, 2008). This may allow a higher structural plasticity in relationship to a variety of environmental influences. Alterations of the UF and connections between UF structural integrity and memory performance have been described in several psychiatric and neurological conditions (Diehl et al., 2008; Sepulcre et al., 2008; Staniloiu & Markowitsch, 2010; Yasmin et al., 2008). Microstructural changes of the UF were also identified in children, who were reared in a deprived, neglectful environment (Govindan, Behen, Helder, Makki, & Chugani, 2010). Interestingly, in children with a history of abuse or neglect, a reduced AEM specificity (overgeneral memory effect) was recently described by Valentino, Toth, and Cicchetti (2009). 5. Clinical aspects of psychogenic memory loss As the description of Ewald Hering implies AEM binds and integrates personal events and emotion with an autonoetic self. One view of retrograde ‘‘organic” AEM recollection impairments is that they may be accounted for by both loss of information and deficits in binding and reassembling details of the past (Rosenbaum, Gilboa, Levine, Winocur, & Moscovitch, 2009). Being the most advanced acquisition ontogenetically (Tulving, 2005), the AEM system displays a higher vulnerability to insults caused by various factors (including physical injuries and stress) in comparison to other memory systems (Markowitsch, 2008; Tulving, 2005). Stressful events can lead to disturbances of the integrated organization of memory, perception, consciousness and identity, causing so called dissociative disorders. Among Dissociative Disorders in DSM-IV-TR (2000), dissociative amnesia has as the central symptom the inability to recall important personal information usually of a traumatic nature (Brandt & van Gorp, 2006; Kopelman, 2000). The disturbance is precipitated by stressful experiences or psychological trauma and is not better explained by other psychiatric or medical conditions. The symptoms of dissociative amnesia cause significant impairment of functioning or distress. The degree of experienced distress depends on many variables, including the cultural views of dissociative experiences, selfhood and past (Seligman & Kirmayer, 2008). Dissociative amnesia is included in the psychological literature under the group of psychogenic amnesic disorders, which have as the main feature the preponderant contribution of psychological factors to their emergence. Psychogenic amnesic disturbances characterize not only dissociative amnesia, but could be part of other dissociative disorders, such as dissociative identity disorder (DID), dissociative fugue, Ganser syndrome, and dissociative trance disorder (possession trance), as well as certain anxiety disorders, such as acute stress disorder and post-traumatic stress disorder (PTSD) or personality disorders–borderline personality disorder (McKay & Kopelman, 2009; Staniloiu et al., 2009; Zanarini, Frankenburg, Jager-Hyman, Reich, & Fitzmaurice, 2008). Ganser syndrome is currently listed under the category of Dissociative Disorders Not Otherwise Specified in DSM-IV-TR and is defined by giving approximate answers to questions (vorbeireden). Ganser’s (1898, 1904) original clinical description of the syndrome was, however, much broader than the current DSM-IV-TR one. It included, in addition to the tendency to give approximate answers, a hysterical semi-trance condition, amnesia and hallucinations, being more consistent with subsequent views of this disorder as a brief reactive psychosis to stress. Dissociative memory symptoms in the form of hypermnesia (‘‘flashbacks”) or amnesia may also occur in PTSD. PTSD conditions which are accompanied by ‘‘positive” dissociative symptoms such as flashbacks and intrusions seem to engage different neural networks than PTSD conditions that are accompanied by ‘‘negative” dissociative symptoms such as amnesia (Markowitsch, 2000). Lanius et al. (2005) found that in general a network of prefronto–temporo-parietal areas was engaged 786 A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 in all patients, but that the group with ‘‘positive” symptoms had – compared to normal subjects – in addition a greater covariation with the right insula and right visual association cortex (compared to the reference in the left ventrolateral thalamus). Between the two groups, that with negative symptoms (amnesia) showed – compared to the ‘‘positive” (flashbacks) group – a more significant covariation in the left inferior frontal gyrus, while vice versa the ‘‘positive” symptoms group had more significant covariations with posterior cingulate/precuneus regions, the right middle temporal and the left inferior frontal gyri. In a recent review of various imaging studies of PTSD, arguments are made for a mechanism of undermodulation of affect via failure of prefrontal inhibition of limbic regions (such as amygdala) underlying the re-experiencing/hyperarousal PTSD subtype and one of overmodulation of the emotional limbic responses in the (‘‘negative”) dissociative PTSD subtype (Lanius et al., 2010). Psychogenic amnesias could further be differentiated according to the timeframe (anterograde versus retrograde), degree (global versus selective) of impairment of autobiographical–episodic memory and the co-existence of deficits in autobiographical–semantic memory and general semantic knowledge (cf. Serra, Fadda, Buccione, Caltagirone, & Carlesimo, 2007). The most frequent manifestations of psychogenic amnesias are forms of retrograde amnesias. A particular type of psychogenic retrograde amnesia has been termed the ‘mnestic block syndrome’ (Markowitsch, Kessler, Russ, et al., 1999). This form is characterized by an AEM block, which may comprise the whole past life. Affected patients otherwise have largely preserved semantic memories; they typically can read, write, calculate and know how to behave in social situations. Furthermore, they can store new long-term AEMs, though the acquisition of these new events may be less emotionally tagged than in normal subjects (Brand & Markowitsch, 2009; Reinhold & Markowitsch, 2009). When retrograde psychogenic amnesia is accompanied by sudden leaving of the customary environment – home and city – and compromised knowledge about personal identity, the condition is referred to as psychogenic (dissociative) fugue (Markowitsch, Fink, Thöne, Kessler, & Heiss, 1997). One example of dissociative fugue is the case of a man in his 30’s, who one morning took his bike to go to the bakery, but then suffered a sudden episode of amnesia. He rode his bike along the river for a whole week, being unable to recall his past and not knowing who he was (Markowitsch, Fink, Thöne, et al., 1997). Fugues have been reported for over a century (Markowitsch, 1992), though they were frequently erroneously associated with epilepsy (e.g., Burgl, 1900). A century ago, the fugue condition was named Wanderlust in Germany (cf. e.g. Burgl, 1900). Most fugues were not found to involve the formation of a new identity and were reported to be brief (Maldonado & Spiegel, 2008). However, prolonged courses of fugues were also described (Hennig-Fast et al., 2008). With regards to the degree of impairment of AEM loss, sometimes the ‘blocked’ autobiographical–episodic material is content-specific (selective) and/or retrograde amnesia is limited to specific life epochs (Markowitsch, Thiel, Kessler, et al., 1997). Although anterograde memory deficits occasionally accompany retrograde psychogenic amnesia, cases of psychogenic anterograde amnesia (inability to store new AEM episodes long-term) with preserved retrograde autobiographical–episodic memory have only been reported rarely (Kessler et al., 1997; Kumar, Rao, Sunny, & Gangadhar, 2007; Markowitsch, Kessler, Kalbe, & Herholz, 1999; Smith et al., in press; Staniloiu, Borsutzky, & Markowitsch, 2010). In a previous article, this condition (anterograde psychogenic amnesia) was described in a young woman who suffered a whiplash injury and who manifested a severe AEM block, which persisted for more than 15 years. After the whiplash injury she displayed an ongoing inability to acquire any new knowledge or AEM long-term, while her memories up to the time of the event remained precise and vivid. Similar to cases reported by Kapur et al. (1997) and O’Connor, Sieggreen, Ahern, Schomer, and Mesulam (1997) this young woman manifested accelerated forgetting, while storing information successfully for one to four hours. Despite that some patients with psychogenic amnesia claim loss of various procedural skills (Serra et al., 2007; van der Hart & Nijenhuis, 2001), objective evidence of procedural memory impairments in psychogenic or functional amnesia remains limited or debatable (Glisky et al., 2004; Huntjens, Postma, Woertman, van der Hart, & Peters, 2005; Smith et al., in press). In comparison to the terms dissociative or psychogenic amnesia, functional amnesia includes also forms of amnesia without significant brain damage as detected by conventional structural brain imaging and no clearly identifiable psychological etiological mechanisms underlying the memory disorder (De Renzi, Lucchelli, Muggia, & Spinnler, 1997; Fujiwara et al., 2008). The lack of clearly identifiable psychological triggers in functional amnesia may be explained by several factors: (a) The memory disturbance for the stressful event due to amnesia. (b) An impaired capacity for emotional awareness and processing in the face of ongoing or repeated stresses that seems to premorbidly characterize some of the patients with psychogenic amnesia and in fact might predispose them to develop this condition (Staniloiu et al., 2009). (c) The possible involvement of mechanisms of kindling sensitization in the face of recurrent stresses, which may trigger an episode of illness after a seemingly minor stress (Post, Weiss, Smith, Rosen, & Frye, 1995). (d) The so called incubation effect of life adversity (see below). (e) The personal view of the trauma within the explanatory model of illness (Kleinman, 1980) that may include direct exposure to trauma as well as vicarious traumatization (Maldonado et al., 2002). (f) The so- called process of ‘‘re-contextualization of health memory” (Modell, 2006) and finally (g) The presence of morphological changes, which could not be captured by conventional structural brain investigations (Bigler, 2004) but may be visualized with other techniques (such as Diffusion Tensor Imaging [DTI]). A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 787 6. The Janus head of amnesia: psychogenic and organic As it is the case with a variety of psychiatric disorders, advances in neurosciences have challenged the validity of the old distinctions in organic and psychogenic (Markowitsch, 1996, 1999a; Ouellet, Rouleau, Labrecque, Bernier, & Scherzer, 2008). Functional imaging studies have delivered evidence for a long searched biological link between the organic and psychogenic amnesia, by demonstrating brain functional and metabolic alterations in psychogenic or functional amnesia, which involve anatomical regions that are agreed upon to exert a crucial role in memory processing (Markowitsch, 1999b; Staniloiu & Markowitsch, in press). In a single case study of a patient with psychogenic amnesia after a traumatic event, involving severe retrograde and anterograde memory deficits, we identified via positron-emission-tomography (PET) an overall brain glucose hypometabolism with a more severe reduction of metabolism in the patient’s anterior and medial temporal lobes and the diencephalon (Markowitsch, Kessler, Van der Ven, Weber-Luxenburger, & Heiss, 1998). The patient’s memory recovery was observed approximately 8–12 months after the initiation of psychotherapeutic and psychopharmacological treatment and was paralleled by a normalization of the brain’s glucose metabolism (Markowitsch et al., 2000). Reductions in the brain’s glucose consumption, which preponderantly involved brain regions known to be part of the AEM network, were also found in subsequent investigations of other patients with psychogenic amnesia and are reviewed in Reinhold, Kühnel, Brand, and Markowitsch (2006), Brand and Markowitsch (2009) and Staniloiu and Markowitsch (in press). In a recent study of Brand, Eggers et al. (2009) that analyzed the results of functional brain imaging performed in resting state in 14 patients with psychogenic amnesia, reductions in metabolism of the right temporofrontal regions were identified, with a common significant hypometabolic zone in the right inferolateral prefrontal cortex (compared to the metabolism of control subjects). Similarly, Tramoni et al. (2009) found evidence of changes in the right prefrontal area in a patient with functional amnesia. By combining functional imaging with structural MRI methods sensitive for assessing white matter integrity, the latter identified subtle structural changes in several long fiber tracts passing through the inferolateral prefrontal cortex or its neighborhood. As the case-report of Tramoni et al. (2009) suggests, the gap between psychogenic and organic has furthermore been narrowed since the advent of newer structural techniques, such as DTI or magnetization transfer ratio measurements, which improved the ability to detect the integrity of brain white matter. Though cases of ‘‘organic” amnesia associated with fiber tract damage have been reported since long, the expanding use of the above mentioned techniques has led to an increase in findings of white matter damage in association with memory impairment or amnesia as well as a renewed interest in the mechanism of disconnection (Markowitsch, 1984). Techniques like DTI (Beaulieu, 2009) or magnetization transfer measurement (Tramoni et al., 2009) may in our opinion prove in the future to be valuable tools for identifying the neural correlates of certain forms of functional amnesia, such as the ones occurring after a mild traumatic brain injury (Staniloiu et al., 2009; Staniloiu, Markowitsch, & Borsutzky, 2010). This might be of importance, given that traumatic brain injuries are a frequent occurrence, but they are, on the one hand, underreported or documented (Ruff, Iverson, Barth, Bush, & Broshek, 2009). On the other hand, a substantial number of patients with mild traumatic brain injury (TBI) experience dissociative symptoms, which typically predict a poor outcome (Mooney & Speed, 2001). Finally, although in most of the cases of mild TBI conventional neuroimaging reveals no objective brain damage, recent studies evidenced via DTI white matter changes in specific brain regions of patients with mild TBI, which correlated with their degree of cognitive dysfunction (Lipton et al., 2008, 2009). The validity of a strict demarcation between organic and psychogenic amnesia has not only been questioned by the results of the functional brain imaging, but also by the observation that in some cases of organic amnesia psychological factors might play a significant role in exacerbating or maintaining the symptomatology (Kopelman, 2000, 2008). Furthermore symptom exaggeration has been reported in a variety of psychiatric and non-psychiatric conditions, including traumatic brain injury, depression, and dissociative disorders (Bass & Halligan, 2007). As suggested already in 1943 by Lennox, feigned amnesia (for primary or secondary gains) may accompany both organic and psychogenic amnesia. For example, ‘‘an epileptic, who also has hysteria and is a malingerer, may have periods of amnesia which exhibit features of all three types” (‘‘pathological”-organic, ‘‘psychological” and ‘‘feigned”; Lennox, 1943, p. 741). Differentiating psychogenic (dissociative) amnesia from disorders, which involve intentional production or feigning of symptoms (such as malingering and factious disorders), constitutes a diagnostic requirement that may pose a variety of challenges, especially in the forensic settings. Part of the challenge arises from the fact that some cases of psychogenic amnesia occur on a forensic background (Kritchevsky et al., 2004; Markowitsch, 1992; Markowitsch, Calabrese, Fink, et al., 1997; Staniloiu & Markowitsch, in press). Although dissociative (conversion) disorders (which were formerly captured under the name of hysteria) have traditionally been associated preponderantly with female gender, recent findings suggest that the frequency of dissociative disorders in men may have been underestimated, arguably partly because men with dissociative disorders may enter more frequently the legal system facilities than psychiatric settings (Spitzer & Freyberger, 2008). On the other hand, subjects who after a traumatic incident malinger cognitive deficits for financial gains, tend to typically claim memory impairments (Serra et al., 2007). In comparison to malingering that involves the intentional feigning of symptoms for legal, financial or economic gains, the intentional production of symptoms in factitious disorder is considered to be solely motivated by the wish to assume the sick role. While intentionality is not part of the diagnostic description of psychogenic (dissociative) disorders, an accurate clinical assessment of someone‘s motivations to assume the sick role remains a difficult task (Bass & Halligan, 2007). 788 A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 This observation has led several authors to propose a working model where dissociative (conversion) disorders, somatoform disorders and factitious disorders are conceptualized as being part of a continuum rather than sharply delineated diagnostic entities (Barbarotto, Laiacona, & Cocchini, 1996; Di Fiorino, 2003; Krahn, Bostwick, & Stonnington, 2008; Mayou, Kirmayer, Simon, Kroenke, & Sharpe, 2005). 7. Stress and psychogenic amnesia Psychogenic amnesias have been described in many cultures and recognized to occur in response to traumatic or psychological stress (Seligman & Kirmayer, 2008; Staniloiu & Markowitsch, in press). They could co-exist with other psychiatric disorders, such as major depressive disorder, bulimia nervosa, alcohol abuse or borderline, histrionic or narcissistic personality disorder (Maldonado & Spiegel, 2008). Psychogenic amnesias tend to affect younger people (Reinhold & Markowitsch, 2007; Yehuda, Schmeidler, Siever, Binder-Brynes, & Elkin, 1997), being more commonly reported in the third and fourth decade of life (Coons & Milstein, 1992; Kanzer, 1939). The age distribution on one hand echoes the findings of a negative correlation between age and dissociation scale scores (Putnam, 1997). On the other hand, it may reflect the differences in windows of vulnerability to stress of the main brain structures involved in AEM processes (Lupien, McEwen, Gunnar, & Heim, 2009) as well as age-related changes in brain connectivity patterns subserving AEM processes (Addis, Leclerc, Muscatell, & Kensinger, 2010; St Jacques, BessetteSymons, & Cabeza, 2009). Although in several cases a direct relationship between the severity of exposure to trauma and incidence of amnesia was reported, there are case reports of psychogenic amnesia after a seemingly minor stressor (Markowitsch et al., 1998; Reinhold & Markowitsch, 2007, 2009). In most of the latter cases the collateral information revealed a history of previous repeated traumatic experiences, suggesting a pathogenetic model of kindling sensitization (Post et al., 1995) or alternatively an incubation effect (Lupien et al., 2009). The concept of kindling was introduced in psychiatry from epileptology, where it described the process by which repeated sub-threshold stimulation of a brain area eventually resulted in a seizure. In psychopathology kindling could explain among other the observed inverse relationship between the number of experienced depressive episodes and precipitating life events. The incubation effect means that the effects of adversity, which occur during a window of vulnerability (such as a critical period of synaptic organization), do not become evident at the time of adversity, but later (such as when the synaptic organization is completed) (Lupien et al., 2009). Prior to elaborating on the connection between stress and psychogenic (dissociative) amnesia, we would like to emphasize that both stress and dissociative responses could be conceptualized as part of a spectrum that ranges from physiological, adaptive or non-pathological to maladaptive and pathological. Some non-pathological dissociative experiences, such as day dreaming, absorption, reverie may involve positive emotions and/or may enhance performance by focused allocation of cognitive resources. Maldonado and Spiegel (2008) give the example of athletes who can perform extremely well while focussing on a particular detail of an event, suspending any critical thinking and allowing their bodies to automatically do what they are supposed to do. Stress hormones, such as glucocorticoids promote under normal conditions brain maturation and remodelling, cell survival and learning and memory. Elevated or decreased levels may, however, have negative consequences for brain function or morphology (Du et al., 2009). Our pathogenetic model of the mnestic block syndrome hypothesizes that increased glucocorticoid levels play a significant role in the memory retrieval block (Markowitsch, 1996, 2000) and has received empirical support from several studies (for a review, see De Kloet & Rinne, 2007). Although glucocorticoids are critically involved in stress responses, a variety of other hormones (including arginin-vasopressin [AVP], oxytocin) may, however, have modulating effects (Joels & Baram, 2009). The degree to which chronic repeated stress or massive acute stress may affect an individual’s homeostasis and lead to psychiatric and/or non-psychiatric disorders depends on a range of factors, such as genetic dispositions, type of stress, duration of stress, developmental phase, age, gender, context, prior experiences, personality characteristics. Genetic factors might influence not only hormonal stress responses (glucocorticoids, AVP), but also the development of brain structures, brain plasticity, brain function (de Quervain et al., 2007; Rasch et al., 2009), and the temperament (including proneness to dissociative experiences; Becker-Blease et al., 2004). Carriers of a functionally relevant deletion variant of ADRA2B, the gene encoding the a2b-adrenoreceptor, were for example found to exhibit enhanced memory for emotional material (de Quervain et al., 2007), likely via modulation of amygdala activity (Rasch et al., 2009). Although Hurlemann et al. (2005) proposed that both hypermnesia and peri-emotional amnesia (decreased memory for neutral events during simultaneous enhanced encoding of an aversive event) are amygdala-dependent and vary as a function of noradrenergic-glucocorticoid input to the amygdala, the genetic underpinnings of psychogenic memory loss remain still unclear. Certain genetic polymorphisms may bias the normal maturation trajectories of key brain structures that are involved in processes of memory and emotion, possibly rendering them more sensitive to the effects of stress. A recent study (Pacheco et al., 2009) that examined the impact of the polymorphism of the serotonin transporter gene promoter region (5-HTTLPR) on white matter tracts connections in young women identified a significant association between the number of low expressing 5-HTTLPR alleles and the microstructural changes of the uncinate fascicle. The creation of brain structure is however not the product of the genes alone, but the result of the frequently synergistic interplay between genes and experience (environment) (Rutter, Moffitt, & Caspi, 2006). Both animal and human studies demonstrate that early life experiences could alter the gene expression via epigenetic modifications and lead to long lasting A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 789 changes in stress hormonal responses, synaptic plasticity and behavior. McGowan et al. (2009) recently showed epigenetic modifications such as increased methylation of the promoters of several genes of interest in the brain of victims of suicide with a history of childhood abuse in comparison to the suicide completers without a history of childhood abuse. These findings mirrored the ones from animal studies, which revealed that early development experiences could have programming effects on the hypothalamo–pituitary–adrenal (HPA) axis (Champagne et al., 2008; Weaver et al., 2005). Accumulating evidence suggests that epigenetic changes might also be involved in processes of learning and memory such as the extinction of fear-related memories, or re-consolidation (Roth & Sweatt, 2009). Several key brain structures for autobiographical memory and emotional processing have been identified as being sensitive to the consequences of exposure to negative or stressful experiences (such as amygdala and hippocampal formation, prefrontal cortex and specific white matter tracts). Reductions in the hippocampal volumes and AEM impairments have been reported in patients with stress-related psychiatric conditions, such as post-traumatic stress disorders (Bremner et al., 1997), recurrent major depressive disorders (especially the ones with a history of trauma) (Campbell, Marriott, Nahmias, & MacQueen, 2004; Vythilingam et al., 2002) and dissociative identity disorder (Vermetten, Schmahl, Lindner, Loewenstein, & Bremner, 2006). The effects of stress on the above mentioned brain structures depend on both their vulnerability and the magnitude of neurotoxic (such as glucocorticoid) cascades (O’Brien, 1997). The stage of development or declining of the respective structures influences their susceptibility to the stress effects (Lupien et al., 2009). This may partly explain why the same type of traumatic experiences may be associated with different brain morphological or functional changes and psychopathology (De Kloet & Rinne, 2007). Along this line, one study found that repeated episodes of sexual abuse were accompanied by reduced hippocampal volume when the traumatic experiences occurred early in childhood, but reduced prefrontal cortex gray matter volume when the traumatic experiences occurred in adolescence (Andersen & Teicher, 2008). The latter findings are explained by the fact that the hippocampus is fully organized during adolescence, while other structures, such as amygdala and prefrontal cortex are still developing. As opposed to the hippocampi, amygdala volumes may increase in response to stress as a result of persistent dendritic growth (Mehta et al., 2009; Roozendaal, McEwen, & Chattarji, 2009). The existence of sexually dimorphic brain structures and functions may get translated in gender differentiated responses to stress (Cahill, 2006; Lanius, Hopper, & Menon, 2003; Piefke, Weiss, Markowitsch, & Fink, 2005; Schulte-Rüther, Markowitsch, Shah, Fink, & Piefke, 2008; Tranel & Bechara, 2009). Changes in white matter tracts have also been reported in response to early stress. In children with a history of early deprivation, microstructural morphological changes of the UF, which were more prominent on the right side, were reported recently (Govindan et al., 2010). These changes correlated with the length of the time spent in the deprived environment. The findings of microstructural alterations in the UF, as detected by DTI, in relation to traumatic stress led us to hypothesize that UF plays a role in some cases of psychogenic amnesia (where a hypometabolism of the right inferior lateral prefrontal cortex was evidenced; Brand et al., 2009). The impact of early experiences on AEM development has also received support from the observation that mothers, who build up secure attachments with their children, engage in more elaborated reminiscing, which facilitates an earlier onset of AEMs and promotes a more elaborated encoding of AEMs in their offspring (Nelson & Fivush, 2004; Siegel, 2004). A reduced AEM specificity was described in children with a history of abuse or neglect (Valentino et al., 2009) and a combination of trauma and disorganized parent–child attachment in early life had empirically been linked to dissociative tendencies in late adolescence (Carlson, 1998; Siegel, 2004). 8. Psychogenic amnesia, emotions/feelings and personal meaning A central feature of AEM is its relationship with emotions. Emotions have an intimate connection to (one-)self and are used for appraising one‘s own position with respect to environment. Emotional evaluation is of importance for the acts of encoding, storage as well as for the recall of AEM. A ‘‘non-specific blurring or flattening” (Markowitsch, 1999b) in AEM (or so called overgeneralization phenomenon; Williams & Scott, 1988) has, for example, been described in patients with major depressive disorder. In contrast to semantic memories, which only display a low correlation with emotion (Welzer & Markowitsch, 2005), most AEM reminiscences are significantly affectively laden. At the brain level, this is reflected by findings that the recall of AEM memories typically engages networks involved in integrating emotion and cognition such as right temporal–frontal areas and right amygdala. The observation that the feelings have motivational functions and assign ‘‘value to what is meaningful” (Modell, 2006, p. 151), may partly explain the dependency of the experience of a coherent personal identity on the explicit episodic memories of elements in one’s own life history. Along this line, our own observations revealed that patients with psychogenic amnesia often displayed a blunted affective disposition along with a resignation to the present situation. Furthermore, several patients with psychogenic amnesia encountered difficulties with judging the feelings (Kritchevsky et al., 2004) and intentions of others (Reinhold & Markowitsch, 2007) as well as interpersonal relationships. This is not surprising as our emotions are molded or modulated by social environment and as ‘‘social animals” (Darwin, 1871/2004) we are particularly dependent on appropriately regulating, expressing and communicating our feelings and emotions in a social context (Leary, 2007). Furthermore certain so called social brain structures, such as the amygdala are also involved in AEM processes. Cahill et al. (1995) found for example that patients with bilateral amygdala degeneration are impaired in getting the gist of a told story; they fail 790 A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 to repeat (and consequently to remember) the emotionally important parts of it and instead tend to concentrate on the emotionally less important parts. These patients furthermore appear socially impaired in everyday life situations (Siebert et al., 2003). Similarly, patients with amygdala damage fail to appropriate judge emotions of others, which may be partly related to the role played by amygdala in directing the gaze towards the eye region of the other person (Gamer & Büchel, 2009; Kawashima et al., 1999; Vuilleumier & Pourtois, 2007). These findings suggest that not only the feeling of a coherent personal identity (being an autonoetic self), but also the ‘‘feeling” and ‘‘imagining” (Modell, 2006) of others (which facilitates the being of oneself with others) may have linkages to AEM. 9. Psychogenic amnesia and self versus other AEM, self-consciousness and theory of mind (ToM) seem to be correlated early during the ontogenetic development (Nelson & Fivush, 2004; Perner, 2000). The ToM capacity, which is defined as the ability to attribute and make inferences about the mental states of other people, such as desires, beliefs, intentions and feelings, and to differentiate between own and others’ mental states, undergoes significant and gradual changes during childhood. Between 16 and 24 months of age children pass the mirror recognition test, which marks the transition from a core self (Damasio, 1999; Nelson & Fivush, 2004) to a cognitive self (Howe & Courage, 1993) and is opined to be a necessary, but not sufficient condition for the development of episodic (Howe & Courage, 1993) or ‘‘evocative” memory (Adler, 1981). An enduring sense of self across time as detected through the delayed self recognition paradigm of Povinelli, Landau, and Perilloux (1996) is unusual in 2–3 year-old children, but is typically described in most 4 and 5 years old ones. Although individual, cultural and gender variations exist, the earliest childhood AEM dates to around age 3.5 years (Harpaz-Rotem & Hirst, 2005; Q. Wang, 2001). In the ToM domain, children seem to understand certain aspects of the mind from a strikingly early age, but it is not until around age 4 when they succeed in a standard false belief task (Baron-Cohen, Leslie, & Frith, 1985; Wimmer & Perner, 1983). These data led to the argument that meta-representational abilities, where there might be an (experiential) awareness – a state of consciousness similar to autonoetic consciousness or meta-representational self-consciousness (Newen and Vogeley, 2003) – of the relation between knowledge sources and present knowledge states (Perner, 2000) play a critical role in the development of ToM abilities. From an evolutionary point of view, it makes sense to believe that AEM, ToM and autonoetic self-consciousness also coevolved phylogenetically, driven by the forces of the social selection process (Corballis, 2009; Spreng, Mar, & Kim, 2008). In addition to directive, self representation and survival functions AEM has been attributed a significant social role (Bluck, Alea, Habermas, & Rubin, 2005). AEM contains a wealth of information about people and social interactions and the exchange of AEM enhances social and romantic bonding, understanding of others’ inner world and perspective. Additionally, it enables the intergenerational sharing of the past experiences, including the moral perspective (Nelson & Fivush, 2004). As Tulving (2005) remarked, possible ties between moral judgment and AEM were suggested by Darwin’s description of ‘‘moral being”, which underlined several features that match essential characteristics of AEM such as the capacity for autonoetic consciousness. While the relationship between AEM and the capacity for moral development remains still largely unexplored (Robertson et al., 2007), AEM impairments (especially for emotional events) were reported in offenders with psychopathy by several authors (Bourget & Whitehurst, 2007; M.C. Craig et al., 2009). Furthermore a recent study found evidence of microstructural integrity in the right uncinate fasciculus of forensic patients with psychopathy (M.C. Craig et al., 2009). Though language is not a necessary condition for AEM (Tulving, 2005), it supports and enriches the development of AEM capacities. It has been argued that language and mental time traveling (Suddendorf, Addis, & Corballis, 2009; Tulving & Kim, 2007) have co-evolved (Corballis, 2010). In contrast to the Chomskian’s view, some authors hypothesized that the languagebased communication system might have been preceded by the mimetic gesture (Corballis, 2009; Tomasello, 2008). Therefore language, similar to certain mechanisms involved in ToM (Boria et al., 2009; Cattaneo & Rizzolatti, 2009; Rizzolatti, Fabbri-Destro, & Cattaneo, 2009) might have had its roots in mirror neurons-containing brain regions such as territories including areas in the inferior frontal gyri. ToM abilities stem from innate neuro-cognitive capacities that are subserved by specific neural networks, which include right and left temporo-parietal junction, medial prefrontal cortex, precuneus, anterior temporal sulci and amygdala (Carrington & Bailey, 2009; Shaw et al., 2004). Congenitally blind people seem to activate similar neural networks during ToM tasks as people with normal vision (Bedny, Pascual-Leone, & Saxe, 2009). Partly dissociable neural systems underlie affective ToM functions (that are related to empathic processes) versus cognitive ToM functions (Shamay-Tsoory & Aharon-Peretz, 2007), with a preponderant role of ventromedial-prefrontal cortex areas for the affective ToM and cognitive aspects of empathy (Shamay-Tsoory, Aharon-Peretz, & Perry, 2009). Similar to AEM the ‘‘emergence” of the ToM capacity of an individual depends on other cognitive functions as well as environmental factors, including early life experiences and socio-cultural milieu. The nature of early attachments with primary care givers and mothers’ reminiscing style influence not only the development of AEM, but also of ToM capacities. The impaired performance on ToM tasks of children who were raised in deprived institutional settings reflects the importance of early life experiences with caregivers for the development of the ToM (Bedny et al., 2009; Colvert et al., 2008). In the field of brain functional imaging significant overlaps in the brain networks that become activated during the assessment of AEM, ToM, prospection (constructive episodic simulation), self projection, scene construction, navigation and the A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 791 default mode (Addis, Pan, Vu, Laiser, & Schacter, 2009; Buckner & Carroll, 2007; Hassabis & Maguire, 2007; Rabin, Gilboa, Stuss, Mar, & Rosenbaum, 2010) have been identified. In a recent meta-analysis the strongest overlap was reported for the neural networks subserving AEM and ToM tasks (Spreng et al., 2008). In spite of these results and sophisticated elaborations in the theoretical domains, the relationship between AEM and ToM has however been insufficiently experimentally investigated (Rabin et al., 2010; Spreng & Grady, 2010). Co-occurring impairments of AEM, ToM and autonoetic self awareness have been described in certain psychiatric disorders, such as schizophrenia (Corcoran & Frith, 2003), but it is unclear to which degree they reflect a developmental arrest of closely in time emerging neuro-cognitive functions or an early functional interdependence. Co-existing impairments of AEM and ToM have also been reported in the autistic spectrum disorders (Lind & Bowler, 2009; Shalom, 2003), as well as in mood disorders, such as major depressive disorder (MDD) and bipolar disorder (Inoue, Tonooka, Yamada, & Kanba, 2004; ShamayTsoory, Harari, Szepsenwol, & Levkovitz, 2009). Interestingly, in all these disorders microstructual changes in white matter tracts – particularly in the uncinate fascicle – were reported or suspected to exist (Kubicki et al., 2007; Pugliese et al., 2009; Taylor, Macfall, Gerig, & Krishnan, 2007; F. Wang et al., 2009). Lesions of the UF as well as of the capsula interna have been linked to memory impairments (Diehl et al., 2008; Fouquet et al., 2009; Markowitsch, von Cramon, Hofmann, Sick, & Kinzler, 1990; Sepulcre et al., 2008); furthermore, there are a few reports of ToM impairments occurring after white matter tract damage (Bach et al., 1998; Happé, Malhi, & Checkley, 2001). Several different theoretical models have attempted to capture the developmental interdependence between theory of mind and AEM (Saxe, Moran, Scholz, & Gabrieli, 2006). One model suggests that AEM depends on ToM and received some support from findings that performance on ToM tasks predicted performance on AEM tasks (Perner, 2001) and good ToM abilities are protective against false memories or contamination of AEM through suggestion (Bright-Paul, Jarrold, & Wright, 2008; Welch-Ross, Diecidue, & Miller, 1997). Other models propose that ToM abilities are at least partly dependent on AEM memory capacities or alternatively that the apparent connection between ToM and AEM tasks is mediated by autonoetic consciousness. Though many researchers would agree – or at least would not reject – that some interdependence between ToM and AEM characterizes early human development, findings from adult patients with brain lesions suggest that adult ToM and AEM abilities are at least partly dissociable (Bird, Castelli, Malik, Frith, & Husain, 2004). Rosenbaum, Stuss, Levine, and Tulving (2007) concluded from their observation in two patients with organic brain damage that a severe impairment of AEM and autonoetic consciousness does not affect the expression of premorbidly acquired ToM capacities; our findings from patients with psychogenic amnesia, however, suggest a certain degree of interdependence between ToM and AEM, which extends beyond childhood development (e.g., Brand et al., 2009; Fujiwara et al., 2008; Reinhold & Markowitsch, 2007, 2008, 2009). We advance several possible explanations for our results: (1) One possibility is that the acquisition of ToM and AEM capacities might initially engage similar neural networks. Once the ToM capacities are fully acquired and matured, they might be subserved by partially different neural networks, but the initial networks might remain available and might still be activated to accomplish or facilitate the performance of more difficult or complex novel tasks (Bird et al., 2004). (2) AEMs represent valuable reservoirs of information that could be flexibly used during social problem solving. In healthy subjects, however, it is unlikely that these reservoirs are accessed for habitual tasks, as for these it is relied on semantic and procedural knowledge (Rabin et al., 2010), which are more cognitively economical. Indeed in Hyman and Faries‘s (1992) research only few self-reports were recorded where AEM was used directively to solve problems. Instead a much more common reason for recalling or retelling memories was the sharing of experience and giving advice. (3) Depending on context, task requirements, gender, developmental stage, experience and cultural affiliations (e.g. Han & Northoff, 2008; Kobayashi, Glover, & Temple, 2007; Ray et al., 2010) individuals might resort to different strategies during ToM tasks. While some subjects might preponderantly rely on semantic memory and general knowledge for solving ToM tasks, others might engage more AEM resources. Several neuroimaging studies that investigated the neural correlates of ToM tasks (such as other minds‘ representation) (Legrand & Ruby, 2009; Schulte-Rüther, Markowitsch, Fink, & Piefke, 2007) produced activations of brain regions that are important for the recall of AEM events and elaboration of future events (Markowitsch, 2005). One possible interpretation of these results is an increased need of autobiographical–episodic recall for another versus self ToM tasks. Evidence for the engagement of different strategies and neural networks during ToM tasks was also provided by studies that investigated the reading of minds of dissimilar others versus similar others and neutral mind reading versus mind reading in emotional context (J.P. Legrand & Ruby, 2009; Mitchell, Macrae, & Banaji, 2006). (4) Patients with impairments of AEM, who perform well on laboratory ToM tasks, might still experience difficulties with ToM tasks in naturalistic settings, when confronted with novel, challenging social environments that may benefit from the use of AEM database or a higher degree of imagination. (5) Certain contexts might be conducive to more engagement of AEM during ToM tasks. In a compelling theoretical review of empathy as applied to psychoanalytical context, Buie (1981) described several types of empathy that could unfold in the therapeutic process: conceptual, resonant, self-experiential, imaginative imitation empathy. The latter occurs when a therapist must resort to his imagination and fantasy to find what it would be like to be in the other person’s shoes (i.e., in approximating experiential understanding) (Nagel, 1974). We would speculate that the professional use 792 A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 of empathy in the psychoanalytical context might offer an example of a setting that would promote the use of AEM resources to modulate empathetic responses, such as during self-experiential empathy or so called imaginative imitation empathy. (6) Gender differences may influence the degree to which the AEMs are engaged during theory of mind tasks and empathic responses, as suggested by a study that showed that women who had a similar experience reported more empathy than women who had not, while men who had a similar experience reported no more empathy than men who had not (Batson et al., 1996). The above described gender differences might partly reflect variations in socialization. Parental expectations of girls versus boys and gender differences in language ability, temperament and interest in social interaction may elicit different maternal reminiscing styles (Nelson & Fivush, 2004). We speculate that this might not only be conducive to gender differences in AEM encoding (such as facilitating a more detailed and emotionally-laden AEM encoding in girls versus boys), but may also promote a gender-differentiated inclination to use of AEM information to modulate ToM functioning and empathic responses. (7) Certain pathological conditions might be accompanied by an increased reliance on AEM or self-reflection for ToM tasks (which may not necessarily lead to an enhanced performance on ToM tasks, but sometimes on the contrary). For example, it has been observed that patients with major depressive disorder display a higher tendency to attribute negative emotions to neutral faces (Phillips, Drevets, Rauch, & Lane, 2003). In addition especially in patients suffering from chronic forms of depression deficits of ToM and capacity for true empathic engagement were described. We conjecture that these deficits may reflect several interacting factors, such as a pervasively egocentric perspective (McCullough, 2000) that might bias these patients towards using their own minds as a template for other people’s experience via simulation, an increased self focus (Grimm et al., 2009), a failure of inhibition of self perspective and/or a poverty of models of the inner world of others (Newen et al., 2009). (8) Age, perhaps as a measure of experience and /or brain networks’ functional connectivity may also influence the degree to which information from AEM is accessed for the performance of ToM tasks (Pfeifer et al., 2010). (9) Another possible explanation is that subtle deficits of ToM capacities might have been present in patients with psychogenic amnesia premorbidly, but that they were significantly exacerbated by the onset of amnesia. Depressive disorders, subclinical depressive symptoms and emotional processing difficulties have been reported to either precede or accompany dissociative disorders (Markowitsch, Kessler, Russ, et al., 1999; Markowitsch et al., 1998). Alexithymic traits – consisting of difficulty with identifying feelings and distinguishing between feelings and the bodily sensations of emotional arousal and verbalizing feelings to other people, reduced imagination and fantasy and externally oriented cognitive style – have been reported to be more common in men and have been linked to increased proneness to dissociation (Modestin, Lotscher, & Erni, 2002). They could accompany several conditions such as conversion disorder, somatoform disorders, fibromyalgia, depression, anxiety and traumatic brain injuries (including mild ones) (Koponen et al., 2005; Mattila et al., 2009). Alexithymic traits have been linked to early traumatic experiences and a dysfunction of the right hemisphere (Moriguchi et al., 2006; Schore, 2002) or inter-hemispheric transfer (Romei et al., 2008). High alexithymic scores have been found to correlate with impaired ToM, deficient self awareness (Moriguchi et al., 2009) and decreased gray matter volumes of brain regions that are known to be important for emotional processing, self awareness, ToM and imagination (e.g., precuneus) (Borsci et al., 2009). (10) Another explanatory avenue is that some, but not all of our patients with psychogenic amnesia and co-existing impairments of AEM and ToM displayed a degree of executive (such as cognitive flexibility) dysfunctions and/or semantic knowledge impairments (Bull, Phillips, & Conway, 2008). (11) Finally, another hypothesis is that the lack of significant impairment on ToM tasks that was observed in some of the patients with ‘‘organic” episodic memory impairment reflected a process of brain reorganization, which allowed other neighboring brain structures to undertake the ToM tasks (Reinhold & Markowitsch, 2008). Degenerative disorders usually develop over a long period of time (Mondadori et al., 2006). An interesting observation is that the two cases presented by Rosenbaum et al. (2007), who experienced significant memory impairments, had their ToM capabilities tested several years after the onset of their AEM problems. Consequently, one could argue that in the meantime significant neural reorganization could have occurred. In fact, brain reorganization has been described in several conditions, including a recent case who despite removal of the whole Broca area for a slowly developing brain tumor, suffered only very subtle speech impairment (Plaza, Gatignol, Leroy, & Duffau, 2009). Similarly, Henke et al. (2003) described a patient who suffered no major memory problems in spite of undergoing one-hemispheric surgical removal of his medial temporal lobe, including the hippocampal region, after epilepsy-related damage of the contralateral medial temporal lobe. We would like to add that employing of AEM information for the ToM tasks might not necessarily get translated into an enhanced accuracy of performance. This might be in a way similar to accessing AEM information to simulate future events. Gilbert and Wilson (2007) described several errors which may accompany future simulation, such as relying on the most recent or salient AEM information. We could speculate that the reliance on most recent AEM to simulate the future may in young adults be related to the stronger connection of these memories to the self, as evidenced by an fMRI study in young women, which demonstrated in this group a specific role of medial prefrontal cortex in retrieving recent AEMs (Oddo et al., 2010). A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 793 10. Psychogenic amnesia and future memories of the self ‘‘Time present and time past Are both perhaps present in time future” (Eliot, 1969, p. 171) An extended self emerges ontogenetically in conjunction with AEM, which throughout lifetime continues to support the self‘s vigorousness and coherence. Through the medium of autonoetic awareness which accompanies AEM, the extended self is capable of subjective mental time traveling in the past, present and future (Markowitsch, 2003; Tulving, 2005). Though traditionally the AEM was viewed as acting primarily as a storehouse of the past (Augustine, 1907/2001), recent data suggest that AEM may serve a significant proscopic function. Neuroimaging studies have revealed that similar networks, which serve AEM are engaged in self projection and construction of future events (Schacter & Addis, 2009). These findings might finally offer the long-time searched explanation to the evolutionists who struggled to find a survival advantage for the existence in humans of an AEM system that only offered an imperfect repository of the past. Indeed, it has since long time been observed that AEM are fragile and susceptible to forgetting and distortion (Kühnel et al., 2008; Loftus, 2005). Freud noted in a letter to Fliess that ‘‘the material present in the form of memory traces (is) being subjected from time to time to a rearrangement in accordance to fresh circumstances-to a retranscription” (Masson, 1985, p. 207). Furthermore several authors remarked that exceptional AEM abilities were not typically perceived as a blessing by their owners (Luria, 1987; Parker, Cahill, & McGaugh, 2006). On the contrary, similarly to the prospect that ‘‘every second of our lives recurs an infinite number of times” (Kundera, 1991, p. 5), the prospect of ‘‘running” (remembering) our entire life on a daily basis (Parker et al., 2006) has seemed terrifying and burdening. Moreover the benefits of an optimum balance between forgetting and remembering have been suggested since Ribot’s time (Hacking, 1995; Markowitsch & Brand, 2010). This provides further evidence that a main function of AEM in a permanently changing environment is to prepare human beings for the future, by reshaping and reconstructing the past to support current aspects of the self and match future goals that are coherent with one individual’s goals, self image and system of beliefs (Conway, 2009). The importance of AEM for mental time traveling (experiencing time versus knowing time) in the future is corroborated by findings from patients with organic or psychogenic amnesia as well other psychiatric disorders that are accompanied by AEM deficits (such as major depressive disorder). Patients with psychogenic or organic amnesia are often unable to plan for their personal future, being imprisoned or trapped in an extended or forever ‘noetic’ present (Suddendorf et al., 2009). Patients with psychogenic amnesia frequently show resignation to their present situation and an apparent lack of concern (la belle indifference; Janet, 1907) about their symptoms (Reinhold & Markowitsch, 2007; Serra et al., 2007). A main characteristic of patients with major depressive disorder is the inability to imagine personal positive future events (Sharot, Riccardi, Raio, & Phelps, 2007), which may lead them to attempt suicide. This inability may reflect a disruption of the balance between the neural networks that subserve the encoding and retrieval of positive versus negative AEMs (Lemogne et al., 2006; Markowitsch, Vandekerckhove, Lanfermann, & Russ, 2003), which may partly be mediated by an impaired sensitivity to positive experiences (Addis et al., 2010; Pizzagalli et al., 2009). Interestingly, several structures involved in AEM processing (such as basal forebrain, amygdala, ventromedial-prefrontal cortex) have also been reported to be engaged in reward-related processing, decision making and future-minded choice behavior (e.g., Daniel & Pollmann, 2010; Peters & Büchel, 2010; Ulrich-Lai & Herman, 2009). These may partly explain descriptions of changes in eating preferences, smoking or drinking habits or other previously rewarding activities (such as car driving) after the onset of psychogenic amnesia, which occurred without evidence for impairments in procedural knowledge (Fujiwara et al., 2008; Kritchevsky et al., 2004; Markowitsch, Fink, Thöne, et al., 1997; Thomas-Antérion, Mazzola, Foyatier-Michel, & Laurent, 2008). 11. Searching for the neural signature of the constricted self in the psychogenic amnesia Despite the risk of being criticized for adding a new form of self to the already existing long list of selves (Legrand & Ruby, 2009), we employed the term constricted self to metaphorically capture the symptomatology of patients with psychogenic amnesia (including the constriction of mental time traveling in the future). Janet (1907) talked about hysteria as a disorder of ‘‘personal synthesis”. As some authors pointed out (Spiegel, 2006; van der Hart & Dorahy, 2006), he often used the term désagregation (in spite of the availability in both his own and French traditional psychological vocabulary of the French term dissociation) (Janet, 1920). This suggests that he might have already understood this condition as involving not only a simple separation, but a failure of integration of various aspects of cognition (memory), self, consciousness and emotion (Maldonado & Spiegel, 2008). Similarly in several patients with psychogenic amnesia not only deficits of AEM, but also impairments of the capacities for self-consciousness, subjective time traveling, executive functions and ‘‘feeling” and ‘‘being” with others were observed (Brand et al., 2009). Repeatedly a neural signature of psychogenic amnesia has been identified in a particular area in the right prefrontal cortex. In a relatively large sample of patients with psychogenic amnesia characterized by severe retrograde AEM deficits and no overt structural abnormalities as detected by conventional imaging methods, evidence of functional changes (hypometabolism) during resting state in the right temporofrontal region with a significant decrease in the inferolateral prefrontal cortex was found (Brand et al., 2009). These changes in our opinion reflect the stress-mediated impairment of AEM retrieval. The detected location is consistent with previous findings from both cases of organic and psychogenic amnesia (Kroll et al., 1997; Piolino et al., 2005). The vicinity of this identified area with regions involved in processing of emotion, self-evaluation, self-regulation, 794 A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 Fig. 6. The place of the insula in the fronto-temporal junction area. consciousness, time perception, executive functions and theory of mind (Brand et al., 2009) offers an explanation for the observed malfunction of personal synthesis in patients with psychogenic amnesia. A neighboring brain region that has recently been receiving increased attention (A.D. Craig, 2010) in stress-related disturbances of self and consciousness (Lanius et al., 2005, 2010), including the ones accompanying the psychogenic memory loss (Brand et al., 2009; Thomas-Anterion, Guedj, Decousus, & Laurent, 2010) is the insula. Via glucose PET evidence of insular hypometabolism in patients with psychogenic amnesia was provided (Brand et al., 2009; Markowitsch et al., 1998; Thomas-Anterion et al., 2010). Hidden in the Sylvian sulcus in the triangle of frontal, parietal, and temporal cortex, the insula (Fig. 6) is connected to a variety of structures, including inferior frontal gyrus, septum and amygdala (Allman et al., 2010; Markowitsch, Emmans, Irle, Streicher, & Preilowski, 1985; Nieuwenhuys, Voogd, & van Huijzen, 1978). It sends projections to hippocampus and receives efferents from the entorhinal cortex (Nieuwenhuys et al., 1978). The most anterior and ventral portion of the insula that is close to the frontal operculum contains the so-called von Economo neurons (VEN) (Allman, Watson, Tetreault, & Hakeem, 2005). In the right hemisphere the human postnatal insular-frontal cortex contains significantly more VENs than in the left (Allman et al., 2010). The VEN neurons, which are also present in the anterior cingular cortex (ACC), have been ascribed functions in consciousness, social cognition and regulation of appetite (Allman et al., 2010). Their degeneration was for example found to be associated with loss of emotional awareness and altered self-consciousness in patients with certain variants of fronto-temporal dementia (Seeley, 2008; Sturm, Rosen, Allison, Miller, & Levenson, 2006) as well as aberrant eating habits (Seeley, 2008). Insula has also been assigned functions in verbal memory tasks (Grasby et al., 1994; Morin, 2009; Morin & Michaud, 2007), inner speech (Vercammen, Knegtering, Bruggeman, & Aleman, 2010), time perception (Rao, Mayer, & Harrington, 2001) and drug (smoking) cravings (Naqvi, Rudrauf, Damasio, & Bechara, 2007.) Recent data from Arzy, Collette, Ionta, Fornari, and Blanke (2009) indicate that also self projection in time and ‘‘facilitation of future judgments with respect to one’s past ‘‘(p. 2016) engage the insula. Distortions of time perception have been described in patients with psychogenic (dissociative) amnesia (Steinberg, 2000). As mentioned above, smoking cessation and changes in eating preferences have also been reported after the onset of psychogenic memory loss (Fujiwara et al., 2008; Thomas-Antérion, Mazzola, Foyatier-Michel, & Laurent, 2008), though their neurobiological underpinnings are still unclear. In normal subjects increased activations of the insula have been evidenced in tasks involving self versus other conditions (Schilbach et al., 2006) and the right insula was particularly found to be activated during self face recognition (Devue et al., 2007). The insula has subsequently been proposed to be a structure that plays a critical role in supporting the so-called ‘‘sentient” (feeling) self (A.D. Craig, 2009). In light of these findings we would speculate that the efferents that the right inferior frontal gyrus receives from the right insula (J.R. Augustine, 1996), might provide gateways from the more primitive ‘‘feeling” self towards the extended or autobiographical self. 12. Conclusions The topic of self and self-consciousness has generated the interest of humanity for centuries. Adherents to Aristotelian or Heraclitean views of self (Moldoveanu & Stevenson, 2001) or opponents to the notion of self have all generated a significant A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 795 amount of sophisticated writings, reinforcing the idea that – in spite of differences of opinions regarding self-specificity or what it means to have a self – self is a topic of universal interest in multiple disciplines. Similar to the topic of selfhood, psychogenic amnesia has also ignited the interest of neurologists, psychiatrists and psychotherapists for more than a century and has generated a lot of debate. By providing a description of psychogenic amnesia in the present article we hope that we were able to reiterate the importance of AEM for supporting and maintaining a robust sense of autonoetic self, and to offer, in addition, some glimpses into the neurobiology that underlies the autobiographical– episodic remembrances of the past, as well as the self‘s memories of the future (Ingvar, 1985). References Addis, D. R., Leclerc, C. M., Muscatell, K. A., & Kensinger, E. A. (2010). There are age-related changes in neural connectivity during the encoding of positive, but not negative information. Cortex, 46, 425–433. Addis, D. R., Pan, L., Vu, M.-A., Laiser, N., & Schacter, D. L. (2009). Constructive episodic simulation of the future and the past: Distinct subsystems of a core brain network mediate imaging and remembering. Neuropsychologia, 42, 2222–2238. Adler, G. (1981). The borderline–narcissistic personality disorder continuum. American Journal of Psychiatry, 138, 46–50. Allman, J. M., Tetreault, N. A., Hakeem, A. Y., Manaye, K. F., Semendeferi, K., Erwin, J. M., et al (2010). The von Economo neurons in frontoinsular and anterior cingulate cortex in great apes and humans. Brain Structure and Function, 214, 495–517. Allman, J. M., Watson, K. K., Tetreault, N. A., & Hakeem, A. Y. (2005). Intuition and autism: A possible role for Von Economo neurons. Trends in Cognitive Sciences, 9, 367–373. Andersen, S. L., & Teicher, M. H. (2008). Stress, sensitive periods and maturational events in adolescent depression. Trends in Neurosciences, 31, 183–191. Arzy, S., Collette, S., Ionta, S., Fornari, E., & Blanke, O. (2009). Subjective mental time: The functional architecture of projecting the self to past and future. European Journal of Neuroscience, 30, 2009–2017. Augustine, J. R. (1996). Circuitry and functional aspects of the insular lobe in primates including humans. Brain Research Reviews, 22, 229–244. Augustine (2001). The confessions. New York: A.A. Knopf. Bach, I., Davis, S., Colvin, C., Wijerante, C., Happé, F., & Howard, R. (1998). A neuropsychological investigation of theory of mind in an elderly lady with frontal leucotomy. Cognitive Neuropsychiatry, 3, 139–159. Baddeley, A. (2000). The episodic buffer: a new component of working memory? Trends in Cognitive Sciences, 4, 417–423. Baddeley, A. D., & Hitch, G. J. (1974). Working memory. In G. A. Bower (Ed.), Recent advances in learning and motivation (pp. 47–89). New York: Academic Press. Barbarotto, R., Laiacona, M., & Cocchini, G. (1996). A case of simulated, psychogenic or focal pure retrograde amnesia: Did an entire life become unconscious? Neuropsychologia, 34, 575–585. Baron-Cohen, S., Leslie, A. M., & Frith, U. (1985). Does the autistic child have a ‘‘theory of mind”? Cognition, 21, 37–46. Bass, C., & Halligan, P. W. (2007). Illness related deception: Social or psychiatric problem? Journal of the Royal Society of Medicine, 100, 81–84. Batson, C. D., Sympson, S. C., Hindman, J. L., Decruz, P., Todd, R. M., Weeks, J. L., et al (1996). ‘‘I’ve been there, too”: Effect on empathy of prior experience with a need. Personality and Social Psychology Bulletin, 22, 474–482. Beaulieu, C. (2009). The biological basis of diffusion anisotropy. In H. Johansen-Berg & T. E. J. Behring (Eds.), Diffusion MRI: From quantitative measurement to in vivo neuroanatomy (pp. 105–126). London, UK: Elsevier. Becker-Blease, K. A., Deater-Deckard, K., Eley, T., Freyd, J. J., Stevenson, J., & Plomin, R. (2004). A genetic analysis of individual differences in dissociative behaviors in childhood and adolescence. Journal of Child Psychology and Psychiatry, 45, 522–532. Bedny, M., Pascual-Leone, A., & Saxe, R. R. (2009). Growing up blind does not change the neural bases of Theory of Mind. Proceedings of the National Academy of Science of the USA, 106, 11312–11317. Bigler, E. D. (2004). Neuropsychological results and neuropathological findings at autopsy in a case of mild traumatic brain injury. Journal of the International Neuropsychological Society, 10, 794–806. Bird, C. M., Castelli, F., Malik, O., Frith, U., & Husain, M. (2004). The impact of extensive medial frontal lobe damage on ‘Theory of mind” and cognition. Brain, 127, 914–928. Bluck, S., Alea, N., Habermas, T., & Rubin, D. (2005). A tale of three functions: The self-reported uses of autobiographical memory. Social Cognition, 23, 91–117. Boria, S., Fabbri-Destro, M., Cattaneo, L., Sparaci, L., Sinigaglia, C., Santelli, E., et al (2009). Intention understanding in autism. PLoS One, 4, e5596. Borsci, G., Boccardi, M., Rossi, R., Rossi, G., Perez, J., Bonetti, M., et al (2009). Alexithymia in healthy women: A brain morphology study. Journal of Affective Disorder, 114, 208–215. Bourget, D., & Whitehurst, L. (2007). Amnesia and crime. Journal of the American Academy of Psychiatry and the Law, 35, 469–480. Brand, M., Eggers, C., Reinhold, N., Fujiwara, E., Kessler, J., Heiss, W.-D., et al (2009). Functional brain imaging in fourteen patients with dissociative amnesia reveals right inferolateral prefrontal hypometabolism. Psychiatry Research: Neuroimaging Section, 174, 132–139. Brand, M., & Markowitsch, H. J. (2003). The principle of bottleneck structures. In R. H. Kluwe, G. Lüer, & F. Rösler (Eds.), Principles of learning and memory (pp. 171–184). Basel: Birkhäuser. Brand, M., & Markowitsch, H. J. (2008). The role of the prefrontal cortex in episodic memory. In E. Dere, A. Easton, L. Nadel, & J. P. Huston (Eds.). Handbook of behavioral neuroscience: Episodic memory research (Vol. 18, pp. 317–342). Amsterdam: Elsevier. Brand, M., & Markowitsch, H. J. (2009). Environmental influences on autobiographical memory: The mnestic block syndrome. In L. Bäckman & L. Nyberg (Eds.), Festschrift in honor of Lars-Göran Nilsson (pp. 229–264). Hove, UK: Psychology Press. Brandt, J., & van Gorp, W. G. (2006). Functional (‘‘psychogenic”) amnesia. Seminars in Neurology, 26, 331–340. Bremner, J. D., Randall, P., Vermetten, E., Staib, L., Bronen, R. A., Mazure, C., et al (1997). Magnetic resonance imaging-based measurement of hippocampal volume in posttraumatic stress disorder related to childhood physical and sexual abuse – A preliminary report. Biological Psychiatry, 41, 23–32. Bright-Paul, A., Jarrold, C., & Wright, D. B. (2008). Theory-of-mind development influences suggestibility and source monitoring. Developmental Psychology, 44, 1055–1068. Buckner, R. L., & Carroll, D. C. (2007). Self-projection and the brain. Trends in Cognitive Sciences, 11, 49–57. Buie, D. H. (1981). Empathy: Its nature and limitations. Journal of the American Psychoanalytic Associations, 29, 281–307. Bull, R., Phillips, L. H., & Conway, C. A. (2008). The role of control functions in mentalizing: Dual-task studies of theory of mind and executive function. Cognition, 107, 663–672. Burgl, G. (1900). Eine Reise in die Schweiz im epileptischen Dämmerzustande und die transitorischen Bewusstseinsstörungen der Epileptiker vor dem Strafrichter [A journey to Switzerland done in epileptic somnolence and the transitory disturbances of consciousness before the criminal judge]. Münchener medizinische Wochenschrift, 37, 1270–1273. Cahill, L. (2006). Why sex matters for neuroscience. Nature Reviews Neuroscience, 7, 477–484. Cahill, L., Babinsky, R., Markowitsch, H. J., & McGaugh, J. L. (1995). Involvement of the amygdaloid complex in emotional memory. Nature, 377, 295–296. Campbell, S., Marriott, M., Nahmias, C., & MacQueen, G. M. (2004). Lower hippocampal volume in patients suffering from depression: A meta-analysis. American Journal of Psychiatry, 161, 598–607. Carlson, E. A. (1998). A prospective longitudinal study of attachment disorganization/disorientation. Child Development, 69, 1107–1128. 796 A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 Carrington, S. J., & Bailey, A. J. (2009). Are there theory of mind regions in the brain? A review of the neuroimaging literature. Human Brain Mapping, 30, 2313–2335. Cattaneo, L., & Rizzolatti, G. (2009). The mirror neuron system. Archives of Neurology, 66, 557–560. Champagne, D. L., Bagot, R. C., van Hasselt, F., Ramakers, G., Meaney, M. J., de Kloet, E. R., et al (2008). Maternal care and hippocampal plasticity: Evidence for experience-dependent structural plasticity, altered synaptic functioning, and differential responsiveness to glucocorticoids and stress. Journal of Neuroscience, 28, 6037–6045. Clayton, N. S., & Dickinson, A. (1998). Episodic-like memory during cache recovery in scrub-jays. Nature, 395, 272–274. Colvert, E., Rutter, M., Kreppner, J., Beckett, C., Castle, J., Groothues, C., et al (2008). Do theory of mind and executive function deficits underlie the adverse outcomes associated with profound early deprivation? Findings from the English and Romanian adoptees study. Journal of Abnormal Child Psychology, 36, 1057–1068. Conway, M. A. (2009). Episodic memories. Neuropsychologia, 47, 2305–2313. Coons, P. M., & Milstein, V. (1992). Psychogenic amnesia: A clinical investigation of 25 cases. Dissociation, 5, 73–79. Corballis, M. C. (2009). Mental time travel and the shaping of language. Experimental Brain Research, 192, 553–560. Corballis, M. C. (2010). Mirror neurons and the evolution of language. Brain and Language, 112, 25–35. Corcoran, R., & Frith, C. D. (2003). Autobiographical memory and theory of mind: Evidence of a relationship in schizophrenia. Psychological Medicine, 33, 897–905. Cowan, N. (2001). The magical number 4 in short-term memory: A reconsideration of mental storage capacity. Behavioral and Brain Sciences, 24, 87–185. Craig, A. D. (2009). How do you feel – Now? The anterior insula and human awareness. Nature Reviews Neuroscience, 10, 59–70. Craig, A. D. (2010). Once an island, now the focus of attention. Brain Structure and Function, 214, 395–396. Craig, M. C., Catani, M., Deeley, Q., Latham, R., Daly, E., Kanaan, R., et al (2009). Altered connections on the road to psychopathy. Molecular Psychiatry, 14, 946–953. Damasio, A. (1999). The feeling of what happens. Body and emotion in the making of consciousness. Orlando, FL: Harcourt, Inc. Daniel, R., & Pollmann, S. (2010). Comparing the neural basis of monetary reward and cognitive feedback during information-integration category learning. Journal of Neuroscience, 30, 47–55. Darwin, C. (2004). The descent of man, and selection in relation to sex. London, UK: Penguin Books. De Kloet, E. R., & Rinne, T. (2007). Neuroendocrine markers of early trauma. In E. Vermetten, M. J. Dorahy, & D. Spiegel (Eds.), Traumatic dissociation. Neurobiology and treatment (pp. 139–156). Arlington, VA: American Psychiatric Publ. De Quervain, D. J., Kolassa, I. T., Ertl, V., Onyut, P. L., Neuner, F., Elbert, T., et al (2007). A deletion variant of the alpha2b-adrenoceptor is related to emotional memory in Europeans and Africans. Nature Neuroscience, 10, 1137–1139. De Renzi, E., Lucchelli, F., Muggia, S., & Spinnler, H. (1997). Is memory loss without anatomical damage tantamount to a psychogenic deficit? The case of pure retrograde amnesia. Neuropsychologia, 35, 781–794. Devue, C., Collette, F., Balteau, E., Degueldre, C., Luxen, A., Maquet, P., et al (2007). Here I am: The cortical correlates of visual self-recognition. Brain Research, 1143, 169–182. Di Fiorino, M. (2003). Factitious disorders: The Münchausen and Ganserian patients. Bridging Eastern and Western Psychiatry, 1, 56–65. Diehl, B., Busch, R. M., Duncan, J. S., Piao, Z., Tkach, J., & Lüders, H. O. (2008). Abnormalities in diffusion tensor imaging of the uncinate fasciculus relate to reduced memory in temporal lobe epilepsy. Epilepsia, 49, 1409–1418. Dorahy, M. J., & Huntjens, R. J. C. (2007). Memory and attentional processes in dissociative identity disorder. In E. Vermetten, M. J. Dorahy, & D. Spiegel (Eds.), Traumatic dissociation. Neurobiology and treatment (pp. 55–75). Arlington, VA: American Psychiatric Publ. DSM-IV-TR (2000). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: American Psychiatric Association. Du, J., Wang, Y., Hunter, R., Wei, Y., Blumenthal, R., Falke, C., et al (2009). Dynamic regulation of mitochondrial function by glucocorticoids. Proceedings of the National Academy of Sciences of the USA, 106, 3543–3548. Ebeling, U., & von Cramon, D. (1992). Topography of the uncinate fascicle and adjacent temporal fiber tracts. Acta Neurochirurgica, 115, 143–148. Eliot, T. S. (1969). The complete poems and plays. London, UK: Faber & Faber Limited. Fink, G. R., Markowitsch, H. J., Reinkemeier, M., Bruckbauer, T., Kessler, J., & Heiss, W.-D. (1996). Cerebral representation of one’s own past: Neural networks involved in autobiographical memory. Journal of Neuroscience, 16, 4275–4282. Fouquet, M., Desgranges, B., Landeau, B., Duchesnay, E., Mezenge, F., de la Sayette, V., et al (2009). Longitudinal brain metabolic changes from amnestic mild cognitive impairment to Alzheimer’s disease. Brain, 132, 2058–2067. Freud, S. (1901). Zum psychischen Mechanismus der Vergesslichkeit [On the psychic mechanism of forgetfulness]. Monatsschrift für Psychiatrie und Neurologie, 4, 436–443. Fujiwara, E., Brand, M., Kracht, L., Kessler, J., Diebel, A., Netz, J., et al (2008). Functional retrograde amnesia: A multiple case study. Cortex, 44, 29–45. Fulton, D. (2006). Reconsolidation: Does the past linger on? Journal of Neuroscience, 26, 10935–10936. Gamer, M., & Büchel, C. (2009). Amygdala activation predicts gaze toward fearful eyes. Journal of Neuroscience, 29, 9123–9126. Ganser, S. J. (1898). Ueber einen eigenartigen hysterischen Dämmerzustand [On a peculiar hysterical state of somnolence]. Archiv für Psychiatrie und Nervenkrankheiten, 30, 633–640. Ganser, S. J. (1904). Zur Lehre vom hysterischen Dämmerzustande [On the theory of the hysterical state of somnolence]. Archiv für Psychiatrie und Nervenkrankheiten, 38, 34–46. Gilbert, D. T., & Wilson, T. D. (2007). Prospection: Experiencing the future. Science, 317, 1351–1354. Glisky, E. L., Ryan, L., Reminger, S., Hardt, O., Hayes, S. M., & Hupbach, A. (2004). A case of psychogenic fugue: I understand, aber ich verstehe nichts. Neuropsychologia, 42, 1132–1147. Govindan, R. M., Behen, M. E., Helder, E., Makki, M. I., & Chugani, H. T. (2010). Altered water diffusivity in cortical association tracts in children with early deprivation identified with tract-based spatial statistics. Cerebral Cortex, 20, 561–569. Grasby, P. M., Frith, C. D., Friston, K. J., Simpson, J., Fletcher, P. C., Frackowiak, R. S., et al (1994). A graded task approach to the functional mapping of brain areas implicated in auditory-verbal memory. Brain, 117, 1271–1282. Grimm, S., Ernst, J., Boesiger, P., Schuepbach, D., Hell, D., Boeker, H., et al (2009). Increased self-focus in major depressive disorder is related to neural abnormalities in subcortical–cortical midline structures. Human Brain Mapping, 30, 2617–2627. Hacking, I. (1995). Rewriting the soul. Multiple personality and the sciences of memory. Princeton, NJ: Princeton University Press. Haist, F. (2001). Consolidation of human memory over decades revealed by functional magnetic resonance imaging. Neuroscience, 4, 1139–1145. Han, S., & Northoff, G. (2008). Culture-sensitive neural substrates of human cognition: A transcultural neuroimaging approach. Nature Reviews Neuroscience, 9, 646–654. Happé, F., Malhi, G. S., & Checkley, S. (2001). Acquired mind-blindness following frontal lobe surgery? A single case study of impaired ‘theory of mind’ in a patient treated with stereotactic anterior capsulotomy. Neuropsychologia, 39, 83–90. Harpaz-Rotem, I., & Hirst, W. (2005). The earliest memories in individuals raised in either traditional and reformed kibbuz or outside the kibbuz. Memory, 13, 51–62. Hassabis, D., & Maguire, E. A. (2007). Deconstructing episodic memory with construction. Trends in Cognitive Sciences, 11, 299–306. Henke, K., Treyer, V., Weber, B., Nitsch, R. M., Hock, C., Wieser, H. G., et al (2003). Functional neuroimaging predicts individual memory outcome after amygdalohippocampectomy. NeuroReport, 14, 1197–1202. Hennig-Fast, K., Meister, F., Frodl, T., Beraldi, A., Padberg, F., Engel, R. R., et al (2008). A case of persistent retrograde amnesia following a dissociative fugue: Neuropsychological and neurofunctional underpinnings of loss of autobiographical memory and self-awareness. Neuropsychologia, 46, 2993–3005. A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 797 Hering, E. (1870/1895). Ueber das Gedächtnis als eine allgemeine Funktion der organisierten Materie. Vortrag gehalten in der feierlichen Sitzung der Kaiserlichen Akademie der Wissenschaften in Wien am XXX. Mai MDCCCLXX. Leipzig: Akademische Verlagsgesellschaft. (Transl.: Hering, E. (1895). Memory as a general function of organized matter. Chicago: Open Court). Highley, J. R., Walker, M. A., Esiri, M. M., Crow, T. J., & Harrison, P. J. (2002). Asymmetry of the uncinate fasciculus: A post-mortem study of normal subjects and patients with schizophrenia. Cerebral Cortex, 12, 1218–1224. Howe, M. L., & Courage, M. L. (1993). On resolving the enigma of infantile amnesia. Psychological Bulletin, 113, 305–326. Huntjens, R. J., Postma, A., Woertman, L., van der Hart, O., & Peters, M. L. (2005). Procedural memory in dissociative identity disorder: When can interidentity amnesia be truly established? Consciousness and Cognition, 14, 377–389. Hurlemann, R., Hawellek, B., Matusch, A., Kolsch, H., Wollersen, H., Madea, B., et al (2005). Noradrenergic modulation of emotion-induced forgetting and remembering. Journal of Neuroscience, 25, 6343–6349. Hyman, I. E., & Faries, J. M. (1992). The functions of autobiographical memory. In M. A. Conway, D. C. Rubin, H. Spinnler, & W. A. Wagenaar (Eds.), Theoretical perspectives on autobiographical memory (pp. 207–221). Dordrecht, Netherlands: Kluwer Academic Publishers. ICD-10 (1992). Classification of mental and behavioral disorders: Clinical descriptions and diagnostic guidelines. Geneva: World Health Organization. Ingvar, D. H. (1985). ‘‘Memory of the future”: An essay on the temporal organization of conscious awareness. Human Neurobiology, 4, 127–136. Inoue, Y., Tonooka, Y., Yamada, K., & Kanba, S. (2004). Deficiency of theory of mind in patients with remitted mood disorder. Journal of Affective Disorder, 82, 403–409. James, W. (1950). The principles of psychology (Vol. I). New York, NY: Dover Publications. Janet, P. (1907). The major symptoms of hysteria. New York: Macmillan. Janet, P. (1920). The major symptoms of hysteria: Fifteen lectures given in the Medical School of Harvard University (2nd ed.). New York: Macmillan. Joels, M., & Baram, T. Z. (2009). The neuro-symphony of stress. Nature Reviews Neuroscience, 10, 459–466. Kanzer, M. (1939). Amnesia, a statistical study. American Journal of Psychiatry, 96, 711–716. Kapur, N., Millar, J., Colbourn, C., Abbott, P., Kennedy, P., & Docherty, T. (1997). Very long-term amnesia in association with temporal lobe epilepsy: Evidence for multiple-stage consolidation processes. Brain and Cognition, 35, 58–70. Kawashima, R., Sugiura, M., Kato, T., Nakamura, A., Hatano, K., Ito, K., et al (1999). The human amygdala plays an important role in gaze monitoring. Brain, 122, 779–783. Kessler, J., Markowitsch, H. J., Huber, M., Kalbe, E., Weber-Luxenburger, G., & Kock, P. (1997). Massive and persistent anterograde amnesia in the absence of detectable brain damage: Anterograde psychogenic amnesia or gross reduction in sustained effort? Journal of Clinical and Experimental Neuropsychology, 19, 604–614. Klein, S. B., & Gangi, C. E. (2010). The multiplicity of self: Neuropsychological evidence and its implications for the self as a construct in psychological research. Annals of the New York Academy of Sciences, 1191, 1–15. Kleinman, A. (1980). Patients and healers in the context of culture. Beverley, CA: University of California Press. Kobayashi, C., Glover, G. H., & Temple, E. (2007). Cultural and linguistic effects on neural bases of ‘Theory of Mind’ in American and Japanese children. Brain Research, 1164, 95–107. Koenigs, M., Huey, E. D., Raymont, V., Cheon, B., Solomon, J., Wassermann, E. M., et al (2008). Focal brain damage protects against post-traumatic stress disorder in combat veterans. Nature Neuroscience, 11, 232–237. Kopelman, M. D. (2000). Focal retrograde amnesia and the attribution of causality: An exceptionally critical review. Cognitive Neuropsychology, 17, 585–621. Kopelman, M. D. (2008). Retrograde memory loss. In G. Goldenberg & B. Miller (Eds.). Handbook of clinical neurology (Vol. 88, pp. 185–202). Amsterdam: Elsevier. Koponen, S., Taiminen, T., Honkalampi, K., Joukamaa, M., Viinamaki, H., Kurki, T., et al (2005). Alexithymia after traumatic brain injury: Its relation to magnetic resonance imaging findings and psychiatric disorders. Psychosomatic Medicine, 67, 807–812. Krahn, L. E., Bostwick, J. M., & Stonnington, C. M. (2008). Looking toward DSM-V: Should factitious disorder become a subtype of somatoform disorder? Psychosomatics, 49, 277–282. Kritchevsky, M., Chang, J., & Squire, L. R. (2004). Functional amnesia: Clinical description and neuropsychological profile of 10 cases. Learning and Memory, 11, 213–226. Kroll, N., Markowitsch, H. J., Knight, R., & von Cramon, D. Y. (1997). Retrieval of old memories – The temporo-frontal hypothesis. Brain, 120, 1377–1399. Kubicki, M., McCarley, R., Westin, C. F., Park, H. J., Maier, S., Kikinis, R., et al (2007). A review of diffusion tensor imaging studies in schizophrenia. Journal of Psychiatric Research, 41, 15–30. Kühnel, S., Woermann, F. G., Mertens, M., & Markowitsch, H. J. (2008). Involvement of the orbitofrontal cortex during correct and false recognitions of visual stimuli. Implications for eyewitness decisions on an fMRI study using a film paradigm. Brain Imaging and Behavior, 2, 163–176. Kumar, S., Rao, S. L., Sunny, B., & Gangadhar, B. N. (2007). Widespread cognitive impairment in psychogenic anterograde amnesia. Psychiatry and Clinical Neuroscience, 61, 583–586. Kundera, M. (1991). The unbearable lightness of being. New York, NY: Harper Perennial. LaBar, K. S., & Cabeza, R. (2006). Cognitive neuroscience of emotional memory. Nature Reviews Neuroscience, 7, 54–64. Lanius, R. A., Hopper, J. W., & Menon, R. S. (2003). Individual differences in a husband and wife who developed PTSD after a motor vehicle accident: A functional MRI case study. American Journal of Psychiatry, 160, 667–669. Lanius, R. A., Vermetten, E., Loewenstein, R. J., Brand, B., Schmahl, C., Bremner, J. D., et al (2010). Emotion modulation in PTSD: Clinical and neurobiological evidence for a dissociative subtype. American Journal of Psychiatry, 167, 640–647. Lanius, R. A., Williamson, P. C., Bluhm, R. L., Densmore, M., Boksman, K., Neufeld, R. W., et al (2005). Functional connectivity of dissociative responses in posttraumatic stress disorder: A functional magnetic resonance imaging investigation. Biological Psychiatry, 57, 873–884. Leary, M. R. (2007). Motivational and emotional aspects of the self. Annual Reviews of Psychology, 58, 317–344. Lebel, C., Walker, L., Leemans, A., Phillips, L., & Beaulieu, C. (2008). Microstructural maturation of the brain from childhood to adulthood. NeuroImage, 20, 1044–1055. Legrand, D., & Ruby, P. (2009). What is self-specific? Theoretical investigation and critical review of neuroimaging results. Psychological Review, 116, 252–282. Lemogne, C., Bergouignan, L., Piolino, P., Jouvent, R., Allilaire, J. F., & Fossati, P. (2009). Cognitive avoidance of intrusive memories and autobiographical memory: Specificity, autonoetic consciousness, and self-perspective. Memory, 17, 1–7. Lemogne, C., Piolino, P., Friszer, S., Claret, A., Girault, N., Jouvent, R., et al (2006). Episodic autobiographical memory in depression: Specificity, autonoetic consciousness, and self perspective. Consciousness and Cognition, 15, 258–268. Lennox, W. G. (1943). Amnesia, real and feigned. American Journal of Psychiatry, 99, 732–743. Levine, B., Black, S. E., Cabeza, R., Sinden, M., McIntosh, A. R., Toth, J. P., et al (1998). Episodic memory and the self in a case of isolated retrograde amnesia. Brain, 121, 1951–1973. Levine, B., Svoboda, E., Turner, G. R., Mandic, M., & Mackey, A. (2009). Behavioral and functional neuroanatomical correlates of anterograde autobiographical memory in isolated retrograde amnesic patient M.L. Neuropsychologia, 47, 2188–2196. Lind, S. E., & Bowler, S. M. (2009). Recognition memory, self-other source memory, and theory-of-mind in children with autism spectrum disorder. Journal of Autism and Developmental Disorders, 39, 1231–1239. Lipton, M. L., Gellella, E., Lo, C., Gold, T., Ardekani, B. A., Shifteh, K., et al (2008). Multifocal white matter ultrastructural abnormalities in mild traumatic brain injury with cognitive disability: A voxel-wise analysis of diffusion tensor imaging. Journal of Neurotrauma, 25, 1335–1342. Lipton, M. L., Gulko, E., Zimmerman, M. E., Friedman, B. W., Kim, M., Gelella, E., et al (2009). Diffusion-tensor imaging implicates prefrontal axonal injury in executive function impairment following very mild traumatic brain injury. Neurology, 252, 816–824. 798 A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 Loewenstein, R. J. (1991). Psychogenic amnesia and psychogenic fugue: A comprehensive review. American Psychiatric Press Review of Psychiatry, 10, 189–222. Loftus, E. F. (2005). Planting misinformation in the human mind: A 30-year investigation of the malleability of memory. Learning and Memory, 12, 361–366. Lupien, S. J., McEwen, B. S., Gunnar, M. R., & Heim, C. (2009). Effects of stress throughout the lifespan on the brain, behaviour and cognition. Nature Reviews Neuroscience, 10, 434–445. Luria, A. R. (1987). The mind of a mnemonist. A little book about a vast memory. New York: Harvard University Press. Maldonado, J. R., Page, K., Koopman, C., Butler, L. D., Stein, H., & Spiegel, D. (2002). Acute stress reactions following the assassination of Mexican presidential candidate Colosio. Journal of Traumatic Stress, 15, 401–405. Maldonado, J. R., & Spiegel, D. (2008). Dissociative disorders. In E. Hales, S. C. Yudofsky, & G. O. Gabbard (Eds.), The American psychiatric publishing textbook of psychiatry (5th ed., pp. 665–710). Arlington, VA: American Psychiatric Publ. Markowitsch, H. J. (1984). Can amnesia be caused by damage of a single brain structure? Cortex, 20, 27–45. Markowitsch, H. J. (1988). Diencephalic amnesia: A reorientation towards tracts? Brain Research Reviews, 13, 351–370. Markowitsch, H. J. (1992). Intellectual functions and the brain. An historical perspective. Toronto: Hogrefe & Huber Publ. Markowitsch, H. J. (1994). The memory storehouse. Trends in Neuroscience, 17, 513. Markowitsch, H. J. (1995). Which brain regions are critically involved in the retrieval of old autobiographic memory? Brain Research Reviews, 21, 117–127. Markowitsch, H. J. (1996). Organic and psychogenic retrograde amnesia: Two sides of the same coin? Neurocase, 2, 357–371. Markowitsch, H. J. (1999a). Neuroimaging and mechanisms of brain function in psychiatric disorders. Current Opinion in Psychiatry, 12, 331–337. Markowitsch, H. J. (1999b). Functional neuroimaging correlates of functional amnesia. Memory, 7, 561–583. Markowitsch, H. J. (2000). Neuroanatomy of memory. In E. Tulving & F. I. M. Craik (Eds.), The Oxford handbook of memory (pp. 465–484). New York: Oxford University Press. Markowitsch, H. J. (2003). Autonoëtic consciousness. In A. S. David & T. Kircher (Eds.), The self in neuroscience and psychiatry (pp. 180–196). Cambridge: Cambridge University Press. Markowitsch, H. J. (2005). Time, memory, and consciousness. A view from the brain. In R. Buccheri, A. C. Elitzur, & M. Saniga (Eds.), Endophysics, time, quantum, and the subjective (pp. 131–147). Singapore: World Scientific Publ. Markowitsch, H. J. (2008). Anterograde amnesia. In G. Goldenberg & B. L. Miller (Eds.). Handbook of clinical neurology (2nd Series, Vol. 88, pp. 155–183). New York: Elsevier. Markowitsch, H. J., & Brand, M. (2010). Forgetting – An historical perspective. In S. Della Sala (Ed.), Forgetting (pp. 23–34). Hove, UK: Psychology Press. Markowitsch, H. J., Calabrese, P., Fink, G. R., Durwen, H. F., Kessler, J., Härting, C., et al (1997). Impaired episodic memory retrieval in a case of probable psychogenic amnesia. Psychiatry Research: Neuroimaging Section, 74, 119–126. Markowitsch, H. J., Calabrese, P., Haupts, M., Durwen, H. F., Liess, J., & Gehlen, W. (1993). Searching for the anatomical basis of retrograde amnesia. Journal of Clinical and Experimental Neuropsychology, 15, 947–967. Markowitsch, H. J., Calabrese, P., Neufeld, H., Gehlen, W., & Durwen, H. F. (1999). Retrograde amnesia for world knowledge and preserved memory for autobiographic events. A case report. Cortex, 35, 243–252. Markowitsch, H. J., Calabrese, P., Würker, M., Durwen, H. F., Kessler, J., Babinsky, R., et al (1994). The amygdala’s contribution to memory – A PET-study on two patients with Urbach–Wiethe disease. NeuroReport, 5, 1349–1352. Markowitsch, H. J., Emmans, D., Irle, E., Streicher, M., & Preilowski, B. (1985). Cortical and subcortical afferent connections of the primate’s temporal pole: A study of rhesus monkeys, squirrel monkeys, and marmosets. Journal of Comparative Neurology, 242, 425–458. Markowitsch, H. J., Fink, G. R., Thöne, A. I. M., Kessler, J., & Heiss, W.-D. (1997). Persistent psychogenic amnesia with a PET-proven organic basis. Cognitive Neuropsychiatry, 2, 135–158. Markowitsch, H. J., Kessler, J., Kalbe, E., & Herholz, K. (1999). Functional amnesia and memory consolidation. A case of persistent anterograde amnesia with rapid forgetting following whiplash injury. Neurocase, 5, 189–200. Markowitsch, H. J., Kessler, J., Russ, M. O., Frölich, L., Schneider, B., & Maurer, K. (1999). Mnestic block syndrome. Cortex, 35, 219–230. Markowitsch, H. J., Kessler, J., Van der Ven, C., Weber-Luxenburger, G., & Heiss, W.-D. (1998). Psychic trauma causing grossly reduced brain metabolism and cognitive deterioration. Neuropsychologia, 36, 77–82. Markowitsch, H. J., Kessler, J., Weber-Luxenburger, G., Van der Ven, C., Albers, M., & Heiss, W. D. (2000). Neuroimaging and behavioral correlates of recovery from mnestic block syndrome and other cognitive deteriorations. Neuropsychiatry, Neuropsychology, and Behavioral Neurology, 13, 60–66. Markowitsch, H. J., Thiel, A., Kessler, J., von Stockhausen, H.-M., & Heiss, W.-D. (1997). Ecphorizing semi-conscious episodic information via the right temporopolar cortex – A PET study. Neurocase, 3, 445–449. Markowitsch, H. J., Vandekerckhove, M. M. P., Lanfermann, H., & Russ, M. O. (2003). Engagement of lateral and medial prefrontal areas in the ecphory of sad and happy autobiographical memories. Cortex, 39, 643–665. Markowitsch, H. J., von Cramon, D. Y., Hofmann, E., Sick, C.-D., & Kinzler, P. (1990). Verbal memory deterioration after unilateral infarct of the internal capsule in an adolescent. Cortex, 26, 597–609. Markowitsch, H. J., von Cramon, D. Y., & Schuri, U. (1993). Mnestic performance profile of a bilateral diencephalic infarct patient with preserved intelligence and severe amnesic disturbances. Journal of Clinical and Experimental Neuropsychology, 15, 627–652. Markowitsch, H. J., & Welzer, H. (2010). The development of autobiographical memory. Hove, UK: Psychology Press. Masson, J. M. (Ed.). (1985). The complete letters of Sigmund Freud to Wilhelm Fliess 1887–1904. Cambridge, MA: The Belknap Press of Harvard University Press. Mattila, A. K., Saarni, S. I., Salminen, J. K., Huhtala, H., Sintonen, H., & Joukamaa, M. (2009). Alexithymia and health-related quality of life in a general population. Psychosomatics, 50, 59–68. Mayou, R., Kirmayer, L. J., Simon, G., Kroenke, K., & Sharpe, M. (2005). Somatoform disorders: Time for a new approach in DSM-V. American Journal of Psychiatry, 162, 847–855. McCullough, J. P. Jr., (2000). Treatment for chronic depression. Cognitive behavioural analysis system of psychotherapy. New York, NY: Guilford Press. McGowan, P. O., Sasaki, A., D’Alessio, A. C., Dymov, S., Labonte, B., Szyf, M., et al (2009). Epigenetic regulation of the glucocorticoid receptor in human brain associates with childhood abuse. Nature Neuroscience, 12, 342–348. McKay, G. C. M., & Kopelman, M. D. (2009). Psychogenic amnesia: When memory complaints are medically unexplained. Advances in Psychiatric Treatment, 15, 152–158. Mehta, M. A., Golembo, N. I., Nosarti, C., Colvert, E., Mota, A., Williams, S. C., et al (2009). Amygdala, hippocampal and corpus callosum size following severe early institutional deprivation: The English and Romanian adoptees study pilot. Journal of Child Psychology and Psychiatry, 50, 943–951. Metzinger, T. (2008). Empirical perspectives from the self-model theory of subjectivity: A brief summary with examples. Progress in Brain Research, 168, 215–245. Miller, G. A. (1956). The magical number seven plus minus two. Some limits on our capacity for processing information. Psychological Review, 63, 244–257. Miller, C. A., & Sweatt, J. D. (2006). Amnesia or retrieval deficit? Implications of a molecular approach to the question of reconsolidation. Learning and Memory, 13, 498–505. Mishkin, M., & Petri, H. L. (1984). Memories and habits: Some implications for the analysis of learning and retention. In L. R. Squire & N. Butters (Eds.), Neuropsychology of memory (pp. 287–296). New York: Guilford Press. Mitchell, S. A., & Black, M. J. (1995). Freud and beyond. A history of modern psychoanalytic thought. New York: Basic Books. Mitchell, J. P., Macrae, C. M., & Banaji, M. R. (2006). Dissociable medial prefrontal contributions to judgments of similar and dissimilar others. Neuron, 50, 655–663. Modell, A. H. (2006). Imagination and the meaningful brain. Cambridge, MA: Bradford. Modestin, J., Lotscher, K., & Erni, T. (2002). Dissociative experiences and their correlates in young non-patients. Psychology and Psychotherapy, 75, 53–64. A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 799 Moldoveanu, M., & Stevenson, H. (2001). The self as a problem: The intra-personal coordination of conflicting desires. Journal of Socio-Economics, 30, 295–330. Mondadori, C. R., Buchmann, A., Mustovic, H., Schmidt, C. F., Boesiger, P., Nitsch, R. M., et al (2006). Enhanced brain activity may precede the diagnosis of Alzheimer’s disease by 30 years. Brain, 129, 2908–2922. Mooney, G., & Speed, J. (2001). The association between mild traumatic brain injury and psychiatric conditions. Brain injury, 15, 865–877. Moriguchi, Y., Ohnishi, T., Decety, J., Hirakata, M., Maeda, M., Matsuda, H., et al (2009). The human mirror neuron system in a population with deficient selfawareness: An fMRI study in alexithymia. Human Brain Mapping, 30, 2063–2076. Moriguchi, Y., Ohnishi, T., Lane, R. D., Maeda, M., Mori, T., Nemoto, K., et al (2006). Impaired self-awareness and theory of mind: An fMRI study of mentalizing in alexithymia. NeuroImage, 32, 1472–1482. Morin, A. (2009). Self-awareness deficits following loss of inner speech: Dr. Jill Bolte Taylor’s case study. Consciousness and Cognition, 18, 524–529. Morin, A., & Michaud, J. (2007). Self-awareness and the left inferior frontal gyrus: Inner speech use during self-related processing. Brain Research Bulletin, 74, 387–396. Nagel, T. (1974). What is it like to be a bat? Philosophical Review, 82, 435–456. Naqvi, N. H., Rudrauf, D., Damasio, H., & Bechara, A. (2007). Damage to the insula disrupts addiction to cigarette smoking. Science, 315, 531–534. Nelson, K. (2003). Narrative and self, myth and memory: Emergence of the cultural self. In R. Fivush & C. Haden (Eds.), The development of autobiographical memory and self-understanding (pp. 3–28). Hillsdale, NJ: Erlbaum Assoc. Nelson, K. (2005). Evolution and development of human memory systems. In B. J. Ellis & D. F. Bjorklund (Eds.), Origins of the social mind (pp. 319–345). New York: Guilford Press. Nelson, K., & Fivush, R. (2004). The emergence of autobiographical memory: A social cultural developmental theory. Psychological Review, 111, 486–511. Newen, A., & Schlicht, T. (2009). Understanding other minds: A criticism of Goldman’s simulation theory and an outline of the person model theory. Grazer Philosophische Studien, 79, 209–242. Newen, A., & Vogeley, K. (2003). Self-representation: Searching for a neural signature of self-consciousness. Consciousness and Cognition, 12, 529–543. Nieuwenhuys, R., Voogd, J., & van Huijzen, C. (1978). The human central nervous system. Berlin: Springer. O’Brien, J. T. (1997). The ‘glucocorticoid cascade’ hypothesis in man. British Journal of Psychiatry, 170, 199–201. O’Connor, M., Sieggreen, M. A., Ahern, G., Schomer, D., & Mesulam, M. (1997). Accelerated forgetting in association with temporal lobe epilepsy and paraneoplastic encephalitis. Brain and Cognition, 35, 71–84. Oddo, S., Lux, S., Weiss, P. H., Schwab, A., Welzer, H., Markowitsch, H. J., et al (2010). Specific role of medial prefrontal cortex in retrieving recent autobiographical memories: An fMRI study of young female subjects. Cortex, 46, 29–39. Ouellet, J., Rouleau, I., Labrecque, R., Bernier, G., & Scherzer, P. B. (2008). Two routes to losing one’s past life: A brain trauma, an emotional trauma. Behavioural Neurology, 20, 27–38. Pacheco, J., Beevers, C. G., Benavides, C., McGeary, J., Stice, E., & Schnyer, D. M. (2009). Frontal-limbic white matter pathway associations with the serotonin transporter gene promoter region (5-HTTLPR) polymorphism. Journal of Neuroscience, 29, 6229–6233. Parker, E. S., Cahill, L., & McGaugh, J. L. (2006). A case of unusual autobiographical remembering. Neurocase, 12, 35–49. Paus, T. (2010). Growth of white matter in the adolescent brain: Myelin or axon? Brain and Cognition, 72, 26–35. Pearce, J. M. (2007). Amnesia. European Neurology, 57, 126. Peigneux, P., Schmitz, R., & Urbain, C. (2010). Sleep and forgetting. In S. Della Sala (Ed.), Forgetting (pp. 165–184). Hove, East Sussex: Psychology Press. Perner, J. (2000). Memory and theory of mind. In E. Tulving & F. I. M. Craik (Eds.), The Oxford handbook of memory (pp. 297–314). New York: Oxford University Press. Perner, J. (2001). Episodic memory: Essential distinctions and developmental implications. In C. Moore & K. Lemmon (Eds.), The self in time: Developmental perspectives (pp. 181–202). Hillsdale, NJ: Erlbaum. Pessoa, L. (2008). On the relationship between emotion and cognition. Nature Reviews Neuroscience, 9, 148–158. Peters, J., & Büchel, C. (2010). Episodic future thinking reduces rewards delay discounting through an enhancement of prefrontal–mediotemporal interactions. Neuron, 66, 138–148. Pfeifer, J. H., Dapretto, M., & Lieberman, M. D. (2010). The neural foundations of evaluative self-knowledge in middle childhood, early adolescence and adulthood. In P. D. Zelazo, M. Chandler, & E. Crone (Eds.), Developmental social cognitive neuroscience (pp. 141–163). New York, NY: Psychology Press. Phillips, M. L., Drevets, W. C., Rauch, S. L., & Lane, R. (2003). Neurobiology of emotion perception. II: Implications for major psychiatric disorders. Biological Psychiatry, 54, 515–528. Piefke, M., Weiss, P. H., Markowitsch, H. J., & Fink, G. R. (2005). Gender differences in the functional neuroanatomy of emotional episodic autobiographical memory. Human Brain Mapping, 24, 313–324. Piolino, P., Desgranges, B., & Eustache, F. (2009). Episodic autobiographical memories over the course of time: Cognitive, neuropsychological and neuroimaging findings. Neuropsychologia, 47, 2314–2329. Piolino, P., Hannequin, D., Desgranges, B., Girard, C., Beaunieux, H., Gittard, B., et al (2005). Right ventral frontal hypometabolism and abnormal sense of self in a case of disproportionate retrograde amnesia. Cognitive Neuropsychology, 22, 1005–1034. Piolino, P., Hisland, M., Ruffeveille, I., Matuszewski, V., Jambaque, I., & Eustache, F. (2007). Do school-age children remember or know the personal past? Consciousness and Cognition, 16, 84–101. Pizzagalli, D. A., Holmes, A. J., Dillon, D. G., Goetz, E. L., Birk, J. L., Bogdan, R., et al (2009). Reduced caudate and nucleus accumbens response to rewards in unmedicated individuals with major depressive disorder. American Journal of Psychiatry, 166, 702–710. Plaza, M., Gatignol, P., Leroy, M., & Duffau, H. (2009). Speaking without Broca’s area after tumor resection. Neurocase, 15, 284–310. Post, R. M., Weiss, S. R., Smith, M., Rosen, J., & Frye, M. (1995). Stress, conditioning, and the temporal aspects of affective disorders. Annals of the New York Academy of Sciences, 771, 677–696. Povinelli, D. J., Landau, K. R., & Perilloux, H. K. (1996). Self-recognition in young children using delayed versus live feedback: Evidence of a developmental asynchrony. Child Development, 67, 1540–1554. Pugliese, L., Catani, M., Ameis, S., Dell’Acqua, F., Thiebaut de Schotten, M., et al (2009). The anatomy of extended limbic pathways in Asperger syndrome: A preliminary diffusion tensor imaging tractography study. NeuroImage, 47, 427–434. Putnam, F. (1997). Dissociation in children and adolescents. A developmental perspective. New York: Guilford Press. Rabin, J. S., Gilboa, A., Stuss, D., Mar, R. A., & Rosenbaum, R. S. (2010). Common and unique neural correlates of autobiographical memory and theory of mind. Journal of Cognitive Neuroscience, 22, 1095–1111. Rao, S. M., Mayer, A. R., & Harrington, D. L. (2001). The evolution of brain activation during temporal processing. Nature Neuroscience, 4, 317–323. Rasch, B., Spalek, K., Buholzer, S., Luechinger, R., Boesiger, P., Papassotiropoulos, A., et al (2009). A genetic variation of the noradrenergic system is related to differential amygdala activation during encoding of emotional memories. Proceedings of the National Academy of Sciences of the USA, 105, 19191–19196. Rathbone, C. J., Moulin, C. J., & Conway, M. A. (2009). Autobiographical memory and amnesia: Using conceptual knowledge to ground the self. Neurocase, 15, 405–418. Ray, R. D., Shelton, A. L., Hollon, N. G., Matsumoto, D., Frankel, C. B., Gross, J. J., et al (2010). Interdependent self-construal and neural representations of self and mother. Social Cognitive and Affective Neuroscience, 5, 318–323. Reinhold, N., & Markowitsch, H. J. (2008). Theory of mind and the sense of self. Science Online [Letter to the editor]. Reinhold, N., Kühnel, S., Brand, M., & Markowitsch, H. J. (2006). Functional neuroimaging in memory and memory disturbances. Current Medical Imaging Reviews, 2, 35–57. Reinhold, N., & Markowitsch, H. J. (2007). Emotion and consciousness in adolescent psychogenic amnesia. Journal of Neuropsychology, 1, 53–64. 800 A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 Reinhold, N., & Markowitsch, H. J. (2009). Retrograde episodic memory and emotion: A perspective from patients with dissociative amnesia. Neuropsychologia, 47, 2197–2206. Rizzolatti, G., Fabbri-Destro, M., & Cattaneo, L. (2009). Mirror neurons and their clinical relevance. Nature clinical practice. Neurology, 5, 24–34. Robertson, D., Snarey, J., Ousley, O., Harenski, K., DuBois Bowman, F., Gilkey, R., et al (2007). The neural processing of moral sensitivity to issues of justice and care. Neuropsychologia, 45, 755–766. Romei, V., De Gennaro, L., Fratello, F., Curcio, G., Ferrara, M., Pascual-Leone, A., et al (2008). Interhemispheric transfer deficit in alexithymia: A transcranial magnetic stimulation study. Psychotherapy and Psychosomatics, 77, 175–181. Roozendaal, B., McEwen, B. S., & Chattarji, S. (2009). Stress, memory and the amygdala. Nature Reviews Neuroscience, 10, 429–433. Rosenbaum, R. S., Gilboa, A., Levine, B., Winocur, G., & Moscovitch, M. (2009). Amnesia as an impairment of detail generation and binding: Evidence from personal, fictional, and semantic narratives in K.C. Neuropsychologia, 47, 2181–2187. Rosenbaum, R. S., Stuss, D. T., Levine, B., & Tulving, E. (2007). Theory of mind is independent of episodic memory. Science, 318, 1257. Roth, T. I., & Sweatt, J. D. (2009). Regulation of chromatin structure in memory formation. Current Opinion in Neurobiology, 19, 336–342. Ruff, R. M., Iverson, G. L., Barth, J. T., Bush, S. S., & Broshek, D. K. (2009). Recommendations for diagnosing a mild traumatic brain injury: A National Academy of Neuropsychology education paper. Archives of Clinical Neuropsychology, 24, 3–10. Rutter, M., Moffitt, T. E., & Caspi, A. (2006). Gene-environment interplay and psychopathology: Multiple varieties but real effects. Journal of Childhood Psychology and Psychiatry and Allied Disciplines, 47, 226–261. Saxe, R., Moran, J. M., Scholz, J., & Gabrieli, J. (2006). Overlapping and non-overlapping brain regions for theory of mind and self reflection in individual subjects. Social Cognitive and Affective Neuroscience, 1, 229–234. Schacter, D. L., & Addis, D. R. (2009). On the nature of medial temporal lobe contributions to the constructive simulation of future events. Philosophical Transactions of the Royal Society, London B, 364, 1245–1253. Schilbach, L., Wohlschlaeger, A. M., Kraemer, N. C., Newen, A., Shah, N. J., Fink, G. R., et al (2006). Being with virtual others: Neural correlates of social interaction. Neuropsychologia, 44, 718–730. Schneider, K. (1912). Über einige klinisch-psychologische Untersuchungsmethoden und ihre Ergebnisse. [On some clinical–pathological methods of investigation and their results]. Zeitschrift für die gesamte Neurologie und Psychiatrie, 8, 553–615. Schneider, K. (1928). Die Störungen des Gedächtnisses [Disturbances of memory]. In O. Bumke (Ed.). Handbuch der Geisteskrankheiten (Vol. 1, pp. 508–529). Berlin: Springer. Schore, A. N. (2002). Dysregulation of the right brain: A fundamental mechanism of traumatic attachment and the psychopathogenesis of posttraumatic stress disorder. Australian and New Zealand Journal of Psychiatry, 36, 9–30. Schulte-Rüther, M., Markowitsch, H. J., Fink, G. R., & Piefke, M. (2007). Mirror neuron and theory of mind mechanisms involved in face-to-face interactions: An fMRI approach to empathy. Journal of Cognitive Neuroscience, 19, 1354–1372. Schulte-Rüther, M., Markowitsch, H. J., Shah, N. J., Fink, G. R., & Piefke, M. (2008). Gender differences in the functional neuroanatomy of emotional perspective taking. NeuroImage, 14, 393–403. Seeley, W. W. (2008). Selective functional, regional, and neuronal vulnerability in frontotemporal dementia. Current Opinion in Neurology, 21, 701–707. Seligman, R., & Kirmayer, L. J. (2008). Dissociative experience and cultural neuroscience: Narrative, metaphor and mechanism. Culture, Medicine and Psychiatry, 32, 31–64. Semon, R. (1904). Die Mneme als erhaltendes Prinzip im Wechsel des organischen Geschehens. Leipzig: Wilhelm Engelmann. Sepulcre, J., Masdeu, J. C., Sastre-Garriga, J., Goni, J., Velez-de-Mendizabal, N., Duque, B., et al (2008). Mapping the brain pathways of declarative verbal memory: Evidence from white matter lesions in the living human brain. NeuroImage, 42, 1237–1243. Serra, L., Fadda, L., Buccione, I., Caltagirone, C., & Carlesimo, G. A. (2007). Psychogenic and organic amnesia: A multidimensional assessment of clinical, neuroradiological, neuropsychological and psychopathological features. Behavioural Neurology, 18, 53–64. Shalom, D. B. (2003). Memory in autism: Review and synthesis. Cortex, 39, 1129–1138. Shamay-Tsoory, S. G., & Aharon-Peretz, J. (2007). Dissociable prefrontal networks for cognitive and affective theory of mind: A lesion study. Neuropsychologia, 45, 3054–3067. Shamay-Tsoory, S. G., Aharon-Peretz, J., & Perry, D. (2009). Two systems for empathy: A double dissociation between emotional and cognitive empathy in inferior frontal gyrus versus ventromedial prefrontal lesions. Brain, 132, 617–627. Shamay-Tsoory, S., Harari, H., Szepsenwol, O., & Levkovitz, Y. (2009). Neuropsychological evidence of impaired cognitive empathy in euthymic bipolar disorder. Journal of Neuropsychiatry and Clinical Neurosciences, 21, 59–67. Sharot, T., Riccardi, A. M., Raio, C. M., & Phelps, E. A. (2007). Neural mechanisms mediating optimism bias. Nature, 450, 102–105. Shaw, P., Lawrence, E. J., Radbourne, C., Bramham, J., Polkey, C. E., & David, A. S. (2004). The impact of early and late damage to the human amygdala on ‘theory of mind’ reasoning. Brain, 127, 1535–1548. Siebert, M., Markowitsch, H. J., & Bartel, P. (2003). Amygdala, affect, and cognition: Evidence from ten patients with Urbach–Wiethe disease. Brain, 126, 2627–2637. Siegel, D. J. (2004). Memory: An overview, with emphasis on developmental, interpersonal, and neurobiological aspects. Journal of the American Academy of Child and Adolescent Psychiatry, 40, 997–1011. Smith, C. N., Frascino, J. C., Kripke, D. L., McHugh, P. R., Treisman, G. J., & Squire, L. R. (2010). Losing memories overnight: A unique form of human amnesia. Neuropsychologia (Epub ahead of print). Spiegel, D. (2006). Recognizing traumatic dissociation. American Journal of Psychiatry, 163, 566–568. Spitzer, C., & Freyberger, H. J. (2008). Geschlechtsunterschiede bei dissoziativen Störungen. Bundesgesundheitsblatt Gesundheitsforschung, Gesundheitsschutz, 51, 46–52. Spreng, R. N., & Grady, C. L. (2010). Patterns of brain activity supporting autobiographical memory, prospection, and theory of mind, and their relationship to the default mode network. Journal of Cognitive Neuroscience, 22, 1112–1123. Spreng, R. N., Mar, R. A., & Kim, A. S. N. (2008). The common neural basis of autobiographical memory, prospection, navigation, theory of mind and the default mode: A quantitative meta-analysis. Journal of Cognitive Neuroscience, 21, 489–510. St Jacques, P. L., Bessette-Symons, B., & Cabeza, R. (2009). Functional neuroimaging studies of aging and emotion: Fronto-amygdalar differences during emotional perception and episodic memory. Journal of the International Neuropsychological Society, 15, 819–825. Staniloiu, A., & Markowitsch, H. J. (2010). Looking at comorbidity through the glasses of neuroscientific memory research: A brain network perspective. Behavioral and Brain Sciences, 33, 170–171. Staniloiu, A., & Markowitsch, H. J. (in press). Searching for the anatomy of dissociative amnesia. Journal of Psychology. Staniloiu, A., Bender, A., Smolewska, K., Ellis, J., Abramowitz, C., & Markowitsch, H. J. (2009). Ganser syndrome with work-related onset in a patient with a background of immigration. Cognitive Neuropsychiatry, 14, 180–198. Staniloiu, A., Borsutzky, S., & Markowitsch, H. J. (2010). Dissociative memory disorders and immigration. In W. Christensen, E. Schier, & J. Sutton (Eds.), ASCS09: Proceedings of the 9th conference of the Australasian Society for Cognitive Science (pp. 316–324). Sydney: Macquarie Centre for Cognitive Science. Steinberg, M. (2000). Dissociative amnesia. In B. J. Sadock & V. A. Sadock (Eds.). Comprehensive textbook of psychiatry (7th ed., Vol. 1, pp. 1544–1549). Philadelphia, PA: Lippincott, Williams & Wilkins. Stickgold, R., & Walker, M. P. (2005). Memory consolidation and reconsolidation: What is the role of sleep? Trends in Neurosciences, 28, 408–415. Sturm, V. E., Rosen, H. J., Allison, S., Miller, B. L., & Levenson, R. W. (2006). Self-conscious emotion deficits in frontotemporal lobar degeneration. Brain, 129, 2508–2516. Suddendorf, T., Addis, D. R., & Corballis, M. C. (2009). Mental time travel and the shaping of the human mind. Philosophical Transactions of the Royal Society London B, 364, 1317–1324. A. Staniloiu et al. / Consciousness and Cognition 19 (2010) 778–801 801 Taylor, W. D., Macfall, J. R., Gerig, G., & Krishnan, R. R. (2007). Structural integrity of the uncinate fasciculus in geriatric depression: Relationship with age of onset. Neuropsychiatric Disease and Treatment, 3, 669–674. Thomas-Anterion, C., Guedj, E., Decousus, M., & Laurent, B. (2010). Can we see personal identity loss? A functional imaging study of typical ‘hysterical amnesia’. Journal of Neurology, Neurosurgery, and Psychiatry, 81, 468–469. Thomas-Antérion, C., Mazzola, L., Foyatier-Michel, N., & Laurent, B. (2008). À la recherche de la mémoire perdue: Nature des troubles et mode de récupération d’un cas d’amnésie rétrograde pure. Revue Neurologique, 164, 271–277. Tomasello, M. (2008). Origins of human communication. Cambridge, MA: MIT Press. Tramoni, E., Aubert-Khalfa, S., Guye, M., Ranjeva, J. P., Felician, O., & Ceccaldi, M. (2009). Hypo-retrieval and hyper-suppression mechanisms in functional amnesia. Neuropsychologia, 47, 611–624. Tranel, D., & Bechara, A. (2009). Sex-related functional asymmetry of the amygdala: Preliminary evidence using a case-matched lesion approach. Neurocase, 15, 1–18. Tulving, E. (1972). Episodic and semantic memory. In E. Tulving & W. Donaldson (Eds.), Organization of memory (pp. 381–403). New York: Academic Press. Tulving, E. (1983). Elements of episodic memory. Oxford: Clarendon Press. Tulving, E. (1985). Memory and consciousness. Canadian Psychology, 26, 1–12. Tulving, E. (1995). Organization of memory: Quo vadis? In M. S. Gazzaniga (Ed.), The cognitive neurosciences (pp. 839–847). Cambridge, MA: MIT Press. Tulving, E. (2000). Concepts of memory. In E. Tulving & F. I. M. Craik (Eds.), The Oxford handbook of memory (pp. 33–43). Oxford: Oxford University Press. Tulving, E. (2002). Episodic memory: From mind to brain. Annual Reviews of Psychology, 53, 1–25. Tulving, E. (2005). Episodic memory and autonoesis: Uniquely human? In H. Terrace & J. Metcalfe (Eds.), The missing link in cognition: Evolution of selfknowing consciousness (pp. 3–56). New York: Oxford University Press. Tulving, E., & Kim, A. (2007). The medium and the message of mental time travel. Behavioral and Brain Sciences, 30, 334–335. Ulrich-Lai, Y. M., & Herman, J. P. (2009). Neural regulation of endocrine and autonomic stress responses. Nature Reviews Neuroscience, 10, 397–409. Valentino, K., Toth, S. L., & Cicchetti, D. (2009). Autobiographical memory functioning among abused, neglected, and nonmaltreated children: The overgeneral memory effect. Journal of Child Psychology and Psychiatry, 50, 1029–1038. Van der Hart, O., & Dorahy, M. (2006). Pierre Janet and the concept of dissociation. American Journal of Psychiatry, 163, 1646 [Letter to the editor]. van der Hart, O., & Nijenhuis, E. (2001). Generalized dissociative amnesia: Episodic, semantic and procedural memories lost and found. The Australian and New Zealand Journal of Psychiatry, 35, 589–600. Vann, S. D. (2009). Gudden’s ventral tegmental nucleus is vital for memory: Re-evaluating diencephalic inputs for amnesia. Brain, 132, 2372–2384. Vargha-Khadem, F., Gadian, D. G., Watkins, K. E., Connelly, A., Van Paesschen, W., & Mishkin, M. (1997). Differential effects of early hippocampal pathology on episodic and semantic memory. Science, 277, 376–380. Vercammen, A., Knegtering, H., Bruggeman, R., & Aleman, A. (2010). Subjective loudness and reality of auditory verbal hallucinations and activation of the inner speech processing network. Schizophrenia Bulletin, 8, 4–23. Vermetten, E., Schmahl, C., Lindner, S., Loewenstein, R. J., & Bremner, J. D. (2006). Hippocampal and amygdalar volumes in dissociative identity disorder. American Journal of Psychiatry, 163, 630–636. Vogeley, K., & Kupke, C. (2007). Disturbances of time consciousness from a phenomenological and a neuroscientific perspective. Schizophrenia Bulletin, 33, 157–165. Vuilleumier, P., & Pourtois, G. (2007). Distributed and interactive brain mechanisms during emotion face perception: Evidence from functional neuroimaging. Neuropsychologia, 45, 174–194. Vythilingam, M., Heim, C., Newport, J., Miller, A. H., Anderson, E., Bronen, R., et al (2002). Childhood trauma associated with smaller hippocampal volume in women with major depression. American Journal of Psychiatry, 159, 2072–2080. Wang, Q. (2001). Culture effects on adults’ earliest childhood recollection and self-description: Implications for the relation between memory and the self. Journal of Personality and Social Psychology, 81, 220–233. Wang, F., Kalmar, J. H., He, Y., Jackowski, M., Chepenik, L. G., Edmiston, E. E., et al (2009). Functional and structural connectivity between the perigenual anterior cingulate and amygdala in bipolar disorder. Biological Psychiatry, 66, 516–521. Weaver, I. C., Champagne, F. A., Brown, S. E., Dymov, S., Sharma, S., Meaney, M. J., et al (2005). Reversal of maternal programming of stress responses in adult offspring through methyl supplementation: Altering epigenetic marking later in life. Journal of Neuroscience, 25, 11045–11054. Welch-Ross, M. K., Diecidue, K., & Miller, S. A. (1997). Young children’s understanding of conflicting mental representation predicts suggestibility. Developmental Psychology, 33, 43–53. Welzer, H., & Markowitsch, H. J. (2005). Towards a bio-psycho-social model of autobiographical memory. Memory, 13, 63–78. Wheeler, M. A., Stuss, D. T., & Tulving, E. (1997). Toward a theory of episodic memory: The frontal lobes and autonoetic consciousness. Psychological Bulletin, 121, 331–354. Williams, J. M., & Scott, J. (1988). Autobiographical memory in depression. Psychological Medicine, 18, 689–695. Wimmer, H., & Perner, J. (1983). Beliefs about beliefs: Representation and constraining function of wrong beliefs in young children’s understanding of deception. Cognition, 13, 103–128. Wood, H. (2003). Consolidating reconsolidation. Nature, 4, 774–775. Yasmin, H., Nakata, Y., Aoki, S., Abe, O., Sato, N., Nemoto, K., et al (2008). Diffusion abnormalities of the uncinate fasciculus in Alzheimer’s disease: Diffusion tensor tract-specific analysis using a new method to measure the core of the tract. Neuroradiology, 50, 293–299. Yehuda, R., Schmeidler, J., Siever, L. J., Binder-Brynes, K., & Elkin, A. (1997). Individual differences in posttraumatic stress disorder symptom profiles in Holocaust survivors in concentration camps or in hiding. Journal of Trauma and Stress, 10, 453–463. Zanarini, M. C., Frankenburg, F. R., Jager-Hyman, S., Reich, D. B., & Fitzmaurice, G. (2008). The course of dissociation for patients with borderline personality disorder and axis II comparison subjects: a 10-year follow-up study. Acta Psychiatrica Scandinavica, 118, 291–296.

