Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Autism Spectrum Disorders and Schizophrenia: Meta-Analysis of the Neural Correlates of Social Cognition

  • Gisela Sugranyes,

    Affiliations Department of Child and Adolescent Psychiatry, Institute of Psychiatry, King's College London, London, United Kingdom, Section of Neurobiology of Psychosis, Department of Psychosis Studies, Institute of Psychiatry, King's College London, London, United Kingdom, Department of Child and Adolescent Psychiatry and Psychology, Institute of Neuroscience, University of Barcelona, Barcelona, Spain

  • Marinos Kyriakopoulos,

    Affiliations Section of Neurobiology of Psychosis, Department of Psychosis Studies, Institute of Psychiatry, King's College London, London, United Kingdom, Child and Adolescent Mental Health Services, South London and Maudsley NHS Foundation Trust, London, United Kingdom

  • Richard Corrigall,

    Affiliation Child and Adolescent Mental Health Services, South London and Maudsley NHS Foundation Trust, London, United Kingdom

  • Eric Taylor,

    Affiliation Department of Child and Adolescent Psychiatry, Institute of Psychiatry, King's College London, London, United Kingdom

  • Sophia Frangou

    Affiliation Section of Neurobiology of Psychosis, Department of Psychosis Studies, Institute of Psychiatry, King's College London, London, United Kingdom



Impaired social cognition is a cardinal feature of Autism Spectrum Disorders (ASD) and Schizophrenia (SZ). However, the functional neuroanatomy of social cognition in either disorder remains unclear due to variability in primary literature. Additionally, it is not known whether deficits in ASD and SZ arise from similar or disease-specific disruption of the social cognition network.


To identify regions most robustly implicated in social cognition processing in SZ and ASD.

Data Sources

Systematic review of English language articles using MEDLINE (1995–2010) and reference lists.

Study Selection

Studies were required to use fMRI to compare ASD or SZ subjects to a matched healthy control group, provide coordinates in standard stereotactic space, and employ standardized facial emotion recognition (FER) or theory of mind (TOM) paradigms.

Data Extraction

Activation foci from studies meeting inclusion criteria (n = 33) were subjected to a quantitative voxel-based meta-analysis using activation likelihood estimation, and encompassed 146 subjects with ASD, 336 SZ patients and 492 healthy controls.


Both SZ and ASD showed medial prefrontal hypoactivation, which was more pronounced in ASD, while ventrolateral prefrontal dysfunction was associated mostly with SZ. Amygdala hypoactivation was observed in SZ patients during FER and in ASD during more complex ToM tasks. Both disorders were associated with hypoactivation within the Superior Temporal Sulcus (STS) during ToM tasks, but activation in these regions was increased in ASD during affect processing. Disease-specific differences were noted in somatosensory engagement, which was increased in SZ and decreased in ASD. Reduced thalamic activation was uniquely seen in SZ.


Reduced frontolimbic and STS engagement emerged as a shared feature of social cognition deficits in SZ and ASD. However, there were disease- and stimulus-specific differences. These findings may aid future studies on SZ and ASD and facilitate the formulation of new hypotheses regarding their pathophysiology.


Autism and psychotic disorders have historically been considered as related diagnostic entities. One of the innovations of the DSM-III in the late 1970s was to separate autism spectrum disorders (ASD) from schizophrenia (SZ) into different diagnostic categories. Although this distinction has many practical advantages, it is currently being reconsidered in view of emerging evidence about common neurobiological processes in both disorders [1][3].

Impairment in social cognition is a cardinal feature of the clinical presentation of both ASD and SZ [4][8]. The term social cognition refers to a complex set of processes subserving adaptive social interaction which depend on “theory of mind”, or in other words, the ability to make correct attributions of the mental states of others [9][10]. Theory of mind (ToM) broadly refers to three types of attributions: attribution of epistemic mental states (e.g. beliefs), attribution of intentions or motivations and attribution of affective states. A range of tasks have been developed to map on these core mentalising domains. Facial emotion recognition (FER) relates to the ability to infer the emotional state of others, and although it measures a core dimension of ToM, it is usually mentioned separately. We will follow this convention here for ease of reference to the existing literature. Similarly, we will use the term ToM to collectively refer to tasks of epistemic (e.g. false beliefs) or intention attribution and tasks that involve more than one ToM domain (e.g. the attribution of mental states and intentions to animated geometric figures).

Neurobiological models of social cognition implicate an extended neural network in processing social stimuli [11][15]. Regions most consistently involved are the medial and ventrolateral prefrontal cortex (PFC), ventral temporal regions around the superior temporal sulcus (STS), occipitotemporal regions (particularly within the fusiform gyrus), the temporoparietal junction (TPJ), and limbic structures, especially the amygdala. These regions are interconnected and have additional connectivity with somatosensory cortices and subcortical structures such as the thalamus. There is emerging consensus for relative regional specialization within this extended network. Involvement of the medial PFC in social cognition is elicited by multiple tasks that require conscious attribution or judgment of mental states, dispositional traits or intentions of one's self or of others [16]. Engagement of the ventrolateral PFC relates primarily to the contextual or social appropriateness of responses to social cues [17][19]. Regions surrounding the TPJ have systematically been associated with social cognition tasks requiring participants to “think about other people's thoughts” or in other words, to take a third person perspective about others' affective or cognitive states [20], [21]. Activation in regions around the STS has been reliably associated with salient biological movement, including changeable characteristics of human facial features, which can be used to infer affective or intentional states [22], [23]. Similarly, somatosensory cortices are thought to contribute to social cognition by invoking or “mirroring” internal representations of affective states [24]. The amygdala is thought to contribute to social cognition by mediating arousal [25] or biological salience [26] associated with stimuli.