Consciousness and Cognition Consciousness and Cognition 14 (2005) 351–376 www.elsevier.com/locate/concog Accuracy of familiarity decisions to famous faces perceived without awareness depends on attitude to the target person and on response latency Anna Stone*, Tim Valentine Department of Psychology, Goldsmiths College, University of London, New Cross, London SE14 6NW, United Kingdom Received 11 July 2004 Available online 2 November 2004 Abstract Stone and Valentine (2004) presented masked 17 ms faces in simultaneous pairs of one famous and one unfamiliar face. Accuracy in selecting the famous face was higher when the famous person was regarded as ‘‘good’’ or liked than when regarded as ‘‘evil’’ or disliked. Experiment 1 attempted to replicate this phenomenon, but produced a different pattern of results. Experiment 2 investigated alternative explanations and found evidence supporting only the effect of response latency: responses made soon after stimulus onset were more accurate to liked than to disliked faces, whereas responses made after a longer delay were equally accurate to disliked faces. It appears that the effect of negative valence was corrected within the space of a few hundred milliseconds. Experiment 3, using an affective priming paradigm, supported the concept that an early-arising effect of valence is corrected if it is misleading to the directed task.
2004 Elsevier Inc. All rights reserved. Keywords: Non-conscious perception; Facial identity; Affective priming; Automatic; Awareness; Visual masking 1. Introduction There is much evidence that facial expressions can be detected, and can influence psychophysiological and behavioural responses, without awareness of the expression (e.g., Dimberg * Corresponding author. E-mail address: pss01as@gold.ac.uk (A. Stone). 1053-8100/$ - see front matter
2004 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2004.09.002 352 A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 & Ohman, 1996; Dimberg, Thunberg, & Elmehed, 2000; Johnsen & Hugdahl, 1991, 1993; Mogg & Bradley, 1999; Murphy & Zajonc, 1993; Niedenthal, 1990; Ohman, Esteves, & Soares, 1995; Robinson, 1998; Saban & Hugdahl, 1999; Whalen et al., 1998; Wong, Shevrin, & Williams, 1994). All of these studies presented masked faces for very brief exposure duration [target-to-mask stimulus onset asynchrony (SOA) of less than 35 ms]. Participants were at chance in two-alternative forced-choice tasks of identifying the facial expression, confirming the absence of awareness of the expression. The recognition without awareness of facial emotional expressions is often interpreted in terms of the importance to the individual of detecting the emotion of others. The question then arises of whether facial identities, like facial expressions, can be recognised without awareness of identity. Literature relevant to this question will be examined. Banse (1999) presented the face or name of the participant or a relationship partner as the prime stimulus, for 10.5 ms with backward masking, followed after SOA of 42 ms by a Chinese letter (re Murphy & Zajonc, 1993). Targets were evaluated more positively when preceded by the partnerÕs face/name than by the participantÕs own face/name: this was the predicted effect, based on the observation that a partner tends to be evaluated more positively than the self. The similarity in the results obtained from face and name primes was taken as implying that the person schemata had been activated. One drawback is that accuracy was well above chance in a two-alternative forced-choice of self or other, with a single face or name presented under the same masked conditions, so it is not clear that participants were entirely unaware of the face or name identity. The limited number of stimuli raises the possibility that even if participants were able to perceive only a vague outline of the masked face, this could have sufficed to enable a correct decision about which of the two persons had been presented. Banse (2001) used an affective priming paradigm to investigate whether famous and personally familiar faces might be recognised without awareness of identity. Primes were presented for the same duration used by Banse (1999) and again, it is not clear that participants were entirely unaware of facial identity. Stone, Valentine, and Davis (2001) reported that responses to famous faces perceived without awareness of facial identity differed according to valence. Experiment 1 found that skin conductance responses to masked 17 ms faces were higher to the faces of famous persons subsequently evaluated ‘‘good’’ than to the faces of persons evaluated ‘‘evil,’’ but did not distinguish between famous and unfamiliar faces. (Responses tended to be higher to good faces than to unfamiliar faces, but tended to be lower to evil faces than to unfamiliar faces.) When faces were exposed for 220 ms, a duration that permits conscious recognition, there was an effect of familiarity but no effect of valence: skin conductance responses were higher to famous faces than to unfamiliar faces with no difference between ‘‘good’’ and ‘‘evil’’ faces. Responses were above chance accuracy in a two-alternative forced-choice of ‘‘good’’ or ‘‘evil’’ to masked 17 ms faces of whose identity participants were unaware (Experiment 3). Stone and Valentine (in press) used a procedure based on Mogg and Bradley (1999). Masked 17 ms faces were presented in simultaneous pairs of a famous and an unfamiliar face, matched on physical characteristics, in LVF and RVF. These were followed by a dot-probe in either LVF or RVF to which participants made a speeded two-alternative forced-choice discrimination response. Orientation of attention towards the famous face would be demonstrated by faster or more accurate responses to the dot-probe when it appeared in the same VF as the famous face. A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 353 Participants were subsequently asked to evaluate each famous person as ‘‘good’’ or ‘‘evil’’ on a 7-point scale from 3 (very evil) to +3 (very good). Fewer errors were made when the dotprobe was presented in the same VF as the famous face compared to the opposite VF, as long as the famous person was evaluated as neutral or good (evaluation from 1 to 3). A reverse effect was observed, with more errors to dot-probes presented in the VF of the famous face, when the famous person was evaluated as evil ( 3 or 2). The within-item analysis, comparing performance between participants who had evaluated the same famous persons as good-neutral or evil, confirmed that the effect was due to participantsÕ reactions to the famous persons and not to any confounding factor. This effect was interpreted as the orientation of attention towards the faces of famous persons evaluated as ‘‘good’’ but not towards those evaluated as ‘‘evil.’’ In a separate awareness check task, the same masked 17 ms famous–unfamiliar face pairs were presented simultaneously while participants attempted to select the famous face. Overall, accuracy at chance supported participant claims of no awareness of facial familiarity or identity. At the same time, responses were more accurate to the faces of famous persons evaluated as ‘‘good’’ than to the faces of persons evaluated as ‘‘evil.’’ This was published as Experiment 1 in Stone and Valentine, 2004. One limitation is that, without specification of the processes underlying the valence effect, the possibility cannot be ruled out that it arose from some particular (and unknown) characteristic of the specific stimulus set. For this reason it was considered advisable to repeat the experiment using a different set of stimuli and a new participant sample. Experiment 1 of the present paper was an attempt to replicate the valence effect with a larger set of stimuli. As will be seen, Experiment 1 did not replicate the valence effect: response accuracy was not consistently lower for negative than positive faces. Experiment 2 contrasted three potential explanations for the results of Experiment 1 and found evidence supporting only one, the influence of response latency. Experiment 3 used an affective priming task to examine another prediction derived from the results of Experiment 1. 2. Experiment 1 Experiment 1 was an attempt to replicate the effect of valence in the awareness check task with a larger set of stimuli. Consequent to the change in the stimulus set, the terms positive and negative were used in the evaluation of the famous persons, replacing the terms good and evil that were used by Stone and Valentine (2004). The stimuli were selected on the expectation that each famous person would be evaluated as positive by some participants and negative by others. Balanced valence ratings would permit a rigorous within-item analysis, in which accuracy could be compared for the same famous–unfamiliar face pair between participants evaluating the famous person positive vs. negative, ruling out confounds arising from systematic differences between famous and unfamiliar faces, or among famous faces. If the same famous–unfamiliar face-pair results in different responses depending on participantsÕ evaluations of the famous person, then this will support the claim that responses depend on participantsÕ attitudes and not on any confounding factor. The prediction was that responses would be more accurate to the faces of famous persons rated positive than negative. 354 A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 2.1. Method 2.1.1. Participants Participants were 46 students at Goldsmiths College, London. Data were excluded from eight participants who failed to evaluate a minimum of 10% of the famous people as negative in the post-experimental evaluation, since it was suspected that these participants might not have been using the scale correctly. Data were excluded from five participants who selected more famous faces than expected by chance (binomial distribution, one-tailed, cut-off = 55.6%) since for these participants, the possibility of some awareness of facial familiarity could not be ruled out. The remaining 33 participants were 24 female and nine male, aged between 18 and 50, mean 23.4, SD 7.9 years. All of the participants had been resident in the UK for at least 10 years by self-report to maximise the likelihood of knowledge of the famous faces. 2.1.2. Stimuli Photographs of famous and unknown faces of a uniform quality were digitised to produce images of 16 greys, 150 · 200 pixels in size. The stimulus set comprised 126 pairs of one famous with one unfamiliar face matched on sex, race and approximate age. The faces in each pair showed a similar pose and facial expression. No data were collected to verify equivalence between the famous and unfamiliar faces on distinctiveness, attractiveness, or any other feature on which the stimuli might vary. The intention was to perform analyses within-items, with each famous person rated as positive by some participants and as negative by others, so that systematic variations between famous faces and their paired unfamiliar faces could not explain any observed experimental result. Names and examples of stimuli are given in Appendices A and B. The average luminance of the famous and unfamiliar faces was approximately 5.5 cd/m2, measured on a Minolta CS100 colour chronometer at a distance of approximately 1 m from the screen, the distance at which participants were seated. The mask was a collage of parts of unfamiliar faces, of the same size as the famous and unfamiliar faces. 2.1.3. Apparatus A personal computer running MEL2 software was used to display the faces at a 640 · 480 screen resolution. Response times and accuracy of response were measured and recorded by the computer. 2.1.4. Design The design comprised two independent within-item factors: valence of famous person (evaluated by experimental participants) and visual field containing the famous face (LVF vs. RVF). The dependent variables were speed and accuracy of response. The stimuli were presented in two blocks of 126 trials each, with each face pair appearing once in each block. Each famous face appeared in the LVF in one block and the RVF in the other block. In each block, there were equal numbers of famous faces in the LVF and the RVF. For each famous face, approximately equal numbers of participants saw it in the LVF and the RVF in each block. Thus, visual field and block were fully counterbalanced across participants. Within each block, the sequence of presentation was randomised by the computer for each participant. The duration of the forward and backward masks was 100 ms. A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 355 Stone and Valentine (2004) suggested that very few faces could be consciously recognised when presented for 17 ms in pairs with a mask similar to that used in the present series of experiments. 2.1.5. Procedure Participants carried out the tasks individually in a darkened, air-conditioned room. The sequence of events on each trial was as follows: central fixation cross for 500 ms, forward masks in LVF and RVF for 100 ms, famous and unfamiliar face for 17 ms, backward masks for 100 ms, then the question ‘‘left or right’’ displayed until a response was received. The response was made by pressing one of two keys: to the left of the keyboard to indicate the LVF, and to the right of the keyboard to indicate the RVF. Each trial was initiated by the response to the previous trial after an inter-trial interval of 1 s. The response time was calculated from the offset of the backward mask. The two faces were approximately 4.5 cm by 6 cm and were presented at a distance of 9 cm apart, subtending a visual angle of approximately 4 from fixation. The masks were presented in the same screen position as the faces. Participants were informed that two faces would be flashed up very briefly, one on either side of the screen, preceded and followed by a mask comprised of a collage of parts of unfamiliar faces. Each pair of faces would contain one famous person and one unfamiliar person, and participants were asked to select on which side of the screen the famous face had appeared. Participants were told they would find it very difficult to see the real faces and this should be no cause for concern, but they should attend carefully to the screen. They were asked to guess if unable to see anything of the stimulus faces. Participants were asked to look at the central fixation cross before each trial and to respond as quickly as possible. At the end of the task, participants were asked whether they had been able to recognise any of the faces displayed during the experiment, and were strongly encouraged to guess. Following this, participants were shown the famous faces used in the experiment, one at a time, and asked to identify each face, either by name or by a combination of biographical information that uniquely identified the individual person. Faces that were uniquely identified were shown again, one at a time, in a different random sequence, and evaluated on a 7-point scale from 3 (very negative) to +3 (very positive). Participants were asked to evaluate the valence of the person, not the face, considering any knowledge they had of the person, and to give their first impression. Finally, participants were debriefed and thanked for their participation. 2.2. Results and discussion Where a participant could not uniquely identify a famous face in the post-experimental evaluation, the responses for this combination of participant and item were excluded from the analysis (19% of trials). No masked 17 ms faces were recognised by any participant. Trials were excluded where the response time was longer than 5000 ms. Some items had missing data, having been evaluated as either positive or negative by all participants, and so 106 items were included in the analysis. The participantsÕ analysis was calculated over these items. The proportion of participants giving a negative evaluation to each of the 106 included items ranged from 0.06 to 0.94, mean 0.30, SD 0.2. For each participant, mean accuracy was calculated over the faces rated as positive and separately over the faces rated as negative (faces evaluated as zero were classified as positive). So two 356 A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 values were calculated for each participant: accuracy-positivep (mean = 50.5, standard error = .7) and accuracy-negativep (mean = 50.5, SE = 1.1). Similarly, for each item, mean accuracy was calculated over all participants rating the famous person as positive and separately over all participants rating the person as negative. So two values were calculated for each item: accuracypositivei (mean = 51.1, SE = 1.0) and accuracy-negativei (mean = 50.1, SE = 1.5). If a famous person was evaluated as positive and negative by unequal numbers of participants, different numbers of participants contributed to the calculation of accuracy-positivei and accuracy-negativei. Overall accuracy for all famous faces did not differ from chance, one-sample ti(105) = .88, ns, and tp(32) = 1.15, ns. Paired-samples t tests revealed no effect of valence, ti(105) = .60, ns, and tp(32) = .06, ns, in contrast to the effect reported by Stone and Valentine (2004). One obvious procedural difference is that participants in Experiment 1 of Stone and Valentine (2004) had performed a separate attention orientation task before the awareness check, whereas participants in the present study had not. The attention orientation task had presented the same masked 17 ms face pairs so that participants had gained experience in perceiving the stimuli. This suggests that a comparison of blocks 1 and 2 might be interesting. ANOVA was performed with two within-participants and within-items factors of block (1 vs. 2) and valence. The main effect of block was non-significant, Fi(1, 105) = 1.43, ns, and Fp(1, 32) = 4.11, p = .051, but the interaction with valence was significant, Fi(1, 105) = 4.78, p < .04, and Fp(1, 32) = 5.50, p < .03. Paired-samples t tests revealed that for negative faces, responses tended to be more accurate in block 1 (meani = 53.4, SE = 2.2) than in block 2 (meani = 46.8, SE = 2.3), ti(105) = 1.96, p < .06, and tp(32) = 2.62, p < .02. For positive faces, accuracy did not differ between block 1 (meani = 50.2, SE = 1.3) and block 2 (meani = 52.1, SE = 1.5), ti(105) = .93, ns, and tp(32) = .66, ns. Mean accuracy of responses to negative and positive faces was calculated for each chunk of five trials (to provide a more stable view of the data) over all participants and items. Accuracy was negatively correlated with chunk for negative faces, r(51) = .33, p < .02, but tended to be positively correlated with chunk for positive faces, r(51) = .24, p < .09, confirming that response accuracy for negative faces, but not positive faces, declined as the task progressed. There are three possible reasons why responses to negative faces were less accurate towards the end of the task than at the start and these will be considered in turn. 2.2.1. Practice The practice explanation proposes that participants became better able to extract information from masked faces as the task progressed, so that the valence associated with each face was more strongly activated, leading to the decline in accuracy of responses to negative faces. The practice explanation would be consistent with Experiment 1 of Stone and Valentine (2004) in which a separate attention orientation task before the awareness check had presented the same masked famous–unfamiliar face pairs. There is some empirical evidence to support the practice account. Dagenbach, Carr, and Wilhelmsen (1989, Experiment 1) measured participantsÕ exposure thresholds for chance performance in a presence/absence decision on briefly exposed masked words, before starting the experimental trials. Participants were tested again on the same presence/absence decision, at the same threshold exposure duration, after the experimental trials were complete. They found that for a sizeable group of participants (15 of 52, approximately 29%) accuracy in the presence/absence decision was above chance in the post-experimental re-test when it had been at chance in the pre-experimental threshold setting procedure, suggesting that some participantsÕ A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 357 ability to detect briefly exposed masked stimuli had improved as a result of practice. In the present experiment, increasing ability to derive valence from masked 17 ms famous faces could explain why the accuracy of responses to negative faces declined throughout the task. 2.2.2. Strategy Comments made by participants during debriefing suggested that they had used different strategies for attempting to perceive the masked faces at the start of the task and towards the end. Many reported that to begin with, they had tried very hard to see anything they could, but this was so difficult they had given up and simply guessed their responses. It makes sense that deliberate effort would be reduced towards the end of a task in which there was no feedback or evidence that effort improved accuracy. Relevant to this possibility, Snodgrass, Shevrin, and Kopka (1993) reported a study investigating the different strategies that participants might adopt for engaging with a task involving the perception of briefly exposed masked stimuli. Participants were asked to decide which of four words, two with pleasant meanings and two with unpleasant meanings, had been presented on each trial, using one of two strategies. The Pop strategy asked participants to ‘‘look where the word is presented and say whatever word pops into your mind.’’ The Look strategy asked participants to ‘‘look very hard where the word is presented. . . for anything you can see. . . use these cues when making your decision’’ (p. 196). Strategy interacted with word meaning: pleasant words tended to be identified more accurately than unpleasant words in the Pop strategy, whereas the converse was observed in the Look strategy. The Pop strategy gave results similar to the pattern observed in Stone and Valentine (2004) and the non-significant trend in the present Experiment 1, block 2, that is, less accurate responses to evil-negative faces than to good-positive faces. The Look strategy gave results similar to those obtained in Experiment 1, block 1, a tendency to more accurate responses to negative faces. This raises the possibility that something similar to the Look strategy was used in Experiment 1, block 1, and something similar to the Pop strategy in Experiment 1, block 2, which would be consistent with the comments made by participants during debriefing. There are other reports of strategy affecting responses in a task involving perception of briefly exposed masked stimuli (e.g., Dagenbach et al., 1989; Kahan, 2000). It seems that the failure to observe the predicted valence effect consistently in Experiment 1 may have been due to participantsÕ strategies. 2.2.3. Response latency In addition to practice and strategy, a third possibility is that response latency measured from stimulus face onset may be a factor, since participants tended to respond faster as the task progressed. A substantial literature on affective priming is relevant to this possibility and will be briefly described. Affective priming refers to the phenomenon in which responses to a valenced target stimulus are facilitated (inhibited) by the prior presentation of a prime stimulus of the same (opposing) valence to the target. Affective priming appears to be a reliable phenomenon, having been observed using words, objects and faces as primes; using words and pictures as targets; using evaluative decision, lexical decision, degraded word identification, and word pronunciation as the task; and using response time and accuracy as the dependent variable (see Fazio, 2001, for a review). Affective priming has been observed with masked primes of whose identity participants were not aware (e.g., 358 A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 Banse, 2001; Bargh, Litt, Pratto, & Spielman, 1989; see Fazio, 2001, for a review). There is substantial evidence that attitudes towards primes are activated automatically, without deliberate evaluation intent (e.g., Bargh, Chaiken, Raymond, & Hymes, 1996; Bargh et al., 1989; Hermans, De Houwer, & Eelen, 1994, 2001). Attitudes appear to be activated very rapidly, with prime to target SOA between 0 and 300 ms (e.g., Banse, 1999, 2001; Bargh et al., 1996; Bargh et al., 1989; De Houwer & Hermans, 1994; De Houwer, Hermans, & Eelen, 1998; Glaser & Banaji, 1999; Hermans et al., 1994, Hermans, De Houwer, & Eelen, 2001; Klauer, Rossnagel, & Musch, 1997; Murphy & Zajonc, 1993; Musch & Klauer, 2001). There is also evidence that the effect of the attitude evoked by the prime is corrected, where it is misleading with regard to the required response to the target. Several studies have reported significant effects at short SOA of up to 300 ms that are absent at longer SOA, which suggests that the correction process requires some duration (e.g., De Houwer et al., 1998; Glaser & Banaji, 1999; Hermans et al., 1994, 2001; Klauer et al., 1997; see Fazio, 2001, for a review). In the present Experiment 1, it is possible that lower accuracy of responses to negative than positive faces appeared in block 2 because responses were faster (block 1 mean response time = 851 ms, block 2 mean = 622 ms), and so were formulated during a period when the negative valence of the famous person influenced participants to select the unfamiliar face. With the longer response latency in Experiment 1, block 1, the influence of negative valence had been corrected before the responses were selected, and so response accuracy was not lower for negative faces than for positive faces. 2.2.4. Re-analysis of Experiment 1 The block design of Experiment 1 permits a comparison of shorter latency responses with longer latency responses in blocks 1 and 2. If only block has a significant effect, this would argue against the response latency account. If only response latency has a significant effect, this would make the practice account less likely. It is less easy to investigate the strategy account since participants were not asked on a trial-by-trial basis which strategy they had used. In addition, strategy may have been confounded with practice such that the Look strategy was preferred in block 1 and the Pop strategy in block 2, as suggested by participants during debriefing, or confounded with response latency, such that the Look strategy was used on longer latency trials and the Pop strategy on shorter latency trials. Responses were divided into short latency and long latency around the grand median response time of 604 ms. The mean accuracy of short and long latency responses was calculated for each participant, and separately for each item, by block and valence. The data were analysed in separate ANOVA for participants and items, with factors of valence, block (1 vs. 2) and response latency (short vs. long). Two participants and 34 items had missing data and were excluded from the ANOVA. See Table 1. The two-way interaction of valence with response latency approached significance, Fi(1, 71) = 3.25, p < .08, and Fp(1, 30) = 8.20, p < .01. See Fig. 1A. Paired-samples t tests investigated the interaction, including the maximum number of items and participants for whom data was complete (collapsing over block). For negative faces, shorter latency responses (meani = 46.1, SE = 2.7) were less accurate than longer latency responses (meani = 57.1, SE = 2.2), ti(100) = 3.25, p < .005, and tp(32) = 2.53, p < .02. For positive faces, there was no difference in accuracy between shorter latency responses (meani = 50.9, SE = 1.4) and longer latency responses A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 359 Table 1 Mean accuracy (and standard error) of fast and slow responses in the first and second block of Experiment 1, for positive and negative items. Positive Negative Fast Slow Total Fast Slow Total Block 1 Block 2 50.2 (2.9) 51.8 (2.3) 49.3 (2.1) 52.0 (3.0) 49.7 (2.6) 51.9 (2.0) 49.7 (4.0) 45.5 (2.9) 56.1 (3.3) 53.0 (3.8) 52.9 (2.5) 49.2 (2.5) Total 51.0 (1.8) 50.6 (1.8) 50.8 (1.3) 47.6 (2.6) 54.6 (2.5) 51.1 (1.9) Fig. 1. Accuracy of responses by valence and response latency (A) and by valence and block (B) in Experiment 1. Bars represent standard errors. (meani = 50.0, SE = 1.5), ti(100) = .46, ns, and tp(32) = 1.26, ns. Shorter latency responses tended to be less accurate to negative than positive faces, ti(100) = 1.69, ns, and tp(32) = 1.60, ns, whereas longer latency responses were more accurate to negative faces, ti(100) = 2.53, p < .02, and tp(32) = 2.87, p < .01. This is consistent with the response latency account, but could also be consistent with the strategy account if the Look strategy tended to be used on longer latency responses and the Pop strategy on shorter latency responses. The two-way interaction of half with valence was non-significant, Fi(1, 71) = 2.08, ns, and Fp(1, 30) = 2.78, ns. See Fig. 1B. Paired-samples t tests have already been reported. Responses tended to be more accurate to negative faces in block 1, and more accurate for positive faces in block 2. This is consistent with the practice account, but could also be consistent with the strategy account if Look strategy was used predominantly in block 1 while the Pop strategy was used predominantly in block 2. This inconclusive pattern of results does not enable a decision between the three accounts. Experiment 2 was designed to investigate the accounts directly by manipulating strategy, practice and response latency as independent factors. 2.2.5. Over-correction of valence? Another observation is worth mentioning: In Experiment 1, longer latency responses were more accurate to negative than positive faces. This could have arisen if the correction for the biasing 360 A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 effect of negative valence went too far and resulted in over-correction. Such over-correction has been observed in studies of affective priming. For example, Glaser and Banaji (1999), in a series of six experiments, reported reverse affective priming, that is, slower responses to targets of valence congruent with the prime than to targets of incongruent valence. They attributed this reverse priming to an over-correction for the biasing effect of the primes (e.g., Greenwald & Banaji, 1995; Stapel, Koomen, & Zeelenberg, 1998; Strack, Schwarz, Bless, Kubler, & Wanke, 1993). Based on the observation of reverse priming with prime-target SOA as short as 150 ms, they theorised that the over-correction was an automatic effect, not requiring deliberate intent by the participants. Klauer et al. (1997, Experiment 2) also reported reverse affective priming in an evaluative decision task, but only at a longer SOA of 1200 ms, and not at shorter SOA of 0 and 100 ms. It seems that the correction for the biasing influence of the primes takes some short time to develop. Glaser and BanajiÕs (1999) reverse priming was obtained at SOA of 150 and 300 ms, falling within the 100 ms SOA and the 1200 ms SOA of Klauer et al. (1997, Experiment 2) that observed congruent and reverse priming, respectively. The Klauer et al. (1997) study also manipulated the proportion of prime-target pairs of consistent evaluation (the consistency proportion). The weakest reverse priming was observed with consistency proportion of 0.75, in which condition the primes had predictive value for the evaluative decision to the targets. Correction for the biasing effects of the primes would have been less useful with such a high consistency proportion, so the weaker reverse priming in this condition fits the correction theory. Under the same logic, the strongest reverse priming should have been observed when the consistency proportion was 0.25, where the prime has least predictive value for the evaluative decision to the targets. In fact, the strongest reverse priming was observed with consistency proportion of 0.50, although the difference between this and the 0.25 consistency proportion was very small. Klauer et al. (1997) suggested that inconsistent pairs would have been more easily observed with 0.50 consistency proportion, since in this condition, consistent and inconsistent pairs would follow each other frequently during the experimental priming task. In Stone and Valentine (2004), the effect of negative valence of the famous person was to cause participants to select the paired unfamiliar face, contrary to the task instruction. There was a similar tendency, though non-significant, in the shorter latency trials of Experiment 1 (see Section 2.2.4). Thus, the effect of valence was misleading to the required response, which implies that the effect of valence should have been corrected, given sufficient time. More accurate responses to negative faces than to positive faces on longer latency trials in Experiment 1 is consistent with the proposal that over-correction, requiring some duration, may occur. This over-correction hypothesis was investigated by introducing a new affective priming task, described as Experiment 3. 3. Experiment 2 This experiment investigated the response latency, practice and strategy accounts that were offered as possible explanations for the failure to observe consistently lower response accuracy to negative faces than positive faces in Experiment 1. The post-experimental evaluation scale was altered to use the terms liked and disliked instead of positive and negative, on the expectation that A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 361 the new terms would focus participantsÕ attention on their personal attitude towards the target famous persons. Response latency was investigated by varying the duration of the backward mask between 500 and 100 ms. The duration of 500 ms should have ensured that responses were selected after the effect of valence had been corrected, and so disliked faces should be selected at least as often as liked faces, consistent with block 1 of Experiment 1. The duration of 100 ms, combined with instructions to respond without thinking too long, should have ensured that responses were decided before the correction took place, and so responses to disliked faces should be less accurate than responses to liked faces. This would be consistent with block 2 of Experiment 1. Half the participants in each backward mask condition were instructed to use the Pop strategy and the other half the Look strategy. If strategy is the major determinant of responses then the Pop strategy should result in higher accuracy for liked faces than disliked faces throughout, while the Look strategy should result in the opposite effect. If the practice account is correct, then disliked faces should be selected at least as often as liked faces in the first half, but less often than liked faces in the second half. 3.1. Method Only the changes from Experiment 1 will be described. 3.1.1. Participants Participants were 64 undergraduate students at Goldsmiths College, London. All had watched UK television for at least 5 years, by self-report, to maximise the likelihood of knowledge of the famous faces. Data were excluded from eight participants who lacked familiarity with the famous faces or failed to comply with experimental instructions. Data were excluded from 18 participants who selected more famous faces than expected by chance (binomial distribution, one-tailed, cutoff = 55%, a = .05) since for these participants the possibility of some awareness of facial familiarity cannot be ruled out. Of these 18 participants, 15 were in the 500 ms backward mask condition and 3 in the 100 ms backward mask condition. This is consistent with the response latency account: disliked faces would be selected less accurately with 100 ms than with 500 ms backward mask, so overall accuracy would be higher at 500 ms. The remaining 38 participants were 35 female and three male, aged between 18 and 36, mean 22.0, SD 5.6 years. There were 23 participants in the 500 ms backward mask condition and 15 in the 100 ms backward mask condition. More participants performed the 500 ms backward mask condition, in which no effect of valence was predicted, than the 100 ms mask condition, in which lower accuracy of responses to disliked faces than to liked faces was predicted, in order that a null effect in the 100 ms backward mask condition could not be attributed to lack of power in the statistical analysis. 3.1.2. Stimuli These comprised facial photographs of 80 celebrities, of whom 40 had been included in the items analysis of Experiment 1, and the remaining 40 were new stimuli. The new celebrities were selected on the expectation that they would be liked by some participants and disliked by others (this entailed avoiding popular comedians, criminals and military dictators). 362 A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 3.1.3. Design There were two within-participant and within-item factors, valence and half, and two betweenparticipant and within-item factors, strategy (Look vs. Pop) and mask duration (500 vs. 100 ms). The dependent variable was accuracy of response, and a correct response was scored by selecting the famous face in a pair. Each face pair was presented once in each of four blocks giving 320 trials altogether. The position of the famous face was counterbalanced as follows. The faces were randomly divided into two sets. Half the participants in each strategy and backward mask condition saw one set in the LVF in blocks 1 and 3 and the RVF in blocks 2 and 4, and the other set in the RVF in blocks 1 and 3 and the LVF in blocks 2 and 4. For the other half of the participants in each condition, these positions were reversed. Thus, face position was fully counterbalanced over the factors of block, strategy and mask duration. This design ensured equal numbers of famous faces in LVF and RVF in each block and minimised the likelihood that the same face would be presented twice in succession. 3.1.4. Procedure The major change from previous experiments was the addition of the specific instructions. In the Look condition, participants were asked to ‘‘try hard to see anything you can of the faces that flash up. Most people find that they can see nothing at first, but as the experiment progresses, they are able to see bits of the faces. Some people can see the outline of a face, or the hair, or maybe the eyes or the mouth. Please try hard to see anything you can, and use what you see to make a guess about which of the 2 faces is famous.’’ In the Pop condition, participants were asked to ‘‘relax, and not make any effort to see the faces. Just look at the cross in the centre, relax, let the faces flash up, and let the answer pop into your head.’’ In both the Look and Pop conditions, participants were told, ‘‘It is very important that you follow this strategy because the purpose of the experiment is to contrast the effects of different strategies.’’ In the 100 ms backward mask condition, participants were asked to respond on each trial without thinking too long, and this instruction was emphasised after the practice trials had been completed. 3.2. Results and discussion Where a participant was unable to uniquely identify a famous face in the post-experimental evaluation, the trials for this combination of participant and item were excluded from the analysis (16% of trials). If a face was correctly identified during the experimental trials then all four trials for this face were excluded (1.3% of trials). Trials were excluded where the response time was longer than 5000 ms or shorter than 200 ms from face offset (<1% of trials). For each participant, liked and disliked faces were selected according to the participantÕs own ratings. For each item, participants who liked and disliked the item were selected according to the same ratings. Faces evaluated as zero were classified as liked in order to distinguish between disliked persons and the rest. Many items had missing data, having been evaluated as liked or disliked by all participants in a condition (mask duration · strategy), so the item analysis consisted of 29 items. The participant analysis was calculated over these items. The proportion of participants evaluating each of the included famous persons as disliked ranged from 0.18 to 0.78, mean = .44, so the included items were roughly balanced between liked and disliked. A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 363 ANOVA was performed by participants and by items, with valence and half as within-participant and within-item factors, and strategy and mask duration as between-participant and withinitem factors. The main effect of mask duration was non-significant, Fi(1, 28) = 2.13, ns, and Fp(1, 34) = 3.27, p < .08, but the interaction of mask duration with valence was significant by items, Fi(1, 28) = 4.85, p < .04, and approached significance by participants, Fp(1, 34) = 3.00, p < .10. See Fig. 2 and Table 2. The interaction was investigated using t tests with a set at 0.025 (one-tailed) for each individual comparison. With mask duration of 500 ms, responses to disliked faces and liked faces were equally accurate, ti(28) = .97, ns, and tp(22) = .64, ns. With mask duration of 100 ms, responses to disliked faces were less accurate than responses to liked faces, ti(28) = 2.05, p = .025, and tp(14) = 2.28, p < .02. Responses to disliked faces were less accurate at 100 ms than at 500 ms mask duration, significant by items and marginal by participants, ti(28) = 2.32, p < .02, and tp(36) = 1.90, p = .033. Responses to liked faces were equally accurate at 500 and at 100 ms, ti(28) = .61, ns, and tp(36) = .16, ns. The two-way interaction of half with mask duration approached significance by participants and items, Fi(1, 28) = 3.52, p < .08, and Fp(1, 34) = 3.43, p < .08. Inspection of Table 2 reveals a tendency for accuracy to increase from the first half to the second half at 500 ms mask duration, but to decrease at 100 ms mask duration. The reason for this interaction is not obvious and so it will not be interpreted. The two-way interaction of mask duration with strategy was non-significant by items, Fi(1, 28) = 2.37, ns, and only marginally significant by participants, Fp(1, 34) = 3.83, p < .06, and so will not be interpreted. No other effects reached significance, all Fi and Fp < 1.9, p > .18. In particular, neither strategy nor task half interacted with valence, all Fi and Fp < 1. These results support the prediction based on the response latency account. With 100 ms backward mask it was predicted that responses would be decided before the effect of negative valence was corrected, so that responses to disliked faces would be less accurate than responses to liked faces, and this effect was observed. In contrast, with 500 ms backward mask, it was predicted that the effect of negative valence would have been corrected before responses were decided, and in- Fig. 2. Accuracy in the 500 ms (long response latency) and 100 ms (short response latency) backward mask conditions of Experiment 2. Bars represent standard errors. 364 Disliked Liked 500 ms 100 ms 500 ms 100 ms Pop Look Total Pop Look Total Pop Look Total Pop Look Total n = 13 n = 10 n = 23 n=8 n=7 n = 15 n = 13 n = 10 n = 23 n=8 n=7 n = 15 51.5 (3.7) 51.0 (2.3) 49.2 (4.8) 47.6 (5.8) 48.4 (4.1) 47.7 (3.3) 52.7 (3.9) 50.2 (2.9) 56.8 (3.9) 54.2 (4.4) 55.5 (3.1) 59.3 (4.3) 54.5 (2.7) 41.7 (4.9) 42.3 (4.3) 42.0 (3.1) 49.5 (3.4) 52.6 (4.0) 51.0 (3.0) 50.1 (3.3) 47.5 (4.8) 48.8 (3.0) 55.4 (2.8) 52.7 (1.8) 45.4 (4.0) 45.0 (4.0) 45.2 (3.0) 48.6 (2.8) 52.6 (2.1) 50.6 (2.1) 53.4 (2.6) 50.8 (3.0) 52.1 (2.1) Half 1 50.4 (3.4) Half 2 49.6 (3.2) Total 50.0 (2.7) A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 Table 2 Mean accuracy (and standard error) of responses to disliked and liked items at 100 and 500 ms backward mask duration, under the Look and Pop strategies, in the first and second half of Experiment 2 A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 365 deed responses to disliked and liked faces were equally accurate. The ANOVA was significant in the items analysis, which is the more important analysis since it rules out potential confounds arising from physical differences between famous and unfamiliar faces, or between liked and disliked famous faces. The ANOVA was marginally significant in the participants analysis, offering weak support for the main items analysis. The factors of strategy and task half showed no reliable effects, in particular failing to interact with valence, providing no support for either the strategy or the practice explanations. Accuracy of responses to disliked faces was not lower under the Pop strategy than under the Look strategy. Accuracy of responses to disliked faces did not decline with practice, revealed by the null effect of task half. A potential confound must be addressed. Response latency was manipulated by varying the duration of the backward mask, so it is possible that the mask duration rather than the response latency was responsible for the pattern of results. Perhaps the 100 ms backward mask was less effective than the 500 ms backward mask, and a less effective mask may have allowed valence to become activated, so that valence influenced participant responses with 100 ms backward mask but not with 500 ms backward mask. This must be regarded as unlikely for three distinct reasons. First, the backward mask duration was 100 ms throughout Experiment 1 and yet there was an interaction of response latency with valence, suggesting that response latency rather than backward mask duration is the key factor. Second, Esteves and Ohman (1993) reported that the duration of the backward mask (between 30 and 120 ms) had no effect on the likelihood of conscious perception of a masked facial emotional expression. Third, in the present experiment, participants were more likely to be aware of the identity of a masked 17 ms face at 500 ms mask duration (item mean percentage conscious recognition = 2.7%) than with 100 ms mask duration (item mean = .2%), Wilcoxon signed-ranks zi(29) = 2.69, p < .01, and Mann–Whitney zp(37) = 2.42, p < .05, which contradicts the proposal that masking is more effective at 500 ms than 100 ms mask duration. Experiment 2 supported the response latency account but not the practice account or the strategy account. It therefore appears that neither practice nor strategy significantly moderated the effect of lower response accuracy to disliked than liked faces, and the only relevant factor was response latency from face onset. This is consistent with the affective priming literature in which effects of valence are swiftly activated and swiftly corrected. It is supposed that the effect of negative valence was corrected because it led participants to make incorrect responses by selecting the paired unfamiliar face instead of the famous face. 4. Experiment 3 A new affective priming task was introduced to investigate the possibility that the higher response accuracy to negative faces than positive faces that was observed on longer latency trials in Experiment 1 was due to over-correction for the negative valence of the famous face. The affective priming task was given to a subset of the participants in Experiment 2 and was always performed after the original task and before participant evaluations of the famous persons. Prime stimuli in Experiment 3 were masked 17 ms faces of famous persons. Targets were clearly visible words of pleasant or unpleasant meaning, and the required response was a 366 A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 pleasant–unpleasant decision to the target word. Normal affective priming suggests facilitation of responses to target words of the same valence as the prime face, that is, liked faces should result in faster responses to pleasant words and disliked faces in faster responses to unpleasant words. However, if the biasing effect of the prime face were corrected, there might be no interaction of prime and target valence. Further, if the biasing effect of the prime face were overcorrected, then the result would be faster responses to targets of valence incongruent with the prime face. It is necessary to examine closely the conditions that might be required for correction to occur. The first condition is that the valence evoked by the prime face should be recognised as misleading with respect to the required response to the target. The required target response was an evaluative decision (pleasant or unpleasant), and the prime and target valence were congruent and incongruent on approximately equal numbers of trials, so the prime valence was in fact misleading with respect to the required response to the target. The question is whether this would be recognised. In this respect, a complication is introduced by the different strategies used in Experiment 2. Given that Experiment 2 always preceded Experiment 3 in the same experimental session, and evidence that the effect of strategy can carry forward into a subsequent task (e.g., Carr & Dagenbach, 1990; Dagenbach et al., 1989; Kahan, 2000), it is necessary to examine whether the Look and Pop strategies predict equivalent correction for the biasing effect of the prime face. The Look strategy in the original task instructed participants to look hard at the stimulus faces and use any partial perception as the basis for their responses. This instruction seems likely to have focused deliberate attention on the masked faces and it may be supposed that this focus of attention carried forward into Experiment 3 (e.g., Carr & Dagenbach, 1990; Dagenbach et al., 1989; Kahan, 2000). So in Experiment 3, the combination of deliberate attention to the masked faces, the instruction to use them as a basis for the response, and equal numbers of valence-congruent and valence-incongruent trials, should have led to the recognition that prime valence was misleading for the required target response. This should have prompted an automatic correction process. In contrast, the Pop strategy asked participants to relax, not make any effort to see the faces, and let the answer pop into their heads. In Experiment 3, this instruction seems likely to have removed the deliberate focus of attention from using the masked faces as a basis for the response, in which case, the irrelevance of the masked faces to the instructed task would have been less obvious. Hence, correction for the effect of prime valence was less likely to occur. This line of reasoning predicts that reverse affective priming (faster responses to target words of valence incongruent with the prime face) should be stronger under the Look strategy than under the Pop strategy. The second condition for correction of the biasing effect of prime valence is sufficient time for the correction to occur. The theorised correction occurred in the original task of Experiment 1 on long latency trials but not on short latency trials. However, faces were presented in simultaneous pairs in Experiment 1, and were presented singly and centrally in Experiment 3. It is quite possible that correction could occur more quickly under the easier conditions of Experiment 3. Given that Glaser and Banaji (1999) reported reverse affective priming with SOA of only 150 ms, it is possible that correction could occur even at the shorter prime-target SOA of 100 ms in the present Experiment 3. Therefore, no prediction was made for the effect of backward mask duration. A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 367 4.1. Method 4.1.1. Participants Participants were 44 undergraduate students at Goldsmiths College, London, comprising all except the first 20 from Experiment 2. Data from 11 participants were not analysed for reasons detailed below in the Results section. The remaining 33 participants were 26 female and seven male, aged between 18 and 41, mean 22.9, SD 6.7 years. There were 17 participants in the 100 ms backward mask condition and 16 participants in the 500 ms backward mask condition; 18 participants in the Look strategy condition and 15 participants in the Pop condition. 4.1.2. Stimuli Ten of the faces used in Experiment 2 were selected as primes on the basis that they were correctly identified by over 90% of the first group of participants in Experiment 2, with roughly equal numbers of participants evaluating each face as liked and disliked. Names are presented in Appendix A. Another five unfamiliar faces were chosen to act as primes on filler trials. The mask was the same as used in Experiment 2. The targets were four words: rose, pleasure, pain, and fighting. The two positive and two negative targets were balanced on word length and word frequency. The targets were those used by Snodgrass et al. (1993). 4.1.3. Design There were three independent factors. Backward mask duration (100 vs. 500 ms) and strategy (Look vs. Pop) were between-participants and within-items, and congruence (face-valence congruent or incongruent with word-valence) was within-participants and within-items. Backward mask duration for each participant was the same as Experiment 2. Strategy was implemented by the instructions given to participants in Experiment 2 that always preceded Experiment 3. Each famous face was rated as liked or disliked by each participant after the affective priming task, by the procedure described in Experiment 2. Each prime face was presented four times, once for each target word, making a total of 60 trials. Trials were presented in a different random sequence for each participant. The required response to the target was a two-alternative forced-choice of pleasant or unpleasant. The keys assigned to the two response options were counterbalanced across participants. 4.1.4. Procedure Participants carried out the affective priming task following the original awareness check task, and before giving their evaluations of the famous persons. Four practice trials preceded the 60 experimental trials. The participant initiated the sequence of trials by pressing a key. The presentation of stimuli on each trial was as follows: forward mask for 100 or 500 ms, prime face for 17 ms, backward mask for 100 or 500 ms, and target word presented until the participant responded. Each subsequent trial was initiated by the response to the previous trial after an interval of 1 s. Participants were informed that a series of words, some pleasant and some unpleasant, would be displayed one at a time on the screen. They were asked to respond by pressing one of two keys depending on whether the target word was pleasant or unpleasant and to respond as quickly as 368 A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 possible. The assignment of keys to responses was counterbalanced across participants. They were asked to report immediately any faces that were consciously recognised, even if recognition was very uncertain and even if the name could not be recalled. After completion of the experimental trials, participants gave their evaluations of the famous persons, as described in Experiment 2. Finally, participants were debriefed and thanked for their participation. 4.2. Results and discussion Trials were excluded if the response to the target word was incorrect, if the response was slower than 1500 ms, if the participant was aware of the identity of the prime face during the experimental trials, or if the participant could not identify the prime face in the post-experimental rating. Eleven participants had only a single face rated liked or disliked of whose identity they were unaware during the experimental trials, and were therefore excluded from the analysis. Other data were considered valid and were included in the analysis. ANOVA was performed by participants and by items with three factors of backward mask (500 vs. 100 ms), strategy (Look vs. Pop) and congruence (face-valence congruent or incongruent with word-valence). The dependent variable was mean response time for correct responses to target words. The approach of collapsing the two factors of word-valence and face-valence into a single factor of congruence is a common approach (e.g., De Houwer & Hermans, 1994; De Houwer et al., 1998; Glaser & Banaji, 1999; Hermans et al., 2001; Klauer et al., 1997; Musch & Klauer, 2001). This approach serves to enhance the clarity of the results and discussion by focusing on the experimental hypothesis. Exposition of any main effects of face-valence and word-valence that might be present would have detracted from the major hypothesis of the experiment. The main effect of backward mask duration was significant, Fi(1, 9) = 124, p < .001, and Fp(1, 29) = 24.9, p < .001, showing slower responses in the 500 ms backward mask condition than in the 100 ms condition. The two-way interaction of strategy with congruence was significant, Fi(1, 9) = 6.91, p < .03, and Fp(1, 29) = 7.03, p < .02. All other effects were non-significant, all Fi and Fp < 1. The interaction was investigated with paired-samples t tests for each strategy. Under the Pop strategy, there was a tendency to faster responses in the congruent condition, ti(9) = 1.77, ns, and tp(14) = 1.80, ns, that is, normal affective priming. Under the Look strategy, there was a tendency to faster responses in the incongruent condition, ti(9) = 2.09, p < .07, and tp(17) = 2.17, p < .05, that is, reverse affective priming. See Table 3 and Fig. 3. The pattern of results supports the prediction that reverse affective priming would be stronger under the Look strategy than under the Pop strategy. This is consistent with the concept that the Table 3 Mean response time (and standard error) for correct responses to valence-congruent and valence-incongruent target words, under the Pop and Look strategies, with 100 and 500 ms backward mask, in Experiment 3 Look Pop Congruent Incongruent Congruent Incongruent 100 ms mask 500 ms mask 639 (19) 813 (32) 602 (15) 785 (23) 609 (14) 803 (21) 643 (9) 825 (17) Mean 726 (21) 693 (14) 706 (13) 734 (11) A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 369 Fig. 3. Mean response time in the valence-congruent and valence-incongruent conditions of Experiment 3 (affective priming), under the Look and Pop strategies. Bars represent standard errors. Look strategy would focus attention on the misleading nature of the primeÕs valence, and so engender a process of correction for the biasing effect of the prime, while the Pop strategy would focus less attention on the prime, and so engender a weaker (or no) correction process. The observation of this result in the item analysis rules out confounds based on variation in some visual characteristics of the faces. Differential priming of valenced target words according to whether the same prime face was liked or disliked (by different participants) confirms that the effect was due to recognition of the unique face identity. The interaction of the direction of affective priming (normal or reverse) with strategy was unaffected by the duration of the backward mask. It appears that the theorised correction for the biasing effect of prime valence that occurred under the Look strategy was as effective at the shorter backward mask duration (SOA of 117 ms) as at the longer backward mask duration (SOA of 517 ms). This might seem inconsistent with the results of Experiment 2, that reported an apparent correction for the biasing effect of negative valence at 500 ms backward mask duration but not at 100 ms mask duration, but it should be noted that Experiment 2 presented faces in simultaneous pairs while Experiment 3 presented faces singly and centrally. It is reasonable to suppose that processing of two faces simultaneously would be slowed in comparison with processing of a single face, so that the theorised over-correction could have occurred with the shorter backward mask duration in Experiment 3 but not in Experiment 2. Another issue is that the over-correction for the biasing effect of prime valence must have occurred before the response on each trial was decided, but it is not clear at exactly what point this happened. In Experiment 3, the response to the target word was formulated some time after the word was presented, but the lag is unknown. The lag would have allowed some extra time for the biasing effect of prime valence to be corrected. 5. General discussion Experiment 1 was an attempt to replicate the effect observed by Stone and Valentine (2004) that detection of familiarity from masked 17 ms faces was more accurate when the 370 A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 famous persons were subsequently evaluated positively than when they were evaluated negatively. A different pattern of results was observed, with the accuracy of responses to negative faces declining throughout the task. Feedback from participants during debriefing and theoretical considerations suggested three potential explanations: practice, strategy and response latency. The results of Experiment 2 supported only the response latency account. With short response latency (100 ms backward mask duration), responses were less accurate to disliked faces than to liked faces, while at long response latency (500 ms backward mask duration), responses were equally accurate to liked and disliked faces. Responses to disliked faces were less accurate at short than at long response latency. There was no effect of practice or strategy. The lower accuracy of responses to negative faces at short response latency was attributed to the biasing effect of negative valence that caused participants to select the paired unfamiliar face, contrary to task instruction. With longer latency, the biasing effect of negative valence was corrected, so that responses to the faces of negative persons were as accurate as responses to positive faces. A speculative explanation for the effect of response latency can be offered, as follows. It is necessary to consider first how facial familiarity is detected. Farah, OÕReilly, and Vecera (1993) suggested that facial familiarity might be detected when any semantic information associated with the face becomes activated. There is an alternative view: the Burton, Bruce, and Johnston (1990) model of face recognition states that familiarity is detected when activation at the Person Identity Node reaches a threshold. This node is an a-model representation of the concept of the person, and does not itself store any semantic information, but is connected to other nodes representing semantics. The concept that familiarity is detected when the Person Identity Node reaches a threshold was based on the observation of a ‘‘familiar-only’’ response that occurs when a face can be deemed familiar but the participant is not aware of any semantic information. However, there is the possibility that several items of semantic information might be activated, each one below threshold for awareness, but together adding up to sufficient activation to declare the face familiar. So, even when the familiar-only response occurred, familiarity may still have been detected from the activation of semantic information. Also, the participant might have access to affective information concerning their attitude towards the person, and this might be sufficient to declare the face familiar. On balance, it seems likely that detection of facial familiarity depends on the (maybe unaware) retrieval of semantic or affective information associated with the person. Given that facial familiarity may be detected from the existence of any semantic or affective information associated with the person, it is relevant to consider the relative timescale of retrieval of these two types of information. ZajoncÕs affective primacy hypothesis (Murphy & Zajonc, 1993; Zajonc, 1980; see Zajonc, 2001, for a review) states that affective information becomes available before semantics. Some support for this hypothesis stems from Bargh et al. (1989) who reported that the valence of masked stimuli could be detected with a shorter exposure duration than semantic meaning, suggesting that valence may be more strongly connected and more readily activated than semantic information. If this is the case, then valence may well become available before semantics. It should also be noted that affective priming has been observed with zero or negative stimulus onset asynchrony (SOA) between prime and target (e.g., De Houwer & Hermans, 1994; Hermans et al., 2001; Klauer et al., 1997; Musch & A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 371 Klauer, 2001) whereas semantic priming requires a positive SOA. All of these lines of evidence suggest that affective information about a stimulus is activated more swiftly than semantic information. Applied to the present series of experiments, the affective primacy hypothesis has the implication that valence exerted a stronger influence on shorter latency responses than on longer latency responses. On shorter latency responses, negative valence associated with the famous person led participants to select the unfamiliar face, and semantics were not sufficiently activated to contradict this decision. On the longer latency trials, semantic information pointed at the famous face, creating a conflict between the valence information and the semantic information. Assuming that a decision mechanism placed more weight on the semantic information, this could explain why responses to negative-disliked faces were at least as accurate as responses to positive-liked faces at longer response latency. The assumption that a decision mechanism placed more weight on semantics than affect seems reasonable—given the perception that a face generates a negative valence, but at the same time the face is known to have, e.g., a certain occupation, it is obvious that the face belongs to a famous person. It remains to be explained why negative valence associated with a famous person should lead to the selection of the paired unfamiliar face. One possible mechanism is based on the mis-application of the mere exposure effect (Zajonc, 1980; see Zajonc, 2001, for a review), which states that a familiar stimulus is preferred over an unfamiliar stimulus, all other things being equal. Participants may have conceived a preference for one face in each pair, assumed that preference indicated familiarity, and so selected the preferred face as likely to be the famous face. Alternatively, participants might have simply selected the face they preferred without any assumption that preference denotes familiarity. Either way, when the famous person was negative-disliked, the unfamiliar face was preferred and so was selected. Preference could have been detected in many ways. In an fMRI study conducted by Pizzagalli, Koenig, Regard, and Lehmann (1998), participants were asked to observe previously unfamiliar faces for 450 ms each, and subsequently rate the faces as liked or disliked. Liked and disliked faces activated different neuronal populations, with the centre of activation for disliked faces being more to the right than the centre of activation for liked faces. Pizzagalli et al. (2002) reported that the N170 ERP component that has been frequently associated with face processing was larger in amplitude for liked than disliked faces. Another possibility is that liked and disliked faces had different effects on the amygdala, e.g., Whalen et al. (1998) observed that fearful facial expressions perceived without awareness of the expression resulted in higher amygdala activation than happy facial expressions. These are just three possibilities: any mechanism for detecting preference between two faces could have generated the observed effect in the present experiments. The observation of more accurate responses to negative than positive faces in the longer latency trials of Experiment 1 suggests that the biasing effect of negative valence might not merely be corrected when semantics were also available, but actually over-corrected. This could occur if the familiarity decision mechanism, assumed to place more weight on semantics than valence, had corrected the valence signal to make it consistent with the semantic signal. If this correction was strong then a negative face might generate a stronger preference than a positive face on the longer latency trials. Combined with the semantic information pointing at the famous face, 372 A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 an over-corrected preference signal could have generated the observed effect of more accurate familiarity detection from negative than positive faces. Experiment 3 examined the possibility of over-correction for the biasing effect of valence in an affective priming paradigm, using masked 17 ms famous faces as primes and clearly visible pleasant or unpleasant words as targets. Results supported the concept of over-correction when participantsÕ strategy in attempting to perceive the masked faces caused recognition of the misleading effect of prime valence for the required response to the target. Several directions for future research are suggested by the experiments reported here. The speed with which the valence associated with a famous person becomes active, and may be subsequently corrected, could be estimated by including various backward mask durations. The conditions of strategic set under which correction is applied could also be systematically examined and may shed light on the mechanism of correction. It would be of interest to invite prosopagnosic participants to perform modified versions of these experiments: given ample evidence of covert face recognition in the form of activation of semantic information, it seems likely that prosopagnosics would be able to activate emotional information associated with famous faces of whose identity there was no awareness. Other factors may have influenced the accuracy of participant choices, for example, face familiarity, attractiveness and distinctiveness, and these could be examined by collecting ratings from experimental participants. The experiments reported here could all be repeated with other masked stimuli; one obvious choice would be names rather than faces of famous and unfamiliar persons. Several conclusions may be drawn. The most basic is that famous faces presented very briefly and masked can be identified without awareness of facial familiarity. A more interesting conclusion is that negative participant attitude towards the famous person has different effects on a familiarity decision at different response latencies: familiarity is less likely to be detected a shorter latency than at longer latency. Consistent with the affective priming literature, it appears that affective responses are influenced by temporal factors. Appendix A. Stimuli A.1. Experiment 1 Film/TV actors: Woody Allen, Gillian Anderson, Jennifer Aniston, Sandra Bullock, Michael Caine, Jim Carrey, Sean Connery, Kevin Costner, Tom Cruise, Jamie Lee Curtis, Robert deNiro, Leonardo Dicaprio, Michael Douglas, David Duchovny, Harrison Ford, Sarah Michelle Gellar, Richard Gere, Mel Gibson, Hugh Grant, Anthony Hopkins, Liz Hurley, Bruce Lee, Marilyn Monroe, Demi Moore, Roger Moore, Jack Nicholson, Brad Pitt, Oliver Reed, Burt Reynolds, Julia Roberts, Arnold Schwarzenegger, William Shatner, Sylvester Stallone, Sharon Stone, Liz Taylor, John Travolta, Bruce Willis, Catherine Zeta-Jones. Pop stars: Damon Albarn, Victoria Beckham, Liam Gallagher, Geri Halliwell, Janet Jackson, Michael Jackson, Mick Jagger, Elton John, John Lennon, Jennifer Lopez, Madonna, Freddie Mercury, George Michael, Kylie Minogue, Elvis Presley, Cliff Richard, Robbie Williams. A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 373 Comedians: Rowan Atkinson, Craig Charles, Martin Clunes, Harry Enfield, Stephen Fry, Joanna Lumley, Nicholas Lyndhurst, Neil Morrissey. Royal family: Sarah Ferguson, Prince Andrew, Prince Charles. Politicians: Gerry Adams, Jeffrey Archer, Cherie Blair, Tony Blair, George W. Bush, Bill Clinton, William Hague, Adolf Hitler, Saddam Hussein, J.F. Kennedy, Ken Livingston, John Major, John Prescott, Ronald Reagan, Margaret Thatcher. TV presenters: Jeremy Beadle, Cilla Black, Paul Daniels, Charlie Dimmock, Chris Evans, Judy Finnigan, Bruce Forsyth, Rolf Harris, Clive James, Des Lynam, Richard Madeley, Michael Parkinson, Anne Robinson, Chris Tarrant, Alan Titchmarsh, Carol Vordeman, Terry Wogan. Sports: Boris Becker, David Beckham, Paul Gascoigne, O.J. Simpson, Mike Tyson. Other: Richard Branson (entrepreneur), Stephen Hawking (scientist), Osama Bin Laden. Faces excluded from analysis: Charles Bronson, Kathy Burke, John Cleese, Glenn Close, Jodie Foster, Dawn French, Tom Hanks, Woody Harrelson, Lenny Henry, Myra Hindley, David Jason, Hugh Laurie, Paul Mccartney, Eddie Murphy, Liam Neeson, Michael Palin, Michelle Pfeiffer, Jennifer Saunders, Robin Williams, Princess Diana. A.2. Experiment 2 Film/TV actors: Russell Crowe, Tom Cruise. Pop stars: Liam Gallagher, Geri Halliwell, Whitney Houston, Michael Jackson, Victoria Beckham, Cliff Richard, Britney Spears, Robbie Williams. Royal family: Sarah Ferguson, Camilla Parker-Bowles, Prince Charles. Politicians: Tony Blair, G.W. Bush, William Hague, John Major, Colin Powell, John Prescott, Margaret Thatcher, Anne Widdecombe. TV presenters: Michael Barrymore, Jeremy Beadle, Bill Clinton, Paul Daniels, Chris Evans, Jeremy Paxman, Anne Robinson. Other: Naomi Campbell (model). Faces excluded from analysis: Woody Allen, Yasser Arafat, David Beckham, Cilla Black, Cherie Blair, Helena Bonham-Carter, Gordon Brown, Jim Carrey, Cher, Puff Daddy, Danny DeVito, Leonardo DiCaprio, David Duchovny, Clint Eastwood, Eminem, Bruce Forsyth, Sarah-Michelle Gellar, Hugh Grant, Rutger Hauer, Liz Hurley, Janet Jackson, Samuel L. Jackson, Mick Jagger, Elton John, Ross Kemp, Annie Lennox, Ken Livingston, Jennifer Lopez, Richard Madeley, Rik Mayall, Ally McBeal, Ewan McGregor, George Michael, Liza Minelli, Mike Myers, Leonard Nimoy, Gwynneth Paltrow, Luciano Pavarotti, Alan Rickman, Jonathan Ross, Arnold Schwarzenegger, O.J. Simpson, Iain Duncan Smith, Sylvester Stallone, Patrick Stewart, Barbra Streisand, Mike Tyson, Oprah Winfrey, Terry Wogan, Princess Anne, The Queen. A.3. Experiment 3 All faces included: Michael Barrymore, Tony Blair, Naomi Campbell, Chris Evans, Liam Gallagher, Geri Halliwell, John Major, Cliff Richard, Anne Robinson, Margaret Thatcher. 374 A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 Appendix B. Examples of stimuli and the mask References Banse, R. (1999). Automatic evaluation of self and significant others: Affective priming in close relationships. Journal of Social and Personal Relationships, 3, 803–821. A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 375 Banse, R. (2001). Affective priming with liked and disliked persons: Prime visibility determines congruency and incongruency effect. Cognition and Emotion, 15, 201–220. Bargh, J. A., Chaiken, S., Raymond, P., & Hymes, C. (1996). The automatic evaluation effect: Unconditional automatic attitude activation with a pronunciation task. Journal of Experimental Social Psychology, 32, 104–128. Bargh, J. A., Litt, J., Pratto, F., & Spielman, L. A. (1989). On the preconscious evaluation of social stimuli. In A. F. Bennett & K. M. McConkey (Eds.), Cognition in individual and social contexts. North-Holland: Elsevier. Burton, A. M., Bruce, V., & Johnston, R. A. (1990). Understanding face recognition with an interactive activation model. British Journal of Psychology, 81, 361–380. Carr, T. H., & Dagenbach, D. (1990). Semantic priming and repetition priming from masked words: Evidence for a centre-surround attentional mechanism in perceptual recognition. Journal of Experimental Psychology: Learning, Memory and Cognition, 16, 341–350. Dagenbach, D., Carr, T. H., & Wilhelmsen, A. L. (1989). Task-induced strategies and near-threshold priming: Conscious influences on unconscious perception. Journal of Memory and Language, 28, 412–443. De Houwer, J., & Hermans, D. (1994). Differences in the affective processing of words and pictures. Cognition and Emotion, 8, 1–20. De Houwer, J., Hermans, D., & Eelen, P. (1998). Affective and identity priming with episodically associated stimuli. Cognition and Emotion, 12, 145–169. Dimberg, U., & Ohman, A. (1996). Behold the wrath: Psychophysiological responses to facial stimuli. Motivation and Emotion, 20, 149–182. Dimberg, U., Thunberg, M., & Elmehed, K. (2000). Unconscious facial reactions to emotional facial expressions. Psychological Science, 11, 86–89. Esteves, F., & Ohman, A. (1993). Masking the face: Recognition of emotional facial expressions as a function of the parameters of backward masking. Scandinavian Journal of Psychology, 34, 1–18. Farah, M. J., OÕReilly, R. C., & Vecera, S. P. (1993). Dissociated overt and covert recognition as an emergent property of a lesioned neural-network. Psychological Review, 100, 571–588. Fazio, R. H. (2001). On the automatic activation of associated evaluations: An overview. Cognition and Emotion, 15, 115–141. Glaser, J., & Banaji, M. R. (1999). When fair is foul and foul is fair: Reverse priming in automatic evaluation. Journal of Personality and Social Psychology, 77, 669–687. Greenwald, A. G., & Banaji, M. R. (1995). Implicit social cognition: Attitudes, self-esteem, and stereotypes. Psychological Review, 102, 4–27. Hermans, D., de Houwer, J., & Eelen, P. (1994). The affective priming effect: Automatic activation of evaluative information in memory. Cognition and Emotion, 8, 515–533. Hermans, D., De Houwer, J., & Eelen, P. (2001). A time course analysis of the affective priming effect. Cognition and Emotion, 15, 143–165. Johnsen, B. H., & Hugdahl, K. (1991). Hemispheric asymmetry in conditioning to facial emotional expressions. Psychophysiology, 28, 154–162. Johnsen, B. H., & Hugdahl, K. (1993). Right hemisphere representation of autonomic conditioning to facial emotional expressions. Psychophysiology, 30, 274–278. Kahan, T. A. (2000). Negative priming from masked words: Retrospective prime clarification of centre-surround inhibition?. Journal of Experimental Psychology: Learning, Memory and Cognition, 26, 1392–1410. Klauer, K. C., Rossnagel, C., & Musch, J. (1997). List-context effects in evaluative priming. Journal of Experimental Psychology: Learning, Memory and Cognition, 23, 246–255. Mogg, K., & Bradley, B. P. (1999). Orienting of attention to threatening facial expressions presented under conditions of restricted awareness. Cognition and Emotion, 13, 713–740. Murphy, S. T., & Zajonc, R. B. (1993). Affect, cognition, and awareness: Affective priming with optimal and suboptimal stimulus exposures. Journal of Personaliy and Social Psychology, 64, 723–739. Musch, J., & Klauer, K. C. (2001). Locational uncertainty moderates affective congruency effects in the evaluative decision task. Cognition and Emotion, 15, 167–188. Niedenthal, P. M. (1990). Implicit perception of affective information. Journal of Experimental Social Psychology, 26, 505–527. 376 A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 351–376 Ohman, A., Esteves, F., & Soares, J. J. F. (1995). Preparedness and pre-attentive associative learning—electrodermal conditioning to masked stimuli. Journal of Psychophysiology, 9, 99–108. Pizzagalli, D., Koenig, T., Regard, M., & Lehmann, D. (1998). Faces and emotions: Brain electric field sources during covert emotional processing. Neuropsychologia, 36, 323–332. Pizzagalli, D., Lehmann, D., Hendrick, A. M., Regard, M., Pascual-Marqui, R. D., & Davidson, R. J. (2002). Affective judjments of faces modulate early activity (160 ms) within the fusiform gyri. Neuroimage, 16, 663–677. Robinson, M. D. (1998). Running from William JamesÕ bear: A review of preattentive mechanisms and their contributions to emotional experience. Cognition and Emotion, 12, 667–696. Saban, S., & Hugdahl, K. (1999). Nonaware classical conditioning to pictorial facial stimuli in a between-groups paradigm. Integrative Physiological and Behavioural Science, 34, 19–29. Snodgrass, M., Shevrin, H., & Kopka, M. (1993). The mediation of intentional judgments by unconscious perceptions: The influences of task strategy, task preference, word meaning, and motivation. Consciousness and Cognition, 2, 169–193. Stapel, D. A., Koomen, W., & Zeelenberg, M. (1998). The impact of accuracy motivation on interpretation, comparison, and correction processes: Accuracy · knowledge accessibility effects. Journal of Personality and Social Psychology, 74, 878–893. Stone, A. M., & Valentine, T. (2004). Better the devil you know? Non-conscious processing of identity and affect of famous faces. Psychonomic Bulletin & Review, 11, 469–474. Stone, A. M., & Valentine, T. (in press). Orientation of attention to non-consciously recognised famous faces. Cognition and Emotion. Stone, A. M., Valentine, T., & Davis, R. (2001). Face recognition and emotional valence: Processing without awareness by neurologically intact participants does not simulate covert recognition in prosopagnosia. Cognitive, Affective and Behavioural Neuroscience, 1, 183–191. Strack, F., Schwarz, N., Bless, H., Kubler, A., & Wanke, M. (1993). Awareness of the influence as a determinant of assimilation versus contrast. European Journal of Social Psychology, 23, 52–62. Whalen, P. J., Rauch, S. L., Etcoff, N. L., McInerney, S. C., Lee, M. B., & Jenike, M. A. (1998). Masked presentations of emotional facial expressions modulate amygdala activity without explicit knowledge. Journal of Neuroscience, 18, 411–418. Wong, P. S., Shevrin, H., & Williams, W. J. (1994). Conscious and nonconscious processes: An ERP index of an anticipatory response in a conditioning paradigm using visually masked stimuli. Psychophysiology, 31, 87–101. Zajonc, R. B. (1980). Feeling and thinking: Preferences need no inferences. American Psychologist, 35, 151–175. Zajonc, R. B. (2001). Mere exposure: A gateway to the subliminal. Current Directions in Psychological Science, 10, 224–228.