Current reviews of the neural basis of social cognition deficits in SZ [27], [28] and ASD [29] have implicated most, if not all the regions within this network. However, the current literature has two significant limitations. Firstly, existing reviews, which are mostly narrative, highlight the variability in the findings from individual studies, but do not provide an integrated model of the functional neuroanatomy of social cognition in either SZ or ASD. Secondly, there are no neuroimaging studies to date directly comparing patients with ASD or SZ, with the single exception of Pinkham et al [30]. The authors compared SZ and ASD patients while performing a functional magnetic resonance imaging (fMRI) task requiring participants to judge the trustworthiness of human faces; both patient groups showed reduced activation in the amygdala and ventrolateral PFC.

The aims of this study were to synthesize existing fMRI data using a meta-analytic approach in order to (a) identify regions most robustly implicated in social cognition processing in SZ and ASD and (b) to draw inferences about differences and similarities in the neural correlates of social cognition between the two disorders. Our key predictions were that during tasks of social cognition both disorders would be associated with reduced engagement within PFC regions associated with mentalising, and that similar functional disruption would also be observed in limbic regions, particularly the amygdala. We also hypothesized that in SZ, PFC dysfunction would be associated with reduced down-regulation of more posterior brain regions involved in the attribution of salience to biological social cues (e.g. facial affect).

Materials and Methods

Data sources and inclusion criteria

The study design and report adhered to the PRISMA Statement guidelines (supporting information Table S1 and Figure S1). The search method and inclusion criteria were specified in advance, informed by existing meta-analyses [1], [31], [32]. All identified articles were reviewed for eligibility by at least two authors, and decisions for inclusion were made by consensus. Data was extracted independently from each study by the first author, and was subsequently reviewed by a second author.

Studies investigating FER or ToM in subjects with ASD (including autism, Asperger's Syndrome and pervasive developmental disorder not otherwise specified) or SZ were identified through a computerized literature search using Medline. We reviewed all papers in English language published up to 2010. The following search keywords were employed: “autism”, “schizophrenia”, “asperger”, and “facial emotion”, “emotional processing”, “social cognition”, “theory of mind”, “mentalization”, “irony”, “empathy”, “fMRI” and their combinations and differing terminations, as well as terms specifying individual facial affect (fear, happiness, sadness, anger and disgust). The reference lists of these papers were searched for additional articles.

In order to meet inclusion criteria, studies were required to (a) report comparisons between ASD or SZ patients and matched healthy controls, (b) employ functional magnetic resonance imaging (fMRI), (c) use image subtraction methodology to identify foci of task-related neural changes contrasting an active and control condition, and (d) report their results in standard stereotactic coordinates (either Talairach or Montreal Neurological Institute space). Studies not fulfilling these requirements were excluded. No age, gender or treatment restrictions were applied.

Additionally, facial affect processing studies were included if they used human facial identities as stimuli in both active and control condition (as opposed to geometric shapes) and were excluded if they used facial stimuli to investigate processes not directly involved in emotional processing, such as working memory or attention. As the majority of FER studies used negatively valence facial expressions (anger, fear, sadness) we focused exclusively on these to minimize heterogeneity due to valence [33].

Information extracted from each included: a) foci of task-related neural changes contrasting an active and control condition between either patient group (ASD or SZ) and healthy controls b) a description of each task and the selected contrast c) clinical Information concerning age, gender, diagnosis, symptom scales, illness duration and treatment, and matching of the healthy control group.

Quantitative meta-analytical voxel-based procedure

We used one contrast from each study as shown in Tables 1 and 2, and we accepted results as significant based on the threshold employed in the original studies.

Table 1. Details of studies included in the meta-analysis for facial emotion recognition tasks.

Table 2. Clinical description of patients with Autism Spectrum Disorders included in facial emotion recognition studies.

Activation Likelihood Estimation procedure

Coordinates from FER and ToM studies were analysed separately following the Activation Likelihood Estimation (ALE) technique implemented in GingerALE 2.0.4 ( This version uses a random effect model and weighting for sample size of the original studies [34]. Coordinates of the foci of activation reported in the original studies were transformed into Talairach space using the Lancaster transform (icbm2tal tool) in GingerALE. For each study, peaks were modelled as the centre of a 3D Gaussian distribution and a modelled activation (MA) map was then computed. ALE scores from the convergent MA maps were then calculated on a voxel-by-voxel basis to test for convergent (random-effects) rather than study specific foci (fixed-effects). All ALE data processing was performed using the BrainMap Search and View software ( The threshold of statistical significance was set at p<0.05, with False Discovery Rate (FDR) correction for multiple comparisons and a minimum cluster size of 80 mm3. Each ALE map was imported into Mango ( and overlaid on an anatomical Template ( for representation purposes. Significant clusters were manually localized and Brodmann areas (BA) were identified using the standard Talairach and Tournoux [35] stereotactic anatomic brain atlas. Finally, in order to compare diagnostic groups, the ALE Maps for SZ and ASD for each experimental design (e.g. ALE map for ASD FER vs ALE map for SZ FER) were directly contrasted using subtraction meta-analysis procedure implemented in GingerALE version 1.1 [34]. Regions in which the two diagnostic groups differed were defined at p<0.05 with FDR correction and a minimum cluster size of 200 mm3.

Demographic data analysis

Student t tests (two-tailed) or chi-square tests were used to compare the distribution of continuous and categorical demographic variables between the two groups. Only variables which displayed group differences were subjected to voxel-wise meta-regression analyses in order to examine their effect on the ALE results.


The initial search returned 415 citations. Of these, 337 studies were discarded after reviewing the abstracts while the full text of the remaining 76 citations was examined in more detail. Thirty three studies fulfilled the inclusion criteria and were included in the meta-analysis. Details are shown in Tables 1 and 2.

Facial emotion recognition

We identified 16 FER studies in ASD and 33 in SZ, of which 5 and 12 respectively were used in the ALE analysis. The total sample comprised 55 ASD and 203 SZ patients and 253 healthy controls (HC) (Table 1). Demographic details for all participants and clinical information of ASD and SZ patients are shown in Tables 2 and 3. Mean age did not differ between diagnostic groups (ASD: mean = 31.96, SD = 4.06; SZ: mean = 31.41, SD = 8.47; t = 0.13, df = 16, p = 0.89,), but there was an over-representation of males in the ASD studies (ASD: 100% males; SZ: 74.93% males; chi-square = 20.4, df = 1, p<0.001).