Consciousness and Cognition Consciousness and Cognition 14 (2005) 548–564 www.elsevier.com/locate/concog Strength of visual percept generated by famous faces perceived without awareness: Effects of affective valence, response latency, and visual field q Anna Stone *, Tim Valentine Department of Psychology, Goldsmiths College, University of London, UK Received 29 September 2004 Available online 5 March 2005 Abstract Participants who were unable to detect familiarity from masked 17 ms faces (Stone & Valentine, 2004, in press-b) did report a vague, partial visual percept. Two experiments investigated the relative strength of the visual percept generated by famous and unfamiliar faces, using masked 17 ms exposure. Each trial presented simultaneously a famous and an unfamiliar face, one face in LVF and the other in RVF. In one task, participants responded according to which of the faces generated the stronger visual percept, and in the other task, they attempted an explicit familiarity decision. The relative strength of the visual percept of the famous face compared to the unfamiliar face was moderated by response latency and participantsÕ attitude towards the famous person. There was also an interaction of visual field with response latency, suggesting that the right hemisphere can generate a visual percept differentiating famous from unfamiliar faces more rapidly than the left hemisphere. Participants were at chance in the explicit familiarity decision, confirming the absence of awareness of facial familiarity.
2005 Elsevier Inc. All rights reserved. Keywords: Non-conscious perception; Facial identity; Awareness; Visual masking; Affect; Attitude; Response latency; Hemisphere; Disgust q The preparation of this paper was supported by an award from the Economic and Social Research Council, UK, reference PTA-026-27-0332. * Corresponding author. E-mail address: pss01as@gold.ac.uk (A. Stone). 1053-8100/$ - see front matter
2005 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2005.01.009 A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 548–564 549 1. Introduction There is much evidence that facial expressions can be detected pre-consciously and can influence psychophysiological and behavioural responses without awareness of the particular expression (e.g., Dimberg & Ohman, 1996; Dimberg, Thunberg, & Elmehed, 2000; Johnsen & Hugdahl, 1991, 1993; Mogg & Bradley, 1999; Murphy & Zajonc, 1993; Niedenthal, 1990; Ohman, Esteves, & Soares, 1995; Robinson, 1998; Saban & Hugdahl, 1999; Whalen et al., 1998; Wong, Shevrin, & Williams, 1994). All of these studies presented masked faces for very brief exposure duration (target-to-mask stimulus onset asynchrony [SOA] of less than 35 ms). Participants were at chance in two-alternative forced-choice tasks of identifying the expression, confirming the absence of awareness. The generation of an appropriate response to a facial expression that is recognised without awareness of the expression is often interpreted in terms of the importance to the individual of detecting and reacting to the emotion of others. Many stimuli in the environment are scanned pre-consciously, and those with the greatest salience, e.g., emotional faces as opposed to neutral faces, are prioritised for processing. This raises the question of whether a famous face would be prioritised for processing in competition with an unfamiliar face when both are perceived without awareness of familiarity. Given evidence that famous faces can be recognised as specific individuals without awareness of facial identity (Banse, 1999, 2001; Stone, Valentine, & Davis, 2001) or familiarity (Stone & Valentine, 2004, in press-a, in press-b), it seems plausible that a known face would be judged more salient than an unknown face. One effect of prioritising a face for processing might be that the face would generate a stronger visual percept, even though the visual percept was vague, partial, and insufficient to permit awareness of facial identity or familiarity. The relative strength of the consciously experienced visual percept generated by famous and unfamiliar faces was investigated in the present experiments. Participants performed three tasks in which masked 17 ms faces were presented in simultaneous pairs of one famous and one unfamiliar face, one face in the left visual field (LVF) and the other in the right visual field. Each pair of faces was matched on age, sex, race, pose, and facial expression. In the perceptual comparison, participants selected the face that generated the stronger consciously experienced visual percept. The rationale was the observation from previous experiments that most participants are able to gain some vague, partial visual impression of the stimulus faces, or at least the impression of ‘‘something there.’’ In the explicit familiarity decision, participants attempted to select the famous face in each pair: this was the task reported in Stone and Valentine (2004, in pressb). Overall performance at chance would indicate the absence of awareness of facial familiarity, and by assumption, the absence of awareness of facial identity. In the attention orientation task, the faces were followed by a dot-probe consisting of two small dots, either horizontal (..) or vertical (:), presented in either the LVF or the RVF, in a location corresponding to the centre of one of the famous–unfamiliar faces. Participants performed a speeded two-alternative forced-choice discrimination on the type of dot-probe. Orientation of attention to the famous face in a pair would be shown by faster or more accurate responses to the dot-probe in the same visual field as the famous face than in the opposite visual field. This task is reported in Stone and Valentine (in press-a). Each participant defined each famous person as either good or evil (Experiment 1) or low- or high-disgust evoking (Experiment 2) in a rating procedure subsequent to the experimental tasks. It was expected that the perceptual comparison would be moderated by the participantsÕ affective 550 A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 548–564 response to the famous face. Several conceptual accounts have been proposed to explain how feedback connections from high-level attributes of a stimulus can modify the strength of earlier perceptual representations of the stimulus (e.g., Di Lollo, Enns, & Rensink, 2000; Kanwisher, 2001; Martens, Wolters, & van Raamsdonk, 2002; Vogel, Luck, & Shapiro, 1998). These accounts propose that a stimulus proceeds through stages of processing from early perceptual analysis to identification and extraction of identity-dependent properties, e.g., affective valence. This can occur before awareness of the stimulus identity is achieved. Feedback connections from representations of high-level properties to earlier perceptual representations can modify the strength of these earlier representations. These conceptual accounts explain how an identity-dependent attribute of a stimulus, e.g., affective valence, can modify the strength of the consciously experienced visual percept. There are grounds for expecting that the relationship between the affect invoked by a stimulus and the strength of the visual percept would vary with response latency, reasoned as follows. Modification of the strength of the consciously experienced visual percept would require some time to become apparent, being dependent on feedback projections. Also, there is evidence from the affective priming literature that automatic effects of stimulus valence are transient (De Houwer, Hermans, & Eelen, 1998; Glaser & Banaji, 1999; Hermans, de Houwer, & Eelen, 1994, 2001; Klauer, Rossnagel, & Musch, 1997; see Fazio, 2001, for a review). Applying these concepts to the present experiments, the activation of affective valence associated with a famous face was expected to modulate the strength of the visual percept within a range of response latencies, not including very fast or slower latencies. What effect would this modulation have? In the explicit familiarity decision (Stone & Valentine, 2004, in press-b) responses were below chance accuracy to evil-disliked faces and tended to be above chance accuracy for good-liked faces. Participants selected the paired unfamiliar face rather than the famous face if they evaluated the famous person as evil or disliked, and tended to select the famous face if they evaluated the person as good or liked. The attention orientation task suggested that attention was oriented towards a famous face if the person was evaluated as good or neutral, but oriented towards the paired unfamiliar face if the famous person was evaluated as evil (Stone & Valentine, in press-a). Stone et al. (2001) had previously reported that physiological responses to famous faces perceived without awareness of identity differed according to valence. Experiment 1 found that skin conductance responses to masked 17 ms faces were higher to the faces of famous persons subsequently evaluated ‘‘good’’ than to the faces of persons evaluated ‘‘evil,’’ but did not distinguish between famous and unfamiliar faces. (Responses tended to be higher to good faces than to unfamiliar faces, but tended to be lower to evil faces than to unfamiliar faces.) When faces were exposed for 220 ms, a duration that permits conscious recognition, there was an effect of familiarity but no effect of valence: skin conductance responses were higher to famous faces than to unfamiliar faces with no difference between ‘‘good’’ and ‘‘evil’’ faces. These results all seem to suggest that participants who regard a famous person as evil tend to process the masked 17 ms face somehow less strongly than those who regard the person as good. In the present study, it seems likely that the indistinct visual percept of the face would be weakened for participants who evaluate the person as evil, and strengthened for participants who evaluate the person as good, relative to an unfamiliar face. The perceptual comparison task asked participants to select which of the famous and unfamiliar faces in each pair yielded the stronger visual percept. ‘‘Accuracy’’ was defined as the selection of the famous face in each pair. From the above reasoning A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 548–564 551 the expectation was derived that responses to evil faces would be less accurate than responses to good faces, for some range of response latencies, not including very fast or slower latencies. The present experiments were also designed to investigate another factor: the right hemisphere superiority in processing facial identity (e.g., Grabowska & Nowicka, 1996; Heider & Groner, 1997; Schweinberger, Sommer, & Stiller, 1994; Sergent, MacDonald, & Zuck, 1994), which suggests that the right hemisphere would be expected to generate a stronger visual percept of a famous face than the left hemisphere. However, things are not this simple. Seeck et al. (1997) reported that early ERPs differed between famous and unfamiliar faces only in the right hemisphere, whereas later ERPs differed between famous and unfamiliar faces in both hemispheres. This suggests that the LH may be able to construct a visual percept that distinguishes between a famous and an unfamiliar face, but more slowly than the RH. It follows that a famous face presented in the left visual field and projected to the right hemisphere (LVF-RH) would generate a stronger visual percept than the paired unfamiliar face (presented in the RVF-LH), on short and long latency trials. In contrast, a famous face presented in the RVF-LH would generate a stronger visual percept than the paired unfamiliar face (in the LVF-RH) only on longer latency trials and not on short latency trials. This leads to the prediction that accuracy (selecting the famous face in each pair as having the stronger visual percept) will be higher for famous faces presented in the LVF than the RVF on short latency responses, with no difference in accuracy between LVF and RVF on longer latency responses. Accuracy for famous faces in the RVF should increase from short to long latency responses, while accuracy for famous faces in the LVF should not change with response latency. The original design intention was to perform analysis within participants, calculating mean accuracy for each participant for the faces rated as good vs. mean accuracy for the faces rated as evil. Experiment 1 posed the problem that the famous persons tended to be rated consistently as either good or evil, so that any differential responding to good and evil faces according to participantsÕ evaluations could be confounded with another factor that differed systematically between the stimuli, e.g., a physical attribute of the faces or the particular photographic image. To overcome this difficulty, the analysis was performed within items, calculating mean accuracy for each item over the participants rating the famous person as good vs. those rating the same famous person as evil. Thus, good and evil stimuli were identical, and the only difference was the participantsÕ attitude towards the famous persons. Uneven numbers of participants contributed to the calculations in the good and evil categories for famous persons whose rating tended to be consistent. The analysis of Experiment 1 was limited by the small number of items (n = 10) and so should be regarded as illustrative and requiring replication. Experiment 2 provides the replication. For convenience, throughout the remainder of this paper, an accurate response will refer to the selection of the famous face in each pair. 2. Experiment 1 2.1. Method 2.1.1. Participants Participants were 34 students, staff and visitors at Goldsmiths College, London. Each participantÕs individual performance was at chance in the explicit familiarity task (binomial distribution, 552 A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 548–564 one-tailed, cut-off at 65%, a = 0.05). Data were excluded from seven participants who failed to identify a minimum of eight faces in total, including two evaluated as good and two evaluated as evil, in the post-experimental evaluation. The remaining 27 participants were 18 female and 9 male, aged between 20 and 51, mean = 27.2, SD = 7.5 years. 2.1.2. Stimuli Photographs of famous and unknown faces of a uniform quality were digitised to produce images of 16 greys, 150 · 200 pixels in size. The stimulus set comprised 10 pairs of one famous with one unfamiliar face. The faces in each pair were matched on sex, race, and approximate age, and showed a similar pose and facial expression. Names and examples of stimuli are given in Appendices A and B. The mask was a collage of parts of unfamiliar faces, of the same size as the famous and unfamiliar faces. Stone et al. (2001) suggested that very few faces could be recognised when presented for 17 ms with a mask similar to that used in the present series of experiments. In that study, faces were presented singly and centrally, and it was expected that conscious identification would be even less likely with faces presented off-centre in simultaneous pairs. 2.1.3. Apparatus A personal computer running MEL2 software was used to display the faces at a 640 · 480 screen resolution. Response times and accuracy of response were measured and recorded by the computer. 2.1.4. Design Participants performed three separate tasks with masked 17 ms faces, always in the sequence of attention orientation, explicit familiarity, and perceptual comparison. The perceptual comparison task was always performed last in order to maximise the likelihood that participants had started to gain some visual percept of the masked faces. The attention orientation task was described in Section 1. The explicit familiarity and perceptual comparison tasks were similar: the explicit familiarity task asked participants to select the famous face in each pair, while the perceptual comparison task asked participants to select the face that yielded the stronger visual impression. The dependent variable was accuracy of response, and a correct response was scored by selecting the famous face. Each face pair was presented four times, with the famous face appearing twice each in left and right visual fields, for a total of 40 trials, presented in a single block. The sequence of presentation was randomised by the computer for each participant. Valence (evil or good) was derived from an evaluation given by each participant for each famous person after the three tasks of attention orientation, explicit familiarity, and perceptual comparison. 2.1.5. Procedure Participants performed individually in a darkened, air-conditioned room at a constant level of background lighting. Stimulus presentation was identical in both tasks. The two faces were each approximately 4.5 cm by 6 cm and were presented at a distance of 9 cm apart, subtending a visual angle of approximately 4 from fixation. The masks were presented in the same screen position as the faces. A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 548–564 553 The sequence of events on each trial was as follows: central fixation cross for 500 ms, forward masks for 100 ms, famous and unfamiliar face for 17 ms, backward masks for 100 ms, question ‘‘left or right’’ in the centre of the screen displayed until the participants responded. The explicit familiarity response was made by pressing one of two keys: to the left of the keyboard to indicate the face in the LVF, and to the right of the keyboard to indicate the face in the RVF. In the perceptual comparison task there was a third response option of ‘‘about equal’’ made by pressing a key in the centre of the keyboard. Response time and accuracy were recorded by the computer. Each trial was initiated by the response to the previous trial after an inter-trial interval of 1 s. Participants were informed that two faces would be flashed up very briefly, one on either side of the screen, preceded and followed by a mask comprised of a collage of parts of unfamiliar faces. Each pair of faces would contain one famous person and one unfamiliar person. In the explicit familiarity task, participants were asked to select on which side of the screen the famous face had appeared. Participants were told they would find it very difficult to see the real faces and this should be no cause for concern, but they should attend carefully to the screen, wait for the question, and respond. They were informed that it was OK to guess if they were unable to see anything of the faces. Participants were asked to look at the central fixation cross before each trial and to respond as quickly and accurately as possible. The perceptual comparison task was similar, with two differences: the instructions asked participants to select the face that generated the stronger visual percept, and a third response option of ‘‘about equal’’ was offered. After each task, participants were asked whether they had been able to recognise any of the faces displayed during the experiment, and were strongly encouraged to guess. The participant was shown the famous faces one at a time in a random sequence and asked to identify each person, either by name or by sufficient biographical detail to uniquely pinpoint the person. After the face identification, the participant was shown the famous faces again, one at a time, in a different random sequence, and asked to rate each person on a 7-point scale from 3 (very evil) through 0 (neutral) to +3 (very good). Participants were asked to evaluate the person, not the face, considering any knowledge they had of the person. Participants were told, ‘‘There are no right or wrong answers, it is entirely your own opinion. Please do not think too long and give your first impression.’’ Finally, participants were debriefed and thanked for their participation. 2.2. Results If a participant could not correctly identify a famous face in the post-experimental identification, all trials for this combination of participant and item were excluded from the analysis (4.8% of trials). All participants insisted they had been unable to recognise any of the faces during the experimental tasks. Trials were excluded if the response time was faster than 100 ms (probable anticipations; including the backward mask this was 200 ms from face offset; 0.1% of trials) and over 2000 ms (5.5% of trials). A face was categorised as evil for a participant if the valence rating was below zero, and as good if the valence rating was 0 or above, to distinguish between evil faces and the rest. A correct response was scored as the selection of the famous face. The presence of a third response option of ‘‘about equal’’ in the perceptual comparison, which could not score as a correct response, resulted in mean accuracy (selecting the famous face in each pair) below 50%. 554 A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 548–564 2.2.1. Explicit familiarity Response accuracy per item (mean = 0.470, SE = 0.016) was obviously not above chance, confirming the absence of awareness of facial familiarity, and by assumption, of facial identity. As reported in Stone and Valentine (2004, Experiment 1), accuracy was lower in participants who evaluated a famous person as evil than in participants who evaluated the same target person as good. 2.2.2. Perceptual comparison: Analysis of valence and response latency Responses were analysed in six ranges measured from face offset: 200–500 ms, 500–700 ms, 700–900 ms, 900–1100 ms, 1100–1500 ms, and over 1500 ms. The selection of these ranges was a compromise between the desire to include as many ranges as possible, to provide a sensitive analysis of the data, and the desire to maximise the number of items without missing data, which requires fewer ranges: 5 of the 10 items has complete data. The expectation was that response accuracy would be lower for participants who evaluated a famous person as evil than for those who evaluated the same person as good, for some range of response latencies, not including the very fast or slower latencies. Fig. 1 presents an illustration of the data. No statistical tests were performed because of the small number of items. Fig. 1 shows a tendency for response accuracy to dip for evil faces in the range 500–700 ms, while response accuracy for good faces shows a peak in the same latency range. A similar pattern was observed when partial data for all 10 items were included. This pattern suggests that affective modulation of the strength of the visual percept may occur for responses in the range of 500–700 ms. 2.2.3. Perceptual comparison: Analysis of visual field and response latency A separate prediction had been made that accuracy for famous faces presented in the RVF should increase from short to long latency responses, while accuracy for famous faces presented Fig. 1. Mean response accuracy in the perceptual comparison task of Experiment 1, by valence and response latency (A) and by visual field and response latency (B). A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 548–564 555 in the LVF should be equivalent on short and long latency responses. Data were calculated in the same latency ranges used above, for famous faces presented in the LVF and RVF. One item had missing data. Fig. 1 presents an illustration; no statistical tests were performed because of the small number of items. The data presented in Fig. 1 appear consistent with expectation. 2.3. Discussion Fig. 1 illustrates the possibility that the strength of the consciously experienced visual percept of a masked 17 ms famous face varied with response latency and affective valence, and with response latency and visual field. The visual percept of an ‘‘evil’’ face may have been weakened, and that of a ‘‘good’’ face may have been strengthened, for responses in the latency range 500–700 ms. The right hemisphere may be able to construct a stronger visual percept of a famous face than an unfamiliar face at all response latencies, while the left hemisphere may be able to do so only at longer latencies. The experiment should be regarded as illustrative owing to the small number of items. Experiment 2 was designed to include a much larger number of items to enable a formal statistical analysis. 3. Experiment 2 The number of stimuli was increased from 10 to 60. The famous persons were selected from a previous study on the criterion that they generated mixed evaluations, which would permit a rigorous within-items analysis. Participants were asked to evaluate the degree of disgust evoked by each famous person. The emotion of disgust was chosen because this was thought to underlie the effects previously observed in the explicit familiarity task (Stone & Valentine, 2004, in press-b) and the attention orientation task (Stone & Valentine, in press-a). The emotion of disgust serves to protect against physical or psychological contamination and motivates avoidance of the object of disgust (e.g., Charash & McKay, 2002; Druschel & Sherman, 1999; Izard, 1977; Levenson, 1994; Nabi, 2002; Newhagen, 1998; Rozin, Haidt, & McCauley, 1999). Psychological contamination could occur because of association with an unpleasant individual, and disgust has been specifically related to the avoidance of ideas or persons regarded as morally corrupt (Izard, 1977; Nabi, 2002; Rozin et al., 1999). Thus, disgust was thought to underlie the below chance accuracy of explicit familiarity responses to the faces of famous persons evaluated as evil, and the orientation of attention away from these faces. For the interaction of disgust rating with response latency, the predictions were as follows: responses to high-disgust evoking faces would be of lower accuracy on trials with latency in the range 500–700 ms than on shorter or longer latency trials; responses to low-disgust evoking faces would be of higher accuracy in the range 500–700 ms than on shorter or longer latency trials. For the interaction of visual field with response latency, the prediction was that accuracy would increase from short to long latency trials for famous faces presented in the RVF-LH, and show no change for famous faces presented in the LVF-RH. The response option ‘‘equal’’ was removed from the perceptual comparison so that participants were compelled to select either the LVF face or the RVF face on each trial. 556 A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 548–564 3.1. Method Only the changes from Experiment 1 will be noted. 3.1.1. Participants Participants were 46 first-year undergraduate students at Goldsmiths College, London. Three participants were excluded whose individual performance was above chance in selecting famous compared to unfamiliar faces in the explicit familiarity decision (binomial distribution, one-tailed, cut-off at 57% correct, a = 0.05) since for these participants, the possibility of some awareness cannot be ruled out. Two more participants were excluded who correctly identified fewer than 40 of the 60 items, and one participant who failed to comply with experimental instructions. The remaining 40 participants were 33 female and 7 male, aged between 18 and 44, mean = 22.1, SD = 6.8 years. All had watched UK television for at least 5 years by self-report to maximise the likelihood of knowledge of the famous faces. 3.1.2. Stimuli The stimulus set comprised 60 pairs of one famous with one unfamiliar face. The faces in each pair were matched on sex, race, and approximate age, and showed a similar pose and facial expression. No data were collected to verify equivalence between the famous and unfamiliar faces on distinctiveness, attractiveness, or any other feature on which the stimuli might vary. The intention was to perform analyses within-items, with each famous person rated as low-disgust evoking by some participants and as high-disgust evoking by others, so that systematic variations between famous faces and their paired unfamiliar faces could not explain any observed experimental result. Names and examples of stimuli are given in Appendix A. 3.1.3. Design Participants performed three tasks on masked 17 ms faces, always in the sequence of attention orientation, explicit familiarity, and perceptual comparison. In each of the explicit familiarity and perceptual comparison, there were two factors of visual field of famous face (LVF or RVF) and evoked disgust (high or low; defined by each participant for each item). The dependent variable was accuracy of response, and a correct response was scored by selecting the famous face. Each face pair was presented twice, with the famous face appearing once each in left and right visual fields, for a total of 120 trials, presented in a single block. The sequence of presentation was randomised by the computer for each participant. 3.1.4. Procedure The procedure for the explicit familiarity and perceptual comparison tasks was the same as Experiment 1, with the following changes. One, participants had to make a response within 2000 ms of the onset of the question ‘‘left or right’’ or the program proceeded to the next trial and no response was accepted. Two, the ‘‘equal’’ response option was not allowed so that participants were compelled to select either the LVF or the RVF face. Three, the evaluation of the famous persons was altered. After identifying the faces, participants were shown the famous faces, one at a time, in a random sequence, and asked to rate on A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 548–564 557 a 7-point scale ‘‘how much each famous person disgusts you’’ (1 = not at all, 4 = moderately, 7 = very much). The emotion of disgust was explained as similar to distaste and disapproval. Participants were asked to evaluate the disgust invoked by the person, not the face, considering any knowledge they had of the person. Participants were told, ‘‘There are no right or wrong answers, it is entirely your own opinion. Please do not think too long and give your first impression.’’ Finally, participants were debriefed and thanked for their participation. 3.2. Results If a participant could not correctly identify a famous face, all trials for this combination of participant and item were excluded from the analysis of all tasks (11.4% of trials). Some faces were recognised during the explicit familiarity or perceptual comparison tasks. Taking a cautious approach, all trials for these combinations of participant and item were excluded from the analysis of both tasks (0.8% of trials). Trials on which no response was made were excluded from the analysis (1.4%/1.2% of trials in the explicit familiarity/perceptual comparison). Trials were excluded as probable anticipations if the response time was faster than 100 ms (including the backward mask this was 200 ms from face offset; 4.6%/6.8% of trials). A correct response was scored as the selection of the famous face. 3.2.1. Explicit familiarity Response accuracy per item (mean = 0.509, SE = 0.009) was at chance, t (59) = 0.99, ns, confirming the absence of awareness of facial familiarity, and by assumption, of facial identity. 3.2.2. Perceptual comparison: Analysis of evoked disgust and response latency A face was categorised as high-disgust evoking for a participant if the rating of disgust was 5 or above, and as low-disgust evoking if the rating was 4 or below (minimum rating was 1 and maximum was 7). Responses were analysed in three latency ranges, measured from stimulus face offset: 200–500 ms, 500–700 ms, and 700–2000 ms. These were chosen because the 500–700 ms range had suggested lowest accuracy for evil faces, and a peak in accuracy for good faces, in Experiment 1. The proportion of trials in each latency range was 0.36, 0.29, and 0.35, respectively. ANOVA was performed with two within-item factors of evoked disgust (high vs. low) and response latency (200–500 ms vs. 500–700 ms vs. 700–2000 ms). The dependent variable was mean response accuracy. Fifteen items had missing data. The two-way interaction of evoked disgust with response latency was significant, F (2, 43) = 4.24, MSE = 0.045, p < .03. Paired comparisons revealed that for the high-disgust faces, 500–700 ms latency responses were less accurate than shorter or longer latency responses, F (1, 44) = 6.80, MSE = 0.128, p = .012, while shorter and longer latency responses did not differ from each other, F = 0. For the low-disgust faces, 500– 700 ms latency responses tended to be more accurate than shorter or longer latency responses, F (1, 44) = 2.49, ns, and shorter and longer latency responses did not differ from each other, F < 1. See Fig. 2 and Table 1. 3.2.3. Perceptual comparison: Analysis of visual field and response latency ANOVA was performed with two within-item factors of famous face visual field (LVF vs. RVF) and response latency (200–500 ms vs. 500–700 ms vs. 700–2000 ms). The dependent variable 558 A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 548–564 Fig. 2. Mean response accuracy in the perceptual comparison task of Experiment 2, by valence and response latency (A) and by visual field and response latency (B). Bars represent 95% confidence intervals calculated according to Eq. (4) of Loftus and Masson (1994). Table 1 Mean accuracy (and standard error) by response latency and degree of evoked disgust (n = 45) or visual field (n = 60) Response latency (ms) Evoked disgust Visual field High Low LVF RVF 200–500 500–700 700–2000 0.579 (0.036) 0.439 (0.044) 0.577 (0.034) 0.529 (0.023) 0.560 (0.021) 0.520 (0.019) 0.583 (0.021) 0.562 (0.021) 0.538 (0.023) 0.476 (0.021) 0.500 (0.022) 0.543 (0.020) Data from Experiment 2. was mean response accuracy. No items had missing data. The two-way interaction of famous face visual field with response latency was significant, F (2, 58) = 4.55, MSE = 0.021, p < .02. Pairedsamples t tests (with a set at 0.0167) revealed that 200–500 ms responses were more accurate when the famous face was in the LVF than the RVF, t (59) = 3.99, p < .001, and 500–700 ms responses tended to be more accurate for famous faces in the LVF than the RVF, t (59) = 2.10, p = .04, while 700–2000 ms responses showed no difference between LVF and RVF, t (59) = 0.17, ns. For famous faces in the LVF, 200–500 ms and 700–2000 ms responses were equally accurate, t (59) = 1.66, ns, while for famous faces in the RVF, 700–2000 ms responses were more accurate than 200–500 ms latency responses, t (59) = 2.51, p = .015. See Fig. 2 and Table 1. 3.3. Discussion The observation of overall accuracy at chance in the explicit familiarity decision confirms that faces were perceived without awareness of familiarity, and by assumption, without awareness of identity. A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 548–564 559 In the perceptual comparison, when the famous person was rated high-disgust evoking, the relative strength of the visual percept of the famous compared to the unfamiliar face changed with response latency, being lower on 500–700 ms latency trials than on shorter or longer latency trials. When the famous person was rated low-disgust evoking, the relative strength of the visual percept for the famous compared to the unfamiliar face did not vary with response latency, although there was a non-significant tendency for response accuracy to be highest at the 500–700 ms range. This pattern of results supports the prediction for the perceptual comparison task regarding high-disgust evoking faces. The perceptual comparison showed higher response accuracy for LVF faces than RVF faces on short latency trials, with no difference on long latency trials. Accuracy increased with response latency for RVF faces but not for LVF faces. This pattern of results supports the concept that the right hemisphere can construct a stronger visual percept of a famous than an unfamiliar face more rapidly than the left hemisphere. 4. General discussion The explicit familiarity decision showed overall accuracy at chance in Experiments 1 and 2, confirming the absence of awareness of facial familiarity and, by assumption, of facial identity. The perceptual comparison showed an interesting pattern: the relative strength of the visual percept of a famous face compared to an unfamiliar face varied with response latency and participantsÕ attitude towards the famous person. For participants who rated the famous person as evoking high disgust, the visual percept of the famous face was relatively strong on short latency trials (below 500 ms), then declined on trials with latency in the range 500–700 ms, before increasing again on longer latency trials (over 700 ms). A tendency for this pattern was observed in Experiment 1 (the small number of items did not permit statistical tests) and confirmed in Experiment 2. The relative weakness of the visual percept of high-disgust evoking faces on 500–700 ms latency trials was attributed to re-entrant feedback from the representation of the affective valence to the earlier representation of the visual image of the face. The feedback required some time to become effective and was also transient, so the visual percept was relatively strong on shorter latency trials (below 500 ms) and on longer latency trials (over 700 ms). Conceptual models of feedback processing, mentioned in Section 1, will be described in some more detail in order to interpret the present findings. Vogel et al. (1998) proposed that processing of visual stimuli proceeds in two stages: the perceptual stage that identifies stimuli and occurs without awareness, and a post-perceptual stage of processing that may result in awareness. They suggested that the visual system is able to identify stimuli faster than they can be processed by post-perceptual systems. One implication is that interference with post-perceptual processing (for example, by backward masking) could impair awareness and the accuracy of overt report without impairing perceptual processing. Another implication is that modulation of post-perceptual processing by identity-dependent affective valence could result in enhanced or weakened awareness of the stimulus. Martens et al. (2002) cite converging evidence that awareness of the presence and identity of a visual stimulus requires an attentional process consisting of a feedback mechanism from 560 A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 548–564 high-level representations to preceding low-level representations. This follows a feedforward cycle that activates representations in subsequent processing levels, up to stimulus identity and meaning. Visual awareness is critically dependent on the feedback cycle re-activating early representations in primary visual cortex. Such feedback can be interpreted as a process of binding the high-level representations to the lower-level representations that caused their activation. This would seem to allow the possibility that high-level identity-dependent stimulus properties could modulate the feedback mechanism and so influence the low-level visual representations. Both of these models (Martens et al., 2002; Vogel et al., 1998) appear to have conceptual similarity with the theorising of Kanwisher (2001) that awareness of a stimulus requires a link between semantic ‘‘type’’ information and spatio-temporal ‘‘token’’ information. This link might occupy the same conceptual function as the post-perceptual stages of Vogel et al and the feedback cycle of Martens et al. Di Lollo et al. (2000) developed an explicit computational model (CMOS) along similar lines of reasoning. The CMOS model explains that processing of a visual stimulus proceeds through sequential levels increasing in abstractness from the visuo-spatial event. Re-entrant neural projections from association cortex attempt to connect with low-level representations in primary visual cortex (a post-perceptual feedback process). Awareness of a stimulus depends on a match between the re-entrant high-level visual representation and ongoing lower-level activity in primary visual cortex. The CMOS model accounts for the effectiveness of backward masking by proposing that the mask replaces the masked stimulus as the object of ongoing lower-level activity, producing a mismatch with the re-entrant visual representation of the stimulus, and so precluding awareness of the stimulus. If the masked stimulus is still generating some attenuated lower-level activity then presumably, a partial match with the re-entrant visual representation can be made, and so a vague, partial visual percept can be experienced. Affective modulation of the re-entrant neural projections from association cortex to primary visual cortex would result in a consciously experienced visual percept whose strength depends on the affect invoked by the stimulus. Any of these conceptual accounts could explain how an identity-dependent attribute of a stimulus (e.g., affective valence) can modify the strength of the consciously experienced visual percept in the absence of awareness of stimulus identity or even of stimulus familiarity. The relative strength of the visual percept of famous faces compared to unfamiliar faces also varied with visual field and response latency. The visual percept was stronger for famous faces presented in the LVF (to the right hemisphere) than for famous faces presented in the RVF at short response latency (up to 500 ms), with no difference at longer latencies; the visual percept strengthened from short to long latency trials for famous faces presented in the RVF, and showed no change for famous faces presented in the LVF. This pattern supports the proposition that the right hemisphere can construct a visual percept that distinguishes between famous and unfamiliar faces more rapidly than the left hemisphere (e.g., Seeck et al., 1997). It is possible that the famous faces may have enjoyed some systematic advantage in the perceptual comparison over the unfamiliar faces, perhaps being more attractive or more distinctive, or perhaps the advantage of perceptual fluency (e.g., Jacoby & Dallas, 1981) since the perceptual system has had previous experience of processing a famous face. Any systematic difference that fa- A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 548–564 561 voured famous faces over unfamiliar faces could have had the effect of elevating overall response accuracy. However, this factor alone could not explain the pattern of accuracy dependent on response latency for high-disgust evoking faces, or the pattern of accuracy varying with visual field and response latency. Regarding future studies, event-related potentials might help to shed light on the temporal pattern of differences in neural activity dependent on the degree of disgust evoked by famous faces. Also, fMRI data could clarify those brain regions in which differences in neural activity occur. The generality of these findings could be tested by repeating the experiment using a different class of stimulus, for example words or pictures, or using faces depicting an expression of disgust, rather than facial identities. In conclusion, when famous faces were presented so briefly that they could not be consciously perceived with sufficient clarity to permit identification or even familiarity detection, they were processed differently according to the participantsÕ emotional reaction to the famous person. The differences reported here appear to relate to the emotion of disgust and are consistent with avoidance or weaker processing of the faces of famous persons evoking a high degree of disgust. Appendix A. Stimuli Experiment 1 Pop stars: Mick Jagger, Cliff Richard Politicians: Adolf Hitler, Saddam Hussein, J.F.Kennedy, Margaret Thatcher TV presenters: Chris Evans Film/TV actors: Richard Gere Others: Myra Hindley (murderess), Mike Tyson (boxer and rapist) Experiment 2 Pop stars: Victoria Beckham, Cher, Eminem, Liam Gallagher, Geri Halliwell, Whitney Houston, Janet Jackson, Michael Jackson, Mick Jagger, Elton John, Jennifer Lopez, Madonna, George Michael, Elvis Presley, Cliff Richard, Britney Spears, Robbie Williams Royal family: Prince Charles, Queen Elizabeth, Sarah Ferguson Politicians: Osama Bin Laden, Tony Blair, George W. Bush, Bill Clinton, William Hague, Adolf Hitler, J.F. Kennedy, John Major, Margaret Thatcher TV presenters: Michael Barrymore, Cilla Black, Paul Daniels, Chris Evans, Bruce Forsyth, Rolf Harris, Richard Madeley, Anne Robinson, Jonathan Ross, Chris Tarrant, Carol Vordeman Film/TV actors: Jim Carrey, Martin Clunes, Russell Crowe, Tom Cruise, Leonardo Dicaprio, Michael Douglas, Callista Flockhart, Sarah Michelle Gellar, Hugh Grant, Ross Kemp, Gwynneth Paltrow, Arnold Schwarzenegger, Sylvester Stallone, Catherine Zeta-Jones Others: Rowan Atkinson (comedian), David Beckham (sports), Richard Branson (entrepreneur), Naomi Campbell (model), Luciano Pavarotti (opera singer), O.J. Simpson (sports) 562 A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 548–564 Appendix B. Examples of stimuli and the mask References Banse, R. (1999). Automatic evaluation of self and significant others: Affective priming in close relationships. Journal of Social and Personal Relationships, 3, 803–821. Banse, R. (2001). Affective priming with liked and disliked persons: Prime visibility determines congruency and incongruency effect. Cognition and Emotion, 15, 201–220. A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 548–564 563 Charash, M., & McKay, D. (2002). Attention bias for disgust. Anxiety Disorders, 16, 529–541. De Houwer, J., Hermans, D., & Eelen, P. (1998). Affective and identity priming with episodically associated stimuli. Cognition and Emotion, 12, 145–169. Di Lollo, V., Enns, J. T., & Rensink, R. A. (2000). Competition for consciousness among visual events: The psychophysics of reentrant visual processes. Journal of Experimental Psychology: General, 129, 481–507. Dimberg, U., & Ohman, A. (1996). Behold the wrath: Psychophysiological responses to facial stimuli. Motivation and Emotion, 20, 149–182. Dimberg, U., Thunberg, M., & Elmehed, K. (2000). Unconscious facial reactions to emotional facial expressions. Psychological Science, 11, 86–89. Druschel, B. A., & Sherman, M. F. (1999). Disgust sensitivity as a function of the Big Five and gender. Personality and Individual Differences, 26, 739–748. Fazio, R. H. (2001). On the automatic activation of associated evaluations: An overview. Cognition and Emotion, 15, 115–141. Glaser, J., & Banaji, M. R. (1999). When fair is foul and foul is fair: Reverse priming in automatic evaluation. Journal of Personality and Social Psychology, 77, 669–687. Grabowska, A., & Nowicka, A. (1996). Visual-spatial-frequency model of cerebral asymmetry: A critical survey of behavioral and electrophysiological studies. Psychological Bulletin, 120, 434–449. Heider, B., & Groner, R. (1997). Backward masking of words and faces: Evidence for different processing speeds in the hemispheres. Neuropsychologia, 35, 1113–1120. Hermans, D., de Houwer, J., & Eelen, P. (1994). The affective priming effect: Automatic activation of evaluative information in memory. Cognition and Emotion, 8, 515–533. Hermans, D., de Houwer, J., & Eelen, P. (2001). A time course analysis of the affective priming effect. Cognition and Emotion, 15, 143–165. Izard, C. E. (1977). Human emotions. New York: Plenum Press, pp 329-354.. Jacoby, L. L., & Dallas, M. (1981). On the relationship between autobiographical memory and perceptual learning. Journal of Experimental Psychology: General, 110, 306–340. Johnsen, B. H., & Hugdahl, K. (1991). Hemispheric asymmetry in conditioning to facial emotional expressions. Psychophysiology, 28, 154–162. Johnsen, B. H., & Hugdahl, K. (1993). Right hemisphere representation of autonomic conditioning to facial emotional expressions. Psychophysiology, 30, 274–278. Kanwisher, N. (2001). Neural events and perceptual awareness. Cognition, 79, 9–113. Klauer, K. C., Rossnagel, C., & Musch, J. (1997). List-context effects in evaluative priming. Journal of Experimental Psychology: Learning, Memory, and Cognition, 23, 246–255. Levenson, R. W. (1994). The search for autonomic specificity. In P. Ekman & R. J. Davidson (Eds.), The nature of emotion: Fundamental questions. Oxford: Oxford University Press. Loftus, G. R., & Masson, M. E. J. (1994). Using confidence intervals in within-subject designs. Psychonomic Bulletin & Review, 1, 476–490. Martens, S., Wolters, G., & van Raamsdonk, M. (2002). Blinks of the mind: Memory effects of attentional processes. Journal of Experimental Psychology: Human Perception and Performance, 28, 1275–1287. Mogg, K., & Bradley, B. P. (1999). Orienting of attention to threatening facial expressions presented under conditions of restricted awareness. Cognition and Emotion, 13, 713–740. Murphy, S. T., & Zajonc, R. B. (1993). Affect, cognition, and awareness: Affective priming with optimal and suboptimal stimulus exposures. Journal of Personality and Social Psychology, 64, 723–739. Nabi, R. L. (2002). The theoretical versus the lay meaning of disgust: Implications for emotion research. Cognition and Emotion, 16, 695–703. Newhagen, J. E. (1998). TV news images that induce anger, fear and disgust: Effects on approach-avoidance and memory. Journal of Broadcasting and Electronic Media, 42, 265–276. Niedenthal, P. M. (1990). Implicit perception of affective information. Journal of Experimental Social Psychology, 26, 505–527. Ohman, A., Esteves, F., & Soares, J. J. F. (1995). Preparedness and pre-attentive associative learning-electrodermal conditioning to masked stimuli. Journal of Psychophysiology, 9, 99–108. 564 A. Stone, T. Valentine / Consciousness and Cognition 14 (2005) 548–564 Robinson, M. D. (1998). Running from William JamesÕ bear: A review of preattentive mechanisms and their contributions to emotional experience. Cognition and Emotion, 12, 667–696. Rozin, P., Haidt, J., & McCauley, C. R. (1999). Disgust: The body and soul emotion. In T. Dalgleish & M. Power (Eds.), Handbook of cognition and emotion (pp. 429–445). Chichester: Wiley. Saban, S., & Hugdahl, K. (1999). Nonaware classical conditioning to pictorial facial stimuli in a between-groups paradigm. Integrative Physiological and Behavioural Science, 34, 19–29. Schweinberger, S. R., Sommer, W., & Stiller, R. M. (1994). Event-related potentials and models of performance asymmetries in face and word recognition. Neuropsychologia, 32, 175–191. Seeck, M., Michel, C. M., Mainwaring, N., Cosgrove, R., Blume, H., Ives, J., et al. (1997). Evidence for rapid face recognition from human scalp and intracranial electrodes. Neuroreport, 8, 2749–2754. Sergent, J., MacDonald, B., & Zuck, E. (1994). Structural and functional organisation of knowledge about faces and proper names—A PET study. Attention and Performance, 15, 203–228. Stone, A. M., & Valentine, T. (2004). Better the devil you know? Non-conscious processing of identity and affect of famous faces. Psychonomic Bulletin & Review, 11, 469–474. Stone, A. M., & Valentine, T. (in press-a). Orientation of attention to non-consciously recognised famous faces. Cognition and Emotion. Stone, A. M., & Valentine, T. (in press-b). Accuracy of familiarity decisions to famous faces perceived without awareness depends on attitude to the target person and on response latency. Consciousness and Cognition doi:10.1016/j.concog.2004.09.002. Stone, A. M., Valentine, T., & Davis, R. (2001). Face recognition and emotional valence: Processing without awareness by neurologically intact participants does not simulate covert recognition in prosopagnosia. Cognitive, Affective and Behavioural Neuroscience, 1, 183–191. Vogel, E. K., Luck, S. J., & Shapiro, K. L. (1998). Electrophysiological evidence for a postperceptual locus of suppression during the attentional blink. Journal of Experimental Psychology: Human Perception and Performance, 24, 1656–1674. Whalen, P. J., Rauch, S. L., Etcoff, N. L., McInerney, S. C., Lee, M. B., & Jenike, M. A. (1998). Masked presentations of emotional facial expressions modulate amygdala activity without explicit knowledge. Journal of Neuroscience, 18, 411–418. Wong, P. S., Shevrin, H., & Williams, W. J. (1994). Conscious and nonconscious processes: An ERP index of an anticipatory response in a conditioning paradigm using visually masked stimuli. Psychophysiology, 31, 87–101.

Consciousness and Cognition 43 (2016) 57–65 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Does sensitivity in binary choice tasks depend on response modality? q Izabela Szumska a,b,⇑, Rob H.J. van der Lubbe b,c, Lukasz Grzeczkowski a, Michael H. Herzog a a Laboratory of Psychophysics, Brain Mind Institute, Ecole Polytechnique Fédérale de Lausanne (EPFL), Switzerland Department of Cognitive Psychology, University of Finance and Management, Warsaw, Poland c Cognitive Psychology and Ergonomics, University of Twente, The Netherlands b a r t i c l e i n f o Article history: Received 7 September 2015 Revised 13 May 2016 Accepted 14 May 2016 Keywords: Metacontrast masking Response modality Saccades Verbal responses Button presses a b s t r a c t In most models of vision, a stimulus is processed in a series of dedicated visual areas, leading to categorization of this stimulus, and possible decision, which subsequently may be mapped onto a motor-response. In these models, stimulus processing is thought to be independent of the response modality. However, in theories of event coding, common coding, and sensorimotor contingency, stimuli may be very specifically mapped onto certain motor-responses. Here, we compared performance in a shape localization task and used three different response modalities: manual, saccadic, and verbal. Meta-contrast masking was employed at various inter-stimulus intervals (ISI) to manipulate target visibility. Although we found major differences in reaction times for the three response modalities, accuracy remained at the same level for each response modality (and all ISIs). Our results support the view that stimulus-response (S-R) associations exist only for specific instances, such as reflexes or skills, but not for arbitrary S-R pairings. Ó 2016 Elsevier Inc. All rights reserved. 1. Introduction The very same stimuli can be mapped on very different motor responses, such as manual button presses, eye movements, or verbal responses. For example, when observers are asked to discriminate a square vs. a diamond, responses can be verbal ‘‘left”/‘‘right” expressions, left/right button presses, or left/right eye movements. Most models assume, implicitly or explicitly, that stimuli are first processed in a sensory stage (e.g., photo-transduction at the retina), followed by a series of visual processing stages, leading to representation and categorization of the stimulus (Donders, 1969; Hubel & Wiesel, 1959; Marr, 1982; Poggio, 1984; Sternberg, 1969). Based on the categorization, a decision is made, and executed as a motor-response. Processing is fully independent of the type of the motor-response. Hence, accuracy for given stimuli and tasks is also independent of the type of the motor-response (as long as motor execution errors do not differ across response modalities). In sharp contrast to these theories, sensori-motor contingency theories propose that certain stimuli may be specifically mapped onto certain motor-responses. For example, a skill like riding a bike comes with distinct sensori-motor contingencies. Similar to the Gibsonian approach (Gibson, 1966), the theory of event coding proposes that stimuli are coded by joint visuo-motor representations (e.g., Hommel & Müsseler, 2001). In the spirit of the behavioristic tradition, sometimes mental q This research was supported by a grant from the Scientific Exchange Programme founded by Rectors’ Conference of the Swiss Universities (CRUS, Project code: 10.262). ⇑ Corresponding author at: Department of Cognitive Psychology, University of Finance and Management, Pawia 55, 01-030 Warsaw, Poland. E-mail address: izabela.szumska@gmail.com (I. Szumska). http://dx.doi.org/10.1016/j.concog.2016.05.005 1053-8100/Ó 2016 Elsevier Inc. All rights reserved. 58 I. Szumska et al. / Consciousness and Cognition 43 (2016) 57–65 representations are entirely abandoned and identified with motor-actions (Nagel, Carl, Kringe, Märtin, & König, 2005; O’Regan & Noë, 2001). In such models, accuracy may differ for the same stimulus when motor responses are different. For example, for certain stimuli, eye movements may be the more ‘‘appropriate” motor-action than verbal responses, leading to superior performance (e.g., Hughes & Kelsey, 1984). Motor (e.g., saccadic) responses are faster than verbal response. They can be more automatic and may even be triggered by unconsciously perceived features. In line with this idea, performance in a near-threshold flash detection task was shown to be better when observers made a saccade than when they pressed a button (Hughes & Kelsey, 1984). Accuracy may also strongly differ when stimuli are processed along different pathways, associated with different motor responses. For example, it is often assumed that a stimulus is processed both in the ‘‘conscious” ventral stream, specialized in shape and color analysis, and in the ‘‘unconscious” dorsal stream, being specialized in motor functions (Goodale, 2014; Goodale & Milner, 1992; Milner & Goodale, 1995; Ungerleider & Mishkin, 1982). Both streams may be differently sensitive to specific stimulus features. Indeed, patients with hemianopia are still able to make a saccade or to point to a target, but they cannot verbally report about this stimulus (Marcel, 1983; Perenin & Jeannerod, 1975; Poppel, Held, & Frost, 1973). Experiments on motor priming also suggest that there are transient task dependent stimulus-response mappings (Jaśkowski, van der Lubbe, Schlotterbeck, & Verleger, 2002; Jáskowski, Skalska, & Verleger, 2003; Klotz & Neumann, 1999, etc.). In these studies, a prime stimulus (e.g., a rectangle on one side and a diamond on the other side) is followed by a slightly larger stimulus pair containing the target. Observers indicate the side on which the target (rectangle or diamond) was presented. Visibility of the prime increases when the ISI increases. Both with invisible and with visible primes, manual responses are commonly faster when the relevant shapes of the prime and the target are on the same side (i.e., congruent) as compared to when they occur on different sides (i.e., incongruent). This effect is commonly denoted as the priming effect and is often understood as being the consequence of activation of the corresponding motor response (e.g., see Klotz & Neumann, 1999). In line with the latter idea, other studies revealed that presentation of a prime already induced hand motor activation (e.g., Jaśkowski et al., 2002, 2003; Klotz, Heumann, Ansorge, & Neumann, 2007). Moreover, invisible primes may not only modulate responses to subsequent targets but may even elicit a wrong response (Jáskowski et al., 2003). Subliminal priming effects were not limited to hand motor reactions as they also facilitated verbal responses (Ansorge, Klotz, & Neumann, 1998; Eimer & Schlaghecken, 2001). Thus, motor priming is not limited to a specific response modality. Here, we asked the question whether accuracy in a visual binary choice task depends on response modality. 2. General method 2.1. Participants In total, 14 participants (ten males, four females; aged 18–28; mean 23.8) recruited from the local student population in Lausanne took part in the experiments (four in Exp. 1 and ten in Exp. 2). Participants were paid 20 CHF per hour. All observers had normal or corrected-to-normal visual acuity as determined by a computerized visual acuity test (Freiburg Visual Acuity Test; Bach, 1996). To participate in the experiment, participants had to reach a value of 1.0 (corresponding to a Snellen fraction of 20/20) for at least one eye. The experiment was conducted with the informed and written consent of each participant. All procedures complied with the Declaration of Helsinki and were approved by the local ethics committee. 2.2. Apparatus Stimuli were presented on a Philips 201B4 CRT monitor driven by a standard accelerated graphic card refreshed at 75 Hz (Exp. 1) or on an ASUS VG248QE LCD monitor with 120 Hz refreshed rate (Exp. 2). Participants were seated in a room dimly illuminated by a background light of about 0.5 lx. The observation distance was 90 cm. Participants were restrained by a chinrest. For eye movement control, we used the iView X-HiSpeed eye tracker from SensoMotoric Instruments (SMI), which was set up for binocular mode with a 500 Hz sampling frequency. Signals of both eyes were averaged in order to reduce noise. Verbal responses were registered by a Hama Stereo Directional Microphone (RMZ-14). 2.3. Stimuli and procedure Trials started with a fixation dot presented for 1 s. Subsequently, a square was presented for 13 ms (Exp. 1) or 17 ms (Exp. 2) pseudo-randomly either to the left or right from the central fixation dot at 4° of eccentricity. A diamond was simultaneously presented at the opposite side (Fig. 1). The square and diamond had an equal width of 0.8°. After a variable ISI, a mask was presented at the square and diamond location for 100 ms. The inner masks contours abutted the contours of the square and diamond. The mask size was 1°
1°. All stimuli were black outlines and were presented on a gray background (Exp. 1: 18 cd/m2, Exp. 2: 52 cd/m2). Half of the participants were instructed to indicate as accurately as possible whether the square was on the left or on the right side, while the other half of the participants had to indicate the side of the diamond. In three different sessions, responses were given with either the left or the right hand, by making a saccade to the left or right, or with verbal responses by saying ‘‘left” or ‘‘right”. A trial ended after the participant’s response or after a maximal time interval of 5 s. The order of 59 I. Szumska et al. / Consciousness and Cognition 43 (2016) 57–65 . . . 1s . 13 ms variable ISI 100 ms Fig. 1. Every trial started with a central fixation dot presented for 1000 ms. Subsequently, a square was presented on one side simultaneously with a diamond on the opposite side, either for a duration of 13 ms (Exp. 1) or 17 ms (Exp. 2). The side of the diamond (or square) varied randomly from trial to trial. After a variable ISI, a mask was presented on both sides for a duration of 100 ms. Participants were instructed to indicate the side of the target by pressing a left or right button, by making a saccade to the left or right, or by giving the verbal responses ‘‘left” or ‘‘right”. These response instructions varied between sessions. the sessions was counterbalanced. Every block consisted of 120 randomized trials (12 ISIs (Exp. 1: 0, 13, 27, 40, 53, 67, 80, 93, 107, 120, 133, 147 ms; Exp. 2: 0, 17, 25, 42, 50, 67, 83, 92, 108, 117, 133, 150 ms)
10 trials each). The experiment was conducted over 3 days with 12 blocks per day (4 blocks for each response modality). Hence, in total there were 4320 trials presented for each participant. 2.4. Data analysis Trials with premature (<50 ms) and late (>1500 ms) responses were discarded and repeated within the same block (<0.5%). Saccade direction was determined online. The eye had to land within a circular window with a radius of 3°. centered on the target. The verbal responses were recorded by microphone and they were classified (for left and right responses) by the experimenter offline. RTs were computed using MATLAB software. Participants’ responses per response modality were averaged for each ISI. For each condition, we determined a psychometric function on the basis of the distribution of the proportion of correct responses as a function of the ISI. Thus, we had nine conditions (3 response modalities
3 days) for each individual. As shown in previous studies (Albrecht, Klapötke, & Mattler, 2010; Francis & Herzog, 2004; Maksimov, Murd, & Bachmann, 2011) the shape of metacontrast masking could be either S-shaped (a so-called Type A masking, where performance increases as a function of ISI) or Ushaped (a Type B masking, good performance for short and long ISIs; Kolers, 1962). Since masking shape might vary between participants and we were mainly interested in the performance increase which occurs for medium ISIs, data for all ISIs except the shortest ISI (0 ms) from each participant and condition were fitted with a sigmoidal logistic S-shaped function. We determined the ISI for which a threshold of 75% correct was reached, and we also determined the slope of the psychometric functions. The threshold allows us to compare performance in the localization of the masked stimulus (lower threshold means higher sensitivity for the masked stimuli). The slope can provide information about the accumulation of the information (e.g., shallower slope may reflect an increase in the range of ISIs for which performance in localization task is better). The advantage using estimated parameters is that the noise per ISI can be reduced. The lapse rate parameter (the probability of an incorrect response which is independent of stimulus intensity) was set at 0.01 in Exp. 1. In Exp. 2, raw results were corrected before determining psychometric function, thus, the lapse rate parameter was set at 0. The goodness of fit of the psychometric function was determined using the method log-likelihood estimation as described in Wichmann and Hill (2001) through the Palamedes routines for MATLAB (Kingdom & Prins, 2010; Neggers, Schölvinck, van der Lubbe, & Postma, 2005). The threshold and the slope for each participant, each response modality (manual, saccadic, verbal) and each testing day (first, second, third) were calculated and statistically analyzed with a 3
3 repeated measures ANOVA. In addition, we analyzed accuracy for each ISI and each response modality (manual, saccadic, verbal) across the examination days in a 12
3 repeated measures ANOVA. When necessary, degrees of freedom were corrected with Greenhouse-Geisser e coefficients. To corroborate our possible null results, we complemented the traditional null hypothesis testing by a more informative Bayesian analysis (Wagenmakers, 2007). We computed the Bayesian Information Criterion (BIC) which indicates the probability pBIC(H0|D) that the null hypothesis is true for the available set of data D. We followed the procedure advocated by Masson (2011; see also Wagenmakers, 2007). 60 I. Szumska et al. / Consciousness and Cognition 43 (2016) 57–65 3. Experiment 1 The major aim of Experiment 1 was to explore whether thresholds and slopes differ as a function of response modality. If sensori-motor contingencies are involved, manual and saccadic responses may be privileged, thus, a lower threshold and possibly also a steeper slope for motor responses is predicted than in the case of verbal responses. Furthermore, perceptual learning might be limited to specific S-R links, thus, effects of testing day may also be restricted to the conditions with a certain response modality. 3.1. Results and discussion We found a main effect of ISI on the percentage of correct responses (PC; Fig. 2A; F(11, 33) = 29.63, p < 0.001, g2 = 0.91). The PC increased from 56% for the shortest ISI (0 ms) to 97% for the longest ISI (147 ms). There was no main effect of response modality (F(2, 6) = 0.70, p = 0.48, g2 = 0.19). All three masking functions did only slightly differ from each other (Fig. 2A). Most importantly, there was no significant interaction between response modality and ISI (F(22, 66) = 0.86, p = 0.48, g2 = 0.22). Response times (RT) decreased with increasing ISI from 1044 ms for an ISI of 0 ms to 770 ms for an ISI of 147 ms (F (11, 33) = 9.57, p < 0.001, g2 = 0.76). RTs differed between the response modalities (F(2, 6) = 23.33, p = 0.001, g2 = 0.89; Fig. 2B). RTs were significantly longer for verbal than for manual responses (1438 vs. 667 ms, t(3) = 5.17, p = 0.014) and saccades (1438 vs. 537 ms, t(3) = 4.82, p = 0.017). RTs for manual responses and saccades did not differ significantly from each other (667 vs. 537 ms, t(3) = 2.1, p = 0.13). The interaction between ISI and response modalities was significant (F(22, 66) = 3.13, p < 0.001, g2 = 0.51). Reaction time decreased with ISI but only for verbal responses (F(1, 3) = 14.48, p < 0.05, g2 = 0.83). B 100% 90% 80% 70% manual saccadic verbal 60% 50% 40% manual saccadic verbal 2000 Reacon Time (ms) Percent of Correct Responses (%) A 0 13 27 40 53 67 80 93 107 120 133 147 1600 1200 800 400 0 0 ISI (ms) 13 27 40 53 67 80 ISI (ms) 93 107 120 133 147 Fig. 2. Results of Experiment 1. A. Percentage of correct responses as a function of ISI for the three response modalities. No significant differences were observed between the different response modalities. B. RTs were significantly longer for verbal than for saccadic and manual responses. There was no significant difference between manual and saccadic response times. Response Times decreased with increasing ISIs for each response modality. Error bars indicate the standard error of the mean for 4 observers. 100 MANUAL SACCADIC VERBAL ISI (ms) 80 60 40 20 0 DAY 1 DAY 2 DAY3 DAY 1 DAY 2 DAY3 DAY 1 DAY 2 DAY3 Fig. 3. Experiment 1. ISI for which a threshold of 75% correct was reached for each day and each modality. The threshold was higher on the second day than on the third day. Error bars indicate the standard error of the mean for 4 observers. I. Szumska et al. / Consciousness and Cognition 43 (2016) 57–65 61 The threshold of the psychometric function did not depend on response modality (Fig. 3; F(2, 6) = 1.16, p = 0.37, g2 = 0.28, p(H0|D) = 0.75).1 The main effect of testing day on the threshold was significant (F(2, 6) = 6.14, p = 0.035, g2 = 0.67, p(H0|D) = 0.05). The ISI for which a threshold of 75% correct was reached, was shorter on the third day than on the second testing day (52 vs. 47 ms, t(3) = 3.82, p = 0.03, p(H0|D) = 0.03). There was no significant interaction between response modality and testing day (F(4, 12) = 0.93, p = 0.44, g2 = 0.24, p(H0|D) = 0.96). The slope of the psychometric function also did not depend on response modality (F(2, 6) = 99, p = 0.39, g2 = 0.25, p(H0|D) = 0.72), nor on the testing day (F(2, 6) = 1.06, p = 0.38, g2 = 0.26, p (H0|D) = 0.70). No significant interaction was observed between response modality and testing day (F(4, 12) = 1.11, p = 0.37, g2 = 0.27, p(H0|D) = 0.95). All non-significant results were examined with Bayesian analyses, which revealed a high probability of the null hypothesis being true. Psychometric functions for each participant and goodness of fit values are available in the Supplementary Materials. Experiment 1 revealed that location discrimination depended on ISI, however, no difference was found of the required response modality. These findings suggests that visual processing is independent from response modality, in line with classic models. An interesting outcome of Experiment 1 concerns the influence of testing day on the thresholds, which, however, did not differ as a function of response modality. Thus, perceptual learning seems to be involved in line with previous reports (Hernandez & Lefton, 1977; Hogben & Di Lollo, 1984; Schwiedrzik, Singer, & Melloni, 2009). As the same effect was found for all response modalities, our results do not support the idea that perceptual learning occurs for specific S-R links only. 4. Experiment 2 Although faster reaction times were observed for manual and saccadic responses as compared to verbal responses, it is possible that responses were not fast enough to benefit from specific S-R links. Therefore, a nearly identical experiment was carried out with speeded responses. 4.1. Methods Methods and instructions were the same as in Experiment 1, except for the explicit instruction to respond as fast and as accurately as possible. The duration of the stimuli and ISIs were slightly different due to the use of a different screen (see General Method). One block consisted of 130 randomized trials (12 ISIs
10 repetitions). The experiment was again conducted over 3 days with 12 blocks per day (4 blocks for each response modality). In total there were 4680 trials for each participant. 4.2. Results and discussion We observed a main effect of ISI (Fig. 4A; F(11, 99) = 110.517, p < 0.001, g2 = 0.92). PC increased from 58% for the shortest ISI (0 ms) to 93% for the longest ISI (150 ms). There was a main effect of response modality (F(2, 18) = 17.198, p < 0.001, g2 = 0.66). Performance was significantly worse for saccadic responses than for manual (73% vs 78%, t(6) = 2.64, p < 0.05) and verbal responses (73% vs 79%, t(6) = 3.12, p < 0.05). The interaction between response modality and ISI was also significant (F(22, 198) = 2.701, p < 0.001, g2 = 0.23). More errors were made in the case of saccadic responses than in case of manual and verbal responses, at least when ISIs were longer than 42 ms. Response time (RT) did not depend on ISI (F(11, 99) = 1.82, p = 0.06, g2 = 0.17). RTs differed between response modalities (F(2, 18) = 97.43, p < 0.001, g2 = 0.91; Fig. 4B). RTs were significantly longer for verbal than for manual responses (803 vs. 554 ms, t(9) = 9.600, p < 0.001), while manual responses were significantly longer than saccades (554 vs. 432 ms, t(9) = 6.096, p < 0.001). The interaction between ISI and response modalities was significant (F(22, 198) = 1.79, p = 0.02, g2 = 0.17). RT decreases with ISI but only for verbal responses (F(11, 99) = 2.36, p = 0.01, g2 = 0.21). In contrast with Experiment 1, we observed lower performance levels for saccades than for manual and verbal responses in Experiment 2. Since this lower performance occurs even for well visible stimuli (e.g. ISI = 150 ms), it seems that saccadic responses were more affected by lapses than manual and verbal responses. To correct for these differences, performance was modified by using the following equation: P0 ¼ pC þ ð1  pCÞðp  pCÞ ðpC  pLÞ where p0 is the probability of correct responses for a certain ISI, p is the probability of correct responses including lapses and constant errors, pC is the chance level of 0.5 and pL is the lapse rate determined as the probability of an incorrect response for the ISI of 150. After correction for differences in lapse rates (Fig. 4C), there is no longer a significant interaction (F(22, 198) = 1.276, p = 0.19, g2 = 0.12). The main effect of ISI was significant (F(11, 99) = 112.938, p < 0.001, g2 = 0.93). The main effect of response modality was also significant (F(2, 18) = 4.583, p = 0.025, g2 = 0.34). Overall, performance was slightly worse for saccadic responses than for manual responses (78% vs 80%, t(9) = 2.721, p = 0.024). 1 P(H0|D) – the probability of the null hypothesis being true for the available set of data D. Strength of the evidence: 0.50–0.75: weak, 0.75–0.95: positive, 0.95–0.99: strong, >0.99 very strong (Raftery, 1995). 62 I. Szumska et al. / Consciousness and Cognition 43 (2016) 57–65 B 100% 2000 90% Reacon Time (ms) Percent of Correct Responses (%) A 80% 70% manual saccadic verbal 60% 50% 1200 800 400 40% 0 17 25 42 50 67 83 92 manual saccadic verbal 1600 0 108 117 133 150 0 17 25 42 C 50 67 83 92 108 117 133 150 ISI (ms) ISI (ms) Percent of Correct Responses (%) 100% 90% 80% 70% manual saccadic verbal 60% 50% 0 17 25 42 50 67 83 92 108 117 133 150 ISI (ms) Fig. 4. Results of Experiment 2. A. Accuracy as a function of ISI for the three response modalities. Accuracy for saccadic responses was lower than for manual and verbal responses. B. RTs changed with ISI only for verbal responses. RTs were significantly longer for verbal than for manual responses which were significantly longer than saccades. C. Corrected accuracy as a function of ISI for the three response modalities. Saccadic responses were shorter than manual and verbal responses. Error bars indicate the standard error of the mean for 10 observers. 100 MANUAL SACCADIC VERBAL DAY 1 DAY 2 DAY3 DAY 1 DAY 2 DAY3 ISI (ms) 80 60 40 20 0 DAY 1 DAY 2 DAY3 Fig. 5. Experiment 2. ISI at which a threshold of 75% correct was reached for each day and each modality. Threshold was larger on the first day than on the second and the third day. Threshold for saccadic response was larger than for manual response but only on the first testing day. Error bars indicate the standard error of the mean for 10 observers. The threshold of the psychometric function did not depend on response modality (Fig. 5; F(2, 18) = 2.46, p = 0.11, g2 = 0.21, p(H0|D) = 0.64). The main effect of testing day on the threshold was significant (F(2, 18) = 25.97, p < 0.001, g2 = 0.74, p(H0|D) < 0.01). The ISI, at which a threshold of 75% correct was reached, was longer on the first day than on the second (75 vs. 56 ms, t(9) = 6.89, p < 0.001, p(H0|D) < 0.01) and the third testing day (75 vs 53 ms, t(9) = 7.81, p < 0.001, p(H0|D) < 0.01). The interaction between the effect of different response modalities and testing day on the threshold was significant (F(4, 36) = 2.79, p = 0.04, g2 = 0.24, p(H0|D) = 0.12). The threshold for saccadic response was larger than for manual response (84 vs 66 ms, t(9) = 3.56, p = 0.006, p(H0|D) = 0.05) but only on the first testing day. The slope of the psychometric function did not depend on the response modality (F(2, 18) = 0.032, p = 0.65, g2 = 0.034, p(H0|D) = 0.93) nor did the testing day (F(2, 18) = 0.28, p = 0.64, g2 = 0.031, p(H0|D) = 0.94). The interaction between response modality and the testing day was also not significant (F(4, 36) = 0.79, p = 0.47, g2 = 0.08, p(H0|D) = 0.99). All non-significant results were also confirmed by Bayesian analyses. They revealed for these effects that there is a high probability of the null hypothesis being true. I. Szumska et al. / Consciousness and Cognition 43 (2016) 57–65 63 An example of the psychometric functions estimated for one participant for each testing day and each response modality and goodness of fit values are available in the Supplementary Materials. In line with the results of Experiment 1, there was no indication that performance is better in the case of button presses or saccadic responses as compared to verbal responses. We found an interaction between response modality and testing day showing that the ISI for which a threshold of 75% correct is reached may be longer for saccadic than manual response, but only on the first testing day. This might indicate that sensitivity in the currently employed shape localization task indeed varies between response modalities under specific conditions. However, worse performance in the case of saccadic responses as compared to manual and verbal responses is opposite to what was shown in previous studies (e.g. Hughes & Kelsey, 1984). Thus, more errors for saccadic responses are unlikely due to changes in stimulus processing but rather by response execution errors, which are more likely for fast responses. Participants mentioned that the block with saccadic responses was very tiring, thus, it may have been more difficult to perform this specific task than the others, which explains the differences in lapse rate. We observed that performance improved over time for all response modalities. The estimated threshold was longer on the first testing day than on the second and the third day, thus, our results again do not support the idea that perceptual learning concerns the selective development of specific S-R links. 5. General discussion It is often assumed that accuracy in binary decision tasks is independent of response modality. However, this seems less obvious for tasks concerning skilled behavior where sensori-motor contingencies are key such as riding a bike. Hence, the question arises to what degree response modality influences response accuracy. We used a simple square/diamond location discrimination task and mapped stimuli on manual, verbal, and saccadic responses. The rationale was that if there are specific sensori-motor contingencies, then, there may be differences in accuracy in the case of near-threshold stimulation. Specifically, fast, saccadic responses may benefit stronger from S-R links than the other two response modalities used here. However, this was not the case. For longer ISIs, the target was clearly visible whereas for short ISIs the mask strongly impeded the visibility of the target. Assuming that verbal responses rely more on conscious processing than manual and saccadic responses, we expected differences in performance between the different response modalities. However, we found no deterioration of performance for verbal as compared to manual and saccadic responses, for all ISIs. Other studies have found differences between response modalities supporting the presence of specific sensori-motor contingencies. For example, Bekkering and Neggers (2002) showed that observers were better in detecting the size and orientation of an object when they were asked to prepare to grasp this object than when they were asked to reach for it. However, these effects may be explained in other ways, as the focus of attention may simply be different depending on the type of the prepared action. Hughes and Kelsey (1984) asked participants to detect a near-threshold flash. Performance was better when observers performed a saccade towards the target than when they had to make a button press. However, the difference may be also explained by the presence of exogenous saccades that originate from the tecto-pulvinar pathway. Jaśkowski and Sobieralska (2004) showed that the difference between saccadic and manual response modalities disappeared when endogenous saccades were made, which is in line with our findings. It has also been proposed that information for more automatic and unconscious reactions (e.g., manual responses) is fast accessible, but also short-lasting (Breitmeyer, 2014; Livingstone & Hubel, 1988; Milner & Goodale, 1995; Schenk & McIntosh, 2010). Thus, fast saccadic reactions may benefit more from specific S-R links. Larger priming effects when observers respond under time pressure (Forster & Davis, 1991; Greenwald, Draine, & Abrams, 1996) may be also explained by an initial boost in prime activity that strongly modulates responses under high time pressure, while in the case of low time pressure this priming effect is already diminished. In line with this idea, when a briefly presented disk is followed by a ring, detection of the disk is better in the case of a direct response than in the case of a delayed response (Lachter, Durgin, & Washington, 2000). In perceptual illusion, manual responses (e.g., grip aperture), in contrast to perceptual judgments, were unaffected by the illusions (Aglioti, DeSouza, & Goodale, 1995; Brenner & Smeets, 1996; Daprati & Gentilucci, 1997; but see Franz, Gegenfurtner, Bulthoff, & Fahle, 2000). However, this was observed only when responses were performed immediately after stimulus onset (Bridgeman, Peery, & Anand, 1997; Creem & Profftt, 1998; Hu & Goodale, 2000). These results support the view that facilitation of processing can be observed for fast motor responses, but not for slower responses. However, we observed the same pattern of data when observers responded without (Experiment 1) and with time pressure (Experiment 2). Reaction time for saccadic responses was shorter than for manual and verbal responses, however, it did not lead to better performance. Hence, the lack of a difference in sensitivity for different response modalities cannot be explained by a decay of information before the response is executed. S-R links may change over time due to perceptual learning (Hernandez & Lefton, 1977; Hogben & Di Lollo, 1984; Schwiedrzik et al., 2009). These effects seem mainly present in the case of localization tasks, while no improvement in target discrimination was observed when observers were asked to withhold their verbal responses for 600 ms (Vorberg, Mattler, Heinecke, Schmidt, & Schwarzbach, 2003). These results indicate that perceptual learning might be limited to direct stimulus-response (S-R) links specific for certain response modalities. However, in our experiments perceptual learning was observed for all response modalities. Thus, our results do not support the view that specific S-R links may improve over time only for certain fast responses. Nevertheless, S-R links could still have been developed but they are comparable for all modalities used in our study. 64 I. Szumska et al. / Consciousness and Cognition 43 (2016) 57–65 Our null results are unlikely to be affected by the low number of participants because, first, the presence of any effect of response modality is not visible in the data. Second, a Bayesian analysis (see Masson, 2011) confirmed that the ‘‘null” hypothesis is a much more likely explanation than the H1 hypothesis. Third, even though there is no effect of response modality, we found learning in both experiments. Thus, the number of participants we used is sufficient to produce significant effects in principle. In conclusion, our findings suggest that response accuracy in visual binary decision tasks does not depend on response modality. Null results are not easy to interpret. One possibility is that indeed in our experiments stimulus processing is independent of the response modality. Another possibility is that there are differences in accuracy but our paradigm was not sensitive enough to reveal them. Another possibility is that there are differences in general but accuracy is identical for the specific settings in our experiment. As mentioned, we do not deny that motor-contingencies exist. They do definitely for reflexes, in all sorts of sports, and wherever fast reactions are required. However, our results suggest that there are no differences in accuracy when there are no specific (trained) stimulus-response mappings. Acknowledgments We thank Marc Repnow for technical support, insightful discussion and useful comments on the manuscript. This work was financially supported by the Scientific Exchange Programme NMS-CH grant (SCIEX). Lukasz Grzeczkowski was funded by the project ‘‘Learning from Delayed and Sparse Feedback” (Project number: CRSII2_147636) of the Swiss National Science Foundation (SNFS). Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.concog. 2016.05.005. References Albrecht, T., Klapötke, S., & Mattler, U. (2010). Individual differences in metacontrast masking are enhanced by perceptual learning. Consciousness and Cognition, 19(2), 656–666. http://dx.doi.org/10.1016/j.concog.2009.12.002. Aglioti, S., DeSouza, J. F., & Goodale, M. A. (1995). Size-contrast illusions deceive the eye but not the hand. Current Biology, 5(6), 679–685. Ansorge, U., Klotz, W., & Neumann, O. (1998). Manual and verbal responses to completely masked (unreportable) stimuli: Exploring some conditions for the metacontrast dissociation. Perception, 27(10), 1177–1189. Bach, M. (1996). The Freiburg Visual Acuity test-automatic measurement of visual acuity. Optometry and Vision Science, 73(1), 49–53. Bekkering, H., & Neggers, S. (2002). Visual search is modulated by action intentions. Psychological Science, 13(4), 370–374. http://dx.doi.org/10.1111/j.09567976.2002.00466.x. Breitmeyer, B. G. (2014). Contributions of magno- and parvocellular channels to conscious and non-conscious vision. Philosophical Transactions of the Royal Society of London. Biological Sciences, 369(1641), 20130213. http://dx.doi.org/10.1098/rstb.2013.0213. Brenner, E., & Smeets, J. B. (1996). Size illusion influences how we lift but not how we grasp an object. Experimental Brain Research, 111(3), 473–476. Bridgeman, B., Peery, S., & Anand, S. (1997). Interaction of cognitive and sensorimotor maps of visual space. Perception & Psychophysics, 59(3), 456–469. http://dx.doi.org/10.3758/BF03211912. Creem, S. H., & Profftt, D. R. (1998). Two memories for geographical slant: Separation and interdependence of action and awareness, 5(1), 22–36. Daprati, E., & Gentilucci, M. (1997). Grasping an illusion. Neuropsychologia, 35(12), 1577–1582. http://dx.doi.org/10.1016/S0028-3932(97)00061-4. Donders, F. C. (1969). On the speed of mental processes. Acta Psychologica, 30, 412–431. Eimer, M., & Schlaghecken, F. (2001). Response facilitation and inhibition in manual, vocal, and oculomotor performance: Evidence for a modality-unspecific mechanism. Journal of Motor Behavior, 33(1), 16–26. http://dx.doi.org/10.1080/00222890109601899. Forster, K., & Davis, C. (1991). The density constraint on form-priming in the naming task: Interference effects from a masked prime. Journal of Memory and Language, 30(1), 1–25. http://dx.doi.org/10.1016/0749-596X(91)90008-8. Francis, G., & Herzog, M. H. (2004). Testing quantitative models of backward masking. Psychonomic Bulletin & Review, 11, 104–112. Franz, V. H., Gegenfurtner, K. R., Bulthoff, H. H., & Fahle, M. (2000). Grasping visual illusions: no evidence for a dissociation between perception and action. Psychological Science, 11(1), 20–25. http://dx.doi.org/10.1111/1467-9280.00209. Gibson, J. J. (1966). The senses considered as perceptual systems. Boston: Houghton Mifflin. Goodale, M. A. (2014). How (and why) the visual control of action differs from visual perception. Proceedings of the Royal Society, 281, 1–9. http://dx.doi.org/ 10.1098/rspb.2014.0337. Goodale, M. A., & Milner, D. A. (1992). Separate visual pathways for perception and action. Trends in Neurosciences, 15(1), 20–25. http://dx.doi.org/10.1016/ 0166-2236(92)90344-8. Greenwald, a. G., Draine, S. C., & Abrams, R. L. (1996). Three cognitive markers of unconscious semantic activation. Science (New York, N.Y.), 273(September), 1699–1702. http://dx.doi.org/10.1126/science.273.5282.1699. Hernandez, L., & Lefton, L. (1977). Metacontrast as measured under a signal detection model. Perception, 6(6), 695–702. Hogben, J. H., & Di Lollo, V. (1984). Practice reduces suppression in metacontrast and in apparent motion. Perception & Psychophysics, 35(5), 441–445. http:// dx.doi.org/10.3758/BF03203920. Hommel, B., & Müsseler, J. (2001). The Theory of Event Coding (TEC): A framework for perception and action planning. Hu, M., & Goodale, M. (2000). Grasping after a delay shifts size-scaling from absolute to relative metrics. Journal of Cognitive Neuroscience, 12(5), 856–868. Hubel, D. H., & Wiesel, T. N. (1959). Receptive fields of single neurones in the cat’s striate cortex. Journal of Physiology, 148, 574–591. Hughes, H. C., & Kelsey, J. V. (1984). Response-dependent effects on near-threshold detection performance: Saccades versus manual responses. Perception & Psychophysics, 35(6), 543–546. Jáskowski, P., Skalska, B., & Verleger, R. (2003). How the self controls its ‘‘automatic pilot” when processing subliminal information. Journal of Cognitive Neuroscience, 15(6), 911–920. http://dx.doi.org/10.1162/089892903322370825. Jaśkowski, P., & Sobieralska, K. (2004). Effect of stimulus intensity on manual and saccadic reaction time. Perception & Psychophysics, 66(4), 535–544. Jaśkowski, P., van der Lubbe, R. H. J., Schlotterbeck, E., & Verleger, R. (2002). Traces left on visual selective attention by stimuli that are not consciously identified. Psychological Science, 13(1), 48–54. http://dx.doi.org/10.1111/1467-9280.00408. Kingdom, F. A., & Prins, N. (2010). Psychophysics. A practical introduction. London: Academic Press: An Imprint of Elsevier. I. Szumska et al. / Consciousness and Cognition 43 (2016) 57–65 65 Klotz, W., Heumann, M., Ansorge, U., & Neumann, O. (2007). Electrophysiological activation by masked primes: Independence of prime-related and targetrelated activities. Advances in Cognitive Psychology, 3(4), 449–465. http://dx.doi.org/10.2478/v10053-008-0008-1. Klotz, W., & Neumann, O. (1999). Motor activation without conscious discrimination in metacontrast masking. Journal of Experimental Psychology: Human Perception and Performance. http://dx.doi.org/10.1037/0096-1523.25.4.976. Kolers, P. A. (1962). Intensity and contour effects in visual masking. Vision Research, 2, 277–294. Lachter, J., Durgin, F., & Washington, T. (2000). Disappearing percepts: Evidence for retention failure in metacontrast masking. Visual Cognition, 7, 269–279. http://dx.doi.org/10.1080/135062800394801. Livingstone, M., & Hubel, D. (1988). Segregation of form, color, movement, and depth: Anatomy, physiology, and perception. Science, 240, 740–749. http:// dx.doi.org/10.1126/science.3283936. Maksimov, M., Murd, C., & Bachmann, T. (2011). Target-mask shape congruence impacts the type of metacontrast masking. Scandinavian Journal of Psychology, 52, 524–529. http://dx.doi.org/10.1111/j.1467-9450.2011.00904.x. Marcel, A. J. (1983). Conscious and unconscious perception: An approach to the relations between phenomenal experience and perceptual processes. Cognitive Psychology, 15(2), 238–300. Marr, D. (1982). Vision: A computational approach. San Francisco: Freeman & Co. Masson, M. E. J. (2011). A tutorial on a practical Bayesian alternative to null-hypothesis significance testing. Behavior Research Methods, 43(3), 679–690. http://dx.doi.org/10.3758/s13428-010-0049-5. Milner, D., & Goodale, M. (1995). The visual brain in action. Oxford: Oxford University Press. Nagel, S. K., Carl, C., Kringe, T., Märtin, R., & König, P. (2005). Beyond sensory substitution-learning the sixth sense. Journal of Neural Engineering, 2(4), R13–R26. http://dx.doi.org/10.1088/1741-2560/2/4/R02. Neggers, S. F. W., Schölvinck, M. L., van der Lubbe, R. H. J., & Postma, A. (2005). Quantifying the interactions between allo- and egocentric representations of space. Acta Psychologica, 118(1–2), 25–45. http://dx.doi.org/10.1016/j.actpsy.2004.10.002. O’Regan, J. K. O., & Noë, A. (2001). A sensorimotor account of vision and visual consciousness. Behavioral and Brain Sciences, 24, 939–1031. Perenin, M., & Jeannerod, M. (1975). Residual vision in cortically blind hemiphelds. Neuropsychologia, 13, 1–7. Poggio, T. (1984). Vision by man and machine. Scientific American, 250(4), 105–116. Poppel, E., Held, R., & Frost, D. (1973). Residual function after brain wounds involving the central visual pathways in man. Nature, 243, 295–296. Raftery, A. E. (1995). Bayesian model selection in social research. Sociological Methodology. Cambridge: Blackwell. Schenk, T., & McIntosh, R. D. (2010). Do we have independent visual streams for perception and action? Cognitive Neuroscience, 1(1), 52–62. http://dx.doi. org/10.1080/17588920903388950. Schwiedrzik, C. M., Singer, W., & Melloni, L. (2009). Sensitivity and perceptual awareness increase with practice in metacontrast masking. Journal of Vision, 9 (18), 1–18. http://dx.doi.org/10.1167/9.10.18. Sternberg, S. (1969). The discovery of processing stages: Extensions of Donders’ method. Acta Psychologica, 30, 276–315. Ungerleider, L. G., & Mishkin, M. (1982). Two cortical visual systems. Analysis of visual behavior. Cambridge, MA: The MIT Press. Vorberg, D., Mattler, U., Heinecke, A., Schmidt, T., & Schwarzbach, J. (2003). Different time courses for visual perception and action priming. Proceedings of the National Academy of Sciences of the United States of America, 100(10), 6275–6280. http://dx.doi.org/10.1073/pnas.0931489100. Wagenmakers, E.-J. (2007). A practical solution to the pervasive problems of p values. Psychonomic Bulletin & Review, 14(5), 779–804. http://dx.doi.org/ 10.3758/BF03194105. Wichmann, F. A., & Hill, N. J. (2001). The psychometric function: I. Fitting, sampling, and goodness of fit. Perception and Psychophysics, 63(8), 1293–1313.

Consciousness and Cognition CONTENTS Volume 14, Number 1, March 2005 Special Issue: The Neurobiology of Animal Consciousness 1 Editorial Wiliam P. Banks, Bernard J. Baars, Antti Revonsuo 3 Dedication IN MEMORIAM 4 In memoriam: Donald R. Griffin INTRODUCTION 7 Subjective experience is probably not limited to humans: The evidence from neurobiology and behavior Bernard J. Baars GUEST EDITORIAL 22 Toward a science of ultimate concern Jaak Panksepp SPECIAL ISSUE ARTICLES 30 Affective consciousness: Core emotional feelings in animals and humans Jaak Panksepp 81 Commentary: Panksepp’s common sense view of affective neuroscience is not the commonsense view in large areas of neuroscience Douglas F. Watt 89 The liabilities of mobility: A selection pressure for the transition to consciousness in animal evolution Björn Merker 115 Commentary: Evolutionary pressures and a stable world for animals and robots: A commentary on Merker Stan Franklin 119 Criteria for consciousness in humans and other mammals Anil K. Seth, Bernard J. Baars, and David B. Edelman 140 Neural Darwinism and consciousness Anil K. Seth and Bernard J. Baars 169 Identifying hallmarks of consciousness in non-mammalian species David B. Edelman, Bernard J. Baars, and Anil K. Seth Continued 188 The threat simulation theory of the evolutionary function of dreaming: Evidence from dreams of traumatized children Katja Valli, Antti Revonsuo, Outi P€alk€as, Kamaran Hassan Ismail, Karzan Jalal Ali, and Raija-Leena Punam€aki 219 Consciousness eclipsed: Jacques Loeb, Ivan P. Pavlov, and the rise of reductionistic biology after 1900 Ralph J. Greenspan and Bernard J. Baars 231 PAPERS TO APPEAR IN FORTHCOMING ISSUES

Consciousness and Cognition CONTENTS Volume 14, Number 2, June 2005 233 The imagination: Cognitive, pre-cognitive, and meta-cognitive aspects Kieron P. OÕConnor, Frederick Aardema 257 Unconscious word-stem completion priming in a mirror-masking paradigm Walter J. Perrig, Doris Eckstein 278 Automatic and controlled semantic processing: A masked prime-task effect B. Valds, A. Catena, P. Mar-Beffa 296 Hypnotic susceptibility, baseline attentional functioning, and the Stroop task Sandro Rubichi, Federico Ricci, Roberto Padovani, Lorenzo Scaglietti 304 Does ÔhypnosisÕ by any other name smell as sweet? The efficacy of ÔhypnoticÕ inductions depends on the label ÔhypnosisÕ Balaganesh Gandhi, David A. Oakley 316 Activating the critical lure during study is unnecessary for false recognition Ren Zeelenberg, Inge Boot, Diane Pecher 327 Event-related potential evidence for multiple causes of the revelation effect P. Andrew Leynes, Joshua Landau, Jessica Walker, Richard J. Addante 351 Accuracy of familiarity decisions to famous faces perceived without awareness depends on attitude to the target person and on response latency Anna Stone, Tim Valentine 377 Procedural memory in dissociative identity disorder: When can inter-identity amnesia be truly established? Rafa0le J.C. Huntjens, Albert Postma, Liesbeth Woertman, Onno van der Hart, Madelon L. Peters 390 Common fronto-parietal activity in attention, memory, and consciousness: Shared demands on integration? Hamid Reza Naghavi, Lars Nyberg 426 PAPERS TO APPEAR IN FORTHCOMING ISSUES

Consciousness and Cognition 19 (2010) 1154–1155 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Reply Methodological and practical issues regarding phenomenological subtypes of highly suggestible individuals: A response to Kumar q Devin Blair Terhune ⇑, Etzel Cardeña Department of Psychology, Lund University, 22100 Lund, Sweden a r t i c l e i n f o Article history: Available online 16 June 2010 Keywords: Hypnosis Hypnotic suggestibility Phenomenology Heterogeneity Subtypes Hypnotherapy a b s t r a c t In his commentary on our article on phenomenological subtypes of highly suggestible individuals (Terhune & Cardeña, 2010), Kumar (2010) argues that methodological differences between our studies and previous research on highly suggestible subtypes temper our ability to link the two and that it is unclear whether the existence of phenomenological subtypes has implications for hypnotherapy. Although we agree that it is premature to make conclusive statements about highly suggestible subtypes, we argue that convergent findings across studies with different methodologies are especially salient because they are unlikely to result from a common set of methodological artifacts. In addition, we describe a number of implications that research on phenomenological subtypes has for hypnotherapy. Ó 2010 Elsevier Inc. All rights reserved. Although there have been speculations about the possibility of highly suggestible (HS) subtypes for quite some time (White, 1937), this research area has remained relatively dormant. In a recent article, we presented evidence for two distinct phenomenological subtypes of HS individuals on the basis of a latent profile analysis of participants’ spontaneous experiential response patterns following a hypnotic induction (Terhune & Cardeña, 2010). Insofar as the profile of one subtype was characterized by greater spontaneous, endogenously-directed attention and vivid imagery, whereas the other was characterized by greater alterations in awareness and volition, we interpreted the findings as supportive of dissociative typological models of high hypnotic suggestibility (Barber, 1999; Barrett, 1996; Carlson & Putnam, 1989). In his commentary on our article, Kumar (2010) highlights two important issues: (1) whether our evidence for putative subtypes corresponds to that provided using other methodologies and (2) that it is unclear whether the existence of discrete phenomenological subtypes has clinical applications. In what follows, we briefly address these issues. We wholeheartedly agree with Kumar that caution should be exerted when contrasting the phenomenological subtypes identified in our paper and the various papers of Pekala and colleagues (see Pekala & Kumar, 2007) with those proposed by Barber and others (Barber, 1999; Barrett, 1996; Carlson & Putnam, 1989) because of the different methodologies across studies. Heterogeneity among HS individuals is a sorely neglected topic and future research needs to focus on integrating research in this area. However, we believe that methodological diversity can also be a strength because convergent evidence for two distinct subtypes from studies with different methodologies strongly suggests that the derivation of distinct subtypes is not the result of an artifact that is unique to one or another methodological approach. For instance, we find it salient that in addition to our study and those of Pekala and colleagues, other authors have presented evidence for non- q Reply to Commentaries on Kumar, V. K. (2010). Reflections on the varieties of hypnotizables: A commentary on Terhune and Cardeña. Consciousness and Cognition, 19, 1151–1153. ⇑ Corresponding author. E-mail address: devin.terhune@psychology.lu.se (D.B. Terhune). 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.05.010 D.B. Terhune, E. Cardeña / Consciousness and Cognition 19 (2010) 1154–1155 1155 overlapping HS subtypes with strong imagery (Kunzendorf & Boisvert, 1996; Wallace, 1990) and elevated dissociative tendencies (King & Council, 1998) that differ on measures of imagery and hypnotic responding. In light of the foregoing, we think it is especially critical to examine whether the HS subtypes are present at multiple explanatory levels. We have made preliminary progress towards this end. In one study, we found that low dissociative HS individuals displayed superior imagery, as measured by a behavioral task involving the identification of degraded static images, whereas high dissociative HS individuals experienced greater involuntariness during hypnotic responding (Terhune, Cardeña, & Lindgren, 2010); these results conceptually replicate the characteristic features of the two phenomenological subtypes (Terhune & Cardeña, 2010). In another behavioral study with the same sample, low dissociative HS individuals displayed marginally improved cognitive control following a hypnotic induction, whereas high dissociative HS individuals exhibited impaired cognitive control (Terhune, Cardeña, & Lindgren, 2009). Although these studies are conceptually consistent with other studies suggesting the existence of discrete HS subtypes (e.g., King & Council, 1998), further research is clearly needed to corroborate the existence of these subtypes and understand their phenomenological and cognitive characteristics. Kumar’s (2010) second concern is the extent to which our findings and those of others concerning putative HS subtypes are applicable in the clinical setting, particularly for individuals of low or medium suggestibility. We maintain that results from empirical hypnosis research need not always benefit the clinical use of hypnosis - the study of hypnosis is a worthwhile pursuit irrespective of its clinical utility. Nevertheless, we believe that the present results could be developed so as to have implications for hypnotherapy. This area of research helps us clarify how dissociation and hypnosis may be related within the hypnotic context, a relationship that is by no means simple (Cardeña, in press). Furthermore, it is important to note that of the four profiles derived in our analysis, the two phenomenological profiles upon which our analyses were focused included individuals in the low to medium range of hypnotic suggestibility. Low and medium suggestible individuals in these profiles displayed greater experiential response to the hypnotic induction than the individuals of comparable hypnotic suggestibility in the other two profiles. We might speculate that the former group would be more responsive to techniques for increasing their hypnotic suggestibility and thus be more suitable for hypnotherapy. Low and medium suggestible individuals in the dissociative profile experienced greater spontaneous negative affect during hypnosis than those in the other profiles. Identifying these individuals by indexing spontaneous phenomenological response to hypnosis may help therapists to tailor their suggestions more precisely and to improve their clients’ experience of hypnosis and their ability to participate in hypnotherapy. References Barber, T. X. (1999). A comprehensive three-dimensional theory of hypnosis. In I. Kirsch, A. Capafons, E. Cardeña-Buelna, & S. Amigo (Eds.), Clinical hypnosis and self-regulation: Cognitive-behavioral perspectives (pp. 21–48). Washington, DC: American Psychological Association. Barrett, D. (1996). Fantasizers and dissociators: Two types of high hypnotizables, two different imagery styles. In R. G. Kunzendorf, N. P. Spanos, & B. Wallace (Eds.), Hypnosis and imagination (pp. 123–146). Amityville, NY: Baywood. Cardeña, E. (in press). Does hypnotic responsiveness come in more than one flavor and, if so, why should clinicians care? ISSTD Newsletter. Carlson, E. B., & Putnam, F. W. (1989). Integrating research on dissociation and hypnotizability: Are there two pathways to hypnotizability? Dissociation, 2, 32–38. King, B. J., & Council, J. R. (1998). Intentionality during hypnosis: An ironic process analysis. International Journal of Clinical and Experimental Hypnosis, 46, 295–313. Kumar, V. K. (2010). Reflections on the varieties of hypnotizables: A commentary on Terhune and Cardeña. Consciousness and Cognition, 19, 1151–1153. Kunzendorf, R. G., & Boisvert, P. (1996). Presence vs. absence of a ‘hidden observer’ during total deafness: the hypnotic illusion of subconsciousness vs. the imaginal attenuation of brainstem evoked potentials. In R. G. Kunzendorf, N. P. Spanos, & B. Wallace (Eds.), Hypnosis and imagination (pp. 223–234). Amityville, NY: Baywood. Pekala, R. J., & Kumar, V. K. (2007). An empirical–phenomenological approach to quantifying consciousness and states of consciousness: With particular reference to understanding the nature of hypnosis. In G. A. Jamieson (Ed.), Hypnosis and conscious states: The cognitive neuroscience perspective (pp. 167–194). Oxford: Oxford University Press. Terhune, D. B., & Cardeña, E. (2010). Differential patterns of spontaneous experiential response to a hypnotic induction: A latent profile analysis. Consciousness and Cognition, 19, 1140–1150. Terhune, D. B., Cardeña, E., & Lindgren, M. (2009). Typological variability in executive attention in high hypnotic suggestibility. Paper presented at the 20th annual convention of the Association for Psychological Science. Terhune, D. B., Cardeña, E., & Lindgren, M. (2010). Dissociative tendencies and individual differences in high hypnotic suggestibility. Unpublished manuscript. Wallace, B. (1990). Imagery vividness, hypnotic-susceptibility, and the perception of fragmented stimuli. Journal of Personality and Social Psychology, 58, 354–359. White, R. W. (1937). Two types of hypnotic trance and their personality correlates. Journal of Psychology, 3, 279–289.