Table 3. Clinical description of patients with Schizophrenia included in facial emotion recognition studies.

Autism spectrum disorders vs. healthy controls.

During the emotional faces>neutral faces contrast (ASD = 34, HC = 34, foci = 9), ASD patients compared to HC showed increased activation in temporal regions near the STS and decreased activation in the primary somatosensory cortex, within the postcentral gyrus. The corresponding anatomical regions and peak ALE maxima are show in Table 4 and Figure 1.

Figure 1. Activation likelihood estimation maps comparing autism spectrum disorder subjects with healthy controls.

Clusters of relative underactivation or overactivation in ASD are shown in blue and red respectively; numbers represent the sagittal (x), coronal (y) and axial (z) coordinates of each slice in Talairach space. p<0.05 False Rate Discovery corrected for multiple comparisons. ASD = Autism Spectrum Disorders, HC = Healthy Controls.

Table 4. Activation likelihood estimation results for facial emotion recognition in patients with schizophrenia or autism spectrum disorders compared to healthy controls.

Schizophrenia vs. healthy controls.

During the same contrast (SZ = 77, HC = 77, foci = 16), SZ patients compared to HC showed decreased activation in the ventrolateral PFC, posterior cingulate cortex, amygdala, occipito-temporal regions (encompassing the fusiform gyrus) and thalamus (Table 4, Figure 2).

Figure 2. Activation likelihood estimation maps comparing schizophrenia patients with healthy controls.

Clusters of relative underactivation or overactivation in SZ are shown in blue and red respectively; numbers represent the sagittal (x), coronal (y) and axial (z) coordinates of each slice in Talairach space. p<0.05 False Rate Discovery corrected for multiple comparisons. SZ = Schizophrenia, HC = Healthy Controls.

Comparison of autism spectrum disorders and schizophrenia.

ASD subjects were significantly more likely than SZ to engage temporal regions near the STS and in the anterior and posterior cingulate cortex. In the inverse contrast, SZ patients expressed greater likelihood than ASD to engage the ventrolateral PFC, the parahippocampal gyrus and regions within the TPJ, inferior occipital gyrus and the cerebellum (Table 5, figure 3).

Figure 3. Activation likelihood estimation maps comparing autism spectrum disorder with schizophrenia patients.

Clusters of relative overactivation and underactivation in ASD are shown in green and yellow respectively; numbers represent the sagittal (x), coronal (y) and axial (z) coordinates of each slice in Talairach space. p<0.05 False Rate Discovery corrected for multiple comparisons. ASD = Autism Spectrum Disorders, SZ = Schizophrenia.

Table 5. Activation likelihood estimation results for facial emotion recognition comparison of schizophrenia and autism spectrum disorders.

Theory of mind

Our search identified 10 studies exploring ToM related processes in ASD subjects and 17 in SZ patients, of which 7 and 9 respectively fulfilled inclusion criteria (Table 6). The total analysis encompassed 91 ASD and 133 SZ patients, and 239 HC subjects. Demographic details for all participants and clinical information corresponding to ASD and SZ patients are shown in Tables 7 and 8. ASD subjects were significantly younger than SZ (ASD: mean = 18.55, SD = 7.03; SZ: mean = 31.36, SD = 4.08; t = 4.1, df = 13, p = .001), and included a higher percentage of males (ASD: mean = 95.2%,; SZ: mean = 63.0%, chi-quare = 47.01, df = 1, p<0.001).

Table 6. Details of studies included in the meta-analysis for theory of mind tasks.

Table 7. Clinical description of patients with Autism Spectrum Disorders included in theory of mind studies.

Table 8. Clinical description of patients with Schizophrenia included in theory of mind studies.

Autism spectrum disorders vs. healthy controls.

In the contrast ToM>control condition (ASD = 88, HC = 98, foci = 33), ASD subjects compared to HC showed decreased activation in six clusters located in the medial prefrontal and in the anterior cingulate cortex, the amygdala and in primary and secondary somatosensory areas in the precentral gyrus, the STS, and in the TPJ within the inferior parietal lobule (Table 9, Figure 1).

Table 9. Activation likelihood estimation results for theory of mind in patients with schizophrenia or autism spectrum disorders compared to healthy controls.

Schizophrenia vs. healthy controls.

In the same contrast (SZ = 133, HC = 140, foci = 43) SZ patients compared to HC showed (a) increased activation in the posterior cingulate cortex and within the somatosensory cortices in the paracentral lobule, and superior temporal gyrus and (b) decreased activation in the medial prefrontal PFC, middle temporal gyrus and the thalamus (Table 9, Figure 2).

Comparison of autism spectrum disorders and schizophrenia.

ASD was associated with greater likelihood of activation in the insula compared to SZ (Table 10). In the inverse contrast, SZ was associated with greater likelihood of engagement in regions within the medial PFC and somatosensory areas in the paracentral gyrus, and in the posterior cingulate cortex (Table 8; Figure 3). The effect size of signal differences in the medial PFC region was positively correlated with age (peak coordinates = −26, 50, 2; r2 = 0.28; p = 0.001; voxels = 59).

Table 10. Activation likelihood estimation results for theory of mind tasks comparison of schizophrenia and autism spectrum disorders.

Examination of confounding factors


The effect size of signal differences in the medial PFC region was positively correlated with age (peak coordinates = −26, 50, 2; r2 = 0.28; p = 0.001; voxels = 59). This would suggest that some of the differences observed between ASD and SZ in this area may be also modulated by age.

Antipsychotic medication.

Details of pharmacological treatments are displayed in tables 2 and 4. The mean dose of antipsychotic drugs was converted into chlorpromazine equivalent doses. Studies providing information on the mean daily chlorpromazine equivalent dose of antipsychotics for their sample were included in voxel-wise meta-regressions exploring the relationship between dose and magnitude of ALE scores. Antipsychotic medications were found to have a normalizing effect on amygdala involvement during FER. The effect size of signal differences in the amygdala was negatively correlated with dose of antipsychotics (peak coordinates: [30, −2, −18]; r2 = 0.48; p = 0.001; number of voxels = 45), and therefore it is unlikely that the differences observed between ASD and SZ are due to the effects of antipsychotic medication. With regards to ToM, we also found an ameliorative effect of medication on the activation in the superior temporal gyrus (peak voxels: [46, 10, −24]; r2 = 0.48; p = 0.001; number of voxels = 237; and [44, 16, 12]; r2 = 0.27; p = 0.001; number of voxels = 171), which we also interpret as unlikely to be the source of the group differences.


We used quantitative analysis to explore the neural correlates of social cognition during facial emotion recognition (FER) and theory of mind (ToM) tasks in ASD and SZ. Differential effects of task and diagnosis were noted in the distribution and direction of the findings. Our key findings can be summarised as follows: (a) Both SZ and ASD patients showed reduced engagement in medial PFC regions, with the extent and degree of deficit being greater in ASD; conversely, ventrolateral PFC disruption was associated mostly with SZ, (b) SZ was associated with hyperactivation in somatosensory cortices, while ASD subjects showed reduced activity in these regions, (c) amygdala hypoactivation was observed in SZ patients during facial emotion processing and in ASD during more complex ToM tasks, (d) both SZ and ASD were associated with decreased engagement of regions around the STS during ToM tasks, but when processing facial affect, overactivation in these regions was noted in ASD, (e) reduced thalamic engagement was uniquely associated with SZ.

Prefrontal and midline cortical involvement

We found both similarities and differences in prefrontal engagement in ASD and SZ. During FER tasks, SZ patients were significantly more likely than ASD to engage the ventrolateral PFC, but the degree of activation was less than in controls. Both SZ and ASD were associated with hypoactivation in medial PFC regions during ToM tasks, although the deficits were quantitatively greater in ASD and extended more posteriorly into the anterior and posterior cingulate cortex.

The role of the ventrolateral PFC in social cognition has begun to be more fully appreciated as it is involved in multiple tasks of social cognition [17]. This region is thought to modulate the influence of emotional stimuli on cognition with respect to contextually (or socially) appropriate behaviour [12], [18], [19], [36]. Our findings of hypoactivation in SZ patients in this region reinforce the results of Leitman and colleagues [37] who demonstrated reduced contextual modulation of the ventrolateral PFC in SZ in connection to impaired affective appraisal. In contrast, activation in the medial prefrontal cortex is thought to underlie emotion generation, particularly when assessing self-relevant attributes or emotional awareness [38][40]. Therefore our results imply that reduced PFC engagement in connection to the social appropriateness of behaviour may be more relevant to SZ, while in both SZ and ASD, reduced medial PFC engagement may contribute to abnormalities in conscious awareness of the emotional states of others.

Somato-sensory cortices

Activation in primary and secondary somato-sensory cortices was reduced in ASD, in comparison to controls and SZ patients, during both FER and ToM tasks. Engagement of somato-sensory cortices is thought to invoke (“mirror”) bodily states associated with pertinent emotions or other internal states, thus facilitating their recognition in oneself or others [11], [41], [42]. Our findings in ASD are consistent with reports of abnormal cortical thickness and grey matter volume patterns in the corresponding anatomical regions [43], [44] and lend further support to the notion that ASD is associated with dysfunction in invoking “mirroring” mechanisms when processing social stimuli [45]. Our evidence of increased involvement of somato-sensory regions during social cognitive tasks in SZ adds to similar findings [46], [47], and may contribute to the over-attribution of agency and “hypermentalization” described in schizophrenia [48], [49].


Decreased engagement of the amygdala was observed in SZ patients during FER tasks, which confirms findings from earlier meta-analytic studies [28], [50]. However, Anticevic et al [50] have reported that in SZ there is increased amygdala activation to neutral stimuli, which may have reduced the contrast between emotional and neutral conditions employed here. In ASD subjects, amygdala hypoactivation was observed during ToM tasks, in line with previous reports of significant amygdala pathology in autism [51][53]. Our results suggest that both SZ and ASD show deficits in engaging the amygdala when processing social stimuli, but this observation may depend on stimulus type. In SZ, reduced amygdala activation was associated with tasks of attribution of affective states, while in ASD it was mostly seen during tasks of epistemic and intentional attributions. We can only speculate as to the underlying mechanisms of this stimulus-specific dissociation of amygdala engagement in SZ and ASD.

The amygdala is not a homogeneous structure but is comprised of 13 nuclei. The functional specificity of the amygdala nuclei remains unclear, since selective lesions in humans have not been reported. Despite evidence that damage to basal and lateral nuclei may be more directly linked to recognition of facial expressions [54], [55] it is not known whether there is regional specialisation within the amygdala for processing other social stimuli. However, our data suggest that the possibility of disease specific regional pathology in SZ and ASD may be worth pursuing in future studies.

Lateral and ventral temporal regions

Both SZ and ASD were associated with similar hypoactivation in temporal regions around the STS during ToM tasks; however ASD patients showed increased engagement in this region compared to controls and to SZ patients during FER.

Temporal lobe regions near the STS are thought to be involved in the representation of biologically salient cues relating to goal-directed movement [22] including the configural and changeable properties of human faces [56]. Processing of this information in regions around the STS is thought to assist in formulating attributions about the intentional states of others [23], [57].

Our results suggest that processing information relating to ToM tasks was associated with similar reduction in engagement within these regions both in SZ and ASD. However, we also found evidence that STS engagement may differ between disorders based on stimulus type, as ASD patients showed overactivation in these regions during FER. Reduced activation in regions around the STS (extending to the fusiform gyrus) has been noted in ASD patients during person identification tasks (i.e. memorising and recalling information about the identity of a face); however person identification is not required in facial emotion recognition paradigms which rely on decoding the “social” (i.e. emotional) cues conveyed by facial expression. Our results resonate with previous reports that patients with ASD rely excessively on facial feature extraction during facial processing [51] and suggest that they probably use a “logical” rather than a “social” approach to dealing with human facial expressions. Our findings reinforce the dependence of functional changes in ASD on the type of stimulus and also the specific task requirements which would be useful in informing and refining future study designs.