Consciousness and Cognition 19 (2010) 1140–1150 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Differential patterns of spontaneous experiential response to a hypnotic induction: A latent profile analysis Devin Blair Terhune *, Etzel Cardeña Department of Psychology, Lund University, 22100 Lund, Sweden a r t i c l e i n f o Article history: Received 17 August 2009 Available online 14 April 2010 Keywords: Hypnosis Hypnotic suggestibility Dissociation Fantasy-proneness Involuntariness Phenomenology Typology Latent profile analysis a b s t r a c t A hypnotic induction produces different patterns of spontaneous experiences across individuals. The magnitude and characteristics of these responses covary moderately with hypnotic suggestibility, but also differ within levels of hypnotic suggestibility. This study sought to identify discrete phenomenological profiles in response to a hypnotic induction and assess whether experiential variability among highly suggestible individuals matches the phenomenological profiles predicted by dissociative typological models of high hypnotic suggestibility. Phenomenological state scores indexed in reference to a resting epoch during hypnosis were submitted to a latent profile analysis. The profiles in the derived four-class solution differed in multiple experiential dimensions and hypnotic suggestibility. Highly suggestible individuals were distributed across two classes that exhibited response patterns suggesting an inward attention subtype and a dissociative subtype. These results provide support for dissociative typological models of high hypnotic suggestibility and indicate that highly suggestible individuals do not display a uniform response to a hypnotic induction. Ó 2010 Elsevier Inc. All rights reserved. 1. Introduction Although considerable attention has been devoted to the striking distortions in agency induced by suggestions administered during hypnosis, a hypnotic induction alone is capable of producing profound alterations in a variety of dimensions of consciousness. A hypnotic induction consists of a set of instructions and suggestions to help a participant become absorbed in the experimenter’s words and reduce their awareness of exogenous stimuli (e.g., Oakley & Halligan, 2009). Unusual spontaneous experiences following a hypnotic induction, omitting particular suggestions, are commonplace but remain understudied (Cardeña, 2005; Pekala & Kumar, 2007; Rainville & Price, 2003). Many individuals, in particular those of high hypnotic suggestibility, frequently report various types of alterations in core phenomenological dimensions of consciousness. Such experiences include vestibular perceptions of floating, marked changes in temporal perception, affect, and internal dialogue, and increased amounts of fantasy-based visual imagery. Some of the variance in these dimensions is attributable to participants’ response expectancies (Henry, 1985; as cited in Kirsch, 1990; Pekala, Kumar, & Hand, 1993). However, alterations in these experiential dimensions are still reported among highly suggestible (HS) individuals when a neutral hypnotic induction, which excludes experience-specific suggestions (e.g., for relaxation), is used (Cardeña, 2005). A consistently replicated finding is that variability in spontaneous experiences during hypnosis covaries with hypnotic suggestibility (Pekala & Kumar, 2007). For instance, HS individuals reliably report greater magnitude alterations in a variety of experiential dimensions than their medium and low suggestible counterparts (Kumar & Pekala, 1988; 1989). However, some studies have observed marked differences in this population (e.g., Barrett, 1996; Pekala & Kumar, 2007). For instance, * Corresponding author. E-mail address: devin.terhune@psychology.lu.se (D.B. Terhune). 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.03.006 D.B. Terhune, E. Cardeña / Consciousness and Cognition 19 (2010) 1140–1150 1141 Barrett (1996) presented evidence for two subtypes of HS individuals, one of which experienced greater alterations in awareness and increased involuntariness during hypnotic responding. HS individuals also exhibit considerable variability in the types of suggestions to which they respond and the strategies they utilize to facilitate responding (e.g., McConkey & Barnier, 2004). In order to resolve outstanding questions regarding heterogeneity in this population, various models have proposed that HS individuals are comprised of distinct subtypes of respondents (e.g., Barber, 1999a; Brown & Oakley, 2004; Kunzendorf & Boisvert, 1996). These subtypes are hypothesized to experience hypnosis through different mechanisms and concomitantly exhibit dissimilar experiential response patterns following a hypnotic induction. Dissociative typological models of high hypnotic suggestibility propose that HS individuals are comprised of dissociative and fantasy-prone respondents (Barber, 1999a; Barrett, 1996; Cardeña, 1996; Carlson & Putnam, 1989; Perry, 2004); Barber (1999a) has also proposed a third subtype: positively-set respondents. According to these models, a hypnotic induction produces a state of experiential detachment in dissociative respondents that is characterized by reduced awareness, attention, episodic memory, imagery, and volitional control relative to other HS individuals. In contrast, fantasy-prone respondents are hypothesized to exhibit alterations in awareness during hypnosis of lower magnitude than dissociative respondents, but to experience greater attentional involvement (absorption), imagery, episodic memory, and volitional control. Positively-set respondents are hypothesized to exhibit relatively minor spontaneous alterations in experiential dimensions of consciousness that do not differ substantially from individuals of low hypnotic suggestibility. (For critical reviews of these models, see Barber (1999b) and accompanying commentaries.) Support for the experiential predictions of the typological models has been provided by cluster analysis studies (Forbes & Pekala, 1996; Pekala, 1991b; Pekala & Forbes, 1997; Pekala, Kumar, & Marcano, 1995; for a review see Pekala & Kumar, 2007). In these studies, participants experienced a short resting epoch embedded within a standardized behavioral measure of hypnotic suggestibility. Participants subsequently completed the Phenomenology of Consciousness Inventory (PCI; Pekala, 1991a) in reference to their spontaneous experiences during the resting epoch. The PCI taps a wide variety of experiential dimensions including body image, temporal perception, positive affect, attentional absorption, and visual imagery. In four studies, Pekala and colleagues used K-means cluster analyses to derive discrete types of respondents at multiple levels of hypnotic suggestibility on the basis of PCI dimension scores (Pekala & Kumar, 2007). In the first study, Pekala (1991b) derived two types of HS participants, labelled fantasy and classic types, both of which were subsequently replicated by Pekala and Forbes (1997). The principal features of the fantasy type’s experiential response were vivid imagery, positive affect, and mild reductions in awareness and memory, whereas the classic type experienced less vivid imagery, reduced control and memory, and greater alterations in awareness. In another study, Pekala et al. (1995) derived two types of HS participants, one that corresponded to the classic type and another labelled compliant, which was similar to the fantasy type except that it exhibited less imagery and positive affect and more internal dialogue. A final study replicated the classic type and found a second type interpreted as a hybrid of the fantasy and compliant types (Forbes & Pekala, 1996) and, in a separate seven-cluster solution, replicated the fantasy and classic types and observed a small percentage of HS participants classified in another cluster who exhibited minor alterations in the measured experiential dimensions. These studies have been criticized for a lack of consistently derived cluster solutions (Lynn Meyer, & Schindler, 2004), but, collectively, provide evidence for distinct patterns of phenomenological response to a hypnotic induction among HS individuals. Further, they suggest that such patterns may be grounded in a latent typology. The classic type was consistently replicated, whereas the characteristics of a second (and possibly third) type are equivocal. Notwithstanding this issue, there are clear parallels between the phenomenological response of the different clusters and the experiential profiles predicted by the dissociative typological models (e.g., Barber, 1999a). The results, however, appear to provide greater support for bifurcated (Barrett, 1996; Brown & Oakley, 2004; Carlson & Putnam, 1989; Kunzendorf & Boisvert, 1996) than trifurcated (Barber, 1999a) typological models. Lack of consistency is neither the only nor most salient limitation of these studies. Although some of the analyses were undertaken on the entire sample, many of the derived cluster solutions were generated by cluster analyses performed on relatively small sample sizes of HS participants (ns < 100). The analyses could also have been strengthened by a validation check of the different types using an independent measure of theoretical significance. Furthermore, the hypnotic suggestibility of the derived types was not contrasted in order to identify their behavioral correlates. Barber (1999a), for instance, proposes that the dissociative subtype is more responsive to posthypnotic amnesia suggestions. A final limitation of these analyses is the use of K-means cluster analysis. Despite its pervasive use, there exists no consensus regarding analytic techniques for class enumeration, that is, the determination of an optimal number of clusters, in a sample using this method (Ruscio & Ruscio, 2008; Vermunt & Magidson, 2002). It follows that the reliability and validity of the derived cluster solutions in these studies may be suspect. Many of the limitations of K-means cluster analysis are circumvented by latent variable modelling techniques such as latent profile analysis (LPA; Goodman, 2002; see also McCutcheon, 1987, 2002). LPA is a method for identifying homogeneous profiles in multivariate continuous data. The central assumption of LPA is that variability in a set of continuous indicator (observed) variables stems from a set of patterns determined by an underlying categorical latent (unobserved) variable comprised of multiple profiles (Vermunt & Magidson, 2002). The principal strength of LPA is that it allows for the computation of model fit statistics that render the process of class enumeration less arbitrary than K-means cluster analysis. In addition, LPA enables the testing of more complex models, such as ones that free restrictions on indicator covariance (Vermunt & Magidson, 2002). In multiple comparative assessments, LPA consistently exhibited superior performance than Kmeans cluster analysis (Magidson & Vermunt, 2002). 1142 D.B. Terhune, E. Cardeña / Consciousness and Cognition 19 (2010) 1140–1150 2. The present study There has been relatively little research on spontaneous phenomenological alterations during hypnosis and their underlying mechanisms (Rainville & Price, 2003). Spontaneous alterations in experiential dimensions of consciousness may reflect mind-wandering and a consequent weakening of executive functioning (Smallwood, Beach, Schooler, & Handy, 2008). Impaired executive functioning during hypnosis has been argued to modulate hypnotic suggestibility and play a critical role in mediating responsiveness to hypnotic suggestions (Egner, Jamieson, & Gruzelier, 2005; Woody & Bowers, 1994; Woody & Sadler, 2008). Accordingly, the examination of individual differences in spontaneous experiential response profiles among HS individuals and whether they exhibit a typological pattern represents a critically important endeavour for understanding the nature of hypnosis and hypnotic suggestibility. This study used LPA to identify the optimal number and principal characteristics of different experiential response profiles following a hypnotic induction. Participants were administered a standardized group measure of hypnotic suggestibility within which was embedded a resting epoch. Following a de-induction, participants retrospectively completed the PCI (Pekala, 1991a) and the Inventory Scale of Hypnotic Depth (ISHD; Field, 1965). The ISHD is a measure of experiential involvement and involuntariness during hypnotic responding and was included to independently validate the dissociative typology because it has been argued to discriminate dissociative and fantasy-prone HS individuals (Barrett, 1996). In addition to predicting that LPA would discern a poly-class solution of experiential profiles on the basis of PCI factor-based scores, we expected HS individuals to fall into two or three phenomenological classes that would exhibit dissimilar experiential profiles, suggesting a typological distribution. Finally, we tested the prediction that the experiential response patterns of the derived profiles would correspond to those predicted by the bifurcated and trifurcated dissociative typological models (Barber, 1999a; Barrett, 1996; Cardeña, 1996; Carlson & Putnam, 1989; Perry, 2004). 3. Method 3.1. Participants Six hundred and forty individuals (375 females [59%]), whose ages ranged from 18 to 65 (M = 23.71, SD = 5.62), consented to participate in this study. Women (MAge = 23.55, SD = 5.56) and men (MAge = 23.93, SD = 5.71) did not differ in age, t < 1. Participants were recruited through advertisements at Lund University and in the city of Lund or volunteered as part of an introductory psychology course. This study was approved by the local ethics committee. 3.2. Materials 3.2.1. Hypnotic suggestibility The Waterloo-Stanford Group Scale of Hypnotic Susceptibility, Form C (WSGC; Bowers, 1993, 1998) was used to measure responsiveness to hypnotic suggestions. The WSGC is a group adaptation of the individually-administered Stanford Hypnotic Susceptibility Scale, Form C (Weitzenhoffer & Hilgard, 1962) and consists of 12 dichotomously-scored items including direct ideomotor (e.g., arm heaviness), challenge motor (e.g., arm immobilization), and cognitive-perceptual (e.g., auditory hallucination) suggestions, with scores ranging from 0 to 12. This measure has strong psychometric properties (Bowers, 1993, 1998). 3.2.2. Experiential dimensions of consciousness The PCI (Pekala, 1991a) is a 53-item self-report scale measuring different dimensions of consciousness that is completed retrospectively in reference to a preceding interval. Each item consists of a pair of bipolar statements anchored on a sevenpoint Likert scale. The PCI consists of 12 dimensions (and 14 sub-dimensions): altered experience (body image, time sense, perception, and meaning); positive affect (joy, sexual excitement, and love); negative affect (anger, sadness, and fear); attention (direction and absorption); imagery (amount and vividness); self awareness; altered state of awareness; arousal; rationality; volitional control; memory; and internal dialogue. Kumar, Pekala, and Cummings (1996) derived five PCI factors: attention to internal processes, dissociated control, negative affect, positive affect, and visual imagery. 3.2.3. Experiential involvement and involuntariness The ISHD (Field, 1965) is a self-report scale composed of 38 dichotomous (true/false) items that measure alterations in awareness, perception, and volition during hypnosis. Representative items include: ‘‘At times I felt completely unaware of being in an experiment” and ‘‘Parts of my body moved without my conscious assistance.” The scale exhibited strong internal consistency in this sample (Cronbach’s a = .89). 3.3. Procedure Participants completed the WSGC in groups ranging in size from four to 40. A clinically-trained consultant was present during all sessions (see Cardeña & Terhune, 2009). A two-minute resting epoch was embedded within the WSGC prior to the 1143 D.B. Terhune, E. Cardeña / Consciousness and Cognition 19 (2010) 1140–1150 administration of items 11 and 12. Prior to the epoch, participants were instructed to sit quietly with their eyes closed and continue to experience hypnosis. Following the de-induction, participants completed the WSGC response booklet, the PCI in reference to the rest epoch, and the ISHD in reference to the whole session. 3.4. Statistical analyses The five PCI state factor-based scores (Kumar, Pekala, & Cummings, 1996) were used as the observable indicators for the derivation of the profiles using LPA. The fit of multiple models (two-class through five-class) was assessed. For each class solution, restricted and unrestricted models were evaluated. In the former, the covariance among indicators is restricted to zero, whereas in the latter it is allowed to be free. Restricted models commonly overestimate the number of profiles and provide less parsimonious solutions (Vermunt & Magidson, 2002). The selection of variables allowed to covary in the unrestricted models was made on the basis of the significance of the correlations among the indicator variables in Table 1. Statistical fit of the different models was evaluated using three information criterion indices: Akaike information criterion (AIC; Akaike, 1987), Bayesian information criterion (BIC; Schwartz, 1978), and the sample-size adjusted BIC (SSABIC; Sclove, 1987). In each case, lower values reflect superior model fit (Vermunt & Magidson, 2002). Two likelihood-ratio based tests were used: the Lo-Mendell-Rubin likelihood-ratio test (LMR-LRT; Lo, Mendell, & Rubin, 2001) and the Bootstrap likelihood-ratio test (BLRT; McLachlan & Peel, 2000). LMR-LRT and BLRT are used to adjudicate between nested models. For both, a non-significant value indicates that a model does not have superior fit than the corresponding model with one less class. The BLRT has consistently outperformed the LMR-LRT in comparative assessments (Nylund, Asparouhov, & Muthén, 2007) and was given preference in class enumeration. Entropy values were calculated on the basis of each model’s posterior probabilities for group membership and range from 0 to 1 with low values indicating poor classification of participants (Ramaswamy, Desarbo, Reibstein, & Robinson, 1993). The analyses were conducted with MPLUS v. 5.0 (Muthén & Muthén, 1998–2007) with secondary analyses performed with SPSS v. 16.0. Non-parametric tests were used for many of the secondary analyses due to violations of the assumptions of homogeneity of variance. Outliers (M ± 2 SDs) were excluded for contrasts among the different profiles. 4. Results 4.1. Intra-test reliability The PCI includes a set of items that allow for the computation of a reliability index (Pekala, 1991a). Twenty-five participants (4%) exhibited unacceptable values (>2); this compares favorably to a previous study (9%; Kumar, Pekala, & Cummings, 1996). These individuals’ data were excluded from the analyses, which thereafter included 615 participants. 4.2. Descriptive statistics Descriptive statistics and correlation coefficients for the research measures are presented in Table 1. All of the correlations were positive. WSGC scores were moderately correlated with ISHD scores and dissociated control, positive affect, and attention to internal processes, weakly correlated with visual imagery, and uncorrelated with negative affect. All other correlations were significant except for that between negative affect and visual imagery. 4.3. Phenomenological profiles All models exhibited high entropy values, indicating acceptable participant classification. Unrestricted models exhibited superior fit to the data for all class solutions, as reflected by lower information criteria values, than restricted models (see Table 2). The four-class unrestricted model had a comparable BIC value to the three-class unrestricted model and lower AIC and SSABIC values, indicating its superior fit. In addition, the former model had a significant BLRT value, indicating that Table 1 Descriptive statistics and correlation matrix for the research measures. Variable M (SD) 1. WSGC 2. ISHD 3. Dissociated control 4. Positive affect 5. Negative affect 6. Visual imagery 7. Attention to internal processes 4.41 (2.18) 15.26 (7.83) 8.43 (5.21) 4.24 (3.26) 1.30 (2.16) 3.32 (1.94) 6.47 (2.33) 1 2 3 4 5 6 7 .57** .53** .80** .42** .53** .63** .07 .15** .25** .21** .29** .30** .35** .38** .08 .49** .74** .77** .58** .10* .28** Note: WSGC = Waterloo-Stanford Group scale of Hypnotic Susceptibility, Form C; ISHD = Inventory Scale of Hypnotic Depth. p < .05. p < .001. * ** 1144 D.B. Terhune, E. Cardeña / Consciousness and Cognition 19 (2010) 1140–1150 Table 2 Evaluation indices and model comparison tests for the latent profile analysis of experiential dimensions during hypnosis. Model AIC BIC SSABIC LMR LRT p BLRT p Entropy 2-class R 2-class UR 3-class R 3-class UR 4-class R 4-class UR 5-class R 5-class UR 14,263 13,903 14,049 13,489 13,920 13,445 13,851 13,404 14,334 14,018 14,146 13,719 14,044 13,724 14,001 13,731 14,283 13,936 14,076 13,554 13,955 13,524 13,893 13,496 760.92 110.26 220.74 166.85 137.29 174.36 79.11 62.72 <.001 <.04 <.001 .03 .08 .11 .05 .12 780.67 113.12 226.47 169.21 140.85 176.83 81.16 – <.001 <.001 <.001 <.001 <.001 <.001 <.001 – .82 .82 .80 .81 .81 .77 .83 – Note. AIC = Akaike information criterion; BIC = Bayesian information criterion; SSABIC = sample-size adjusted BIC; LMR-LRT = Lo-Mendell-Rubin likelihoodratio test; BLRT = Bootstrap likelihood-ratio test; R = restricted; UR = unrestricted; BLRT values for the 5-class unrestricted model failed to replicate and are not provided; the optimal model is in bold. it is a better model than the latter. The unrestricted five-class model had superior AIC and SSABIC values than the four-class unrestricted model. However, its BIC was lower and its BLRT value was not consistently replicated, indicating its instability and the unreliability of its p-value. Moreover, the replicability of BLRT values declined with the inclusion of increased starting values. Because of these replicability failures and for the sake of parsimony, we selected the four-class unrestricted model as the optimal model. Participants were assigned to a profile on the basis of posterior probabilities. Table 3 presents descriptive statistics for the different profiles. Profile 2 was the largest class, whereas the rest exhibited comparable sample sizes. The profiles did not differ in age, F < 2.5, but there was a significant relationship between sex and profile, v2 (3, N = 615) = 25.20, p < .001. Profile 2 had a greater proportion of women than the other profiles, profiles 1 and 3 had comparable sex distributions, and profile 4 had the largest proportion of males. The profiles were also found to differ as a function of categorical hypnotic suggestibility level (low, medium, high), v2 (6, N = 615) = 87.27, p < .001. Profiles 1 and 2 were primarily comprised of participants in the medium range of hypnotic suggestibility, whereas profiles 3 and 4 were primarily comprised of those in the low range of hypnotic suggestibility. HS participants were distributed across profiles 1 and 2 with none in profile 3 and two in profile 4.1 To identify their characteristic features, we first contrasted the four profiles on the five PCI state scores. Kruskal–Wallis tests yielded main effects of Profile for all five PCI state scores: dissociated control, H(3) = 332.52, p < .001, positive affect, H(3) = 327.81, p < .001, negative affect, H(3) = 334.58, p < .001, visual imagery, H(3) = 67.38, p < .001, and attention to internal process, H(3) = 279.24, p < .001. Bonferroni-corrected post hoc Mann–Whitney tests indicated a clear demarcation between the first two and last two profiles, that is, profiles 1 and 2 differed from 3 and 4, on all five PCI state scores. Profile 1 was found to have lower negative affect and greater attention to internal processes than profile 2, whereas profile 4 exhibited greater dissociated control, positive affect and negative affect than profile 3. These findings indicate that profiles 1 and 2 represent participants who exhibited marked experiential responses to a hypnotic induction, whereas profiles 3 and 4 were comprised of participants who experienced relatively minimal and moderate responses, respectively. Next, we sought to further examine variability in PCI state scores between the first two profiles as a function of hypnotic suggestibility. We restricted this analysis to profiles 1 (inward attention) and 2 (dissociative), which were the only two profiles that included HS participants (see Table 4 for descriptive statistics). A 2 (Profile: inward attention v. dissociative)  3 (Hypnotic suggestibility: low, medium, high) multivariate analysis of variance on the five PCI state factors revealed main effects of Profile, F(5, 379) = 43.64, p < .001, g2 = .37, and Hypnotic suggestibility, F(10, 758) = 4.93, p < .001, g2 = .06, but no interaction, F < 2. In addition to the main effects of Profile on negative affect and attention to internal processes reported above, main effects of Hypnotic suggestibility were found for dissociated control, positive affect, visual imagery, and attention to internal processes, all Fs > 5, all ps < .007, g2 range: .03–.09. These effects were mediated by Profile  Hypnotic suggestibility interactions for dissociated control, visual imagery, and attention to internal processes, all Fs > 3, all ps < .05, all g2s = .02. In the inward attention profile, medium suggestible and HS participants exhibited greater dissociated control than low suggestible participants, but the former two did not differ from one another. HS participants in this profile also exhibited greater visual imagery than low suggestible participants, but did not differ from medium suggestible participants. No differences were found for attention to internal processes in this profile. In contrast, dissociated control and attention to internal processes increased in a significant linear fashion as a function of hypnotic suggestibility in the dissociative profile, whereas visual imagery increased from low to medium hypnotic suggestibility and did not differ between medium suggestible and HS participants. This indicates that variability in dissociated control, visual imagery, and attention to internal processes is differentially influenced by hypnotic suggestibility in the two profiles. 4.4. Assessment of the typological models As profile 1 exhibited greater internally-directed attention and lower negative affect than profile 2, and the two profiles included all of the HS participants, we next examined whether they exhibited experiential response patterns consonant with 1 In another study (Terhune, Cardeña, & Lindgren, 2010), the two HS participants in class 4 were both found to be false positives, that is, they failed to meet screening criteria for high hypnotic suggestibility as measured by individually-administered scales (Weitzenhoffer & Hilgard, 1967). 1145 D.B. Terhune, E. Cardeña / Consciousness and Cognition 19 (2010) 1140–1150 Table 3 Distributional data and descriptive statistics for PCI state factor scores in the four profiles: number or mean (percentage or standard deviation). Variable Profile Sex (female) Hypnotic suggestibility Low Medium High Dissociated control Positive affect Negative affect Visual imagery Attention to internal processes Inward attention 1 (n = 131) Dissociative 2 (n = 258) Minimal response 3 (n = 127) Moderate response 4 (n = 99) 70 (53%) 177 (69%) 65 (51%) 42 (42%) 29 (22%) 78 (60%) 24 (18%) 5.42 (3.07)a 5.44 (3.06)a 0.33 (0.29)a 3.88 (2.02)a 8.25 (1.23)a 68 (26%) 157 (61%) 33 (12%) 5.73 (4.58)a 5.99 (3.01)a 2.94 (2.32)b 3.80 (1.82)a 7.26 (2.05)b 78 (61%) 49 (39%) 0 (0%) 13.89 (2.91)b 0.66 (0.53)b 0.04 (0.70)c 2.43 (1.79)b 4.34 (1.68)c 55 (56%) 42 (42%) 2 (2%) 12.44 (2.25)c 2.67 (1.00)c 0.86 (1.15)d 2.47 (1.61)b 4.78 (1.68)c Note: Different superscripted letters indicate cell means significantly differ according to Mann-Whitney Tests after a Bonferroni correction (a = .002). Table 4 Descriptive statistics (Mean and Standard Deviation) for PCI state factor scores in the inward attention and dissociative profiles as a function of hypnotic suggestibility. PCI state factor Profile Inward attention Dissociated control Positive affect Negative affect Visual imagery Attention to internal processes Dissociative Low Medium High Low Medium High 7.08 (2.72)a 4.75 (2.73) 0.34 (0.24) 3.90 (2.00)a,b 7.79 (1.51)a 4.97 (3.12)b 5.60 (3.20) 0.33 (0.31) 3.58 (2.01)a 8.34 (1.10)a 4.86 (2.69)b 5.74 (2.99) 0.29 (0.31) 4.85 (1.86)b 8.48 (1.18)a 8.22 (4.75)a 4.79 (2.91) 3.07 (2.28) 3.17 (1.83)a 6.27 (2.03)a 5.29 (4.07)b 6.26 (2.95) 2.90 (2.29) 3.98 (1.77)b 7.43 (1.88)b 2.75 (4.20)c 7.17 (2.74) 2.86 (2.59) 4.23 (1.78)b 8.52 (2.02)c Note: Different superscripted letters indicate cell means in each profile significantly differ according to Tukey HSD tests. the fantasy-prone and dissociative types, respectively. We tested specific directional predictions of the dissociative typological models in HS participants using PCI dimensions and sub-dimensions (see Table 5). The dissociative profile was expected to exhibit greater distortions of awareness and reduced attention, imagery, memory, and volitional control than the inward attention profile. In line with these predictions, the dissociative profile exhibited greater scores on the altered experience dimension, F(1, 55) = 4.35, p < .05, g2 = .07, and lower scores on attention, F(1, 55) = 4.25, p < .05, g2 = .07, including direction of attention, F(1, 55) = 4.27, p < .05, g2 = .07, but not absorption, F < 1.5. The inward attention profile displayed greater imagery vividness, F(1, 54) = 4.88, p < .05, g2 = .08, but did not score higher on the general imagery dimension, nor amount of imagery, Fs < 1. The dissociative profile was also found to exhibit suggestively lower volitional control than the inward attention profile, F(1, 55) = 3.84, p = .055, g2 = .07. However, in contrast with the predictions of some variants of the dissociative typological model (Barber, 1999a; Barrett, 1996), the two profiles did not differ in memory, F < 2. Barrett (1996) presented evidence indicating that dissociative HS individuals exhibit greater ISHD scores than their nondissociative counterparts. We next sought to test the effectiveness of the ISHD for discriminating the two profiles of HS participants. In line with Barrett’s findings, HS participants in the dissociative profile exhibited significantly greater ISHD scores (M = 24.77, SD = 5.10) than those in the inward attention profile (M = 21.76, SD = 4.55), F(1, 54) = 5.21, p < .05, g2 = .09. This Table 5 Descriptive statistics (Mean and Standard Deviation) for the PCI dimensions and sub-dimensions in the two profiles of highly suggestible participants. Variable Altered experience Attention Direction Absorption Imagery Amount Vividness Memory Volitional control Positive affect Negative affect Profile Inward attention (n = 24) Dissociative (n = 33) 2.42 (0.70) 4.58 (0.80) 4.76 (0.80) 4.31 (1.04) 3.40 (1.29) 3.60 (1.51) 3.33 (1.19) 3.83 (1.15) 2.91 (1.10) 0.38 (1.01) 0.70 (0.09) 2.93 (1.02) 4.00 (1.19) 4.07 (1.50) 3.91 (1.41) 3.06 (1.25) 3.53 (1.53) 2.59 (1.25) 3.53 (1.05) 2.30 (1.18) 0.70 (0.90) 0.74 (1.19) 1146 D.B. Terhune, E. Cardeña / Consciousness and Cognition 19 (2010) 1140–1150 finding was followed up with a 2 (Profile)  3 (Hypnotic suggestibility) analysis of variance (ANOVA) on ISHD scores to examine whether the relationship between involuntariness and hypnotic suggestibility differs across profiles. There was no main effect of Profile, F < 1.5, but a main effect of Hypnotic suggestibility, F(2, 383) = 26.10, p < .001, g2 = .12, which was qualified by a Profile  Hypnotic suggestibility interaction, F(2, 383) = 4.18, p < .05, g2 = .02. Subsidiary one-way ANOVAs revealed main effects of Hypnotic suggestibility in the inward attention, F(2, 128), p < .001, g2 = .10, and dissociative, F(2, 255) = 29.41, p < .001, g2 = .19, profiles. Although low suggestible participants (M = 17.35, SD = 4.45) in the inward attention profile exhibited lower ISHD scores than medium (M = 20.51, SD = 4.74) and HS participants, ps < .01, the latter two did not differ, p > .05. In contrast, low (M = 13.63, SD = 7.57) and medium (M = 19.00, SD = 6.61) suggestible participants in the dissociative profile differed from one another as well as HS participants, ps < .001. This indicates that the relationship between hypnotic suggestibility and involuntariness during hypnotic responding is linear in the dissociative profile but plateaus in the inward attention profile in medium to high levels of hypnotic suggestibility. We next report analyses examining differential affective response between the two subtypes of HS participants. A mixedmodel ANOVA with Affect as a repeated-measures variable (positive vs. negative) and Profile (inward attention vs. dissociative) as a between-groups variable using Z-score transformed values for the PCI dimension scores revealed main effects of Affect, F(1, 55) = 7.85, p < .01, g2 = .13, Profile, F(1, 55) = 27.60, p < .001, g2 = .33, and an Affect  Profile interaction, F(1, 55) = 9.06, p < .01, g2 = .14. Follow-up repeated measures ANOVAs revealed a main effect of Affect in the inward attention profile, with lower negative than positive affect, F(1, 23) = 26.91, p < .001, g2 = .54, but no effect in the dissociative profile, F < 0.5. These findings indicate that the dissociative profile exhibits an elevated level of general affect, relative to participants in profiles 3 and 4, whereas the inward attention profile only exhibits elevated positive affect. 4.5. Hypnotic suggestibility as a function of profile We finally undertook a series of exploratory analyses to discern differences in hypnotic suggestibility between the two profiles. The inward attention profile exhibited significantly greater WSGC total scores (M = 5.41, SD = 2.14) than the dissociative profile (M = 4.94, SD = 2.14), F(1, 387) = 4.20, p < .05, but the magnitude of this difference was negligible: g2 = .01. After a Bonferroni correction (a = .004), the inward attention profile was found to more frequently respond to the direct ideomotor (arm heaviness) suggestion (90%) than the dissociative profile (74%), v2(1, N = 389) = 13.16, p < .001, phi = .18. There were also trends for the inward attention profile (44%) to exhibit greater responsiveness than the dissociative profile (34%) on the posthypnotic drawing item, v2(1, N = 389) = 3.83, p = .050, phi = .10, but less responsiveness to the negative visual hallucination item (inward attention: 17%, dissociative: 25%), v2(1, N = 389) = 3.24, p = .072, phi = .09. The two profiles did not differ on WSGC total scores or any individual WSGC items when the analyses were restricted to HS participants. 5. Discussion This study sought to identify discrete experiential profiles in response to a hypnotic induction and examine whether the profiles of HS participants corresponded to the patterns predicted by dissociative typological models of high hypnotic suggestibility (e.g., Barber, 1999a). The results indicate that phenomenological response to hypnosis can be classified in terms of four experiential profiles. Two involve marked alterations in a variety of experiential dimensions, whereas the other two are characterized by relatively minor experiential shifts. All HS participants fell into the first two profiles, whereas medium and low suggestible participants were distributed among the four profiles. The first two profiles differed in endogenous attention and negative affect, suggesting that they corresponded to the fantasy-prone and dissociative subtypes, respectively, predicted by the dissociative typological models (e.g., Barber, 1999a; Barrett, 1996). Upon closer inspection, HS participants in these two profiles were found to exhibit differential levels of awareness, affect, attention, imagery, and volitional control. All observed findings were in the direction predicted by the dissociative typological models (Barber, 1999a; Barrett, 1996; Carlson & Putnam, 1989). Critically, in replication of a previous finding (Barrett, 1996), the two subtypes were also found to differ in involuntariness during hypnotic responding, as measured by the ISHD. In particular, the fact that the ISHD correlated strongly with dissociated control, but only weakly with imagery scores provides further support for the utility of this measure for discriminating the two subtypes (Barrett, 1996). The results also corroborate many of the findings of previous cluster analyses on spontaneous experiential response to a hypnotic induction (Pekala & Kumar, 2007), as well as the relationships between hypnotic suggestibility and the PCI state scores (Kumar, Pekala, and Cummings, 1996; Kumar, Pekala, and Marcano, 1996). In sum, the results provide strong support for the proposal that HS individuals are comprised of two distinct subtypes of respondents. Despite the support found for the dissociative typological models, our results diverge from the models’ predictions in multiple respects that are worth considering. First, no evidence was found for a third HS subtype, positively-set respondents (Barber, 1999a). It is plausible that the inward attention and dissociative profiles had members with minimal alterations in awareness that correspond to the positively-set subtype but which were either too few in number or not sufficiently unique in their displayed experiential response patterns to be classified as a discrete phenomenological profile. This possibility notwithstanding, the results favor bifurcated variants of the dissociative typological model (Barrett, 1996; Cardeña, 1996; Carlson & Putnam, 1989; Perry, 2004) rather than the trifurcated version (Barber, 1999a). In addition, the dissociative profile did not exhibit reduced episodic memory during hypnosis, as predicted by Barber (1999a; see also Barrett, 1996), and found in previous cluster analyses (Pekala & Kumar, 2007). Further, despite reporting less vivid imagery than the inward attention D.B. Terhune, E. Cardeña / Consciousness and Cognition 19 (2010) 1140–1150 1147 profile, the dissociative profile still exhibited greater vividness of imagery than profiles 3 and 4. This finding is inconsistent with the hypothesis that this subtype experiences minimal imagery following a hypnotic induction (Barber, 1999a; see also Pekala & Kumar, 2007). These disparities may stem from cultural differences (e.g., expectancies) in our sample, relative to previous North American samples. Alternatively, a large proportion of cognitive-perceptual suggestions in the measure of hypnotic suggestibility that we used (Bowers, 1993, 1998) may have invoked greater amounts of imagery, which carried over into the resting epoch. At the very least, these disparities suggest that HS individuals are comprised of two distinct subtypes and that spontaneous episodic memory deficits during hypnosis should not be regarded as a critical marker of typological variability in high hypnotic suggestibility. A novel finding of this study is that the strongest discriminator of the experiential response profiles of the two types of HS participants was negative affect. Specifically, the dissociative profile exhibited greater negative affect following a hypnotic induction than the inward attention profile. Although previous cluster analyses of PCI dimension scores during hypnosis did not observe greater negative affect in the dissociative subtype of HS participants (Pekala & Kumar, 2007), dissociative tendencies have been found to predict negative affect during hypnosis (Kumar, Pekala, and Marcano, 1996; Pekala et al., 2009). More broadly, this may suggest a greater proneness to psychopathology in this profile (Pekala et al., 2009), as has been argued elsewhere (e.g., Lynn, Lilienfeld, & Rhue, 1999). One explanation for this finding is that the distortions in awareness produced by the hypnotic induction induced state-dependent memory intrusions in dissociative participants corresponding to negative events to which the participants had previously responded with experiential detachment (e.g., Spiegel & Cardeña, 1990). A second possibility is that greater negative affect may have resulted from participants’ response expectancies (Kirsch, 1999; Lynn, Kirsch, & Hallquist, 2008). However, a recent study, which found that expectations for negative affect during hypnosis were unrelated to its occurrence during hypnosis after controlling for baseline negative affect (Cardeña, Jönsson, Terhune, & Lehmann, 2009), casts doubt on this interpretation. A final explanation for increased negative affect during hypnosis in the dissociative profile is that the participants in the two profiles experienced increased general affect during hypnosis, but the dissociative profile was unable to sufficiently regulate negative affect due to weakened executive control. For instance, both profiles of HS participants experienced elevated positive affect, but only the dissociative profile experienced elevated negative affect. A finding in another study, that high dissociative HS participants displayed impaired cognitive control during hypnosis relative to a control condition, whereas low dissociative HS participants exhibited marginally superior cognitive control (Terhune, Cardeña, & Lindgren, 2009), clearly supports this interpretation. Increased involuntariness during hypnotic responding among the dissociative profile is also consistent with weakened control during hypnosis in this subtype. In non-hypnotic contexts, a negative mood has been found to increase mind wandering (Smallwood, Fitzgerald, Miles, & Phillips, 2009), which is associated with attentional lapses (Smallwood et al., 2008). Future work should attempt to directly link impaired cognitive control during hypnosis in the dissociative profile with increases in negative affect. Importantly, the present study failed to identify unequivocal behavioral signatures of the two subtypes of HS individuals. The only robust behavioral difference between the two profiles was the increased level of responsiveness to the direct ideomotor suggestion of the WSGC in the inward attention profile. This difference may have been caused by the fact that this suggestion was administered first, as dissociative HS individuals may require a longer hypnotic induction before achieving an optimal level of hypnotic suggestibility (Barber, 1999a; Barrett, 1996; Brown & Oakley, 2004). More broadly, the WSGC (and a group environment) may be insufficient for discerning differences among HS individuals. Measures of hypnotic suggestibility with larger proportions of cognitive-perceptual suggestions may be better suited to this task (Weitzenhoffer & Hilgard, 1967; Woody & Barnier, 2008). In addition, future research may consider measuring factorial invariance, that is, equivalence of factor structures, of hypnotic suggestibility scales across the two profiles. Another suitable place to look for differences between the two subtypes may be in their utilization of cognitive strategies during hypnotic responding. Studies on response strategy utilization have provided evidence for distinct subtypes of respondents (Danziger et al., 1998; Kunzendorf & Boisvert, 1996; Winkel, Younger, Tomcik, Borckardt, & Nash, 2006). For instance, Kunzendorf and Boisvert (1996) found that suggestions for negative hallucinations and hyperaesthesia modulated brainstem evoked potentials in only a subset of HS individuals despite the fact that all reported the phenomenal impression of responding to the suggestions. The reconciliation of individual differences in phenomenological response to a hypnotic induction with differential response strategy utilization should be afforded greater attention in future research (see also Brown & Oakley, 2004). A question raised by critics of the typological models is whether the different subtypes are dimensional, that is, extending from low to high hypnotic suggestibility or taxonic, that is, reflective of a discrete subtype of HS respondent (e.g., Lynn et al., 1999). Barber’s (1999a) model is somewhat equivocal in this regard (see commentaries accompanying Barber, 1999b). Although by no means conclusive, the present results suggest that typological variability in experiential response is dimensional rather than taxonic. HS individuals were classified into two profiles, which also included participants of low and medium suggestibility, whereas a taxonic typological pattern would predict that HS individuals would form two or more distinct profiles. In so far as many of the participants in the dissociative profile exhibited low or medium hypnotic suggestibility, the present findings are also broadly consistent with previous research that demonstrated that high dissociative individuals uniformly experience high experiential involvement during hypnotic responding, but only some display high hypnotic suggestibility (Kumar, Pekala, and Marcano, 1996). However, the dimensional structure of each profile and in particular, its relationship to hypnotic responding, may differ. For instance, involuntariness during hypnotic responding, and dissociated control and attention to internal processes during the resting epoch, increased linearly as a function of hypnotic suggestibility in the dissociative profile, but did not increase from medium to high hypnotic suggestibility in the inward attention 1148 D.B. Terhune, E. Cardeña / Consciousness and Cognition 19 (2010) 1140–1150 profile. These results could be interpreted as reflecting a taxonic distribution in the dissociative profile and a dimensional distribution in the inward attention profile. Dimensional and taxonic variants of the typological models, and their attendant predictions, require closer inspection in future studies. An alternative to the typological models is the componential model (Woody & Barnier, 2008; Woody, Barnier, & McConkey, 2005). According to this account, hypnotic suggestibility is determined by a single latent factor and individual differences at specific levels of hypnotic suggestibility, in particular those among HS individuals, are modulated by ancillary ‘componential’ abilities. For instance, imagery ability may not correlate with general hypnotic suggestibility but may facilitate responsiveness to particular types of suggestions among HS individuals (see, e.g., Laurence, Beaulieu-Prévost, & du Chéné, 2008). On this account, the two experiential profiles of HS individuals observed in this study do not constitute discrete subtypes per se, but rather different ancillary aptitudes (e.g., for altering awareness and agency) that, in turn, affect particular features of hypnotic responding (e.g., involuntariness), but which are not indicative of differential underlying mechanisms. Although the componential model is a valuable alternative to the typological models and warrants greater attention, the different abilities that contribute to individual differences among HS individuals remain underspecified (Woody et al., 2005; see also Laurence et al., 2008). This model requires refinement before it can generate testable predictions that clearly diverge from those of the typological models. The present study is limited in at least four respects. First, in so far as the results are dependent upon self-reports, some participants, particularly those in the dissociative profile, who exhibited greater distortions in awareness during hypnosis, may have had greater difficulty quantifying their experiential responses. Although the high level of inter-item reliability speaks against this limitation, it would be useful to corroborate self-reported lapses in attention with performance on a behavioral task (e.g., Smallwood et al., 2008; Terhune et al., 2009). Second, because there are no Swedish-language equivalents of the measures included in this study, all of the measures were administered in English. However, previous work indicates that deflation of hypnotic suggestibility because of English measurement with a Swedish sample is negligible (Cardeña, Kallio, Terhune, Buratti, & Lööf, 2007). Furthermore, our measures exhibited strong internal consistency and reliability and the observed WSGC scores are comparable to those of a recent study with a British sample (Dienes et al., 2009). This renders unlikely the possibility that English-language administration of the measures represents a serious confound. A third potential limitation stems from the selection of an LPA model that allowed for class-dependent unrestricted covariance matrices. Allowing local dependencies between indicator variables possesses a number of strengths, such as the prevention of selecting a model with too many profiles, but it may also function to hide additional meaningful profiles (Vermunt & Magidson, 2002). A final limitation concerns the small number of experiential dimensions included as indicator variables in the LPA models. The success of any clustering technique is dependent upon the extent to which the selected indicator variables measure the dimensions of interest. As a result, the present analyses may not have included all relevant experiential dimensions that could discriminate among the different profiles. Future research should consider including a wider variety of experiential dimensions. 6. Conclusions This study used LPA to identify discrete experiential profiles in response to a hypnotic induction to test the prediction that there are distinct subtypes of HS individuals (e.g., Barber, 1999a). We identified a homogeneous subset of dissociative HS participants who exhibit pronounced distortions in awareness, affect, and volitional control and reduced attention and imagery during hypnosis relative to a second profile of HS participants who were primarily characterized by endogenously-directed attention. The former was also found to exhibit increased involuntariness during hypnotic responding. We maintain that these experiential responses can be understood as reflecting a weakening of executive functioning following a hypnotic induction that is isolated to the dissociative profile (see also Barber, 1999a; Brown & Oakley, 2004). In so far as there is consensus that involuntariness is the core phenomenological feature of hypnotic responses and the primary explanandum of experimental hypnosis research (Kihlstrom, 2008; Kirsch & Lynn, 1998; Weitzenhoffer, 1980), these findings have critical implications. They indicate that the relationship between involuntariness and hypnotic suggestibility is modulated by (typological) experiential response to a hypnotic induction. That is, increased hypnotic suggestibility among HS, relative to low and medium suggestible individuals, appears to only be coupled with increased involuntariness in the dissociative profile. These findings further suggest that the mechanisms underlying hypnotic responding in dissociative HS individuals are either different from or more pronounced than those underlying the responses of individuals in the inward attention subtype. Acknowledgment This research was supported by Research Bursary 54/06 from the Bial Foundation and the David Caul Graduate Research Award from the International Society for the Study of Dissociation. The research assistance of Tina Koch is gratefully acknowledged. References Akaike, H. (1987). Factor analysis and the AIC. Psychometrika, 52, 317–332. Barber, T. X. (1999a). A comprehensive three-dimensional theory of hypnosis. In I. Kirsch, A. Capafons, E. Cardeña-Buelna, & S. Amigo (Eds.), Clinical hypnosis and self-regulation: Cognitive-behavioral perspectives (pp. 21–48). Washington, DC: American Psychological Association. D.B. Terhune, E. Cardeña / Consciousness and Cognition 19 (2010) 1140–1150 1149 Barber, T. X. (1999b). Hypnosis: A mature view. Contemporary Hypnosis, 16, 123–127. Barrett, D. (1996). Fantasizers and dissociaters: Two types of high hypnotizables, two different imagery styles. In R. G. Kunsendorf, N. P. Spanos, & B. Wallace (Eds.), Hypnosis and imagination (pp. 123–135). Amityville, NY: Baywood. Bowers, K. S. (1993). The Waterloo-Stanford Group C (WSGC) scale of hypnotic susceptibility: Normative and comparative data. International Journal of Clinical and Experimental Hypnosis, 41, 35–46. Bowers, K. S. (1998). Waterloo-stanford group scale of hypnotic susceptibility, form C: Manual and response booklet. International Journal of Clinical and Experimental Hypnosis, 46, 250–268. Brown, R. J., & Oakley, D. A. (2004). An integrative cognitive theory of hypnosis and high hypnotizability. In M. Heap, R. J. Brown, & D. A. Oakley (Eds.), The highly hypnotizable person: Theoretical, experimental and clinical issues (pp. 152–186). NY: Routledge. Cardeña, E. (1996). Just floating in the sky. A comparison of shamanic and hypnotic phenomenology. In R. Quekelbherge & D. Eigner (Eds.), 6th Jahrbuch fur transkulturelle medizin und psychotherapie [6th Yearbook of cross-cultural medicine and psychotherapy] (pp. 367–380). Berlin: Verlag fur Wissenschaft und Bildung. Cardeña, E. (2005). The phenomenology of hypnosis: Physically active and quiescent. International Journal of Clinical and Experimental Hypnosis, 53, 1–23. Cardeña, E., Jönsson, P., Terhune, D. B., & Lehmann, D. (2009). Spontaneous hypnotic phenomenology across the hypnotizability spectrum. Manuscript in preparation. Cardeña, E., Kallio, S., Terhune, D. B., Buratti, S., & Lööf, A. (2007). The effect of translation and sex on hypnotizability testing. Contemporary Hypnosis, 27, 154–160. Cardeña, E., & Terhune, D. B. (2009). A note of caution on the Waterloo Stanford Group scale of hypnotic susceptibility. International Journal of Clinical and Experimental Hypnosis, 45, 222–226. Carlson, E. B., & Putnam, F. W. (1989). Integrating research on dissociation and hypnotizability: Are there two pathways to hypnotizability? Dissociation, 2, 32–38. Danziger, N., Fournier, E., Bouhassira, D., Michaud, D., De Broucker, T., Santarcangelo, E., et al (1998). Different strategies of modulation can be operative during hypnotic analgesia: A neurophysiological study. Pain, 75, 85–92. Dienes, Z., Brown, E., Hutton, S., Kirsch, I., Mazzoni, G., & Wright, D. B. (2009). Hypnotic suggestibility, cognitive inhibition, and dissociation. Consciousness and Cognition, 18, 837–847. Egner, T., Jamieson, G., & Gruzelier, J. (2005). Hypnosis decouples cognitive control from conflict monitoring processes of the frontal lobe. Neuroimage, 27, 969–978. Field, P. B. (1965). An inventory of hypnotic depth. International Journal of Clinical and Experimental Hypnosis, 13, 238–249. Forbes, E. J., & Pekala, R. J. (1996). Types of hypnotically (un)susceptible individuals as a function of phenomenological experience: A partial replication. Australian Journal of Clinical & Experimental Hypnosis, 24, 92–109. Goodman, L. A. (2002). Latent class analysis: The empirical study of latent types, latent variables, and latent structures, and some notes on the history of this subject. In J. A. Hagenaars & A. L. McCutcheon (Eds.), Applied latent class analysis (pp. 3–55). Cambridge: Cambridge University Press. Henry, D. (1985). Subjects’ expectancies and subjective experience of hypnosis. Unpublished doctoral dissertation, University of Connecticut. Kihlstrom, J. F. (2008). The domain of hypnosis, revisited. In M. R. Nash & A. J. Barnier (Eds.), Oxford handbook of hypnosis (pp. 21–52). Oxford: Oxford University Press. Kirsch, I. (1990). Changing expectations: A key to effective psychotherapy. Pacific Grove, CA: Brooks/Cole. Kirsch, I. (1999). How expectancies shape experience. Washington, DC: American Psychological Association. Kirsch, I., & Lynn, S. J. (1998). Dissociation theories of hypnosis. Psychological Bulletin, 123, 100–115. Kumar, V. K., & Pekala, R. J. (1988). Hypnotizability, absorption, and individual differences in phenomenological experience. International Journal of Clinical and Experimental Hypnosis, 36, 80–88. Kumar, V. K., & Pekala, R. J. (1989). Variations in phenomenological experience as a function of hypnotic susceptibility: A replication. British Journal of Experimental and Clinical Hypnosis, 6, 17–22. Kumar, V. K., Pekala, R. J., & Cummings, J. (1996). Trait factors, state effects, and hypnotizability. International Journal of Clinical and Experimental Hypnosis, 44, 232–249. Kumar, V. K., Pekala, R. J., & Marcano, G. (1996). Hypnotizability, dissociativity, and phenomenological experience. Dissociation, 9, 143–153. Kunzendorf, R. G., & Boisvert, P. (1996). Presence vs. absence of a ‘hidden observer’ during total deafness: The hypnotic illusion of subconsciousness vs. the imaginal attenuation of brainstem evoked potentials. In R. G. Kunzendorf, N. P. Spanos, & B. Wallace (Eds.), Hypnosis and imagination (pp. 223–234). Amityville, NY: Baywood. Laurence, J.-R., Beaulieu-Prévost, D., & du Chéné, T. (2008). Measuring and understanding individual differences in hypnotizability. In M. R. Nash & A. Barnier (Eds.), Oxford handbook of hypnosis (pp. 225–253). Oxford: Oxford University Press. Lo, Y., Mendell, N., & Rubin, D. B. (2001). Testing the number of components in a normal mixture. Biometrika, 88, 767–778. Lynn, S. J., Kirsch, I., & Hallquist, M. N. (2008). Social cognitive theories of hypnosis. In M. R. Nash & A. Barnier (Eds.), Oxford handbook of hypnosis (pp. 111–139). Oxford: Oxford University Press. Lynn, S. J., Lilienfeld, S., & Rhue, J. W. (1999). An evaluation of Barber’s three-dimensional theory of hypnosis: Promise and pitfalls. Contemporary Hypnosis, 16, 160–164. Lynn, S. J., Meyer, E., & Schindler, K. (2004). Clinical correlates of high hypnotizability. In M. Heap, R. J. Brown, & D. A. Oakley (Eds.), The highly hypnotizable person: Theoretical, experimental and clinical issues (pp. 187–212). NY: Routledge. Magidson, J., & Vermunt, J. K. (2002). Latent class models for clustering: A comparison with K-means. Canadian Journal of Marketing Research, 20, 37–44. McConkey, K. M., & Barnier, A. J. (2004). High hypnotizability: Unity and diversity in behavior and experience. In M. Heap, R. J. Brown, & D. A. Oakley (Eds.), The highly hypnotizable person: Theoretical, experimental and clinical issues (pp. 61–84). NY: Routledge. McCutcheon, A. C. (1987). Latent class analysis. Beverly Hills, CA: Sage. McCutcheon, A. L. (2002). Basic concepts and procedures in single and multiple group latent class analysis. In J. A. Hagenaars & A. L. McCutcheon (Eds.), Applied latent class analysis (pp. 56–85). Cambridge: Cambridge University Press. McLachlan, G., & Peel, D. (2000). Finite mixture models. NY: Wiley. Muthén, L. K., & Muthén, B. O. (1998–2007). Mplus user’s guide (5th ed.). Los Angeles, CA: Muthén & Muthén. Nylund, K. L., Asparouhov, T., & Muthén, B. O. (2007). Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural Equation Modeling, 14, 535–569. Oakley, D. A., & Halligan, P. W. (2009). Hypnotic suggestion and cognitive neuroscience. Trends in Cognitive Sciences, 13, 264–270. Pekala, R. J. (1991a). The phenomenology of consciousness inventory. West Chester, PA: Mid-Atlantic Educational Institute, Inc. Pekala, R. J. (1991b). Hypnotic types: Evidence from a cluster analysis of phenomenal experience. Contemporary Hypnosis, 8, 95–104. Pekala, R. J., & Forbes, E. J. (1997). Types of hypnotically (un)susceptible individuals as a function of phenomenological experience. Towards a typology of hypnotic types. American Journal of Clinical Hypnosis, 39, 212–224. Pekala, R. J., & Kumar, V. K. (2007). An empirical–phenomenological approach to quantifying consciousness and states of consciousness: With particular reference to understanding the nature of hypnosis. In G. A. Jamieson (Ed.), Hypnosis and conscious states: The cognitive neuroscience perspective (pp. 167–194). NY: Oxford University Press. Pekala, R. J., Kumar, V. K., & Hand, J. (1993). Subjective experience, expectancy, and hypnosis: Interacting effects. Contemporary Hypnosis, 10, 133–143. Pekala, R. J., Kumar, V. K., & Marcano, G. (1995). Hypnotic types: A partial replication concerning phenomenal experience. Contemporary Hypnosis, 12, 194–200. 1150 D.B. Terhune, E. Cardeña / Consciousness and Cognition 19 (2010) 1140–1150 Pekala, R. J., Kumar, V. K., Maurer, R. L., Sr., Elliott-Carter, N., Moon, E., & Mullen, K. (2009). Positive affect, negative affect, and negative effects during a phenomenological hypnotic assessment within a substance abuse population. International Journal of Clinical and Experimental Hypnosis, 57, 64–93. Perry, C. (2004). Can anecdotes add to an understanding of hypnosis? International Journal of Clinical and Experimental Hypnosis, 52, 218–231. Rainville, P., & Price, D. D. (2003). Hypnosis phenomenology and the neurobiology of consciousness. International Journal of Clinical and Experimental Hypnosis, 51, 105–129. Ramaswamy, V., Desarbo, W. S., Reibstein, D. J., & Robinson, W. T. (1993). An empirical pooling approach for estimating marketing mix elasticities with PIMS Data. Marketing Science, 12, 103–124. Ruscio, J., & Ruscio, A. M. (2008). Categories and dimensions: Advancing psychological science through the study of latent structure. Current Directions in Psychological Science, 17, 203–207. Schwartz, G. (1978). Estimating the dimension of a model. Annals of Statistics, 6, 461–464. Sclove, S. L. (1987). Application of model-selection criteria to some problems in multivariate analysis. Psychometrika, 52, 333–343. Smallwood, J., Beach, E., Schooler, J. W., & Handy, T. C. (2008). Going AWOL in the brain: Mind wandering reduces cortical analysis of external events. Journal of Cognitive Neuroscience, 20, 458–469. Smallwood, J., Fitzgerald, A., Miles, L. K., & Phillips, L. H. (2009). Shifting moods, wandering minds: Negative moods lead the mind to wander. Emotion, 9, 271–276. Spiegel, D., & Cardeña, E. (1990). New uses of hypnosis in the treatment of PTSD. Journal of Clinical Psychiatry, 51(Suppl), 39–43. Terhune, D.B., Cardeña, E., & Lindgren, M. (2009). Typological variability in executive attention in high hypnotic suggestibility. Poster presented at the 20th Annual Convention of the Association for Psychological Science. Terhune, D.B., Cardeña, E., & Lindgren, M. (2010). Dissociative tendencies and individual differences in high hypnotic suggestibility. Submitted for publication. Vermunt, J. K., & Magidson, J. (2002). Latent class cluster analysis. In J. A. Hagenaars & A. L. McCutcheon (Eds.), Applied latent class analysis (pp. 89–106). Cambridge: Cambridge University Press. Weitzenhoffer, A. M. (1980). Hypnotic susceptibility revisited. American Journal of Clinical Hypnosis, 22, 130–146. Weitzenhoffer, A. M., & Hilgard, E. R. (1962). Stanford hypnotic susceptibility scale: Form C. Palo Alto, CA: Consulting Psychologists Press. Weitzenhoffer, A. M., & Hilgard, E. R. (1967). Revised Stanford profile scales of hypnotic susceptibility: Forms I and II. Palo Alto, CA: Consulting Psychologists Press. Winkel, J. D., Younger, J. W., Tomcik, N., Borckardt, J. J., & Nash, M. R. (2006). Anatomy of a hypnotic response: Self-report estimates, actual behavior, and physiological response to the hypnotic suggestion for arm rigidity. International Journal of Clinical and Experimental Hypnosis, 54, 186–205. Woody, E. Z., & Barnier, A. J. (2008). Hypnosis scales for the twenty-first century: What do we need and how should we use them? In M. R. Nash & A. Barnier (Eds.), Oxford handbook of hypnosis (pp. 255–281). Oxford: Oxford University Press. Woody, E. Z., Barnier, A. J., & McConkey, K. M. (2005). Multiple hypnotizabilities: Differentiating the building blocks of hypnotic response. Psychological Assessment, 17, 200–211. Woody, E. Z., & Bowers, K. S. (1994). A frontal assault on dissociated control. In S. J. Lynn & J. W. Rhue (Eds.), Dissociation: Clinical and theoretical perspectives (pp. 52–79). NY: Guilford Press. Woody, E. Z., & Sadler, P. (2008). Dissociation theories of hypnosis. In M. R. Nash & A. Barnier (Eds.), Oxford handbook of hypnosis (pp. 81–110). Oxford: Oxford University Press.

Consciousness and Cognition 19 (2010) 1122–1123 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Reply Reply to Josipovic: Duality and non-duality in meditation research q Frederick Travis a,b,⇑, Jonathan Shear c a b c Center for the Brain, Consciousness, and Cognition, Maharishi University of Management, 1000 North 4th Street, Fairfield, IA 52557, United States Maharishi University of Management Research Institute, Maharishi Vedic City, IA 52557, United States Department of Philosophy, Virginia Commonwealth University, 817 West Franklin Street, Richmond, VA 23284-9002, United States a r t i c l e i n f o Article history: Available online 14 May 2010 Keywords: Focus attention Open monitoring Non-dual experiences Meditation Automatic self-transcending a b s t r a c t We agree with Josipovic that a fundamental differentiating feature of meditation techniques is whether they remain within the dualistic subject–object cognitive structure, or they transcend this structure to reveal an underlying level of non-dual awareness. Further discussion is needed to delineate the basic non-dual experience in meditation, where all phenomenal content is absent, from the more advanced experience of non-duality in daily life, where phenomenal content is obviously present as well. In this discussion, it is important to recognize that the experiencer–object relation makes the experience dual or nondual, rather than the nature of the object experienced. Ó 2010 Elsevier Inc. All rights reserved. Josipovic defined a second important dimension to categorize meditation practices, namely whether meditation experiences are primarily dual or non-dual: ‘‘In terms of the actual goal of meditation practice, the fundamental differentiating feature of a meditation technique is whether it remains within the dualistic subject-object cognitive structure, or whether it transcends this structure to reveal the underlying non-dual awareness.’’ This experience-dimension maps roughly onto the three categories of cognitive processing in meditation practice—focused attention, open monitoring and automatic self-transcending—expanded by Travis and Shear (2010) from Lutz, Slagter, Dunne, and Davidson (2008). We agree with Josipovic that: ‘‘Both focused attention and open monitoring styles of meditation contain an essentially dualistic orientation of subject-observing-object.’’ Focus on a single object of experience or an orientation to monitoring changing objects of experience keeps the meditator involved with the procedures of the technique. In contrast, the basic non-dual experience (‘‘pure consciousness’’ or ‘‘emptiness’’) is devoid of phenomenological content, and by its nature requires transcending of the processes and objects of meditation. If processes or objects were there, it would not be this widely described experience. We also agree that ‘‘. . .there may be, initially, various degrees of focused attention deployment, until one can access this non-dual awareness.’’ This could be true within a single meditation sitting as well in an individual’s meditation experience over time. We were happy to see Josipovic mention the distinction between (i) the basic non-dual experience in meditation, where all phenomenal content is absent and (ii) the more advanced experience of non-duality in daily life, where phenomenal content is obviously present as well. This is a very important phenomenological distinction, often missed in contemporary discussions of non-duality and the development of consciousness. Our discussion of non-dual experiences here was only with the basic type of non-duality as experienced in meditation, rather than later, more advanced experience. We also agree with his observation that non-dual awareness is not a level of consciousness easily operationalized in cognitive neuroscience. We explored the nature of non-dual experiences during practice of the Transcendental MeditationÒ q Reply to Commentary on Josipovic, Z. (2010). Duality and non-duality in meditation research. Consciousness and Cognition, 19, 1119–1121. ⇑ Corresponding author at: Center for the Brain, Consciousness, and Cognition, Maharishi University of Management, 1000 North 4th Street, Fairfield, IA 52557, United States. Fax: +1 515 472 1189. E-mail address: ftravis@mum.edu (F. Travis). 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.04.003 F. Travis, J. Shear / Consciousness and Cognition 19 (2010) 1122–1123 1123 technique, a meditation practice that fell in the ‘‘automatic self-transcending’’ category. Fifty-two TM practitioners were asked to describe their deepest experiences during TM. They were asked to use their own words, as if they were describing the taste of a strawberry. Content analysis of their descriptions yielded three themes—the absence of time, absence of space, and absence of body sense (Travis & Pearson, 2000). Time, space and body sense are the framework for giving meaning to waking experience. It is understandable that the basic non-dual meditation experience, characterized by the absence of this meaning-making framework, would be outside of usual cognitive models. We’d like to address Josipovic’s discussion of gamma EEG. First, we agree that gamma EEG during meditation needs to be carefully distinguished from eye saccades (Yuval-Greenberg, Tomer, Keren, Nelken, & Deouell, 2008) and effects of muscle tone (Whitham et al., 2007). Second, Josipovic questioned whether loving kindness meditation should be in the focused attention category with Qigong, Zen (focusing on the 3rd ventricle) and Diamond Way Buddhism. He suggests that ‘‘. . .the non-referentiality of compassion makes it more akin to meditations in the non-dual or ‘automatic self-transcending’ category.’’ As we discussed, gamma EEG reflects local processing within short-range connections responsible for object recognition and so construction of the content of experience (Lubar, 1997; Singer, 1999). The description of the procedure of loving kindness and compassion—to create a specific experience, pure compassion—suggests that the subject actively constructs in awareness a specific experience different from other experiences. As gamma EEG is reported during ‘‘. . .construction of the content of experience,’’ it appears that loving kindness would fit in this category. In our opinion, it is not the nature of the object experienced, but the experiencer–object relation that makes an experience non-dual, and Josipovic’s observation about the non-referentiality of the object of experience (compassion) does not itself speak to this relation. In addition to their focused attention practice, the monks studied may well have been having non-dual experiences of the advanced type. Here, as elsewhere, teasing out the reality will require close attention to the details of the phenomenology involved, as Josipovic so rightly emphasizes. Last, Josipovic mentions that Cahn, Delorme, and Polich (2010) reported ‘‘parietal–occipital’’ gamma EEG during Mindfulness Meditation. In our paper, we commented: ‘‘Gamma was reported in the four sensors on the periphery of the electro-cap, with P7 and P8 being over temporal rather than parietal cortices. This study is included in the open monitoring category, marked by theta activity, because theta (also) significantly increased during this meditation and the meaning of gamma power on the periphery of the recorded EEG is not clear.’’ We echo Josipovic’s ending comment that EEG signals during meditation may more reflect the overall levels of arousal in the brain and the specifics of various cognitive processes associated with the techniques of meditation rather than the end state of meditation practice itself. Adding this second dimension—the nature of the subjective experiences during meditation along a dual/non-dual continuum—is a valuable step towards building a taxonomy to guide meditation research. References Cahn, B. R., Delorme, A., & Polich, J. (2010). Occipital gamma activation during Vipassana meditation. Cognitive Processing, 11, 39–56. Josipovic, Z. (2010). Duality and non-duality in meditation research. Consciousness and Cognition, 19, 1119–1121. Lubar, J. F. (1997). Neocortical dynamics: Implications for understanding the role of neurofeedback and related techniques for the enhancement of attention. Applied Psychophysiology Biofeedback, 22(2), 111–126. Lutz, A., Slagter, H. A., Dunne, J. D., & Davidson, R. J. (2008). Attention regulation and monitoring in meditation. Trends in Cognitive Science, 12(4), 163–169. Singer, W. (1999). Neuronal synchrony: A versatile code for the definition of relations? Neuron, 24(1), 49–65. Travis, F., & Shear, J. (2010). Focused attention, open monitoring and automatic self-transcending: Categories to organize meditations from Vedic, Buddhist and Chinese traditions. Consciousness and Cognition, 19(4), 1110–1118. Travis, F., & Pearson, C. (2000). Pure consciousness: Distinct phenomenological and physiological correlates of ‘‘consciousness itself’’. International Journal of Neuroscience, 100(1–4), 77–89. Whitham, E. M., Pope, K. J., Fitzgibbon, S. P., Lewis, T., Clark, C. R., Loveless, S., et al (2007). Scalp electrical recording during paralysis: Quantitative evidence that EEG frequencies above 20 Hz are contaminated by EMG. Clinical Neurophysiology, 118(8), 1877–1888. Yuval-Greenberg, S., Tomer, O., Keren, A. S., Nelken, I., & Deouell, L. Y. (2008). Transient induced gamma-band response in EEG as a manifestation of miniature saccades. Neuron, 58(3), 429–441.

Consciousness and Cognition 19 (2010) 1110–1118 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Review Focused attention, open monitoring and automatic self-transcending: Categories to organize meditations from Vedic, Buddhist and Chinese traditions Fred Travis a,b,*, Jonathan Shear c a Center for the Brain, Consciousness, and Cognition, Maharishi University of Management, 1000 North 4th Street, Fairfield, IA 52557, United States Maharishi University of Management Research Institute, Maharishi Vedic City, IA 52557, United States c Department of Philosophy, Virginia Commonwealth University, 817 West Franklin Street, Richmond, VA 23284-9002, United States b a r t i c l e i n f o Article history: Received 29 March 2008 Available online 18 February 2010 Keywords: Meditation Mindfulness TM Transcendental Meditation Coherence Zen Tibetan Buddhism Gamma Alpha a b s t r a c t This paper proposes a third meditation-category—automatic self-transcending— to extend the dichotomy of focused attention and open monitoring proposed by Lutz. Automatic selftranscending includes techniques designed to transcend their own activity. This contrasts with focused attention, which keeps attention focused on an object; and open monitoring, which keeps attention involved in the monitoring process. Each category was assigned EEG bands, based on reported brain patterns during mental tasks, and meditations were categorized based on their reported EEG. Focused attention, characterized by beta/gamma activity, included meditations from Tibetan Buddhist, Buddhist, and Chinese traditions. Open monitoring, characterized by theta activity, included meditations from Buddhist, Chinese, and Vedic traditions. Automatic self-transcending, characterized by alpha1 activity, included meditations from Vedic and Chinese traditions. Between categories, the included meditations differed in focus, subject/object relation, and procedures. These findings shed light on the common mistake of averaging meditations together to determine mechanisms or clinical effects. Ó 2010 Elsevier Inc. All rights reserved. 1. Introduction Meditation practices are embedded in different cultures, worldviews, and traditions, which confounds discussions between meditation traditions. Neuroscience provides the language of brain functioning to discuss meditation practices. Brain patterns reflect the cognitive processes used in meditation practices (attention, feeling, reasoning, visualization), the way these processes are used (minimal- to highly-controlled cognitive processing), and the objects of meditation (thoughts, images, emotions, breath) (see Shear, 2006). Thus, brain patterns could provide an objective ‘‘language” to discuss procedures and experiences resulting from different meditation practices. Lutz has divided meditation practices into two categories: focused attention meditation, which entails voluntary and sustained attention on a chosen object, and open monitoring meditation, which involves non-reactive monitoring of the moment-to-moment content of experience (Lutz, Slagter, Dunne, & Davidson, 2008). We suggest a third category of meditation practice, automatic self-transcending, which includes techniques designed to transcend their own activity. * Corresponding author. Address: Center for the Brain, Consciousness, and Cognition, Maharishi University of Management, 1000 North 4th Street, Fairfield, IA 52557, United States. E-mail address: ftravis@mum.edu (F. Travis). 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.01.007 F. Travis, J. Shear / Consciousness and Cognition 19 (2010) 1110–1118 1111 The category of automatic self-transcending is marked by the absence of both (a) focus and (b) individual control or effort. Focus on a single object of experience and an orientation to monitoring changing objects of experience keeps the meditator involved with the procedures of the technique—these practices are not designed to transcend their activity. Focus and monitoring experience are active mental processes, which keep the brain engaged in specific processing—individual activity keeps the mind from transcending. Thus, automatic self-transcending appears to define a class of meditations distinct from both focused attention and open monitoring. These three categories are not mutually exclusive within a single session or over the course of a life-time of meditation practice. Focused attention and open monitoring are combined in Zen, Vipassana and Tibetan Buddhism meditation traditions (Austin, 2006; Gyatso & Jinpa, 1995; Lutz et al., 2008). Also, with diligent practice over many years, focused attention meditations may lead to reduced cognitive control and could result in ‘‘effortless” concentration (Lutz et al., 2008; Wallace, 1999). Each meditation-category can be distinguished by its associated cognitive processes. Since different cognitive processes are associated with activity in different frequency bands (von Stein & Sarnthein, 2000), each category can be assigned characteristic EEG frequency band(s). The brainwave patterns reported during each meditations technique could be used to assign meditations to categories. These grouping of meditations will allow us to understand these three categories in terms of differences in attentional control, subject/object relation, and the nature of different meditation procedures. The purpose of this categorization is to appreciate the nature of different practices and not to assign a ‘‘grade” or value judgment to each one. 2. Cognitive processing and EEG frequency bands Different processing modules work in parallel during information processing (Varela, Lachaux, Rodriguez, & Martinerie, 2001). Low frequency rhythms (theta and alpha) reflect top-down information processing involving attention and workingmemory retention, whereas high frequency rhythms (beta2 and gamma) reflect bottom-up processing of the contents of experience (Razumnikova, 2007). While all frequencies work in concert, individual frequencies can be associated with specific cognitive processes (von Stein & Sarnthein, 2000). 2.1. Gamma bands (30–50 Hz) and Beta 2 (20–30 Hz) Gamma activity reflects local processing within short-range connections responsible for object recognition and so construction of the content of experience (Lubar, 1997; Singer, 1999). Synchronized gamma serves as a gain control for mental processing (Salinas & Sejnowski, 2001), enabling postsynaptic potentials to integrate and so direct downstream networks to bind the elements of sensory processing into a perceptual object (von Stein & Sarnthein, 2000). Gamma band activity closely follows local changes in brain blood flow and increases synaptic plasticity important for long term memories (Niessing et al., 2005). Gamma activity is higher when actively maintaining abstract visual shapes in short-term memory, and is higher in attended compared to unattended stimuli (Jensen, Kaiser, & Lachaux, 2007). Beta2 activity also is reported during focused executive processing. For instance, EEG patterns during Remote Associate Tasks in comparison to Simple Associate Tasks—Remote Associate Tasks require more attention—are characterized as widespread increases of beta2 activity and decreases of frontal alpha1 coherence (Razumnikova, 2007). Consequently, meditation practices that involve highly focused attention to a specific object in the experiential field might lead to higher activity in the beta2 and gamma bands. It should be noted, that gamma activity reported during meditation practice might be confounded by EMG artifacts. Researchers recorded EEG during an auditory oddball task on two subjects in the presence (one subject) and absence (other subject) of complete neuromuscular blockade, sparing the dominant arm. The recordings were conducted in a Faraday cage to eliminate external sources of high frequency EEG. EEG rhythms in the paralyzed subject had six times less power in the 25–30 Hz band, and 100 times less power in the 40–100 Hz band (Whitham et al., 2007). Thus, muscle contamination of the EEG is an important consideration in the interpretation of gamma EEG during meditation practices. 2.2. Beta1 band (13–20 Hz) Cortico-thalamic feedback loops modulating attention operate in the beta1 frequency. Beta1 bursts shift the system to an attention state that consequently allows for gamma synchronization and perception (Wrobel, 2000). Beta1 activity arises from ‘‘regional” processes that develop between nearby macrocolumns (Lubar, 1997). Beta1 activity has been associated with binding of sensory qualities into a unified perception, such as the integration of visual and auditory information (Hanslmayr et al., 2007; von Stein, Rappelsberger, Sarnthein, & Petsche, 1999; von Stein & Sarnthein, 2000). Increase of temporal and parietal 13–18 Hz beta1 coherence was seen across recognition tasks involving pictures, spoken words and written words. Consequently, beta1 activity during meditation practices may play a role in creating the unity of meditation experiences and could be part of all three categories. 2.3. Alpha band (8–12 Hz) Alpha power has been considered a sign of cortical idling (Pfurtscheller, Stancak, & Neuper, 1996). High alpha activity in sensory and motor areas has been correlated with lower thalamic activity and lower posterior cerebral metabolic rate during 1112 F. Travis, J. Shear / Consciousness and Cognition 19 (2010) 1110–1118 eyes-closed rest (Oakes et al., 2004). The association of higher alpha activity with reduced cortical excitability is supported by reports of longer latency evoked potentials (Sauseng, Klimesch, Gerloff, & Hummel, 2009) and degraded perception of incoming stimuli (Thut & Miniussi, 2009) with higher posterior alpha power. Higher posterior alpha power may play a role in perceptual tuning of sensory areas in anticipation of visual events by deactivating brain areas involved in irrelevant processing (Ergenoglu et al., 2004; Rihs, Michel, & Thut, 2007). This classical understanding of alpha as cortical idling holds for simple sensori-motor tasks. In contrast, alpha activity in association cortices is reported to positively covary with task demands. This so-called ‘paradoxical’ frontal alpha is reported during tasks involving internally directed attention (Shaw, 1996) such as imagining a tune compared to listening to the same tune (Cooper, Burgess, Croft, & Gruzelier, 2006). Alpha activity in association areas may represent liveliness of the ‘‘screen of consciousness,” which provides a context for grouping isolated elements into the unity of experience. For instance, when solving a problem by intuition or insight, alpha activity increases first, followed by increases in the gamma band when the idea comes to mind (Kounios & Beeman, 2009). Also, cross frequency coherence—the synchrony between alpha, beta and gamma—increases with higher cognitive load on a continuous mental arithmetic task. Cross frequency coherence is considered important for integrating anatomically distributed processing in the brain (Palva & Palva, 2007; Palva, Palva, & Kaila, 2005). Alpha1 (8–10 Hz) versus alpha2 (10–12). Alpha1 and alpha2 activity desynchronize at different stages in an oddball task, with a warning signal preceding target and non-target stimuli. Alpha1 desynchronized in response to the two warning signals and target; alpha2 desynchronized only after the target was presented (Klimesch, Doppelmayr, Russegger, Pachinger, & Schwaiger, 1998). In another study, EEG complexity negatively correlated with frontal power in the theta and alpha1 bands, and positively correlated with power in the alpha2, beta, and gamma bands (Aftanas & Golocheikine, 2002). Alpha1 appears to index level of internalized attention, alertness and expectancy (Takahashi et al., 2005); alpha2 appears to index engagement of specific brain modules used in task performance. Consequently, we expect posterior alpha2 in any meditation that involves sitting with eyesclosed—classical cortical idling—and frontal alpha1 in any meditation that transcends its own activity. 2.4. Theta band (4–8 Hz) Frontal midline theta, which originates in medial prefrontal and anterior cingulate cortices, is a neural index of monitoring inner processes (Vinogradova, 2001). Frontal midline theta is reported during tasks requiring self-control, internal timing, and assessment of reward (Ishii et al., 1999); during working memory tasks (Sarnthein, Petsche, Rappelsberger, Shaw, & von Stein, 1998); and during tasks requiring memory retention and mental imagery (von Stein & Sarnthein, 2000). Frontal midline theta activity increases a few seconds before a self-initiated hand-movement and reaches a peak immediately after the movement (Tsujimoto, Shimazu, & Isomura, 2006). Theta activity dynamically coordinates central executive circuits during serial subtraction (Mizuhara & Yamaguchi, 2007). Consequently, we expect frontal midline theta in a meditation that involves monitoring of ongoing experience without high levels of control or manipulation of the contents of experience. 2.5. Delta band (1–4 Hz) Delta EEG around 1 Hz is the hallmark of slow wave sleep. The restorative value of sleep is linked to the periods of delta EEG when human growth hormone is secreted (Feinberg, 2000). The delta wave form is generated intracortically and supports synaptic plasticity in cortical neurons (Steriade, 2003). Delta activity during waking is most often associated with pathology such as dementia (Babiloni et al., 2008). However, two meditation studies have reported changes in the delta band during meditation practice. One reported lower frontal delta power (Mindfulness Meditation, (Cahn, Delorme, & Polich, 2010)) and the other reported higher delta source localization as measured by eLORETA (Qigong (Tei et al., 2009)). 3. EEG patterns during different meditation practices Meditation practices were assigned to a category by their reported EEG patterns. Table 1 contains the three meditationcategories and their suggested associated EEG frequency bands (left column), characteristic procedures of this meditationcategory (middle band), and meditation practices that fit into each category as determined by reported EEG. Studies are listed in the table if they included a control group; if the meditation group had been meditating at least a few weeks to have begun to master the meditation practice; and if the technique practiced was a traditional meditation technique rather than a clinically-derived practice. Studies using single-group designs or case studies are summarized at the end of each category on the table. The studies are discussed in depth after the table. The implications of the groupings of meditation practices are explored in the general discussion. 3.1. Category: focused attention In focused attention or concentrative styles of meditations, voluntary sustained attention is focused on a given object, and attention is brought back to the object of attention when the mind has wandered (Cahn & Polich, 2006; Raffone & Srinivasan, 2009). The meditator is controlling the contents in the beam of attention. F. Travis, J. Shear / Consciousness and Cognition 19 (2010) 1110–1118 1113 Table 1 Summary of meditation-categories and associated EEG frequency bands (left column), characteristic elements of each meditation-category (middle band), and meditation practices that fit into each category as determined by the published EEG patterns. Meditation-category and EEG Band Elements of these categories Different meditation practices Focused attention Gamma (30–50 Hz) and Beta2 (20–30 Hz)  Voluntary control of attention and cognitive processes Open Monitoring Theta (5–8 Hz)  Dispassionate, non-evaluative awareness of ongoing experience Automatic Self-Transcending Alpha1 (8–10 Hz)  Automatic transcending of the procedures of the meditation practice  Loving-kindness-compassion (Lutz, Greischar, Rawlings, Ricard, & Davidson, 2004): increased frontal-parietal gamma coherence and power  Other studies with single group or case study designs s Qigong: (Litscher, Wenzel, Niederwieser, & Schwarz, 2001) s Zen–3rd ventricle: (Huang and Lo, 2009) s Diamond Way Buddhism: (Lehmann et al., 2001)  Vipassana meditation (Cahn et al., 2010): decreased frontal delta, increased frontal midline theta and increased occipital gamma power  Zen meditation (ZaZen) (Murata, Koshino, & Ormari, 1994): increased frontal midline theta  Sahaja Yoga (Aftanas and Golocheikine, 2001): increased frontal midline theta and frontal-parietal theta coherence  Sahaja Yoga (Baijal & Srinivasan, 2009): increased frontal midline theta and coherence  Concentrative Qigong (Pan, Zhang, & Xia, 1994): increased frontal midline theta  Transcendental Meditation technique (Dillbeck & Bronson, 1981): increased frontal alpha coherence  Transcendental Meditation technique (Travis et al., 2010): increased frontal alpha1 power and decreased beta1 and gamma power; increased alpha1 and beta1 frontal coherence; and increased activation in the default mode network  Transcendental Meditation technique (Travis & Wallace, 1999): increased frontal coherence in the first minute of TM practice and continued high coherence throughout the session  Transcendental Meditation technique (Travis, 2001): higher frontal alpha coherence during transcending  Transcendental Meditation technique (Travis & Arenander, 2006): higher frontal alpha1 coherence (cross-sectional design) and increasing frontal alpha coherence (1 year longitudinal design)  Transcendental Meditation technique (Hebert, Lehmann, Tan, Travis, & Arenander, 2005): enhanced anterior/posterior alpha phase synchrony  Other case study s Qigong (Qin, Jin, Lin, & Hermanowicz, 2009) Note: All studies reported here used non-equivalent or matched control group designs, except for the first four studies on practice of the Transcendental Meditation technique, which used random assignment designs. Loving-kindness-compassion (Tibetan Buddhist tradition) – higher gamma power and coherence. This study tested EEG patterns recorded during a neutral period and during meditation on unconditional loving-kindness-compassion described as an ‘‘unrestricted readiness and availability to help living beings’’ (Lutz, Greischar, Rawlings, Ricard, & Davidson, 2004). EEG patterns in eight long-term Buddhist practitioners (average age 49 years), with 10,000–50,000 h meditation practice over 15– 40 years, were compared to 10 healthy student volunteers (average age 21 years), with 1 week of meditation training. This was a non-equivalent control group design, with significant age differences between subjects. During meditation compared to controls, the long term subjects had (1) higher global gamma/theta-alpha (4–13 Hz) ratios; (2) 30 times higher parietal, temporal, and frontal gamma power; and (3) higher frontal-parietal gamma synchrony. There was a strong positive correlation between years of practice and relative gamma power (r > 0.6). Other papers reporting activity in the gamma band—single group and case study designs. Three other papers reported increased gamma or beta2 activity during meditation from Chinese and Buddhist traditions. In one study, EEG was measured in a Qigong master (57 years old) and in an experienced female Qigong meditator (47 years old). Compared to eyes-closed rest, gamma activity increased and alpha activity dropped to near zero during meditation in both individuals. Brain blood flow also increased during meditation, as measured by near infrared spectroscopy (Litscher, Wenzel, Niederwieser, & Schwarz, 2001). The next study investigated EEG in 23 experienced Zen meditators (average age 31.3 years) when concentrating on the third ventricle, the so-called ‘‘Zen chakra.” This was a single-group design. During meditation, there was higher complexity in the EEG that was correlated with higher beta2 (20–30 Hz) activity. Frontal alpha power and occipital beta power were higher at the beginning of meditation; while beta2 (20–30 Hz) power was higher in the middle and end (Huang & Lo, 2009). A third study investigated EEG patterns in a long term, advanced meditator, 59 years old, who was a Buddhist Lama of the Karma Kagyu lineage, teaching Diamond Way Buddhism. EEG was recorded during five meditations: (1) Buddha in front; (2) Buddha above, (3) verbalization of a 100-Syllable Mantra, (4) concentration on the experience of dissolution of the self into a boundless unity or emptiness; and (5) concentration on the experience of the reconstitution of the self (Lehmann et al., 2001). Increased gamma activity was seen in brain areas known to be involved in each type of cognitive processing: visual areas, verbal areas, and frontal areas for the 4th and 5th meditations. 1114 F. Travis, J. Shear / Consciousness and Cognition 19 (2010) 1110–1118 Summary: In conclusion, beta2 and gamma activity have been reported during meditation practices from the Tibetan Buddhist, Chinese, and Buddhist traditions. Beta2 and gamma activity were reported when individuals created a vivid inner emotion, sustained focused on an area of the body, or created a strong visual image, and strictly monitored deviation of attention from that object. 3.2. Category: open monitoring Open monitoring or mindfulness-based meditations, involve the non-reactive monitoring of the content of ongoing experience, primarily as a means to become reflectively aware of the nature of emotional and cognitive patterns (Raffone & Srinivasan, 2009). Open monitoring practices are based on an attentive set characterized by an open presence and a nonjudgmental awareness of sensory, cognitive and affective fields of experience in the present moment and involves a higher-order meta-awareness of ongoing mental processes (Cahn & Polich, 2006). Vipassana Meditation (Buddhist Meditation)—frontal theta and occipital gamma. One study has reported EEG patterns during Vipassana (Mindfulness) Meditation. This was a single-group design. It is included in this table, since it is the only published study on EEG patterns during Vipassana or Mindfulness Meditation. EEG was recorded during a neutral thinking period and during Vipassana meditation in 16 individuals who had practiced Vipassana meditation for an average of 20.0 years. This meditation included attentional scanning of sensations throughout the body in an iterative and cyclic fashion—scanning body sensations from the top of the head to the toes and back again repeatedly—with the concomitant adoption of an attitude of detached observation and non-reactivity to any sensations and thoughts that may arise. In both neutral and meditation conditions, posterior alpha power was higher than central or frontal power—with no differences in alpha between the two conditions. During meditation, (1) frontal theta (4–8 Hz) was higher; and (2) occipital (P8, O1, O2, and P7) gamma power was higher (Cahn et al., 2010). Gamma appeared in the four sensors on the periphery of the electro-cap, with P7 and P8 being in temporal rather than parietal cortices. This study is included in the open monitoring category, marked by theta activity, because theta increased during this meditation and the meaning of gamma power on the periphery of the recorded EEG is not clear. Zen meditation (Zazen: Buddhist tradition)—higher frontal midline theta. This is the only published study using control groups that reports EEG findings during Zen meditation (Chiesa, 2009). This study tested EEG patterns recorded in 10 Buddhist monks with extensive experience, 10 monks with moderate experience, and 10 non-meditating controls prior to and during Zazen meditation. Only the long term group showed frontal midline theta during meditation compared to the short-term and control groups (Murata, Koshino, & Ormari, 1994). The level of theta correlated positively with years of experience. Sahaja Meditation (Vedic tradition)—higher theta power and coherence. This study tested EEG patterns recorded in 11 short-term (<1 year) and 16 long-term practitioners (3–7 years) of Sahaja Meditation, a concentrative form of meditation that is part of Sudarshan Kriya yoga. Specific details of this practice are not available. The key experience during Sahaja meditation is a state called ‘thoughtless awareness’ or ‘mental silence’ defined as a state of alertness and awareness free of unnecessary mental activity. EEG was recorded before, during and after meditation. An eightsec artifact-free window was selected from each period, and power and coherence were calculated and compared. The long-term meditators had significantly higher 3.8–5.7 Hz frontal power and frontal-parietal coherence, and significantly higher frontal 5.7–7.5 Hz power (Aftanas & Golocheikine, 2001). (The paper reported these results as changes in the ‘‘theta2” and ‘‘alpha1” bands. However, as you see, they were in the typical theta1 (4.0–6.0 Hz) and theta2 (6.0–8.0 Hz) bands.) Sahaja Samadhi Meditation (Vedic tradition)—higher theta power and coherence. The finding of higher theta (5–8 Hz) power and coherence during Sahaja meditation was replicated in 10 teachers of Sahaja Samadhi meditation at the Art of Living Foundation compared to 10 age-matched subjects who did not practice any meditation techniques. The controls relaxed with eyes closed. The meditating subjects showed increased percentage frontal theta power and frontal theta coherence. Occipital and parietal theta power decreased (Baijal & Srinivasan, 2009). Concentrative Qi-Cong (Chinese tradition)—higher midline frontal theta. This study tested EEG patterns recorded in 20 practitioners of concentrative Qigong, 30 practitioners of non-concentrative Qigong, and 23 control subjects. Relative to the other two groups, frontal midline theta rhythm was higher during the concentrative Qigong state (Pan, Zhang, & Xia, 1994). Other papers reporting activity in the theta band—single-group designs. Three studies from the same laboratory have reported changes during a beginning Zen practice called Su-Soku—counting breathes guided by a metronome. These studies used the same subject population, undergradute students with no meditation experience, and measured EEG during their first day of meditation practice. These three papers reported different findings—theta in one study (Kubota et al., 2001), theta and alpha power in another (Murata et al., 2004), and alpha coherence in the third (Takahashi et al., 2005). Since the same design reported different findings with the same subject population, these finding may not be reliable. Thus, we did not include these three studies in the table and in the later discussion. Summary. In conclusion, Vipassana meditation, as expected, fell in this category of open monitoring, and was characterized by higher frontal theta power. Higher theta power was also seen during ZaZen meditation, and higher theta power and coherence were reported in two studies on Sahaja meditation and one on concentration Qigong. F. Travis, J. Shear / Consciousness and Cognition 19 (2010) 1110–1118 1115 3.3. Category: automatic self-transcending Automatic self-transcending practices involve transcending of the procedures of the meditation. Since cognitive control increases mental activity, self-transcending procedures would need to involve minimal cognitive control—be automatic or effortless. This is explored more in the discussion. Transcendental Meditation (TM) practice—higher frontal alpha1 coherence. This random assignment study reported EEG patterns in eight subjects during Transcendental Meditation practice after 2 weeks practice compared to eyes-closed rest in seven subjects who rested twice a day for the 2 week period (Dillbeck & Bronson, 1981). Transcendental Meditation practice was marked by significantly higher frontal alpha coherence. There were no differences in alpha power or occipital power and coherence. Transcendental Meditation (TM) practice—higher frontal alpha1 and lower beta1 and gamma power, higher frontal alpha1 and beta1 coherence, and greater activation in the Default Mode Network (eLORETA). This random assignment study reported EEG patterns in 19 subjects during Transcendental Meditation technique after 3 months practice compared to eyes-closed rest in 19 control subjects (Travis et al., 2010). Compared to eyes-closed rest, TM practice led to higher alpha1 frontal and lower beta1 and gamma frontal log-power; higher frontal and parietal alpha1 coherence and higher beta1 coherence; and eLORETA analysis identified sources of alpha1 activity in midline cortical regions that overlapped with the default mode network (DMN). The DMN is defined as an intrinsic default brain state, which is lower during goal-directed behaviors requiring executive control and higher during low cognitive load periods such as eyes-closed rest (Gusnard, Raichle, & Raichle, 2001; Raichle & Snyder, 2007), during self-referential mental tasks (Gusnard et al., 2001; Kelley et al., 2002; Vogeley et al., 2001), during tasks involving self-projection—mentally projecting oneself into alternative situations such as envisioning the future (Kelley et al., 2002; Vogeley et al., 2001), and when considering the viewpoint of others (Theory of Mind) (Buckner & Carroll, 2007). Greater activation in the DMN during TM practice suggests that the experience gained during this meditation may be one of greatly reduced cognitive load and heightened sense-of-self. This meditation state could be a foundational or ‘ground’ state of cerebral functioning that may underlie eyes-closed rest and more focused cognitive processes. Transcendental Meditation (TM) practice—higher frontal alpha1 coherence. This random assignment within-subject study reported EEG patterns in 20 subjects during 10 min practice of the Transcendental Meditation technique compared to orderbalanced 10 min eyes-closed rest periods (Travis & Wallace, 1999). The TM periods were distinguished by higher alpha1 frontal and frontal-parietal coherence. These differences were seen in the first minute of TM practice compared to eyesclosed rest, and continued throughout the 10 min session. No differences were seen in alpha power. Transcendental Meditation (TM) practice—higher global alpha power and higher frontal alpha1 coherence. This random assignment within-subject study reported EEG patterns in 25 subjects during transcending compared to order-balanced periods of thought-filled experiences during Transcendental Meditation practice. Transcending was subjectively characterized by the absence of time, space and body sense (Travis & Pearson, 2000), and objectively characterized by: (1) significantly lower breath rates; (2) higher respiratory sinus arrhythmia amplitudes; (3) higher global alpha power; and (4) higher frontal alpha coherence (Travis, 2001). Transcendental Meditation (TM) practice—higher alpha frontal coherence (cross-sectional and longitudinal designs). This study reported two experiments–one used a cross-sectional design and the other used a single group 1 year longitudinal design. In the cross-sectional study, frontal alpha1 coherence was higher in 13 TM subjects with 7 years average TM experience compared to 12 matched controls sitting with eyes closed. In the longitudinal study, frontal alpha1 coherence increased after 2 months’ TM practice and reached that high level at 6 and 12 months posttests (Travis & Arenander, 2006). Transcendental Meditation (TM) practice—enhanced anterior/posterior alpha phase synchrony. This matched control study compared alpha phase synchrony in 15 TM subjects with an average of 25 years TM practice to alpha phase synchrony in twelve control subjects without meditation experience. Significant increases in phase synchrony were found during the meditation condition as compared to the eyes-closed resting condition in long-range electrode pairings between frontal and parietal-occipital sites. There were no significant differences in alpha phase synchrony between the two eyes-closed resting periods in the controls. Enhanced alpha anterior/posterior phase synchrony during TM practice may improve functional integration and may have implications for performance and mind-body health (Hebert et al., 2005). Other papers reporting activity in the alpha1 band—case study. A case study of a single practitioner of Qigong tested once and then 45 years later, reported that global alpha1 power increased immediately during Qigong practice on posttest, and remained higher at rest after the Qigong practice (Qin, Jin, Lin, & Hermanowicz, 2009). Caveat on coherence. Since EEG is a difference-measure, coherence estimates are influenced by activity at both the active and reference electrodes. As the power at the active sensors becomes lower, so the contribution of a common reference will be larger in both signals leading to larger coherence values (Fein, Raz, Brown, & Merrin, 1988). While converting the data to a Laplacian or averaged reference removes the effect of the reference (Travis, 1994), it disrupts the network dynamics. The common reference—though leading to higher coherence values—maintains the coherence and phase synchrony relations among leads and so makes them more interpretable (Thatcher, North, & Biver, 2005). Thus, using a common reference allows better comparison between coherence pairs over the scalp, and any inflation of the coherence from the common reference would equally effect all conditions compared in the research. Summary. Two different meditations were in this category—the Transcendental Meditation technique, described as a technique for automatic transcending (Maharishi Mahesh Yogi, 1969) and Qigong, which included a single meditator with extensive practice (45 years). The idea of automaticity is explored in the discussion. This category of automatic 1116 F. Travis, J. Shear / Consciousness and Cognition 19 (2010) 1110–1118 self-transcending refers to meditation techniques rather than subjective experiences. While these meditations may be automatic, the rate of transcending may vary, person-to-person, and often meditation-to-meditation, owing to differences in the mind and body when one sits to meditate. 4. Discussion The three meditation-categories had different EEG patterns that were associated with different groups of meditation practices with distinct procedures involving different cognitive processes, different ranges of attention, and different subject/object content of experience. The meditations that were grouped under focused attention included meditations that involved voluntary sustained attention on a specific experience—creating a vivid emotion, a strong visual image, or focusing on a body area. The meditations grouped under open monitoring included Vipassana and ZaZen that involve open monitoring. Sahaja and Qigong meditations were also in this category though the details of these practices were not given in the papers. The meditations grouped under automatic self-transcending included Transcendental Meditation and Qigong. TM is described as being automatic in that one uses the ‘‘natural tendency of the mind” to transcend (Maharishi Mahesh Yogi, 1969; Travis et al., 2010). Qigong practice may have become automatic after diligent practice for 45 years. The implications of these groupings are discussed below. It is interesting that the Transcendental Meditation technique, the most researched of the automatic self-transcending procedures, is often placed in the category of focused attention (Cahn & Polich, 2006; Raffone & Srinivasan, 2009). This raises question of the relation of the TM technique and focused attention. Also, it will be important to examine the notion of ‘‘automatic” transcending. 4.1. What is the relation of Transcendental Meditation to focused attention? The reader may ask: why focused attention meditations and the Transcendental Meditation technique have different EEG patterns? TM can be superficially described as thinking or repeating a mantra—a word without meaning—and going back to it when it is forgotten (Raffone & Srinivasan, 2009)—this sounds similar to descriptions of focused attention techniques. A deeper analysis reveals that the TM technique is a technique for transcending its own procedures—appreciating the mantra at ‘‘finer” levels in which the mantra becomes increasingly secondary in experience and ultimately disappears and selfawareness becomes more primary (Maharishi Mahesh Yogi, 1969; Travis & Pearson, 2000). While focused arousal involves voluntary sustained attention, TM practice involves automatic moving of attention to mental silence. During TM practice, the subject-object relation that defines customary experiences is transcended. In focused attention the object of experience is sustained in awareness—the subject (experience) and object co-exist, they are independent but interact. In TM, the object of experience fades away—you use the mantra to lose it. When the mantra disappears, the subject, or the experiencer, as Maharishi puts it, ‘‘finds him/herself awake to his/her own existence” (Maharishi Mahesh Yogi, 1997). 4.2. Investigation of automatic transcending Automatic and controlled processing have been investigated in psychology for the last three decades. In psychology, a task is automatic (1) if the response does not require controlled processing, does not require attention to the steps of responding; and (2) if performance is not affected by increasing task loads (Schneider, Pimm-Smith, & Worden, 1994). Research on automaticity has investigated automaticity through extensive rehearsal. Focused attention meditations when practiced over time are described as leading to ‘‘effortless” concentration (Lutz et al., 2008). This is automaticity resulting from long practice. If the category automatic self-transcending captures automatic transcending of the steps of techniques in general, then one might expect to find the EEG signature of this category–elevated alpha1 power and coherence–in any meditation after long practice. The case study reported here of 45 year Qigong practice supports this prediction (Qin et al., 2009). Future research should further investigate this prediction. The TM technique is automatic at the outset, rather than through extensive practice. This is said to be because it uses the ‘‘natural tendency of the mind” to transcend perception of the mantra (Maharishi Mahesh Yogi, 1969). The automaticity of the TM technique is reflected in research reporting the lack of a novice/expert dichotomy among TM meditators, in contrast to research on other meditation traditions (Brefczynski-Lewis, Lutz, Schaefer, Levinson, & Davidson, 2007). For instance, two cross-sectional studies —comparing individuals with 4 months versus 8 years TM practice (Travis, 1991) or individuals with 7 years versus 32 years TM practice (Travis, Tecce, Arenander, & Wallace, 2002)—report that brain wave patterns reach high levels during TM practice after a few months practice, and that progressive changes in EEG patterns are seen in activity after the meditation session, reflecting experience-related neuroplasticity integrating the meditation experience with daily activity (Travis, Tecce, & Durchholz, 2001). 5. Conclusion Each of the three meditation-categories—focused attention, open monitoring and automatic self-transcending— included different meditation practices with different degrees of attention control, different degree of subject/object relations, and F. Travis, J. Shear / Consciousness and Cognition 19 (2010) 1110–1118 1117 different procedures. Each category appears orthogonal to the others, and together they appear to reflect the wide range of possible meditation practices. These explicit differences between meditation techniques need to be respected when researching physiological patterns or clinical outcomes of meditation practices. If they are averaged together, then the resulting phenomenological, physiological, and clinical profiles cannot be meaningfully interpreted (see Luders, Toga, Lepore, & Gaser, 2009). Attention to these differences in meditation practices will clarify the results gained from researching the power of meditation practices to enhance development of mind and body. Acknowledgment We thank Steve Guich for his comments on earlier drafts of this manuscript. References Aftanas, L. I., & Golocheikine, S. A. (2001). Human anterior and frontal midline theta and lower alpha reflect emotionally positive state and internalized attention: High-resolution eeg investigation of meditation. Neuroscience Letters, 310(1), 57–60. Aftanas, L. I., & Golocheikine, S. A. (2002). Non-linear dynamic complexity of the human EEG during meditation. Neuroscience Letters, 330(2), 143–146. Austin, J. H. (2006). Zen-brian reflections. Cambridge, MA: MIT Press. Babiloni, C., Visser, P. J., Frisoni, G., De Deyn, P. P., Bresciani, L., Jelic, V., et al (2008). Cortical sources of resting EEG rhythms in mild cognitive impairment and subjective memory complaint. Neurobiology of Aging. Baijal, S., & Srinivasan, N. (2009). Theta activity and meditative states: Spectral changes during concentrative meditation. Cognitive Processing. Brefczynski-Lewis, J. A., Lutz, A., Schaefer, H. S., Levinson, D. B., & Davidson, R. J. (2007). Neural correlates of attentional expertise in long-term meditation practitioners. Proceedings of the National Academy of Science, 104(27), 11483–11488. Buckner, R. L., & Carroll, D. C. (2007). Self-projection and the brain. Trends in Cognitive Sciences, 11(2), 49–57. Cahn, B. R., Delorme, A., & Polich, J. (2010). Occipital gamma activation during vipassana meditation. Cognitive Processing, 11(1), 39–56. Cahn, B. R., & Polich, J. (2006). Meditation states and traits: EEG, ERP, and neuroimaging studies. Psychological Bulletin, 132(2), 180–211. Chiesa, A. (2009). Zen meditation: An integration of current evidence. Journal of Alternative and Complementary Medicine, 15(5), 1–8. Cooper, N. R., Burgess, A. P., Croft, R. J., & Gruzelier, J. H. (2006). Investigating evoked and induced electroencephalogram activity in task-related alpha power increases during an internally directed attention task. Neuroreport, 17(2), 205–208. Dillbeck, M. C., & Bronson, E. C. (1981). Short-term longitudinal effects of the transcendental meditation technique on EEG power and coherence. International Journal of Neuroscience, 14(3–4). Ergenoglu, T., Demiralp, T., Bayraktaroglu, Z., Ergen, M., Beydagi, H., & Uresin, Y. (2004). Alpha rhythm of the EEG modulates visual detection performance in humans. Brain Research, Cognitive Brain Research, 20(3), 376–383. Fein, G., Raz, J., Brown, F. F., & Merrin, E. L. (1988). Common reference coherence data are confounded by power and phase effects. Electroencephalogr Clin Neurophysiol, 69(6), 581–584. Feinberg, I. (2000). Slow wave sleep and release of growth hormone. Jama, 284(21), 2717–2718. Gusnard, D. A., Raichle, M. E., & Raichle, M. E. (2001). Searching for a baseline: Functional imaging and the resting human brain. Natural Review Neuroscience, 2(10), 685–694. Gyatso, T., & Jinpa, T. (1995). The world of Tibetan Buddhism: An overview of its philosophy and practice. Somerville, MA: Wisdom Publications. Hanslmayr, S., Klimesch, W., Sauseng, P., Gruber, W., Doppelmayr, M., Freunberger, R., et al (2007). Alpha phase reset contributes to the generation of erps. Cerebral Cortex, 17(1), 1–8. Hebert, R., Lehmann, D., Tan, G., Travis, F., & Arenander, A. (2005). Enhanced EEG alpha time-domain phase synchrony during Transcendental Meditation: Implications for cortical integration theory. Signal Processing, 85, 2213–2232. Huang, H. Y., & Lo, P. C. (2009). EEG dynamics of experienced Zen meditation practitioners probed by complexity index and spectral measure. Journal of Medical Engineering and Technology, 33(4), 314–321. Ishii, R., Shinosaki, K., Ukai, S., Inouye, T., Ishihara, T., Yoshimine, T., et al (1999). Medial prefrontal cortex generates frontal midline theta rhythm. Neuroreport, 10(4), 675–679. Jensen, O., Kaiser, J., & Lachaux, J. P. (2007). Human gamma-frequency oscillations associated with attention and memory. Trends in Neurosciences, 30(7), 317–324. Kelley, W. M., Macrae, C. N., Wyland, C. L., Caglar, S., Inati, S., & Heatherton, T. F. (2002). Finding the self? An event-related fmri study. Journal of Cognitive Neurosciences, 14(5), 785–794. Klimesch, W., Doppelmayr, M., Russegger, H., Pachinger, T., & Schwaiger, J. (1998). Induced alpha band power changes in the human EEG and attention. Neuroscience Letters, 244(2), 73–76. Kounios, J., & Beeman, M. (2009). The aha! moment: The cognitive neuroscience of insight. Current Directions in Psychological Science, 18(4), 210–216. Kubota, Y., Sato, W., Toichi, M., Murai, T., Okada, T., Hayashi, A., et al (2001). Frontal midline theta rhythm is correlated with cardiac autonomic activities during the performance of an attention demanding meditation procedure. Brain Research, Cognitive Brain Research, 11(2), 281–287. Lehmann, D., Faber, P. L., Achermann, P., Jeanmonod, D., Gianotti, L. R., & Pizzagalli, D. (2001). Brain sources of EEG gamma frequency during volitionally meditation-induced, altered states of consciousness, and experience of the self. Psychiatry Research, 108(2), 111–121. Litscher, G., Wenzel, G., Niederwieser, G., & Schwarz, G. (2001). Effects of qigong on brain function. Neurological Research, 23(5), 501–505. Lubar, J. F. (1997). Neocortical dynamics: Implications for understanding the role of neurofeedback and related techniques for the enhancement of attention. Applied Psychophysiol Biofeedback, 22(2), 111–126. Luders, E., Toga, A. W., Lepore, N., & Gaser, C. (2009). The underlying anatomical correlates of long-term meditation: Larger hippocampal and frontal volumes of gray matter. Neuroimage, 45(3), 672–678. Lutz, A., Greischar, L. L., Rawlings, N. B., Ricard, M., & Davidson, R. J. (2004). Long-term meditators self-induce high-amplitude gamma synchrony during mental practice. Proceedings of the National Academy of Science, 101(46), 16369–16373. Lutz, A., Slagter, H. A., Dunne, J. D., & Davidson, R. J. (2008). Attention regulation and monitoring in meditation. Trends in Cognitive Sciences, 12(4), 163–169. Maharishi Mahesh Yogi (1969). Maharishi Mahesh Yogi on the bhagavad gita. New York: Penguin. Maharishi Mahesh Yogi (1997). Celebrating perfection in education (2nd ed.). Noida, India: Maharishi Vedic University Press. Mizuhara, H., & Yamaguchi, Y. (2007). Human cortical circuits for central executive function emerge by theta phase synchronization. Neuroimage, 36(1), 232–244. Murata, T., Koshino, Y., & Ormari, M. (1994). Quantitative EEG study on Zen meditation (zazen). Japanese Journal of Psychiatry and Neurology, 48, 881–890. Murata, T., Takahashi, T., Hamada, T., Omori, M., Kosaka, H., Yoshida, H., et al (2004). Individual trait anxiety levels characterizing the properties of Zen meditation. Neuropsychobiology, 50(2), 189–194. Niessing, J., Ebisch, B., Schmidt, K. E., Niessing, M., Singer, W., & Galuske, R. A. (2005). Hemodynamic signals correlate tightly with synchronized gamma oscillations. Science, 309(5736), 948–951. Oakes, T. R., Pizzagalli, D. A., Hendrick, A. M., Horras, K. A., Larson, C. L., Abercrombie, H. C., et al (2004). Functional coupling of simultaneous electrical and metabolic activity in the human brain. Human Brain Mapping, 21(4), 257–270. 1118 F. Travis, J. Shear / Consciousness and Cognition 19 (2010) 1110–1118 Palva, S., & Palva, J. M. (2007). New vistas for alpha-frequency band oscillations. Trends in Neuroscience, 30(4), 150–158. Palva, J. M., Palva, S., & Kaila, K. (2005). Phase synchrony among neuronal oscillations in the human cortex. Journal of Neuroscience, 25(15), 3962–3972. Pan, W., Zhang, L., & Xia, Y. (1994). The difference in EEG theta waves between concentrative and non-concentrative qigong states – A power spectrum and topographic mapping study. Journal of Traditional Chinese Medicine, 14(3), 212–218. Pfurtscheller, G., Stancak, A., Jr., & Neuper, C. (1996). Event-related synchronization (ers) in the alpha band–an electrophysiological correlate of cortical idling: A review. International Journal of Psychophysiology, 24(1–2), 39–46. Qin, Z., Jin, Y., Lin, S., & Hermanowicz, N. S. (2009). A forty-five year follow-up EEG study of qigong practice. International Journal of Neuroscience, 119(4), 538–552. Raffone, A., & Srinivasan, N. (2009). The exploration of meditation in the neuroscience of attention and consciousness. Cognitive Process. Raichle, M. E., & Snyder, A. Z. (2007). A default mode of brain function: A brief history of an evolving idea. Neuroimage, 37(4), 1083–1090 [discussion 1097– 1089]. Razumnikova, O. M. (2007). Creativity related cortex activity in the remote associates task. Brain Research Bulletin, 73(1–3), 96–102. Rihs, T. A., Michel, C. M., & Thut, G. (2007). Mechanisms of selective inhibition in visual spatial attention are indexed by alpha-band EEG synchronization. European Journal of Neuroscience, 25(2), 603–610. Salinas, E., & Sejnowski, T. J. (2001). Correlated neuronal activity and the flow of neural information. Nature Reviews Neuroscience, 2(8), 539–550. Sarnthein, J., Petsche, H., Rappelsberger, P., Shaw, G. L., & von Stein, A. (1998). Synchronization between prefrontal and posterior association cortex during human working memory. Proceedings of the National Academy of Science, 95(12), 7092–7096. Sauseng, P., Klimesch, W., Gerloff, C., & Hummel, F. C. (2009). Spontaneous locally restricted EEG alpha activity determines cortical excitability in the motor cortex. Neuropsychologia, 47(1), 284–288. Schneider, W., Pimm-Smith, M., & Worden, M. (1994). Neurobiology of attention and automaticity. Current Opinion Neurobiology, 4(2), 177–182. Shaw, J. C. (1996). Intention as a component of the alpha-rhythm response to mental activity. International Journal of Psychophysiology, 24(1–2), 7–23. Shear, J. (2006). The experience of meditation. St. Paul, MN: Paragon House. Singer, W. (1999). Neuronal synchrony: A versatile code for the definition of relations? Neuron, 24(1), 49–65. Steriade, M. (2003). The corticothalamic system in sleep. Front Bioscience, 8, d878–899. Takahashi, T., Murata, T., Hamada, T., Omori, M., Kosaka, H., Kikuchi, M., et al (2005). Changes in EEG and autonomic nervous activity during meditation and their association with personality traits. International Journal of Psychophysiology, 55(2), 199–207. Tei, S., Faber, P. L., Lehmann, D., Tsujiuchi, T., Kumano, H., Pascual-Marqui, R. D., et al (2009). Meditators and non-meditators: EEG source imaging during resting. Brain Topography, 22(3), 158–165. Thatcher, R. W., North, D., & Biver, C. (2005). EEG and intelligence. Relations between EEG coherence, EEG phase delay and power. Clinical Neurophysiology, 116(9), 2129–2141. Thut, G., & Miniussi, C. (2009). New insights into rhythmic brain activity from tms-EEG studies. Trends in Cognitive Sciences, 13(4), 182–189. Travis, F. T. (1991). Eyes open and TM EEG patterns after one and after 8 years of tm practice. Psychophysiology, 28(3a), S58. Travis, F. (1994). A second linked-reference issue: Possible biasing of power and coherence spectra. International Journal of Neuroscience, 75(1–2), 111–117. Travis, F. (2001). Autonomic and EEG patterns distinguish transcending from other experiences during Transcendental Meditation practice. International Journal of Psychophysiology, 42(1), 1–9. Travis, F., & Arenander, A. (2006). Cross-sectional and longitudinal study of effects of Transcendental Meditation practice on interhemispheric frontal asymmetry and frontal coherence. International Journal of Neuroscience, 116(12), 1519–1538. Travis, F., Haaga, D., Hagelin, J., Arenander, A., Tanner, M., & Schneider, R. (2010). Self-referential awareness: Coherence, power, and eloreta patterns during eyes-closed rest, Transcendental Meditation and TM-sidhi practice. Journal of Cognitive Processing, 11(1), 21–30. Travis, F., & Pearson, C. (2000). Pure consciousness: Distinct phenomenological and physiological correlates of ‘‘consciousness itself”. International Journal of Neuroscience, 100, 77–89. Travis, F. T., Tecce, J., Arenander, A., & Wallace, R. K. (2002). Patterns of EEG coherence, power, and contingent negative variation characterize the integration of transcendental and waking states. Biological Psychology, 61, 293–319. Travis, F., Tecce, J. J., & Durchholz, C. (2001). Cortical plasticity, cnv, and transcendent experiences: Replication with subjects reporting permanent transcendental experiences. Psychophysiology, 38, S95. Travis, F., & Wallace, R. K. (1999). Autonomic and EEG patterns during eyes-closed rest and Transcendental Meditation (TM) practice. The basis for a neural model of tm practice. Consciousness and Cognition, 8, 302–318. Tsujimoto, T., Shimazu, H., & Isomura, Y. (2006). Direct recording of theta oscillations in primate prefrontal and anterior cingulate cortices. Journal of Neurophysiology, 95(5), 2987–3000. Varela, F., Lachaux, J. P., Rodriguez, E., & Martinerie, J. (2001). The brainweb: Phase synchronization and large-scale integration. Natural Review of Neuroscience, 2(4), 229–239. Vinogradova, O. S. (2001). Hippocampus as comparator: Role of the two input and two output systems of the hippocampus in selection and registration of information. Hippocampus, 11(5), 578–598. Vogeley, K., Bussfeld, P., Newen, A., Herrmann, S., Happe, F., Falkai, P., et al (2001). Mind reading: Neural mechanisms of theory of mind and self-perspective. Neuroimage, 14(1 Pt 1), 170–181. von Stein, A., Rappelsberger, P., Sarnthein, J., & Petsche, H. (1999). Synchronization between temporal and parietal cortex during multimodal object processing in man. Cerebral Cortex, 9(2), 137–150. von Stein, A., & Sarnthein, J. (2000). Different frequencies for different scales of cortical integration: From local gamma to long range alpha/theta synchronization. International Journal of Psychophysiology, 38(3), 301–313. Wallace, A. (1999). The Buddhist tradition of samatha: Methods for refining and examining consciousness. Journal of Consciousness Studies, 6, 175–187. Whitham, E. M., Pope, K. J., Fitzgibbon, S. P., Lewis, T., Clark, C. R., Loveless, S., et al (2007). Scalp electrical recording during paralysis: Quantitative evidence that EEG frequencies above 20 Hz are contaminated by EMG. Clinical Neurophysiology, 118(8), 1877–1888. Wrobel, A. (2000). Beta activity: A carrier for visual attention. Acta Neurobiologia Experimentalis, 60, 247–260.