Decreased engagement of the thalamus during social cognition was found mostly in connection to SZ, implicating the dorsomedial and pulvinar nuclei. The thalamus provides a nodal link for multiple functional circuits coordinating the flow of information in a range of higher cognitive and sensory processes. The pulvinar is is considered a “higher order” nucleus because of its widespread bidirectional cortical connections; it is directly involved in visual perception, particularly in directing and maintaining attention towards salient stimuli. The involvement of the dorsomedial thalamus in social cognition is mostly considered within the context of its general contribution to executive functions through its connections with the prefrontal cortex. Extensive thalamic pathology both in terms of microstructure, gross anatomy and functional engagement have been consistently reported in SZ [49]. Our findings extend this literature to include thalamic dysfunction to social cognition tasks.

Methodological considerations

Although Activation Likelihood Estimation represents a powerful approach for the meta-analytic treatment of neuroimaging data, a number of factors should be considered in the interpretation of the current set of findings. First, comparison of neuroimaging studies between SZ and ASD is complicated by the variability of the activation paradigms used. We attempted to minimise this by grouping together paradigms that map on to different domains of social cognition and then examining them separately. This was possible for studies investigating attribution of affective states using facial expressions. However, given the variability of the ToM paradigms we followed the approach of other meta-analytic studies, which have also pooled several related domains of cognition together [31], [58]. Second, we accepted the results of individual original studies as reported, since ALE analyses do not allow for weighting based on the threshold of significance employed in each original study. Some studies have reported coordinates extracted from pre-specified regions of interest; this may have inflated the weight assigned to the findings regarding the amygdala and fusiform gyrus (FER studies), and medial prefrontal cortex, cingulate and STS (ToM studies). Third, ASD and SZ patients differ in their symptom profiles, with psychotic symptoms being predominantly associated with the diagnosis of SZ. Previous studies have suggested that the presence or absence of positive symptoms may contribute to the distribution and degree of functional disruption in SZ during social cognition [59]. The contribution of psychotic symptoms to the present findings is unclear. Nevertheless, the majority of SZ studies included patients that would be generally regarded as being in remission. Fourth, we have documented effects of antipsychotic medication on signal differences in several brain regions; however, these effects are predominantly ameliorative (i.e. reduced the effect size of group differences), and are thus unlikely to account for the differences observed between groups. Fifth, although gender differences have been found during social cognition tasks [58] it was not possible to examine this directly because the predominance of male participants in ASD studies did not allow gender-specific analysis of the available data. Fifth, age differences between the diagnostic groups may have influenced our findings, particularly in ToM tasks, where ASD patients were significantly younger than SZ patients. Indeed, an effect of age was observed in prefrontal regions, favouring SZ. Sixth, the average sample size per study was generally small. This suggests that individual studies may be prone to Type II errors, but also that our results represent the most robustly replicated group differences. Finally, our meta-analysis provides an estimate of how consistently clusters of differential activation occur in particular brain regions when comparing groups of individuals (e.g. SZ vs. controls), and not the mean activation difference in these regions. Therefore traditional measures of heterogeneity and publication bias that are based on the effect size of group differences are not applicable.


In sum, the present results shed light on the neural correlates of social cognition in SZ and ASD and identify differences and similarities between the disorders. In the context of the current debate with regard to diagnostic boundaries, we hope that these findings will facilitate the formulation of new pathophysiological hypotheses and aid the design of future studies.

Author Contributions

Conceived and designed the experiments: SF ET. Performed the experiments: SF GS MK. Analyzed the data: SF GS MK. Contributed reagents/materials/analysis tools: SF GS MK. Wrote the paper: SF GS MK RC ET.


  1. 1. Cheung C, Yu K, Fung G, Leung M, Wong C, et al. (2010) Autistic disorders and schizophrenia: related or remote? An anatomical likelihood estimation. PLoS One 5(8): e12233.
  2. 2. King BH, Lord C (2011) Is schizophrenia on the autism spectrum? Brain Res 1380: 34–41.
  3. 3. Rapoport J, Chavez A, Greenstein D, Addington A, Gogtay N (2009) Autism spectrum disorders and childhood-onset schizophrenia: clinical and biological contributions to a relation revisited. J Am Acad Child Adolesc Psychiatry 48(1): 10–18.
  4. 4. Baron-Cohen S, Leslie AM, Frith U (1985) Does the autistic child have a “theory of mind”? Cognition 21(1): 37–46.
  5. 5. Bora E, Yucel M, Pantelis C (2009) Theory of mind impairment in schizophrenia: meta-analysis. Schizophr Res 109(1–3): 1–9.
  6. 6. Couture SM, Penn DL, Losh M, Adolphs R, Hurley R, et al. (2010) Comparison of social cognitive functioning in schizophrenia and high functioning autism: more convergence than divergence. Psychol Med 40(4): 569–579.
  7. 7. Frith CD, Corcoran R (1996) Exploring ‘theory of mind’ in people with schizophrenia. Psychol Med 26(3): 521–530.
  8. 8. Tremeau F (2006) A review of emotion deficits in schizophrenia. Dialogues Clin Neurosci 8(1): 59–70.
  9. 9. Brothers L (1990) The social brain: a project for integrating primate behavior and neurophysiology in a new domain. Concepts Neurosci 1: 27–51.
  10. 10. Frith U, Frith CD (2003) Development and neurophysiology of mentalizing. Philos Trans R Soc Lond B Biol Sci 358(1431): 459–473.
  11. 11. Adolphs R (2002) Neural systems for recognizing emotion. Curr Opin Neurobiol 12(2): 169–177.
  12. 12. Bechara A (2002) The neurology of social cognition. Brain 125(Pt 8): 1673–1675.
  13. 13. Frith CD (2007) The social brain? Philos Trans R Soc Lond B Biol Sci 362(1480): 671–678.
  14. 14. Gallagher H, Frith C (2003) Functional imaging of ‘theory of mind’. Trends Cogn Sci 7(2): 77–83.
  15. 15. Phillips ML, Drevets WC, Rauch SL, Lane R (2003) Neurobiology of emotion perception I: The neural basis of normal emotion perception. Biol Psychiatry 54(5): 504–514.
  16. 16. Amodio DM, Frith CD (2006) Meeting of minds: the medial frontal cortex and social cognition. Nat Rev Neurosci 7(4): 268–277.
  17. 17. Spreng RN, Mar RA, Kim AS (2009) The common neural basis of autobiographical memory, prospection, navigation, theory of mind, and the default mode: a quantitative meta-analysis. J Cogn Neurosci 21(3): 489–510.
  18. 18. Quirk GJ, Beer JS (2006) Prefrontal involvement in the regulation of emotion: convergence of rat and human studies. Curr Opin Neurobiol 16(6): 723–727.
  19. 19. Berthoz S, Armony JL, Blair RJ, Dolan RJ (2002) An fMRI study of intentional and unintentional (embarrassing) violations of social norms. Brain 125(Pt 8): 1696–1708.
  20. 20. Saxe R, Powell LJ (2006) It's the thought that counts: specific brain regions for one component of theory of mind. Psychol Sci 17(8): 692–699.
  21. 21. Saxe R, Kanwisher N (2003) People thinking about thinking people. The role of the temporo-parietal junction in “theory of mind”. Neuroimage 19(4): 1835–1842.
  22. 22. Grosbras MH, Beaton S, Eickhoff SB (2011) Brain regions involved in human movement perception: A quantitative voxel-based meta-analysis. Hum Brain Mapp.
  23. 23. Allison T, Puce A, McCarthy G (2000) Social perception from visual cues: role of the STS region. Trends Cogn Sci 4(7): 267–278.
  24. 24. Rizzolatti G, Craighero L (2004) The mirror-neuron system. Annu Rev Neurosci 27: 169–192.
  25. 25. Critchley HD, Rotshtein P, Nagai Y, O'Doherty J, Mathias CJ, et al. (2005) Activity in the human brain predicting differential heart rate responses to emotional facial expressions. Neuroimage 24(3): 751–762.
  26. 26. Santos A, Mier D, Kirsch P, Meyer-Lindenberg A (2010) Evidence for a general face salience signal in human amygdala. Neuroimage 54(4): 3111–3116.
  27. 27. Pinkham AE, Penn DL, Perkins DO, Lieberman J (2003) Implications for the neural basis of social cognition for the study of schizophrenia. Am J Psychiatry 160(5): 815–824.
  28. 28. Li H, Chan RC, McAlonan GM, Gong QY (2009) Facial emotion processing in schizophrenia: a meta-analysis of functional neuroimaging data. Schizophr Bull 36(5): 1029–1039.
  29. 29. Pelphrey K, Adolphs R, Morris JP (2004) Neuroanatomical substrates of social cognition dysfunction in autism. Ment Retard Dev Disabil Res Rev 10(4): 259–271.
  30. 30. Pinkham AE, Hopfinger JB, Pelphrey KA, Piven J, Penn DL (2008) Neural bases for impaired social cognition in schizophrenia and autism spectrum disorders. Schizophr Res 99(1–3): 164–175.
  31. 31. Di Martino A, Ross K, Uddin LQ, Sklar AB, Castellanos FX, et al. (2009) Functional brain correlates of social and nonsocial processes in autism spectrum disorders: an activation likelihood estimation meta-analysis. Biol Psychiatry 65(1): 63–74.
  32. 32. Fusar-Poli P, Placentino A, Carletti F, Landi P, Allen P, et al. (2009) Functional atlas of emotional faces processing: a voxel-based meta-analysis of 105 functional magnetic resonance imaging studies. J Psychiatry Neurosci 34(6): 418–432.
  33. 33. Vytal K, Hamann S (2010) Neuroimaging support for discrete neural correlates of basic emotions: a voxel-based meta-analysis. J Cogn Neurosci 22(12): 2864–2885.
  34. 34. Laird AR, Fox PM, Price CJ, Glahn DC, Uecker AM, Lancaster JL, et al. (2005) ALE meta-analysis: controlling the false discovery rate and performing statistical contrasts. Hum Brain Mapp 25(1): 155–164.
  35. 35. Talairach J, Tournoux P (1988) Co-planar stereotaxic atlas of the human brain. New York: Thieme.
  36. 36. Beer JS, Knight RT, D'Esposito M (2006) Controlling the integration of emotion and cognition: the role of frontal cortex in distinguishing helpful from hurtful emotional information. Psychol Sci 17(5): 448–453.