Consciousness and Cognition Consciousness and Cognition 14 (2005) 633–640 www.elsevier.com/locate/concog Split-brain reveals separate but equal self-recognition in the two cerebral hemispheres Lucina Q. Uddin *, Jan Rayman, Eran Zaidel Department of Psychology, University of California Los Angeles, Los Angeles, CA, USA Received 9 November 2004 Available online 10 March 2005 Abstract To assess the ability of the disconnected cerebral hemispheres to recognize images of the self, a splitbrain patient (an individual who underwent complete cerebral commissurotomy to relieve intractable epilepsy) was tested using morphed self-face images presented to one visual hemifield (projecting to one hemisphere) at a time while making ‘‘self/other’’ judgments. The performance of the right and left hemispheres of this patient as assessed by a signal detection method was not significantly different, though a measure of bias did reveal hemispheric differences. The right and left hemispheres of this patient independently and equally possessed the ability to self-recognize, but only the right hemisphere could successfully recognize familiar others. This supports a modular concept of self-recognition and other-recognition, separately present in each cerebral hemisphere. Ó 2005 Elsevier Inc. All rights reserved. Keywords: Laterality; Self-awareness; Faces; Hemispheric specialization 1. Introduction Hemispheric specialization of self-recognition has become critical evidence for the self as a distinct cognitive construct. The ability to recognize oneÕs own face has often been used as an opera* Corresponding author. Fax: +1 310 206 3655. E-mail address: lucina@ucla.edu (L.Q. Uddin). 1053-8100/$ - see front matter Ó 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2005.01.008 634 L.Q. Uddin et al. / Consciousness and Cognition 14 (2005) 633–640 tional definition of higher-order self-awareness (Keenan, Wheeler, Gallup, & Pascual-Leone, 2000). Phylogenetic (Gallup, 1970) and ontogenetic (Amsterdam, 1972) trends suggest that the capacity for visual facial self-recognition develops along with a variety of other cognitive functions that are unique to adult humans and to certain non-human primates. The neuroanatomical substrates and possible lateralization of this ability have only recently been explored. Some studies found a right hemisphere (RH) bias or selective activation for self-recognition (Keenan, Nelson, OÕConner, & Pascual-Leone, 2001; Keenan, Wheeler, Platek, Lardi, & Lassonde, 2003; Preilowski, 1977; Uddin, Kaplan, Molnar-Szakacs, Zaidel, & Iacoboni, 2005), others reported a left hemisphere (LH) bias (Turk et al., 2002), and yet others report significant bilateral contributions to the task (Kircher et al., 2001). However, not all of these studies used lateralized stimuli to restrict visual input to one hemifield (Keenan, Freund, Hamilton, Ganis, & Pascual-Leone, 2000; Keenan et al., 2001; Kircher et al., 2001). Further, some of these reports have presumed that responses with the right or left hand are sufficient to reflect contralateral hemispheric performance (Keenan et al., 2000), though data to the contrary exist (Weems, 2003; Zaidel, 1998). Finally, some studies failed to distinguish decision bias and sensitivity in self-recognition (Turk et al., 2002). Some inconsistencies in the literature are due to conceptual issues. For example, one can use split-brain patients to test theories of positive hemispheric competence, while functional imaging studies in normal subjects can shed light on which hemisphere is more involved during a given task. Together, these kinds of studies can address the issue of whether a hemisphere is necessary and/or sufficient for a particular task. Here we redress these methodological concerns to more rigorously investigate the claim that the ability to self-recognize is lateralized to one hemisphere. The present study investigated the positive competence of the two hemispheres on a task of visual self-recognition by assessing the sensitivity and bias of NG, a split-brain patient. 2. Materials and methods 2.1. Participant NG is a 70-year-old woman who underwent complete forebrain commissurotomy (single stage midline section of the anterior commissure, corpus callosum, hippocampal commissure, and massa intermedia) in 1963 to minimize the spread of epileptic seizures. The patientÕs case history is documented elsewhere (Bogen & Vogel, 1976). The experimental protocol was approved by the UCLA general campus IRB. Consent from the patient was secured prior to the study, and she was compensated for her participation. 2.2. Stimuli and procedure Pictures of both the subject and a highly familiar gender-matched associate known by the subject for over 30 years were taken with a Kodak 3400C digital camera. Morphed face images were generated using software called Morph Editor (SoftKey, Cambridge, MA), which allows the experimenter to define the percent of image 1 (based on luminance) to overlay onto image 2. Alignment between faces is achieved by using at least 40 reference points placed on common landmarks such as the eyes, nose, and mouth, as well as tracing the overall head shape. These points L.Q. Uddin et al. / Consciousness and Cognition 14 (2005) 633–640 635 guide the shape of both internal facial features and external head size. Morphed images were created in 5% increments, such that there were 21 self-to-unknown morphs and 21 familiar-to-unknown morphs. These were then converted to a different format using Graphic Converter (Lemke Software, Germany) and given a uniform gray background with photo retouching software (Adobe Photoshop 5.0). The subject was seated 57.3 cm from a high resolution RGB color monitor of a MacIntosh computer, with her chin in a chinrest to ensure consistent viewing. She was instructed to keep her eyes focused on a central fixation cross throughout the experiment. The experiment was run with the software package MacProbe (Hunt, 1994). Morphed images were presented in a random order either to the left or right side of a fixation point on a computer screen for a duration of 180 ms to prevent involuntary saccades and ensure stimulation of the appropriate hemisphere (Zaidel, 1979). In one condition, the subject was asked to press the top button on a response box if the image presented ‘‘looks more like yourself’’ and the bottom button if the image looked more like ‘‘an unknown other face.’’ This condition included images morphed between the subject and an unknown, gender-matched face. Another condition consisted of viewing morphs between the subjectÕs familiar associate, DZ, and an unknown, gender-matched face. Here the subject was asked to press the top button if the image presented ‘‘looks more like DZ’’ and the bottom if the image looked more like ‘‘an unknown other face.’’ This condition was used as a control for familiarity and overlearnedness (Fig. 1B). The subject was tested over six sessions, with each morphed image presented four times/visual field per session. For the first four testing sessions, the subject responded unimanually. Response hand and condition were counterbalanced, such that the subject alternated between starting a session with the right or left hand and doing the ‘‘self’’ or ‘‘other’’ block in alternate sessions. For the last two sessions, the subject was instructed to respond with the left hand if the stimulus appeared in the left visual field (LVF) and the right hand if the stimulus appeared in the right visual field (RVF). As there was no main effect of session, final analyses collapsed data across all sessions, utilizing only responses made in the ‘‘pure hemispheric’’ conditions where stimuli and response hand were congruent (e.g., LVF stimulus, left-hand response; RVF stimulus, right-hand response). The pairing of ‘‘crossed’’ response hand/visual field conditions (e.g., LVF stimulus, right-hand response) introduces noise due to resource limitations resulting from interference between response programming and decision processes (Zaidel, White, Sakurai, & Banks, 1988). Consequently, one routinely excludes ‘‘crossed’’ trials when assessing independent hemispheric contributions. Illustrative examples of the stimuli and task are presented in Figs. 1A and B. Note that the images shown here are example faces, not those actually used in the experiment. 3. Results Data from ‘‘uncrossed’’ or ‘‘pure hemispheric’’ responses (responses where response hand was congruent to VF of stimulus presentation) were analyzed using a hierarchical v2 (Winer, Brown, & Michels, 1991). The hierarchical v2 was a four-way, mixed model v2, with response (yes/no) random and all other factors fixed. Additionally, signal detection analysis was used to compute d 0 as a measure of sensitivity for detecting ‘‘self’’ or ‘‘familiar’’ stimuli, independent of bias (McMichol, 1972). 636 L.Q. Uddin et al. / Consciousness and Cognition 14 (2005) 633–640 Fig. 1. (A) Morphed images between the subject and an unfamiliar face (‘‘self’’ condition) and between a highly familiar associate and an unfamiliar face (‘‘familiar’’ condition) were created in 5% increments. (B) Images were randomly flashed to the right or left of a central fixation cross for 180 ms while the subject made ‘‘self’’/(‘‘familiar’’) or ‘‘other’’ judgments. Analysis revealed an overall bias to affirm the presence of the target face, regardless of condition (54.7%, v2ð1Þ ¼ 11:88, p < .001). Overall, this affirmation bias was stronger in the LH (64.4%, v2ð1Þ ¼ 51:04). As shown in Fig. 2, the subject affirmed the presence of the target face more frequently in the ‘‘familiar’’ condition (64.2%) than in the ‘‘self’’ condition (45.1%, v2ð1Þ ¼ 49:98, p < .001), but only when the stimulus was in the RVF (82%, v2ð1Þ ¼ 38:81, p < .001). The effect of morph interval differed for the two tasks (v2ð20Þ ¼ 42:78, p < .005). The percentage of affirmative responses to both LVF and RVF trials in the ‘‘self’’ condition followed an orderly descent as the stimuli contained less ‘‘self’’ and more ‘‘other’’ (Fig. 3A). In the ‘‘familiar’’ condition, this orderly progression was present for LVF stimuli, but absent for RVF stimuli, producing a Task, VF, and Morph Interval interaction (v2ð20Þ ¼ 39:05, p < .01, Fig. 3B). Signal detection methods were used to obtain d 0 , a measure of subjectsÕ sensitivity to detecting targets independent of bias. These calculations assume the underlying signal and signal + noise distributions are normal and of equal variance. Here, we used d 0 as a measure of sensitivity to detection of ‘‘self’’ and ‘‘familiar’’ stimuli. Images morphed towards ‘‘other’’ more than 50% were designated as non-target or target-absent trials. A response was considered a ‘‘target-present’’ response if the subject responded ‘‘self’’ in the self condition or ‘‘familiar’’ in the familiar condition. Hits were defined as ‘‘target-present’’ responses to target-present trials. False alarms were defined L.Q. Uddin et al. / Consciousness and Cognition 14 (2005) 633–640 637 Fig. 2. The target was either a picture of NGÕs own face (‘‘self’’ condition) or of the familiar associate DZ (‘‘familiar’’ condition). NG responded ‘‘yes’’ to the presence of the target more frequently in the ‘‘familiar’’ task than in the ‘‘self’’ task, but only when responding to stimuli in the left hemisphere. Fig. 3. (A) Proportion of ‘‘self’’ responses as a function of the percentage of NGÕs image contained in the stimulus and the cerebral hemisphere that perceived the image and made the response. Both hemispheres were able to perform this task, though the left hemisphere showed more of a ‘‘self’’ bias. (B) Proportion of ‘‘familiar’’ responses as a function of the percentage of a familiar associateÕs image contained in the stimulus and the cerebral hemisphere that perceived the image and made the response. Only the right hemisphere of NG could successfully perform this task. as ‘‘target-present’’ responses to target-absent trials. In other words, a response was considered a hit if the subject identified a <50% morph as ‘‘self’’ in the self condition or as ‘‘familiar’’ in the familiar condition. The responses to the 50% morphed images were excluded from the analysis. Thus, the proportion of hits was defined as the proportion of <50% morph stimuli responded to as ‘‘self’’ in the self condition or as ‘‘familiar’’ in the familiar condition. This analysis revealed that both hemispheres were significantly above chance (d 0 > 2, p < 0.05) at performing the task of correctly identifying images predominantly containing self-elements (‘‘self-present’’ targets). However, only the RH (d 0 = 1.67, p < 0.05) was sensitive for detecting ‘‘familiar-present’’ targets; familiar images presented to the LH were not detected above chance (Table 1). 638 L.Q. Uddin et al. / Consciousness and Cognition 14 (2005) 633–640 Table 1 Both hemispheres of NG showed sensitivity for detecting ‘‘self’’ images. In contrast, only the right hemisphere of NG was able to detect images of the familiar associate d0 Self Left hemisphere Right hemisphere a Familiar a 2.277 2.019a 0.6115 1.666a Significant as one-tailed z scores, p < 0.05. 4. Discussion Previous work has suggested a double dissociation wherein the disconnected LH shows a recognition bias for self, and the RH shows a bias for familiar others (Turk et al., 2002) or vice versa (Keenan et al., 2003). Instead, our results show that the LH exhibits a selective ‘‘target-present’’ bias in the ‘‘familiar’’ recognition task. We interpret this pattern as an inability of the disconnected LH of patient NG to perform the ‘‘familiar’’ recognition task. In fact, the LH of NG demonstrates a high false alarm rate in the ‘‘familiar’’ task. That is a characteristic strategy that this patient has used when unable to perform a cognitively challenging task (Iacoboni, Rayman, & Zaidel, 1996). By contrast, it appears that the RH of NG can perform both the ‘‘self’’ and ‘‘familiar’’ recognition tasks. The previously reported selective hemispheric bias for self is not supported in our data. Our findings do suggest a selective deficit in the LH for detecting familiar face stimuli. Indeed, we see no evidence for hemispheric specialization of the capacity for self-recognition in this patient. Our finding that both cerebral hemispheres of this patient equally possess the capacity for self-recognition is consistent with previous reports of independent and similar self-awareness in the two disconnected hemispheres (Sperry, Zaidel, & Zaidel, 1979). On the basis of the case study of this single subject alone, we hesitate to generalize. However, an identical hemifield presentation study in 40 normal subjects also revealed no difference in sensitivity to detection of ‘‘self’’ faces for LVF (mean d 0 = 2.556) and RVF (mean d 0 = 2.574) presentations (F = .094, p = .761), suggesting that this finding may apply to a more general population (Uddin, Keenan, Mooshagian, Rayman, & Zaidel, 2002). The remarkable fact that NGÕs disconnected LH was able to recognize the (100%) unknown other face in the ‘‘self’’ condition but unable to do so in the ‘‘familiar’’ condition is further evidence that the self is a distinct cognitive construct. We suggest that the representation of the self that allows for self-recognition is not restricted to a particular hemisphere, but is rather available to each cerebral hemisphere independently. There is an alternative interpretation of these data, which cannot be completely ruled out at present. It is possible that in distinguishing the self from an unknown, mere familiarity is sufficient. Thus the RH, which is able to recognize both the familiar face and the self-face may just be performing the task based on familiarity. The LH, on the other hand, is unable to recognize the familiar face, yet recognizes the self-face, thus it cannot be using familiarity alone. This interpretation suggests that only the LH truly recognizes the self, while the RH may be able to complete the task by other means. It should be acknowledged that previous research has suggested a special role for the RH in novelty detection (Martin, 1999) and that novel stimuli can be regarded as unfamiliar. Thus, novelty detection can be the basis for familiarity detection. However, if only novelty detection were operating during both tasks, with the ‘‘familiar’’ task being just a more L.Q. Uddin et al. / Consciousness and Cognition 14 (2005) 633–640 639 difficult novelty detection task, we would not expect the LH to perform qualitatively differently in the two tasks. We suggest that this familiarity/novelty interpretation is unlikely, based on previous data on this patient. NG has previously shown appropriate emotional reactions when pictures of herself were introduced unexpectedly amongst test items shown to the RH (Sperry et al., 1979). Moreover, running an experiment that requires the subject to directly distinguish the self from the familiar other face may still not adequately resolve this objection. It is likely that the self face is more familiar to the subject than any familiar other, so that the classification could be made by degree of familiarity alone. Further careful studies are required to address the confound of familiarity. While Sperry et al. (1979) report the presence of a similar and well developed self-concept in both hemispheres, other work with commissurotomy patients further suggests different cognitive and emotional characteristics, with respect to the self, present in the two disconnected hemispheres (Schiffer, Zaidel, Bogen, & Chasan-Taber, 1998). We report here the independent ability of each cerebral hemisphere to distinguish self and other, but make no claims as to the nature of the underlying self-concept which gives rise to self-recognition. When the abilities of the disconnected cerebral hemispheres are independently assessed, no differences with respect to detection of self-images are seen. Previous reports of hemispheric specialization for self-recognition in both split-brain patients and normals have primarily used measures of bias, questionable means of ascertaining behavioral laterality, and centrally presented visual stimuli. When these methodological concerns are addressed, indeed we find no evidence for hemispheric differences in the ability to self-recognize. Our findings suggest that the cognitive capacity that gives rise to visual facial self-recognition is robustly represented in both cerebral hemispheres. Acknowledgments This work was supported by National Institutes of Health Grant RO1 NS20187 and a National Science Foundation Graduate Research Fellowship. The authors thank Julian Keenan for providing the MorphEditor software, Eric Mooshagian for help with programming, Karin Foerde for her role in construction of stimuli, and Marco Iacoboni, Joseph Bogen, Bernard Baars, and Bruno Preilowski for helpful comments and discussions. References Amsterdam, B. (1972). Mirror self-image reactions before age two. Developmental Psychobiology, 5, 297–305. Bogen, J. E., & Vogel, P. J. (1976). Neurological status in the long term following complete cerebral commissurotomy. In B. Schott (Ed.), Les syndromes de disconnexion calleuse chez lÕhomme (pp. 227–251). Lyon: Hopital Neurologique. Gallup, G. G. (1970). Chimpanzees: Self-recognition. Science, 167, 86–87. Hunt, S. M. J. (1994). MacProbe: A Macintosh-based experimenterÕs workstation for the cognitive sciences. Behavior Research Methods, Instruments and Computers, 26(3), 345–351. Iacoboni, M., Rayman, J., & Zaidel, E. (1996). Left brain says yes, right brain says no: Normative duality in the split brain. In S. R. Hameroff, A. W. Kaszniak, & A. C. Scott (Eds.), Toward a science of consciousness: The first Tucson discussions and debates (pp. 197–202). Cambridge: MIT Press. Keenan, J. P., Freund, S., Hamilton, R. H., Ganis, G., & Pascual-Leone, A. (2000). Hand response differences in a selfface identification task. Neuropsychologia, 38(7), 1047–1053. 640 L.Q. Uddin et al. / Consciousness and Cognition 14 (2005) 633–640 Keenan, J. P., Nelson, A., OÕConner, M., & Pascual-Leone, A. (2001). Self-recognition and the right hemisphere. Nature, 409, 305. Keenan, J. P., Wheeler, M., Platek, S. M., Lardi, G., & Lassonde, M. (2003). Self-face processing in a callosotomy patient. European Journal of Neuroscience, 18(8), 2391–2395. Keenan, J. P., Wheeler, M. A., Gallup, G. G., Jr., & Pascual-Leone, A. (2000). Self-recognition and the right prefrontal cortex. Trends in Cognition Science, 4(9), 338–344. Kircher, T. T., Senior, C., Phillips, M. L., Rabe-Hesketh, S., Benson, P. J., Bullmore, E. T., et al. (2001). Recognizing oneÕs own face. Cognition, 78(1), B1–B15. Martin, A. (1999). Automatic activation of the medial temporal lobe during encoding: Lateralized influences of meaning and novelty. Hippocampus, 9(1), 62–70. McMichol, D. (1972). A primer on signal detection theory. London: Allen and Unwin. Preilowski, B. (1977). Self-recognition as a test of consciousness in left and right hemisphere of split-brain patients. Activitas Nervosa Superior (Praha), 19(Suppl. 2), 343–344. Schiffer, F., Zaidel, E., Bogen, J., & Chasan-Taber, S. (1998). Different psychological status in the two hemispheres of two split-brain patients. Neuropsychiatry Neuropsychology and Behavioral Neurology, 11(3), 151–156. Sperry, R. W., Zaidel, E., & Zaidel, D. (1979). Self recognition and social awareness in the deconnected minor hemisphere. Neuropsychologia, 17(2), 153–166. Turk, D. J., Heatherton, T. F., Kelley, W. M., Funnell, M. G., Gazzaniga, M. S., & Macrae, C. N. (2002). Mike or me? Self-recognition in a split-brain patient. Nature Neuroscience, 5(9), 841–842. Uddin, L. Q., Kaplan, J. T., Molnar-Szakacs, I., Zaidel, E., & Iacoboni, M. (2005). Self-face recognition activates a frontoparietal mirror network in the right hemisphere: An event-related fMRI study. Neuroimage, in press. Uddin, L. Q., Keenan, J. P., Mooshagian, E., Rayman, J., & Zaidel, E. (2002). Self-recognition in the two cerebral hemispheres. Annual Meeting of the Society for Neuroscience. Weems, S. (2003). Behavioral and electrophysiological examination of lexical processing in the two cerebral hemispheres. Unpublished Dissertation. University of California, Los Angeles, Los Angeles. Winer, J. J., Brown, D. R., & Michels, K. M. (1991). Statistical principles in experimental design (34th ed.). New York: McGraw Hill. Zaidel, E. (1979). On measuring hemispheric specialization in man. In B. Rybak (Ed.), Advanced Technobiology, pp. 365–404. Zaidel, E. (1998). Stereognosis in the chronic split brain: Hemispheric differences, ipsilateral control and sensory integration across the midline. Neuropsychologia, 36(10), 1033–1047. Zaidel, E., White, H., Sakurai, E., & Banks, W. (1988). Hemispheric locus of lexical congruity effects: Neuropsychological reinterpretation of psycholinguistic results. In C. Chiarello (Ed.), Right hemisphere contributions to lexical semantics (pp. 71–88). New York: Springer.

Consciousness and Cognition 19 (2010) 1135–1137 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Commentary Ceiling effects make Hughes and Nicholson’s data analyses and conclusions inconclusive q Bob Uttl ⇑, Alain Morin Department of Psychology, Mount Royal University, 4825 Mount Royal Gate S.W., Calgary, Alberta, Canada T3E 6K6 a r t i c l e i n f o Article history: Available online 11 June 2010 Keywords: Self-recognition Face Voice Hemispheric specialization Laterality Ceiling effects a b s t r a c t Hughes and Nicholson (2010) suggest that recognizing oneself is easier from face vs. voice stimuli, that a combined presentation of face and voice actually inhibits self-recognition relative to presentation of face or voice alone, that the left hemisphere is superior in self-recognition to the right hemisphere, and that recognizing self requires more effort than recognizing others. A re-examination of their method, data, and analyses unfortunately shows important ceiling effects that cast doubts on these conclusions. Ó 2010 Elsevier Inc. All rights reserved. Hughes and Nicholson (2010) present a new study supposedly demonstrating that recognizing oneself is easier from face vs. voice stimuli, that a combined presentation of face and voice actually inhibits self-recognition relative to presentation of face or voice alone, that the left hemisphere is superior in self-recognition to the right hemisphere, and that recognizing self requires more effort than recognizing others. We re-examined Hughes and Nicholson’s method, data, and analyses and were forced to conclude that severe ceiling effects, inappropriate analyses, and methodological confounds render Hughes and Nicholson’s conclusions unwarranted. Hughes and Nicholson (2010) devote several pages to reviewing the background literature but the review appears superficial. For example, the authors discuss an eclectic collection of studies on self-recognition, conclude that the right hemisphere is dominant for self-recognition of one’s face, and propose that the right hemisphere may also be dominant for self-recognition of one’s voice. The authors, however, did not consider much of the evidence showing no such right hemisphere dominance in self-recognition (see Morin, in press) as well as in other forms of self-awareness such as autobiographical memory and emotion awareness (Gillihan & Farah, 2005; Northoff et al., 2006; Phan, Wager, Taylor & Liberzon; 2004; Svoboda, McKinnon, & Levine, 2006). Even more critically, the authors appear to had missed Rosa, Lassonde, Pinard, Keenan, and Belin’s (2008) conclusion that the ceiling effects in their first experiment were responsible for their null findings, and in turn, required them to design a new experiment that would avoid ceiling effects by using a more difficult recognition task. As we demonstrate below, by missing the lesson learned by Rosa et al., Hughes and Nicholson set themselves up for the same predicament: the ceiling effects in their first experiment render their data non-interpretable and their conclusions unwarranted. Hughes and Nicholson’s (2010) key data are shown in Figs. 1 and 2. The figures show participants’ accuracy in deciding whether stimuli depict or do not depict themselves when faces and voices are presented separately (Fig. 1) and when they q Commentary on Hughes, S.M., & Nicholson, S.E. (2010). The processing of auditory and visual recognition of self stimuli. Consciousness and Cognition, 19, 1124–1134. ⇑ Corresponding author. Fax: +1 403 440 7027. E-mail address: buttl@mtroyal.ca (B. Uttl). 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.05.012 1136 B. Uttl, A. Morin / Consciousness and Cognition 19 (2010) 1135–1137 are presented simultaneously (Fig. 2). Our own Figs. 1 and 2 present the exact same data as shown in Hughes and Nicholson’s Figs. 1 and 2, respectively, but with y-axis ranging from 0% to 100% and the bars depicting one SDs (directly reported in the text or calculated from SEs reported by Hughes and Nicholson). The replotted figures highlight the severe ceiling effects in all but one condition (Fig. 1, Self Voice/Right Hand) using one SD criteria for detecting ceiling effects, and in all conditions using 1.5 SD criteria (Uttl, 2005). Accordingly, the means do not reflect participants’ ability to decide whether stimuli do or do not represent them but primarily ceiling effects, and, similarly, the SDs do not reflect true individual variability in recognition abilities but are merely an artifact of ceiling effects (see Uttl, 2005). Second, due to ceiling effects, the ratios between the largest and the smallest SD within each figure’s data are huge and the corresponding ratio between the smallest and the largest variances is 285 for the data in Fig. 1 and 625 for the data in Fig. 1. Participants accuracy (means with bars indicating one standard deviation) in deciding whether stimuli depict or do not depict themselves when faces and voices were presented separately. Fig. 2. Participants accuracy (means with bars indicating one standard deviation) in deciding whether stimuli depict or do not depict themselves when faces and voices were presented simultaneously. B. Uttl, A. Morin / Consciousness and Cognition 19 (2010) 1135–1137 1137 Fig. 2. Moreover, correlations between condition means and condition standard deviations highlight that the high means are associated with low standard deviation (r = .85 for the data in Fig. 1 and r = .99 for the data in Fig. 2). In turn, these important differences in variability of scores across the experimental conditions and the strong association between the condition means and their standard deviations invalidate the parametric analyses (i.e., ANOVAs) reported by Hughes and Nicholson (see, for example, Howell, 2007, for a comprehensive discussion of assumptions that need to be satisfied to analyze the data using ANOVA). In combination, the ceiling effects and invalid parametric analyses of the ceiling limited data render Hughes and Nicholson’s key conclusions unwarranted. Hughes and Nicholson’s data tell us nothing about participants’ abilities except that the task was so easy that many participants (in some conditions nearly all of them) obtained perfect scores in all or nearly all conditions. The data and the analyses provide no conclusive evidence that ‘‘visual self-recognition processing is superior to auditory self-recognition processing’’ that ‘‘a cross-modal exposure to self stimuli . . . inhibits self-recognition’’ that ‘‘there may be lateralization effects for vocal self-recognition alone and when combined with visual information’’ and that ‘‘selfinformation required more effort to process than information about others’’. Although Hughes and Nicholson (2010) acknowledged that participants’ voice recognition accuracy was very high relative to the previous studies, they noted that their very high means were ‘‘consistent’’ with the very high means recently reported by Rosa et al. (2008). If they read Rosa et al.’s article, however, they learned that Rosa et al. (2008) concluded that ‘‘this [their] task was too easy and that the results were confounded by a ceiling effect.’’ and ‘‘decided to design a second experiment’’ (p. 207). We think Hughes and Nicholson ought to have followed Rosa et al.’s example and also design a second experiment that would avoid the widespread ceiling effects that render their study difficult to interpret. The ceiling effects problem aside, the authors make several puzzling claims. First, they claim that ‘‘participants were overall much more accurate at identifying stimuli presented that was not themselves versus when it was, suggesting it may take more effort to process self-information’’. However, the participants’ task was to decide whether each stimuli was of self vs. not of self, and thus, participants never had to identify stimuli of others as those of specific people, for example, Joe or Jane. Arriving to a decision that a stimulus is vs. is not of a particular person is more difficult and requires more extensive cognitive processing. Even if the data were not limited by the ceiling effects and even if they were interpretable, the authors could only conclude that participants were more accurate in deciding that stimuli were not representing them than deciding that stimuli were representing them. Second, the authors claim that ‘‘auditory-stimuli presented concomitantly with visual self-stimuli inhibited performance and reaction times on the self-recognition tasks’’. Again, arriving to a decision that both voice and face are of a particular person is much more difficult and requires more cognitive processing than a decision that a face or a voice is not of a particular person (e.g., the first feature that does not fit allows me to conclude it is not me). Thus, one task requires processing of more information than the other task and the inhibition/interference interpretation is unwarranted. In conclusion, Hughes and Nicholson’s claims are largely unwarranted due to the ceiling limited data, invalid parametric analyses of the ceiling limited data, and the failure to recognize that the participants’ task did not at all require identification of specific others. References Gillihan, S. J., & Farah, M. J. (2005). Is self special? A critical review of evidence from experimental psychology and cognitive neuroscience. Psychological Bulletin, 131, 76–97. Howell, D. C. (2007). Statistical methods for psychology. Belmont, CA: Thomson-Wadsworth. Hughes, S. M., & Nicholson, S. E. (2010). The processing of auditory and visual recognition of self stimuli. Consciousness and Cognition, 19, 1124–1134. Morin, A. (in press). Self-recognition, theory-of-mind, and self-awareness: What side are you on? Laterality. Northoff, G., Heinzel, A., de Greck, M., Bermpohl, F., Dobrowolny, H., & Panksepp, J. (2006). Self-referential processing in our brain: A meta-analysis of imaging studies on the self. NeuroImage, 31, 440–457. Phan, K. L., Wager, T. D., Taylor, S. F., & Liberzon, I. (2004). Functional neuroimaging studies of human emotions. CNS Spectrums, 9(4), 258–266. Rosa, C., Lassonde, M., Pinard, C., Keenan, J. P., & Belin, P. (2008). Investigation of hemispheric specialization of self-voice recognition. Brain and Cognition, 68, 204–214. Svoboda, E., McKinnon, M. C., & Levine, B. (2006). The functional neuroanatomy of autobiographical memory: A meta-analysis. Neuropsychologia, 44, 2189–2208. Uttl, B. (2005). Measurement of individual differences: Lessons from memory assessment in research and clinical practice. Psychological Science, 16, 460–467.

Consciousness and Cognition 43 (2016) 128–132 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Review article Attentional bias to pain-relevant body locations: New methods, new challenges Stefaan Van Damme ⇑, Charlotte Vanden Bulcke, Wouter Durnez, Geert Crombez Department of Experimental-Clinical and Health Psychology, Faculty of Psychology and Educational Sciences, Ghent University, Henri Dunantlaan 2, 9000 Ghent, Belgium a r t i c l e i n f o Article history: Received 18 May 2016 Accepted 21 May 2016 Keywords: Attention Temporal Order Judgment Experimental pain Bodily threat a b s t r a c t In a recent issue of Consciousness and Cognition, Filbrich, Torta, Vanderclausen, Azanon, and Legrain (2016) commented on a paper in which we used a tactile Temporal Order Judgment (TOJ) task to show that expecting pain on a specific body location biased attention to that location (Vanden Bulcke, Crombez, Durnez, & Van Damme, 2015). Their main criticism is that the effects are likely to reflect response bias rather than genuine attentional bias. We agree that the TOJ task used may be susceptible to response bias, and welcome the authors’ methodological suggestions to control for such bias. However, we feel that certain aspects of our work are misrepresented in their paper. Most importantly, we contest their argument that our instructions made the threat location task-relevant, thereby increasing risk of response bias. Further, we reply to other methodological and theoretical issues raised by these authors. Ó 2016 Published by Elsevier Inc. Contents Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 In a recent commentary on the use of Temporal Order Judgment (TOJ) tasks to investigate pain-related attentional bias, Filbrich, Torta, Vanderclausen, Azanon, and Legrain (2016) discussed a study we have published in this domain (Vanden Bulcke, Crombez, Durnez, & Van Damme, 2015). We agree with some of the issues raised, and welcome critical appraisal and constructive debate, which eventually will advance the field. Nevertheless, we wish to reply to some of their comments, which we feel, misrepresent aspects of our work. Before focusing on these comments, we will first briefly discuss the theoretical background and aims of the conducted studies. ⇑ Corresponding author. E-mail address: Stefaan.Vandamme@UGent.be (S. Van Damme). http://dx.doi.org/10.1016/j.concog.2016.05.012 1053-8100/Ó 2016 Published by Elsevier Inc. S. Van Damme et al. / Consciousness and Cognition 43 (2016) 128–132 129 Biases in processing of pain-related information, such as excessive focus of attention on somatosensory signals, have been proposed as predisposing and/or maintaining factors in pain-related disability and distress in several models of pain suffering (Crombez, Van Damme, & Eccleston, 2005; Legrain et al., 2009; Pincus & Morley, 2001; Todd et al., 2015; Van Damme, Legrain, Vogt, & Crombez, 2010; Vlaeyen & Linton, 2012). While this idea has attracted significant research effort, it is worth noting that the study of pain-related attentional bias has long been dominated by experimental paradigms using visual stimulus material (i.e., pain-related words or pictures). For instance, in a typical study using the dot-probe paradigm, pain-related stimuli (words or pictures) and neutral stimuli are simultaneously presented at different locations of a display, after which one of the stimuli is being replaced by a dot. The reaction time to respond to the location of the dot is measured. Pain-related attentional bias typically results in faster responses when the dot is presented at the ‘‘pain location”. Recent meta-analyses of such studies showed that attentional bias was overall smaller in magnitude than would have been expected, and that its manifestation was dependent upon specific procedural aspects, such as type of stimuli and presentation time of stimuli (Crombez, Van Ryckeghem, Eccleston, & Van Damme, 2013; Schoth, Nunes, & Liossi, 2012). One potential explanation for these underwhelming results is that visual representations of pain may not sufficiently capture bodily threat, and that the adaptive value of bias to (the location of) such stimuli is very limited (Crombez et al., 2013; Van Damme et al., 2010). Consequently, it was recommended to develop new research paradigms implementing actual bodily threats, for example by experimentally inducing pain, and to measure responses to actual somatosensory inputs, to assess pain-related attentional bias. There have been several attempts to develop somatosensory attention paradigms. For instance, Peters and colleagues (Peters, Vlaeyen, & Kunnen, 2002; Peters, Vlaeyen, & van Drunen, 2000) investigated biases attention in chronic pain patients by looking at the detection of electrical stimuli of increasing intensity in combination with a second attention-demanding task. They hypothesized that patients, compared with healthy controls, would display facilitated detection of these somatosensory stimuli. This hypothesis was not confirmed. However, a potential problem with these studies is that participants were explicitly instructed to detect the pain stimuli, which was likely to induce a strong focus on the target locations, thereby possibly wiping out individual differences in somatosensory attention. A possible solution is to design paradigms in which pain stimuli are task-irrelevant, and in which the effect of threat of pain on attention is examined. Indeed, it is likely that attention is specifically biased to body locations that are pain-relevant or threatened. This seems to be confirmed in a study by Crombez and colleagues (Crombez, Eccleston, Baeyens, & Eelen, 1998). They presented mildly painful stimuli on the left and right arm during performance of an auditory task in healthy volunteers, and led participants to falsely believe that stimulation on one of the arms (either left or right) would occasionally be increased to a pain stimulus of high intensity. Especially participants scoring high on catastrophic thinking about pain rated the pain stimuli at the threatened arm as more intense and unpleasant. Of particular interest, in these participants, auditory task performance was most strongly disrupted on trials in which the pain stimulus was administered on the threatened arm, possibly indicating threat-induced attentional bias. Demonstrating pain-related attentional bias in chronic pain patients in this way may, however, be challenging. Although the use of reaction times may be useful in homogenous samples of undergraduate students, it may be less suitable in more heterogeneous clinical samples. Chronic pain populations are often characterized by cognitive dysfunction resulting in overall slowing of and high variability in reaction times (Moriarty, McGuire, & Finn, 2011), which may obscure the typically short-lived and subtle effects of attention (Van Damme, Crombez, & Eccleston, 2002; Van Hulle, Durnez, Crombez, & Van Damme, 2015). In order to avoid this problem, we went out to develop new approaches to assess attentional bias for pain-related body locations. In one such approach, we adopted a TOJ task, which has been previously used to assess bodily threat-related shifts in attention (Moseley, Gallace, & Spence, 2009; Van Damme, Gallace, Spence, Crombez, & Moseley, 2009; Zampini et al., 2007). The TOJ methodology is based upon Titchener’s law of prior entry, stating that attended stimuli come to consciousness more quickly than unattended stimuli (Spence & Parise, 2010). We adapted this methodology to examine if threat of pain at a certain body location biased attention to that location (Vanden Bulcke, Van Damme, Durnez, & Crombez, 2013). Specifically, we asked participants to report which one of two tactile stimuli, one administered to each hand at a range of different stimulus onset asynchronies (SOAs), was perceived first. Performance measures on such task, especially the point of subjective similarity (PSS; virtual SOA at which both stimuli are perceived as occurring simultaneously), may provide information about which hand is attended to (Shore, Gray, Spry, & Spence, 2005). Crucial in this study, participants were informed that the color of the cue preceding each TOJ trial would either signal the possible delivery (threat trials) or absence (neutral trials) of painful electrical stimulus on one hand. Analysis of PSS values indicated that in trials during which pain was expected, there was a shift of attention to the threatened hand (Vanden Bulcke et al., 2013). Follow-up studies replicated this effect, and additionally showed that pain-related bias was not limited to the exact pain location but even generalized to other body parts of the same body half (Vanden Bulcke, Crombez, Spence, & Van Damme, 2014), and that similar effects could be found when using visual instead of tactile TOJ suggesting that prioritization is not limited to somatosensory information (Vanden Bulcke et al., 2015). While our approach, which generated consistent findings over studies, is relatively straightforward, and was maximally aimed at potential application in clinical samples, this may have a drawback in terms of experimental control. A possible criticism of the TOJ task is its susceptibility to response bias. Participants are typically required to make a choice between a ‘‘left-first” or ‘‘right-first” response, with no possibility to report that they did not perceive a temporal difference. However, especially in trials with very short SOAs, perceptually undecided participants could be inclined to respond with the location that was most salient in the experiment, namely the location in which they expected pain. A shift in PSS values could then 130 S. Van Damme et al. / Consciousness and Cognition 43 (2016) 128–132 reflect a bias in the decision process, rather than a genuinely perceptual effect of pain anticipation (for a more extensive discussion of response bias in TOJ tasks, see García-Pérez & Alcalá-Quintana, 2012; Spence & Parise, 2010). Filbrich et al. (2016) argued that it is likely that the instructions used in our studies (Vanden Bulcke et al., 2013, 2014, 2015) have strongly induced such response bias. Specifically, they state that ‘‘participants were explicitly instructed to attend one side of space”, that ‘‘participants reported more often the side of space they had been instructed to attend, that is, the side of the threat”, and that ‘‘participants simply resolved temporal uncertainty by choosing the side of space they expected to be the most taskrelevant” (page 136). We object to these statements, and argue that their suggestion that the pain location in our studies was task-relevant is misleading. In fact, the pain stimulus was task-irrelevant. In none of our studies we have provided any explicit instruction that participants were required to attend to the threatened hand or the side of the threat. We only provided information about which hand could receive pain just before the start of each block. This message was not repeated during performance of the block, and the visual cue preceding each TOJ trial was presented centrally on a display and was spatially uninformative for the task (i.e., indicating which hand was stimulated first). It may be premature to conclude that effects of cueing a body location by means of threat of pain are non-perceptual. In fact, two lines of research using paradigms in which response bias was unlikely, suggest that such effect could be genuinely perceptual. First, recent studies examining the effect of visual cues presented near the left or right hand on a TOJ using nociceptive electrical stimuli, showed that attention was biased to the cued location (De Paepe, Crombez, & Legrain, 2015; De Paepe, Crombez, Spence, & Legrain, 2014). In these studies each TOJ trial was preceded by a lateral cue, making it very likely that participants, in case of uncertainty, would select the cued location. However, to prevent such response bias, the authors included blocks in which participants were asked to indicate ‘‘which is second” instead of ‘‘which is first”. This had no impact at all upon the results, suggesting that the effect was perceptual in nature. Although the design of this study is not identical to the design in our studies, the results at least indicate that cueing a location does not necessarily result in response bias to the cued location. Second, recent studies demonstrated spatial prioritization of a threatened location by means of a tactile change detection (TCD) paradigm (Durnez & Van Damme, 2015; Van Hulle et al., 2015). Participants were requested to judge whether or not they perceived a change between two consequently presented spatial patterns of 3 tactile stimuli on 8 possible body locations. In half of the trials, a painful electrical stimulus could be administered on one of these locations, and this was announced by a cue (threat trials). In the other half of the trials, no threat was induced (neutral trials). Changes in tactile patterns either did or did not involve the threatened body location. Results of both studies showed that, in trials during which pain was expected, change detection was better when it involved the threatened location. It is evident that the response format in this study design (‘‘change” or ‘‘no change”) was orthogonal to the spatial dimension (threat location involved or not), and that results thus cannot be accounted for by response bias. Of course, the results of this study cannot be automatically generalized to the TOJ methodology, but at least they indicate that the effect of expecting pain at a certain body location on responses to stimuli presented at that location could be perceptual in nature. Nevertheless, as is the case in many TOJ studies in different research areas, we cannot exclude that response bias may have contributed to the effects reported in our studies to some extent. We therefore welcome the strategies suggested by Filbrich et al. (2016) to control for this by (a) using a response organization that is orthogonal to the spatial dimension of the stimuli (which should prevent bias in selecting the response option that is equivalent with the experimental manipulation), (b) including blocks in which participants are asked to indicate ‘‘which is second” instead of ‘‘which is first” (which should - in case of response bias - result in reversed effects), and (c) using simultaneity judgments or allowing a third response ‘‘both stimuli came at the same time” (which should prevent participants to guess in case they did not perceive a difference in timing between stimuli). We also welcome studies that demonstrate that our results are owing to response bias. Indeed, as yet it remains a hypothesis, which still needs to be confirmed. The studies by Vanden Bulcke and colleagues represent the first steps in developing a new approach to assess pain-related attentional bias that may be suitable for application in clinical pain populations. One obvious advantage over previous studies is that our approach does not depend on reaction times, which could be problematic in detecting biases in chronic pain patients due to overall slowing and variability. Another potential advantage in the context of research in clinical populations and settings is that the straigtforward setup allows collecting data in one relatively short test session. Admittedly, our studies are not perfect, and methodological refinement requires time and involve progressive insight. There are three further comments on our work by Filbrich et al. (2016) that we would like to reply to. First, referring to the TOJ study of Vanden Bulcke et al. (2015), they challenge our conclusion in terms of a multisensory effect. In that study, it was found that induction of pain biased PSS values not differently when using tactile versus visual TOJ, suggesting that effects of pain anticipation may not be specific for somatosensory processing. The message we intended to convey was that anticipation of pain might also prioritize processing of non-somatosensory stimuli when these are presented at the threatened location. While we discussed that multisensory integration may be in line with such finding, this is not necessarily the same as saying that there was actual cross-modal integration between pain and visual stimuli in this specific study. Indeed, the experiment was never designed to examine such cross-modal integration, but rather as a test of an attentional theory (Legrain et al., 2009; Van Damme et al., 2010). Specifically, we wanted to examine the idea that anticipation of pain may result in activation of ‘attentional control settings’, biasing attention to certain stimulus features that are relevant for adequate reaction to threat. The location where one expects pain to occur is likely to be an important feature, as a result of which all sensory input, irrespective of its modality, may be prioritized when experiencing threat. We agree, though, that our framing in terms of a ‘‘multisensory effect” may be confusing, and that this is best avoided. Note that Filbrich et al. (2016) also questioned our multisensory effect from a statistical perspective. They argued that the effect of threat on the PSS was mainly S. Van Damme et al. / Consciousness and Cognition 43 (2016) 128–132 131 driven by the tactile TOJ condition, because in the visual TOJ condition the values ‘‘were smaller and not significantly different from 0”. This is, however, an invalid interpretation. Two separate tests of which one is significant (tactile TOJ) and the other is non-significant (visual TOJ), do not allow one to conclude that the effects are different, especially when the crucial interaction involving modality is not significant (for an elaborate discussion of this error, see Gelman & Stern, 2006). Second, Filbrich and colleagues suggest that we have incorrectly concluded that our effects are pain-specific. We can only express strong disagreement with their interpretation. In fact, we have consistently and explicitly warned against conclusions in terms of pain-specificity (Van Damme et al., 2002; Vanden Bulcke et al., 2013). Even more, we have been amongst the first to systematically argue that a search for what is unique to pain, has led to an unduly focus on the sensory characteristics of the experience, distracting from the central role of its affective-motivational (Eccleston & Crombez, 1999). We may not confuse two levels of explanation. The first explanation is on a descriptive and operational level: there is no doubt that the delivery of a pain stimulus brings along particular effects. The second explanation is on a mechanistic or mediating level. Here, we hypothesize that the arousal dimensions is critical. In that respect, we believe that if a stimulus is equally arousing (or salient) as a pain stimulus, we will observe equal effects. Indeed, it has been demonstrated that attentional bias to certain stimuli depends on the level of arousal they evoke, irrespective of whether they have a negative or positive valence (Vogt, De Houwer, Crombez, Koster, & Van Damme, 2008). From this perspective, it is not helpful to search for attentional bias effects that are unique to pain. Third, they argue that similar TOJ tasks have been used in patients with complex regional pain syndrome (CRPS). Moseley et al. (2009) found indications for an attentional bias away from the painful body part. The reasoning of Filbrich and colleagues that this finding contradicts our results and questions our hypothesis of increased attention to threatened body parts in chronic pain patients is invalid. On the one hand, the study of Moseley et al. (2009) investigated the effect of the presence of persistent pain in a limb on tactile TOJ in CRPS patients, whereas our studies went out to examine the effects of anticipated phasic pain stimulation in healthy volunteers. The substantial differences in research question, study design, type of participants, and type of pain, render any direct comparison between the results of our study and the study of Moseley et al. (2009) meaningless. Also note that the sample in the study of Moseley et al. (2009) was small (10 patients), and that their interpretation awaits further corroboration. On the other hand, it is well known that CRPS patients are an atypical population that may not be representative for the broader chronic pain population. Specifically, we should be very cautious in generalizing the neglect-like avoidance of the space in which the painful limb usually resides in CRPS patients (Legrain, Bultitude, De Paepe, & Rossetti, 2012) to other chronic pain populations, and to question theoretical assumptions on over-attentiveness that are present in several pain models based upon these CRPS studies. In sum, our work presents a promising, but inevitably imperfect, first step into a new generation of paradigms assessing pain-related attentional bias. Although we regret that certain aspects of our work were misrepresented in the paper by Filbrich and colleagues, we welcome critical discussion, and appreciate some of their methodological recommendations to exclude alternative explanations. Ultimately, we hope that this discussion will result in further methodological progress, and that it will invite new studies advancing the field. Acknowledgements This work was supported by Research Foundation Flanders (FWO; Grant 3G005611) and by Ghent University (Special Research Fund; Grant BOF11/STA/004). References Crombez, G., Eccleston, C., Baeyens, F., & Eelen, P. (1998). When somatic information threatens, catastrophic thinking enhances attentional interference. Pain, 75, 187–198. http://dx.doi.org/10.1016/S0304-3959(97)00219-4. Crombez, G., Van Damme, S., & Eccleston, C. (2005). Hypervigilance to pain: An experimental and clinical analysis. Pain, 116, 4–7. http://dx.doi.org/10.1016/ j.pain.2005.03.035. Crombez, G., Van Ryckeghem, D. M. L., Eccleston, C., & Van Damme, S. (2013). Attentional bias to pain-related information: A meta-analysis. Pain, 154, 497–510. http://dx.doi.org/10.1016/j.pain.2012.11.013. De Paepe, A. L., Crombez, G., & Legrain, V. (2015). From a somatotopic to a spatiotopic frame of reference for the localization of nociceptive stimuli. PLoS ONE, e0137120. http://dx.doi.org/10.1371/journal.pone.0137120. De Paepe, A. L., Crombez, G., Spence, C., & Legrain, V. (2014). Mapping nociceptive stimuli in a peripersonal frame of reference: Evidence from a temporal order judgment task. Neuropsychologia, 56, 219–228. http://dx.doi.org/10.1016/j.neuropsychologia.2014.01.016. Durnez, W., & Van Damme, S. (2015). Trying to fix a painful problem: The impact of pain control attempts on the attentional prioritization of a threatened body location. The Journal of Pain, 16, 135–143. http://dx.doi.org/10.1016/j.jpain.2014.10.012. Eccleston, C., & Crombez, G. (1999). Pain demands attention: A cognitive-affective model of the interruptive function of pain. Psychological Bulletin, 125, 356–366. http://dx.doi.org/10.1037//0033-2909.125.3.356. Filbrich, L., Torta, D. M., Vanderclausen, C., Azanon, E., & Legrain, V. (2016). Using temporal order judgments to investigate attention bias toward pain and threat-related information: Methodological and theoretical issues. Consciousness and Cognition, 41, 135–138. http://dx.doi.org/10.1016/ j.concog.2016.02.008. García-Pérez, M. A., & Alcalá-Quintana, R. (2012). On the discrepant results in synchrony judgment and temporal-order judgment tasks: A quantitative model. Psychonomic Bulletin Review, 19, 820–846. http://dx.doi.org/10.3758/s13423-012-0278-y. Gelman, A., & Stern, H. (2006). The difference between ‘‘significant” and ‘‘non significant” is not itself statistically significant. The American Statistician, 60, 328–331. http://dx.doi.org/10.1198/000313006X152649. Legrain, V., Bultitude, J. H., De Paepe, A. L., & Rossetti, Y. (2012). Pain, body, and space: What do patients with complex regional pain syndrome really neglect? Pain, 156, 948–951. http://dx.doi.org/10.1016/j.pain.2011.12.010. 132 S. Van Damme et al. / Consciousness and Cognition 43 (2016) 128–132 Legrain, V., Van Damme, S., Eccleston, C., Davis, K. D., Seminowicz, D. A., & Crombez, G. (2009). A neurocognitive model of attention to pain: Behavioral and neuroimaging evidence. Pain, 144, 230–232. http://dx.doi.org/10.1016/j.pain.2009.03.020. Moriarty, O., McGuire, B. E., & Finn, D. P. (2011). The effect of pain on cognitive function: A review of clinical and preclinical research. Progress in Neurobiology, 93, 385–404. http://dx.doi.org/10.1016/j.pneurobio.2011.01.002. Moseley, G. L., Gallace, A., & Spence, C. (2009). Space-based, but not arm-based, shift in tactile processing in complex regional pain syndrome and its relationship to cooling of the affected limb. Brain, 132, 3142–3151. http://dx.doi.org/10.1093/brain/awp224. Peters, M. L., Vlaeyen, J. W. S., & Kunnen, A. M. W. (2002). Is pain-related fear a predictor of somatosensory hypervigilance in chronic low back pain patients? Behaviour Research and Therapy, 40, 85–103. http://dx.doi.org/10.1016/S0005-7967(01)00005-5. Peters, M. L., Vlaeyen, J. W. S., & van Drunen, C. (2000). Do fibromyalgia patients display hypervigilance for innocuous somatosensory stimuli? Application of a body scanning reaction time paradigm. Pain, 86, 283–292. http://dx.doi.org/10.1016/S0304-3959(00)00259-1. Pincus, T., & Morley, S. (2001). Cognitive-processing bias in chronic pain: A review and integration. Psychological Bulletin, 127, 599. http://dx.doi.org/ 10.1037//0033-2909.127.5.599. Schoth, D. E., Nunes, V. D., & Liossi, C. (2012). Attentional bias towards pain-related information in chronic pain patients: A meta-analysis of visual-probe investigations. Clinical Psychology Review, 32, 13–25. http://dx.doi.org/10.1016/j.cpr.2011.09.004. Shore, D., Gray, K., Spry, E., & Spence, C. (2005). Spatial modulation of tactile temporal-order judgments. Perception, 34, 1251–1262. http://dx.doi.org/ 10.1068/p3313. Spence, C., & Parise, C. (2010). Prior-entry: A review. Consciousness and Cognition, 19, 364–379. http://dx.doi.org/10.1016/j.concog.2009.12.001. Todd, J., Sharpe, L., Johnson, A., Perry, N. K., Colagiuri, B., & Dear, B. F. (2015). Towards a new model of attentional biases in the development, maintenance, and management of pain. Pain, 156, 1589–1600. http://dx.doi.org/10.1097/j.pain.0000000000000214. Van Damme, S., Crombez, G., & Eccleston, C. (2002). Retarded disengagement from pain cues: The effects of pain catastrophizing and pain expectancy. Pain, 100, 111–118. http://dx.doi.org/10.1016/S0304-3959(02)00290-7. Van Damme, S., Gallace, A., Spence, C., Crombez, G., & Moseley, G. L. (2009). Does the sight of physical threat induce a tactile processing bias? Modalityspecific attentional facilitation induced by viewing threatening pictures. Brain Research, 1253, 100–106. http://dx.doi.org/10.1016/j. brainres.2008.11.072. Van Damme, S., Legrain, V., Vogt, J., & Crombez, G. (2010). Keeping pain in mind: A motivational account of attention to pain. Neuroscience and Biobehavioral Reviews, 34, 204–213. http://dx.doi.org/10.1016/j.neubiorev.2009.01.005. Van Hulle, L., Durnez, W., Crombez, G., & Van Damme, S. (2015). Detection of tactile change on a bodily location where pain is expected. Perceptual and Motor Skills: Perception, 120, 219–231. http://dx.doi.org/10.2466/24.PMS.120v13x1. Vanden Bulcke, C., Crombez, G., Durnez, W., & Van Damme, S. (2015). Is attentional prioritization on a location where pain is expected modality-specific or multisensory? Consciousness and Cognition, 36, 246–255. http://dx.doi.org/10.1016/j.concog.2015.07.003. Vanden Bulcke, C., Crombez, G., Spence, C., & Van Damme, S. (2014). Are the spatial features of bodily threat limited to the exact location where pain is expected? Acta Psychologica, 153, 113–119. http://dx.doi.org/10.1016/j.actpsy.2014.09.014. Vanden Bulcke, C., Van Damme, S., Durnez, W., & Crombez, G. (2013). The anticipation of pain at a specific location of the body prioritizes tactile stimuli at that location. Pain, 154, 1464–1468. http://dx.doi.org/10.1016/j.pain.2013.05.009. Vlaeyen, J. W. S., & Linton, S. J. (2012). Fear-avoidance model of chronic musculoskeletal pain: 12 years on. Pain, 153, 1144–1147. http://dx.doi.org/10.1016/ j.pain.2011.12.009. Vogt, J., De Houwer, J., Crombez, G., Koster, E. H. W., & Van Damme, S. (2008). Allocation of spatial attention to emotional stimuli depends upon arousal and not valence. Emotion, 8, 880–885. http://dx.doi.org/10.1037/a0013981. Zampini, M., Bird, K. S., Bentley, D. E., Watson, A., Barrett, G., Jones, A. K., & Spence, C. (2007). ‘Prior entry’ for pain: Attention speeds the perceptual processing of painful stimuli. Neuroscience Letters, 414, 75–79. http://dx.doi.org/10.1016/j.neulet.2006.12.006.

Consciousness and Cognition 22 (2013) 420–429 Contents lists available at SciVerse ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Attention modulates sensory suppression during back movements Lore Van Hulle a,⇑, Georgiana Juravle b, Charles Spence b, Geert Crombez a, Stefaan Van Damme a a b Department of Experimental-Clinical and Health Psychology, Ghent University, Ghent, Belgium Crossmodal Research Laboratory, Department of Experimental Psychology, Oxford University, Oxford, UK a r t i c l e i n f o Article history: Received 28 June 2012 Available online 1 March 2013 Keywords: Sensory suppression Tactile perception Attention Back movement a b s t r a c t Tactile perception is often impaired during movement. The present study investigated whether such sensory suppression also occurs during back movements, and whether this would be modulated by attention. In two tactile detection experiments, participants simultaneously engaged in a movement task, in which they executed a back-bending movement, and a perceptual task, consisting of the detection of subtle tactile stimuli administered to their upper or lower back. The focus of participants’ attention was manipulated by raising the probability that one of the back locations would be stimulated. The results revealed that tactile detection was suppressed during the execution of the back movements. Furthermore, the results of Experiment 2 revealed that when the stimulus was always presented to the attended location, tactile suppression was substantially reduced, suggesting that sensory suppression can be modulated by top-down attentional processes. The potential of this paradigm for studying tactile information processing in clinical populations is discussed. Ó 2013 Elsevier Inc. All rights reserved. 1. Introduction Bending over to lift your shopping or reaching forward in order to grasp the television remote control are but two examples highlighting the fact that back movements are part of everyday life and are involved in many functional behaviors. In order to make sure that these movements are adequately accomplished and are not constantly interrupted by stimuli that we may become aware of, the mass of sensory information (e.g., tactile, proprioceptive) that is associated with the execution of such movements has to be selectively filtered (Bays & Wolpert, 2007; Gallace, Zeeden, Röder, & Spence, 2010). Relevant here is the finding that the detection of subtle, near-threshold tactile stimuli is impaired during the execution of movement, a phenomenon that has been referred to as sensory suppression (Chapman & Beauchamp, 2006; Juravle, Deubel, & Spence, 2011; Juravle, Deubel, Tan, & Spence, 2010; Juravle & Spence, 2011; Voss, Ingram, Wolpert, & Haggard, 2008; Wasaka, Hoshiyama, Nakata, Nishihira, & Kakigi, 2003; Williams, Shenasa, & Chapman, 1998). The suppression of tactile information seems to be related to the movement of the specific body part where the stimulation happens to be delivered. Previous studies have shown that the effect of sensory suppression decreases as the distance between the site of stimulation and the site of movement increases (Williams et al., 1998; Post, Zompa, & Chapman, 1994). The phenomenon of sensory suppression can be explained both by feed-forward motor signals that predict and modulate the activity evoked by incoming sensory signals, and by re-afferent sensations resulting from body movements, leading to backward masking (for a detailed discussion, see Chapman & Beauchamp, 2006; Voss et al., 2008). ⇑ Corresponding author. Address: Department of Experimental-Clinical and Health Psychology, Ghent University, Henri Dunantlaan 2, 9000 Ghent, Belgium. Fax: +32 (0)9 2646489. E-mail address: Lore.VanHulle@UGent.be (L. Van Hulle). 1053-8100/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.concog.2013.01.011 L. Van Hulle et al. / Consciousness and Cognition 22 (2013) 420–429 421 Even though sensory suppression appears to serve the goal of efficient movement execution (Bays & Wolpert, 2007; Gallace et al., 2010), it remains important that this sensory suppression is not absolute, and that the movement can be interrupted by more important demands or goals. Interruption may occur as a result of the bottom-up selection of salient information, such as highly intense or unexpected stimuli (Chapman & Beauchamp, 2006; Coulter, 1976; Williams et al., 1998), but also in a more top-down fashion as, for example, when a stimulus is considered as relevant or informative for a current goal or concern. Stimulus features that are related to an individual’s objectives are assumed to receive more attention (Corbetta & Shulman, 2002; Folk, Remington, & Johnston, 1992; Yantis, 1998). In a recent study reported by Juravle et al. (2011), the participants had to move one arm in order to grasp an object, while keeping their other arm still. During this movement, the participants received a tactile stimulus to either their moving or stationary hand and were instructed to detect its presence. As expected, tactile reaction times (RTs) were slower for stimuli delivered to the moving hand as compared to those stimuli delivered to the stationary hand. Interestingly, though, when the probability that stimulation would be administered to either the moving or stationary hand was raised, tactile RTs were correspondingly shorter, thus suggesting that attention can modulate tactile perception. Thus far, studies on sensory suppression have mostly been documented in the context of the movement of the fingers (e.g., Chapman & Beauchamp, 2006; Wasaka et al., 2003; Williams & Chapman, 2000, 2002; Williams et al., 1998) or the arms (e.g., Chapman & Beauchamp, 2006; Gallace et al., 2010; Juravle & Spence, 2011; Juravle et al., 2010, 2011). The present study was designed to address the question of whether sensory suppression also occurs during movements involving the back, and further to investigate whether this can be modulated in a top-down manner by attention. Indeed, although one might well expect similar findings as in previous work (see Juravle et al., 2011), it has to be noted that there might be differences between the processing of sensory information in the region of the back as compared to the front of the body. For example, there is usually no visual information available in the region around our back and, as such, vision typically does not provide additional information about the location from which tactile stimuli have been presented. Indeed, providing visual information has, on occasion, been shown to improve the processing of tactile information (Gillmeister & Forster, 2010; Kóbor, Füredi, Kovács, Spence, & Vidnyánzky, 2006; Press, Taylor-Clarke, Kennett, & Haggard, 2004). Moreover, Tipper et al. (2001) reported that the additive effect of vision was larger when viewing more familiar body parts (e.g., the face) as compared to viewing less familiar (e.g., the neck or back) non-directly visible body sites. One might therefore expect that participants would be less sensitive in terms of detecting stimuli presented to the back, as they are not so familiar with being stimulated in this region. On the other hand, a study reported by Weinstein (1968), measuring tactile sensitivity for different body parts by means of point localization, two-point discrimination, and pressure sensitivity, suggested that the back is clearly less sensitive than regions such as the head and the fingers, but more or less equally sensitive as compared to other body regions (e.g., the limbs). It is well-known that the cortical representation of the back is smaller as compared to more innervated regions like the fingers or the face (Banich, 2004). With regard to attentional processing in the space around the back, one study (Gillmeister & Forster, 2012) suggested that attentional processes may affect tactile information processing not only in the front but also in the back space of the body. It should, however, be noted that in this study, the tactile stimuli were applied to the hands, which were held behind the back, and not to the back itself. We are especially interested in sensory suppression during back movements due to its potential clinical relevance. It is well-known that low back pain patients are typically concerned about pain and injury during activities that involve back strain, and report being over-attentive to pain and even non-painful sensations in their back (Crombez, Vervaet, Lysens, Baeyens, & Eelen, 1998). This fearful anticipation might be expected to lead to a stronger focusing of their attention (i.e., hypervigilance) to the region of the back, especially during back movements, in order to rapidly detect signals of potential harm (Crombez, Van Damme, & Eccleston, 2005; Legrain et al., 2009; Van Damme, Legrain, Vogt, & Crombez, 2010; Vlaeyen & Linton, 2000). Accordingly, low intensity somatosensory input presented to the back may be processed more thoroughly as compared to individuals without this condition. As yet, however, this has not been investigated by means of a valid experimental paradigm (Van Damme et al., 2010). We consider our study as a first step in developing such a paradigm. Two tactile detection experiments are presented in which the perception of tactile information at rest and while performing back-bending movements are investigated. Tactile stimuli could be delivered to either the lower or upper back, as such creating a situation in which attention had to be divided between multiple stimulus locations. The focus of participants’ attention was manipulated by means of raising the probability that either the upper or the lower back would be stimulated. First, in line with previous research, tactile perception was hypothesized to be suppressed during the execution of back movements. Second, it was also hypothesized that the suppression of tactile stimulation during movement would be reduced when a participant’s attention was manipulated toward the stimulated location. 2. Experiment 1 2.1. Methods 2.1.1. Participants Twelve participants (10 females, 2 males; mean age = 24 years, age range 18–35 years), both students and PhD students from the Psychology Department who had no previous knowledge about the experiment, received a £5 voucher in return for taking part in the experiment. The study was conducted in accordance with the Declaration of Helsinki. All participants gave 422 L. Van Hulle et al. / Consciousness and Cognition 22 (2013) 420–429 Fig. 1. Representation of experimental set-up and stimulus locations used in Experiment 1. The participant is depicted with both hands next to the starting mice. informed consent and were free to terminate the experiment at any time. They all reported normal tactile perception (absence of nerve damage or injuries) at those locations where the tactile stimuli would be delivered, and reported no history of back pain problems. All data were considered appropriate for further statistical analyses. 2.1.2. Apparatus and materials Two computer mice (the start mice) were affixed to the surface of the table in front of the participant, at a distance of 50 cm from two other mice (the goal mice) that were placed on the other side of the table (see Fig. 1). The participants were seated so that their arms were stretched when holding the start mice. This meant that the participants always had to make a back-bending movement whenever they reached for the goal mice. Tactors (VBW32 skin stimulators, 1.6  2.4 cm vibrating surface, Audiological Engineering Corp., Somerville, MA, USA) were attached with tape to both the lower (lumbar curve, ±L4) and the upper back (upper thoracic curve, ±T2) of the participant. The participants wore a pair of headphones for the duration of the experiment in order to reduce the possibility that they would hear the operation of the tactors (Beyer Dynamic DT 531). The tactors were controlled by means of a custom-built tactor box connected to the main computer (Dell Technologies) and interfaced through Matlab (Psychophysics Toolbox 3; Brainard, 1997; Pelli, 1997) on Windows XP. Auditory stimuli (start signal, 100 ms, 800 Hz; stop signal, 50 ms, 400 Hz) were delivered by means of two loudspeakers, placed on both sides of the table, and could be clearly perceived by the participants despite wearing headphones. Participants responded by means of two foot pedals connected to the computer. 2.1.3. Task In this dual-task paradigm, the participants simultaneously engaged in a movement task and a perceptual task. The movement task consisted of moving both hands from the start mice toward the goal mice (see Fig. 1). More specifically, a trial started when the participant pressed the fingers of both hands on the buttons of the start mice. Immediately thereafter, a start signal indicated that the reach-to-grasp movement had to be initiated. When grasping the goal mice, the participants also needed to press them with both hands. Successful accomplishment of the task was indicated by a stop signal. In the rest condition, in which no back movement was required, the participants kept their hands at rest on top of the start mice. Each trial started with a start signal, which was followed, 600–900 ms later, by the stop signal. The perceptual task consisted of an unspeeded detection of subtle tactile stimuli administered on either the upper or lower back. In half of the trials, a tactile stimulus (11 dB, 250 Hz, 2 ms) was presented (signal trials), whereas, in the remainder of the trials, no stimulus was presented (noise trials). The intensity of the tactile stimulus was tested prior to the experiment. For this purpose, two collaborators performed a number of trials of the movement/perceptual task with different stimulus intensities. A stimulus intensity was chosen that could be perceived not only during rest but also during movement. The participants were instructed to indicate whether they felt a stimulus by pressing their left (right) foot down on a foot pedal when the signal was present (absent). Response assignments were counterbalanced across participants. The stimuli could be presented either during the preparation phase of the movement (10–100 ms following the start signal) or during its’ execution phase (300– 600 ms after the start signal). In order to reduce expectancy effects, stimuli were randomly delivered within these time windows and could be delivered with equal probability either during the preparation or execution stage of the movement (e.g., a stimulus could presented at 10, 11, . . . , 100 ms during the preparation phase or at 300, 301, . . . , 600 ms during the execution phase). In the rest condition, tactile stimuli were delivered at the same points in time. 2.1.4. Design The experimental design was blocked with six experimental conditions1 each consisting of 64 trials: Movement (rest, movement)  Attention (divided, focused-up, focused-low). The order of the blocks was counterbalanced across participants. 1 For the purpose of this study, the location of the stimulation (upper or lower back) and the timing of the stimulation (preparation or execution phase) were not analyzed as separate experimental conditions. However, other studies have demonstrated that movement execution causes sensory suppression during both movement preparation and execution phases (Juravle et al., 2011; Williams et al., 1998). L. Van Hulle et al. / Consciousness and Cognition 22 (2013) 420–429 423 In half of the blocks, the participants only had to perform the perceptual task (rest). In the other half of the blocks, the participants executed both the perceptual and movement tasks (movement). In order to manipulate attention to a specific body site, there were three different block types. In one block type (divided), the stimuli were in 50% of the signal trials delivered to the upper/lower back. In the second block type (focused-up), the stimuli were in 75% of the signal trials presented to the upper back and in 25% delivered to the lower back. In a third block type (focused-low), the stimuli were presented to the lower back in 75% of the signal trials and in 25% of trials to the upper back. The participants were informed about the proportion of stimuli that would be presented to the lower or upper back within each block. The experimental session took between 60 and 75 min to complete. 2.1.5. Data analysis Signal detection theory was used in order to calculate the hit and false alarm rates; which further allowed differentiation between perceptual sensitivity and response bias (Macmillan & Creelman, 2005; Stanislaw & Todorov, 1999; Wickens, 2002). A signal trial was defined as a trial in which a tactile stimulus was delivered, whereas a noise trial was defined as a trial in which no tactile stimulus was delivered. The hit rates (H) and false alarm rates (F) were computed for each experimental condition (for an overview of the calculations, see Appendix A). As a measure of perceptual sensitivity, A0 was calculated for each experimental condition, by using the following equation: A0 ¼ 0:5 þ ½signðH  FÞðH  FÞ2 þ jH  Fj=½4maxðH; FÞ  4HF ð1Þ 0 Values of A range between zero and one, with values of 0.5 or above indicating that perceptual sensitivity exceeds chance level, thus showing that participants were able to distinguish signals from noise. As a measure of response bias, c was calculated for each experimental condition, by using the following equation2: c ¼ ðU1 ðHÞ þ U1 ðFÞÞ=2 ð2Þ In this study, the sign of c was reversed in order to simplify its interpretation. Therefore, a value of zero here indicates no response bias, a positive value indicates a bias toward ‘yes’-responses (i.e., a bias to respond that a signal was present), and a negative value a bias toward ‘no’-responses (i.e., a bias to respond that no signal had been presented).3 For ease of comparison with the divided attention condition, the data of the focus-up and focus-low condition were merged (focused). The data were analyzed using repeated measures Analyses of Variance (ANOVAs). To obtain an objective and standardized measure of the magnitude of the observed effects, namely, a standardized difference between two means, effect sizes (Cohen’s d) for independent samples were calculated using Morris and DeShon’s (2002) formula (as cited in Borenstein, Hedges, Higgins, & Rothstein, 2009). The 95% Confidence Interval (95% CI) was also calculated. Cohen’s d is an effect size that is not design-dependent and conventional norms are available (Field, 2005). We determined whether Cohen’s d was small (0.20), medium (0.50), or large (0.80) (Cohen, 1988). 2.2. Results 2.2.1. Perceptual sensitivity Overall, perceptual sensitivity measures differed significantly from chance level, indicating that the participants were able to distinguish the signal from the noise. The means and standard deviations, and one-sample t-tests are presented in Table 1. A separate comparison of the focused-up or focused-low condition with the divided condition cannot be interpreted, as no distinction could be made for stimuli delivered at the upper vs. lower back in the divided attention condition. A repeated measures ANOVA was performed with Attention (divided, focused) and Movement (rest, movement) as independent variables, and perceptual sensitivity as the dependent variable. There was a main effect of Movement ((F(1, 11) = 22.58, p = .001); d = 0.98, 95% CI [0.48, 1.47]), revealing a decreased perceptual sensitivity for detecting tactile stimulation in the movement condition (M = 0.84, SD = 0.11), as compared to the rest condition (M = 0.97, SD = 0.02). There was no main effect of Attention (F(1, 11) < 1, ns; d = 0.00, 95% CI [0.28, 0.28]), indicating that perceptual sensitivity did not differ significantly when attention was divided between the two body locations (M = 0.91, SD = 0.06) as compared to when attention was focused on the location where the stimulation was most likely to be delivered (M = 0.91, SD = 0.07). The Attention  Movement interaction also failed to reach statistical significance (F(1, 11) = 1.00, p = .34). See Fig. 2 for a graphical representation of the data. 2.2.2. Response bias One sample t-tests revealed that all but one of the measures of response bias was significantly different from zero. The negative values indicated that overall participants were conservative in reporting the presence of a stimulus. The means and standard deviations, and one-sample t-tests are presented in Table 1. A repeated measures ANOVA was performed with 2 3 U Represents the ‘phi’ score used to convert z scores into probabilities. No sensitivity and response bias measures can be calculated for the location (or the timing of the stimulation) separately. More specifically, when a stimulus is present, a distinction can be made between stimuli delivered to the upper or the lower back (or during the preparation or execution phase) and a hit rate can be calculated. However, when no stimulus is present, no such distinction can be made regarding the upper and lower back (or the preparation or execution phase) in order to calculate separate false alarm rates. 424 L. Van Hulle et al. / Consciousness and Cognition 22 (2013) 420–429 Table 1 Hit (H) and false alarm (F) rates for each condition together with means and standard deviations for perceptual sensitivity (A0 ) and response bias (c) in Experiment 1. One sample t-tests were used to assess whether perceptual sensitivity exceeded chance level (A > 0.50) and whether there was any indication of response bias (c – 0). A0 Rest divided Movement divided Rest focused Movement focused c M SD t(11) p M SD t(11) p 0.96 0.85 0.98 0.83 0.04 0.10 0.01 0.14 37.82 12.25 118.74 8.38 .00 .00 .00 .00 0.23 0.77 0.15 0.75 0.33 0.57 0.24 0.70 2.43 4.68 2.19 3.90 .03 .001 .05 .002 H FA .90 .60 .95 .58 .02 .05 .02 .07 Attention (divided, focused) and Movement (rest, movement) as independent variables and response bias (c) as the dependent variable. There was a main effect of Movement (F(1, 11) = 18.84, p = .001; d = 1.12, 95% CI [0.48, 1.75]), indicating that participants were more inclined to report the presence of a stimulus in the rest condition (M = 0.19, SD = 0.26) than in the movement condition (M = 0.76, SD = 0.56). There was no main effect of Attention (F(1, 11) < 1, ns; d = 0.03, 95% CI [0.35, 0.40]), indicating that participants were no more inclined to report the presence of a stimulus in the focused attention condition (M = 0.45, SD = 0.42) than in the divided attention condition (M = 0.44, SD = 0.34), nor was there a significant Attention  Movement interaction effect (F(1, 11) < 1). 2.3. Interim discussion The present results clearly demonstrate that the detection of tactile information is suppressed while participants perform a back movement as compared to rest. These results therefore extend the findings from previous research that has investigated the processing of tactile information during hand or arm movements (e.g., Gallace et al., 2010; Juravle et al., 2010, 2011; Wasaka et al., 2003; Williams & Chapman, 2000, 2002; Williams et al., 1998). There was, however, no effect of the attention manipulation. With their attention divided between the two locations, participants performed equally well as compared to when they were instructed to focus their attention on a specific location. This result might be explained by the fact that in the focused attention condition, there was also a 25% chance of receiving a stimulus at the other location. This might have led the participants to divide their attention between both locations in the ‘focused attention’ condition as well. Therefore, in Experiment 2, a stronger attentional manipulation was used. In the focused attention condition, all (i.e., 100%) of the stimuli were delivered to one location, i.e., either the upper or the lower back. In a control condition (divided attention), the probability of receiving tactile stimulation to the lower vs. upper back was equalized. It should be noted that the perceptual sensitivity of the participants was near ceiling in the rest condition. This might be the result of the particular stimulus intensity that was chosen. In order to examine the influence of attention on detection performance, participants would ideally need to show a ‘medium’ level of performance in the absence of any attentional manipulation. However, prior testing revealed that when a stimulus intensity was chosen that gave rise to an intermediate level of performance at rest, performance dropped to zero during movement. On the other hand, when a stimulus intensity was chosen that gave rise to an intermediate level of performance during movement, performance approached ceiling under conditions of rest. As we were especially interested in the influence of attention on stimulus detection during movement, the second option was preferred. In the pre-experimental phase of Experiment 2, we used a standard psychophysical procedure in order to improve the determination of the stimulus intensity for the experimental phase. As in the first experiment, the participants were expected to have a higher perceptual sensitivity while at rest as compared to while performing a movement. In addition, it was expected that the participants would show better performance Perceptual sensitivity A' 1 0.9 0.8 Rest Movement 0.7 0.6 0.5 Divided Focused Fig. 2. Perceptual sensitivity measures (A0 ) depending on Movement and Attention in Experiment 1. Vertical error bars represent the standard errors of the mean. L. Van Hulle et al. / Consciousness and Cognition 22 (2013) 420–429 425 during movement when their attention was focused on one body location as compared to when their attention was divided equally between the upper and the lower back. 3. Experiment 2 3.1. Methods 3.1.1. Participants Fourteen participants (13 females, 1 male; mean age = 26 years, age range 19–31 years), all students and PhD students from the Psychology Department who had no previous knowledge of the experiment, received a £5 voucher for taking part in the study. The study was conducted in accordance with the Declaration of Helsinki. All of the participants gave their informed consent and were free to terminate the experiment at any time. The participants reported normal tactile perception (and the absence of nerve damage or injuries) at the locations where the tactile stimuli would be delivered, and reported no history of back pain problems. One participant was excluded from further analysis because of technical problems (no tactile stimuli were delivered during the experimental blocks). 3.1.2. Task and procedure In a pre-experimental phase, the intensity of the tactile stimulation was determined at rest by means of an adaptive procedure comprising two different interleaved staircases for each of the two stimulation locations (upper or lower back). Each of the two staircases consisted of a one-up four-down adaptive procedure designed to keep performance at a level of 90% correct (Levitt, 1971). For the first trial, each of the two staircases started with an above threshold stimulation intensity (6 dB). The presentation of trials from each of the staircases was randomized throughout this pre-experimental phase. The participants were instructed to respond whenever they felt the presence of a stimulus. The staircase changed direction after one incorrect response (i.e., increasing the corresponding location stimulation by one step – ‘UP’) or a sequence of four correct responses (i.e., decreasing the corresponding location stimulation by one step – ‘DOWN’). Changes in the direction of the staircase are referred to as ‘reversals’. The pre-experimental phase required the participants to complete a maximum of 240 trials. After the completion of every 60 trials, the participants were informed by three consecutive beeps that the block had finished and that they could take a break if they so desired. During the breaks, the experimenter could monitor a progress bar presented on the screen behind their chair. This provided an estimate of the number of trials remaining, calculated on the basis of the total number of possible reversals (34). The experimenter pressed a key on the keyboard in order to continue onto the next block. The staircase for each stimulation location terminated once the total number of reversals (17) or the total number of trials (120) had been reached. The first five reversals were excluded from the final threshold calculations, which consisted of the average value of upward and downward reversals. The average intensity was found to be the same for each location and for each individual (14 dB, 250 Hz). This intensity was then used for the experimental trials described next. Both the perceptual and the movement task were similar to those used in Experiment 1. Only the attention manipulation differed slightly. The experimental design was blocked with six experimental conditions each consisting of 64 trials: Movement (movement vs. rest)  Attention (focused vs. divided). The order of the blocks was counterbalanced across participants. In half of the blocks, the participants executed both the perceptual and the movement tasks (movement). In the other half of the blocks, the participants only had to perform the perceptual task (rest). In order to manipulate attention to a specific body location, there were three different block types. In one block type, all (i.e., 100%) of the stimuli were presented to the participant’s upper back (focused-up); in the second block type, all of the stimuli were presented to the participant’s lower back (focused-low); and in a third block type, 50% of the stimuli were presented to the upper back and 50% to the lower back (divided). The participants were informed about the proportion of stimuli presented at the lower or the upper back within each block. Signal detection measures of perceptual sensitivity and response bias were computed (see, Section 2.1.5.; for an overview of the calculations, see Appendix A). 3.2. Results 3.2.1. Perceptual sensitivity Overall, measures of perceptual sensitivity differed significantly from chance level, indicating that the participants were able to distinguish the signal from the noise. See Table 2 for means, standard deviations, and one-sample t-tests. A repeated measures ANOVA was performed with Attention (divided, focused) and Movement (rest, movement) as independent variables and perceptual sensitivity (A0 ) as the dependent variable. The analysis revealed a significant main effect of Movement (F(1, 12) = 36.80, p < .001; d = 2.24, 95% CI [0.91, 3.57]), indicating that participants exhibited a lower perceptual sensitivity in the movement condition (M = 0.70, SD = 0.12) than in the rest condition (M = 0.93, SD = 0.08) than in the movement condition. There was also a significant main effect of Attention (F(1, 12) = 6.26, p = .03; d = 0.47, 95% CI [0.04, 0.99]), with higher perceptual sensitivity being observed when participants’ attention was focused on the location of stimulation (M = 0.84, SD = 0.09) as compared to when it was divided between the two locations (M = 0.80, SD = 0.08). Of particular interest, the analysis revealed a significant Attention  Movement interaction (F(1, 12) = 6.03, p = .03). To further explore this interaction, contrast analyses were carried out. These analyses revealed that, in the rest condition, perceptual sensitivity did not differ 426 L. Van Hulle et al. / Consciousness and Cognition 22 (2013) 420–429 Table 2 Hit (H) and false alarm (F) rates for each condition together with means and standard deviations for perceptual sensitivity (A0 ) and response bias (c) in Experiment 2. One sample t-tests were used to assess whether perceptual sensitivity exceeded chance level (A0 > 0.50) and whether there was any indication of response bias (c – 0). A0 Rest divided Movement divided Rest focused Movement focused c M SD t(12) p M SD t(12) p 0.93 0.66 0.94 0.75 0.08 0.12 0.09 0.14 18.56 4.82 17.39 6.23 <.001 <.001 <.001 <.001 0.26 0.94 0.23 1.23 0.35 0.60 0.33 0.28 2.69 5.68 2.49 16.10 .02 <.001 .03 <.001 H FA .85 .34 .86 .33 .06 .13 .05 .04 significantly when participants’ attention was focused on the location where the stimuli would be administered as compared to when participants’ attention was divided between the two body locations (F(1, 12) = 0.12, p = .74; d = 0.11, 95% CI [0.25, 0.49]). However, in the movement condition, perceptual sensitivity was significantly higher when participants’ attention was focused on the location where the stimuli would be administered as compared to when their attention during movement was divided between the two body locations (F(1, 12) = 8.08, p = .02; d = 0.68, 95% CI [0.21, 1.15]). See Fig. 3 for a graphical representation of the data. 3.2.2. Response bias One sample t-tests revealed that all response bias measures were significantly different from zero. The negative values indicated that, overall, participants were conservative in reporting the presence of a stimulus. See Table 2 for means, standard deviations, and the results of one-sample t-tests. A repeated measures ANOVA was performed with Attention (divided, focused) and Movement (rest, movement) as independent variables and response bias (c) as the dependent variable. There was a significant main effect of Movement (F(1, 12) = 55.17, p < .001; d = 2.61, 95% CI [1.18, 4.04]), indicating that participants were more inclined to report the presence of a stimulus in the rest condition (M = 0.24, SD = 0.27) than in the movement condition (M = 1.09, SD = 0.37). There was no significant main effect of attention (F(1, 12) = 2.26, p = .16; d = 0.42, 95% CI [0.16, 1.01]), indicating that participants were no more inclined to report the presence of a stimulus when their attention was divided between two body locations (M = 0.60, SD = 0.35), as compared to when it was focused on one body site (M = .73, SD = 0.24). The Attention  Movement interaction was not significant (F(1, 12) = 2.50, p = .14). 4. General discussion The aim of the present study was to investigate: (1) whether the performance of a back movement, i.e., bending over in order to grasp an object, leads to the sensory suppression of tactile perception on the back; and (2) whether attention modulates the processing of tactile information during the execution of a back movement. Two experiments were conducted in which participants were instructed to detect the presence of subtle tactile stimuli delivered to either their lower or upper back, while either performing a back-bending movement or while at rest. The focus of participants’ attention was manipulated by informing them that the probability of stimulation of either the lower or the upper back would be raised during the upcoming experimental block. First of all, it was expected that participants’ perceptual sensitivity for the detection of tactile stimuli would be lower while they were simultaneously performing a back-bending movement as compared to rest. This hypothesis was confirmed by the results of both experiments. The findings clearly demonstrate a sensory suppression effect during the execution of the back movement. Our results extend the findings of previous research that has investigated the processing of tactile Perceptual sensitivity A' 1 0.9 0.8 Rest Movement 0.7 0.6 0.5 Divided Focused Fig. 3. Perceptual sensitivity (A0 ) measures depending on Movement and Attention in Experiment 2. Vertical error bars represent the standard errors of the mean. L. Van Hulle et al. / Consciousness and Cognition 22 (2013) 420–429 427 information during the movement of the hand or arm (e.g., Gallace et al., 2010; Juravle et al., 2010, 2011; Wasaka et al., 2003; Williams & Chapman, 2000, 2002; Williams et al., 1998), as sensory suppression also seems to play a role when performing a back-movement. Although there might be reasons to assume that the processing of sensory information in front and rear space may not be identical because, for example, no visual information is available for the region of the back (Gillmeister & Forster, 2010; Kóbor et al., 2006; Press et al., 2004; Tipper et al., 2001; but see Weinstein, 1968), it appears that the effect of sensory suppression was quite similar. Furthermore, it was hypothesized that the suppression effect would be reduced when participants’ attention was focused toward the stimulated location. This effect was not observed in Experiment 1, where the results revealed that participants performed equally well no matter whether their attention was divided between the two body locations or focused on a specific location. In Experiment 2, in which a stronger attentional manipulation was used, the results clearly demonstrated that participants were better able to detect the stimuli when their attention was focused on one body location as compared to when it was divided equally between two body locations, but only when they were executing the movement. Thus, the suppression seemed to be counteracted by the effects of attention, presumably because attention increased perceptual sensitivity. Gillmeister and Forster (2012) already suggested that attentional processes may affect both the space in front as well as behind the body. Intriguingly, our findings suggest that the suppression mechanism described above might not only be regulated by bottom-up (or stimulus-driven), but also by top-down (or goal-driven) attentional processes. From an evolutionary perspective, it makes sense that it is not only highly intense or novel stimulation that captures our attention. It is widely accepted that the processing of sensory information results from an interaction between stimulus features and personal, or task, goals (Corbetta & Shulman, 2002; Folk et al., 1992; Legrain et al., 2009; Santangelo & Spence, 2008). It has been proposed that individuals adopt ‘attentional control settings’ including certain stimulus features or characteristics that are relevant for their goals and that will receive more attention if they are present in the environment (Yantis, 1998). In Experiment 2, the information that the tactile stimuli would only be delivered to the upper back might have activated the features ‘tactile’ and ‘upper back’ in the participants’ attentional set, resulting in higher attention for that specific location. However, when participants were informed that the tactile stimuli could be delivered both at the upper and lower back, the location feature in their attentional set was defined less precisely and participants therefore needed to divide their attention between the two locations on their back. This may also explain why the attention manipulation used in Experiment 1, in which the location feature was also defined less precisely, did not result in higher perceptual sensitivity. The results of both experiments revealed that the execution of a movement also affected response bias. When they did not have to perform any movement, participants were more inclined to report the presence of a tactile stimulus as compared to when they were executing the back movement. One explanation that has been proposed for the response bias is that the suppression phenomenon might involve a decision-based component (Juravle & Spence, 2011). Besides this, the fact that movement execution is accompanied by more noise (as compared to rest) might have led to an altered decision criterion (i.e., a conservative response strategy) because of a high level of uncertainty during the task. This was the case particularly because the importance of accuracy rather than response speed was stressed before the experiment. The fact that the perceptual task was more difficult during movement might have resulted in participants having different expectations about the task during movement vs. at rest. It has been suggested before that expectations might affect the response criterion used by participants (Summerfield & Egner, 2009). It should be noted that there are some limitations with regard to the interpretation of the results of the present study. First, Juravle et al.’s study (2011), by utilizing the speed of response as a dependent variable, illustrated an additional effect of attention on the detection of tactile stimuli that were delivered to participants’ stationary hand. In contrast, neither of the experiments described here demonstrate that attention increased participants’ perceptual sensitivity while they were at rest. However, these results should be interpreted with caution. Participants’ performance in the rest condition was very good, suggesting the presence of a ceiling effect. The interaction between the attention manipulation and the movement conditions in Experiment 2 might thus have resulted from the fact that the performance in the rest condition was too high to show any gains in the focused attention condition. Seemingly, the standard psychophysical procedure used in the pre-experimental phase of Experiment 2 did not get round the problem of ceiling effects. A more time-consuming procedure that might, however, avoid ceiling effects and, as such, make all potential effects of attention visible would be to determine stimulus intensities separately during movement and rest (see Juravle et al., 2010; Williams & Chapman, 2000) in a pre-experimental phase. Second, it has been suggested previously that the simultaneous execution of two tasks may explain the deterioration in stimulus detection in sensory suppression experiments. Indeed, a rest-condition in which participants are given a dual task that doesn’t require movement would make it possible to rule out this possibility. Available studies, however, have shown that decrements in detection performance during movement can only partly be explained by dual-task effects (Gallace et al., 2010; Williams & Chapman, 2000, 2002; Williams et al., 1998). Gallace et al. (2010), for example, investigated tactile change detection performance in three conditions: a dual-task condition in which a verbal response was required, a dual-task condition in which a motor-response was required, and a single-task. The results revealed that a secondary task did indeed diminish participants’ sensitivity for the detection of change. However, the execution of a motor response resulted in a much larger drop in participants’ change detection performance as compared to the verbal response condition. Future research might well be advised to try and avoid these two limitations by including, besides a movement condition, a rest condition in which participants have to make a movement that is irrelevant to the stimulus locations (e.g., eye movements). That way, both conditions would constitute dual task conditions, which makes it easy to interpret the results in terms of 428 L. Van Hulle et al. / Consciousness and Cognition 22 (2013) 420–429 movement-related suppression. Furthermore, such a condition might eventually result in a lower detection performance in the rest condition and thus avoid ceiling effects. Third, although one might suspect that the current study suffers from limited statistical power due to the relatively low number of participants tested, a closer look at the data of Experiment 2 for each participant separately revealed that for most participants, perceptual sensitivity during movement was higher when their attention was focused at the location of stimulation (9 out of 13 participants). Furthermore, this effect has a large effect size according to Cohen’s (1988) norms. This underlines the robustness of the findings reported here. In conclusion, the results of the two experiments reported here expand our understanding concerning the processing of tactile information during movement execution. More research is certainly needed before any firm conclusions can be drawn, but sensory suppression seems to be present in the execution of many body movements. Nevertheless, our results extend previous findings by showing that, in the back region just as elsewhere on the body surface, this mechanism is flexible as it allows modulation by top-down attentional processes. Furthermore, future research on sensory suppression could be expected to advance our knowledge on the processing of bodily sensations in certain clinical populations, such as those individuals suffering from chronic lower back pain. One particular benefit of this paradigm involves the focus on accuracy rather than RTs. As it been demonstrated that certain clinical populations such as chronic pain patients are characterized by cognitive dysfunction and psychomotor slowing (e.g., Dick, Eccleston, & Crombez, 2002; Glass, 2009; Veldhuijzen, Sondaal, & Oosterman, 2012), paradigms relying on response speed may prove less reliable in these populations (Van Damme, Crombez, & Notebaert, 2008). Multiple hypotheses can be specified when applying this paradigm in persons with chronic low back pain. The hypervigilance hypothesis assumes that individuals with chronic low back pain spontaneously focus their attention on the region of the back, especially in situations that evoke bodily threat, such as the execution of a movement involving the back (Crombez et al., 2005; Vlaeyen & Linton, 2000). It is known that the presence of threat leads to the facilitated processing of threat-related information (Van Damme et al., 2010). When applying this paradigm – without an experimental manipulation of attention – to chronic low back pain patients, it might be expected that they will spontaneously focus attention to the back because back-related movements are threatening for them. As such, a better detection of tactile information during movement as compared to a control group without back pain could be expected. The paradigm is thus primarily intended to investigate the presence of hypervigilance, rather than reductions in pain experience during movement or physical activity as a result of distraction. Of course, other hypotheses can be specified. For example, it has been hypothesized that individuals with chronic low back pain might suffer from tactile dysfunction (e.g., Moseley, Gallagher, & Gallace, 2012), which might result in an overall decreased detection performance in this group as compared to a control group. Acknowledgments L. Van Hulle is funded by the Special Research Fund of Ghent University (BOF09/DOC/013). The study reported here was funded by the Special Research Fund for a Bilateral Scientific Cooperation Oxford-Ghent University (BOF10/BIP/015). Appendix A Calculation of the hit (H) and false alarm (F) rates for each condition in both experiments. Where the hit rate was perfect (H = 1), or where there were no false alarms (F = 0), the proportions 1 and 0 were adjusted by 1/2N and 1/(1–2N) respectively (Juravle & Spence, 2011). H F Rest Focused-up # Correctly reported up-signals # Up-signals # Incorrectly reported signals # Noise trials Movement Focused-low # Correctly reported low-signals # Low-signals # Incorrectly reported signals # Noise trials Divided # Correctly reported signals # Signals # Incorrectly reported signals # Noise trials References Banich, M. T. (2004). Cognitive neuroscience and neuropsychology. Boston: Houghton Mifflin Company. Bays, P. M., & Wolpert, D. M. (2007). Computational principles of sensorimotor control that minimize uncertainty and variability. Journal of Physiology, 578, 387–396. Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2009). Introduction to meta-analysis. West Sussex (UK): Wiley. Brainard, D. H. (1997). The psychophysics toolbox. Spatial Vision, 10, 433–436. Chapman, C. E., & Beauchamp, E. (2006). Differential controls over tactile detection in humans by motor commands and peripheral reafference. Journal of Neurophysiology, 96, 1664–1675. L. Van Hulle et al. / Consciousness and Cognition 22 (2013) 420–429 429 Cohen, J. (1988). Statistical power analysis for the behavioral sciences. San Diego (CA): McGraw-Hill. Corbetta, M., & Shulman, G. L. (2002). Control of goal-directed and stimulus-driven attention in the brain. Nature Reviews Neuroscience, 3, 201–215. Coulter (1976). Sensory transmission through lemniscal pathway during voluntary movement in cat. Journal of Neurophysiology, 37, 831–845. Crombez, G., Van Damme, S., & Eccleston, C. (2005). Hypervigilance to pain: An experimental and clinical analysis. Pain, 116, 4–7. Crombez, G., Vervaet, L., Lysens, R., Baeyens, F., & Eelen, P. (1998). Avoidance and confrontation of painful, back-straining movements in chronic back pain patients. Behavior Modification, 22, 62–77. Dick, B., Eccleston, C., & Crombez, G. (2002). Attentional functioning in fibromyalgia, rheumatoid arthritis, and musculoskeletal pain patients. Arthritis & Rheumatism, 47, 639–644. Folk, C. L., Remington, R. W., & Johnston, J. C. (1992). Involuntary covert orienting is contingent on attentional control settings. Journal of Experimental Psychology: Human Perception and Performance, 18, 1030–1044. Gallace, A., Zeeden, S., Röder, B., & Spence, C. (2010). Lost in the move? Secondary task performance impairs tactile change detection on the body. Consciousness and Cognition, 19, 215–229. Gillmeister, H., & Forster, B. (2010). Vision enhances selective attention to body-related information. Neuroscience Letters, 483, 184–188. Gillmeister, H., & Forster, B. (2012). Hands behind your back: Effects of arm posture on tactile attention in the space behind the body. Experimental Brain Research, 216, 489–497. Glass, J. M. (2009). Review of cognitive dysfunction in fibromyalgia: A convergence on working memory and attentional control impairments. Rheumatic Disease Clinics of North America, 35, 299–311. Juravle, G., Deubel, H., & Spence, C. (2011). Attention and suppression affect tactile perception in reach-to-grasp movements. Acta Psychologica, 138, 302–310. Juravle, G., Deubel, H., Tan, H. Z., & Spence, C. (2010). Changes in tactile sensitivity over the time-course of a goal-directed movement. Behavioural Brain Research, 208, 391–401. Juravle, G., & Spence, C. (2011). Juggling reveals a decisional component to tactile suppression. Experimental Brain Research, 213, 87–97. Kóbor, I., Füredi, L., Kovács, G., Spence, C., & Vidnyánzky, Z. (2006). Back-to-front: Improved tactile discrimination performance in the space you cannot see. Neuroscience Letters, 400, 163–167. Legrain, V., Van Damme, S., Eccleston, C., Davis, K. D., Seminowicz, D. A., & Crombez, G. (2009). A neurocognitive model of attention to pain: Behavioral and neuroimaging evidence. Pain, 144, 230–232. Levitt, H. (1971). Transformed up-down methods in psychoacoustics. Journal of the Acoustical Society of America, 49, 467–477. Macmillan, N. A., & Creelman, C. D. (2005). Detection theory: A user’s guide (2nd ed.). Mahwah, New Jersey: Lawrence Erlbaum Associates. Moseley, G. L., Gallagher, L., & Gallace, A. (2012). Neglect-like tactile dysfunction in chronic back pain. Neurology, 79, 327–332. Pelli, D. G. (1997). The video toolbox software for visual psychophysics: Transforming numbers into movies. Spatial Vision, 10, 437–442. Post, L. J., Zompa, I. C., & Chapman, C. E. (1994). Perception of vibrotactile stimuli during motor activity in human subjects. Experimental Brain Research, 100, 107–120. Press, C., Taylor-Clarke, M., Kennett, S., & Haggard, P. (2004). Visual enhancement of touch in spatial body representation. Experimental Brain Research, 154, 238–245. Santangelo, V., & Spence, C. (2008). Is the exogenous orienting of spatial attention truly automatic? Evidence from unimodal and multisensory studies. Consciousness and Cognition, 17, 989–1015. Stanislaw, H., & Todorov, N. (1999). Calculation of signal detection theory measures. Behaviour Research Methods, Instruments, & Computers, 31, 137–149. Summerfield, C., & Egner, T. (2009). Expectation (and attention) in visual cognition. Trends in Cognitive Sciences, 13, 403–409. Tipper, S. P., Philips, N., Dancer, C., Lloyd, D., Howard, L. A., & McGlone, F. (2001). Vision influences tactile perception at body sites that cannot be viewed directly. Experimental Brain Research, 139, 160–167. Van Damme, S., Crombez, G., & Notebaert, L. (2008). Attentional bias to threat: A perceptual accuracy approach. Emotion, 8, 820–827. Van Damme, S., Legrain, V., Vogt, J., & Crombez, G. (2010). Keeping pain in mind: A motivational account of attention to pain. Neuroscience and Biobehavioral Reviews, 34, 204–213. Veldhuijzen, D. S., Sondaal, S. F. V., & Oosterman, J. M. (2012). Intact cognitive inhibition in patients with fibromyalgia but evidence of declined processing speed. Journal of Pain, 13, 507–515. Vlaeyen, J. W. S., & Linton, S. J. (2000). Fear-avoidance and its consequences in chronic musculoskeletal pain: A state of the art. Pain, 85, 317–332. Voss, M., Ingram, J. N., Wolpert, D. M., & Haggard, P. (2008). Mere expectation to move causes attenuation of sensory signals. PLoS One, 3, e2866. Wasaka, T., Hoshiyama, M., Nakata, H., Nishihira, Y., & Kakigi, R. (2003). Gating of somatosensory evoked magnetic fields during the preparatory period of self-initiated finger movement. NeuroImage, 20, 1830–1838. Weinstein, S. (1968). Intensitive and extensive aspects of tactile sensitivity as a function of body part, sex and laterality. In D. R. Kenshalo (Ed.), The skin senses (pp. 195–200). Springfield (IL): Thomas. Wickens, T. D. (2002). Elementary signal detection theory. New York: Oxford University Press. Williams, S. R., & Chapman, E. C. (2000). Time course and magnitude of movement-related gating of tactile detection in humans. II. Effects of stimulus intensity. Journal of Neurophysiology, 84, 863–875. Williams, S. R., & Chapman, E. C. (2002). Time course and magnitude of movement-related gating of tactile detection in humans. III. Effect of motor tasks. Journal of Neurophysiology, 88, 1968–1979. Williams, S. R., Shenasa, J., & Chapman, E. (1998). Time course and magnitude of movement-related gating of tactile detection in humans. I. Importance of stimulus location. Journal of Neurophysiology, 79, 947–963. Yantis, S. (1998). Control of visual attention. In H. Pashler (Ed.), Attention (pp. 223–256). Hove, England: Psychology Press.