  37. 37. Leitman DI, Wolf DH, Loughead J, Valdez JN, Kohler CG, et al. (2010) Ventrolateral prefrontal cortex and the effects of task demand context on facial affect appraisal in schizophrenia. Soc Cogn Affect Neurosci 6(1): 66–73.
  38. 38. Gallagher HL, Jack AI, Roepstorff A, Frith CD (2002) Imaging the intentional stance in a competitive game. Neuroimage 16(3 Pt 1): 814–821.
  39. 39. Johnson SC, Baxter LC, Wilder LS, Pipe JG, Heiserman JE, et al. (2002) Neural correlates of self-reflection. Brain 125(Pt 8): 1808–1814.
  40. 40. Zysset S, Huber O, Ferstl E, von Cramon DY (2002) The anterior frontomedian cortex and evaluative judgment: an fMRI study. Neuroimage 15(4): 983–991.
  41. 41. Wild B, Erb M, Bartels M (2001) Are emotions contagious? Evoked emotions while viewing emotionally expressive faces: quality, quantity, time course and gender differences. Psychiatry Res 102(2): 109–124.
  42. 42. Adolphs R (2010) Social cognition: feeling voices to recognize emotions. Curr Biol 20(24): R1071–1072.
  43. 43. Amaral DG, Schumann CM, Nordahl CW (2008) Neuroanatomy of autism. Trends Neurosci 31(3): 137–145.
  44. 44. Nickl-Jockschat T, Habel U, Michel TM, Manning J, Laird AR, et al. (2011) Brain structure anomalies in autism spectrum disorder—a meta-analysis of VBM studies using anatomic likelihood estimation. Hum Brain Mapp. DOI:
  45. 45. Rizzolatti G, Fabbri-Destro M (2009) Mirror neurons: from discovery to autism. Exp Brain Res 200(3–4): 223–237.
  46. 46. Huang MX, Lee RR, Gaa KM, Song T, Harrington DL, Loh C, et al. (2010) Somatosensory system deficits in schizophrenia revealed by MEG during a median-nerve oddball task. BrainTopogr 23(1): 82–104.
  47. 47. Norra C, Waberski TD, Kawohl W, Kunert HJ, Hock D, et al. (2004) High-frequency somatosensory thalamocortical oscillations and psychopathology in schizophrenia. Neuropsychobiology 49(2): 71–80.
  48. 48. Walter H, Ciaramidaro A, Adenzato M, Vasic N, Ardito RB, et al. (2009) Dysfunction of the social brain in schizophrenia is modulated by intention type: an fMRI study. Soc Cogn Affect Neurosci 4(2): 166–176.
  49. 49. Meyer-Lindenberg A (2010) From maps to mechanisms through neuroimaging of schizophrenia. Nature 468(7321): 194–202.
  50. 50. Anticevic A, Van Snellenberg JX, Cohen RE, Repovs G, Dowd EC, et al. (2010) Amygdala Recruitment in Schizophrenia in Response to Aversive Emotional Material: A Meta-analysis of Neuroimaging Studies. Schizophr Bull. DOI:
  51. 51. Schultz RT (2005) Developmental deficits in social perception in autism: the role of the amygdala and fusiform face area. Int J Dev Neurosci 23(2–3): 125–141.
  52. 52. Baron-Cohen S, Ring HA, Bullmore ET, Wheelwright S, Ashwin C, et al. (2000) The amygdala theory of autism. Neurosci Biobehav Rev 24(3): 355–364.
  53. 53. Baron-Cohen S, Ring HA, Wheelwright S, Bullmore ET, Brammer MJ, et al. (1999) Social intelligence in the normal and autistic brain: an fMRI study. Eur J Neurosci 11(6): 1891–1898.
  54. 54. Reynders HJ, Broks P, Dickson JM, Lee CE, Turpin G (2005) Investigation of social and emotion information processing in temporal lobe epilepsy with ictal fear. Epilepsy Behav 7(3): 419–429.
  55. 55. Adolphs R, Tranel D, Damasio AR (1998) The human amygdala in social judgment. Nature 393(6684): 470–474.
  56. 56. Haxby JV, Hoffman EA, Gobbini MI (2000) The distributed human neural system for face perception. Trends Cogn Sci 4(6): 223–233.
  57. 57. Friston K, Mattout J, Kilner J (2011) Action understanding and active inference. Biol Cybern 104(1–2): 137–160.
  58. 58. Minzenberg MJ, Laird AR, Thelen S, Carter CS, Glahn DC (2009) Meta-analysis of 41 functional neuroimaging studies of executive function in schizophrenia. Arch Gen Psychiatry 66(8): 811–822.
  59. 59. Pinkham AE, Brensinger C, Kohler C, Gur RE, Gur RC (2010) Actively paranoid patients with schizophrenia over attribute anger to neutral faces. Schizophr Res 125(2–3): 174–178.
  60. 60. Hall GB, Doyle KA, Goldberg J, West D, Szatmari P (2010) Amygdala engagement in response to subthreshold presentations of anxious face stimuli in adults with autism spectrum disorders: preliminary insights. PLoS One 5(5): e10804.
  61. 61. Monk CS, Weng SJ, Wiggins JL, Kurapati N, Louro HM, et al. (2010) Neural circuitry of emotional face processing in autism spectrum disorders. J Psychiatry Neurosci 35(2): 105–114.
  62. 62. Deeley Q, Daly EM, Surguladze S, Page L, Toal F, et al. (2007) An event related functional magnetic resonance imaging study of facial emotion processing in Asperger syndrome. Biol Psychiatry 62(3): 207–217.
  63. 63. Ashwin C, Baron-Cohen S, Wheelwright S, O'Riordan M, Bullmore ET (2007) Differential activation of the amygdala and the ‘social brain’ during fearful face-processing in Asperger Syndrome. Neuropsychologia 45(1): 2–14.
  64. 64. Critchley HD, Daly EM, Bullmore ET, Williams SC, Van Amelsvoort T, et al. (2000) The functional neuroanatomy of social behaviour: changes in cerebral blood flow when people with autistic disorder process facial expressions. Brain 123(Pt 11): 2203–2212.
  65. 65. Rauch AV, Reker M, Ohrmann P, Pedersen A, Bauer J, et al. (2010) Increased amygdala activation during automatic processing of facial emotion in schizophrenia. Psychiatry Res 182(3): 200–206.
  66. 66. Dowd EC, Barch DM (2010) Anhedonia and emotional experience in schizophrenia: neural and behavioral indicators. Biol Psychiatry 67(10): 902–911.
  67. 67. Seiferth NY, Pauly K, Kellermann T, Shah NJ, Ott G, et al. (2009) Neuronal correlates of facial emotion discrimination in early onset schizophrenia. Neuropsychopharmacology 34(2): 477–487.
  68. 68. Reske M, Habel U, Kellermann T, Backes V, Jon Shah N, et al. (2009) Differential brain activation during facial emotion discrimination in first-episode schizophrenia. J Psychiatr Res 43(6): 592–599.
  69. 69. Michalopoulou PG, Surguladze S, Morley LA, Giampietro VP, Murray RM, et al. (2008) Facial fear processing and psychotic symptoms in schizophrenia: functional magnetic resonance imaging study. Br J Psychiatry 192(3): 191–196.
  70. 70. Hall J, Whalley HC, McKirdy JW, Romaniuk L, McGonigle D, et al. (2008) Overactivation of fear systems to neutral faces in schizophrenia. Biol Psychiatry 64(1): 70–73.
  71. 71. Williams LM, Das P, Liddell BJ, Olivieri G, Peduto AS, et al. (2007) Fronto-limbic and autonomic disjunctions to negative emotion distinguish schizophrenia subtypes. Psychiatry Res 155(1): 29–44.
  72. 72. Gur RE, Loughead J, Kohler CG, Elliott MA, Lesko K, et al. (2007) Limbic activation associated with misidentification of fearful faces and flat affect in schizophrenia. Arch Gen Psychiatry 64(12): 1356–1366.
  73. 73. Das P, Kemp AH, Flynn G, Harris AW, Liddell BJ, et al. (2007) Functional disconnections in the direct and indirect amygdala pathways for fear processing in schizophrenia. Schizophr Res 90(1–3): 284–294.
  74. 74. Surguladze S, Russell T, Kucharska-Pietura K, Travis MJ, Giampietro V, et al. (2006) A reversal of the normal pattern of parahippocampal response to neutral and fearful faces is associated with reality distortion in schizophrenia. Biol Psychiatry 60(5): 423–431.
  75. 75. Holt DJ, Kunkel L, Weiss AP, Goff DC, Wright CI, et al. (2006) Increased medial temporal lobe activation during the passive viewing of emotional and neutral facial expressions in schizophrenia. Schizophr Res 82(2–3): 153–162.
  76. 76. Kosaka H, Omori M, Murata T, Lidaka T, Yamada H, et al. (2002) Differential amygdala response during facial recognition in patients with schizophrenia: an fMRI study. Schizophr Res 57(1): 87–95.
  77. 77. Lord C, Rutter M, Le Couteur A (1994) Autism Diagnostic Interview-Revised: a revised version of a diagnostic interview for caregivers of individuals with possible pervasive developmental disorders. J Autism Dev Disord 24(5): 659–85.
  78. 78. Lord CR, Dilavore M, P. Risi S (1999) Autism Diagnostic Observation Schedule (ADOS). Los Angeles, CA: Western Psychological Services.
  79. 79. Kay SR, Fiszbein A, Opler LA (1987) The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophr Bull 13(2): 261–276.
  80. 80. Andreasen N (1984) Scale for the Assessment of Positive Symptoms. Iowa City: University of Iowa.
  81. 81. Andreasen N (1984) Scale for the Assessment of Negative Symptoms. Iowa City: University of Iowa.
  82. 82. Lombardo MV, Chakrabarti B, Bullmore ET, Sadek SA, Pasco G, et al. (2010) Atypical neural self-representation in autism. Brain 133(Pt 2): 611–624.
  83. 83. Uddin LQ, Davies MS, Scott AA, Zaidel E, Bookheimer SY, et al. (2008) Neural basis of self and other representation in autism: an FMRI study of self-face recognition. PLoS One 3(10): e3526.
  84. 84. Wang AT, Lee SS, Sigman M, Dapretto M (2007) Reading affect in the face and voice: neural correlates of interpreting communicative intent in children and adolescents with autism spectrum disorders. Arch Gen Psychiatry 64(6): 698–708.
  85. 85. Dapretto M, Davies MS, Pfeifer JH, Scott AA, Sigman M, et al. (2006) Understanding emotions in others: mirror neuron dysfunction in children with autism spectrum disorders. Nat Neurosci 9(1): 28–30.
  86. 86. Williams JH, Waiter GD, Gilchrist A, Perrett DI, Murray AD, et al. (2006) Neural mechanisms of imitation and ‘mirror neuron’ functioning in autistic spectrum disorder. Neuropsychologia 44(4): 610–621.
  87. 87. Pelphrey KA, Morris JP, McCarthy G (2005) Neural basis of eye gaze processing deficits in autism. Brain 128(Pt 5): 1038–1048.
  88. 88. Park IH, Ku J, Lee H, Kim SY, Kim SI, et al. (2011) Disrupted theory of mind network processing in response to idea of reference evocation in schizophrenia. Acta Psychiatr Scand 123(1): 43–54.
  89. 89. Lee SJ, Kang do H, Kim CW, Gu BM, Park JY, et al. (2010) Multi-level comparison of empathy in schizophrenia: an fMRI study of a cartoon task. Psychiatry Res 181(2): 121–129.
  90. 90. Holt DJ, Cassidy BS, Andrews-Hanna JR, Lee SM, Coombs G, et al. (2011) An Anterior-to-Posterior Shift in Midline Cortical Activity in Schizophrenia During Self-Reflection. Biol Psychiatry 69(5): 415–23.
  91. 91. Benedetti F, Bernasconi A, Bosia M, Cavallaro R, Dallaspezia S, et al. (2009) Functional and structural brain correlates of theory of mind and empathy deficits in schizophrenia. Schizophr Res 114(1–3): 154–160.
  92. 92. Dollfus S, Razafimandimby A, Maiza O, Lebain P, Brazo P, et al. (2008) Functional deficit in the medial prefrontal cortex during a language comprehension task in patients with schizophrenia. Schizophr Res 99(1–3): 304–311.
  93. 93. Brune M, Lissek S, Fuchs N, Witthaus H, Peters S, et al. (2008) An fMRI study of theory of mind in schizophrenic patients with “passivity” symptoms. Neuropsychologia 46(7): 1992–2001.
  94. 94. Lee KH, Brown WH, Egleston PN, Green RD, Farrow TF, et al. (2006) A functional magnetic resonance imaging study of social cognition in schizophrenia during an acute episode and after recovery. Am J Psychiatry 163(11): 1926–1933.
  95. 95. Russell TA, Rubia K, Bullmore ET, Soni W, Suckling J, et al. (2000) Exploring the social brain in schizophrenia: left prefrontal underactivation during mental state attribution. Am J Psychiatry 157(12): 2040–2042.
  96. 96. Baron-Cohen S, Wheelwright S, Skinner R, Martin J, Clubley E (2001) The autism-spectrum quotient (AQ): evidence from Asperger syndrome/high-functioning autism, males and females, scientists and mathematicians. J Autism Dev Disord 31(1): 5–17.