Consciousness and Cognition 19 (2010) 829–837 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog It just felt right: The neural correlates of the fluency heuristic q Kirsten G. Volz a,b,*, Lael J. Schooler a, D. Yves von Cramon b,c a b c Max Planck Institute for Human Development, Berlin, Germany Max Planck Institute for Neurological Research, Cologne, Germany Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany a r t i c l e i n f o Article history: Received 30 November 2009 Available online 16 June 2010 Keywords: Fluency heuristic Decision making fMRI Claustrum a b s t r a c t Simple heuristics exploit basic human abilities, such as recognition memory, to make decisions based on sparse information. Based on the relative speed of recognizing two objects, the fluency heuristic infers that the one recognized more quickly has the higher value with respect to the criterion of interest. Behavioral data show that reliance on retrieval fluency enables quick inferences. Our goal with the present functional magnetic resonance imaging study was to isolate fluency-heuristic-based judgments to map the use of fluency onto specific brain areas that might give a better understanding of the heuristic’s underlying processes. Activation within the claustrum for fluency heuristic decisions was found. Given that claustrum activation is thought to reflect the integration of perceptual and memory elements into a conscious gestalt, we suggest this activation correlates with the experience of fluency. Ó 2010 Elsevier Inc. All rights reserved. 1. Introduction Simple heuristics can be used to exploit basic human abilities, such as recognition memory, to make decisions based on sparse information. One such heuristic is the fluency heuristic. Building on a long tradition of research on fluency (Jacoby & Dallas, 1981; Kelley & Jacoby, 1998), Schooler and Hertwig (2005) defined their fluency heuristic this way: If two objects are recognized, and one of the objects is more fluently retrieved, then infer that this object has the higher value with respect to the criterion, where retrieval fluency is defined as how long it takes to retrieve a trace from long-term memory. The fluency heuristic can help us make good inferences when all relevant objects are recognized and when there is a substantial correlation— in either direction—between the criterion and the retrieval fluency. For convenience, we assume a positive correlation for the remainder of this paper. For example, when choosing which of two recognized cities, say, the Japanese cities of Yokohama and Kyoto, has more inhabitants, one could achieve a reasonable level of accuracy if there is a substantial correlation between the ease of retrieving a city’s name and its population. The fluency heuristic works well to the extent that it can exploit relevant characteristics of the environment that are encoded in the relative accessibility of memory traces. The rationale behind the fluency heuristic is that memory performance reflects the patterns with which stimuli appear and reappear in the environment (Anderson & Schooler, 1991; Schooler & Anderson, 1997). Accordingly, the retrieval fluency associated with the recognition of stimuli correlates to a large extent with how frequently and recently relevant stimuli have been experienced (Hertwig, Herzog, Schooler, & Reimer, 2008). Since it is not currently possible to measure retrieval fluency directly, it has been operationalized by recognition latency, that is, how long it takes people to judge whether they recognize q This article is part of a special issue of this journal on Self, Other and Memory. * Corresponding author. Address: Max Planck Institute for Human Development Research, Lentzeallee 94, 14195 Berlin, Germany. Fax: + 49 (0)30 82406 394. E-mail address: kirsten.volz@cin.uni-tuebingen.de (K.G. Volz). URL: http://www.cin.uni-tuebingen.de (K.G. Volz). 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.05.014 830 K.G. Volz et al. / Consciousness and Cognition 19 (2010) 829–837 an object or not (e.g., Hertwig et al., 2008). In short, differences in recognition latencies are taken as differences in retrieval fluencies. In a series of experiments, Hertwig et al. (2008) pursued several basic questions about how the fluency heuristic operates. More precisely, the authors showed that (a) recognition latencies are indeed indicative of criteria, such as city population or the wealth of athletes, (b) people can discriminate the recognition latencies of two objects that exceed a difference of 100 ms, (c) people’s judgments adhere more to the predictions of the fluency heuristic when differences between recognition latencies are large, which tends to happen when fluency is most predictive, (d) inferences in line with the fluency heuristic take less time than those that conflict with the heuristic, and (e) people’s inferences frequently accord with those predicted by the fluency heuristic, given that it is applicable. It can be argued, however, that accordance to the fluency heuristic is spurious, perhaps depending on a correlation between fast retrieval and knowledge about an object. So while people’s behavior may correspond to the fluency heuristic, the underlying process may be an entirely different knowledge-based strategy (Marewski & Schooler, 2010). Through a priming study, Hertwig et al. provided evidence that fluency does indeed guide inferences irrespective of other factors associated with fluency, such as the amount of knowledge known about the objects. Hertwig et al.’s main goal was to establish that the fluency heuristic does, in fact, guide decisions. Our aim was to uncover more about the processes underlying fluency-based decisions. One hypothesis draws on previous neuroscientific results on phenomenologically similar kinds of decisions. A second hypothesis focuses on an account of retrieval fluency that is based on successfully binding memory traces. The phenomenological experience of decisions based on the fluency heuristic has not been studied. The experience may be one of familiarity, the ‘‘felt-rightness” of a specific response, or an intuitive feeling about which option to choose. We start with Volz et al.’s (2006) investigation of the neural correlates of the recognition heuristic, which can be stated as follows: ‘‘If one of two objects is recognized and the other is not, then infer that the recognized object has the higher value with respect to a criterion” (Goldstein & Gigerenzer, 2002, p. 76). For this functional magnetic resonance imaging (fMRI) study, Volz et al. adapted Goldstein and Gigerenzer’s city judgment task in which participants had to infer which of two cities was larger. There were three critical types of trials, depending on whether both cities were recognized (RR trials), one of the cities was recognized (RU), or both cities were unrecognized (UU). After the fMRI experiment, participants were asked about what strategies they used in the three types of critical trials. For RR trials participants said they relied on knowledge if available, yet they said that such knowledge-based inferences were possible in only a fraction of these trials. When participants said they could not reach a decision based on knowledge, 17 of the 18 participants reported choosing the city that felt larger or more familiar or that they had made an informed guess. Together, these anecdotal data suggest that if knowledge was unavailable, participants relied on some other information. Their descriptions suggest that their judgments resemble the ‘‘intuitive assessment of the felt-rightness of a memory” (Schnyer, Nicholls, & Verfaellie, 2005, p. 837). The idea of an intuitive assessment of the felt-rightness of a memory is found in the literature on the ventromedial prefrontal cortex (VMPFC). For example, Schnyer et al. (2005) investigated feeling-of-knowing (FOK) judgments in an episodic memory task. Participants had to indicate the probability that they would recognize the final word of a sentence from a list of previously studied sentences. Prior to being shown the final word, the participants made an FOK judgment about whether they could retrieve the word on a 5-point scale. The instructions for using the rating scale were as follows: ‘‘Only press 5 if the answer pops effortlessly to mind. If you feel the answer is just there under the surface, press 4. . . . If you have no recollection of ever having seen the sentence, then press 1.” The authors found the VMPFC to be specifically engaged during accurate FOK judgments, defined as the four ratings on the scale that did not indicate certain knowledge of successful retrieval (i.e., 1–4 on the scale). In addition, the authors found a positive correlation between VMPFC activation and FOK judgments. When ‘‘know” ratings (i.e., 5 on the scale) were added to this analysis of retrievability, the amplitude levels for these ‘‘know” ratings did not continue the linear trend in the VMPFC but rather dropped off. Accordingly, the authors concluded that the VMPFC monitors the output of retrieval processes. Supporting this hypothesis, patients with lesions to the prefrontal cortex encompassing the VMPFC showed a clear impairment when making predictions about their subsequent recognition performance (Schnyer et al., 2004). Since the patients’ familiarity-based assessment was demonstrated to be intact, the results were taken to suggest that the VMPFC crucially subserves the assessment of the accessibility of memory contents. In addition, clinical and experimental findings showed the VMPFC to be crucially involved in ‘‘emotional decision making,” for example, when responses were based on feelings of rightness when deciding which option to pursue to gain or lose money or which action to judge as moral or immoral (Bechara, Tranel, & Damasio, 2002; Damasio, 2004; Greene & Haidt, 2002). Accordingly, previous findings may point to a crucial role of the VMPFC when decisions follow fluency. Schooler and Hertwig (2005) defined retrieval fluency in terms of how long it takes to retrieve a trace form long-term memory. Besides retrieval latency, other cognitive processes could contribute to a sense of retrieval fluency (Oppenheimer, 2008). One such process could be the ease with which associated memories are bound together. There is evidence that such a binding process would be associated with activation in the claustrum. Although activation within this area has rarely been reported in fMRI studies, Volz and von Cramon (2006), in a study on intuitive perceptual decisions, observed activation within the claustrum. They presented participants with fragmented line drawings of common objects. Within 400 ms, the participants had to indicate whether they perceived a coherent gestalt. A functional connectivity analysis revealed that the perceived gestalt was correlated with activation in the claustrum and medial orbitofrontal cortex (OFC). Medial OFC activation has been associated with a positive affective valence known to bias decisions (Bar et al., 2006; Kringelbach & Rolls, 2004; Volz, Rübsamen, & von Cramon, 2008). Further support for the critical role of the claustrum in binding can be found in anatomical and connectivity data (e.g., Crick & Koch, 2005; Fernández-Miranda, Rhoton, Kakizawa, Choi, & Alvarez-Linera, 2008; K.G. Volz et al. / Consciousness and Cognition 19 (2010) 829–837 831 Kowianski, Dziewiatkowski, Kowianska, & Morys, 1999). Because the claustrum integrates information within and across various modalities, Crick and Koch (2005) argued that the claustrum plays a key role in conscious experience. Additionally, the two-way connections between the claustrum and most, if not all, parts of the cortex as well as subcortical structures, which in turn have been suggested to subserve emotional processes, suggest that the claustrum binds disparate events into a single percept. Thus, the goal of the present fMRI study was to isolate judgments based on the fluency heuristic, so that we could map the use of fluency onto specific brain areas, which might give us a better understanding of the processes underlying adherence to this heuristic. As outlined above, it is difficult to tell whether a particular answer is based on knowledge or on fluency. To distill fluency-based judgments, we calculated two interaction contrasts specifically designed to tap fluency-based judgments (as described below) and the conjunction of these two interaction contrasts (test for a positive AND; Nichols, Brett, Andersson, Wager, & Poline, 2005). By drawing on previous results found by Hertwig et al. (2008), that people rely on the fluency heuristic when differences in recognition latencies are large and that inferences in line with the fluency heuristic take less time than inferences conflicting with the fluency heuristic, the first interaction is between differences in recognition latencies (large vs. small differences) and response time (individually determined fast vs. slow responses). The second interaction focuses on the factors that lead to participants’ judgments adhering to the predictions of the fluency heuristic (decisions agreeing vs. conflicting with the fluency heuristic) and incorrectness (incorrect vs. correct responses). Assuming knowledge-based strategies are more often right than wrong, looking at trials in which participants’ judgments adhered to the fluency heuristic but were wrong lowers the chances that their judgments were based on knowledge-based strategies. A conjunction of the two interaction contrasts should pull out trials on which judgments are more likely to be based on fluency than on knowledge-based strategies. As we are reporting new analyses of largely unanalyzed data from Volz and von Cramon (2006), we provide details of their methods. 2. Method 2.1. Participants Healthy, right-handed volunteers participated in the fMRI experiment (10 women, 8 men, mean age 25.6 years, SD 3.4, range 20–32 years). Informed consent was obtained prior to the experiment from each participant according to the Declaration of Helsinki. The local ethics committee of the University of Leipzig approved the experimental standards. Data were handled anonymously. 2.2. Stimuli, task, and experimental session On the left and right side of a screen two city names were presented simultaneously (horizontal visual angle 11°; vertical visual angle 1.7°). Participants had their left and right index fingers on left and right response buttons spatially corresponding to the stimulus locations on the screen. Within each trial a cue was presented for 500 ms, indicating the beginning of the next trial, followed by the presentation of a fixation cross for 500 ms; thereafter the two city names were presented for a maximum of 4 s during which participants’ response and reaction time were recorded. As soon as participants indicated their choice with a button press, the city names disappeared and a fixation cross was presented until the next trial started. No performance feedback was delivered whatsoever. The participants’ task (i.e., the inference task) was to indicate which city in each pair had the larger population. Each session contained 218 trials, consisting of 140 critical trials plus 48 filler trials and 30 null events, in which no stimulus was presented and the Blood Oxygen Level-Dependent BOLD response was allowed to return to a baseline state. In the filler trials participants had to indicate which of two presented words contained more vowels. All trials lasted for 8 s (i.e., four scans at a repetition time (TR) of 2 s). The onset of each stimulus presentation relative to the beginning of the first of the four scans was randomly varied (0, 500, 1500 ms) to enhance the temporal resolution of the signal captured (Birn, Cox, & Bandettini, 2002; Miezin, Maccotta, Ollinger, Petersen, & Buckner, 2000). Participants were unaware of this modulation. Following the fMRI session (i.e., outside the scanner), participants completed a recognition test in which they were presented with each particular city name and had to indicate whether they knew each city already before the experimental session. It was emphasized that participants should declare as recognized only those cities that they had heard of before the functional session. The data of the recognition test were used first to individually determine trial types in the inference task, for example, whether both cities were recognized (recognize–recognize, or RR trials), neither city was recognized (unrecognized–unrecognized, or UU trials), or one of the two cities was recognized and the other not (recognized–unrecognized, or RU trials), and second to determine recognition latency for each stimulus in the inference task. On the basis of these data, we determined which of the RR trials in the inference task were solved as predicted by the fluency heuristic, that is, when the city that was recognized faster (as determined from the recognition test) was chosen as the larger city. Following the recognition test, participants were requested to fill out a questionnaire asking for strategies; subsequently they were debriefed and thanked. Our experimental design called for the specific order of city task and recognition test, rather then counterbalancing the task order. Having the recognition test before the city task could have biased participants by making salient that we (the 832 K.G. Volz et al. / Consciousness and Cognition 19 (2010) 829–837 experimenters) were interested in whether they recognized the stimuli. In addition, by keeping this fixed task order, we were able to reliably measure the hemodynamic activity elicited by recognition judgments. Had the recognition test come first, the participants would have had to judge not only whether they recognized the city, but also whether the source of the recognition was just from the experiment or possibly from elsewhere. The additional demands of this discrimination task mean that the recognition judgments from the two task orders could draw on somewhat different brain structures and so involve brain structures that would not otherwise be involved in the application of the fluency heuristic. Nevertheless, we expected that in the recognition test, participants could reliably report whether they recognized the city from the experiment or from elsewhere. That is, they would rarely mis-categorize as recognized a city that they had only encountered in the experiment. Support for this assumption come from the studies by Pohl (2006) and Goldstein and Gigerenzer (2002). Neither study found any differences in the recognition rates that depended on the task order. However, we would expect that the recognition latencies would be faster, because the cities would have been recently seen in the inference task. As a result, we may be underestimating the absolute differences in retrieval fluency between items, but the relative differences should be preserved. For a detailed description of how the city pairs were generated, please see Volz and von Cramon (2006). 2.3. Data acquisition Imaging was performed on a 3T scanner (Siemens TRIO, Erlangen, Germany). Twenty-two axial slices (4 mm thickness, 20% spacing, field of view [FOV] 19.2 cm, data matrix of 64  64 voxels, and in-plane resolution of 3  3 mm) parallel to the bicommissural plane (AC–PC) covering the whole brain were acquired using a single-shot echo-planar imaging (EPI) sequence (TR 2 s, echo time [TE] 30 ms, flip angle 90°). One functional run with 872 time points was run with each time point sampling over the 22 slices. Prior to the functional runs, 22 anatomical T1-weighted modified driven equilibrium Fourier transform (MDEFT; Norris, 2000; Ugurbil et al., 1993) images (data matrix of 256  256 voxels, TR 1.3 s, TE 10 ms) were acquired as well as 22 T1-weighted EPI images with the same spatial orientation as the functional data. The latter were used to coregister the functional scans with previously acquired high-resolution full-brain 3-dimensional brain scans. 2.4. Data evaluation The functional imaging data were processed using the software package LIPSIA (Lohmann et al., 2001). Functional data were motioncorrected off-line with the Siemens motion–correction protocol. To correct for the temporal offset between the slices acquired in one scan, a cubic spline interpolation was applied. A temporal high-pass filter with a cut-off frequency of 1/160 Hz was used for baseline correction of the signal and a spatial Gaussian filter with 5.65 mm full-width half-maximum (FWHM) was applied. The anatomical slices were coregistered with the high-resolution full-brain scan that resided in the stereotactic coordinate system and then transformed by linear scaling to a standard size (Talairach & Tournoux, 1988). The transformation parameters obtained from this step were subsequently applied to the preprocessed functional slices so that the functional slices were also registered into the stereotactic space. This linear normalization process was improved by a subsequent processing step that performed an additional nonlinear normalization known as ‘‘demon matching.” In this type of nonlinear normalization, an anatomical 3-dimensional data set (i.e., the model) is deformed such that it matches another 3-dimensional anatomical data set (i.e., the source) that serves as a fixed reference image (Thirion, 1998). Voxel size was interpolated during coregistration from 3  3  4 mm to 3  3  3 mm. The statistical evaluation was based on a leastsquares estimation using the general linear model (GLM) for serially autocorrelated observations (random effects model; Friston, Frith, Turner, & Frackowiak, 1995; Worsley & Friston, 1995). The general linear regression performs a ‘‘precoloring” of the data; that is, it applies a temporal Gaussian smoothing with a user-specified kernel width given by the parameter FWHM. The smoothing imposes a temporal autocorrelation that determines the degrees of freedom. An event-related design was implemented; that is, the hemodynamic response function was modeled by means of the experimental conditions for each stimulus (event being onset of stimulus presentation). The design matrix was generated using a synthetic hemodynamic response function and its first and second derivative (Friston et al., 1998) and a response delay of 6 s. The model equation, including the observation data, the design matrix, and the error term, was convolved with a Gaussian kernel of dispersion of 4 s FWHM to deal with the temporal autocorrelation (Worsley & Friston, 1995). Contrast images, that is, estimates of the raw score differences between specified conditions were generated for each participant. The single-subject contrast images were entered into a second-level analysis based on Bayesian statistics (Neumann & Lohmann, 2003). In Neumann and Lohmann’s approach, posterior probability maps and maps of the effect size for the effects of interest in groups of participants are calculated on the basis of the resulting least-squares estimates of parameters for the GLM. The output of the Bayesian second-level analysis is a probability map showing the probability of the contrast being larger than zero. For visualization, a threshold of 99% was applied to the probability maps. For each participant all contrasts of interest were calculated. Reasons to use Bayesian second-level analysis for fMRI data are manifold: A comparison between the established analysis based on t statistics and Bayesian second-level analysis showed that the latter is more robust against outliers. Furthermore, the Bayesian approach overcomes some problems of null hypothesis significance testing, such as the need to correct for multiple comparisons, and this approach provides estimates for the size of an effect of interest as well as for the probability that the effect occurs in the population (Neumann & Lohmann, 2003). 833 K.G. Volz et al. / Consciousness and Cognition 19 (2010) 829–837 The main purpose of the fMRI study was to reveal brain areas involved in fluency-heuristic-based decision processes. The first analysis therefore included regressors for trials with large differences in recognition latencies, trials with small differences in recognition latencies, trials that were answered quickly (individually determined as trials below the median response time (RT)), trials that were answered slowly (individually determined as trials above the median RT), RU trials, and UU trials. Concerning the factor differences in recognition latencies, we classified trials with latencies > 400 ms as trials with large differences and trials with latencies > 100 ms and 6400 ms as trials with small differences following Hertwig et al. (2008). A second analysis included regressors for trials that were and were not answered according to the fluency heuristic (factor fluency), correctly and incorrectly answered trials (factor incorrectness), RU trials, and UU trials. The analysis that was conducted as a manipulation check for recognition- and retrieval-related activations was conducted by building special contrasts within the two described models. Note that in all analyses only trials in which the difference in recognition latencies between two recognized cities equaled or exceeded a just noticeable difference (JND) of 100 ms were included (see below for the rationale). 3. Results 3.1. Behavioral results On average, participants faced 55 RR trials (range: 34–76), 46 RU trials (range: 38–61), and 39 UU trials (range: 19–62). Volz and von Cramon (2006) specifically investigated the RU trials, where by definition the recognition heuristic is applicable but the fluency heuristic is not. Here we focus on the RR trials, where the fluency heuristic is applicable but the recognition heuristic is not. As outlined above, in this contribution, we are exclusively interested in how people deal with situations in which they recognize both of two presented objects and have to arrive at a decision about a criterion, which is very likely unknown (e.g., city size). Hence, in the following section on behavioral data we concentrate on RR trials only. For an overview of the behavioral results, please refer to Table 1. 3.2. The validity of the fluency heuristic in the present sample To determine how ecologically valid retrieval fluency was in our sample, we quantified the strength of the relationship between the retrieval fluency and the criterion as the proportion of times a more quickly recognized city indeed had a higher criterion value than the city that required more time to be recognized. Thus, fluency validity is calculated as: vf = Rf/(Rf + Wf), where Rf is the number of correct inferences made by relying on the fluency heuristic, and Wf is the number of incorrect inferences made by relying on the fluency heuristic. In the present study, the mean fluency validity was .55 (95% confidence interval [CI] = 0.497, 0.588) and exceeded chance level (.50), t(17) = 2.13; p = .048. Note that in this analysis only trials in which the difference in recognition latencies between two recognized cities equaled or exceeded a JND of 100 ms were included. The JND of 100 ms was determined following Fraisse’s (1984) suggestion, based on a thorough review of the timing literature, that durations of less than 100 ms are perceived as instantaneous. Hertwig et al. (2008) provided supporting results by showing that when differences in recognition latencies were shorter than 100 ms, people’s ability to discriminate between recognition latencies (of the two objects) dropped close to chance level. To be able to directly compare our results with those of Hertwig et al. (2008), we categorized the objective difference in recognition latencies into four equal bins: 0–99 ms, 100–399 ms, 400–699 ms, and > 700 ms and calculated fluency validity as a function of these four bins (see Table 1). For the first three bins, we replicated Hertwig et al.’s findings showing that there is a tendency that the larger the objective difference in recognition latencies, the higher the fluency validity. Yet, the linear trend did not continue for the last bin (>700 ms), F(3, 14) = .508; p = .634. Given recent results on fluency validities in different environments, fluency validity in the present sample can be considered moderate and hence participants could at least theoretically infer the distal properties of the world (Hertwig et al., 2008). 3.3. Participants’ accordance with the fluency heuristic To determine to what degree people’s inferences agreed with the fluency heuristic in the present sample, we computed for each participant the percentage of inferences that were in line with the fluency heuristic among all cases in which it could Table 1 Means (SD) of fluency heuristic validity, fluency heuristic accordance, percent correct overall, and percent correct when applying the fluency heuristic. Fluency heuristic validity Fluency heuristic accordance % Correct % Correct with the fluency heuristic a Overalla 0–99 ms 100–399 ms 400–699 ms >700 ms 0.55 (0.21) 0.68 (0.05) 62.8 (19.5) 81.3 (14.8) 0.49 (0.13) 0.48 (0.13) 60.4 (20.4) 54.2 (32.9) 0.53 (0.13) 0.64 (0.09) 65.3 (11.1) 79.3 (10.2) 0.58 (0.24) 0.75 (0.17) 63.6 (23.3) 83.7 (14.9) 0.53 (0.26) 0.74 (0.17) 59.5 (24.2) 80.9 (19.2) Data for this category exclude trials with a just noticeable difference of 100 ms or less. 834 K.G. Volz et al. / Consciousness and Cognition 19 (2010) 829–837 Interaction difference in re cognition latencies & response time A x = 31 claustrum Interaction fluency and incorrectness B RCZ y=6 claustrum x = 34 claustrum Conjunction analysis of the contrasts shown in A & B C y=6 claustrum x = 31 Fig. 1. Group averaged activations are shown on four slices taken from an individual brain normalized and aligned to the Talairach stereotactic space. For visualization a threshold of 99% was applied to all probability maps. (A) Results of the interaction contrast differences in recognition latencies (large vs. small) and response time (individually determined fast slow responses). To understand which factor combination is driving the effects, the mean percent signal changes with standard errors are reported for all four conditions. (B) Results of the interaction contrast fluency heuristic (decisions agreeing vs. conflicting with the fluency heuristic) and incorrectness (incorrect vs. correct responses). To understand which factor combination is driving the effects, the mean percent signal changes with standard errors are reported for all four conditions. (C) Result of the conjunction analysis of the two interaction contrasts. Abbreviations: amla: above median and large differences in recognition latencies; amsm: above median and small differences in recognition latencies; bmla: below median and large differences in recognition latencies; bmsm: below median and small differences in recognition latencies; RCZ: rostral cingulate zone; flc: fluency heuristic trials and correct; fli: fluency heuristic trials and incorrect; nflc: no fluency heuristic trials and correct; nfli: no fluency heuristic trials and incorrect. 835 K.G. Volz et al. / Consciousness and Cognition 19 (2010) 829–837 be applied (i.e., among all RR trials), excluding trials with a JND of 100 ms or smaller. The mean fluency heuristic accordance in the present sample was .678 (CI: 0.649, 0.707). The interindividual variation in the proportion of judgments that agreed with the fluency heuristic was rather moderate. The rate of individuals’ fluency heuristic accordance ranged between .58 and .79. Hence, none of the participants appeared to have systematically decided against fluency (i.e., fluency accordance rate below .50), nor did anyone use the fluency heuristic all the time. The fluency heuristic applicability ranged between .24 and .54 with a mean of .39 (CI: 0.343, 0.437) and hence in more than a third of all inferences, the fluency heuristic was applicable. In Table 1, we report fluency heuristic accordance rates as a function of differences in recognition latencies. Accordance rates increased with larger differences in recognition latencies, F(3, 14) = 11.67; p = .0001, but did not differ much between the last (>700 ms) and the second to last (400–699 ms) bin, t(17) = .876. Again, except for the last bin, we replicated the results of Hertwig et al. (2008). 3.4. Participants’ performance in the inference task Overall, participants scored a median .63 accuracy (CI: 0.580, 0.683) calculated as percent correct of all RR trials irrespective of response strategy (trials with a JND of 100 ms or smaller were excluded). In Table 1, we report percent correct of all RR trials as a function of differences in recognition latencies. Overall, performance did not differ substantially subject to differences in recognition latencies, F(3, 14) = .322; p = .76. Yet, this pattern changes when we look at percent correct made by the fluency heuristic: There is a clear tendency that the larger the objective difference in recognition latencies, the higher the percent correct, F(3, 11) = 5.50; p = .010. Again, the linear trend did not continue for the last bin. Inferences in line with the fluency heuristic were made faster (mean RT: 2842 ms) than inferences conflicting with the fluency heuristic (mean RT: 2944 ms), t(17) = 3.47, p = .003. In all analyses where the variable of interest was split according to differences in recognition latencies, the last bin (>700 ms) seemed to be the odd one out. One reason for this finding might be the restriction of response time. Before the experimental session, participants were informed that the two city names would be presented for a maximum of 4 s during which their response would be recorded. If no response was given within the 4 s, then the trial would be counted as a noresponse trial. Participants’ mean response time was 3055 ms when differences in recognition latencies exceeded 700 ms. Thus, when it took participants fairly long to recognized one of the city names, the overall decision time was almost expended. Accordingly, this time pressure might have eventually led to guessing. 3.5. Imaging results To test for the specific neural correlates of fluency-heuristic-based decision processes, we calculated two interaction contrasts and their conjunction. Based on previous findings on fluency-heuristic-based decisions (Hertwig et al., 2008), the interaction contrasts were designed to capture fluency-based inferences, relatively uncontaminated by extensive knowledge about the options. 3.5.1. Interaction between differences in recognition latencies and response time The interaction contrast of the factors large versus small differences in recognition latencies and fast versus slow response times revealed activation bilaterally within the dorsal claustrum and the left anterior insula; bilaterally within the anterior thalamus region and amygdala; and within the anterior portion of the left superior temporal gyrus, left supramarginal gyrus, and right cuneus (see Fig. 1A and Table 2). When we plotted the mean percent signal change in the dorsal claustrum, results revealed specifically those trials with large differences in recognition latencies and fast response times to elicit activation within this area. 3.5.2. Interaction between fluency and incorrectness The interaction contrast of the factors fluency heuristic trials versus no fluency heuristic trials and incorrect versus correct responses revealed activation within the right dorsal claustrum and bilaterally within the posterior rostral cingulate zone Table 2 Anatomical specification, cluster size (mm3), and Talairach (x, y, z) coordinates of significantly activated voxels of the interaction contrast differences in recognition latencies (large vs. small differences) and response time (fast vs. slow responses). Anatomical specification mm3 L dorsal claustrum R dorsal claustrum L anterior insula R anterior thalamus region L anterior thalamus region L superior temporal gyrus R amygdala L amygdala L supramarginal gyrus R cuneus (occipital lobe) 81 81 162 1755 621 324 108 81 162 351 Note: Only clusters of at least five contiguous voxels are reported. L: left, R: right. x y 35 31 26 13 11 50 19 23 47 4 z 6 4 15 3 12 36 6 9 48 84 3 3 12 12 9 9 9 9 27 27 836 K.G. Volz et al. / Consciousness and Cognition 19 (2010) 829–837 (RCZ) and the left superior and middle temporal gyrus (see Fig. 1B and Table 3). When we plotted the mean percent signal change in the right dorsal claustrum, results revealed both incorrect fluency trials and correct nonfluency trials to specifically elicit activation within this area. 3.5.3. Conjunction analysis of the two interaction contrasts To test for those regions that are commonly activated across the two interaction contrasts, we calculated a conjunction analysis, that is, a test for a logical AND (Nichols et al., 2005). Activation was revealed solely within the right dorsal claustrum (x = 31, y = 6, z = 9; mm2 = 135; see Fig. 1C). 4. Discussion The present imaging study was conducted to investigate the cognitive processes underlying the fluency heuristic. Simply stated, when people use the fluency heuristic they will infer that the more easily retrieved item has the higher criterion value (Schooler & Hertwig, 2005). Our method was to isolate fluency-heuristic-based judgments, to facilitate mapping specific brain areas subserving the heuristic. We started with two hypotheses: The first one assumed that the VMPFC is a neural correlate of fluency-heuristic-based decisions and was derived from previous imaging findings on phenomenologically similar kinds of decisions. The second hypothesis suggested that the claustrum is a neural correlate of fluency-heuristic-based decisions and was derived from an account of retrieval fluency based on successfully binding memory traces. We ran two interaction contrasts specifically designed to pull out trials that were very likely based on the fluency heuristic. The first interaction included the factor difference in recognition latencies (large vs. small differences) and the factor response time (fast vs. slow responses). The second interaction included the factor adherence to the fluency heuristic and the factor correctness. Both specific interaction contrasts elicited activation within the dorsal claustrum, as did a conjunction analysis of the two interaction contrasts. No activation within the VMPFC was observed. We take the latter result to indicate that fluency-heuristic-based decisions are dissimilar to metacognitive judgments about the felt-rightness of a memory. Thus, the present results revealing the dorsal claustrum as the most consistent neural correlate of decisions that are most likely to rely on the fluency heuristic support our second hypothesis. Hitherto, comparative anatomical, tractographic, electrophysiological, tracing, histological studies and the like have dominated the investigation of the claustrum, yet claustral activation has only seldom been reported in fMRI studies (Volz & von Cramon, 2006). Recently, Crick and Koch (2005) speculated that the claustrum gives rise to integrated conscious percepts. Morphologically, the claustrum is a thin, irregular band of gray matter, hidden beneath the inner surface of the neocortex and in close proximity to the insula. Macroscopically, the claustrum is divided into a dorsal (or posterosuperior) part and a ventral (anteroinferior) part (Fernández-Miranda et al., 2008; Kowianski et al., 1999). The dorsal claustrum lies between the putamen, from which it is separated by the external capsule, and the insular cortex, from which it is separated by the extreme capsule (Edelstein & Denaro, 2004; Fernández-Miranda et al., 2008). Considering the claustrum’s tractography may hint at its function. Tractography studies have revealed extensive neocortical connections and a topographical organization in the dorsal claustrum, where posterior cortical areas project to the posterior part of the dorsal claustrum and more anterior parts of cortical areas converge onto the anterior part of the claustrum, thereby forming (overlapping) cortical projection zones (Fernández-Miranda et al., 2008; Morys, Narkiewicz, & Wisniewski, 1993). Almost all of the claustrum-to-cortex projections are reciprocated, with only a few exceptions (e.g., V1; Sherk, 1986). In addition, the claustrum maintains two-way connections with subcortical structures involved in the limbic system, such as the amygdala and prepiriform cortex (Fernández-Miranda et al., 2008). These data were taken to suggest that the claustrum is a structure that interconnects the senses, providing them direct access to each other (Ettlinger & Wilson, 1990). In doing so, the claustrum plays a crucial role as a polymodal structure engaged in the transfer of information to and from various cortical regions (Kowianski et al., 1999). Crick and Koch’s (2005) metaphor for the claustrum is that of a conductor coordinating the players in an orchestra, where the musicians are the various cortical regions. The conductor is responsible for binding the performances of individual musicians into an integrated, synchronous whole. Similarly, the claustrum rapidly combines the different attributes of objects both within and across modalities so that an integrated whole is experienced, rather than a collection of isolated attributes. Given the findings on the function of the claustrum so far, claustral activation during fluency-heuristic-based decisions may reflect the experience of an integrated signal associated with the retrieval of one or more memory records. For example, Table 3 Anatomical specification, cluster size (mm3), and Talairach (x, y, z) coordinates of significantly activated voxels of the interaction contrast fluency (decisions agreeing vs. conflicting with the fluency heuristic) and incorrectness (incorrect vs. correct responses). Anatomical specification mm3 R claustrum L rostral cingulate zone, posterior part R rostral cingulate zone L superior temporal gyrus L middle temporal gyrus, posterior part 216 1107 216 297 243 Note: Only clusters of at least five contiguous voxels are reported. L: left, R: right. x y 34 10 12 53 42 z 6 6 3 30 62 9 39 42 21 9 K.G. Volz et al. / Consciousness and Cognition 19 (2010) 829–837 837 episodic, semantic, and visual information may all be rapidly combined and bound in the claustrum. This signal may then suffice to guide inferences, and accordingly, be experienced as fluency. This differs from the way fluency was treated by Schooler and Hertwig (2005), where it was taken to simply represent the speed with which a single memory record is retrieved. The present finding corresponds to findings on perceptual fluency: In the study by Volz and von Cramon (2006), participants had to judge the coherence of pictorial stimuli, and claustral activation was suggested to represent the temporal synchronization process during perceptual fluency judgments. Together, we take the present results to indicate that inferences that are made according to the fluency heuristic specifically rely on an integrated signal at the fast time scale. Acknowledgments We thank Thomas Dratsch for helping with the behavioral data analyses and Anita Todd for editing a draft of this article. References Anderson, J. R., & Schooler, L. J. (1991). Reflections on the environment in memory. Psychological Science, 2, 396–408. Bar, M., Kassam, K. S., Ghuman, A. S., Boshyan, J., Schmid, A. M., Dale, A. M., et al (2006). Top-down facilitation of visual recognition. Proceedings of the National Academy of Sciences of the United States of America, 103, 449–454. Bechara, A., Tranel, D., & Damasio, A. R. (2002). The somatic marker hypothesis and decision-making. In F. Boller & J. Grafman (Eds.), Handbook of neuropsychology (2nd ed., pp. 117–143). Amsterdam, Netherlands: Elsevier. Birn, R. M., Cox, R. W., & Bandettini, P. A. (2002). Detection versus estimation in event-related fMRI: Choosing the optimal stimulus timing. NeuroImage, 15, 252–264. Crick, F. C., & Koch, C. (2005). What is the function of the claustrum? Philosophical Transactions of the Royal Society of London B: Biological Sciences, 360, 1271–1279. Damasio, A. R. (2004). William James and the modern neurobiology of emotion. In D. Evans & P. Cruse (Eds.), Emotion, evolution, and rationality (pp. 3–14). Oxford, England: Oxford University Press. Edelstein, L. R., & Denaro, F. J. (2004). The claustrum: A historical review of its anatomy, physiology, cytochemistry and functional significance. Cellular and Molecular Biology, 50, 675–702. Ettlinger, G., & Wilson, W. A. (1990). Cross-modal performance. Behavioral processes, phylogenetic considerations and neural mechanisms. Behavioural Brain Research, 40, 169–192. Fernández-Miranda, J. C., Rhoton, A. L., Kakizawa, Y., Choi, C., & Alvarez-Linera, J. (2008). The claustrum and its projection system in the human brain: A microsurgical and tractographic anatomical study. Journal of Neurosurgery, 108, 764–774. Fraisse, P. (1984). Perception and estimation of time. Annual Review of Psychology, 35, 1–36. Friston, K. J., Fletcher, P., Josephs, O., Holmes, A., Rugg, M. D., & Turner, R. (1998). Event-related fMRI: Characterizing differential responses. NeuroImage, 7, 30–40. Friston, K. J., Frith, C. D., Turner, R., & Frackowiak, R. S. (1995). Characterizing evoked hemodynamics with fMRI. NeuroImage, 2, 157–165. Goldstein, D. G., & Gigerenzer, G. (2002). Models of ecological rationality: The recognition heuristic. Psychological Review, 109, 75–90. Greene, J., & Haidt, J. (2002). How (and where) does moral judgment work? Trends in Cognitive Sciences, 6, 517–523. Hertwig, R., Herzog, S. M., Schooler, L. J., & Reimer, T. (2008). Fluency heuristic: A model of how the mind exploits a by-product of information retrieval. Journal of Experimental Psychology: Learning, Memory, and Cognition, 34, 1191–1206. Jacoby, L. L., & Dallas, M. (1981). On the relationship between autobiographical memory and perceptual learning. Journal of Experimental Psychology: General, 110, 306–340. Kelley, C. M., & Jacoby, L. L. (1998). Subjective reports and process dissociation: Fluency, knowing and feeling. Acta Psychologica, 98, 127–140. Kowianski, P., Dziewiatkowski, J., Kowianska, J., & Morys, J. (1999). Comparative anatomy of the claustrum in selected species: A morphometric analysis. Brain, Behavior and Evolution, 53, 44–54. Kringelbach, M. L., & Rolls, E. T. (2004). The functional neuroanatomy of the human orbitofrontal cortex: Evidence from neuroimaging and neuropsychology. Progress in Neurobiology, 72, 341–372. Lohmann, G., Muller, K., Bosch, V., Mentzel, H., Hessler, S., Chen, L., et al (2001). LIPSIA—A new software system for the evaluation of functional magnetic resonance imaging of the human brain. Computerized Medical Imaging and Graphics, 25, 449–457. Marewski, J. N., & Schooler, L. J. (2010). Cognitive niches: An ecological model of emergent strategy selection, submitted for publication. Miezin, F. M., Maccotta, L., Ollinger, J. M., Petersen, S. E., & Buckner, R. L. (2000). Characterizing the hemodynamic response: Effects of presentation rate, sampling, procedure, and the possibility of ordering brain activity based on relative timing. NeuroImage, 11, 735–759. Morys, J., Narkiewicz, O., & Wisniewski, H. M. (1993). Neuronal loss in the human claustrum following ulegyria. Brain Research, 616, 176–180. Neumann, J., & Lohmann, G. (2003). Bayesian second-level analysis of functional magnetic resonance images. NeuroImage, 20, 1346–1355. Nichols, T., Brett, M., Andersson, J., Wager, T., & Poline, J. B. (2005). Valid conjunction inference with the minimum statistic. NeuroImage, 25, 653–660. Norris, D. G. (2000). Reduced power multislice MDEFT imaging. Journal of Magnetic Resonance Imaging: JMRI, 11, 445–451. Oppenheimer, D. M. (2008). The secret life of fluency. Trends in Cognitive Sciences, 12, 237–241. Pohl, R. F. (2006). Empirical tests of the recognition heuristic. Journal of Behavioral Decision Making, 19, 251–271. Schnyer, D. M., Nicholls, L., & Verfaellie, M. (2005). The role of VMPFC in metamemorial judgments of content retrievability. Journal of Cognitive Neuroscience, 17, 832–846. Schnyer, D. M., Verfaellie, M., Alexander, M. P., Lafleche, G., Nicholls, L., & Kaszniak, A. W. (2004). A role for right medial prefrontal cortex in accurate feeling of knowing judgments: Evidence from patients with lesions to frontal cortex. Neuropsychologia, 42, 957–966. Schooler, L. J., & Anderson, J. R. (1997). The role of process in the rational analysis of memory. Cognitive Psychology, 32, 219–250. Schooler, L. J., & Hertwig, R. (2005). How forgetting aids heuristic inference. Psychological Review, 112, 610–628. Sherk, H. (1986). The claustrum and the cerebral cortex. In E. G. Jones & A. Peters (Eds.). Cerebral cortex (Vol. 5, pp. 467–499). New York, NY: Plenum Press. Talairach, P., & Tournoux, J. (1988). Co-planar stereotactic atlas of the human brain. Stuttgart, Germany: Thieme. Thirion, J. P. (1998). Image matching as a diffusion process. An analogy with Maxwell’s demons. Medical Image Analysis, 2, 243–260. Ugurbil, K., Garwood, M., Ellermann, J., Hendrich, K., Hinke, R., Hu, X. P., et al (1993). Imaging at high magnetic fields: Initial experiences at 4 T. Magnetic Resonance Quarterly, 9, 259–277. Volz, K. G., Rübsamen, R., & von Cramon, D. Y. (2008). Cortical regions activated by the subjective sense of perceptual coherence of environmental sounds: A proposal for a neuroscience of intuition. Cognitive, Affective, & Behavioral Neuroscience, 8, 318–328. Volz, K. G., Schooler, L. J., Schubotz, R. I., Raab, M., Gigerenzer, G., von Cramon, D. Y., et al (2006). Why you think Milan is larger than Modena: Neural correlates of the recognition heuristic. Journal of Cognitive Neuroscience, 18, 1924–1936. Volz, K. G., & von Cramon, D. Y. (2006). What neuroscience can tell about intuitive processes in the context of perceptual discovery? Journal of Cognitive Neuroscience, 18, 1–11. Worsley, K. J., & Friston, K. J. (1995). Analysis of fMRI time-series revisited—Again. NeuroImage, 2, 173–181.

Consciousness and Cognition 19 (2010) 838–846 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Memory and content q Gottfried Vosgerau * Institut für Philosophie, 23.21.00.46B, Heinrich-Heine-Universität, Universitätsstr. 1, 40225 Düsseldorf, Germany a r t i c l e i n f o Article history: Available online 23 August 2010 Keywords: Memory Mental representation Content Behavior Functionalism a b s t r a c t The paper argues that any theory of content has to adopt a ‘‘functionalistic core” to concord with the cognitive sciences. This functionalistic core requires that representations are defined as substitutes in functions that describe the flexible behavior to be explained by the representation. The content of a representation can thus only be determined if the representation is ‘‘in use”, i.e. if it is an argument in such a function. The stored entities in memory are not in use while they are stored, and hence cannot be assigned a specific content. The term ‘‘template” is introduced to describe stored entities in memory. The discussion of some implications of this result show that some deep philosophical problems follow from this argument as well as consequences for empirical research on memory. Ó 2010 Elsevier Inc. All rights reserved. 1. Introduction Memory is commonly assumed to be preservative storage of content.1 Just like a story can be ‘‘stored” by writing it down, the contents of our mental states can be stored in memory. Naïvely speaking, we take a content as it is and put it somewhere safe to be retrieved whenever needed. Starting with this picture, we can proceed with differentiating between different kinds of memories that fulfill different functions—some examples are: short-term vs. long-term memory, episodic vs. semantic memory, implicit vs. explicit memory, procedural vs. declarative memory, working memory, visual buffers. Although this naïve picture of memory is somewhat neat, there are some deep problems with it. The general strategy of this paper is to argue that any theory of (representational) content has to adopt what I call a ‘‘functionalistic core” if it is to capture the explanatory role of representations in behavioral sciences. This means that if the content of a representation is taken to explain (flexible) behavior, then we have to take the target of the behavior as the starting point for determining the content. The very abstract characterization of this idea is formulated in terms of substitution in functions. This abstract formulation cannot be taken as a theory of representational content itself, but it is sufficiently general to be a frame for theories of representational contents. On the basis of this general frame, it is then argued that we cannot assign certain contents to stored entities (whatever they are). Thus, it is concluded, there is no content in memory (and so, strictly speaking, no representations). Based on this negative outcome of the discussion, I will introduce the term ‘‘template” to sketch—again in a general way— a strategy of overcoming the obvious contradiction with the classical intuition regarding memory. Thereby, the introduced notion cannot provide a theory but only some constraints for a theory of memory. However, already these general considerations have some important implications which will be discussed in the last section. q This article is part of a special issue of this journal on Self, Other and Memory. * Fax: +49 (0)211 81 14338. E-mail address: gottfried.vosgerau@uni-duesseldorf.de URL: http://www.phil-fak.uni-duesseldorf.de/philo/personal/thphil/vosgerau/. 1 See, for example, Burge (1993) and, for a critical discussion and rebuttal of this picture: Matthen (2010). 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.06.021 G. Vosgerau / Consciousness and Cognition 19 (2010) 838–846 839 d d nest Fig. 1. Homing behavior of ants. The problem I want to discuss in this paper is, however, confined to long-term memory and does not apply to short-term memory or working memory, let alone ultra-short-term buffers (like the visual buffer). This has to be qualified even further: The discussion is confined to ‘‘entries” in long-term memory that are not actually remembered, i.e. to non-occurrent memory items.2 Indeed, the points made in the paper address all forms of non-occurrent ‘‘mental representations” (I assume that all of these, which include the beliefs in our ‘‘belief-system”, can be said to be ‘‘in memory”3). Thereby, ‘‘occurrent” is not understood in terms of consciously available (background-beliefs that play a role in the tokening of a certain chain of reasoning, e.g. do not have to be consciously available but are occurrent in this sense) but rather in the sense of being currently involved in some mental process (consciously or unconsciously). To be more precise: a thought is occurrent if it has to be mentioned in an explanation of the behavior. My ‘‘stored belief” that Paris is the capital of France, e.g. is non-occurrent most of the time, since most of the time my behavior can be explained without reference to this belief. Of course, sometimes the explanation of my behavior has to refer to this belief, and in these cases we have to assume that this belief is indeed occurrent in the sense that it is involved in my reasoning. However, this does not mean that it necessarily is a conscious thought in such cases—it is very well possible that it stays an unconscious background belief. To sum up, this paper is concerned with non-occurrent ‘‘stored” representations that are not involved in any way in actual mental processes.4 Furthermore, let me add some notes on what is not in the scope of the paper. Some may think that having content (being intentional, if you want) is the hallmark of being mental. Thus, if something does not have content, it cannot be mental either. So, for some it might seem that I argue that memory is not part of the mind. However, this issue is, in large parts, independent of the arguments presented here and thus cannot be discussed in this paper. Therefore, I will be completely silent about the definition of the mental and its boundaries. (Indeed, I think that the mental cannot be defined with the notion of intentionality, and I would even be tempted to regard the arguments presented here as a reductio of such definitions.) Moreover, memory has often been discussed in epistemological context where the main question is whether and how memory can justify an occurrent belief. My considerations are not carried out in such a context—this paper is interested in whether the notion of stored content is compatible with the notion of representation as it is used in the behavioral sciences. However, the arguments presented here have some implications for the epistemological discussion, but they are not based and thus independent of this discussion. 2. Theories of content The problem of determining content5 is as old as philosophy itself. Accordingly, many more or less detailed theories of content have been formulated, and I will not discuss any account here. In contrast, let me start with some reflections on why we talk about representations and contents at all which will lead to a very general constraint for every theory of representation, namely that it has to acknowledge that the content of representations (and thereby the status of something as a representation) can only be determined if the representation is ‘‘in use”. 2.1. The role of representations Consider the so-called homing behavior of the ant (Gallistel, 1993): After an unsystematic search for food, the ant is able to go back to its nest on a straight line (see Fig. 1 left). If the ant is displaced before returning, it will take a parallel path to a point where the nest would have been, if it had been displaced in the same direction and distance d (see Fig. 1 right). 2 I do not want to suggest that the term ‘‘memory” should not be used for occurrent memories, i.e. occurrent mental representations based on memory. However, for the sake of readability, I will refer to non-occurrent memories with the term ‘‘memory” if not otherwise explicitly mentioned. 3 Thus, the discussion concerns all kinds of memory, not just episodic memory. 4 Thus, if you have a holistic picture of the mind in the sense that you belief that every potential representation of a subject has to figure in every explanation of her or his behavior, then my distinction between occurrent and non-occurrent representations does not make any sense to you, and so the whole paper is worthless to you. However, I will argue for the role of representations in explanations of behavior below in a way that makes the holisitic picture very unplausible. 5 In this paper, I will use ‘‘content” to refer to representational content exclusively. 840 G. Vosgerau / Consciousness and Cognition 19 (2010) 838–846 In order to explain this kind of behavior, we cannot use simple stimulus–response-patterns, because the ant obviously does not rely on nest-stimuli in finding its way home (otherwise, it would not take the parallel path in the displacement condition). Hence, the behavior can only be explained if some internal states of the ant are assumed. These internal states, whatever they are in detail, are what is called a (mental) representation. In this case, the ant is said to have a representation of the location of the nest. (In fact, detailed research on the homing behavior revealed that the representation of the ant consists of a registration of the path’s angle to the sun and the number of steps the ant has to make.) The reason why the concept of representation is introduced in cognitive science is, as demonstrated with the case of the homing behavior, to explain flexible behavior, i.e. behavior that cannot be explained by rigid stimulus–response-patterns. The explanation of flexible behavior requires the assumption that some internal state of the behaving system exists which is then said to be a representation. Contrast this with the behavior of the moth, which flies towards light sources. If the moth is displaced on its way, it will not take a parallel route but continue to fly directly to the light source. If the light source is removed, the moth will simply stop this kind of behavior and not search for it or continue to fly to the point where the light source had been. Thus, this behavior of the moth is not flexible since it can be explained by a simple rigid response (flying) to a stimulus (light). Therefore, in this kind of rigid behavior, no representations are involved. (The difference between rigid behavior and flexible behavior can also be described in terms of misrepresentation: Flexible behavior is the kind of behavior which allows for unsuccessful cases, e.g. the displacement condition above. Rigid behavior, on the other hand, does not allow for such cases—there is no way in which the moth can err in flying towards a light source. The possibility of explaining such unsuccessful cases depends on the possibility of misrepresentation. If there is no representation, there cannot be misrepresentation either. And—as argued for by Dretske (1986), there is no representation without the possibility of misrepresentation. In other words: The possibility of misrepresentation is a conceptual necessity for every representation.) Why is the internal state of the ant a representation of the location of the nest, and not, say, of trees? Or: How is the content of a representation determined? The internal state of the ant, the representation, is assumed to explain the homing behavior of the ant. Since the homing behavior is directed towards the nest (the target object of the behavior is the nest), we assume that the internal state represents this very target object (namely the nest). It is, to borrow a term from Cummins (1996), a proxy or a stand-in for the nest.6 Note that the unsuccessful behavior displayed in the displacement condition is still of the same type, namely of the type ‘‘homing behavior”. If it would not be, empirical research on the homing behavior could not draw inferences from the displacement conditions. In particular, the content of misrepresentation is the same as in the case of successful representations: Misrepresentations are characterized as cases where content and the represented reality fall apart.7 Thus, even in cases where the ant does not reach the nest, it still has a nest-representation, which is a misrepresentation, because the nest is not there (and thus content and represented world fall apart). Hence, the target object of the homing behavior is still the nest, even if the behavior is not successful and the nest is not reached.8 Let me illustrate this very important point with another example. In now classical experiments, Cheng (1986) trained rats in a rectangular arena to find food in one specific corner. In the experimental phase, the rats showed in most of the trails the following behavior: they went either to the correct corner or to the corner diagonally opposite of the correct corner. Cheng (1986) concluded that rats are able to navigate according to the geometrical layout of the arena. The rationale of this (typical behavioral) experiment is the following: after training, the setup is so modified that systematic cases of unsuccessful behavior are shown. These cases are then interpreted as cases of misrepresentation. Because of the systematic nature of these misrepresentations, such cases allow insight into the structure of the representations. Two things are important here: first, the behavior has to be interpreted as being of the same kind, namely being a search-for-food-behavior. If the cases of misrepresentation are not interpreted in this way, they could not give us any insight into the structure of the rats’ representations. Therefore, even in these unsuccessful cases, the target of the behavior still is the food (and this is the reason why we can assume that it is also the target or the content of the (mis-)representation). Thus, the individuation of behavior is already the crucial move for determining the content of the representation which ultimately explains the (kind of) behavior. Second, although this experiment is based on learning and so on memory, we cannot conclude from these data that this or that content is stored in memory. The representation that explains the behavior is the occurrent representation that the rat forms in the experimental trails. This representation is, surely, based on memory. However, this experiment does not tell us anything about what is stored and how the occurrent representation is based on memory. The explanation of flexible behavior thus runs along the following lines. A cognitive system shows a certain kind of behavior (e.g. homing behavior, food-search-behavior). The classification of a behavior token is made according to the criteria for behavior individuation, which are certainly not easy to determine, but which are not to be discussed here (see also footnote 8). Importantly, the individuation of behavior is independent from the question whether the target of the behavior is present or not: the ant, e.g. shows homing behavior even in cases where the nest has been removed (or the ant is displaced), and the 6 The use of the term ‘‘target” is also inspired by Cummins (1996), although there is the difference that Cummins speaks of the target of a representation and I speak of the target of behavior. These two are not the same, since an unsuccessful behavior can still have a well defined target. 7 Otherwise, the famous disjunction problem (Fodor 1987) could not even be formulated—it arises if some X causes a (mis-)representation with content Y. 8 Of course, the individuation criteria for behavior types are not at all clear or easy to describe. However, ‘‘behavior” is the more basic term since this is ultimately the explanandum of cognitive sciences. In this sense, the individuation of representations is dependent on the individuation of behavior. This is also the reason why Cummins (1996) general critique of use-based theories of representation does not apply to the present account: What Cummins rightly criticizes is theories that rely on a notion of successful use; however, ‘‘use” here is meant to refer to the use in the guidance of behavior—successful or not. For place limitations, I cannot discuss the problem of behavior individuation here. G. Vosgerau / Consciousness and Cognition 19 (2010) 838–846 841 rat shows food-search-behavior even if no food was hidden. Thus, the target of behavior is determined by identifying a behavior as being of a certain kind. This behavior can now be explained by claiming that the cognitive system has a representation of the target of behavior: the ant shows homing behavior because it has a representation with the content that can be linguistically expressed by ‘‘the location of its nest”; the rat shows search-food-behavior because it has an (occurrent) representation with the content that can be linguistically expressed by ‘‘the location of food”. The content of a representation thus has a specific explanatory role: it explains why cognitive systems are able to show behavior that is directed at targets about which the cognitive system does not have sensory information. If the content of the representation would not be ultimately determined by the target of the behavior, it could not fulfill this explanatory role. This kind of explanation might sound vacuous (cf. Cummins, 1996), because the content of the representation is dependent of the behavior which is to be explained. I do not think that this is a serious problem for two reasons9: First, there are other explanations that involve similar twists and which, nevertheless, seem unproblematic. The gravitational mass of an object, for example, is determined by the force between the object and other masses. Nevertheless, the following explanation is unproblematic: the object attracts other masses because it has this and that mass. Second, the explanations I gave above are not the end of the story, of course. Behavioral experiments try to further analyze the structure of the representation, and thereby provide more and more detailed explanations. For example, experiments revealed that the ant is able to represent the nest in terms of the path’s angle to the sun and the number of steps it has to make to reach the nest. However, such refinements in explanation do not touch the core, namely that this representation really represents the nest (and not just angles and steps), and that the ant reaches the nest because it has this nest-representation. 10 To sum up, the content of a representation is ultimately determined by the target object of the behavior to be explained. Of course, this kind of determination is quite coarse-grained. Therefore, further criteria are often introduced (e.g. the criterion of simultaneous believability; Frege, 1892); nevertheless, none of these criteria can explain why the ant has a nest-representation rather than a tree-representation. In other words: Every determination of content has to start with the target object of the behavior to be explained, whatever further criteria are then added. In this sense, the ‘‘core” of every content individuation is based on the target object of the behavior in question (see also Vosgerau, Schlicht, & Newen, 2008). This ‘‘core” of content determination can be formulated in functionalistic terms (Vosgerau, 2008, 2009): If the behavior is described in terms of functions (in the mathematical sense), representations can be defined as substitutes in these functions. For example, the behavior of the ant might be described by a function mapping the actual location of the ant and the location 0 of the nest onto a new location of the ant which is identical to the location of the nest f: hlocant, locnesti # locant , whereby (usu0 ally) locant ¼ locnest : The representation of the location of the nest, which guides the ant, can now be defined as a substitute for h i 0 nest : hlocant ; repnest i # locant : Note that this formulation is functionalistic in the widest sense, as it is not confined locnest : f rep locnest to inputs, states, and outputs of brains (as in ‘‘classical” functionalism; Putnam, 1975; Block, 1978), nor to teleosemantics (Dretske, 1986; Millikan, 1984). It is formulated in such an abstract way that it is possible to cover different cases, e.g. evolutionary functions like the mapping of the presence of danger onto fleeing behavior, where the presence of danger can be substituted by the beaver’s splash (see Millikan, 1984, 1994); or behavioristic functions like the mapping of red light onto the press of a button in an experiment, where the red light can be substituted by the firing of certain cells in the retina. Moreover, it is able to include cases of misrepresentation (as, e.g. the case in which the ant is displaced), since in such cases the same kind of behavior is displayed which is described by the same function.11 Of course, this abstract functionalistic core is not apt to determine content (see Vosgerau, 2008, 2009 for details), but it is a core without which representation cannot be defined—at least not in a way that it could fulfill its role in the explanation of behavior.12 However, the first and ultimate reason to introduce the talk about representations is to explain flexible behavior. Thus, any definition of representation that is not able to cover the explanatory role of representations for behavior cannot be adequate with respect to the concept of representation as it is used in the behavioral sciences. Therefore, I take this functionalistic characterization of representations to be a necessary core feature of every definition of representation and content. 2.2. The ‘‘use” of representations The basic idea of mental representation as captured in the ‘‘functionalistic core” implies that representations and their contents can only be determined when they are ‘‘in use”, i.e. when they are actually substituting the represented entity 9 Due to space limitations, I cannot discuss the nature of explanation in detail here. There might be cases, where further investigations do change the content we ascribe. However, in these cases the classification of the behavior is changed as well. Take, e.g. the snapping-behavior of the frog: Lettvin, Maturana, McCulloch, & Pitts (1959) showed that frogs do not react to flies but to black moving dots. Thus, their behavior cannot be described as snapping at flies, but has to be re-described as snapping at black moving dots. (You might then add a story of why this behavior makes evolutionary sense . . .) 11 As argued for above, it is crucial to the empirical investigation of representation that cases of unsuccessful behavior/misrepresentation are still classified as the same kind of behavior, since otherwise the systematic variation of conditions that produce such cases would not provide information about successful cases. 12 In this sense, it is not a theory of representation but rather a frame for theories of representation. Thus, it does not compete with other theories—rather, it presents constraints for theories such that theories that do not include, in some way or another, this functionalistic core are ruled out as good candidates for successful theories of representation. 10 842 G. Vosgerau / Consciousness and Cognition 19 (2010) 838–846 in a function to guide behavior. As long as they (the vehicles) are not used, they cannot be identified as representations with this and that content; this is to say that they are not representations per se. Rather, the property of being a representation with a certain content is a relational property. There are many things that can be used to represent something, but which cannot be said to represent when they are not used. One famous example is the trace of an ant in the sand which incidentally resembles a picture of Churchill (Putnam, 1981). Likewise, I can use (a representation of) the postal address system to explain somebody the IP system used to identify computers in networks. In this case, I will use the postal address system to represent the IP system (which it does, of course, not by itself). The reason is that something that is not actually playing the role of a substitute in a certain function does not fall under the definition of representation. The fact that a given thing (a certain physical object, a certain structure, or a certain mental entity) can be used to represent very different things is the reason why dispositional definitions of content will not succeed. The crucial problem here is basically the disjunction problem (discussed foremost in relation to causal theories of mental representation; e.g. in Fodor, 1987): If we define the content of a representation to be what it possibly could stand for (in the dispositional sense), then we end up with a huge (maybe even infinite) disjunction of all different possible targets for which the representation could stand (e.g. the content expressed by ‘‘the postal address system or the IP system or . . .”). Therefore, a dispositional definition of content can only be successful if there is only one way the representation is disposed to be used, which would imply that the content was independent of usage. This, however, is—as argued for—not the case. Thus, dispositional definitions of contents cannot succeed. To sum up, the behavior of a system can be described with functions (in the mathematical sense), which are describing processes.13 In these functions, representations can substitute the entities they represent. This very abstract and general description of representations expresses the fact that representations have to be in use and so reflects the basic constraint on any theory of content that hopes to capture the explanatory role of representations played in behavioral sciences. In particular, it is not sufficient to be usable (not even to be disposed to be used) as a representation in order for something to be a representation. Therefore, we can assign a certain content to a representation only if the representation is in use, i.e. involved in a process. 2.3. No theory of stored content The lesson from the discussion so far is that we do not have a theory of content and representation that would allow for ‘‘stored contents”: As soon as a representation is stored, it is not involved in processes anymore (until it is retrieved from memory); as long as it is not involved in any process, it cannot be assigned a content and thus ceases to be a representation. Therefore, as soon as a representation is stored, it is not a representation anymore. In particular, we cannot assign contents to whatever it is that gets stored. Although it is intuitively appealing to speak of stored content, there is no theory of content that could justify this talk. Quite on the contrary, the investigation of memory has to be conducted in different terms—the representational picture of the mind is, I assume, very well suited to describe occurrent mental processes and states, but it fails to give us an account of non-occurrent states (memory). Thus, the representational picture of mind is severely limited in its explanatory value. Note that a dispositional reading does not help here neither (as argued for above, it does not help to determine content in general): A stored ‘‘representation” can be ‘‘reactivated” when retrieved from memory and can be used in a process where it represents the same as it did before storage. In this sense, we could say that it is disposed to be used in the same way during storage. However, it can be used in very different ways as well. If the disposition to be used in this and that way would determine their content, then we would be faced with the classical ‘‘disjunction problem” (cf. Fodor, 1987), i.e. we would have to say that their content is the disjunction of all the different contents which we would assign in all different possible usages. To illustrate this point, a little story that I heard somewhere might help (the story is so nice that it is probably not true): The NASA used to record data on huge magnetic bands which they stored in some place. In order to read these bands, big machines were necessary. At some time when the bands were not used anymore, they decided to get rid of the reading machines. Only later they discovered that with the reading machines the stored information was literally gone for them— although the bands were still there, nobody could read them and retrieve the stored information. The question here is: Is there really any sense in saying that a certain specific content (as opposed to information, which is, of course, on the band) is stored in the bands when all reading machines are gone? I do not think so. The bands could be, for example, used by a composer with a regular music tape player to create a specific sound which figures as a part of a symphony. The band, which the composer stores in his studio, has now the (at least additional if not sole new) musical content. Therefore, the content of some storing device cannot be determined without the retrieval mechanism, which means that the content can ultimately only be determined when the representation is in use, i.e., when it is actually retrieved. Applying these results to human memory leads to the conclusion that there is no content stored in memory. The content of memory can only be determined when the use after retrieval is known. Thus, to say that specific contents are stored in memory presupposes that the use of the representation is always known. This can only be the case if the use after retrieval is always the same. However, this presupposition is far away from being warranted. Above, I already gave some example for 13 Here, nothing is assumed about the processes; in particular, it is not assumed that they are computational processes. By saying that force is a function of mass and acceleration, we give a functional description and we assume that some process (namely the acting of the force on an object) takes place; however, nothing is implied about the nature of this process. G. Vosgerau / Consciousness and Cognition 19 (2010) 838–846 843 how stored ‘‘representations” can be used in new ways (the postal address system for representing the IP system): This kind of new ways of usage could be subsumed under the term ‘‘analogies”. Because this kind of usage is not only possible but often actual, I will speak of templates that are stored in memory and that can be used to construct mental representations with content. The templates are, however, not representations themselves because they lack content (just like the business letter template on the computer is not a business letter). Rather, they are used for the construction of representations. Retrieval from memory is then the construction of a representation with the help of a stored template. 3. Templates in memory In long-term memory, templates are stored. They cannot be assigned a specific content, because they are not in use and because they can be used to construct different representations. Of course, templates cannot be used to construct any kind of representation: For each template, there are constraints on what representations can be constructed on its basis. These constraints are best understood in terms of certain relations between the templates and the possible represented things. For example, the ant’s trace in the sand (see above) can be used to represent Churchill because of a certain similarity between the two. Likewise, the structure of the postal address system stands in a certain similarity relation to the IP system, which is the reason that it can be used to represent it. I have argued that this relation is (at least for conceptual representations) best understood as a partial isomorphism (Vosgerau, 2006, 2008, 2009). Partial isomorphism is thereby defined as an isomorphism that holds between the structure of the representation and a part of the structure of the represented object. Which part of the represented object is relevant for this relation is determined by the function in which the representation plays its functional role. The crucial point here is that partial isomorphism does not suffice to determine content or to make something a representation since there is always a multitude of isomorphic objects (and structures). As argued for above, we need the functionalistic core—or: the use of the representation—for this.14 Templates are thus not just representing abstract entities (like structures) or fuzzy things. They do not represent at all, although their structure might be well-defined. However, the structure alone does not make them representations, although the structure highly constrains possible usages as representations. Therefore, they stand in contrast to the notion of schemata, which is widely used in memory research (cf. Alba & Hasher, 1983). Schemata, as scripts (Schank & Abelson, 1977) or frames (Barsalou, 1992; Minsky, 1975), are conceived of as representing abstract entities, e.g. not a specific birthday party but the general procedure of a birthday party or not a specific car but the general (possible) properties of a car. In this sense, schemata are taken to have a defined content (which can be true or false), although they are not supposed to represent single events or objects. Templates, in contrast, do not represent anything—not even abstract entities or generalized objects. Another term that is often used for memory is the term ‘‘trace” (e.g. Bernecker, 2008, 2010; Matthen, 2010). A memory trace is said to exist between a mental state (e.g. a perception) and the mental state we are in when we remember the first one. It connects the two mental states in a causal manner, and so it is able to build the basis for justification. However, a trace itself is not a mental state but a sequence of (mental?) states in which one state is caused by its predecessor (cf. Bernecker, 2008, Fig. 3.1 on p. 43). Therefore, the notion of a memory trace can be fruitfully applied to problems discussed in epistemology (where the trace can build the basis for justification for the remembered state), but it is accompanied with assumptions regarding causal connections that are not relevant for the present argument. The notion of a template is, in contrast, introduced as a more general notion which does not rely on specific assumptions about causation or sequences. In fact, traces can be viewed as one possible form of templates (if they are interpreted as bearing no content, as done, e.g. by Matthen, 2010). However, there are many other possible forms or ‘‘realizations” of templates, e.g. weights in neural networks, symbolic strings, mental models, . . . Since templates can be used to construct very different representations, memory recall has to be understood as a constructive processes that creates content rather than a conservative processes that re-activates content. The fact that at least large parts of memory recall are indeed constructive is well documented in the empirical literature (see, e.g. Schacter, Norman, & Koutstaal, 1998; Schank & Abelson, 1977). The claim of this paper goes beyond the empirical claim in that it claims that all kinds of memory retrieval are constructive. It is to be understood in the sense that in principle, every stored template can be the basis of different representation constructions. Thus, the point is clearly a philosophical point: Even if a specific template could be empirically shown to yield always the same kind of representation (the same content), the thesis could still be true, since a de facto existing rigid connection between a specific template and a specific kind of content does not contradict the in principle possibility of connections to different contents. The crucial point here is that the stored template alone can never explain behavior; only when retrieval mechanisms are also taken into account, such that the resulting occurrent representation is in focus, explanation of behavior can be successful (see above). Nevertheless, the argument has implications not only for philosophy but also for theorizing about memory and thus for empirical research on memory. In the next section, some of the most important implications will be sketched. 14 Bartels and May (2009) argue that ‘‘structural resemblance” (which is spelled out in terms of homomorphism; Bartels, 2005) is an integral part of functionalism. In Vosgerau (in press, chap. 8) I argue that the functionalistic story and the structural resemblance story are two different stories that in principle come apart. 844 G. Vosgerau / Consciousness and Cognition 19 (2010) 838–846 4. Implications Ramsey (2007) presents a detailed discussion of different understandings of the term ‘‘representation”. His powerful (and in my view successful) arguments against the coherence of the notion of ‘‘implicit” or ‘‘tacit” representations shows parallels to the arguments presented here. However, the notion of tacit representation is ‘‘characterize[d] as based on the idea that there is a close link between the dispositional properties of a cognitive system, and a type of knowledge or information that the system represents ‘inexplicitly’ or ‘tacitly.’” (Ramsey, 2007, p. 151). Thus, tacit or implicit contents are conceived of as being present in the system (as dispositions) without there being a specific vehicle to which the specific content could be assigned. In this sense, they are stored (unless the dispositions manifest themselves)15 and therefore build a special subclass of what I called non-occurrent representations. Therefore, one implication of this paper is the claim of Ramsey: There is no such thing as implicit or tacit representations, since no content can be assigned to such non-occurrent ‘‘representations”. Another central implication is that stored templates are neither true nor false. Since stored templates are not representations and have no contents, they cannot be assigned truth-values either. A memory can thus not be correct by itself, but only when it is remembered. This means that only ‘‘occurrent” mental states have truth-values, and that true memories have to be constructed as occurrent representations of past events. There is no doubt that memory and stored templates play a crucial role for the explanation of cognition (not only human cognition) because memories shape the way we represent the world. Depending of what we have learned, we will represent the world in different ways. This fact can be easily explained with the idea of stored templates: Since templates constrain the structure of the constructed representations, occurrent representations will differ with the templates that are used for their construction. If, e.g. I have never learned how motorbikes work, my representation of a specific motorbike will not contain very specific information about the different parts of it. If, however, I am an expert in motorbikes, my representation of the same motorbike will have a much richer structure that allows me to identify different functional units and so on. The reason for this difference lies in the differently organized templates (in memory) that are available to me to construct the specific representation. However, the consequence of this discussion is that this very important part of cognition—memory—cannot be described in ‘‘classical” truth-apt terms. Such notions as ‘‘proposition” or ‘‘misrepresentation” cannot be applied by a theory of memory. Therefore, there can be no causal connection between the content of a stored representation and the content of an occurrent (remembered) memory because the first just does not exist. Bernecker (2008, 2010) argues for memory traces that explain the causal relation between past mental states (that are stored) and present states of remembering the past event that was represented by the past mental state. He distinguishes between a physical and a ‘‘mental” description, the first referring to some neural states and processes and the second referring to ‘‘content causation”. While there is, from the point of view of this paper, nothing wrong with the idea that occurrent memories are physically caused (at least in part) by some physical states or processes that are causally linked to former states (they are linked so by templates in my terminology), the idea that the content of the occurrent state is causally connected to the content of the past state, however, cannot be successful. Indeed, it is not at all clear why we need such a form of ‘‘content causation” at all (i.e. I see no explanatory advantage of assuming such a causal connection): Contentful states can be caused by non-contentful states (e.g. the tree in front of me and my sense apparatus are not contentful but cause my percept which is contentful), and there is no reason to assume that memory retrieval should work another way. Note that I do not argue against ‘‘mental causation”: Of course, contents can have causal powers since they explain behavior (this was the starting point in the paper). And perhaps it is even useful to think that contents can cause other contents by triggering some mental processes that produce other contentful states, e.g. in inferences. In this way, the past occurrent content could be said to indirectly cause the present (remembered) content through a processes involving non-contentful states (e.g. traces). Nevertheless, it does not follow that templates or traces have content. More generally, the so-called ‘‘belief system” of a cognitive agent, which is traditionally assumed to be a fairly stable system of contents which are believed by the agent, does not exist according to the position defended here. Rather, the stored items are templates without contents. The often adopted view (e.g. Bernecker, 2010) that stored items have a dispositional content does not help here as argued for above.16 Taken together, in the epistemological literature on memory, the term ‘‘content” is used to apply to non-occurrent states as well as occurrent states. However, such a notion of content is not compatible with the explanatory role this notion plays in behavioral sciences. Thus, the epistemic status of memory cannot be captured within a representational framework, especially not in a framework that postulates truth-apt stored representations.17 This point is also related to the discussion about ‘‘knowing how” and ‘‘knowing that” going back to Ryle (1945): Knowinghow is often conceived of as a kind of implicit knowledge not being analyzable in terms of truth-apt (propositional) representations (e.g. Jung & Newen, 2010). However, knowing-how is certainly part of our memory since we have learned how to 15 It could be discussed whether the manifestation of such dispositions always involves some kind of representation with content or not—I would suspect that in many cases the contents of the occurrent representations that build the basis of the manifestation have to be assigned contents that are not reducible to the disposition. However, due to space limitations this has to remain a suspicion at the moment. 16 In fact, Bernecker (2010) makes a distinction between conceptual and non-conceptual contents, the first being explained in terms of dispositional beliefs and the second being explained in terms of ‘‘subdoxastic” or implicit states. The notion of implicit representation is, however, also not compatible with the view presented here as shown above. 17 This does, of course, not entail that everything is wrong about such epistemological accounts—however, since this paper is not concerned with the epistemology of memory, I will not discuss these accounts further. G. Vosgerau / Consciousness and Cognition 19 (2010) 838–846 845 do things. The present discussion yielded that a representational analysis of stored ‘‘knowledge” is not possible at all. Thus, even the so-called ‘‘knowing that” kind of knowledge cannot be described in terms of content. In this sense, it might turn out that the distinction between the two kinds of knowledge vanishes, although in quite another sense than often assumed: Knowing-how is not reducible to knowing-that (e.g. Stanley & Williamson, 2001) but knowing-that is nothing but a kind of knowing-how in the sense that it is a certain ability to produce something (a representation in the one case and a certain kind of behavior in the other case). This ability, however, cannot be described or explained in terms of content. Therefore, according to the argument of this paper large parts of the classical picture of the ‘‘representational mind” have to be changed: contents are not the kind of things that can be stored, and—although memory is central to cognition—memory cannot be described in terms of representations. This does not necessarily mean that theories of memory have to be re-written from the very start. Instead, the aim of this paper is to point to one important factor in explanations involving memory that is often completely left out: In fact, the arguments given in this paper try to show that arguments of the following form are no good explanations: the cognitive system shows this and that behavior because it has stored this and that content. They are no good explanations because they at least leave out a very important part, namely the reference to the constructive retrieval mechanisms. In this sense, the (full) explanation should read: the cognitive system shows this and that behavior because it constructed this and that content out of this and that template with the help of this and that retrieval mechanism. Probably, some explanations of the above form that are actually given in behavioral sciences can be reinterpreted as ‘‘abbreviations” of full explanations. However, this will not be the case for all such explanations. To illustrate this point, consider one of the central paradigms of empirical research in memory which is focused on false recall and false recognition (cf. Roediger, McDermott, & Robinson, 1998). Thereby, memory recall is shown to be often influenced by associative processes. This effect perfectly fits the picture presented here: Since templates only serve to construct representations, whereby the content of the constructed representation is not fully determined but only constrained by the template, so-called ‘‘false” memories are even likely to occur. This line of research explicitly makes reference to the retrieval mechanisms and shows that they cannot be assumed to be stable (i.e. they cannot be assume to involve the same associations every time they are employed). However, the focus on false memory is not justified if we want to learn about the structure of the stored templates, since they cannot be true or false. If an occurrent representation that is constructed on the basis of a stored template turns out to be false, this does not mean that something wrong has been stored (it could equally well be that a different retrieval mechanism would have generated a true representations). Thus, if investigations should be informative about the structure of the stored templates, the strong focus on ‘‘false” memories has to be abandoned: Both true and false occurrent representations based on memory are constructed with the same mechanisms (this means that ‘‘true” memories are no less based on associations). Therefore, the content and structure of both true and false occurrent memories equally contribute to an understanding of the underlying templates, since they constrain the results of the constructive retrieval mechanisms in the same way in every case. However, these constraints can only be studied if we focus on the whole range of contents that are constructed from memory, and not only on false contents. 5. Conclusion I have argued that each theory of representation has to have a ‘‘functionalistic” core, i.e. it has to define representations as substitutes for the represented objects in functions which describe the behavior to be explained. Only such theories of representation can do justice to the use of the notion of representation as it is used in the cognitive sciences. This implies that the content of a representation can only be determined if the representation is in use. The consequence of this view is that we do not have an adequate theory of stored content. Quite on the contrary, contents cannot be stored since they cease to be contents as soon as they are no longer in use. Because they can be used to construct different representations when retrieved, a dispositional theory of stored content cannot succeed either (it would run into the disjunction problem). I have proposed to use the term ‘‘template” to describe the entities that are stored in long-term memory. They are not representations and do not have contents, but they can be used to construct representations (including representations about past events = memories). Some implications from this conclusion are shortly discussed. Thereby, I hope to have shown that the usual way of talking about memory as storing content is central to philosophical issues about memory and to empirical methods of investigating memory. However, this usual way is fundamentally mistaken such that we need some detailed account of memory that can deal with the absence of content for stored templates. This paper is certainly not able to present such an account. It rather tries to set up the stage for the problem itself and for a possible direction for a solution. Acknowledgments This paper is based on a presentation given at the international conference ‘‘Memory and Self-Understanding” (Delmenhorst, Germany, June 2009); I would like to thank the organizers and participants for this stimulating event and fruitful discussions. Moreover, I would like to thank Christoph Michel and three anonymous reviewers for very helpful remarks which improved earlier versions of the paper dramatically. References Alba, J. W., & Hasher, L. (1983). Is memory schematic? Psychological Bulletin, 93, 203–231. 846 G. Vosgerau / Consciousness and Cognition 19 (2010) 838–846 Barsalou, L. W. (1992). Frames, concepts and conceptual fields. In Adrienne Lehrer & Eva Feder Kittay (Eds.), Frames, fields and contrasts (pp. 21–74). Hillsdale, Hove and London: Erlbaum. Bartels, A. (2005). Strukturale Repräsentation, Mentis, Paderborn. Bartels, A., & May, M. (2009). Functional role theories of representation and content explanation: With a case study from spatial cognition. Cognitive Processing, 10, 63–75. Bernecker, S. (2008). The metaphysics of memory. New York: Springer. Bernecker, S. (2010). Memory: A philosophical study. Oxford: Oxford University Press. Block, N. (1978). Troubles with functionalism. In C. W. Savage (Ed.), Perception and cognition: Issues in the foundation of psychology. Minneapolis: University of Minnesota Press. Burge, T. (1993). Content preservation. The Philosophical Review, 102, 457–488. Cheng, K. (1986). A purely geometric module in the rat’s spatial representation. Cognition, 23, 149–178. Cummins, R. (1996). Representations, targets, and attitudes. Cambridge, MA, London: The MIT Press. Dretske, F. (1986). Misrepresentation. In Radu J. Bogdan (Ed.), Belief, form, content, and function (pp. 17–36). Oxford: Clarendon Press. Fodor, J. (1987). Psychosemantics. Cambridge, MA, London: The MIT Press. Frege, G. (1892). Über Sinn und Bedeutung. Zeitschrift für Philosophie und philosophische Kritik, 100, 20–50. Gallistel, C. (1993). The organization of learning. Cambridge, MA: The MIT Press. Jung, E.-M., & Newen, A. (2010). Knowledge and abilities: The need for a new understanding of knowing-how. Phenomenology and the Cognitive Sciences, 9, 113–131. Lettvin, J. Y., Maturana, H. R., McCulloch, W. S., & Pitts, W. H. (1959). What the frog’s eye tells the frog’s brain. Proceedings of the IRE, 47(11), 1940–1951. Matthen, M. (2010). Is memory preservation? Philosophical Studies, 148, 3–14. Millikan, R. G. (1984). Language, thought, and other biological categories. Cambridge, MA, London: The MIT Press. Millikan, R. (1994). Biosemantics. In Stephen Stich (Ed.), Mental representation: A reader (pp. 243–258). Cambridge, MA, Oxford: Blackwell. Minsky, M. (1975). A framework for representing knowledge. In P. H. Winston (Ed.), The psychology of computer vision. New York: McGraw-Hill. Putnam, H. (1975), The Meaning of ‘‘Meaning”, in ‘Mind, Language, and Reality’ (pp. 215–271). Cambridge: Cambridge University Press. Putnam, H. (1981). Reason, truth, and history. Cambridge: Cambridge University Press. Ramsey, W. M. (2007). Representation reconsidered. Cambridge: Cambridge University Press. Roediger, H. L., McDermott, K. B., & Robinson, K. J. (1998). The role of associative processes in creating false memories. In M. A. Conway, S. E. Gathercole, & C. Cornoldi (Eds.), Theories of memory II (pp. 187–245). Hove, UK: Psychological Press. Ryle, G. (1945). Knowing how and knowing that. Proceedings of the Aristotelian Society, 46, 1–16. Schacter, D. L., Norman, K. A., & Koutstaal, W. (1998). The cognitive neuroscience of constructive memory. Annual Review of Psychology, 49(1), 289–318. Schank, R. C., & Abelson, R. P. (1977). Scripts, plans, goals and understanding: An inquiry into human knowledge structures. Oxford, England: Lawrence Erlbaum. Stanley, J., & Williamson, T. (2001). Knowing how. The Journal of Philosophy, 98, 411–444. Vosgerau, G. (2006). The perceptual nature of mental models. In C. Held, M. Knauff, & & G. Vosgerau (Eds.), Mental models and the mind: Current developments in cognitive psychology, neuroscience, and philosophy of mind (pp. 255–275). Amsterdam: Elsevier. Vosgerau, G. (2008). Adäquatheit und Arten mentaler Repräsentationen. Facta Philosophica, 10, 67–82. Vosgerau, G. (2009). Mental representation and self-consciousness. From basic self-representation to self-related cognition, Mentis, Paderborn. Vosgerau, G. (in press). Varieties of representation. In A. Newen, A. Bartels, & E. Jung, (Eds.), Knowledge and representation. Mentis, Stanford, Paderborn: CSLI Publications. Vosgerau, G., Schlicht, T., & Newen, A. (2008). Orthogonality of phenomenality and content. American Philosophical Quarterly, 45, 309–328.

Consciousness and Cognition 43 (2016) 133–142 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog The relation between the sense of agency and the experience of flow Matti Vuorre, Janet Metcalfe ⇑ Columbia University, United States a r t i c l e i n f o Article history: Received 30 January 2016 Revised 1 June 2016 Accepted 2 June 2016 Keywords: Sense of agency Flow a b s t r a c t This article investigates the relation between people’s feelings of agency and their feelings of flow. In the dominant model describing how people are able to assess their own agency— the comparator model of agency—when the person’s intentions match perfectly to what happens, the discrepancy between intention and outcome is zero, and the person is thought to interpret this lack of discrepancy as being in control. The lack of perceived push back from the external world seems remarkably similar to the state that has been described as a state of flow. However, when we used a computer game paradigm to investigate the relation between people’s feelings of agency and their feelings of flow, we found a dissociation between these two states. Although these two states may, in some ways, seem to be similar, our data indicate that they are governed by different principles and phenomenology. Ó 2016 Published by Elsevier Inc. 1. Introduction Gallagher (2007, p. 348) defined the sense of agency as ‘‘the pre-reflective experience or sense that I am the cause or author of the movement (e.g., an experience that I am in control of my action)”. For example, if I reach to pick up a glass, I may actually have a sense of control over the movement and so have a sense of agency for this movement; if I am then asked, did I reach for the glass, I can correctly attribute agency to myself: ‘Yes, I was the one who reached for the glass.’ Cognitive models have made considerable progress in pointing to the informational basis for such feelings of control. The most widely accepted computational model addressing this question is the comparator model (Blakemore, Frith, & Wolpert, 1999; Wolpert, Ghahramani, & Jordan, 1995). This model was initially proposed to solve the problem of how people are able to achieve fast, finely-tuned and flexible online control of movements. Wolpert and colleagues (Wolpert & Flanagan, 2001; Wolpert et al., 1995) proposed that the actual sensory feedback of movement is compared to its predicted feedback. When a discrepancy between the predicted and actual feedback occurs, it can be used to immediately correct the movement. But the comparison can also be co-opted for other purposes: Blakemore, Frith and others (Blakemore & Frith, 2003; Blakemore, Wolpert, & Frith, 1998; Blakemore et al., 1999) proposed that this comparison information could serve as the basis for people’s sense of agency. When a large discrepancy at the locus of convergence for the actual and predicted feedback (the comparator) is detected, it means that there was a large difference between the person’s intentions and what happened and therefore that the person was not in control. In contrast, when there is little to no difference between the two signals, it ⇑ Corresponding author at: Columbia University, Department of Psychology, 406 Schermerhorn Hall, 1190 Amsterdam Avenue, New York, NY 10027, United States. E-mail address: jm348@columbia.edu (J. Metcalfe). http://dx.doi.org/10.1016/j.concog.2016.06.001 1053-8100/Ó 2016 Published by Elsevier Inc. 134 M. Vuorre, J. Metcalfe / Consciousness and Cognition 43 (2016) 133–142 means that the person’s intentions are being smoothly actualized, and the person is in control. The smoothness that sometimes accompanies and is a cue for the positive sense of agency—where the person’s intentions play out without apparent opposition from the outside—is highly reminiscent of the reports that people give when they are experiencing a state of effortless flow. A common model describes feelings of flow as a function of skills and task demands. When the task demands match the actor’s skills, people are prone to experience flow. When demands exceed skills, people become anxious. They become bored when they are too skilled for the current demands (the balance hypothesis, Csikszentmihalyi, 2009; Rheinberg & Vollmeyer, 2003). Recently, this model was augmented to include people’s judgments of performance. Flow judgments are highest when skills and demands are in balance and when they think they are performing well (the balance plus hypothesis; Kennedy, Miele, & Metcalfe, 2014). The experience of flow is reported across a range of activities, including sports, music, mountaineering, painting, and gaming, and is described as a pleasurable and motivating immersion in the current activity (Csikszentmihalyi, 2009; Massimini & Carli, 1988). The flow state is similar to the heightened state of consciousness that is sometimes, especially in sports, described as a feeling of being in the zone, or having a ‘‘hot hand” in basketball (Gilovich, Vallone, & Tversky, 1985; Young & Pain, 1999). In the literature, the terms zone and flow are often used interchangeably (Young & Pain, 1999), and we do so here as well. Crucially, when the person is ‘in the zone’ or in a state of flow, there is a feeling of effortlessness, and lack of resistance from the world. For instance, Roger Bannister, in describing his four-minute mile breaking run says: ‘‘Brasher went into the lead and I slipped in effortlessly behind him. My legs seemed to meet no resistance, as if propelled by some unknown force” (Bannister, 2014). Similarly, Hales (1999, p. 79) reported that when five time Wimbledon champion, Björn Borg, was at the height of his powers, he described the feeling that he ‘‘could do anything—put the ball on a dime at any angle, anywhere on the court, at any speed he chose, with the spin he wanted.” Note that Borg’s description emphasizes precise control over his actions and their outcomes; this sense of control is considered an important characteristic of the flow state (Nakamura & Csikszentmihalyi, 2002). There have been no studies in which the relation between agency and flow has been explicitly tested. However, Wenke, Fleming, and Haggard (2010, and see Chambon & Haggard, 2012; Chambon, Sidarus, & Haggard, 2014; Sidarus, Chambon, & Haggard, 2013; Stenner et al., 2014) investigated the effect of subliminal priming of actions on people’s feelings of agency. They found that the increases in action selection fluency due to such priming resulted in increases in people’s sense of control over the effects of their action. Wenke et al. (2010, p. 36) noted: ‘Our key finding is that this smoothness produces a heightened sense of control. This is in keeping with the notion that during a well-learned, skilled task such as playing the piano, people often report mastery of what they are doing, and a feeling of ‘‘flow” (Csikszentmihalyi, 2000).’ Although the previous considerations suggest similarities between flow and agency, there also may be important differences between these two states. For instance, the relationship between experienced control and the flow state is sometimes called paradoxical (Young & Pain, 1999). Flow requires control over actions, but is thought to also involve a loss of awareness of oneself as a (social) actor (Nakamura & Csikszentmihalyi, 2002, p. 90), suggesting that these two experiences might not be the same. The hypothesis that motivated the present study was that there would be convergence between people’s sense of agency and feelings of flow. We began with the findings of Kennedy et al. (2014) who demonstrated that the feeling of flow can be manipulated by having participants play a computer game in which X’s and O’s scroll from the top of the screen, and the participants’ task is to move a mouse controlling a cursor in such a way as to catch the X’s while avoiding the O’s. They varied the speed of the scroll, and found a non-monotonic relation between people’s feelings of flow and speed. People tended to experience the highest levels of flow at moderate speeds—speeds at which they were maximizing the number of X’s that they hit over the 20 s game interval. Less flow was experienced when the speed was either too high or too low. People also experienced an increase in flow when they had the metacognitive feeling that their performance was particularly good on that trial (regardless of whether their performance actually had been particularly good). Kennedy et al. (2014), did not look at agency judgments. However, the computer game that they used has been used in other studies that have investigated people’s feelings of agency (see, Metcalfe, Eich, & Castel, 2010; Metcalfe, Eich, & Miele, 2013; Metcalfe & Greene, 2007; Metcalfe, Snellenberg, DeRosse, Balsam, & Malhotra, 2012; Miele, Wager, Mitchell, & Metcalfe, 2011). In the first experiment that we present here, we simply took exactly the same game that Kennedy et al. (2014) had used, but instead of asking people for judgments of flow (what Kennedy et al. called ‘Z’ scores), on each trial, we asked them for judgments of agency (how in control they felt). Our expectation was that we would obtain agency functions that were virtually identical to those reported for flow by Kennedy et al. (2014). 2. Experiment 1 2.1. Method 2.1.1. Participants The sample of 14 participants (3 males and 11 females, Mage = 20 years) were students in an Introductory Psychology course at Columbia University and received partial course credit for their participation. All participants were treated in accordance with APA regulations, and the ethical guidelines of the Psychonomic Society. M. Vuorre, J. Metcalfe / Consciousness and Cognition 43 (2016) 133–142 135 2.1.2. Design Each 20-s trial of the computer game task involved using a mouse to move a cursor along a horizontal track at the bottom of the screen as stimuli (Xs and Os) that were randomly distributed from left to right scrolled down from the top of the screen to the bottom (the total number of stimuli depended on the speed of the game). The Xs or Os disappeared as soon as they were ‘‘touched” by the cursor, but continued to scroll past the horizontal track if they were not touched. In addition, a distinctive BONK sound occurred each time an X was hit and a THUD sound occurred each time an O was hit. Seven different experimental conditions were created by manipulating the speed with which the Xs and Os scrolled down the screen. The entire session included 21 trials grouped into 3 blocks, such that each block contained trials from all 7 speed conditions. The order of conditions within each block was pseudorandom. 2.1.3. Procedure After completing a demographic questionnaire (which included questions about their experience playing video games), participants were told that they would be playing a game in which the purpose was to use the mouse to touch the Xs when they came into range of the cursor box and to avoid touching the Os. Participants then read a brief passage about feelings of agency: ‘‘In this experiment, we are interested in people’s metacognitions of control or their feelings about when they are causing things to happen. You have probably heard stories, and maybe even have had the experience of going to an arcade and you start playing a video game—thinking you are controlling what is happening. But when you let go of the levers it just keeps on happening: you weren’t actually controlling anything, but you thought you were. Metacognition of control can differ in more mundane circumstances too. Sometimes if you are driving, the steering may be such that you don’t feel like you are in control. At other times you simply know that you are controlling every move. Regardless, then, of whether you actually are in control or not (which is not our question here) you may sometimes feel like you are in control (and hence have a high metacognition of control) or feel like you are not (and have a low metacognition of control).” Note that this judgment was substituted into the program used by Kennedy et al. (2014), to replace the judgments of flow or of ‘being in the zone’ that had been requested following each game, in their experiment. Otherwise, this experiment was identical to their Experiment 1. Following a practice trial, participants completed the 3 blocks of experimental trials. After each trial, they were asked to make both agency judgments, indicating how in control they felt, and judgments of performance (JOPs). For agency judgments, participants were presented with a visual analogue scale labeled ‘‘Amount of CONTROL” and were asked to pull the slider toward the left end if the amount of control they experienced was ‘‘None” or to the right end if the amount they experienced was ‘‘Complete.” For performance, they were presented with a scale labeled ‘‘Performance” and were instructed to pull the slider to the left end if they felt that they had gotten ‘‘None correct” or to the right end if they felt they were ‘‘Completely correct.” In between each block of trials, participants completed a 20-s distractor task that involved subtracting by 3’s from a randomly presented three-digit number. 2.2. Results 2.2.1. Statistics We specified a multilevel linear regression model (e.g. Gelman & Hill, 2007) to estimate the main effects of speed (linear and quadratic) and JOPs (standardized within participants) on participant’s agency judgments. All effects were estimated at the average level (fixed effects) and as varying at the participant level (random effects) to appropriately account for the clustering of data and possible between-participants heterogeneity in effects. We used orthogonal polynomials to estimate the linear and quadratic effects of speed on the agency judgments (the linear predictor of speed was scaled to range from
0.57 to 57). We adopted a Bayesian approach to estimating the model parameters, because it is especially suitable for multilevel models (Gelman & Hill, 2007), and approximated the posterior distribution by Hamiltonian Monte Carlo sampling as implemented in the Stan programming language (Stan Development Team, 2015.) We ran four HMC chains of 30,000 iterations, discarding the first 5000 values of each chain, leading to a total of 100,000 samples from the posterior distribution. The samples from the posterior distribution were analyzed with the R programming language (R Core Team, 2015). We used a uniform prior for the intercept parameter, Normal(M = 0, SD = 50) distributions as priors for the average level regression parameters, and half-Cauchy(location = 0, scale = 50) priors for the hierarchical variance components (random effects’ standard deviations) Because of the scale of the data (judgment range: 0–100), these priors were minimally informative, but we also estimated the model with a variety of priors (including uniform) to confirm the results. We also estimated the model using standard maximum likelihood methods (Bates, Mächler, Bolker, & Walker, 2015), and found the same results. For ^ and their associated 95% Credible Intervals parameters estimated with the Bayesian model, we report posterior means ðbÞ (CI; the central 95% of values in the respective posterior distribution.) 2.2.2. Results As is shown in Fig. 1, people’s judgments of agency were a monotonically decreasing function of speed (linear effect of ^ =
17.14, 95% CI [
29.54,
4.51]; quadratic effect of speed b ^ =
2.34, 95% CI [
7.18, 2.46]). Additionally, higher speed b ^ = 10.46, 95% CI [4.68, 16.39]). judgments of agency were associated with higher judgments of performance (b 136 M. Vuorre, J. Metcalfe / Consciousness and Cognition 43 (2016) 133–142 (a) (b) 100 ● Speed linear ● ● ● ● 75 ● ● Rating ● ● ● ● 50 ● ● ● Speed quadratic 25 JOP 0 1 2 3 4 5 6 7 −40 −20 Speed level 0 20 40 Effect estimate Fig. 1. Bayesian multilevel linear regression model of Agency judgments from Experiment 1. (a) Agency ratings as a function of game speed and Judgments of Performance (blue = high JOP; red = low JOP). Circles are means from raw data, regression lines at ±1 SD of JOP are displayed with 95% CIs as gray shades. (b) Histograms of 100,000 samples from the posterior distribution of the regression parameters with 95% CIs (gray shades). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) (a) (b) 100 Speed linear 75 Rating ● ● ● ● ● 50 ● ● ● Speed ● ● ● ● 25 ● JOP ● 0 1 2 3 4 Speed level 5 6 7 −40 −20 0 20 40 Effect estimate Fig. 2. Bayesian multilevel linear regression model of Zone judgments from Kennedy et al. (2014). (a) Zone judgments as a function of game speed and Judgments of Performance (blue = high JOP; red = low JOP). Circles are means from raw data, regression lines at ±1 SD of JOP are displayed with 95% CIs as gray shades. (b) Histograms of 100,000 samples from the posterior distribution of the regression parameters with 95% CIs (gray shades). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) For comparison, we reanalyzed the data from Kennedy et al.’s (2014) Experiment 1 in the same manner. As is shown in Fig. 2, the zone judgments in Kennedy et al. (2014) followed the inverted u-shaped function as was reported in the original ^ = 33.36, 95% CI [24.3, 42.39]; quadratic effect of speed b ^ =
23.44, 95% CI [
29.35,
17.51]; paper (linear effect of speed b ^ = 10.74, 95% CI [5.59, 16.07]). While both types of judgments (agency and flow) were positively associated with judgJOP b ments of performance, the basic shape of the agency function (Fig. 1a) was quite different from that of the flow function (Fig. 2a): The former was a decreasing linear function with no quadratic component, and the latter was quadratic with a linear increase. 2.3. Discussion The first experiment suggested that our initial conjecture—that the experiences of flow and agency were isomorphic—was wrong. However, different participants were run in our experiment and in that of Kennedy et al. (2014). Furthermore, the number of participants in Experiment 1 was small, and we wanted to be sure that those results replicated. To compare the phenomenology of flow and agency within individuals, then, we conducted Experiment 2 in which the same participants gave both types of judgments in a within-participant design. M. Vuorre, J. Metcalfe / Consciousness and Cognition 43 (2016) 133–142 137 3. Experiment 2 3.1. Method 3.1.1. Participants We recruited 25 participants (18 females and 7 males, Mage = 24 years) from an Introductory Psychology course at Columbia University who received partial course credit for their participation. All participants were treated in accordance with APA regulations, and the ethical guidelines of the Psychonomic Society. 3.1.2. Procedure and design All participants did both the agency condition, with 21 trials, including 3 blocks with each of the 7 speeds, each followed by judgments of agency and then judgments of performance, and the flow condition, in which, following each trial they gave their ‘Z’ scores, followed by judgments of performance. We used the same instructions to illustrate what flow (the feeling of being in the zone) was, that had been used by Kennedy et al. (2014). In particular, before beginning the flow trials, participants read a brief passage about zone, or ‘‘Z-factor”, that included descriptions of flow (Csikszentmihalyi, 2000) and the hot hand (Gilovich et al., 1985), as well as a quote from the soccer player Pelé (Pelé & Fish, 2013), and other instructions used by Kennedy et al. (2014): ‘‘In this part of the experiment, we are interested in your experiences of being fully immersed in a feeling of energized focus, of full involvement, and of enjoyment which is characterized by a complete absorption in what you were just doing. In sports this may be known as ‘being on fire’ or in music as ‘being in the groove.’ It is a specific feeling in which everything seems to flow in such a way that it feels like everything is going your way. The legendary soccer player Pelé described his experience as feeling ‘. . .a strange calmness. . . a kind of euphoria. I felt I could run all day without tiring, that I could dribble through any of their team or all of them, that I could almost pass through them physically.’ For the present purposes we will call this feeling, or this state of being ‘in the zone,’ the Z-factor. We will ask you to rate how much of this feeling you had—from a very low z-score to a very high z-score.” After each trial in this block, participants reported their Z-score on a 10 cm visual analogue scale ranging from ‘‘Very Low” to ‘‘Very High.” Half of the participants did the 21 agency trials first followed by the 21 Z-score trials; the other half did the Z-score trials first, followed by the agency trials. 3.2. Results We analyzed the data in four ways: (1) a Bayesian analysis as described for Experiment 1, (2) a factorial ANOVA in which type of judgment was treated as if it were an independent variable, and (3) using two separate multilevel regression models similar to the Bayesian analysis, but additionally using the participants’ average agency judgments to predict flow judgments, and vice versa. Finally, (4) we examined the relationships between the two types of judgments and two performance metrics. 3.2.1. Bayesian regression model We specified a multilevel linear regression model identical to the one described for Experiment 1, but with an additional dummy predictor for judgment type (0 = agency judgment, 1 = flow judgment), and interaction terms between the judgment type dummy predictor, and the other predictors. These results replicated the findings of Experiment 1 and those of Kennedy et al. (2014). Participants’ agency judgments were a monotonically decreasing function of game speed, but there was an ^ =
15.11, 95% CI [
23.31,
6.80]; quadratic effect of speed additional small quadratic component (linear effect of speed b ^ =
3.73, 95% CI [
6.68,
0.78]). Participants’ flow judgments displayed a strong linear increase and quadratic component b ^ = 15.75, 95% CI [4.97, 26.34]; quadratic effect of speed b ^ =
13.75, 95% CI [
21.54,
5.99]). Impor(linear effect of speed b tantly, the interaction terms indicated that the differences in the linear and quadratic effects of game speed between agency ^ = 30.87, 95% CI [17.84, and flow judgments were large and plausibly non-zero (linear effect by judgment type interaction b ^ =
10.02, 95% CI [
17.99,
2.04]). Additionally, JOPs were positively 43.66]; quadratic effect by judgment type interaction b ^ = 8.65, 95% CI [5.64, 11.83]) and flow (b ^ = 9.89, 95% CI [6.62, 13.26]) judgments; judgment associated with higher agency (b ^ type and JOP did not interact (b = 1.24, 95% CI [
1.41, 3.93]). These results are displayed in Fig. 3. 3.2.2. ANOVA We also conducted a 2 (Judgment type: Agency, Zone)  7 (Speed level) ANOVA (Greenhouse-Geisser corrected for violations of sphericity) confirming that speed exerted a significant effect on judgments (F(2.1, 50.9) = 11.7, p < 0.001, g2p = 0.33), and that the effect of judgment type was significant (F(1, 24) = 15.43, p < 0.001, g2p = 0.39). Importantly, the speed by judgment type interaction was also significant (F(2.2, 52.4) = 15.5, p < 0.001, g2p = 0.39). Bayes factors for all three effects indicated decisive evidence for the effects over their respective null models (models including all lower-level terms; all BF10s > 100; Love et al., 2015; Rouder, Morey, Speckman, & Province, 2012). These results are displayed in Fig. 4; while agency ratings 138 M. Vuorre, J. Metcalfe / Consciousness and Cognition 43 (2016) 133–142 (b) (a) Agency Zone Agency speed 100 ● ● 75 Rating ● ● Agency speed quadratic ● ● ● ● ● ● ● ● ● ● 50 ● ● ● ● ● ● ● ● ● ● ● ● Agency JOP Zone speed ● ● Zone speed quadratic 25 Zone JOP 0 1 2 3 4 5 6 7 1 2 3 4 5 6 7 −30−20−10 0 10 20 30 Speed level Effect estimate Fig. 3. Bayesian multilevel linear regression model of Agency and Zone judgments from Experiment 2. (a) Judgments as a function of game speed and JOP (blue = high JOP; red = low JOP). Circles are means from raw data, regression lines at ±1 SD of JOP are displayed with 95% CIs as gray shades. (b) Histograms of 100,000 samples from the posterior distribution of the regression effects with 95% CIs (gray shades). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) 100 ● 80 ● ● ● Rating ● ● 60 ● ● ● ● ● ● ● 40 ● 20 0 1 2 3 4 5 6 7 Speed level Fig. 4. Average Agency (solid line) and Zone (dashed line) judgments at each of the 7 levels of speed, in Experiment 2. Errors bars are 95% CIs. were significantly higher than zone judgments for low speeds, these two types of judgments converged, on average, for faster game speeds. 3.2.3. Separate regressions We then asked whether participants’ mean agency judgments predicted their zone judgments, and vice versa, by analyzing each judgment type with a separate multilevel linear regression model. These models included as predictors the linear and quadratic game speed, JOPs as standardized between- and within-person predictors (Bolger & Laurenceau, 2013), and the subject-specific mean of the other judgment as a standardized between-person predictor. Linear and quadratic speed, and the within-person JOPs were modeled as fixed and varying effects between participants. These regression models were fitted with standard maximum likelihood methods using the lme4 package in the R statistical programming environment (Bates et al., 2015; R Core Team, 2015); p-values for the regression parameters were computed using the Satterthwaite approximation (Kuznetsova, Brockhoff, & Christensen, 2015). These results, again, confirmed that the sense of agency was a monotonically decreasing function of game speed—albeit with a slight quadratic trend that was confirmed to be much smaller than the quadratic effect for zone judgments in the ^ =
14.54, 95% Bayesian analysis—and that JOPs were positively associated with agency judgments (linear effect of speed b ^ =
3.83, 95% CI [
6.79,
0.89], p = 0.016; JOP within-subjects CI [
22.96,
6.08], p = 0.002; quadratic effect of speed b ^ b = 10.31, 95% CI [7.38, 13.32], p < 0.001). Additionally, we found that JOPs predicted agency ratings between individuals: Participants who perceived their performance to be better, on average, gave higher judgments of agency than people who 139 M. Vuorre, J. Metcalfe / Consciousness and Cognition 43 (2016) 133–142 ^ = 18.65, 95% CI [5.66, 31.51], p = 0.005). Participants’ average judged their performance to be poorer (JOP between-subjects b ^ =
6.06, 95% CI [
18.92, 6.79], p = 0.26). zone judgments did not predict how in control they felt (b ^ = 16.57, 95% CI [7.80, 25.39], As for the zone judgments, we confirmed the earlier results (linear effect of speed b ^ =
13.94, 95% CI [
21.22,
6.66], p < 0.001; JOP within-subjects b ^ = 12.16, 95% CI p < 0.001; quadratic effect of speed b [9.44, 15.05], p < 0.001), and additionally found that JOPs are positively associated with zone judgments between individuals ^ = 22.50, 95% CI [16.82, 28.23], p < 0.001). This model also revealed that participants’ average agency (JOP between-subjects b judgments were negatively associated with their zone judgments: Participants who, on average, felt more in control, ^ =
7.54, 95% CI [
12.78,
2.06], p = 0.004). These results are shown reported lower levels of feelings of being in the zone (b in Fig. 5. 3.2.4. Relationship with judgments and performance Finally, we asked how agency and zone judgments relate to hit rate (proportion of Xs touched, of all possible Xs, in a trial) and ‘reward’ (an approximation of the absolute amount of reward the person obtained on each trial, calculated as the num- Agency Zone 30 Estimate 15 0 −15 −30 Speed Speed JOP JOP Zone linear quadratic within between between Speed Speed JOP JOP Agency linear quadratic within between between Effect Fig. 5. Average regression coefficients (with 95% Confidence Intervals) from regression analyses where agency judgments (left) and zone judgments (right) were regressed on linear and quadratic speed, within- and between-subjects standardized judgments of performance (JOP) and the subject-specific mean of the other judgment (between-subjects effect). Agency Zone Rating / Hit rate / Reward 100 80 ● ● ● ● ● 60 ● ● ● ● ● ● ● ● ● 40 20 0 1 2 3 4 5 6 7 1 2 3 4 5 6 7 Speed level Fig. 6. Average Agency and Zone judgments (solid lines), hit rates (dashed lines), and reward (dotted lines) at each 7 speed levels, Experiment 2. Error bars are 95% CIs. 140 M. Vuorre, J. Metcalfe / Consciousness and Cognition 43 (2016) 133–142 ber of hits). Based on the results of Kennedy et al. (2014) we expected the zone judgments to be most similar to reward, and that the pattern of agency judgments would be most similar to proportion correct. As observed in Fig. 6, Zone judgments mirrored participants’ reward: both displayed a quadratic association with speed. In contrast, agency judgments were more similar to proportion correct, which both displayed a linear association with speed. We then examined the withinparticipant relationship between reward and the two types of judgments with separate multilevel linear regressions for each ^ =
4.33, SE = 0.82, (standardized) judgment predicting reward: Reward was negatively associated with Agency judgments (b ^ = 3.85, SE = 1.23, p = 0.004). In contrast, hit rate was positively p < 0.001), but positively associated with Zone judgments (b ^ = 12.03, SE = 1.30, p < 0.001), but not significantly associated with Zone judgments associated with Agency judgments (b ^ =
0.21, SE = 2.12, p = 0.92). Judgments of performance, and actual performance (hit rate) were highly positively corre(b lated (mean within-participant correlation r = 0.56, 95% CI [0.48, 0.65]). These results suggested another qualitative dimension in which Zone and Agency experiences differ: Agency closely tracks the person’s objective task performance whereas Zone is more sensitive to the absolute rewards obtained in the game. 4. General discussion We found a robust dissociation between the sense of agency—the experience that ’I am in control of my actions’ and their outcomes—and the feeling of flow—the effortless, pleasurable absorption in a task. Specifically, the sense of agency decreased as the task became more difficult, whereas the feeling of flow peaked at middle values of difficulty, where task demands and the person’s abilities are in balance (Kennedy et al., 2014; Rheinberg & Vollmeyer, 2003). Furthermore, we found that participants who, on average, felt a stronger sense of agency, felt lower levels of flow than individuals who reported lower average levels of agency, but more research is needed to specifically address this relationship. However, we also found one dimension in which these metacognitive evaluations were similarly influenced: Both experiences were positively influenced by judgments of performance: Higher judgments of performance were associated with an elevated sense of agency and feeling of flow. Although previous experimental work has not directly addressed the relationship between the sense of agency and the experience of flow, it has been suggested that they are, to some degree, similarly influenced by the ‘‘smoothness” or experienced ease of the current task or action (Wenke et al., 2010). Contrary to this hypothesis, we showed that while the sense of agency is at its highest when the task is easy it decreases linearly as the task becomes more difficult. The experience of flow shows an entirely different pattern: Evaluations of flow increase as the task becomes more difficult, and peak at middle values of difficulty, where skills and demands are in balance. The present results are consistent with previous findings showing that the sense of agency is reduced by difficulty or dysfluency (Sidarus & Haggard, 2016, but note, this study provided no evidence on whether or not the sense of agency was increased by increased fluency). Other previous studies, however, have suggested a relationship between increased experiences of agency and increased flow, suggesting that both require a ‘smoothness’, or ease of action (Wenke et al., 2010). The present results indicate that the relationship between the experience of control and ’flow’ is more complicated than initially suggested, and these complex results may reflect an important, but unexplored, distinction between action fluency and the feeling of flow. In another study investigating individuals’ proneness to experience flow in the popular game Tetris, Keller and Blomann (2008) found that individuals who had a greater internal Locus of Control (LOC, e.g. Rotter, 1990) were more prone to experience flow states. Internal LOC is thought to index the degree to which individuals believe that their actions and efforts are instrumental in obtaining desired outcomes (in contrast to external LOC, reflecting beliefs that outcomes are not contingent on one’s efforts). Insofar as internal LOC reflects beliefs associated with a high degree of sense of agency, our results are in contradiction with Keller and Blomann’s (2008) results. We showed that when the experience of agency in a task is directly assessed, people who, on average, felt a stronger sense of agency reported lower levels of flow, and that the sense of agency and feelings of flow are differently influenced by task demands. This consideration suggests that the general belief in the causal impact of one’s actions (internal LOC), and the experience of controlling one’s actions (sense of agency) might not be directly related. Furthermore, the sense of being able to control one’s actions is often thought to be an important, though it is sometimes described as a paradoxical (see Young & Pain, 1999), characteristic of the flow experience (Keller & Blomann, 2008; Nakamura & Csikszentmihalyi, 2002). Our results further complicate this component of the phenomenology of flow. The fact that agency ratings decrease monotonically with game difficulty, but flow ratings do not, suggests that the sense of agency is at least partly a metacognitively accurate judgment of people’s ability to control their actions and those actions’ outcomes in the game (Metcalfe & Greene, 2007). Indeed, agency ratings were strongly related to both judgments of performance and to actual proportion correct. In contrast, the quadratic relationship between flow ratings and game difficulty suggest that the flow state has less to do with metacognitive evaluations of one’s control in the task, but instead reflects the match of skills and task demands, a positive assessment of one’s performance, and perhaps how enjoyable the experience is (Csikszentmihalyi, 2009; Kennedy et al., 2014). And, indeed, flow ratings were strongly associated with reward – the absolute number of X’s the person exploded during the entire trial – but not performance when the latter is computed as a proportion correct. Notably, on slow games performance and reward are quite different. A person may hit all of the X’s in slow game (and hence have a high proportion correct, and a high sense of agency) but if there were only 3 or 4 such X’s, because the game was so slow, the reward (and hence the sense of flow) may be small. At the fast speeds, proportion correct M. Vuorre, J. Metcalfe / Consciousness and Cognition 43 (2016) 133–142 141 decreases because the game is very difficult, but it turns out that reward, or the number of X’s hit over the entire trial, decreases as well because the X’s pass by so quickly that the person cannot explode them (Fig. 6). Another important distinction between the experiences of flow and agency is the degree to which the experienced self is involved in either state. An important characteristic of the flow state is a ‘‘Loss of reflective self-consciousness (i.e., loss of awareness of oneself as a social actor)” (Nakamura & Csikszentmihalyi, 2002, p. 90). In contrast, the central involvement of the self as the actor is a defining feature of the sense of agency (knowing that ‘‘I am the actor”; Gallagher, 2007). These different descriptions would be difficult to reconcile, if the sense of agency and the experience of flow were indeed one and the same—which they are not, as shown by the current results. Acknowledgments The authors would like to thank Judy Xu and Kelsey McLeod for help with data collection, and John Dewey and an anonymous reviewer for helpful comments. This research was funded by James S. McDonnell grant 220020166. References Bannister, R. (2014, March 30). The day I broke the four-minute mile. <http://www.telegraph.co.uk/sport/10731234/Roger-Bannister-The-day-I-broke-thefour-minute-mile.html> Retrieved 30 January 2016. Bates, D., Mächler, M., Bolker, B., & Walker, S. (2015). Fitting linear mixed-effects models using lme4. Journal of Statistical Software, 67(1), 1–48. http://dx.doi. org/10.18637/jss.v067.i01. Blakemore, S.-J., & Frith, C. (2003). Self-awareness and action. Current Opinion in Neurobiology, 13(2), 219–224. http://dx.doi.org/10.1016/S0959-4388(03) 00043-6. Blakemore, S.-J., Frith, C. D., & Wolpert, D. M. (1999). Spatio-temporal prediction modulates the perception of self-produced stimuli. Journal of Cognitive Neuroscience, 11(5), 551–559. http://dx.doi.org/10.1162/089892999563607. Blakemore, S.-J., Wolpert, D. M., & Frith, C. D. (1998). Central cancellation of self-produced tickle sensation. Nature Neuroscience, 1(7), 635–640. Bolger, N., & Laurenceau, J.-P. (2013). Intensive longitudinal methods: An introduction to diary and experience sampling research : . Guilford Press. Chambon, V., & Haggard, P. (2012). Sense of control depends on fluency of action selection, not motor performance. Cognition, 125(3), 441–451. http://dx. doi.org/10.1016/j.cognition.2012.07.011. Chambon, V., Sidarus, N., & Haggard, P. (2014). From action intentions to action effects: How does the sense of agency come about? Frontiers in Human Neuroscience, 8, 320. http://dx.doi.org/10.3389/fnhum.2014.00320. Csikszentmihalyi, M. (2009). Flow: The psychology of optimal experience : . Harper Collins. Gallagher, S. (2007). The natural philosophy of agency. Philosophy Compass, 2(2), 347–357. http://dx.doi.org/10.1111/j.1747-9991.2007.00067.x. Gelman, A., & Hill, J. (2007). Data analysis using regression and multilevel/hierarchical models : . Cambridge University Press. Gilovich, T., Vallone, R., & Tversky, A. (1985). The hot hand in basketball: On the misperception of random sequences. Cognitive Psychology, 17(3), 295–314. http://dx.doi.org/10.1016/0010-0285(85)90010-6. Hales, S. D. (1999). An epistemologist looks at the hot hand in sports. Journal of the Philosophy of Sport, 26(1), 79–87. http://dx.doi.org/10.1080/ 00948705.1999.9714580. Keller, J., & Blomann, F. (2008). Locus of control and the flow experience: An experimental analysis. European Journal of Personality, 22(7), 589–607. http://dx. doi.org/10.1002/per.692. Kennedy, P., Miele, D. B., & Metcalfe, J. (2014). The cognitive antecedents and motivational consequences of the feeling of being in the zone. Consciousness and Cognition, 30, 48–61. Kuznetsova, A., Brockhoff, P. B., & Christensen, R. H. B. (2015). lmerTest: Tests in linear mixed effects models Retrieved from<http://CRAN.R-project. org/package=lmerTest>. Love, J., Selker, R., Marsman, M., Jamil, T., Dropmann, D., Verhagen, J., . . . Wagenmakers, E.-J. (2015). JASP (Version 0.7). Massimini, F., & Carli, M. (1988). The systematic assessment of flow in daily experience. In Optimal experience. http://dx.doi.org/10.1017/ CBO9780511621956.016. Metcalfe, J., Eich, T. S., & Castel, A. D. (2010). Metacognition of agency across the lifespan. Cognition, 116(2), 267–282. http://dx.doi.org/10.1016/j. cognition.2010.05.009. Metcalfe, J., Eich, T. S., & Miele, D. B. (2013). Metacognition of agency: Proximal action and distal outcome. Experimental Brain Research, 229(3), 485–496. http://dx.doi.org/10.1007/s00221-012-3371-6. Metcalfe, J., & Greene, M. J. (2007). Metacognition of agency. Journal of Experimental Psychology: General, 136(2), 184–199. http://dx.doi.org/10.1037/00963445.136.2.184. Metcalfe, J., Snellenberg, J. X. V., DeRosse, P., Balsam, P., & Malhotra, A. K. (2012). Judgements of agency in schizophrenia: An impairment in autonoetic metacognition. Philosophical Transactions of the Royal Society of London B: Biological Sciences, 367(1594), 1391–1400. http://dx.doi.org/10.1098/ rstb.2012.0006. Miele, D. B., Wager, T. D., Mitchell, J. P., & Metcalfe, J. (2011). Dissociating neural correlates of action monitoring and metacognition of agency. Journal of Cognitive Neuroscience, 23(11), 3620–3636. Nakamura, J., & Csikszentmihalyi, M. (2002). The concept of flow. In Handbook of positive psychology (pp. 89–105). Pelé & Fish, R. L. (2013). My life and the beautiful game : . Skyhorse Publishing, Inc. R Core Team (2015). R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. Retrieved from <https://www.R-project.org/>. Rheinberg, F., & Vollmeyer, R. (2003). Flow-Erleben in einem Computerspiel unter experimentell variierten Bedingungen. Zeitschrift Für Psychologie/Journal of Psychology, 211(4), 161–170. http://dx.doi.org/10.1026//0044-3409.211.4.161. Rotter, J. B. (1990). Internal versus external control of reinforcement: A case history of a variable. American Psychologist, 45(4), 489–493. http://dx.doi.org/ 10.1037/0003-066X.45.4.489. Rouder, J. N., Morey, R. D., Speckman, P. L., & Province, J. M. (2012). Default Bayes factors for ANOVA designs. Journal of Mathematical Psychology, 56(5), 356–374. http://dx.doi.org/10.1016/j.jmp.2012.08.001. Sidarus, N., Chambon, V., & Haggard, P. (2013). Priming of actions increases sense of control over unexpected outcomes. Consciousness and Cognition, 22(4), 1403–1411. http://dx.doi.org/10.1016/j.concog.2013.09.008. Sidarus, N., & Haggard, P. (2016). Difficult action decisions reduce the sense of agency: A study using the Eriksen flanker task. Acta Psychologica, 166, 1–11. http://dx.doi.org/10.1016/j.actpsy.2016.03.003. Stan Development Team (2015). Stan: A C++ library for probability and sampling, version 2.9.0 Retrieved from<http://mc-stan.org/>. Stenner, M.-P., Bauer, M., Sidarus, N., Heinze, H.-J., Haggard, P., & Dolan, R. J. (2014). Subliminal action priming modulates the perceived intensity of sensory action consequences. Cognition, 130(2), 227–235. http://dx.doi.org/10.1016/j.cognition.2013.11.008. 142 M. Vuorre, J. Metcalfe / Consciousness and Cognition 43 (2016) 133–142 Wenke, D., Fleming, S. M., & Haggard, P. (2010). Subliminal priming of actions influences sense of control over effects of action. Cognition, 115(1), 26–38. http://dx.doi.org/10.1016/j.cognition.2009.10.016. Wolpert, D. M., & Flanagan, J. R. (2001). Motor prediction. Current Biology, 11(18), R729–R732. http://dx.doi.org/10.1016/S0960-9822(01)00432-8. Wolpert, D. M., Ghahramani, Z., & Jordan, M. I. (1995). An internal model for sensorimotor integration. Science, 269(5232), 1880–1882. http://dx.doi.org/ 10.1126/science.7569931. Young, J. A., & Pain, M. D. (1999). The zone: Evidence of a universal phenomenon for athletes across sports. Athletic Insight: The Online Journal of Sport Psychology, 1(3), 21–30.

Consciousness and Cognition 43 (2016) 102–112 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Implicit learning: A way to improve visual search in spatial neglect? Murielle Wansard a,⇑, Marie Geurten a, Catherine Colson b, Thierry Meulemans a a b Department of Psychology, Cognition and Behavior, Neuropsychology Unit, University of Liège, Belgium Service de Revalidation Neurologique, CHU Brugmann, Bruxelles, Belgium a r t i c l e i n f o Article history: Received 12 November 2015 Revised 25 May 2016 Accepted 25 May 2016 Available online 2 June 2016 Keywords: Visual neglect Spatial attention Association learning Consciousness Exploration of space Right-hemisphere a b s t r a c t Studies have shown that neglect patients are able to use stimulus regularities to orient faster toward the neglected side, without necessarily being aware of that information, or at the very least without being able to verbalize their knowledge. In order to better control for the involvement of explicit processes, the present study sought to test neglect patients’ ability to detect more complex associations between stimuli using tasks similar to those used in implicit learning studies. Our results demonstrate that neglect patients had difficulties implicitly learning complex associations, contrary to what we found with controls. The possible influence of attentional and working memory impairments are discussed. Ó 2016 Elsevier Inc. All rights reserved. 1. Introduction Right hemisphere damage following stroke is well known to lead to spatial neglect, particularly when distributed fronto-parietal networks are involved (Bartolomeo, Thiebaut de Schotten, & Doricchi, 2007; Bartolomeo, Zieren, Vohn, Dubois, & Sturm, 2008; Corbetta & Shulman, 2002; Ptak & Schnider, 2011). Neglect is characterized by an inability to detect events or to initiate motor responses towards the contralateral side of the brain lesion (Heilman, Watson, & Valenstein, 1984). Patients suffering from spatial neglect are frequently unaware of their difficulty (anosognosia). Consequently, they do not develop compensatory strategies in their daily life to orient their attention toward the left (Bartolomeo, 2014; Vuilleumier & Vocat, 2011). Although different impairments may contribute to the syndrome (including non-spatial deficits such as low levels of vigilance or arousal; see e.g., Malhotra, Coulthard, & Husain, 2009; Robertson, 2001), deficits in spatial attention are the core symptoms of neglect, including an inability to orient attention toward the contralesional left hemispace (Bartolomeo & Chokron, 2002) and difficulty disengaging attention from items on the right side (for a review, see Bartolomeo, Thiebaut de Schotten, & Chica, 2012). To deal with the massive amount of information available in our environment, our visual system must rapidly prioritize and select stimuli based on their pertinence either in terms of salience (e.g., potentially dangerous events) or their relevance to the observer’s internal goals and expectancies (Chica, Bartolomeo, & Lupianez, 2013; Ruz & Lupiáñez, 2002). This process of selection thus involves two modes of attentional orienting: attention can be directed to an object in space either automatically, in a relatively reflexive way (exogenous orienting), or in a more voluntary and controlled mode (endogenous ⇑ Corresponding author at: Department of Psychology, Cognition & Behavior, Boulevard du Rectorat, B33, B-4000 Liège, Belgium. E-mail address: m.wansard@ulg.ac.be (M. Wansard). http://dx.doi.org/10.1016/j.concog.2016.05.011 1053-8100/Ó 2016 Elsevier Inc. All rights reserved. M. Wansard et al. / Consciousness and Cognition 43 (2016) 102–112 103 orienting; Bartolomeo et al., 2012; Chica, Lupianez, & Bartolomeo, 2006; Fecteau & Munoz, 2006; Itti, Koch, Way, & Angeles, 2001). Attentional orienting has been extensively studied in spatial neglect using the simple manual reaction tasks developed by Posner (1980). These studies showed that exogenous orienting is particularly impaired in spatial neglect, with an initial orienting of attention towards the right side, followed by a difficulty in disengaging attention in order to reorient it leftward (Bartolomeo & Chokron, 2002; Bartolomeo, Decaix, Siéroff, & Chokron, 2001; Chica et al., 2011, 2013; Gainotti, D’Erme, & Bartolomeo, 1991; Losier & Klein, 2001; Natale, Posteraro, Prior, & Marzi, 2005; Posner, Walker, Friedrich, & Rafal, 1984; Smania et al., 1998). Interestingly, endogenous orienting seems to be relatively spared, albeit slowed, in neglect patients. The use of central spatial cues (e.g., an arrow indicating target location) or peripheral cues that are spatially predictive of the future location of the target (e.g., following a cue, the target will appear with high probability in the uncued location) can help neglect patients to orient leftward (Bartolomeo et al., 2001; Ladavas, Carletti, & Gori, 1994; Natale et al., 2005; Smania et al., 1998). However, although intact endogenous processes improve response times and accuracy in the detection of stimuli on the left side, they are not sufficient to completely abolish signs of neglect (Chica & Bartolomeo, 2012; Natale et al., 2005; Siéroff, Decaix, Chokron, & Bartolomeo, 2007). Moreover, in Bartolomeo et al. (2001), although participants were informed of the cue’s predictability before the experiment, some claimed not to have used this information to respond more quickly or precisely to targets. Thus, motivated strategic considerations or voluntary control (referred as endogenous orienting) may not be required to observe these attentional effects. In this sense, Lambert, Naikar, McLachlan, and Aitken (1999) questioned the purely controlled nature of endogenous orienting by demonstrating that visual orienting can occur independently of explicit knowledge of the predictive value of the cue (Chica & Bartolomeo, 2010; Chun, 2000; Lambert, 2003; Lopez-Ramon, Chica, Bartolomeo, & Lupiáñez, 2011; Risko & Stolz, 2010; Turk-Browne, Jungé, & Scholl, 2005; Zhao, Al-Aidroos, & Turk-Browne, 2013). Similarly, Bartolomeo, Decaix, and Siéroff (2007) found that subjects’ sensitivity to the probabilistic association between cues and targets in the Posner paradigm might be sustained by more implicit mechanisms, operating independent of consciousness or verbal reports (see also Decaix, Siéroff, & Bartolomeo, 2002; Rieth & Hubert, 2013). Recently, we demonstrated that some neglect patients were able to exploit the predictive value of cues presented on the right side to respond more quickly and accurately toward the neglected space without necessarily being able to describe the cue-target relationship (Wansard, Bartolomeo, Vanderaspoilden, Geurten, & Meulemans, 2015). In a Posner spatial cueing task, we distributed presentation of a peripheral cue such that the target appeared with high probability on the opposite side of space (80% invalid trials). The most effective strategy was therefore to orient attention towards the side opposite to the cue. Interestingly, in this study, neglect patients were able to inhibit the capture of attention from right ipsilateral cues and reorient toward contralesional targets without necessarily being aware of the cue-target relationship. This finding suggests that the probabilistic association between cues and targets location might be sustained by implicit processes, which might be preserved in some neglect patients. Other studies have investigated implicit processes (and more specifically spatial priming and statistical learning) in right brain-damaged patients with or without spatial neglect (for a recent review, see Shaqiri, Anderson, & Danckert, 2013). Using a visual search task in which a target’s location was repeated across successive trials, Kristjánsson, Vuilleumier, Husain, Malhotra, and Driver (2005) demonstrated that neglect patients were sensitive to spatial priming for position, but only if they had an unlimited amount of time to respond (see also Geng & Behrmann, 2002). Similarly, Shaqiri and Anderson (2012) showed that a chronic neglect patient was able to benefit from the statistical structure of target locations to improve his performance toward the neglected side when he was exposed to regularities over a long period of time (3 days, Shaqiri & Anderson, 2012). Together, these studies suggest that some forms of implicit processes can promote attentional orienting toward the left side in neglect patients in the absence of explicit awareness of the related contingencies. However, these findings are based on the use of relatively simple experimental designs, far from those encountered in the literature on implicit learning (Cleeremans, Destrebecqz, & Boyer, 1998). The paradigms generally used in implicit learning studies involve exposure to some complex rule-governed environment, such as a finite-state grammar, a deterministic spatial sequence, or a dynamic system (Cleeremans & Dienes, 2008). In addition, the absence of verbal report observed in studies involving neglect patients does not necessarily mean that explicit processes were not involved during learning; it could be that the patients were not confident enough in their knowledge to report it verbally (see Perruchet & Vinter, 2002; Shanks, 2003). Other measures, such as generation tasks, must be used in addition to the questionnaire to better characterize the nature (implicit vs. explicit) of the acquired knowledge. The objective of the present study was twofold: (1) to explore the ability of neglect patients to implicitly learn complex conditional associations between stimuli, using tasks that are closer to those usually employed in the implicit learning literature; (2) to determine whether knowledge of these conditional associations would help them to respond to stimuli located in their neglected side. For this purpose, given the difficulty of applying the classical serial reaction time (SRT) task in neglect patients, we developed a new paradigm, an adapted SRT task, in which sequences of letters presented at the center of the screen are predictive (or non-predictive) of the target location (see Section 2 for details). More specifically, we investigated neglect patients’ ability to use the predictive letters to orient faster toward the left and to increase their rate of detection of left targets. If participants learn about associations between letters and target location, this should produce a facilitation effect for targets following predictive letters compared with non-predictive letters over time. We expected the facilitation effect to be particularly clear at the end of the learning task, i.e., in the last block. 104 M. Wansard et al. / Consciousness and Cognition 43 (2016) 102–112 2. Method 2.1. Participants Ten patients with right-hemisphere damage suffering from left spatial neglect were recruited from hospitals in the French-speaking part of Belgium (mean age: 63.4 years; SD: 10.9, range: 44–79). The presence of neglect was assessed with the Batterie d’Evaluation de la Négligence (BEN; Azouvi et al., 2002) and was defined as performance outside the normal range on one or more of these clinical tests. The mean time interval between stroke onset and experimental testing was 7 months (range: 1–31 months). Demographic and clinical data for the neglect patients are shown in Table 1. Exclusion criteria were bilateral lesions, evidence of previous neurological diseases or psychiatric disorders, and the presence of left hemianopia. Twenty healthy controls matched for age, t(28) = 0.62, p = 0.54, and educational level, t(28) = 0.37, p = 0.71., constituted the control group. All participants were right-handed and reported to have normal or corrected-to-normal vision. Control participants had no known neurological or psychiatric history at the time of testing. The study was carried out following the guidelines of the University of Liège ethics review board. All participants gave their written informed consent prior to their participation in the study. 2.2. Apparatus and stimuli Stimulus presentation and response collection were performed using the E-prime software (Version 2.0, Psychology Software Tools, Pittsburgh, PA). All participants were seated comfortably at approximately 50 cm from the computer screen. Care was taken to align the vertical midline of the screen with the body midline. At the beginning of each trial, an empty black horizontal box (60 mm in width by 30 mm in height) was displayed centrally, on the top of the screen, against a silver background. As a cue, a sequence of uppercase consonant letters (M, P, T, B, and F) was presented in the center of the box, in black. On each trial, the letter ‘F’ terminated the sequence, thus announcing the appearance of the target. The target was an asterisk (12 mm in diameter) that appeared on the left or right side of the screen. 2.3. Procedure Fig. 1 displays the sequence and timing of a trial in the SRT task. Each trial began with the appearance of the box, in which a sequence of 2, 3, or 4 letters was presented, one letter at a time. Each letter was displayed for 700 ms. Participants had to read the letters aloud to ensure that they were centrally fixated and that they read each stimulus in the sequence. The target was presented until a response was given, for a maximum duration of 4 s to minimize the possibility of omissions. If no response was given within 4 s, the next trial was automatically initiated. Participants were instructed to press the ‘‘1” key on the numeric keypad if the target was presented on the left and the ‘‘2” key if the target appeared on the right. They were asked to detect the presence of the target as quickly and accurately as possible. The next trial began after a 200-ms responsestimulus interval. At the beginning of the experiment, participants completed a brief training period of 8 practice trials to familiarize them with the procedure. The experiment consisted of 4 blocks of 105 trials each. Unknown to the participants, the target location was related to the identity of the letter preceding the terminal letter ‘F’. When the letter ‘P’ preceded the letter ‘F’, the target appeared on the right, while the letter ‘M’ predicted the appearance of target on the left (predictive letters). By contrast, if the letter ‘T’ or ‘B’ preceded the letter ‘F’, the target appeared on either the left or the right side with equal probability (non-predictive letters). The predictability of the letters was counterbalanced between participants: for half of the participants, the predictive letters were ‘P’ and ‘M’ (‘T’ and ‘B’ being non-predictive); for the other half, the predictive letters were ‘T’ and ‘B’ (with ‘P’ and ‘M’ as non-predictive letters). Predictive and non-predictive trials were intermixed during the experiment. A particular letter could not appear more than once in the same sequence. Participants were allowed to take a short break after 210 trials, i.e., between Blocks 2 and 3. Both accuracy and reaction time (RTs) were collected for each trial. It is important to note that, because participants were exposed to predictive as well as non-predictive letters for both the left and the right side, our procedure allowed us to control the decrease in RTs for the left side, even in patients showing left neglect. With this procedure, rule-learning should be attested by faster RTs for predictive trials (compared to non-predictive ones) presented to the left side (as well as to those presented to the right side), even for patients showing a general slowdown for the stimuli presented on the left side. After the SRT task, participants’ explicit awareness of the predictive value of the letters was tested. First, they filled out a questionnaire asking them whether (1) they noticed anything special about the experiment, (2) the side that the target appeared on was determined by chance, and (3) there was any relationship between the sequence of letters and the target location. Afterward, subjects were administered a generation task in which they were required to predict the location of the target after each sequence of letters by pressing the ‘1’ or ‘2’ key for left and right, respectively. The generation task stimuli consisted of sequences of letters (of length 2, 3, or 4) similar to those presented during the SRT task. The task consisted of a single block of 40 trials (20 predictive vs. 20 non-predictive). Patient Sex/age/ education Days post stroke onset Etiology/ localisation Neglect severity Bells cancellation (left/right found targets, max = 15/15) Bells cancellation (time in seconds) Letter cancellation (left/right found targets, max = 30/30) Line bisection (mm of rightward deviation for 20-cm lines) Landscape drawing score Reading (left/right words, max = 61/55) Clock drawing P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 F/44/12 M/64/17 F/62/15 F/62/16 M/70/12 M/68/12 F/71/9 F/79/12 M/46/16 M/68/12 28 580 47 52 218 32 47 102 944 206 I/FTP (sylvian) I+H/FP Ins (sylvian) I+H/Bg ic H/T Th H/Bg I/FP ic I+H/OP I+H/Bg Th I/PT (sylvian) I/FP 0.38 0.56 0.13 0.44 0.19 0.38 0.38 0.24 0.56 0.63 10/12⁄ 14/15 15/15 9/12⁄ 12/14 10/8⁄ 11/14⁄ 10/15⁄ 13/13 12/12 81 231⁄ 150 131 150 215 178 218 255⁄ 293⁄ 26/29⁄ 6/30⁄ 30/29 29/30 24/27⁄ 8/23⁄ 29/26⁄ 20/30⁄ 25/29⁄ 28/27⁄ +3
7.5⁄ +5.5 +15.5⁄ +7.5⁄ +6
8⁄ +14⁄ +10⁄
10 1⁄ 2⁄ 1⁄ 1⁄ 0 1⁄ 1⁄ 1⁄ 1⁄ 2⁄ 61/55 60/55⁄ 61/55 61/55 61/55 51/55⁄ 56/55⁄ 3/40⁄ 22/55⁄ 46/42⁄ 0 0 1⁄ 0 0 0 0 0 2⁄ 1⁄ I: ischemic; H: hemorrhagic; F: frontal; P: parietal; T: temporal; O: Occipital; Th: thalamus; Ins: insula; Bg: basal ganglia; ic: internal capsule. For line bisection, positive values indicate rightward deviation, and negative values indicate leftward deviation. Score for landscape drawing indicates the number of omitted left-sided details. Score > 0 for clock drawing indicates left-sided details missing. Asterisks denote pathological performance. M. Wansard et al. / Consciousness and Cognition 43 (2016) 102–112 Table 1 Demographic, neurological and clinical data on neglect patients. 105 106 M. Wansard et al. / Consciousness and Cognition 43 (2016) 102–112 Fig. 1. Schematic representation of the sequence and timing of a trial in the SRT task. 3. Results 3.1. Response time Only correct responses were included in the RT analyses. Trials with RTs outside the range of 150–2500 ms were also discarded. For control participants, this resulted in the exclusion of 1% of responses. Neglect patients either omitted to respond or gave an anticipated response on 12% of trials. Because neglect patients differ significantly from age-matched controls on both median RTs, F(1, 28) = 15.71, p < 0.001, g2p = 0.36, and variances, F(1, 28) = 29.71, p < 0.001, g2p = 0.51, for the right side, results from the two groups of participants were analyzed separately (Bartolomeo et al., 2001). For each group (controls and neglect patients), median RTs were entered into a repeated-measures analysis of variance (ANOVA) with block (1, 2, 3, 4), predictability (predictive, non-predictive), and side (left, right) as within-participant factors (see Table 2). 3.1.1. Controls The analysis showed a significant main effect of block, F(3, 57) = 6.79, p < 0.001, g2p = 0.26, indicating a gradual decrease in RTs over the blocks. There was also a main effect of predictability, F(1, 19) = 11.45, p = 0.003, g2p = 0.38, with slower RTs for nonpredictive trials (488 ms) compared to predictive trials (481 ms). The interaction between block and predictability did not reach significance, F(3, 57) = 0.50, p = 0.69, g2p = 0.03. However, the critical hypothesis in the present study was that the learning effect would mainly appear at the end of the experiment. For this reason, we chose to examine whether participants demonstrated lower RTs for predictive than non-predictive sequences in the last block of the task. Planned comparisons confirmed that a significant difference between predictive (mean = 463 ms) and non-predictive (474 ms) trials appeared in the fourth block, F(1, 19) = 16.53, p < 0.001, g2p = 0.47. The results were not significant for the three other blocks, [Fs < 2.55] (see Fig. 2). No other effect or interaction reached significance. 3.1.2. Neglect patients The RT analysis for the neglect group revealed a significant main effect of block, F(3, 27) = 3.6, p = 0.03, g2p = 0.29, showing that RTs decreased progressively with successive blocks. A main effect of side was also found, F(1, 9) = 9.70, p = 0.01, g2p = 0.52, indicating that RTs were slowed for left-side stimuli. As in the control group, the interaction between block and predictability was not significant, F(3, 27) = 0.12, p = 0.94, g2p = 0.01. Planned comparisons did not reveal a difference between predictive and non-predictive trials in Block 4, either for the left side, F(1, 9) = 0.48, p = 0.51, g2p = 0.05, or the right, F(1, 9) = 0.00, p = 0.99, g2p = 0.00 (see Fig. 3). 3.2. Omissions The percentage of missed trials for each condition was analyzed through a repeated-measures ANOVA. Only neglect patients’ scores were analyzed because the data showed no variance for controls, who only missed 0.09% of targets. The results showed a main effect of predictability, F(1, 9) = 5.31, p = 0.04, g2p = 0.37, with more omissions on predictive (6%) than on non-predictive trials (5%). There was also a main effect of side, F(1, 9) = 4.8, p = 0.05, g2p = 0.35, showing that neglect patients made more omissions on the left side (8%) than on the right (3%). Neglect participants’ omissions on the left side decreased numerically between the first (14%) and the fourth block (6%) on predictive trials, but this effect did not reach statistical significance, F(1, 9) = 2.53, p = 0.15, g2p = 0.22 (see Fig. 4) and is not statistically different from 0 (t = 1.59, p = 0.15). No other main effect or interaction was significant [Fs < 1.59]. Interestingly, the absence of significant block  predictability  side interaction suggests that the higher rate of omissions for the predictive than the non-predictive trials did not results from differences for a specific block or a specific side. Group RTs Block 1 Block 2 Predictive Control group Neglect group Non-predictive Block 3 Predictive Non-predictive Block 4 Predictive Non-predictive Predictive Non-predictive Left Right Left Right Left Right Left Right Left Right Left Right Left Right Left Right 491 (78.59) 953 (306.9) 495 (86.43) 702 (218.43) 493 (86.86) 928 (300.52) 503 (94.02) 704 (232.41) 512 (104.6) 977 (381.68) 509 (91.76) 805 (395.81) 520 (87.67) 1029 (410.78) 517 (95.02) 765 (315.11) 458 (67.23) 899 (300.29) 458 (82.02) 669 (219.15) 464 (73.25) 899 (305.75) 459 (76.79) 648 (155.95) 459 (73.24) 887 (298.22) 467 (90.77) 705 (261.49) 465 (79.71) 906 (340.75) 483 (92.97) 705 (235.5) M. Wansard et al. / Consciousness and Cognition 43 (2016) 102–112 Table 2 Reaction times (means and SDs) by side and condition for the neglect and control groups. 107 108 M. Wansard et al. / Consciousness and Cognition 43 (2016) 102–112 Fig. 2. Response times (in ms) for controls by block (1, 2, 3, 4) and predictability (predictive vs. non-predictive). Fig. 3. Response times (in ms) for neglect patients by block (1, 2, 3, 4), side (left vs. right), and predictability (predictive vs. non-predictive). Fig. 4. Percentage of missed trials for the left side by block (1, 2, 3, 4) and predictability (predictive vs. non predictive). 3.3. Level of consciousness 3.3.1. Post-experiment questionnaire On the post-experiment questionnaire, all participants reported having noticed that the sequences were always formed with the same letters, or that the letter ‘F’ terminated the sequence in each trial. Some participants reported having tried to predict the side of the target based on the number of letters included in the sequence, without success. None of the control participants correctly detected the relationship between the sequence of letters and the target’s location. One control subject reported that there was a relationship between the identity of the letters and the target location, M. Wansard et al. / Consciousness and Cognition 43 (2016) 102–112 109 Table 3 Mean of correct responses and standard deviation on the generation task by group. Group Control group Neglect group Patient P2 % Total Left Right 52 (0.09) 51 (0.12) 75 54 (0.17) 46 (0.23) 80 51 (0.13) 56 (0.21) 70 but was unable to describe this relationship. Thus, it appears that control participants had very limited reportable knowledge of the predictive value of the letters in our study. By contrast, one neglect patient (P2) correctly described the association between the letter and the target location for the right side. To establish whether, in light of this ability to verbalize in the post-questionnaire, P20 s learning performance might be closer to (or even better than) that of the control group, we calculated a learning score using data from the fourth block by subtracting median RT on predictive trials from median RT on non-predictive trials. Due to the massive amount of individual variability in RTs, we corrected each participant’s learning score by dividing it by their average RT, in order to be able to compare neglect patient P20 s performance to that of the control group. We performed a modified one-tailed t-test to compare P20 s learning scores for the left and the right side to those of the normative sample (Crawford, Garthwaite & Porter, 2010). Neglect patient P2 learned significantly better than controls on the SRT task for targets presented on the left side, t = 2.56, p = 0.001, Z-CC = 2.62; and tended to perform better for the targets presented on the right side, t = 1.46, p = 0.08, Z-CC = 1.49. 3.3.2. Generation task To determine whether participants were able to predict the side that the target would appear on following the predictive letters, we determined the percentage of correct responses on the generation task for each participant. Chance level was 50%. It appears that neither control participants, t(19) = 0.93, p = 0.37, nor neglect patients, t(9) = 0.26, p = 0.80, performed better than chance for the predictive letters on the generation task (see Table 3). Interestingly, neglect patient P2, who correctly detected the predictability value of the letter for the right side, performed better than chance at completing the sequences, t(19) = 2.52, p = 0.02. In addition, he showed a tendency to perform better than the control group in the generation task, both for the right, t = 1.59, p = 0.06, Z-CC = 1.63, and the left side, t = 1.62, p = 0.06, Z-CC = 1.66. 4. Discussion Recent studies have demonstrated that some neglect patients are able to use spatial contingencies to orient faster toward the left, without being explicitly aware that they are doing so (Shaqiri & Anderson, 2013; Wansard et al., 2015). However, neglect patients’ ability to implicitly learn more complex associations between stimuli, referred to as implicit learning, has not yet been investigated. The present study was designed to explore this question, and to observe whether neglect patients are able to exploit implicit learning mechanisms to orient faster toward the left or increase their detection rate of left targets. In the control group, there appears to be evidence of a gradual increase of sensitivity to the predictability value of the letters, which results in a statistical difference between response times to predictive and non-predictive letters in the last block of trials. Importantly, this learning effect appeared without awareness of the association between letters and target location. Indeed, none of the participants was able either to correctly describe the relationship between the two types of stimuli on the post-questionnaire or to achieve better than chance performance in the generation task, indicating that the experimental paradigm used in the present experiment allowed us to implicitly manipulate visual orienting in healthy controls. Our results in neglect patients showed the expected effect of side, with slower RTs and higher numbers of omissions on the left side than on the right. The neglect patients did not show any benefit of predictability: we observed no significant improvement in RTs or omissions following the predictive letters across the four blocks, regardless of target hemifield. However, we did find a numerical (but non-significant) decrease for left-omitted targets between the first (14%) and the last (5%) block of trials, specific to predictive trials. Curiously, one neglect patient (P2) became explicitly aware of the letter-position associations, but only for right-side targets, leading to better performance than the control group in both the SRT task and the generation task. Perhaps even more surprising was this patient’s ability to perform better than healthy controls for the targets presented on the left side. However, statistical analysis revealed that P2 performed substantially better than chance in the generation task (80%, see Table 3). In this context, we would suggest that this patient’s performance for left targets was also supported by explicit processes, even if this knowledge was not sufficiently precise or strong to allow him to describe the relationship for the left side of space in the post-questionnaire. 110 M. Wansard et al. / Consciousness and Cognition 43 (2016) 102–112 Contrary to what we observed in our previous study with a simple manual reaction time task (Wansard et al., 2015), the overall results of neglect patients failed to reveal any implicit learning of complex associations. However, we observed that the ability to exploit the statistical relationships between stimuli appeared late in the control group. It is thus quite possible that patients with spatial neglect need a longer learning time to show the same facilitation effect, as found by Shaqiri and Anderson (2013). It would be interesting, in a future study, to increase the number of learning sessions and distribute them over several days to observe whether neglect patients also show successful implicit learning. Furthermore, the fact that, in our neglect group, the predictive cues did not exert a significant effect on reaction times could be explained by a combination of multiple impairments. Currently, theories of attentional orienting refer to two kinds of independent processes: exogenous orienting, driven by salient stimuli, and endogenous orienting, guided by internal goals and expectancies (Chica et al., 2013; Ruz & Lupiáñez, 2002). Some authors have proposed that implicit learning could also be used to guide visual attention, and therefore be considered as a third source of attentional orienting, separate from both reflexive (i.e., bottom-up) and voluntary (i.e., top-down) orienting (Jiang, Swallow, & Rosenbaum, 2013; Lambert, Norris, Naikar, & Aitken, 2000; Lambert et al., 1999). Although they are considered independent, a number of studies have highlighted mutual interference between these different modes of orienting (Berger, Henik, & Rafal, 2005; Jiang et al., 2013; Santangelo & Spence, 2008). In particular, it has been found that while regularities in implicit learning can guide attentional orienting (Jiang & Chun, 2003; Zhao et al., 2013), selective attentional processes can also determine what information can be implicitly learned (Baker, Olson, & Behrmann, 2004; Jimenez & Mendez, 1999; Lambert, 2003; Turk-Browne et al., 2005). Therefore, one possible explanation for the lack of results is that the attentional deficits which characterize spatial neglect (i.e., exogenous orienting, but also endogenous orienting for some patients; see Siéroff et al., 2007) prevent the operation of implicit learning. On the other hand, Umemoto, Scolari, Vogel, and Awh (2010) found a close relationship between working memory resources and statistical learning. In a change detection task, these authors observed that regularities in target position influence encoding in visual working memory, in the absence of explicit knowledge of these regularities. It has recently been suggested that impairments in spatial working memory might exacerbate the neglect syndrome, particularly the trend to repeatedly return to the same items located on the right side of space during search (Husain et al., 2001; Kennard et al., 2001; Kristjánsson & Vuilleumier, 2010; Mannan et al., 2005; Wansard et al., 2014), or failure to retain visual patterns and/or their serial order (Malhotra, Mannan, Driver, & Husain, 2004; Malhotra et al., 2005; Wansard et al., 2015). In keeping with these views, it can be suggested that the combination of attentional and working memory impairments frequently encountered in spatial neglect could prevent the extraction of the cue-target relationship required for implicit learning. However, establishing the precise nature of the interactions between attention, working memory, and implicit learning will require further investigations. In addition to these functional hypothesis, another explanation – not incompatible with the former – could be structural. Some recent studies suggested that some forms of statistical learning require processes specific to the right hemisphere (Roser, Fiser, Aslin, & Gazzaniga, 2011; Turk-Browne, Scholl, Chun, & Johnson, 2009) or on inter-hemispheric cooperation (Schmitz, Pasquali, Cleeremans, & Peigneux, 2013). Therefore, these processes could be impaired in neglect patients who are all right brain damaged subjects, either by focal brain damage to the right hemisphere or from dysfunction of largescale networks (inter or intra-hemispheric disconnection; see Lunven et al., 2015). Future studies will need to examine to what extent implicit learning is dependent on attention, working memory, or lesion location (study of anatomical data and/or structural connectivity), which in turn would contribute to better understanding the potential benefits of this method of learning in the rehabilitation of neglect patients. The observation that some more simple implicit attentional processes may help neglect patients to improve their exploration of the left side has important implications for rehabilitation (Wansard et al., 2015). As previously mentioned, spatial neglect is often associated with patients’ unawareness of difficulties (anosognosia), making this disorder particularly difficult to rehabilitate (Beschin, Cocchini, Allen, & Della Sala, 2012). Currently, the methods used in the rehabilitation of spatial neglect allow a slow and often temporary improvement of symptoms, except when different therapeutic approaches are combined (Saevarsson, Halsband, & Kristjánsson, 2011; Saevarsson, Kristjánsson, & Halsband, 2010). Following Shaqiri et al. (2013), we suggest that implicit learning processes (provided that the conditions for their effectiveness are determined) could serve as a new tool for the rehabilitation of neglect patients, since they could lead to involuntary attention shifts toward the neglected side. In order to maximize the benefits of implicit learning for neglect patients, a subliminal cue1 could be included in the paradigm used here, such as an arrow pointing toward the location predicted by the letter, or a brief brightening of the predictive letters. Lack of awareness of the cue prevents participants from directly taking advantage of the predictive information that it provides, but the abrupt luminance change would lead to attentional capture (Ruz & Lupiáñez, 2002). By making the predictive stimuli more salient, we might observe additional effects on learning and potentially, as a consequence, develop effective new tools to rehabilitate spatial neglect. In conclusion, the results of the present study suggest that neglect patients have difficulty implicitly learning complex associations between stimuli in order to orient faster toward predictive targets, contrary to unimpaired controls. We argue that attentional and working memory impairments, which are frequently encountered in spatial neglect could explain, at least in part, this absence of learning. It remains unclear what kind of attentional processes are necessary to or influence 1 Cue presented below participants’ subjective threshold of consciousness. M. Wansard et al. / Consciousness and Cognition 43 (2016) 102–112 111 the acquisition of implicit knowledge. Considering the substantial advantage that this mode of learning could provide in the rehabilitation of neglect patients, this question is of great practical but also theoretical value, and therefore deserves to be further explored in future studies. Acknowledgements This research was supported by a Grant from the National Fund for Scientific Research (FRS – FNRS, Belgium) to MG. The authors do not have any conflicts of interest to disclose. References Azouvi, P., Samuel, C., Louis-Dreyfus, A., Bernati, T., Bartolomeo, P., Beis, J. M., ... Rousseaux, M. (2002). Sensitivity of clinical and behavioral tests of spatial neglect after right hemisphere stroke. Journal of Neurology Neurosurgery and Psychiatry, 73, 160–166. Baker, C. I., Olson, C. R., & Behrmann, M. (2004). Role of attention and perceptual grouping in visual statistical learning. Psychological Science, 15, 460–466. Bartolomeo, P. (2014). Attention disorders after right brain damage: Living in halved worlds. London: Springer-Verlag. Bartolomeo, P., & Chokron, S. (2002). Orienting of attention in left unilateral neglect. Neuroscience and Biobehavioral Reviews, 26(2), 217–234. Bartolomeo, P., Decaix, C., & Siéroff, E. (2007). The phenomenology of endogenous orienting. Consciousness and Cognition, 16(1), 144–161. http://dx.doi.org/ 10.1016/j.concog.2005.09.002. Bartolomeo, P., Decaix, C., Siéroff, E., & Chokron, S. (2001). Modulating the attentional bias in unilateral neglect: The effects of the strategic set. Experimental Brain Research, 137(3–4), 432–444. http://dx.doi.org/10.1007/s002210000642. Bartolomeo, P., Thiebaut de Schotten, M., & Chica, A. B. (2012). Brain networks of visuospatial attention and their disruption in visual neglect. Frontiers in Human Neuroscience, 6. http://dx.doi.org/10.3389/fnhum.2012.00110, pp. 1–10. Bartolomeo, P., Thiebaut de Schotten, M., & Doricchi, F. (2007). Left unilateral neglect as a disconnection syndrome. Cerebral Cortex, 17(11), 2479–2490. Bartolomeo, P., Zieren, N., Vohn, R., Dubois, B., & Sturm, W. (2008). Neural correlates of primary and reflective consciousness of spatial orienting. Neuropsychologia, 46(1), 348–361. http://dx.doi.org/10.1016/j.neuropsychologia.2007.07.005. Berger, A., Henik, A., & Rafal, R. (2005). Competition between endogenous and exogenous orienting of visual attention. Journal of Experimental Psychology: General, 134(2), 207–221. http://dx.doi.org/10.1037/0096-3445.134.2.207. Beschin, N., Cocchini, G., Allen, R., & Della Sala, S. (2012). Anosognosia and neglect respond differently to the same treatments. Neuropsychological Rehabilitation, 22(4), 550–562. http://dx.doi.org/10.1080/09602011.2012.669353. Chica, A. B., & Bartolomeo, P. (2010). Unconscious strategies? Commentary on Risko and Stolz (2010): ‘‘the proportion valid effect in covert orienting: Strategic control or implicit learning?”. Consciousness and Cognition, 19(1), 443–444. http://dx.doi.org/10.1016/j.concog.2009.09.010. Chica, A. B., & Bartolomeo, P. (2012). Attentional routes to conscious perception. Frontiers in Psychology, 3, 1–12. http://dx.doi.org/10.3389/ fpsyg.2012.00001. Chica, A. B., Bartolomeo, P., & Lupianez, J. (2013). Two cognitive and neural systems for endogenous and exogenous spatial attention. Behavioral Brain Research, 237, 107–123. http://dx.doi.org/10.1016/j.bbr.2012.09.027. Chica, A. B., Lupianez, J., & Bartolomeo, P. (2006). Dissociating inhibition of return from endogenous orienting of spatial attention: Evidence from detection and discrimination tasks. Cognitive Neuropsychology, 23(7), 1015–1034. http://dx.doi.org/10.1080/02643290600588277. Chica, A. B., Thiebaut de Schotten, M., Toba, M., Malhotra, P., Lupianez, J., & Bartolomeo, P. (2011). Attention networks and their interactions after righthemisphere damage. Cortex, 48(6), 654–663. http://dx.doi.org/10.1016/j.cortex.2011.01.009. Chun, M. M. (2000). Contextual cueing of visual attention. Trends in Cognitive Sciences, 4(5), 170–178. http://dx.doi.org/10.1016/S1364-6613(00)01476-5. Cleeremans, A., Destrebecqz, A., & Boyer, M. (1998). Implicit learning: News from the front. Trends in Cognitive Sciences, 2(10), 406–416. http://dx.doi.org/ 10.1016/S1364-6613(98)01232-7. Cleeremans, A., & Dienes, Z. (2008). In R. Sun (Ed.), Computational models of implicit learning (pp. 396–421). Cambridge, UK: Cambridge University Press. Corbetta, M., & Shulman, G. L. (2002). Control of goal-directed and stimulus-driven attention in the brain. Nature Reviews Neuroscience, 3(3), 201–215. http:// dx.doi.org/10.1038/nrn755. Decaix, C., Siéroff, E., & Bartolomeo, P. (2002). How voluntary is ‘voluntary’ orienting of attention? Cortex, 38, 841–845. http://dx.doi.org/10.1016/S00109452(08)70053-4. Fecteau, J. H., & Munoz, D. P. (2006). Salience, relevance, and firing: A priority map for target selection. Trends in Cognitive Sciences, 10(8), 382–390. http://dx. doi.org/10.1016/j.tics.2006.06.011. Gainotti, G., D’Erme, P., & Bartolomeo, P. (1991). Early orientation of attention toward the half space ipsilateral to the lesion in patients with unilateral brain damage. Journal of Neurology Neurosurgery and Psychiatry, 54, 1082–1089. Geng, J. J., & Behrmann, M. (2002). Probability cuing of target location facilitates visual search implicitly in normal participants and patients with hemispatial neglect. Psychological Science, 13(6), 520–525. http://dx.doi.org/10.1111/1467-9280.00491. Heilman, K. M., Watson, R. T., & Valenstein, E. (1984). In K. M. Heilman & E. Valenstein (Eds.), Neglect and related disorders (pp. 243–293). New York, NY: Oxford University Press. Husain, M., Mannan, S., Hodgson, T., Wojciulik, E., Driver, J., & Kennard, C. (2001). Impaired spatial working memory across saccades contributes to abnormal search in parietal neglect. Brain, 124(5), 941–952. http://dx.doi.org/10.1093/brain/124.5.941. Itti, L., Koch, C., Way, W., & Angeles, L. (2001). Computational modelling of visual attention. Nature Reviews Neuroscience, 2(3), 194–203. Jiang, Y., & Chun, M. M. (2003). Contextual cueing: Reciprocal influences between attention and implicit learning. In L. Jimenez (Ed.), Attention and implicit learning (pp. 277–296). Amsterdam, Netherlands: John Benjamins Publishing Company. Jiang, Y., Swallow, K. M., & Rosenbaum, G. M. (2013). Guidance of spatial attention by incidental learning and endogenous cuing. Journal of Experimental Psychology: Human Perception and Performance, 39(1), 285–297. http://dx.doi.org/10.1037/a0028022. Jimenez, L., & Mendez, C. (1999). Which attention is needed for implicit sequence learning? Journal of Experimental Psychology: Learning, Memory, and Cognition., 25(1), 236–259. Kennard, C., Husain, M., Mannan, S., Hodgson, T., Wojciulik, E., & Driver, J. (2001). Abnormal visual search in parietal neglect: A defect of spatial working memory. Annals of Neurology, 50(3) (p. S39). Kristjánsson, Á., & Vuilleumier, P. (2010). Disruption of spatial memory in visual search in the left visual field in patients with hemispatial neglect. Vision Research, 50(14), 1426–1435. http://dx.doi.org/10.1016/j.visres.2010.03.001. Kristjánsson, Á., Vuilleumier, P., Husain, M., Malhotra, P., & Driver, J. (2005). Priming of colour and position during visual search in unilateral spatial neglect. Journal of Cognitive Neuroscience, 17(6), 859–873. Ladavas, E., Carletti, M., & Gori, G. (1994). Automatic and voluntary orienting attention in patients with visual neglect: Horizontal and vertical dimensions. Neuropsychologia, 32, 1195–1208. Lambert, A. (2003). In L. Jimenez (Ed.), Visual orienting, learning, and conscious awareness (pp. 253–275). Amsterdam: John Benjamins Publishing. Lambert, A., Naikar, N., McLachlan, K., & Aitken, V. (1999). A new component of visual orienting: Implicit effects of peripheral information and subtreshold cues on covert attention. Journal of Experimental Psychology: Human Perception and Performance, 25(2), 231–240. 112 M. Wansard et al. / Consciousness and Cognition 43 (2016) 102–112 Lambert, A., Norris, A., Naikar, N., & Aitken, V. (2000). Effects of informative peripheral cues on eye movements: Revisiting William James derived attention. Visual Cognition, 7(5), 545–569. http://dx.doi.org/10.1080/135062800407194. Lopez-Ramon, M. F., Chica, A. B., Bartolomeo, P., & Lupiáñez, J. (2011). Attentional orienting and awareness: Evidence from a discrimination task. Consciousness and Cognition, 20(3), 745–755. http://dx.doi.org/10.1016/j.concog.2010.10.024. Losier, B. J. W., & Klein, R. M. (2001). A review of the evidence for a disengage deficit following parietal lobe damage. Neuroscience and Biobehavioral Reviews, 25, 1–13. Lunven, M., Thibeaut de Schotten, M., Bourlon, C., Duret, C., Migliaccio, L., Rode, G., & Bartolomeo, P. (2015). White matter lesional predictors of chronic visual neglect: A longitudinal study. Brain, 138(3), 746–760. http://dx.doi.org/10.1093/brain/awu389. Malhotra, P., Coulthard, E., & Husain, M. (2009). Role of right posterior parietal cortex in maintaining attention to spatial locations over time. Brain, 132, 645–660. http://dx.doi.org/10.1093/brain/awn350. Malhotra, P., Jager, H., Parton, A., Greenwood, R., Playford, E., Brown, M. M., ... Husain, M. (2005). Spatial working memory capacity in unilateral neglect. Brain, 128(2), 424–435. http://dx.doi.org/10.1093/brain/awh372. Malhotra, P., Mannan, S., Driver, J., & Husain, M. (2004). Impaired spatial working memory: One component of the visual neglect syndrome? Cortex, 40(4–5), 667–676. http://dx.doi.org/10.1016/S0010-9452%2808%2970163-1. Mannan, S., Mort, D., Hodgson, T., Driver, J., Kennard, C., & Husain, M. (2005). Revisiting previously searched locations in visual neglect: Role of right parietal and frontal lesions in misjudging old locations as new. Journal of Cognitive Neuroscience, 17(2), 340–354. Natale, E., Posteraro, L., Prior, M., & Marzi, C. (2005). What kind of visual spatial attention is impaired in neglect? Neuropsychologia, 43(7), 1072–1085. http:// dx.doi.org/10.1016/j.neuropsychologia.2004.10.002. Perruchet, P., & Vinter, A. (2002). The self-organizing consciousness. Behavioral and Brain Sciences, 25, 297–388. Posner, M. (1980). Orienting of attention. Quarterly Journal of Experimental Psychology, 32, 3–25. Posner, M., Walker, J. A., Friedrich, F. J., & Rafal, R. (1984). Effects of parietal injury on covert orienting of attention. The Journal of Neuroscience, 4(7), 1863–1874. Ptak, R., & Schnider, A. (2011). The attention network of the human brain: Relating structural damage associated with spatial neglect to functional imaging correlates of spatial attention. Neuropsychologia, 49(11), 3063–3070. http://dx.doi.org/10.1016/j.neuropsychologia.2011.07.008. Rieth, C. A., & Hubert, D. E. (2013). Implicit learning of spatiotemporal contingencies in spatial cueing. Journal of Experimental Psychology: Human perception and Performance, 39(4), 1165–1180. http://dx.doi.org/10.1037/a0030870. Risko, E. F., & Stolz, J. A. (2010). The proportion valid effect in covert orienting: Strategic control or implicit learning? Consciousness and Cognition, 19(1), 432–442. http://dx.doi.org/10.1016/j.concog.2009.07.013. Robertson, I. H. (2001). Do we need the ‘‘lateral” in unilateral neglect? Spatially nonselective attention deficits in unilateral neglect and their implications for rehabilitation. NeuroImage, 14(1 Pt 2), S85–S90. http://dx.doi.org/10.1006/nimg.2001.0838. Roser, M. E., Fiser, J., Aslin, R. N., & Gazzaniga, M. S. (2011). Right hemisphere dominance in visual statistical learning. Journal of Cognitive Neuroscience, 23(5), 1089–1099. http://dx.doi.org/10.1162/jocn.2010.21508. Ruz, M., & Lupiáñez, J. (2002). A review of attentional capture: On its automaticity and sensitivity to endogenous control. Psicologica, 23(2), 283–369. Saevarsson, S., Halsband, U., & Kristjánsson, Á. (2011). Designing rehabilitation programs for neglect: Could 2 be more than 1 + 1? Applied Neuropsychology, 18, 95–106. http://dx.doi.org/10.1080/09084282.2010.547774. Saevarsson, S., Kristjánsson, Á., & Halsband, U. (2010). Strength in numbers: Combining neck vibration and prism adaptation produces additive therapeutic effects in unilateral neglect. Neuropsychological Rehabilitation, 20(5), 704–724. http://dx.doi.org/10.1080/09602011003737087.Strength. Santangelo, V., & Spence, C. (2008). Is the exogenous orienting of spatial attention truly automatic? Evidence from unimodal and multisensory studies. Consciousness and Cognition, 17(3), 989–1015. http://dx.doi.org/10.1016/j.concog.2008.02.006. Schmitz, R., Pasquali, A., Cleeremans, A., & Peigneux, P. (2013). Lateralized implicit sequence learning in uni- and bi-manual conditions. Brain and Cognition, 81(1), 1–9. http://dx.doi.org/10.1016/j.bandc.2012.09.002. Shanks, D. R. (2003). Attention and awareness in ‘‘implicit” sequence learning. In L. Jimenez (Ed.), Attention implicit learning (pp. 11–42). Amsterdam: John Benjamins Publishing. Shaqiri, A., & Anderson, B. (2012). Spatial probability cuing and right hemisphere damage. Brain and Cognition, 80(3), 352–360. http://dx.doi.org/10.1016/j. bandc.2012.08.006. Shaqiri, A., & Anderson, B. (2013). Priming and statistical learning in right brain damaged patients. Neuropsychologia, 51(13), 2526–2533. Shaqiri, A., Anderson, B., & Danckert, J. (2013). Statistical learning as a tool for rehabilitation in spatial neglect. Frontiers in Human Neuroscience, 7. http://dx. doi.org/10.3389/fnhum.2013.00224 (pp. 223–237). Siéroff, E., Decaix, C., Chokron, S., & Bartolomeo, P. (2007). Impaired orienting of attention in left unilateral neglect: A componential analysis. Neuropsychology, 21(1), 94–113. http://dx.doi.org/10.1037/0894-4105.21.1.94. Smania, N., Martini, M. C., Gambina, G., Tomelleri, G., Palamara, A., Natale, E., & Marzi, C. A. (1998). The spatial distribution of visual attention in hemineglect and extinction patients. Brain, 121, 1759–1770. Turk-Browne, N. B., Jungé, J., & Scholl, B. J. (2005). The automaticity of visual statistical learning. Journal of Experimental Psychology. General, 134(4), 552–564. http://dx.doi.org/10.1037/0096-3445.134.4.552. Turk-Browne Scholl, B. J., Chun, M. M., & Johnson, M. K. (2009). Neural evidence of statistical learning: Efficient detection of visual regularities without awareness. Journal of Cognitive Neuroscience, 21(10), 1934–1945. http://dx.doi.org/10.1162/jocn.2009.21131. Umemoto, A., Scolari, M., Vogel, E. K., & Awh, E. (2010). Statistical learning induces discrete shifts in the allocation of working memory resources. Journal of Experimental Psychology: Human perception and Performance, 36(6), 1419–1429. http://dx.doi.org/10.1037/a0019324. Vuilleumier, P., & Vocat, R. (2011). In P. Azouvi, Y. Martin, & G. Rode (Eds.), Anosognosie de l’hémiplégie après lésions de l’hémisphère droit: Aspects cognitifs, affectifs et neuroanatomiques (pp. 45–67). Marseille: Solal. Wansard, M., Bartolomeo, P., Bastin, C., Segovia, F., Gillet, S., Duret, C., & Meulemans, T. (2015). Support for distinct subcomponents of spatial working memory: A double dissociation between spatial – Simultaneous and spatial – Sequential performance in unilateral neglect. Cognitive Neuropsychology, 32(1), 14–28. http://dx.doi.org/10.1080/02643294.2014.995075. Wansard, M., Bartolomeo, P., Vanderaspoilden, V., Geurten, M., & Meulemans, T. (2015). Can the exploration of left space be induced implicitly in unilateral neglect? Consciousness and Cognition, 31, 115–123. http://dx.doi.org/10.1016/j.concog.2014.11.004. Wansard, M., Meulemans, T., Gillet, S., Segovia, F., Bastin, C., Toba, M. N., & Bartolomeo, P. (2014). Visual neglect: Is there a relationship between impaired spatial working memory and re-cancellation? Experimental Brain Research, 232(10), 3333–3343. Zhao, J., Al-Aidroos, N., & Turk-Browne, N. B. (2013). Attention is spontaneously biased toward regularities. Psychological Science, 24(5), 667–677. http://dx. doi.org/10.1177/0956797612460407.

Consciousness and Cognition 19 (2010) 751–761 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Descartes discarded? Introspective self-awareness and the problems of transparency and compositionality q Markus Werning Ruhr University Bochum, Department of Philosophy, Universitätsstr. 150, 44801 Bochum, Germany a r t i c l e i n f o Article history: Available online 16 August 2010 Keywords: Introspection Self-awareness Inner self Semantic compositionality Phenomenal transparency Mental quotation Extrospection Inner speech Phonology Descartes a b s t r a c t What has the self to be like such that introspective awareness of it is possible? The paper asks if Descartes’s idea of an inner self can be upheld and discusses this issue by invoking two principles: the phenomenal transparency of experience and the semantic compositionality of conceptual content. It is assumed that self-awareness is a second-order state either in the domain of experience or in the domain of thought. In the former case self-awareness turns out empty if experience is transparent. In the latter, it can best be conceived of as a form of mental quotation. Various proposed analyses of direct and indirect quotation are discussed and tested regarding their applicability to thought. It is concluded that, on the assumption of compositionality, the inner self is only insofar accessible to awareness as it has an accessible phonological (or otherwise subsymbolic) structure, as apparently only inner speech does. Ó 2010 Elsevier Inc. All rights reserved. 1. Introduction ‘‘I am aware that I exist; I ask who is that I of which I am aware”.1 When René Descartes poses this question right after his seminal cogito argument of the Second Meditation, he raises the issue of self-awareness. He does not question the existence of such a self. The existence of the self, rather, is asserted in advance. His question is also not one about the epistemic status of the process of self-awareness: Is it knowledge that we have or not? The Latin original indicates that Descartes, indeed, takes it to be knowledge.2 No, what Descartes asks is: What has the self to be like such that reflexive awareness of it is possible? It is thus presupposed that the essential property of the self is an epistemic property, viz. being the object of reflexive awareness. Descartes’s own answer is twofold. The negative part consists in the claim that the object of self-awareness in the intended sense is not the body. The positive part is rather indirect: ‘‘that which is doubting, understanding, affirming, denying, q This article is part of a special issue of this journal on Self, Other and Memory. E-mail address: markus.werning@rub.de 1 Latin original: ‘‘Novi me existere; quaero, quis sim ego ille, quem novi” (Descartes, 1641/1904, p. 27) – A frequently noted worry roots in a grammatical peculiarity that concerns the unusual application of a demonstrative – ille ‘that’ – to a personal pronoun – ego ‘I’. Some philosophers have blamed Descartes for using ungrammatical language and even accused him of ‘‘a linguistic derailment with terrible consequences” (Beckermann, 2009, p.7). Demonstratives are akin to definite articles in that they are typically applied to generic terms as in that/the man. The Latin ego as the English I is an indexical term – whose referent varies with the speaker and the situational context – and normally neither allows for a demonstrative nor for a definite or indefinite article. The application of a demonstrative to the first person singular pronoun coerces a grammatical type shift and treats the pronoun as a generic term. It so, in a question-begging way, hypostatizes a class of entities: the Is – and consequently my I, your I, her I, etc. To excuse Descartes, one might point out that natural language is very flexible and provides many examples for forced shifts of grammatical type. There, e.g., do exist accepted phrases like the here and now where indexicals occur with a definite article. Grammar alone does not suffice to discard Descartes. 2 Novi is the first person present perfect of noscere which means ‘get to know’. 1053-8100/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2010.07.003 752 M. Werning / Consciousness and Cognition 19 (2010) 751–761 willing, refusing, imagining, and experiencing.”3 Descartes, so we may justly read, identifies the object of self-awareness with the subject of mental states and he thereby individuates the self as the unity that carries mental contents. This would make selfawareness a case of introspective self-awareness or awareness of an inner self. The idea of an inner self contrasts with the idea of a bodily self. The latter can be regarded as the object of perceptual or doxastic attitudes turned towards one’s own body. These make up the so-called body image. The bodily self can also be regarded as given by a body schema: a collection of motor programs or habits that enable and constrain movement and the maintenance of posture (Gallagher, 2005). It is yet an open question whether and to which extent body image and body schema are intertwined. Experiments known as the rubber hand illusion and the out-of-body illusion induce an illusionary body schema – either a mislocation or a partial or global misidentification of one’s own body – on the basis of visual and tactile stimulation alone (Botvinick & Cohen, 1998; Lenggenhager, Tadi, Metzinger, & Blanke, 2007). This indicates that body image and body schema may in fact be more closely related to each other than the conceptual differentiation suggests. An essentially diachronic view to be distinguished from the notions of the inner and the bodily self is the proposal of a narrative self. It has been suggested to conceive of it is as an abstraction from an autobiographical narrative process that brings together hitherto only minimally coherent pieces of behavior ‘‘enhanced by an illusion of greater unity” (Dennett, 1992, see also Ricoeur, 1992). The present paper takes up Descartes’s original question regarding the nature of the object of introspective self-awareness and asks whether Descartes’s idea of an inner self as a unitary carrier of mental contents can be upheld or whether it has to be discarded. In the latter case the self would very likely reduce to nothing but the bodily self or would have to be replaced by the quasi fictional idea of a narrative self. I would like to stress that metaphysical questions regarding the legacy of a res cogitans as a substance ontologically independent from the physical world are important in their own right, but will not be of concern to us here. The idea that a carrier of mental contents be the object of self-awareness does per se not contradict physicalism, as the various contemporary naturalistic accounts of mental content seem to show (see Greenberg, 2005, for review). Instead, we will discuss Descartes’s legacy by invoking two principles that have been moving ever closer into the center of the debate in the philosophy of mind over the past two decades: the principle of the phenomenal transparency of experience and the principle of the semantic compositionality of conceptual content. Experience is said to be phenomenally transparent just in case having an experience of certain objects, events or situations is for the subject just as if the objects, events, or situations were present (see the following section for elaboration). By saying that a conceptual representation is semantically compositional we mean that the content of a complex concept is a structure-dependent function solely of the contents of the parts of the concept (see below). The two principles cover two aspects – the conceptual and the phenomenal – that are somewhat heterogeneously distributed over the class of mental states. There are mental states like beliefs, desires, recollections, or expectations that are often regarded as conceptual – these will be called thoughts in the course of the paper. There are mental states like perceptions, hallucinations, and proprioceptions that are widely regarded as phenomenal – they will subsequently be discussed under the label of experience. There may also be mental states that have both a conceptual and a phenomenal aspect: many emotions are eventual candidates here. One may even hold that seemingly phenomenal states after some scrutiny reduce to conceptual states. Even though the transparency of experience and the compositionality of conceptual content regard two different aspects in the class of mental states, they, from a more abstract point of view, are somewhat akin because they both deny any particular role that access to the intrinsic properties of the vehicles of mental content might have in the determination of content. As we will see later on, this is the main reason why the idea of introspective self-awareness is so difficult to accommodate for if the two principles hold. Given the main mutually non-exclusive, but probably exhaustive dichotomy in the class of mental states – experiences with phenomenal qualities on the one hand and thoughts with a conceptual structure on the other hand – there are altogether three ways to account for introspective self-awareness: (i) It can be construed as a phenomenon completely in the domain of experience, such that the introspective and the introspected states are experiences. For that case will we show that introspective self-awareness is empty if experience is phenomenally transparent. (ii) Introspective self-awareness can be analyzed fully in the domain of thought. In that case we will argue that introspective self-awareness is best conceived of as a mental analogue of direct quotation. Provided that thought is compositional, we will then show that the inner self must have a cognitively accessible subsymbolic structure which is very likely akin to the phonological structure of natural language. Thus mental states would have to be linguistically structured thoughts in order to be introspected. The inner self would reduce to a stream of inner speech. (iii) Introspective self-awareness is a hybrid phenomenon crossing the domains of thought and experience: it would thus either be a non-conceptual experience directed towards a non-phenomenal thought or a non-phenomenal thought representing a non-conceptual experience. Since this is a rather hypothetical option and to my knowledge nobody has ever tried to analyze introspective self-awareness this way, I will spare a discussion of this option due to space limits. The paper concludes with the conditional claim that if experience is phenomenally transparent and thought is semantically compositional, then introspective self-awareness is either empty or reduces to a representation of inner speech. 3 Latin original: ‘‘Nempe dubitans, intelligens, affirmans, negans, volens, nolens, imaginans quoque, et sentiens‘‘ (Descartes, 1641/1904, p. 28). M. Werning / Consciousness and Cognition 19 (2010) 751–761 753 Although I will focus more on an explication than on a detailed justification of the conjunctive antecedent of the conditional in this paper, I expect that only few philosophers will be ready to reject the if-clause of the conditional tout court. Whereas compositionality is almost unanimously accepted as a constraint on thought, the situation with regard to the phenomenal transparency of experience, I concede, is more complex. Here the main question is whether perception and eventually other experiential states are regarded as having conceptual or non-conceptual content. If experiences were construed as conceptual (see, e.g., McDowell, 1996), they would count as instances of thought in the sense of this paper. The compositionality conjunct of the antecedent would apply and the truth of the conditional would make its consequent unavoidable even if one were to deny the phenomenal transparency of experience. Only for non-conceptualist views of experience the transparency conjunct is essential to establish the consequent. Note, however, that most non-conceptualists, prominently Dretske (1981/2003b,2003a), do endorse the transparency thesis or related claims (e.g., direct realism). So even if the main claim of this paper is limited to a conditional, the foregoing remarks illustrate that logical space leaves remarkably little room to avoid the antecedent of the conditional and hence its consequent. 2. Phenomenal transparency This issue of phenomenal transparency was first advanced by G. E. Moore. In his Refutation of Idealism he writes: [...] that which makes the sensation of blue a mental fact seems to escape us; it seems, if I may use a metaphor, to be transparent – we look through it and see nothing but the blue; we may be convinced that there is something, but what it is no philosopher, I think, has yet clearly recognized. (Moore 1903, p. 446) In a slightly more poetic way Martin Heidegger formulated the same idea independently and made it a cornerstone of his philosophy: We never originally [...] perceive a throng of sensations, e.g., tones and noises, in the appearance of things [...]; rather we hear the storm whistling in the chimney, we hear the three-engine aeroplane, we hear the Mercedes in immediate distinction from the Volkswagen. Much closer to us than any sensations are the things in themselves. (Heidegger, 1935/1977, p. 156) Gilbert Harman re-introduced the notion of phenomenal transparency into modern analytic philosophy: When you see a tree, you do not experience any features as intrinsic features of your experience. Look at a tree and try to turn your attention to intrinsic features of your visual experience. I predict you will find that the only features there to turn your attention to will be features of the presented tree, [...] (Harman, 1990, p. 39). What the three authors point to is the fact that, when we experience something, we do not experience the intrinsic properties of the vehicles of the experience – what Moore and Heidegger call ‘‘sensations”.4 To give a positive account of phenomenal transparency – what it is for the subject to have such an experience rather than what it is not – and to avoid metaphorical language, I propose a definition that uses a counterfactual as-if construction: Definition phenomenal transparency A subject’s S experience of a scenario X is called phenomenally transparent just in case S’s having an experience of X is for S just as if X were present. As the scenario somebody experiences we understand the totality of the objects and events with their properties and relations that make up the content of the experience at a time. The as-if analysis of transparency exhibits a number of features that allude to aspects prominent in traditional philosophical characterizations of experience and are worth closer investigation. First, the definition presupposes a relation of ownership between the experience of the scenario and the experiencing subject. Moreover, it also covers the aspect of subjectivity in that it claims that, for the owner of the experience, having the experience is as if the scenario were present. For a third person, who studies the experience, say, neurophysiologically, the experience may well be different. Second, our definition contrasts with the position of direct realism which would license the inference from S experiences X to X is present. This inference is blocked because the as-if clause is counterfactual. So even if the scenario is not present, having the experience might well be for the subject as if the scenario were present. ‘‘Experience” in our context is not a success word and the definition is not restricted to veridical experience but may as well hold for hallucination. However, what is sometimes called the ‘‘feeling of realness” (Brown, 2004) is intended to be entailed. This feeling of realness is typical for perceptions and proprioceptions, but also comes with many hallucinations and illusions. The notion of presence is even stronger 4 Metzinger formulates what he calls the ‘‘standard definition” of phenomenal transparency as follows: ‘‘Phenomenal states are transparent in that only their content-properties are introspectively accessible to the subject of experience” (Metzinger, 2003, p. 354). In my eyes it is misleading to presuppose that the content-properties are accessible introspectively. For, it is not introspection, but rather perception or – as in the case of bodily states of one’s own – proprioception that makes the content-properties available to the subject. However, since the idea of phenomenal transparency denies that there are accessible intrinsic properties of the vehicles of experiences, the whole idea of introspection becomes at least questionable and we should not define phenomenal transparency in terms of introspective accessibility in the first place. 754 M. Werning / Consciousness and Cognition 19 (2010) 751–761 than that of realness since it does not only entail that the scenario is real for the subject, but also that there is some spatiotemporal immediacy of the scenario for the subject. Finally, the definition contains an important minimality constraint. The use of the word just implies that in the counterfactual condition nothing is present for the subject, but the objects and events with their properties and relations of the scenario that make up the content of the experience. As we will see, the minimality constraint is crucial for the problem phenomenal transparency creates for purely experiential accounts of introspective self-awareness. 3. The experiential account of introspective self-awareness One way to think of introspection is to view it as a phenomenon fully in the domain of experience. Introspection might be regarded as an experience of the experiences one has at a certain time. The idea is that a person might, e.g., see a tree and then in addition experience that she sees a tree. She might have a toothache or feel a sore tooth and then experience that she has a toothache or feels a sore tooth. Introspection according to this view would be a second-order experience. However, would introspection construed as a second-order experience lead to self-awareness? Would it license the inference to a self? The answer very likely is ‘‘no” if experience is phenomenally transparent. To see this, let’s first turn to proprioceptions like pain. There are two ways of thinking about pain (and most other proprioceptions). According to the first conception, pain does not necessarily involve consciousness, but simply is a state of the body that one occasionally experiences when one feels pain. Under this view feeling pain is a first-order experience and does not at all qualify as an introspective state. According to the alternative conception, pain is conceived of as a state that necessarily involves consciousness. Here, pain is itself a state of experience rather than a content of experience. Having a toothache, according to this view, is the experience of a sore tooth (of one’s own). Turning to the second-order experience we now face a dilemma: Experiencing the presence of a sore tooth and experiencing a toothache – i.e., having the second-order experience – are for the subject either the same or not the same. If they are the same for the subject, there is no reason for the subject to infer the existence of anything over and above the tooth and its soreness on the basis of her second-order experience just as experiencing the presence of a sore tooth would not license any such inferences. Any inference to an internal vehicle of the proprioception as part of an inner self would be unmotivated by the second-order experience. If experiencing the presence of a sore tooth and experiencing a toothache, on the other hand, are not the same for the subject, then phenomenal transparency is violated. For, if the experience of a toothache were phenomenally transparent it should, by definition, be for the subject just as if a toothache were present. Now again, the presence of a toothache for the subject under the second conception of pain should be just the same as experiencing the presence of a sore tooth. Unfortunately, the states of proprioception do not by introspection contribute to an awareness of the self if phenomenal transparency holds. The same is true for perception: Assume that the perception of a tree be itself an object of introspective experience. The phenomenal transparency of experience would entail that being in the perceptive state and the introspective state at the same time would be for the subject just as if a tree and the perception of a tree were present. After another application of the transparency property we see that the latter – the perception of a tree – is for a subject just as if a tree were present. It follows that being in the combined – perceptive plus introspective – state is for the subject just as if a tree and the same tree were present. It hence is just like being in the state of perception without being in the state of introspection. Introspection adds nothing to perception and gives no cues to an inner self if experience in general is phenomenally transparent.5 4. Second-order thoughts Introspective self-awareness might also be analyzed in terms of conceptually structured thoughts rather than as a form of phenomenal experience. The property of phenomenal transparency would not apply. In those terms an introspective state would be a second-order thought of sorts that reports another thought of one’s own. What we are looking for seems closely analogous to – yes, indeed, structurally the same as – quotation in natural language. A quotation in natural language is an utterance that reports another utterance. One generally distinguishes two kinds of quotation: direct quotation and indirect quotation. The reason is that an utterance can either be reported by reference to its vehicle – the actual expression uttered – or by reference to its content or meaning – what was said by the utterance. Compare the following: Donald said; ‘In Mark’s garden furze is more widespread than iv y’: ð1Þ Donald said that in Mark’s garden furze is more widespread than iv y: ð2Þ The truth-conditions of the two sentences differ. The words that were actually used in the reported utterance make a difference regarding direct quotation whereas they do not have an effect on the truth of the indirect quotation if content is left unchanged. This holds for all kinds of verbal variations that would leave content unchanged: (a) Choice of language: Had Donald said what he said in German, (1) would no longer be an accurate report of the utterance, while (2) would remain true. (b) Substitution of synonyms: For the truth of (2) it does not matter whether Donald used the word gorse or furze whereas it does matter for (1). (c) Substitution of co-referential rigid designators: Assume that Mark and Samuel be proper 5 The argument of this section largely builds on Dretske (2003). For a closer discussion see Werning (2004a). M. Werning / Consciousness and Cognition 19 (2010) 751–761 755 names that refer to the same person – as would be the case for Mark Twain and Samuel Clemens. Then the truth of (2) would remain unaffected even if Donald actually used the word Samuel to refer to the garden in question.6 (d) Analytical or logical equivalence: The utterance In Mark’s garden, ivy is less widespread than furze would be truly reported by (2), but not by (1). Furthermore, in specific contexts – where e.g., conversational focus, metrics, dialect, or speech impediments are at issue – the accuracy standards for direct quotation may be raised so that even word order, prosody, phonology, or phonetics play a role. All those aspects are irrelevant for indirect quotation because here only the content, not the vehicle of the utterance is at issue. There are numerous other aspects in which direct and indirect quotation differ – the role of indexicals, anaphoric reference, tense, etc. – but these issues will not be dealt with here. When we now turn back to self-reporting second-order thoughts, we have a choice between conceiving of them as reports of the internal vehicles of other thoughts of the subject – i.e., as direct mental quotations – or as reports only of the contents of those thoughts – as indirect mental quotations. What would those choices imply for the possibility of self-awareness? For second-order thoughts of the latter kind Fernández (2003, 2005) has coined the term extrospection. He argues that any piece of evidence – perception, memory, testimony – that justifies a subject’s belief that p also justifies the subject’s belief that she believes that p if the subject has formed the belief that p and the belief that she believes that p on the basis of this evidence. His argument presupposes a reliabilist notion of justification and identifies the justifiedness of a belief with being the product of a belief formation process that is grounded in some reliable regularity. The argument can roughly be summarized as follows: If a properly functioning epistemic subject has good evidence for the truth of the proposition p, she normally comes to believe that p. Due to this mind-to-mind regularity, a subject that forms the belief that p on the basis of good evidence for the proposition p is also justified, on the basis of this evidence, to believe that she believes that p. Access to the internal vehicle of the first-order thought is thus not required to justify the second-order thought. One only has to have access – this is the role of good evidence – to the content of the first-order thought, the proposition p, to justify the secondorder thought. Thus self-knowledge would be sufficiently grounded in extrospection. I am in general very sympathetic with this account of self-knowledge. The question in our context however is whether extrospective self-knowledge qualifies as knowledge of an inner self. Only thus Descartes’s original idea could be defended. I take it that this was not even in the intentions of Fernández’s. Just as indirect linguistic quotation gives us only few hints as to which linguistic expressions – which language, which words, and which syntactic structures – were actually used by the reported speaker, extrospection, viewed as indirect mental quotation, provides us with almost no clue with respect to the nature, the constituents, and the structure of the reported thought. All it provides us with is the external content of the latter. The inner self remains in the dark. Self-reporting second-order thoughts might also be construed in analogy to direct quotation. Here we may justly speak of introspective thoughts because thoughts of one’s own are reported by reference to their internal vehicles. The challenge here is to find a proper way to analyze direct quotation and apply this analysis to thought. In this regard it is of particular significance that thought – as is widely assumed – is compositional. The principle of semantic compositionality is defined in most general terms for representational structures like language or thought that consist of (i) a set of representational vehicles and (ii) a set of combinational operations mapping arrays of constituent vehicles onto a complex vehicle, and for which (iii) a function of semantic evaluation is defined on the set of vehicles: Definition semantic compositionality A representational structure (i.e., language or thought) is called semantically compositional just in case the semantic value (meaning, or respectively, content) of any complex vehicle (grammatical term or, respectively, mental concept) is a function of the semantic values of its constituent vehicles where the choice of the function depends solely on the operation by which the constituents combine to form the complex.7 As Fodor (2001) has pointed out, compositionality is much less controversial for concepts (including truth-evaluable thoughts) with respect to their contents than it is for linguistic expressions with regard to their meanings. Even though many a priori justifications for the compositionality of meaning in natural language have been put forward more or less convincingly – learnability, productivity, systematicity, communicative efficiency (for a critical discussion of those arguments see Werning, 2005; Pagin & Westerståhl, in press) – it is not keen to say that a final verdict has not yet been spoken. In particular many tenacious linguistic counterexamples have yet to be defused. In the case of thought the case seems clearer: What could matter for the contents of the constituent concepts to determine the content of the complex concept other than their combinational structure? In natural language extralinguistic factors that lie neither in syntactic structure nor in the meanings of the constituent expressions may eventually matter to determine the meaning of a complex expression. Think of ostensive gestures that help to determine the reference of demonstratives (see Kaplan, 1989). Or imagine background situations that play a role for the determination of truth-conditions. In sentences like All guests have gone uttered at a party or Some passengers have died said after a train accident, the situation obviously restricts the domain of the quantifier – if quantifiers ranged over an unrestricted domain, sentences of this kind would either be trivially false or trivially true. In those cases one is typically required to allow for explicit or implicit 6 It is widely, though not unanimously accepted that co-referential rigid designators may be replaced in indirect quotations without changing their truthconditions. This view grounds in the possible-worlds semantics of propositional attitudes. 7 For formalizations see Hodges (2001) and Werning (2004b). 756 M. Werning / Consciousness and Cognition 19 (2010) 751–761 contextual arguments in order to force the semantic analysis of a sentence into a compositional framework. This strategy only works because one may assume that both the speaker and the hearer of such a sentence do represent the relevant factors mentally – typically on the basis of perception or memory – even if the factors are not contained in the learned meanings of the constituent expressions themselves. In the domain of thought this strategy is not viable because contextual factors may matter to determine the content of a complex concept only if they are mentally represented. They are thus not extra-representational, but belong to a larger conceptual structure internal to the mind of the subject. To thought, compositionality applies without exception. If we are hence to construe introspective thoughts in analogy to direct quotations, we have to make sure that the analysis of direct quotation used abides by the principle of compositionality. This is an important point because many accounts of direct quotation in natural language take it virtually as an axiom that quotation be non-compositional. Cappelen and Lepore (2009), e.g., identify the first ‘‘basic quotational [feature] of particular importance” with the principle that ‘‘in quotation you cannot substitute co-referential or synonymous terms salva veritate.” If you cannot substitute synonymous terms without a change in the truth-conditions of a sentence, you cannot substitute synonymous terms in a sentence without a change in meaning. Now, the latter, the substitutability of synonyms salva significatione is logically equivalent to compositionality (as proven by Hodges, 2001). Cappelen and Lepore’s view of quotation hence logically entails that direct quotation be non-compositional. The widespread conviction that direct quotation should be treated in a non-compositional way – quasi as an exception to the principle of compositionality – is probably the reason why so few explicitly compositional treatments of direct quotation can be found in the literature. To my knowledge only three papers, Potts (1997), Pagin and Westerståhl (in press), and Werning (2005), have ever tried so. Due to space and time limits, I will not discuss the first two. A closer investigation, however, would reveal that the analyses Potts and Pagin and Westerståhl provide for direct quotation in natural language cannot be transferred to the domain of thought because they are either not properly compositional or they essentially make use of extralinguistic contextual arguments. 5. Direct quotation and compositionality To illustrate what kind of questions the compositionality requirement raises for the treatment of direct quotation, I will first turn to the most natural and still very widespread, so-called disquotational analysis of quotation. To make this analysis explicit, we first need a rudimentary formal framework. According to the standard view, syntactic operations are not operative on the set E of expressions of the language directly, but on a set T of underlying grammatical terms. The distinction between expressions and terms is useful to deal with lexically or syntactically ambiguous expressions such as bank or The boy watches the girl with the telescope. By convention, expressions are set in italics. An expressing function e with e:T!E ð3Þ maps the set of terms onto the set of expressions and a meaning function l with l:T!M ð4Þ maps the set of terms onto the set of meanings. The grammar of a language is a pair hT, Si where S is the set of syntactic operations. Each syntactic operation is a partial function on some Cartesian product of T into T. As a useful convention that allows us to reduce notational complexity we set primitive or complex expressions in indexed angel brackets in order to disambiguate them and to denote the underlying grammatical terms, e.g., hbanki1 and hbanki2. If an expression is unambiguous the index, by convention, is left away. In the disquotational analysis a syntactic operation of quotation q is adopted into the grammar. It is a total function defined as follows: q : T ! T such that eðqðsÞÞ ¼ ‘eðsÞ’ and lðqðsÞÞ ¼ eðsÞ: ð5Þ The grammatical term s that underlies the quoted expression e(s) is the argument of the syntactic operation of quotation q. The expression of a quotation, i.e., e(q(s)), is the concatenation of the onset quotation mark, the quoted expression, and the offset quotation mark: ‘e(s)’ – as is common, the operation of concatenation is left implicit to simplify notation. The meaning of a quotation is the quoted expression. This account of quotation might be called holophrastic because it takes the quoted expression as a whole and sets it in quotation marks. It can be shown that the disquotational analysis violates compositionality, provided the language to be considered contains synonymous expressions, e.g., the pair gorse and furze: lðhgorseiÞ ¼ lðhfurzeiÞ ð6Þ Although the two expressions are synonymous, they are not identical: gorse – furze: ð7Þ From our definition of q we derive the following: lðqðhgorseiÞÞ ¼ gorse: ð8Þ lðqðhfurzeiÞ ¼ furze: ð9Þ M. Werning / Consciousness and Cognition 19 (2010) 751–761 757 From (7), (8), and (9) we may infer: lðqðhgorseiÞÞ – lðqðhfurzeiÞ: ð10Þ If we, for the sake of the argument, assume compositionality, the meaning of a syntactically complex term q(s) should be some function fq of the meanings of its syntactic part s: lðqðhgorseiÞÞ ¼ fq ðlðhgorseiÞÞ: ð11Þ Substitution of identicals according to (6) yields: lðqðhgorseiÞÞ ¼ fq ðlðhfurzeiÞÞ: ð12Þ Due to compositionality it also holds that: lðqðhfurzeiÞÞ ¼ fq ðlðhfurzeiÞÞ ð13Þ From (12) and (13), we hence get: lðqðhgorseiÞÞ ¼ lðqðhfurzeiÞÞ: ð14Þ We have derived a contradiction with (10). The hypothetical assumption that the language be compositional must consequently be rejected. Similar arguments can be made for all holophrastic analyses of quotation provided they treat the contents of expressions in a context-invariant unambiguous way. To see this, one may adapt the argument to Davidson’s (1984) demonstrative analysis of direct quotation, which is another case of holophrastic quotation. As in any such analysis, quotation is taken as a syntactic operation q: T ? T such that e(q(s)) = ‘e(s)’. In the demonstrative analysis, for any sentential term u(n) in which a quotation may occur at position n, the meaning is given as the meaning of a discourse with backward demonstrative reference:8 lðuðqðsÞjnÞÞ ¼ lðs uðhthatijnÞÞ: ð15Þ Example: The meaning of John said; ‘Furze is beautiful’ ð16Þ is given as the meaning of the two-sentence discourse Furze is beautiful: John said that: ð17Þ If we were licensed to assume compositionality, we should be allowed to replace synonyms in the discourse without changing its meaning. (17) should hence be synonymous to Gorse is beautiful: John said that: ð18Þ However, this discourse according to (15) would have the same meaning as John said; ‘Gorse is beautiful’: ð19Þ We have again derived a contradiction since (16) and (19) obviously have different truth-conditions. Compositionality must not be presupposed in the demonstrative analysis either. 6. Phonological quotation In two representative examples we have shown that any holophrastic treatment of direct quotation leads to a violation of the principle of compositionality unless one is ready to allow for extralinguistic context arguments.9 Permitting the latter would, however, block any transfer of the analysis from language to thought. Any analysis of direct quotation applicable to the domain of thought and thus possibly accounting for introspective self-awareness, must hence be non-holophrastic. It must refer to features of the quoted expression that are below the level of the whole expression. Such an analysis is justly called subsymbolic. To show that quotation in natural language can be analyzed in a compositional way without extralinguistic context arguments, I will here introduce the method of phonological quotation as an example of a subsymbolic analysis of direct quotation. The idea partly builds on Geach’s (1970) descriptive analysis of quotation. What’s new in the phonological analysis is that the means of reference to expressions is description by imitation. An expression is described by imitating its phonological structure. The phonological analysis is on par with alternative subsymbolic analyses that refer to expressions by imitating and describing graphical features of inscription or gestures of sign languages. 8 The substitution operator | in a term u(s|n) reads: replace the variable n – i.e. a placeholder that indicates a certain position – in the term u(n) with the term s. The syntactic full-stop operator . builds discourses from its left and right argument terms. 9 The formal proof is immediate and left to the reader. 758 M. Werning / Consciousness and Cognition 19 (2010) 751–761 Let us assume that each expression of our language be a temporal sequence of primitive phonological parts, e.g., phonemes, from a finite set. The expression dog, e.g., is a sequence of the plosive voiced dental consonant d, the closed-mid back vowel o, and the plosive voiced velar consonant g. To enable reference to the expressions of the language within the language, the set of terms T of the language has to include as a subset a set of atomic terms P that denote the elementary phonological parts necessary to generate each and every expression of the language. Neglecting the details of phonology, a crude approximation might be the following (capital letters – except the labels of sets – are used to refer to object language terms that are names of phonemes, whereas lower-case characters in our meta-language are used to refer to phonemes when they are not used as variables. The space bar _ denotes a distinctive length of silence): P ¼ fA; B; . . . ; Z; g ð20Þ lðAÞ ¼ a; lðBÞ ¼ b; . . . ; lðZÞ ¼ z; lð Þ ¼ x milliseconds of silence: ð21Þ The phonological analysis of quotation to be developed here refers to phonemes and sequences of phonemes by imitation. Formally, the idea of imitation is captured by the fact that the phonemic terms are themselves expressed by the phonemes they denote: eðAÞ ¼ a; eðBÞ ¼ b; . . . ; eðZÞ ¼ z; eð Þ ¼ x milliseconds of silence: ð22Þ To achieve reference to complex phonological sequences, a binary syntactic operation of concatenation is adopted into the grammar. The set of phonological descriptions is the closure P* of the set P with respect to concatenation.10 The syntactic operation of concatenation has a unique semantic counterpart operation: the function fs that maps pairs of phonological sequences to the sequence of them. Compositionality is warranted by the following homomorphism, with the variables s and t in P*: _ lðs_ tÞ ¼ fs ðlðsÞ; lðtÞÞ: ð23Þ To enable description by imitation also for complex expressions, the expressing function is to satisfy a similar constraint: eðs_ tÞ ¼ fs ðeðsÞ; eðtÞÞ: ð24Þ Notice that all expressions of the language can be described by phonological descriptions, but usually not all concatenations do actually describe some word or some sequence of words of the language: The concatenation O_ G denotes the phonological sequences og, which is not a meaningful English word, while D_ ðO_ GÞ denotes the English word dog. The meaning of the quotation is derived compositionally: lðD_ ðO_ GÞÞ ¼ fs ðlðDÞ; lðO_ GÞÞ ¼ fs ðD; fs ðlðOÞ; lðGÞÞÞ ¼ dog: ð25Þ Likewise the expression of the quotation is generated stepwise: eðD_ ðO_ GÞÞ ¼ fs ðeðDÞ; eðO_ GÞÞ ¼ fs ðD; fs ðeðOÞ; eðGÞÞÞ ¼ dog: ð26Þ Quotation is hence expressed by an imitation of the phonological structure of the quoted expression. Be aware, however, that the expression of the quotation and the quoted expression are given different syntactic analyses, even though they exhibit the same phonological structure. In (25) the quoted expression dog is regarded as syntactically unstructured. It is the expression of the primitive, lexical term hdogi1: dog ¼ eðhdogi1 Þ: ð27Þ _ _ In (26) dog is regarded as the expression of the quotation D (O G): dog ¼ eðD_ ðO_ GÞÞ: ð28Þ Keep in mind that in spoken language there is no phonological expression of the quotation marks we use in writing. The written expressions ‘dog’ and dog correspond to one and the same phonological sequence, and hence to one and the same expression in spoken language (however, see option three in the passage on higher-order quotation below). Notice that phonology is all there is to the identity of expressions in spoken language. The phonological analysis of quotation is here developed for spoken language and, thus, systematically introduces ambiguities on the level of expressions. These ambiguities, however, are syntactically resolved and do not pose any threat to a strict application of the principle of compositionality. For, compositionality regards the relation between terms and meanings, rather than the one between expressions and meanings. We may sum up that unlike holophrastic analyses, the phonological treatment provides a compositional analysis of direct quotation and avoids context arguments and syntactically unresolved ambiguities. I am aware that the phonological analysis is not a silver bullet against all sorts of difficulties quotation brings about in natural language. As with other analyses of quotation the treatment of mixed quotation is problematic (see Reimer, 10 To guarantee that terms are uniquely identified by a nested structure of syntactic operations, we adopt the convention that phonological sequences are described from the end. The innermost concatenation refers to the sequence of the final two phonemes of an expression. Thus D_ (O_ G) is a well formed term while ðD_ OÞ_ G is not. M. Werning / Consciousness and Cognition 19 (2010) 751–761 759 2003). Here the quoted expression is used and mentioned at the same time (as in: Quine said that quotation ‘has a certain anomalous feature’). However, many other problems often brought forward against the phonological theory are less severe.11 Aside from being compositional, the phonological analysis has considerable further advantages over holophrastic analyses. Whereas holophrastic analyses only allow for the quotation of expressions with an existing underlying term, the phonological analysis immediately explains why non-words can be quoted, as in: Hans always says ‘shnoff ’ when he speaks of snow: ð29Þ Holophrastic analyses furthermore have problems with explaining the quotation of syntactically ambiguous expressions. For, the syntactic operation of quotation takes as arguments only syntactically disambiguated terms. The phonological analysis has no problems here because it just recurs to the phonological surface of an expression. In a holophrastic framework there is, e.g., no syntactic analysis of the following sentence that renders the sentence true: ‘The boy watches the girl with the telescope’ is ambiguous: ð30Þ Of particular interest is the question whether phonological quotation as a way of description by imitation is consistent with the assumption that quotations can be recursively embedded, as in: \dog" refers to ‘dog’; which refers to dogs: ð31Þ At first glance there seems to be no difference between the imitation of the imitation of the phonological structure of an expression and the simple imitation of the phonological structure of the expression. Both sound the same and, more importantly, have the same underlying syntactic structure. One option is to bite the bullet and argue that higher order quotation is in fact a notorious problem in spoken language. Only auxiliary conventions like reading the quotation marks aloud – quote dog unquote vs. quote quote dog unquote unquote – might be a means to circumvent the problem. A second option is to appeal to speech act theory and argue that, even though the phonological and syntactic structure of the expressions of a second and a first order quotation are the same, the illocutionary force and consequently the intentions of speaker and hearer are different. The quotation marks in writing could then be interpreted as signs of an imitative illocutionary act. Their role would thus be somewhat analogous to punctuation marks like the question mark or the exclamation mark. In many languages question marks and exclamation marks can, e.g., be placed after sentences that are otherwise indistinguishable from assertions. In those cases the punctuation marks indicate nothing but the change of illocutionary force. A third, rather technical option could be the following: We stipulate as a rule that each imitation start and end with a certain period of silence – or some other phonological marker: _dog_ would be the expression of the first order quotation D_ (O_ G); __dog__ the expression of the second order quotation _ (D_ (O_ (G_ _))); and so on. Notice that higher order quotations have to make explicit reference to the periods of silence in the expressions of lower-level quotations. This is because a full description of the lowerlevel expressions from the start to the end of the imitation has to be given. In this approach first and second order quotations now differ phonologically and syntactically. The quotation marks we use in writing could be interpreted as expressions of those periods of silence or simply as analogous tools. Regardless of whether the phonological analysis is the optimal treatment of direct quotation in natural language, what matters in the context of this paper is that it is the only strictly compositional analysis and as such the only one that may be transferred to the domain of thought – on a par with other comparable subsymbolic analyses. If self-awareness is regarded as self-reporting second-order thought that achieves access to an inner self, it has to be treated in analogy to direct quotation. Introspective thought is a form of direct mental quotation. However, if phonological or subsymbolic quotation is the only viable analysis of quotation that may be applied to the domain of thought, we may infer that awareness of an inner self taken as a second-order thought must be a form of phonological quotation or closely analogous to it. 7. Introspective self-awareness as the awareness of inner speech The conclusion that awareness of an inner self must be akin to phonological or otherwise subsymbolic quotation – think of representing or simulating the gestures of sign languages – has far reaching consequences. The most immediate consequence is that thoughts reported by introspective thoughts must have an accessible phonological or quasi-phonological structure (e.g., a subsymbolic gestural structure). The best and as far as I can see only candidate here is inner speech: the mental simulation of phonologically (or gesturally) structured speech. Inner speech seems to be a prerequisite for the awareness of an inner self. A second consequence is that thoughts cannot be introspectively reported if their constituents fail to 11 It is true, for instance, that in the phonological analysis the relation between the quotation and the quoted expression is less close than in holophrastic analyses – an often heard objection (Cappelen & Lepore, 2009) – because the quoted expression is not a syntactic constituent of the quotation, but merely referred to. I take this more as a strength than as a weakness. A sentence like ‘Dog’ refers to dogs, e.g., can no longer be regarded as a priori. Recall that the alleged aprioricity of such sentences generates various philosophical paradoxes. Take, e.g., Putnam’s (1981) famous brain-in-a-vat argument in the reconstruction of Wright (1992), shortest version: In my language ‘dog’ refers to dogs. In the language of a brain in a vat ‘dog’ does not refer to dogs. Therefore, I am not a brain in a vat. The problem with this argument is not its validity, but the alleged aprioricity of its two premises – the first justified by disquotation, the second by the causal theory of reference. Isn’t it paradoxical that I should be able to know a priori that I am not a brain in a vat? According to the phonological analysis, the first premise would not be a priori anymore, but boils down to something like the empirical statement A word referring to dogs in my language is the sequence of the plosive voiced dental consonant, the closed-mid back vowel, and the plosive voiced velar consonant. This sentence after some phonological scrutiny may eventually even turn out false. 760 M. Werning / Consciousness and Cognition 19 (2010) 751–761 have a phonological or otherwise subsymbolic accessible surface. This implies that the inner self is accessible only insofar as it is constituted by inner speech. If one furthermore makes the assumption that epistemic accessibility is an essential property of the inner self, as Descartes apparently did, the inner self reduces to that part of our mind that simulates phonologically or otherwise subsymbolically structured speech. One may thirdly infer that some competence of natural language is necessary to have introspective self-awareness. It is needless to say that this has numerous ontogenetic and phylogenic consequences. A fourth implication is that any impairment of our capacity to simulate inner speech should lead to an impairment regarding the inner self. There is indeed ample evidence for such a correlation (see Morin, 2005, for review). Such dysfunctions might be due to a stroke in language-related cortical areas, be caused by developmental retardation, or be grounded in psychiatric pathologies.12 Before we conclude, let me briefly summarize the overall argumentation of the paper: (P1) Self-awareness is a second-order state either in the domain of experience or in the domain of thought. (P2) If self-awareness is a second-order experience, and experience is phenomenally transparent, then self-awareness fails to achieve awareness of an inner self and is as such empty. (P3) If self-awareness is a second-order thought, then it is either (a) identified with a thought that is analogous to indirect quotation or (b) with one that is analogous to direct quotation. (P4) In the case of (a) self-awareness is an instance of extrospection, fails to achieve awareness of an inner self and is as such empty. (P5) Direct quotation is either analyzed holophrastically or phonologically (or otherwise subsymbolically). (P6) Holophrastic analyses of direct quotation fail to be strictly compositional whereas the phonological (or any likewise subsymbolic) analysis of quotation is strictly compositional. (P7) Thought is strictly compositional. (P8) Experience is phenomenally transparent. (C1) Hence: self-awareness is either empty as awareness of an inner self or analogous to phonological (or otherwise subsymbolic) quotation. (C2) Corollary: the inner self is only insofar accessible to awareness as it has an accessible phonological (or otherwise subsymbolic) structure, as apparently only inner speech does. (C3) Corollary: if being the object of self-awareness is essential to the inner self, then it apparently reduces to nothing but a stream of inner speech. I am aware that the argument of the paper explores only some portion of the relevant logical space. One might, e.g., question the exhaustiveness of the dichotomy in P1. Are experiences and thoughts – both understood in a broad sense – the only major categories of mental states? What about emotions?13 Even though there is no room for a closer discussion here, I would suggest distinguishing at least two groups: We have the group of low-level emotions like hunger and pain, which qualify as proprioceptive experiences and might be regarded as phenomenally transparent (see the foregoing treatment of pain). We also have the group of high-level emotions like hope and disappointment. These most likely are predominantly composed of thought-like states that involve beliefs, desires, and expectations. Here the compositionality requirement applies. In the context of introspective self-awareness and in the light of the argument in this paper, it should be of major concern whether the two groups exhaust the class of emotions or whether there are emotions that are neither experience-like nor thought-like. One may of course also raise more principled objections and question the compositionality of thought or the transparency of experience. It is a remarkable fact, though, that in the argument the property of phenomenal transparency does a similar job for experience as the property of compositionality does for thought. Despite potential points of criticism, each of the premises P1 to P8 seems to be rather well justified to me in its own light. This is especially important since the conclusions strike me as rather far-reaching. What about Descartes’s idea of an inner self? We have largely, however, not completely discarded this idea. Much of what Descartes suggested as belonging to the inner self can no longer be regarded as a part of it if being the object of introspective awareness is essential to the inner self. Still, there is a remainder: linguistically structured inner speech. This conclusion, of course, kicks off a whole cascade of follow-up questions: Do we have to postulate an author behind inner speech that remains unobserved in respects other than the phonological/subsymbolic structure we are introspectively aware of? Is there an issue of ownership? Does the phonological/subsymbolic structure of which one becomes aware have to be attributed to oneself in order to constitute the inner self of one’s own? May this attribution of ownership eventually even go wrong so that the speech-like structure is attributed to nobody or to somebody else? Do features internal to inner speech suffice to establish the unity of an inner self? Speech-internal features that indicate unity might be semantic such as coherence, syntactic as, e.g., anaphora and other means to interweave sentences into a discourse, and finally phonological and phonetic such as accent or voice. Similar questions arise for the uniqueness of the inner self. I am optimistic that the issues of authorship, ownership, unity, and uniqueness can in principle be dealt with in an inner speech account of the self. We may hence hope to 12 A good candidate for a correlation between inner speech deficits and an impaired inner self in psychiatry are some forms of schizophrenia as pointed out by Hoffman (1986). See Stephens and Graham (2000) for discussion. 13 For a discussion of the phenomenal transparency of emotions see Metzinger (2003). Imaginations also constitute a possible exception to transparency, for discussion see Werning (2004a). M. Werning / Consciousness and Cognition 19 (2010) 751–761 761 save at least some fragment of Descartes’s idea of an inner self without falling back into the Cartesian metaphysics of a thinking substance and related conceptions of the ‘‘I”. References Beckermann, A. (2009). Es gibt kein Ich, doch es gibt mich. In M. Fürst, W. Gombocz, & C. Hiebaum (Eds.), Gehirne und Personen–Beiträge zum 8. Internationalen Kongress der Österreichischen Gesellschaft für Philosophie in Graz (pp. 1–17). Frankfurt, Germany: Ontos. Botvinick, M., & Cohen, J. (1998). Rubber hands ‘feel’ touch that eyes see. Nature, 391, 756. Brown, J. W. (2004). The illusory and the real. Mind and Matter, 2, 37–59. Cappelen, H., & Lepore, E. (2009). Quotation. In E. N. Zalta (Ed.), The Stanford encyclopedia of philosophy. Stanford, CA. Davidson, D. (1984). Quotation. In Truth and interpretation (pp. 79–92). Oxford: Clarendon Press. Dennett, D. C. (1992). The self as a center of narrative gravity. In F. Kessel, P. Cole, & D. Johnson (Eds.), Self and consciousness: Multiple perspectives. Hillsdale, NJ: Erlbaum. Descartes, R. (1904). Meditationes de prima philosophia. In C. Adam & P. Tannery (Eds.), Oeuvres de Descartes (Vol. VII, pp. 1–90). Paris (Year of original: 1641). Dretske, F. (2003a). How do you know you are not a zombie? In B. Gertler (Ed.), Privileged access. Philosophical accounts of self-knowledge (pp. 1–14). Burlington: Ashgate Publishing. Dretske, F. (2003b). Sensation and perception. In Y. H. Gunther (Ed.), Essays on nonconceptual content. Cambridge, MA: MIT Press (Year of original: 1981). Fernández, J. (2003). Privileged access naturalized. The Philosophical Quarterly, 53, 352–372. Fernández, J. (2005). Privileged access revisited. The Philosophical Quarterly, 55, 102–105. Fodor, J. A. (2001). Language, thought and compositionality. Mind & Language, 16(1), 1–15. Gallagher, S. (2005). How the body shapes the mind. Oxford: Oxford University Press. Geach, P. (1970). Quotation and quantification. In Logic matters (pp. 205–208). Oxford: Basil Blackwell. Greenberg, M. (2005). A new map of theories of mental content. Nous, 39, 299–320. Harman, G. (1990). The intrinsic quality of experience. Philosophical Perspectives, 4, 31–51. Heidegger, M. (1977). The origin of the work of art. In Basic writings (pp. 139–211). New York: Harper and Row (Year of original: 1935). Hodges, W. (2001). Formal features of compositionality. Journal of Logic, Language and Information, 10, 7–28. Hoffman, R. E. (1986). Verbal hallucinations and language production processes in schizophrenia. Behavioral and Brain Sciences, 9, 503–517. Kaplan, D. (1989). Demonstratives. In Themes from Kaplan (pp. 481–563). Oxford: Oxford University Press. Lenggenhager, B., Tadi, T., Metzinger, T., & Blanke, O. (2007). Video ergo sum: Manipulating bodily self-consciousness. Science, 317, 1096–1099. McDowell, J. (1996). Mind and world. Cambridge, MA: Harvard University Press. Metzinger, T. (2003). Phenomenal transparency and cognitive self-reference. Phenomenology and the Cognitive Sciences, 2, 353–393. Moore, G. E. (1903). The refutation of idealism. Mind, 12, 433–453. Morin, A. (2005). Possible links between self-awareness and inner speech: Theoretical background, underlying mechanisms, and empirical evidence. Journal of Consciousness Studies, 12, 115–134. Pagin, P., & Westerståhl, D. (in press). Compositionality. In K. von Heusinger, C. Maienborn, & P. Portner (Eds.), Semantics. An international handbook of natural language meaning. Berlin: Mouton de Gruyter. Potts, C. (1997). The dimensions of quotation. In C. Barker & P. Jacobsen (Eds.), Direct compositionality (pp. 405–431). Oxford: Oxford University Press. Putnam, H. (1981). Brains in a vat. In Reason, truth, and history (pp. 1–21). Cambridge: Cambridge University Press. Reimer, M. (2003). Too counter-intuitive to believe? Pragmatic accounts of mixed quotation. Belgian Journal of Linguistics, 17, 167–186. Ricoeur, P. (1992). Oneself as another. Chicago: University of Chicago Press. Stephens, G. L., & Graham, G. (2000). When self-consciousness breaks: Alien voices and inserted thoughts. Cambridge, MA: MIT Press. Werning, M. (2004a). Self-awareness and imagination. In J. Sàágua (Ed.), The explanation of human interpretation (pp. 369–377). Lisbon: Colibri. Werning, M. (2004b). Compositionality, context, categories and the indeterminacy of translation. Erkenntnis, 60, 145–178. Werning, M. (2005). Right and wrong reasons for compositionality. In M. Werning, E. Machery, & G. Schurz (Eds.), The compositionality of meaning and content (Vol. I: Foundational Issues, (pp. 285–309). Frankfurt: Ontos Verlag. Wright, C. (1992). On Putnam’s proof that we are not brains-in-a-vat. Proceedings of the Aristotelian Society, 92, 67–94.

Consciousness and Cognition 19 (2010) 505–519 Contents lists available at ScienceDirect Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog Tactile expectations and the perception of self-touch: An investigation using the rubber hand paradigm Rebekah C. White a,*, Anne M. Aimola Davies a,b,c,d, Terri J. Halleen d, Martin Davies a,b a Department of Experimental Psychology, University of Oxford, United Kingdom Faculty of Philosophy, University of Oxford, United Kingdom c NIHR Biomedical Research Centre, University of Oxford, United Kingdom d Department of Psychology, The Australian National University, Australia b a r t i c l e i n f o Article history: Received 12 March 2009 Available online 24 February 2010 Keywords: Body illusion Expectation violation Self-touch Sensory Tactile a b s t r a c t The rubber hand paradigm is used to create the illusion of self-touch, by having the participant administer stimulation to a prosthetic hand while the Examiner, with an identical stimulus (index finger, paintbrush or stick), administers stimulation to the participant’s hand. With synchronous stimulation, participants experience the compelling illusion that they are touching their own hand. In the current study, the robustness of this illusion was assessed using incongruent stimuli. The participant used the index finger of the right hand to administer stimulation to a prosthetic hand while the Examiner used a paintbrush to administer stimulation to the participant’s left hand. The results indicate that this violation of tactile expectations does not diminish the illusion of self-touch. Participants experienced the illusion despite the use of incongruent stimuli, both when vision was precluded and when visual feedback provided clear evidence of the tactile mismatch. Ó 2009 Elsevier Inc. All rights reserved. 1. Introduction How do you know that the hand you see before you is your own? The systematic study of body ownership poses a challenge to cognitive psychologists for the simple reason that the body is always there (James, 1890), existing as ‘‘a backdrop to whatever one is thinking, experiencing, or doing, though its various parts are not being monitored” (Kinsbourne, 1995, p. 217). Over the past decade, the rubber hand paradigm (Botvinick & Cohen, 1998) has provided an invaluable tool for experimental investigations into body ownership. With this paradigm, individuals experience a sense of ownership over a prosthetic limb. Researchers have used this illusion of ownership to understand body awareness better. They manipulate factors thought to underlie self-attribution of body parts, and they assess the impact of these manipulations on the sense of ownership of the prosthetic limb. Candidate factors which may play a role in self-attribution of body parts include: visual signals, somatosensory signals, proprioceptive signals and higher-order representations of the body. In the visual rubber hand paradigm (Botvinick & Cohen, 1998), the participant views a prosthetic hand being stimulated while her own hand – hidden from view – receives synchronous stimulation. The participant may feel as if the touch on her own hand is occurring at the location where she sees the prosthetic hand being touched. The participant may also experience the illusion that the prosthetic hand is her own hand. The rubber hand illusion is most often assessed using questionnaires that measure the participant’s subjective experience (Botvinick & Cohen, 1998; Mussap & Salton, 2006; Pavani, Spence, & Driver, 2000; Schaefer, Flor, Heinze, & Rotte, 2007). The illusion can also be assessed by measuring: (1) proprioceptive drift * Corresponding author. Address: Department of Experimental Psychology, University of Oxford, South Parks Road, OX1 3UD, United Kingdom. Fax: +44 (0) 1865 310447. E-mail address: Rebekah.White@psy.ox.ac.uk (R.C. White). 1053-8100/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2009.08.003 506 R.C. White et al. / Consciousness and Cognition 19 (2010) 505–519 – the participant may mislocalise her own hidden hand, indicating that the felt position is closer to the observed prosthetic hand (Botvinick & Cohen, 1998; Costantini & Haggard, 2007; Durgin, Evans, Dunphy, Klostermann, & Simmons, 2007; Holmes, Crozier, & Spence, 2004; Tsakiris & Haggard, 2005; Tsakiris, Hesse, Boy, Haggard, & Fink, 2007; Tsakiris, Prabhu, & Haggard, 2006); (2) skin temperature – the participant may undergo a drop in the temperature of her own hidden hand because taking ownership of the observed prosthetic hand disrupts ownership of her own hand (Moseley, Olthof, Venema, Don et al., 2008); and (3) galvanic skin response – the participant may show an increased galvanic skin response if the observed prosthetic hand is threatened (e.g., by a needle) because it has been assimilated into the participant’s body image (Armel & Ramachandran, 2003; Hägni, Eng, Hepp-Reymond, Holper et al., 2008). The following conditions are best for demonstrating the visual rubber hand illusion: (1) stimulation on the participant’s hand is synchronous with stimulation on the observed prosthetic hand (Armel & Ramachandran, 2003; Botvinick & Cohen, 1998; Shimada, Fukuda, & Hiraki, 2009); (2) orientation of the strokes on the participant’s hand is matched to those on the observed prosthetic hand (Costantini & Haggard, 2007); and (3) postural alignment of the participant’s hand is matched to that of the observed prosthetic hand (Costantini & Haggard, 2007; Ehrsson, Spence, & Passingham, 2004; Pavani et al., 2000; Tsakiris & Haggard, 2005), with the observed prosthetic hand positioned (3a) in a biologically-plausible location relative to the participant’s own body (Armel & Ramachandran, 2003) and (3b) close to the participant’s own hidden hand (Lloyd, 2007). It may also be important for the observed prosthetic hand to correspond to the participant’s pre-existing higher-order body representation (Armel & Ramachandran, 2003; Costantini & Haggard, 2007; Haans, IJsselsteijn, & de Kort, 2008; Pavani & Zampini, 2007; Tsakiris et al., 2007). However, the research literature is divided on this point. Armel and Ramachandran (2003) found that the rubber hand illusion was elicited despite visual inconsistencies between the participant’s own hand and the prosthetic hand, including skin tone, hand size and distinguishing visual features such as nail polish (see also Longo, Schüür, Kammers, Tsakiris, & Haggard, 2009). Moreover, they reported that the illusion could even be elicited in the absence of a prosthetic hand, insofar as participants ‘‘often reported sensations arising from the table surface” (p. 1499) when it was stroked and tapped in precise synchrony with the stimulation administered to the participant’s hidden hand. This finding was not supported by Haans and colleagues (2008), who found that the subjective experience of the rubber hand illusion (as measured by questionnaire items) was significantly diminished when the participant viewed the table (rather than a prosthetic hand) being stimulated. Likewise, Tsakiris and Haggard (2005) found that synchronous stimulation of a wooden stick and the participant’s hand did not produce the illusion. Tsakiris and colleagues (2007) have also shown that the illusion is not elicited if the participant is observing stimulation on a prosthetic left hand while her own (hidden) right hand is stimulated. How do we account for these findings that the rubber hand illusion may be diminished when the participant views stimulation on a non-hand or wrong-hand object? One possible explanation is that it may be more difficult for the participant to experience illusory ownership over an object that does not correspond to her higher-order body representation, simply because the object conflicts with her stored representation of what belongs to her own body or to the human body more generally. An alternative explanation is that the participant may expect the stimulation on her own hand to feel a certain way and this expectation may depend on the viewed object. If the viewed object has a non-skin texture, stimulation on the participant’s own skin may feel different from what she would expect given the texture of the viewed object (Armel & Ramachandran, 2003; Haans et al., 2008) and this violation of tactile expectations may be sufficient to break the illusion. These alternative views are best conceptualised using cases from the literature. Armel and Ramachandran described four participants (out of 120) with particularly hairy hands who spontaneously indicated that ‘‘the illusion was ruined when their hand was touched in areas of high hair density” (p. 1504). The researchers suggested that it was ‘‘a mismatch in the expected (from visual information) versus felt type of touch, rather than just the visual inconsistencies of hair versus no hair, that diminished the illusion” (p. 1504). Not dissimilarly, Haans and colleagues (2008) found that, when the illusion was evoked using a prosthetic hand wearing a latex glove rather than a prosthetic hand of natural skin texture, ‘‘several participants remarked that their tactile sensations did not match those generally perceived while wearing gloves” (p. 393). But, as noted by Haans and colleagues, visual similarity and tactile expectation are confounded in these examples – the sensations on the participant’s hand match expectations when viewing the visually similar object but not when viewing the visually dissimilar object. Hence, it is not clear whether the illusion is affected by use of a visually dissimilar hand that the participant is not able to incorporate into a higher-order body representation, or by the violation of tactile expectation. Schütz-Bosbach, Tausche, and Weiss (2009) have very recently isolated the role of tactile expectation, thus addressing the concern that higher-order body representation and tactile expectation are confounded in previous research. Rather than manipulating the match between tactile properties of the observed prosthetic hand and of the participant’s own hand, the researchers manipulated the match between tactile properties of the stimulus that was administered to the observed prosthetic hand and of the stimulus administered to the participant’s own hand. For example, the participant viewed a prosthetic hand being touched with a soft fabric while receiving stimulation from a rough fabric on her own (hidden) hand. In this case there was no confound. The participant’s tactile expectation was violated but the participant was viewing stimulation upon a realistic prosthetic hand that could be incorporated into her higher-order body representation. Participants experienced the rubber hand illusion even when their tactile expectations were violated. When considered together with previous studies demonstrating that the rubber hand illusion is diminished if the participant views stimulation on a nonhand or wrong-hand object (or a hand that is markedly dissimilar to the participant’s own hairy hand; Armel & Ramachandran, 2003), Schütz-Bosbach and colleagues’ finding suggests that the disruption to the illusion occurs because the viewed R.C. White et al. / Consciousness and Cognition 19 (2010) 505–519 507 object cannot be incorporated into the participant’s higher-order body representation, rather than because the participant’s tactile expectations are violated. Here we propose that the interplay of higher-order body representation and tactile expectation may also be investigated using the non-visual rubber hand paradigm (Ehrsson, Holmes, & Passingham, 2005). With vision precluded, the Examiner guides the participant’s hand in administering stimulation to a prosthetic hand while the Examiner administers synchronous stimulation to the participant’s other hand. The participant may experience the illusion that she is touching her own hand, even when the two hands are separated by 15 cm. The illusion of self-touch can be assessed using questionnaires as well as measures of proprioceptive drift. As with the visual rubber hand illusion, a participant who experiences the non-visual rubber hand illusion, that is, the self-touch illusion, may mislocalise her ‘touched’ hand, indicating that its felt position is closer to the prosthetic hand than is actually the case (Ehrsson et al., 2005). In the elegant study by Ehrsson and colleagues (2005), the ‘‘participants, the experimenter, and the rubber hand all wore identical plastic surgical gloves to make the tactile surfaces of the two hands [participant and prosthetic] as similar as possible to each other” (p. 10565). When the participant’s stimulation of the prosthetic hand and the Examiner’s stimulation of the participant’s hand were synchronous, the participant experienced the compelling illusion of touching her own hand. In a control condition, the Examiner guided the participant instead to tap the bristles of a small dish brush with one hand while the Examiner synchronously stimulated the participant’s other hand. In this condition, ‘‘no illusion of self-touch was typically elicited” (p. 10566), but it is not clear whether this was due to the violation of higher-order body representation or the violation of tactile expectation since these factors were confounded. Higher-order body representation was violated because a dish brush cannot be assimilated into one’s body image, and tactile expectations were violated because under conditions of self-touch the participant would expect her administering index finger to feel a skin-like surface rather than the bristles of a brush. In the current study, our aim is to investigate the role of tactile expectation in the self-touch illusion, without confounding higher-order body representation and tactile expectation. We manipulate tactile properties of the administering stimulus (as in Schütz-Bosbach et al., 2009) rather than tactile properties of the object to which the participant administers stimulation. If the illusion of self-touch requires a match between expected and felt sensations, the participant will not experience the illusion when she is touching the prosthetic hand with one stimulus and receiving stimulation from a different stimulus. Alternatively, if expectations about tactile sensations do not affect the illusion, the impression of self-touch should persist provided there is synchronous stimulation of the two hands. 1.1. Overview of experiments Three experiments are presented. In Experiment 1, we assessed whether the illusion of self-touch could be elicited when participants used instruments, such as a paintbrush or a stick, as well as when they used an index finger to administer touch. In the only previous experiment using the non-visual rubber hand paradigm (Ehrsson et al., 2005), stimulation was always administered with the index finger. Experiment 1 was therefore a necessary precursor to later experiments, which introduced incongruent stimulation. To examine whether the non-visual rubber hand illusion breaks down under conditions of incongruent stimulation, it was first necessary to show that the illusion could be elicited using a variety of stimuli administered under congruent conditions and with vision precluded. In Experiment 2, we assessed whether the illusion of self-touch could be elicited when the participant administered stimulation to the prosthetic hand using her index finger while her other hand was touched with – an incongruent stimulus – a paintbrush. This tactile mismatch has an effect on both of the participant’s hands: the participant’s administering hand is receiving tactile sensations consistent with touching a prosthetic hand while the other hand is receiving tactile sensations consistent with being touched with a paintbrush. This experiment was conducted with vision precluded. In Experiment 3, we combined the methodologies of the traditional visual (Botvinick & Cohen, 1998) and non-visual (Ehrsson et al., 2005) rubber hand paradigms. The participants were involved in the administration of touch as they are in the non-visual paradigm but the paradigm was conducted with vision permitted. Stimulation was either congruent or incongruent. In the congruent condition, the participant administered stimulation to the prosthetic hand using her index finger while her other hand was touched with – a congruent stimulus – the Examiner’s index finger. In the incongruent condition, the participant administered stimulation to the prosthetic hand using her index finger while her other hand was touched with – an incongruent stimulus – a paintbrush. On a given trial, the participant had vision either of her right hand administering stimulation to the prosthetic hand or of her left hand receiving stimulation from the Examiner. 2. Experiment 1 2.1. Method 2.1.1. Participants Thirty-six right-handed (Oldfield, 1971) participants (aged 19–33 years; 19 females) took part in Experiment 1. Participants gave informed consent and the study was approved by the University of Oxford Research Ethics Committee and conducted in accordance with the ethical standards laid down in the 2008 Declaration of Helsinki. 508 R.C. White et al. / Consciousness and Cognition 19 (2010) 505–519 2.1.2. Materials and procedure The participant was seated at a testing table with a prosthetic left hand positioned at body midline. The participant’s own left hand was positioned to the left of the prosthetic hand with the two index fingers 15 cm apart. The Examiner sat on the opposite side of the table to the participant. The Examiner guided the participant’s right hand in administering stimulation to the index finger of the prosthetic hand, using (a) the participant’s index finger, (b) a paintbrush or (c) a stick. The paintbrush and the stick were mounted on identical handles. The participant was instructed to keep her eyes closed throughout the experimental trial, and to relax and allow her right hand to be guided by the Examiner. At the same time, the Examiner administered stimulation to the index finger of the participant’s left hand using the same (congruent) stimulus as was used to stimulate the prosthetic hand. For example, the participant was guided to administer stimulation to the prosthetic hand using a paintbrush while the Examiner used an identical paintbrush to administer stimulation to the participant’s left hand (Fig. 1). Stimulation comprised strokes and taps administered to the proximal phalanx of the index finger (including the metacarpal phalangeal joint and the proximal interphalangeal joint). Strokes were always unidirectional towards the finger tip. The Examiner’s stimulation of the participant’s left hand was timed to be either (a) synchronous or (b) asynchronous with the stimulation that the participant administered to the prosthetic hand. Each trial lasted for 60 s, and the order of conditions (i.e., finger, paintbrush, stick) was counterbalanced across participants using a Latin-Square design. 2.1.3. Measure of proprioceptive drift To examine the extent of proprioceptive drift associated with the self-touch illusion, we adapted the calibration method used by Ehrsson and colleagues (2005). Three measurements were obtained: two pre-stimulation measurements and one post-stimulation measurement. The pre-stimulation measurements were averaged to give a baseline measure of felt position. Proprioceptive drift was calculated as the change in felt position of the index finger from baseline. The prosthetic hand was removed from the table when these measurements were taken, immediately before and after each 60-s trial. The participant used her right index finger to point to the felt position of her left index finger. The participant extended her right arm at 45° to the right of the body’s midsagittal plane, and slid her right index finger along the testing table until it was in line with the felt position of her left index finger. Participants were instructed to point quickly using a single movement. The Examiner recorded the distance between the participant’s indicated and actual finger position. Note Fig. 1. Experimental set-up for stimulation administered using the paintbrush. Note that the index finger of the participant’s left hand is positioned 15 cm from the index finger of the prosthetic hand. R.C. White et al. / Consciousness and Cognition 19 (2010) 505–519 509 that the participant closed her eyes prior to the two pre-stimulation measurements, and was asked to maintain eyes closed for the duration of the experimental trial, that is, during the stimulation period and until the post-stimulation measure of felt position was obtained. 2.1.4. Rubber hand illusion questionnaire Following each trial, the participant indicated her level of agreement with two statements: (1) It felt as if I were touching my own hand. (2) It felt as if my left hand were shrinking. The first statement was adapted from Ehrsson and colleagues (2005) and was used to measure the participant’s subjective experience of the rubber hand illusion. This is the only statement used previously in the research literature to quantify the non-visual rubber hand illusion. The second statement served as a control for suggestibility, that is, a statement about an experience that the paradigm was not designed to evoke. A seven-point visual analogue scale (0 = not at all; 6 = very strongly agree) was used to rate agreement with the statements.1 2.2. Results 2.2.1. Proprioceptive drift We performed a mixed between- and within-subjects analysis of variance (ANOVA) on proprioceptive drift, with one between-subjects factor and two within-subjects factors. The between-subjects factor was order of the stimuli used to administer stimulation (order one: finger, paintbrush, stick; order two: paintbrush, stick, finger; order three: stick, finger, paintbrush). The within-subjects factors were mode of stroking (synchronous, asynchronous) and the stimulus used (finger, paintbrush, stick). There was a main effect of administration order (F(2, 33) = 5.336, p = .01, multivariate partial eta-squared = .244). Post-hoc comparisons using the Scheffé test indicated significantly greater proprioceptive drift (p = .014) for participants exposed to order two (paintbrush, stick, finger: M = 2.281 cm) compared with order one (finger, paintbrush, stick: M = .500 cm). There was a main effect of mode of stroking (F(1, 33) = 7.922, p = .008, multivariate partial eta-squared = .194), with greater proprioceptive drift for synchronous (M = 1.829 cm) compared with asynchronous (M = .648 cm) stimulation. There was no main effect of stimulus (F(2, 32) = 1.369, p = .269, multivariate partial eta-squared = .079), and only one interaction (administration order and stimulus) approached significance (p = .091; all other p values >.190). 2.2.2. Subjective experience of the rubber hand illusion We performed a mixed between- and within-subjects ANOVA on the participant’s subjective experience of the rubber hand illusion, with one between-subjects factor (administration order) and two within-subjects factors (mode of stroking and stimulus). There was no main effect of administration order (F(2, 33) = .178, p = .838, multivariate partial etasquared = .011). There was a main effect of mode of stroking (F(1, 33) = 88.743, p < .001, multivariate partial etasquared = .729), with higher illusion ratings for synchronous (M = 3.106) compared with asynchronous (M = .602) stimulation. There was also a main effect of stimulus (F(2, 32) = 8.237, p = .001, multivariate partial eta-squared = .340). There were significant two-way interactions of administration order and stimulus (F(4, 64) = 2.807, p = .033, multivariate partial etasquared = .149) and mode of stroking and stimulus (F(2, 32) = 14.170, p < .001, multivariate partial eta-squared = .470). There was also a significant three-way interaction of administration order, mode of stroking and stimulus (F(4, 64) = 2.891, p = .029, multivariate partial eta-squared = .153). Separate ANOVA were conducted for synchronous and asynchronous stimulation to investigate these interactions. For synchronous stimulation (Fig. 2: top panel), there was no main effect of administration order (F(2, 33) = .247, p = .783, multivariate partial eta-squared = .015) but there was a main effect of stimulus (F(2, 32) = 13.105, p < .001, multivariate partial eta-squared = .450) and a significant interaction of administration order and stimulus (F(4, 64) = 3.795, p = .008, multivariate partial eta-squared = .192). As depicted in the top panel of Fig. 2, participants in administration order 2 provided equally high illusion ratings for the three stimuli, (F(2, 10) = 1.254, p = .327, multivariate partial eta-squared = .201). In contrast, participants in administration order 1 had a significant difference in the strength of illusion ratings for the three stimuli (F(2, 10) = 6.442, p = .016, multivariate partial eta-squared = .563). Pairwise comparisons (adjusting for multiple comparisons) indicated higher illusion ratings for the paintbrush (M = 3.750) than for the finger (M = 1.833, p = .010). Participants in administration order 3 also had a significant difference in the strength of illusion ratings for the three stimuli (F(2, 1 The strength of the rubber hand illusion is frequently measured using a seven-point scale with negative through positive values. Participants provide an agreement rating with statements probing experience of the illusion (e.g., 3 = strongly disagree; +3 = strongly agree). However, it has been argued recently (Miles, 2008, April) that participants who do not experience the illusion may find it difficult to rate non-agreement (i.e., 1 vs. 2 vs. 3). Therefore we use a scale in which failure to experience the illusion receives a zero rating (not at all) and experience of the illusion is rated 1 (slightly agree) through 6 (very strongly agree). This scale provides greater range for rating relative strength for the vast majority of participants who do experience the illusion. Notably, results from our scale (in the conditions in which stimulation was administered with a finger) are comparable to the pioneering study by Ehrsson et al. (2005). In the current study, 31 of 36 participants (86%) provided ‘agree’ ratings compared with 25 of 32 participants (78%) in the study by Ehrsson et al. (v2(1, N = 68) = .295, p = .587). Recent studies using a similar (positive) scale to ours include: Haans et al. (2008): 0 = not at all; 10 = completely; Moseley et al. (2008): 0 = not at all vivid; 10 = completely vivid. 510 R.C. White et al. / Consciousness and Cognition 19 (2010) 505–519 Fig. 2. Top panel – synchronous stimulation; Bottom panel – asynchronous stimulation. Illusion ratings as a function of administration order and stimulus. Error bars represent Standard Error of the Mean. 10) = 7.485, p = .010, multivariate partial eta-squared = .600). Pairwise comparisons (adjusting for multiple comparisons) indicated higher illusion ratings for the paintbrush (M = 4.250) than for the stick (M = 2.167, p = .006) or the finger (M = 2.500, p = .014). Thus the interaction of administration order and stimulus indicates that the paintbrush may have a significant effect on the illusion ratings of the finger and stick stimuli if they follow paintbrush in the administration order. For asynchronous stimulation (Fig. 2: bottom panel), there was no main effect of administration order (F(2, 33) = .082, p = .922, multivariate partial eta-squared = .005) or stimulus (F(2, 32) = .283, p = .755, multivariate partial etasquared = .017) and there was no interaction of administration order and stimulus (F(4, 64) = .791, p = .536, multivariate partial eta-squared = .047). 2.2.3. Suggestibility The mean rating for the control statement was less than .5 for each stimulus (finger, paintbrush, stick), regardless of whether stimulation was synchronous or asynchronous. Paired t-tests (using an adjusted alpha level of .0167) confirmed that there was no difference in agreement ratings for synchronous versus asynchronous stimulation: finger (t(35) = .211, p = .834); paintbrush (t(35) = .452, p = .654); stick (t(35) = .298, p = .768). R.C. White et al. / Consciousness and Cognition 19 (2010) 505–519 511 A further comparison (using an adjusted alpha level of .0167) confirmed higher agreement ratings for the illusion statement (It felt as if I were touching my own hand) compared with the control statement (It felt as if my left hand were shrinking) in the synchronous stimulation conditions: finger (t(35) =7.032, p < .001); paintbrush (t(35) = 10.736, p < .001); stick (t(35) = 7.289, p < .001). 2.2.4. Correlation of proprioceptive drift and questionnaire ratings For each stimulus, there was a significant correlation between proprioceptive drift and the difference between agreement ratings for the illusion and control statements on the questionnaire: finger (r = .534, p < .001); paintbrush (r = .371, p = .013); stick (r = .287, p = .045). 3. Experiment 2 3.1. Method 3.1.1. Participants Twenty-three new right-handed participants took part in Experiment 2. One participant was excluded due to failure to maintain eyes closed. Data from 22 participants (aged 18–31 years; 16 females) were analysed. 3.1.2. Materials and procedure The procedure was closely modelled on Experiment 1. Experiment 2 comprised four 60-s trials, two with congruent stimulation and two with incongruent stimulation. The Examiner guided the participant to use the index finger of one hand to administer stimulation to the prosthetic hand while the Examiner administered stimulation to the participant’s other hand, either with an index finger (congruent stimulation – Trials 1 and 2) or with a paintbrush (incongruent stimulation – Trials 3 and 4). Stimulation was timed so as to be (a) synchronous (Trials 1 and 3) or (b) asynchronous (Trials 2 and 4). To ensure that the participant was relaxed with the procedure for administering stimulation, the Examiner and participant practised the guided stimulation technique before the experimental trials began. The participant was instructed to keep her eyes closed throughout the experimental trial, and to relax and allow her right hand to be guided by the Examiner. 3.1.3. Measure of proprioceptive drift Proprioceptive drift was assessed with the pointing judgment task used in Experiment 1. As with Experiment 1, the participant closed her eyes prior to the two pre-stimulation measurements, and she was asked to maintain eyes closed for the duration of the experimental trial, that is, during the stimulation period and until the post-stimulation measure of felt position was obtained. 3.1.4. Rubber hand illusion questionnaires Following each trial, the participant completed two short questionnaires. The first questionnaire comprised two statements as in Experiment 1: (1) It felt as if I were touching my own hand. (2) It felt as if my left hand were shrinking. The second questionnaire comprised eight statements designed to measure the participant’s perception of the stimuli used in the experimental trials. The statements were based on the experiences reported by participants who took part in earlier experiments piloting the use of incongruent stimulation: (1) It felt as if I were administering touch with a finger. (2) It felt as if I were administering touch with a brush. (3) It felt as if I were administering touch with a finger and a brush simultaneously. (4) My left index finger felt as if it were touched by a finger. (5) My left index finger felt as if it were touched by a brush. (6) My left index finger felt as if it were touched by a finger and a brush simultaneously. (7) My right index finger felt like a finger. (8) My right index finger felt like a brush. The order of statements was randomised across trials, and a seven-point visual analogue scale (0 = not at all; 6 = very strongly agree) was used to rate agreement with the statements. Once the participant had completed these subjective items assessing felt experience, that is, her perception of the stimuli used, she was asked to name both the stimulus used to administer stimulation to the prosthetic hand and the stimulus used to administer stimulation to her own hand. This question was designed to verify that the participant understood that the stimuli were incongruent on the critical trials. 512 R.C. White et al. / Consciousness and Cognition 19 (2010) 505–519 3.2. Results 3.2.1. Proprioceptive drift We performed an ANOVA on proprioceptive drift. The two within-subjects factors were congruence of stimulation (congruent, incongruent) and mode of stroking (synchronous, asynchronous). There was no main effect of congruence (F(1, 21) = .020, p = .890, multivariate partial eta-squared = .001). There was a main effect of mode of stroking (F(1, 21) = 16.220, p = .001, multivariate partial eta-squared = .436), with greater proprioceptive drift for synchronous (M = 2.381 cm) compared with asynchronous (M = .034 cm) stimulation. There was no interaction of congruence and mode of stroking (F(1, 21) = .317, p = .579, multivariate partial eta-squared = .015). See Fig. 3. 3.2.2. Subjective experience of the rubber hand illusion We performed an ANOVA on the participant’s subjective experience of the rubber hand illusion. The within-subjects factors were congruence of stimulation (congruent, incongruent) and mode of stroking (synchronous, asynchronous). There was no main effect of congruence (F(1, 21) = 1.634, p = .215, multivariate partial eta-squared = .072). There was a main effect of mode of stroking (F(1, 21) = 36.426, p < .001, multivariate partial eta-squared = .634), with higher overall illusion ratings for synchronous (M = 3.091) compared with asynchronous (M = .795) stimulation. There was no interaction of congruence and mode of stroking (F(1, 21) = .009, p = .926, multivariate partial eta-squared = .000). See Fig. 4. 3.2.3. Suggestibility The mean rating for the control statement was less than .5 for congruent and incongruent stimulation trials, regardless of whether stimulation was synchronous or asynchronous. Paired t-tests (using an adjusted alpha level of .025) confirmed that there was no difference in agreement ratings for synchronous versus asynchronous stimulation for either congruent stimulation (t(21) = .826, p = .418) or incongruent stimulation (t(21) = 1.449, p = .162). A further comparison (using an adjusted alpha level of .025) confirmed higher agreement ratings for the illusion statement (It felt as if I were touching my own hand) compared with the control statement (It felt as if my left hand were shrinking) in each of the synchronous stimulation conditions: congruent stimulation (t(21) = 6.757, p < .001); incongruent stimulation (t(21) = 5.376, p < .001). 3.2.4. Correlation of proprioceptive drift and questionnaire ratings There was no significant correlation found, for either congruent stimulation or incongruent stimulation, between proprioceptive drift and the difference between agreement ratings for the illusion and control statements on the first questionnaire: congruent stimulation (r = .200, p = .373); incongruent stimulation (r = .251, p = .259). 3.2.5. Perception of the stimulus On the congruent trial with synchronous stimulation, 20 of 22 participants correctly reported receiving stimulation from an index finger. On the incongruent trial with synchronous stimulation, all 22 participants correctly reported receiving Fig. 3. Mean proprioceptive drift for each of the stimulus conditions in Experiment 1 (collapsing across administration order) and Experiment 2. Error bars represent Standard Error of the Mean. R.C. White et al. / Consciousness and Cognition 19 (2010) 505–519 513 Fig. 4. Illusion ratings as a function of stimulus congruence and mode of stroking (synchronous versus asynchronous). Error bars represent Standard Error of the Mean. stimulation from a brush. (Not surprisingly, all participants correctly reported administering stimulation with an index finger on the congruent and incongruent stimulation trials.) Under conditions of incongruent stimulation, the participant may bind the tactile sensations from each hand when she experiences the illusion of self-touch. This may lead either to the perception that the receptive left hand is being simultaneously touched by two stimuli or, alternatively, to the perception that the index finger is brush-like. Two planned analyses (using an adjusted alpha level of .025) were conducted to assess the impact of stimulus incongruence on tactile perception. First, a paired t-test was conducted on ratings provided for the statement ‘My left index finger felt as if it were touched by a finger and a brush simultaneously’. There was a significant difference (t(21) = 2.42, p = .025) in the level of agreement between the congruent (M = .318) and incongruent stimulation trials (M = 1.41), despite the fact that in both trials the participant received touch from one stimulus only. Second, a paired t-test was conducted on ratings provided for the statement ‘My right index finger felt like a brush’. There was a significant difference (t(21) = 4.143, p < .001) in the level of agreement between the congruent (M = .318) and incongruent stimulation trials (M = 1.64), despite the fact that in both trials the participant’s right index finger was administering touch to the prosthetic hand. 4. Experiment 3 4.1. Method 4.1.1. Participants Twelve participants (aged 19–27 years; 9 females) took part in Experiment 3. The participants were recruited from amongst 16 individuals who had experienced a strong (or very strong) illusion of self-touch in any of the stimulus conditions (finger, paintbrush, stick) used in Experiment 1. Participants were tested at least 1 week following participation in Experiment 1. 4.1.2. Experiment overview Experiment 3 comprised six trials. The baseline conditions (Trials 1 and 4) were conducted with vision precluded, and the remaining trials (Trials 2, 3, 5, 6) with vision permitted. 4.2. Baseline conditions 4.2.1. Materials and procedure Trials 1 and 4 were congruent (finger–finger) 60-s stimulation trials, conducted with vision precluded. The Examiner guided the participant’s right index finger to administer stimulation to the prosthetic hand while the Examiner administered synchronous stimulation to the participant’s left hand. The procedure for stimulation was as in Experiments 1 and 2 except that only strokes of the proximal phalanx of the index finger were used, rather than strokes and taps, and the strokes were always unidirectional towards the wrist, rather than towards the finger tip. 514 R.C. White et al. / Consciousness and Cognition 19 (2010) 505–519 4.2.2. Rubber hand illusion questionnaire Following each trial, the participant indicated her level of agreement with two statements as in Experiment 1: (1) It felt as if I were touching my own hand. (2) It felt as if my left hand were shrinking. A seven-point visual analogue scale (0 = not at all; 6 = very strongly agree) was used to rate agreement with the statements. 4.3. Vision conditions 4.3.1. Materials and procedure Trials 2, 3, 5 and 6 were conducted with vision permitted. Two examiners were required to conduct these trials. Examiner 1 sat to the left of the participant and Examiner 2 was positioned to the right. Trials 2 and 5 comprised congruent stimulation: Examiner 1 used her right index finger to administer stimulation to the participant’s left hand while Examiner 2 guided the participant’s right index finger to administer stimulation to the prosthetic hand (Fig. 5). Trials 3 and 6 comprised incongruent stimulation: Examiner 1 used a paintbrush to administer stimulation to the participant’s left hand while Examiner 2 guided the participant’s right index finger to administer stimulation to the prosthetic hand. Timing was guided by the stimulation administered by Examiner 1. Six participants were permitted vision of their left hand on Trials 2 and 3 (view-left condition) and vision of their right hand on Trials 5 and 6 (view-right condition). The remaining six participants were permitted vision of their right hand on Trials 2 and 3 (view-right condition) and vision of their left hand on Trials 5 and 6 (view-left condition). In the view-left condition, the participant had vision of her own left hand being stroked by Examiner 1. In the view-right condition, the participant had vision of her own right hand as it was guided by Examiner 2 to stroke the prosthetic hand. A partition was placed between the participant’s left hand and the prosthetic hand to ensure that the participant could only see the hand(s) she was permitted to view. Once Examiner 1 and Examiner 2 had achieved synchronous stimulation, as determined by the participant and Examiner 2, the stopwatch was activated. Sixty seconds of synchronous stimulation followed. 4.3.2. Rubber hand illusion questionnaire Following each trial, the participant indicated her level of agreement with three statements: (1) It felt as if I were touching my own hand. (2) It seemed as if I were observing my right hand stroking my left hand. (3) It felt as if my left hand were shrinking. The first statement was used to measure the participant’s subjective experience of the rubber hand illusion. The second statement was used to measure whether participants experienced ownership of the Examiner’s administering hand (viewleft condition) or the prosthetic hand (view-right condition). The third statement served as a control for suggestibility. A seven-point visual analogue scale (0 = not at all; 6 = very strongly agree) was used to rate agreement with the statements. Fig. 5. Experimental set-up in a congruent stimulation trial in which the participant was permitted vision of her own left hand as it was touched by Examiner 1. Black fabric was draped over the arms on the viewed side to hide irrelevant cues from clothing (view-left condition) or from the stump of the prosthetic hand (view-right condition). R.C. White et al. / Consciousness and Cognition 19 (2010) 505–519 515 4.4. Results Experiment 2 demonstrated that there was no difference in illusion ratings between the congruent and incongruent stimulation conditions. Since we predicted similar non-significant findings for Experiment 3 (see also Schütz-Bosbach et al., 2009), we used an adjusted alpha level of .10 to control for the possibility that a non-significant result may be due to insufficient power (Stevens, 1996). 4.4.1. Vision and the illusion of self-touch To assess the effect of vision of the left (touched) or right (touching) hand, on the illusion that one is touching one’s own hand, we conducted a mixed between- and within-subjects ANOVA on the illusion ratings (Statement 1 ‘It felt as if I were touching my own hand’) for conditions of congruent stimulation (Trials 1, 2, 4, 5). There was one between-subjects factor and two within-subjects factors. The between-subjects factor was order – participants viewed their left hand or their right hand first. The within-subjects factors were vision (vision precluded, vision permitted) and hand viewed by the participant (view-left, view-right). There was no main effect of order (F(1, 10) = .049, p = .830, multivariate partial eta-squared = .005). There was a main effect of vision (F(1, 10) = 5.090, p = .048, multivariate partial eta-squared = .337), with higher illusion ratings for trials conducted with vision precluded (M = 3.271) compared to trials with vision permitted (M = 2.729). There was no main effect of viewed-hand (F(1, 10) = .057, p = .816, multivariate partial eta-squared = .006) but there was a significant interaction of vision and viewed-hand (F(1, 10) = 4.447, p = .061, multivariate partial eta-squared = .308) at the adjusted alpha level of .10. To assess this interaction, we conducted two separate paired t-tests (using an adjusted alpha level of .05). The first t-test demonstrated no difference in illusion ratings between the non-visual trial before participants looked at their left hand (M = 3.04) and the visual trial in which participants looked at their left hand (M = 3.04) (t(11) = .000, p = 1.000). The second t-test demonstrated a difference in illusion ratings between the non-visual trial before participants looked at their right hand (M = 3.5) and the visual trial in which participants looked at their right hand (M = 2.41) (t(11) = 2.493, p = .03). There were no other interactions (all p values >.149). 4.4.2. Stimulus incongruence and the illusion of self-touch To assess whether stimulus incongruence affects the felt illusion that one is touching one’s own hand under conditions permitting vision, we conducted a mixed between- and within-subjects ANOVA on illusion ratings (Statement 1 ‘It felt as if I were touching my own hand’). There was one between-subjects factor and two within-subjects factors. The between-subjects factor was order – participants viewed their left or their right hand first. The within-subjects factors were congruence of stimulation (congruent, incongruent) and hand viewed by the participant (view-left, view-right). There were no main effects of order (F(1, 10) = .000, p = .987, multivariate partial eta-squared = .000), congruence (F(1, 10) = .254, p = .625, multivariate partial eta-squared = .025) or viewed-hand (F(1, 10) = 1.549, p = .242, multivariate partial eta-squared = .134), and there were no significant interactions (all p values >.199). 4.4.3. Stimulus incongruence and the illusion of ownership To assess whether stimulus incongruence affects the illusion of ownership of the prosthetic hand or of the Examiner’s administering hand, we conducted a mixed between- and within-subjects ANOVA on ownership ratings (Statement 2 ‘It seemed as if I were observing my right hand stroking my left hand’). There was one between-subjects factor and two within-subjects factors. The between-subjects factor was order – participants viewed their left or their right hand first. The within-subjects factors were congruence of stimulation (congruent, incongruent) and hand viewed by the participant (view-left, view-right). There was no main effect of order (F(1, 10) = .561, p = .471, multivariate partial eta-squared = .053). There was a main effect of congruence (F(1, 10) = 7.840, p = .019, multivariate partial eta-squared = .439), indicating higher levels of agreement when stimulation was congruent (M = 2.062) compared with when stimulation was incongruent (M = 1.479). There was no main effect of viewed-hand (F(1, 10) = 1.257, p = .289, multivariate partial eta-squared = .112), and there were no significant interactions (all p values >.124). 4.4.4. Suggestibility The mean rating for the control statement was less than .5 for all trials. Paired t-tests (using an adjusted alpha level of .025) confirmed that, in all conditions, participants indicated higher levels of agreement with the illusion statement (It felt as if I were touching my own hand) compared with the control statement (It felt as if my left hand were shrinking): view-left congruent (t(11) = 3.959, p = .002); view-left incongruent (t(11) = 3.006, p = .012); view-right congruent (t(11) = 3.694, p = .004); view-right incongruent (t(11) = 3.764, p = .003). 5. General discussion We set out to investigate whether the illusion of self-touch is affected by violations of tactile expectation. We used the non-visual rubber hand paradigm devised by Ehrsson and colleagues (2005) to address this question. In this paradigm, the Examiner guides the participant to administer stimulation to a prosthetic hand while the Examiner administers synchronous stimulation to the participant’s hand. If the participant experiences the illusion, she reports feeling as if she is touching 516 R.C. White et al. / Consciousness and Cognition 19 (2010) 505–519 her own hand. Three experiments were conducted. In Experiment 1, we assessed whether the illusion of self-touch could be elicited when participants used instruments, such as a paintbrush or a stick, as well as when they used an index finger to administer touch. In the only previous experiment using the non-visual rubber hand paradigm (Ehrsson et al., 2005), stimulation was always administered with the index finger. In Experiment 2, we assessed whether the illusion of self-touch could be elicited when the participant administered stimulation to the prosthetic hand using her right index finger while her left hand was touched with – an incongruent stimulus – a paintbrush. In Experiment 3, we assessed whether the illusion of self-touch would be diminished when the participant had visual feedback indicating incongruent stimulation. This experiment combined the methodologies of the traditional visual (Botvinick & Cohen, 1998) and non-visual (Ehrsson et al., 2005) rubber hand paradigms. The participants were involved in the administration of touch as they are in the non-visual paradigm but the paradigm was conducted with vision permitted. The primary objective of Experiment 1 was to investigate whether the illusion of self-touch can be elicited when the participant uses an instrument (i.e., paintbrush or stick) instead of an index finger to administer touch to the prosthetic hand while the Examiner uses an identical instrument to administer touch to the participant’s other hand, that is, the hand not involved in administration. The results, showing that synchronous stimulation produced greater proprioceptive drift and higher illusion ratings on the rubber hand questionnaire than asynchronous stimulation, indicate that the self-touch illusion can be elicited when touch is administered with instruments. We also found that order of administration was important to the magnitude of the illusion, so that participants who first received stimulation with the paintbrush (administration order 2) displayed high illusion ratings for all three stimuli and overall greater proprioceptive drift. We suggest that the illusion of self-touch is most readily evoked using the paintbrush, possibly because the flexible nature of the stimulus makes it less likely to be affected by small differences in timing, pressure and pattern of stimulation to the two hands. Once a participant has experienced the illusion with the paintbrush, she may be more prepared to experience it with subsequent stimuli. It may otherwise take time for the participant to be relaxed about the experimental procedure and to allow her hand to be guided by the Examiner. Potential problems were addressed in Experiment 2 by having the participant and Examiner practise the guided stimulation technique before the experimental trials and in Experiment 3, in which two Examiners were involved in administering stimulation, by starting the stopwatch only when the participant and Examiner 2 agreed that synchronous stimulation of the participant’s and the prosthetic hand had been achieved. Having demonstrated that the illusion of self-touch can be elicited using a variety of stimuli in the non-visual rubber hand paradigm, we designed Experiment 2 to assess whether the illusion persists if there is a mismatch in the stimuli used to administer stimulation. Incongruence was set up by having the participant administer touch to the prosthetic hand using her right index finger while the Examiner administered touch to the participant’s left hand using a paintbrush. The results, showing that synchronous stimulation produced greater proprioceptive drift and higher illusion ratings on the rubber hand questionnaire than asynchronous stimulation, indicate that the illusion of touching one’s own hand persists, even when the participant is aware that she is administering touch with a different type of stimulus from the one touching her other hand.2 There was no significant difference in proprioceptive drift when the participant administered touch with a finger and was synchronously touched with a finger (congruent stimulation trial: M = 2.443 cm) compared with when the participant administered touch with a finger and was synchronously touched with a paintbrush (incongruent stimulation trial: M = 2.318 cm). This finding indicates that participants experience mislocalisation in the felt position of the receptive left hand, whether stimulation is congruent or incongruent. Similarly, in response to the statements in the questionnaire, there was no significant difference in the illusion rating between the congruent (M = 3.182) and incongruent (M = 3.000) stimulation trials, indicating that the subjective illusion of self-touch is equally compelling whether stimulation is congruent or incongruent. Schütz-Bosbach and colleagues (2009) demonstrated that the visual rubber hand illusion was resistant to violations of tactile expectation. In the incongruent trial of their study, the participant viewed a prosthetic hand being touched with either a soft or rough fabric while she received incongruent tactile stimulation on her hidden hand. We build on the results of Schütz-Bosbach and colleagues by demonstrating that the non-visual rubber hand illusion is also resistant to violations of tactile expectation. Participants experience the illusion of self-touch even under conditions of incongruent stimulation: the participant administers stimulation to the prosthetic hand using the index finger of her right hand while stimulation is administered to her left hand using a paintbrush. In the only prior study using the non-visual paradigm (Ehrsson et al., 2005), the incongruent stimulation trial took a different format. Specifically, the Examiner guided the participant to tap the bristles of a small dish brush while the Examiner synchronously stimulated the participant’s other hand. The illusion of self-touch was not typically elicited in this condition. There are some important differences between our Experiment 2 and the study by Ehrsson and colleagues (2005). In Experiment 2, the participant always administered stimulation to a prosthetic hand using her index finger but we manipulated the stimulus – index finger or paintbrush – the Examiner was using to touch the participant’s other hand. In contrast, in the study by Ehrsson and colleagues (2005), the researchers manipulated the object to which the participant administered stimulation – prosthetic hand or dish brush, but the Examiner always used an index finger to touch the participant’s other hand. 2 Earlier pilot experiments demonstrated that the illusion of self-touch could also be elicited when the participant administered stimulation with a paintbrush while her own hand was touched with a stick (and vice versa). However, in these conditions, the participant may have simply attended to administering touch with, and receiving touch from, an instrument; thus failing to register the stimulus incongruence. In the experiment proper, we used the finger-paintbrush incongruence condition, and so avoided this possible confound. R.C. White et al. / Consciousness and Cognition 19 (2010) 505–519 517 In short, we introduced incongruent stimulation by varying the instrument of stimulation while Ehrsson and colleagues (2005) introduced it by varying the object of stimulation. In our incongruent stimulation condition, the participant’s tactile expectations were violated but there was no conflict with her higher-order body representation. In this condition, participants experienced the illusion of self-touch. We suggest that, in the study by Ehrsson and colleagues (2005), participants did not experience the illusion of self-touch in the incongruent stimulation condition because they were unable to incorporate a dish brush into their higher-order body representation, whereas the prosthetic hand was more easily incorporated. One might question how the participant derives information that allows the prosthetic hand to be assimilated into her body image when vision of the hand is precluded. We believe that this information can be derived through stimulation of the prosthetic hand itself. The prosthetic hand used for the current experiments was exceptionally lifelike in shape and detail. When the participant administered stimulation, she was guided between the metacarpal phalangeal joint (the knuckle where the index finger meets the hand) and the proximal interphalangeal joint (the first knuckle of the index finger). The matching shape and structure of the stimulated object would have been sufficient to indicate its correspondence with the participant’s own stimulated hand. In Experiment 3, we introduced vision to assess whether visual feedback regarding stimulus incongruence would break the illusion of self-touch. There are two key differences in experimental design between the traditional visual rubber hand paradigm and the paradigm used in Experiment 3. First, in the traditional visual rubber hand paradigm, the participant is not involved in the administration of touch so there can be no illusion of self-touch, whereas in Experiment 3, the Examiner guided the participant to administer stimulation to the prosthetic hand. Second, in the traditional visual rubber hand paradigm, the participant is permitted vision of the prosthetic hand only, whereas in Experiment 3, the participant was permitted vision of either her right hand administering stimulation to the prosthetic hand or her left hand while it was being touched by the Examiner. Experiment 3 addressed two important questions: Can the illusion of self-touch be evoked when the participant has visual feedback? And, if the illusion can be evoked, will it persist when visual feedback indicates incongruent stimulation? A failure to experience the self-touch illusion in the visual condition may occur either because the illusion cannot be evoked with vision or because the participant does not experience the self-touch illusion irrespective of the visual manipulation. This second interpretation was discounted by testing participants from Experiment 1 who had been shown to experience the illusion in the non-visual version of the paradigm. In the congruent stimulation condition, participants did experience the illusion of self-touch when vision was permitted. Five participants (42%) indicated strong or very strong agreement with the statement ‘It felt as if I were touching my own hand’ even when they had visual feedback indicating that this was not the case. Note though that the illusion of self-touch was diminished by vision when the participant was looking at her right hand touching the prosthetic hand. How do we explain the finding that the illusion of self-touch was elicited whether the participant looked at her right hand touching the prosthetic hand or at her left hand being touched by the Examiner? In the view-right condition, the participant observed touch on the prosthetic hand which corresponded to the touch on her (hidden) left hand. The viewing conditions were thus matched to the traditional visual rubber hand paradigm, in which the Examiner administers stimulation to the prosthetic hand, but now the Examiner guided the participant to administer the observed stimulation to the prosthetic hand. When the participant views touch on the prosthetic hand that corresponds to touch on her own hand, she may experience visual capture of touch – the illusion that she is experiencing tactile sensations in the location of the viewed prosthetic hand (Botvinick & Cohen, 1998). The novel finding is that participants can also experience a visual rubber hand illusion when they are not looking at a prosthetic hand. In the view-left condition, the participant observed touch being administered to her own left hand by the Examiner. The action of the Examiner’s administering hand corresponded to the action of her (hidden) right hand which was administering touch to the prosthetic hand. We propose that, to experience the illusion that her hands are in contact, the participant may experience visual capture of action – the illusion that the action of her administering hand is in the location of the Examiner’s administering hand. Both types of displacement (touch and action) may lead to the visual illusion of self-touch. In the first case, tactile sensations are displaced to the location of the prosthetic hand and thus the location of the participant’s administering hand. In the second case, action is displaced to the location of the Examiner’s administering hand and thus the location of the participant’s touched hand. Note that the concept of visual capture of action is not new. Nielsen (1963) conducted an elegant experiment in which the participant completed a simple line-drawing task. Unbeknownst to the participant, a mirror was inserted into the experimental set-up so that the participant was viewing another person’s hand drawing the lines, rather than her own hand. Participants experienced the so-called alien hand as their own, making compensatory movements when the alien hand performed in an unpredictable manner. For example, when the task was to draw a straight line but the viewed hand veered rightward, most participants compensated for this error with a leftward adjustment of their own hand. Nielsen concluded that the ‘‘alien ‘visual hand’ dominate[d] the subject’s ‘kinesthetical/tactile hand’” (p. 230). He noted that most participants did not initially realise that they were viewing someone else’s hand. In considering Nielsen’s findings, we suggest that this pioneering study essentially evoked what would today be regarded as a rubber hand illusion, whereby participants took ownership of the actions of a viewed hand. (For a recent study demonstrating the illusion of ownership elicited using a moving rubber hand, see: Dummer, Picot-Annand, Neal, & Moore, 2009.) Having established that the illusion of self-touch can be elicited under congruent stimulation conditions which permit visual feedback, we next assessed whether the illusion was diminished under incongruent stimulation conditions which permit visual feedback. Specifically, we compared (a) the congruent trial in which the participant administered stimulation 518 R.C. White et al. / Consciousness and Cognition 19 (2010) 505–519 with, and received stimulation from, an index finger and (b) the incongruent trial in which the participant administered stimulation with an index finger, and received stimulation from a paintbrush. There was no difference in illusion ratings between the congruent (M = 2.729) and incongruent (M = 2.583) stimulation trials and three participants (25%) indicated strong or very strong agreement with the statement ‘It felt as if I were touching my own hand’ even when they had visual feedback indicating incongruent stimulation. Experiment 3 demonstrates that when vision is permitted in the ‘non-visual’ rubber hand paradigm (in which the participant is involved in administering stimulation), participants experience an equally compelling illusion of self-touch for congruent and incongruent stimulation. A further question that we asked in Experiment 3 was whether participants would experience the illusion that they were observing their own right hand touching their own left hand, and whether this illusion would be affected by incongruent stimulation. The sense that one is observing one’s two hands in contact may be taken to suggest that the participant is experiencing illusory ownership over the ‘alien’ hand (i.e., the Examiner’s hand in the view-left condition or the prosthetic hand in the view-right condition). The illusion that the participant was observing her right hand touching her left hand was diminished by stimulus incongruence. Participants indicated higher levels of agreement with the statement ‘It seemed as if I were observing my right hand stroking my left hand’ in the congruent stimulation condition (M = 2.333) than in the incongruent stimulation condition (M = 1.479). How do we account for the diminished sense of ownership in the incongruent condition? When the participant looked to her left hand, the visual image was of a paintbrush touching her left index finger. This visual image was inconsistent with the tactile perceptions of the right hand, not only because of the incongruent stimulus but also because there was no guiding Examiner’s hand.3 Note that this latter point was also true of the congruent stimulation condition. When the participant looked to her right hand, the visual image was of her right index finger being guided to touch the prosthetic hand. This visual image was inconsistent with the tactile perceptions of the left hand. The results suggest that, although visual feedback of stimulus incongruence does not affect the felt illusion of self-touch, it does affect the illusion that the observed ‘alien’ hand is one’s own. 6. Conclusion In the current study, we investigated the role of tactile expectations in the self-touch illusion, without confounding tactile expectation and higher-order body representation. This was achieved by manipulating tactile properties of the administering stimulus (as in Schütz-Bosbach et al., 2009) rather than tactile properties of the object to which the participant administered stimulation. Experiment 1 established that the self-touch illusion was elicited when the participant administered stimulation to a prosthetic hand using a paintbrush or a stick while the Examiner administered stimulation to the participant’s other hand using a congruent stimulus. Previously, the self-touch illusion has only been evoked using stimulation administered with an index finger. In Experiments 2 and 3 we manipulated tactile expectations: the participant administered stimulation to the prosthetic hand using her index finger while the participant’s own hand was touched with either (a) a congruent stimulus – a finger or (b) an incongruent stimulus – a paintbrush. The self-touch illusion was not diminished by incongruent stimulation, and this was true whether the procedure was conducted with vision precluded or permitted. Armel and Ramachandran (2003) suggest that the traditional visual rubber hand illusion may be explained using a Bayesian model of perceptual learning, in which ‘‘two perceptions from different modalities are ‘bound’ when they co-occur with a high probability” (p. 1505). In the traditional visual rubber hand paradigm, the relevant modalities are vision (the observed touching of the prosthetic hand) and touch (the felt sensations in the participant’s hand). Thus, ‘‘the seen and felt touch were bound because of their temporal synchrony” (ibid.). A Bayesian model may likewise explain the illusion of self-touch, whereby the proprioceptive cues from the administering hand are bound with the tactile sensations of the receptive hand. Participants experience the illusion of self-touch when they administer touch to a prosthetic hand while receiving synchronous touch from the Examiner (see Ramachandran & Hirstein, 1997, 1998). According to Bayesian logic, there is a very low probability that the actions of the administering hand and the sensations on the receptive hand could correspond so precisely by chance; thus participants experience the illusion that the administering hand is touching the receptive hand. Self-as-active and self-as-receptive are experienced as participants in a single event of self-touch. Using incongruent stimuli (and indeed introducing vision) does not change the temporal correspondence between the actions of the administering hand and the sensations of the receptive hand, and the improbability that this correspondence occurred by chance. Armel and Ramachandran (2003) note that the brain takes advantage of statistical correlations, ‘‘even when they do not ‘make sense’ from the cognitive point of view” (p. 1505). In participants who experience the illusion of self-touch, information from multiple sensory sources is integrated into a single event file. Under conditions of incongruent stimulation, the dissonance between proprioceptive information from the administering right hand (I am administering touch with my finger) and tactile information from the receptive left hand (I am being touched with a brush) is apparent in participants’ descriptions of their experience. They may agree with the statement ‘‘My left index finger felt as if it were touched by a finger and a brush simultaneously” or with the statement ‘‘My right index finger felt like a brush”. 3 We thank an anonymous reviewer for this observation. R.C. White et al. / Consciousness and Cognition 19 (2010) 505–519 519 Acknowledgments The authors would like to thank Mr Graham Thew for invaluable discussions, and two anonymous reviewers for helpful comments. References Armel, K. C., & Ramachandran, V. S. (2003). Projecting sensations to external objects: Evidence from skin conductance response. Proceedings of the Royal Society of London B, 21, 1499–1506. Botvinick, M., & Cohen, J. (1998). Rubber hands ‘feel’ touch that eyes see. Nature, 391, 756. Costantini, M., & Haggard, P. (2007). The rubber hand illusion: Sensitivity and reference frame for body ownership. Consciousness and Cognition, 16, 229–240. Dummer, T., Picot-Annand, A., Neal, T., & Moore, C. (2009). Movement and the rubber hand illusion. Perception, 38, 271–280. Durgin, F. H., Evans, L., Dunphy, N., Klostermann, S., & Simmons, K. (2007). Rubber hands feel the touch of light. Psychological Science, 18, 152–157. Ehrsson, H. H., Holmes, N. P., & Passingham, R. E. (2005). Touching a rubber hand: Feeling of body ownership is associated with activity in multisensory brain areas. The Journal of Neuroscience, 25, 10564–10573. Ehrsson, H. H., Spence, C., & Passingham, R. E. (2004). That’s my hand! Activity in premotor cortex reflects feeling of ownership of a limb. Science, 305, 875–877. Haans, A., IJsselsteijn, W. A., & de Kort, Y. A. W. (2008). The effect of similarities in skin texture and hand shape on perceived ownership of a fake limb. Body Image, 5, 389–394. Hägni, K., Eng, K., Hepp-Reymond, M.-C., Holper, L., Keisker, B., Siekierka, E., et al (2008). Observing virtual arms that you imagine are yours increases the galvanic skin response to an unexpected threat. Public Library of Science ONE, 3, e3082. Holmes, N. P., Crozier, G., & Spence, C. (2004). When mirrors lie: ‘Visual capture’ of arm position impairs reaching performance. Cognitive, Affective, and Behavioral Neuroscience, 4, 193–200. James, W. (1890). The principles of psychology. Cambridge, MA: Harvard University Press. Kinsbourne, M. (1995). Awareness of one’s own body: An attentional theory of its nature, development, and brain basis. In J. L. Bermúdez, A. Marcel, & N. Eilan (Eds.), The body and the self (pp. 205–224). England: MIT Press. Lloyd, D. M. (2007). Spatial limits on referred touch to an alien limb may reflect boundaries of visuo-tactile peripersonal space surrounding the hand. Brain and Cognition, 64, 104–109. Longo, M. R., Schüür, F., Kammers, M. P. M., Tsakiris, M., & Haggard, P. (2009). Self awareness and the body image. Acta Psychologica, 132, 166–172. Miles, E. (2008). What can the rubber hand illusion tell us about body representation in people with unexplained symptoms. In Symposium conducted at the Body Representation Workshop, University of Nottingham, UK. Moseley, G. L., Olthof, N., Venema, A., Don, S., Wijers, M., Gallace, A., et al (2008). Psychologically induced cooling of a specific body part caused by the illusory ownership of an artificial counterpart. Proceedings of the National Academy of Sciences, 105, 13169–13173. Mussap, A. J., & Salton, N. (2006). A ‘rubber-hand’ illusion reveals a relationship between perceptual body image and unhealthy body change. Journal of Health Psychology, 11, 627–639. Nielsen, T. I. (1963). Volition: A new experimental approach. Scandinavian Journal of Psychology, 4, 225–230. Oldfield, R. C. (1971). The assessment and analysis of handedness: The Edinburgh inventory. Neuropsychologia, 9, 97–113. Pavani, F., Spence, C., & Driver, J. (2000). Visual capture of touch: Out-of-body experiences with rubber gloves. Psychological Science, 11, 353–359. Pavani, F., & Zampini, M. (2007). The role of hand size in the fake-hand illusion paradigm. Perception, 36, 1547–1554. Ramachandran, V. S., & Hirstein, W. (1997). The three laws of qualia: What neurology tells us about the biological functions of consciousness, qualia and the self. Journal of Consciousness Studies, 4, 429–458. Ramachandran, V. S., & Hirstein, W. (1998). The perception of phantom limbs: The D. O. Hebb lecture. Brain, 121, 1603–1630. Schaefer, M., Flor, H., Heinze, H.-J., & Rotte, M. (2007). Morphing the body: Illusory feeling of an elongated arm affects somatosensory homunculus. NeuroImage, 36, 700–705. Schütz-Bosbach, S., Tausche, P., & Weiss, C. (2009). Roughness perception during the rubber hand illusion. Brain and Cognition, 70, 136–144. Shimada, S., Fukuda, K., & Hiraki, K. (2009). Rubber hand illusion under delayed visual feedback. Public Library of Science ONE, 4, e6185. Stevens, J. (1996). Applied multivariate statistics for the social sciences (3rd ed.). Mahwah, NJ: Lawrence Erlbaum. Tsakiris, M., & Haggard, P. (2005). The rubber hand illusion revisited: Visuotactile integration and self-attribution. Journal of Experimental Psychology: Human Perception and Performance, 31, 80–91. Tsakiris, M., Hesse, M. D., Boy, C., Haggard, P., & Fink, G. R. (2007). Neural signatures of body ownership: A sensory network for bodily self-consciousness. Cerebral Cortex, 17, 2235–2244. Tsakiris, M., Prabhu, G., & Haggard, P. (2006). Having a body versus moving your body: How agency structures body-ownership. Consciousness and Cognition, 15, 423–432.

Consciousness and Cognition Consciousness and Cognition 14 (2005) 316–326 www.elsevier.com/locate/concog Activating the critical lure during study is unnecessary for false recognition René Zeelenberga,*, Inge Bootb, Diane Pechera a Department of Psychology, Erasmus University Rotterdam, Woudestein J5-65, P.O. Box 1738, 3000 DR Rotterdam, The Netherlands b University of Leiden, The Netherlands Received 15 April 2004 Available online 23 November 2004 Abstract Participants studied lists of nonwords (e.g., froost, floost, stoost, etc.) that were orthographic-phonologically similar to a nonpresented critical lure, which was also a nonword (e.g., ploost). Experiment 1 showed a high level of false recognition for the critical lure. Experiment 2 showed that the false recognition effect was also present for forewarned participants who were informed about the nature of the false recognition effect and told to avoid making false recognition judgments. The present results show that false recognition effects can be obtained even when the critical lure itself is not stored during study. This finding is problematic for accounts that attribute false memories to implicit associative responses or spreading activation but is easily explained by global familiarity models of recognition memory. Ó 2004 Elsevier Inc. All rights reserved. Keywords: False recognition; Global Familiarity Models; Nonconscious spreading activation; Orthographic-phonological similarity; Retrieval 1. Introduction Human memory is not perfect and some conditions can give rise to highly inaccurate memories. One such condition that creates inaccurate memories is the DRM paradigm (Deese, 1959; Roedi* Corresponding author. Fax: +31 10 408 9009. E-mail address: zeelenberg@fsw.eur.nl (R. Zeelenberg). 1053-8100/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2004.08.004 R. Zeelenberg et al. / Consciousness and Cognition 14 (2005) 316–326 317 ger & McDermott, 1995). In the standard DRM paradigm, participants study lists of words (e.g., thread, pin, eye, etc.) that are semantically related to a nonpresented word, the critical lure (e.g., needle). Although the critical lure is not presented during study participants often falsely recognize it in an episodic recognition test. The percent ÔoldÕ responses to the critical lure is often as high as or not far below the percent ÔoldÕ responses to studied items. Such findings are obtained even for warned participants who are informed about the false memory effect and told not to make such errors (e.g., Gallo, Roediger, & McDermott, 2001). Moreover, participants are usually quite confident that the critical lure was presented during study (e.g., McDermott & Roediger, 1998). Thus, the false recognition effect is quite powerful and robust. Different explanations for false memories have been proposed. One explanation attributes false memories to implicit associative responses (e.g., Underwood, 1965; also see Lövdén & Johansson, 2003; McDermott, 1997; Roediger & McDermott, 1995). According to this account, when a word is studied a related word may consciously come to mind. For example, presentation of the words thread, pin, eye, etc. during study may elicit the covert verbal response needle and, consequently, needle may be stored in memory. On a later memory test, participants may think needle was actually presented on the study list if they fail to retrieve the correct source of the memory trace for needle. Recently, however, it has been argued that false memories may also be obtained when the critical lure does not consciously come to mind during study. At least two findings have been taken to support this claim. First, false memories have been found when list items are presented for 100 ms or less (e.g., McDermott & Watson, 2001; Seamon, Luo, & Gallo, 1998). At these presentation rates, it has been argued, it is unlikely that the critical lure will consciously come to mind. The second relevant finding was obtained by Seamon et al. (2002) who asked participants to think out loud and say the words that came to mind during study. Although participants spontaneously rehearsed about half the critical lures during study false memories were also found for critical lures that had not been rehearsed during study. Seamon et al. (2002) concluded that thinking of the critical lure was not necessary for producing false memories. Findings such as these have been explained by nonconscious spreading activation (e.g., Roediger, Balota, & Watson, 2002; Seamon et al., 1998). Spreading activation theories (e.g., Collins & Loftus, 1975) assume that when a word is presented activation automatically spreads to related words in a semantic-associative network, resulting in the activation of these words. In the DRM paradigm, multiple words presented during study are related to the nonstudied critical lure. Convergence of the activation spreading from these words may result in a high level of activation of the critical lure even though the critical lure was not presented. Assuming that this high level of activation has a long-lasting effect on memory this could lead to a large false recognition effect. To summarize, a currently popular view is that false memories can be obtained when the critical lure does not consciously come to mind during study. Instead, false memories may be due to nonconscious activation of the critical lure. In this paper, we take the issue one step further and ask the question whether false memories can be obtained even when the critical lure itself is not activated during study. At first this might seem unlikely but this possibility is suggested by global familiarity or global matching theories of recognition memory such as, for example, TODAM (Murdock, 1982), SAM (Gillund & Shiffrin, 1984), MINERVA2 (Hintzman, 1988) and REM (Shiffrin & Steyvers, 1997). According to global familiarity models (for a review, see Clark & Gronlund, 1996), a recognition judgment is based on the match between the test item and all (list) 318 R. Zeelenberg et al. / Consciousness and Cognition 14 (2005) 316–326 items in memory. Because the match depends on the similarity between the test item and the items in memory, participants are more likely to make an ÔoldÕ recognition judgment if (semantically, orthographically or phonologically) similar items were presented on the study list. In the standard DRM paradigm, many items on the study list are semantically similar to the critical lure and this will result in a relatively strong match between the lure and the items in memory. As a result, participants will be likely to think the critical lure was presented during study. To be a bit more specific, consider, for example, the REM model for recognition memory (we present a brief verbal description; mathematical details can be found in various sources, e.g., Criss & Shiffrin, 2004; Malmberg, Zeelenberg, & Shiffrin, 2004; Shiffrin & Steyvers, 1997). REM assumes that memory traces consist of features representing orthographic, phonological, semantic and contextual information. When an item is studied features are stored in an episodic memory trace.1 Recognition judgments are based on a comparison of the test item to the studied list items in episodic memory. The test item is matched in parallel to each memory trace and the system notes the matching and nonmatching features. Feature values in the test item that match corresponding feature values in an episodic trace contribute evidence that an item is old. Mismatching features contribute evidence that an item is new. Based on the matching and mismatching features the system calculates the odds that the test item is old (the odds equal the probability that the test item is old divided by the probability that the test item is new). If the odds exceed the criterion (the default criterion is 1.0, but participants could deviate from this value) an ÔoldÕ response is made. False alarms arise when the odds of a nonstudied item exceed the criterion for an ÔoldÕ response. This is more likely to happen when items similar to the nonstudied test item have been studied, because in this case there will be relatively many matching features. An important feature of global familiarity models is that they attribute false recognition to the matching process that takes place during the recognition test. Thus, global familiarity models need not assume that the critical lure was stored during study to account for false recognition. Hence, if we would find a false recognition effect under conditions in which it is unlikely that the critical lure was stored during study this would provide evidence supporting the global familiarity account of false recognition. 2. Experiment 1 To test the prediction that false recognition effects can be obtained even when the critical lure itself is not activated during study, participants studied nonwords (e.g., froost, floost, stoost, etc.) that were orthographic-phonologically similar to a nonpresented critical lure (e.g., ploost). Global familiarity models predict false recognition for a nonword that is orthographic-phonologically similar to multiple list items because the similarity will result in a relatively strong match between the nonword lure presented at test and the orthographic-phonologically similar nonwords in memory. Previously published studies (e.g., Schacter, Verfaillie, & Anes, 1997; Shiffrin, Huber, 1 Episodic memory traces are incomplete and error prone. That is, not all features of a stimulus are stored in the episodic trace and those features that are stored, may be stored incorrectly (i.e., an incorrect feature value is stored). The number of features stored and the accuracy with which they are stored depend on a number of factors such as study time, attention, hippocampal lesions, and the influence of psychopharmaca (e.g., Malmberg et al., 2004). R. Zeelenberg et al. / Consciousness and Cognition 14 (2005) 316–326 319 & Marinelli, 1995) have shown that false recognition can be obtained for critical lures (e.g., fate) that are orthographic-phonologically (instead of semantically) related to multiple list items (e.g., hate, mate, etc.). However, these studies have used existing words and activation may spread not only between semantically related words but also between orthographic-phonologically related words (e.g., Collins & Loftus, 1975; Sommers & Lewis, 1999). Therefore, these results do not show that false recognition can be obtained when the critical lure was not activated during study. In the present study, however, we used nonword stimuli. Because nonwords have no representation in lexical-semantic memory it is unlikely that activation will spread from one nonword to another (orthographic-phonologically related) nonword. It is also unlikely that the critical lure will consciously come to mind during study because prior to the test phase the critical lure has never been encountered by the participant. The finding of a false memory effect in the present study would therefore show that false recognition does not rely on the conscious or nonconscious storage of the critical lure during study. 2.1. Method 2.1.1. Participants The participants were 22 students of the Erasmus University Rotterdam and Leiden University who received course credit or a small monetary reward for their participation. All participants were native speakers of Dutch. None of them had previously participated in a related memory experiment. 2.1.2. Materials and design The materials consisted of 16 lists of orthographically legal, pronounceable nonwords.2 Each list consisted of a critical lure (e.g., ploost) and 12 list items (e.g., froost, floost, stoost, koost, noost, spoost, moost, soost, boost, poost, broost, and droost) that were all orthographic-phonologically similar to each other and the critical lure. The critical lure was not necessarily the item that was most similar to the other items on the list (i.e., the critical lure was not the ÔprototypeÕ or central item). The list items and critical lure differed from the other items on the same list by the addition, deletion or change of one or two letters. All items varied in length from 4 to 8 letters. For counterbalancing purposes the nonword lists were divided into two sets of eight lists, set A and set B. Half of the participants studied set A, the other half studied set B. The same test list was used for all participants, hence items that were targets for one half of the participants were distractors for the other half, and vice versa. 2.1.3. Procedure Participants were told that they would participate in an experiment that was about memory and were instructed to memorize as many items as possible for an unspecified memory test that would follow at the end of the experiment. No mention was made of the structure of the study list and the fact that many of the items were orthographic-phonologically similar to each other. 2 Pronunciation of a nonword is straightforward because there is a close correspondence between orthography and phonology in the Dutch language. Similar nonwords have been used in numerous lexical decision studies (e.g., de Groot, Thomassen, & Hudson, 1982; Zeelenberg & Pecher, 2003). 320 R. Zeelenberg et al. / Consciousness and Cognition 14 (2005) 316–326 During study, 96 nonwords (8 lists of 12 items) were presented in a randomly intermixed order (e.g., floost, zoes, trapel, praaf, froost, bleun, etc.). The critical lures were never studied. Each nonword was displayed for 3 s and followed by a 500 ms ISI. After presentation of the 96 nonwords, participants were given a recognition test in which they had to judge whether or not a test item had been presented during study. Participants were instructed to respond as accurately as possible and were informed that there was no time limit for responding. The test items consisted of three list items from each of the 16 lists that were used in the experiment, as well as the critical lure belonging to each list. Thus, of the 64 nonwords presented at test, 24 nonwords had been studied, 8 nonwords were the critical lures belonging to the 8 studied lists and the remaining 32 nonwords belonged to the lists that were not studied (24 list items and 8 critical lures). If participants thought they recognized a nonword as a studied item they were instructed to press the m-key of the keyboard (ÔoldÕ response) and if they did not recognize the nonword as a studied item they were instructed to press the z-key (ÔnewÕ response). All items in the recognition test were presented in a different random order for each participant and remained on the screen until a response was made. 2.2. Results and discussion The results of Experiment 1 are summarized in Table 1, which shows the percentage of ÔoldÕ responses as a function of study condition. Participants gave an ÔoldÕ response to 74.4% of the critical lures from studied lists and to only 17.0% of the critical lures from nonstudied lists. In other words, participant were much more likely to indicate erroneously that the nonword ploost had been studied if they had studied froost, floost, stoost, etc. than if they had not studied froost, floost, stoost, etc. This false memory effect was highly significant, t (21) = 11.37, p < .0001. Participants correctly recognized 86.4% of the list items from studied lists, while the false alarm rate for nonstudied list items was 14.6.%. This veridical memory effect was also significant, t (21) = 17.58, p < .0001. The finding of a false memory effect for nonwords that are orthographic-phonologically related to studied nonwords is consistent with the idea that activation of the critical lure during study is not a necessary condition for false recognition effects. In Experiment 2, we wanted to replicate and extend the findings of Experiment 1. False memory effects using the standard DRM paradigm have been obtained under a variety of study and test conditions. Notably, false recognition effects are present even when participants are informed about the nature of false memory experiments and told to avoid false memories (Gallo, Roberts, & Seamon, 1997; Gallo et al., 2001; McDermott & Roediger, 1998). The aim of Experiment 2 was to investigate if the false memory effect that we obtained for nonwords in Experiment 1 would still be present for such forewarned participants. Table 1 Percentage of ÔoldÕ responses in Experiment 1 as a function of study condition Condition Studied list Nonstudied list Critical lure List item 74.4 86.4 17.0 14.6 R. Zeelenberg et al. / Consciousness and Cognition 14 (2005) 316–326 321 3. Experiment 2 3.1. Method The participants were 18 students of the Erasmus University Rotterdam who received course credit or a small monetary reward for their participation. All participants were native speakers of Dutch and none of them had previously participated in a related memory experiment. The materials and procedure were identical to that of Experiment 1, except that prior to the study phase participants were informed about the false memory effect. Participants were told that prior research has shown that participants often show false memories for stimuli that are orthographic-phonologically related to studied list items and were given examples of stimuli that might cause a false memory effect. They were told to avoid making such errors. Immediately before test participants were reminded of the false memory phenomenon and told to respond ÔoldÕ only to items that were actually presented during study. 3.2. Results and discussion Table 2 summarizes the results of Experiment 2, displaying the percentage of ÔoldÕ responses on the recognition test as a function of study condition. Participants gave an ÔoldÕ response to 58.6% of the critical lures from studied lists and to only 12.5% of the critical lures from nonstudied lists. This false memory effect was again highly significant, t (17) = 6.48, p < .0001. The difference in the number of ÔoldÕ responses to studied and nonstudied list items was also significant, t (17) = 15.54, p < .0001, showing a veridical memory effect. Thus, even though participants were informed about the false memory effect and told to avoid false memories we still obtained a large false memory effect. The primary aim of Experiment 2 was to determine whether a false recognition effect would still be present in forewarned participants and not to determine exactly how warning participants affects the false memory effect. Nevertheless, it might be interesting to take a closer look at the effect of warning on performance in the recognition test. To compare performance in Experiments 1 and 2, we calculated A 0 , a measure of memory discrimination,3 and B00D , a measure of response criterion (Donaldson, 1992). When calculating A 0 and B00D for critical lures, an ÔoldÕ response to a critical lure from a studied list was considered a ÔhitÕ and an ÔoldÕ response to a critical lure from an unstudied list was considered a false alarm. The values of A 0 and B00D are displayed in Table 3. For critical lures, A 0 did not significantly differ between Experiment 1 and Experiment 2, t (38) = 1.38, p > .15. Likewise, for list items, A 0 did not significantly differ between Experiment 1 and Experiment 2, t (38) = 1.30, p > .15. Thus, warning participants did not affect A 0 for either false or veridical recognition.4 A different picture was obtained for B00D . For critical lures, B00D was significantly higher in Experiment 2 than in Experiment 1, t (38) = 3.86, p < .001. For list items too, B00D was significantly higher in Experiment 2 than in Experiment 1, t (38) = 3.57, p < .001. Thus, it seems that the main effect of forewarning was to make the participantsÕ recognition response criterion 3 4 For critical lures, A 0 indicates false recognition susceptibility or false recognition strength. The same result was obtained when we computed d 0 instead of A 0 . 322 R. Zeelenberg et al. / Consciousness and Cognition 14 (2005) 316–326 Table 2 Percentage of ÔoldÕ responses in Experiment 2 as a function of study condition Condition Studied list Nonstudied list Critical lure List item 54.2 71.3 11.1 10.0 Table 3 Memory discrimination (A 0 ) and response criterion ðB00D Þ in Experiments 1 and 2 B00D Condition A0 CL LI CL LI Experiment 1: no warning Experiment 2: forewarning .86 .81 .91 .89 .20 .85 .01 .58 Note. CL, critical lure; LI, list item. more conservative. In other words, forewarned participants were overall (i.e., for both critical lures and list items) less likely to make an ÔoldÕ response. It is interesting to note that in terms of A 0 the false recognition effect was about equally large in Experiments 1 and 2. Previous studies (Gallo et al., 2001; Neuschatz, Benoit, & Payne, 2003) using semantically related lists of words have shown that forewarning participants can sometimes substantially reduce the size of the false recognition effect (although false memories are still present for such forewarned participants). It might be worth speculating why warning was not very effective in our study. One possibility is that participants in our study were not able to identify the critical lure. A recent study by Neuschatz et al. (2003) showed that forewarning was much more effective for high identifiable DRM lists (i.e., lists for which a high proportion of participants can identify the critical lure) than for low identifiable DRM lists. If, during study, participants identify the critical lure and tag it as ÔnonpresentedÕ they may use this information at test to withhold an erroneous ÔoldÕ response. In our study, however, the critical lure was probably not highly identifiable. Because there is a large number of nonwords that could potentially be the critical lure (remember that the critical lure was not the ÔprototypeÕ or central item), the probability of identifying the lure is probably rather small and hence it could not be tagged as ÔnonpresentedÕ. Of course this explanation is somewhat speculative and must await further experimentation. The main point of Experiment 2, however, is that even when participants are forewarned false recognition effects are still obtained. 4. General discussion In two experiments, we investigated false recognition for nonwords that were orthographicphonologically similar to nonwords presented at study. Experiment 1 showed that participants more often erroneously indicated that they had studied the critical lure (e.g., ploost) if they had studied orthographic-phonologically similar list items (e.g., froost, floost, stoost, etc.) than if they had not studied orthographic-phonologically similar list items. Experiment 2 showed that the false recognition effect persisted even for forewarned participants indicating that, like the false recog- R. Zeelenberg et al. / Consciousness and Cognition 14 (2005) 316–326 323 nition effect obtained in the standard DRM paradigm for semantically related lists of words, the false recognition effect for nonwords is powerful and robust. The results of the present study are predicted by global familiarity models of recognition memory (e.g., Gillund & Shiffrin, 1984; Hintzman, 1988; Murdock, 1982; Shiffrin & Steyvers, 1997). These models assume that the test item is matched in parallel to all studied items in memory. Because the familiarity of the test item depends on the similarity between the test item and the items in memory test items that are similar to multiple studied items will be highly familiar. Hence, participants will be likely to think the test item was studied even though it was not. The present results are more difficult to explain, however, by the implicit associative response account and extant spreading activation accounts, because these accounts assume that the critical lure itself is stored during study. According to the implicit associative response account, false memories occur because the critical lure consciously comes to mind during study of the list items. This seems unlikely to happen, however, for nonwords that have never been encountered by the participant. Likewise, because nonwords do not have representations in lexical-semantic memory it is difficult to see how activation can spread from the list items to the critical lure.5 Proponents of the spreading activation account might argue that although the present results are problematic for spreading activation accounts of false recognition they do not show that spreading activation is not responsible for false recognition in the standard DRM paradigm in which participants study lists of words that are semantically related to a nonstudied critical lure. The standard DRM paradigm differs from the present study in that stimuli are semantically related instead of orthographic-phonologically and in that words are used instead of nonwords. It is, of course, possible that different mechanisms are responsible for the false recognition effects for different types of stimuli, but it does not strike us as a very attractive explanation. First, it should be noted that the spreading activation mechanism has been used as an explanation for false recognition of orthographic-phonologically related words (e.g., Roediger et al., 2002; Sommers & Lewis, 1999). Second, and maybe more important, global familiarity models can explain false recognition for different types of stimuli, not only for words and nonwords, but also for other stimuli, such as, for example, faces (e.g., Criss & Shiffrin, 2004). As long as a nonstudied test item is similar to studied list items on a relevant dimension global familiarity models predict a false recognition effect. Thus, global familiarity models provide a simple and parsimonious account of false recognition for different types of stimuli and different types of similarity. 5 Although the present results cannot be explained by extant spreading activation accounts of false memories, one might argue they could possibly be explained by a modified spreading activation account that assumes sublexical representations (cf. Dorfman, 1994; Treiman, Mullennix, Bijeljac-Babic, & Richmond-Welty, 1995). Such an explanation would have to assume that during study activation converges on sublexical components. If these sublexical components are subsequently accessed during test they may affect memory performance. One important question would be what kind of sublexical components mediate the false memory effect. Since most stimuli used in the present study consisted of a single syllable it is unlikely that the effect is mediated by syllables (or morphemes) common to the stimuli on the lists. The false memory effect could possibly be mediated, however, by orthographic or phonological body units. It should be noted that such an explanation differs from the explanations that have been put forward to explain false memory effects and it remains to be seen whether it can account for orthographicphonologically based false memory effects (as well as veridical memory effects) for both words and nonwords. Most important for the present study, however, is that such a modified spreading activation account does not assume that the critical lure itself has to be activated to obtain false memory effects and hence it is consistent with the main claim of the present study. 324 R. Zeelenberg et al. / Consciousness and Cognition 14 (2005) 316–326 As we mentioned in Section 1, nonconscious spreading activation accounts of false memories have been proposed to explain findings that are supposedly not easily explained by conscious activation of the critical lure during study. One such finding was reported by Seamon et al. (2002) who found false memory effects for lures not rehearsed during study. In another study, Seamon et al. (1998) obtained a false recognition effect with very fast presentation rates during study. According to Seamon and colleagues these results indicate that false memories can be obtained even when the critical lure is not consciously activated during study (but see Raaijmakers & Zeelenberg, 2004; Zeelenberg, Plomp, & Raaijmakers, 2003). But do these results necessitate an explanation in terms of nonconscious storage of the critical lure in memory? The answer is that these results can also be explained by global familiarity models of recognition memory. The only thing we need to assume is that in these studies some semantic features of the list items were stored in memory. This will result in a relatively strong match at test between the critical lure and the semantically related list items in memory. Hence, a higher false alarm rate is predicted for a critical lure belonging to a studied list than for a critical lure belonging to a nonstudied list. Another finding that has been explained by spreading activation is that false memory effects increase with the number of semantically related list items on the study list. Such effects are also predicted, however, by global familiarity accounts because the overall match between the critical lure and the list items will be stronger the more items on the study list are similar to the lure (see Arndt & Hirshman, 1998). To conclude, the results of the present study show that global familiarity models provide a viable account of false recognition. The present paper is, of course, not the only one to mention the ability of global familiarity models to account for false recognition. In their original paper, Roediger and McDermott (1995) already discussed the ability of global familiarity models to account for false recognition effects. Moreover, some studies have been explicitly designed to test the predictions of global familiarity models (e.g., Arndt & Hirshman, 1998). Nevertheless, global familiarity models are often overlooked when explaining false recognition. The present study shows that global familiarity models should be taken into account when trying to explain false memories. In our view it would be interesting if future studies of the false memory effect would try to test the predictions of global familiarity accounts against those of alternative accounts such as spreading activation. Acknowledgments We thank Jeroen Raaijmakers, Mark Rotteveel and Eric-Jan Wagenmakers for helpful discussions. References Arndt, J., & Hirshman, E. (1998). True and false recognition in MINERVA2: Explanations of a global matching perspective. Journal of Memory and Language, 39, 371–391. Clark, S. E., & Gronlund, S. D. (1996). Global matching models of recognition memory: How the models match the data. Psychonomic Bulletin & Review, 3, 37–60. R. Zeelenberg et al. / Consciousness and Cognition 14 (2005) 316–326 325 Collins, A. M., & Loftus, E. F. (1975). A spreading activation theory of semantic memory. Psychological Review, 82, 407–428. Criss, A. H., & Shiffrin, R. M. (2004). Context noise and item noise jointly determine recognition memory: A comment on Dennis and Humphreys (2001). Psychological Review, 111, 800–807. Deese, J. (1959). On the prediction of the occurrence of particular verbal intrusions of words in immediate recall. Journal of Experimental Psychology, 58, 17–22. de Groot, A. M. B., Thomassen, A. J. W. M., & Hudson, P. T. M. (1982). Associative facilitation of word recognition as measured from a neutral prime. Memory & Cognition, 10, 358–370. Donaldson, W. (1992). Measuring recognition memory. Journal of Experimental Psychology: General, 121, 275–277. Dorfman, J. (1994). Sublexical components in implicit memory for novel words. Journal of Experimental Psychology: Learning, Memory, and Cognition, 5, 1108–1125. Gallo, D. A., Roberts, M. J., & Seamon, J. G. (1997). Remembering words not presented in lists: Can we avoid creating false memories. Psychonomic Bulletin & Review, 4, 271–276. Gallo, D. A., Roediger, H. L., III, & McDermott, K. B. (2001). Associative false recognition occurs without criterion shifts. Psychonomic Bulletin & Review, 8, 579–586. Gillund, G., & Shiffrin, R. M. (1984). A retrieval model for both recognition and recall. Psychological Review, 91, 1–67. Hintzman, D. (1988). Judgments of frequency and recognition memory in a multiple trace memory model. Psychological Review, 95, 528–551. Lövdén, M., & Johansson, M. (2003). Are covert verbal responses mediating false implicit memory?. Psychonomic Bulletin & Review, 10, 724–729. Malmberg, K. J., Zeelenberg, R., & Shiffrin, R. M. (2004). Turning up the noise or turning down the volume? On the nature of the impairment of episodic recognition memory by midazolam. Journal of Experimental Psychology: Learning, Memory, and Cognition, 30, 540–549. McDermott, K. B. (1997). Priming on perceptual implicit tests can be achieved through presentation of associates. Psychonomic Bulletin & Review, 4, 582–586. McDermott, K. B., & Roediger, H. L. III, (1998). Attempting to avoid illusory memories: Robust false recognition of associates persists under conditions of explicit warnings and immediate testing. Journal of Memory and Language, 39, 508–520. McDermott, K. B., & Watson, J. M. (2001). The rise and fall of false recall: The impact of presentation duration. Journal of Memory and Language, 45, 160–176. Murdock, B. B. (1982). A theory for the storage and retrieval of item and associative information. Psychological Review, 89, 609–626. Neuschatz, J. S., Benoit, G. E., & Payne, D. G. (2003). Effective warnings in the Deese–Roediger–McDermott falsememory paradigm: The role of identifiability. Journal of Experimental Psychology: Learning, Memory, and Cognition, 29, 35–41. Raaijmakers, J. G. W., & Zeelenberg, R. (2004). Evaluating the evidence for nonconscious processes in producing false memories. Consciousness and Cognition, 13, 169–172. Roediger, H. L., & McDermott, K. B. (1995). Creating false memories: Remembering words not presented in lists. Journal of Experimental Psychology: Learning, Memory and Cognition, 21, 803–814. Roediger, H. L., Balota, D. A., & Watson, J. M. (2002). Spreading activation and arousal of false memories. In H. L. Roediger, J. S. Nairne, I. Neath, & A. M. Surprenant (Eds.), The nature of remembering: Essays in the honor of Robert G. Crowder (pp. 95–115). Washington, DC: American Psychological Association. Seamon, J. G., Luo, C. R., & Gallo, D. A. (1998). Creating false memories of words with or without recognition of list items: Evidence for nonconscious processes. Psychological Science, 9, 20–26. Seamon, J. G., Lee, I. A., Toner, S. K., Wheeler, R. H., Goodkind, M. S., & Birch, A. D. (2002). Thinking of critical words during study is unnecessary for false memories in the Deese, Roediger, and McDermott procedure. Psychological Science, 13, 526–531. Schacter, D. L., Verfaillie, M., & Anes, M. D. (1997). Illusory memories in amnesic patients: Conceptual and perceptual false memories. Neuropsychology, 11, 331–342. Shiffrin, R. M., Huber, D. E., & Marinelli, K. (1995). Effects of category length and strength on familiarity in recognition. Journal of Experimental Psychology: Learning, Memory, and Conition, 21, 267–287. 326 R. Zeelenberg et al. / Consciousness and Cognition 14 (2005) 316–326 Shiffrin, R. M., & Steyvers, M. (1997). A model for recognition memory: REM-retrieving effectively from memory. Psychonomic Bulletin & Review, 4, 145–166. Sommers, M. S., & Lewis, B. P. (1999). Who really lives next door: Creating false memories with phonological neighbors. Journal of Memory and Language, 40, 83–108. Treiman, R., Mullennix, J., Bijeljac-Babic, R., & Richmond-Welty, E. D. (1995). The special role of rimes in the description, use, and acquisition of English orthography. Journal of Experimental Psychology: General, 2, 107–136. Underwood, B. J. (1965). False recognition produced by implicit associative verbal responses. Journal of Experimental Psychology, 70, 122–129. Zeelenberg, R., & Pecher, D. (2003). Evidence for long-term cross-language repetition priming in conceptual implicit memory tasks. Journal of Memory and Language, 49, 80–94. Zeelenberg, R., Plomp, G., & Raaijmakers, J. G. W. (2003). Can false memories be created through nonconscious processes?. Consciousness and Cognition, 12, 403–412.