Key PointsQuestion
How do deficits in social cognition diverge or overlap between individuals with schizophrenia spectrum disorders (SSDs) and autism spectrum disorder (ASD)?
Findings
In this systematic review and meta-analysis of 36 studies directly comparing social cognitive performance in individuals with SSDs vs ASD, there were no statistically significant differences in emotion processing or theory of mind.
Meaning
Similar levels of social cognitive impairment may be present in individuals with SSDs and ASD, but cross-disorder studies probing social cognitive domains with larger samples, consistent reporting of clinical measures, and neuroimaging are needed to substantiate these findings, clarify underlying mechanisms, and parse heterogeneity.
Importance
Schizophrenia spectrum disorders (SSDs) and autism spectrum disorder (ASD) both feature social cognitive deficits; however, these disorders historically have been examined separately using a range of tests and subdomain focus and at different time points in the life span. Moving beyond diagnostic categories and characterizing social cognitive deficits can enhance understanding of shared pathways across these disorders.
Objective
To investigate how deficits in social cognitive domains diverge or overlap between SSDs and ASD based on the extant literature.
Data Sources
Literature searches were conducted in MEDLINE, PsycInfo, Embase, and Web of Science from database inception until July 26, 2020.
Study Selection
Original research articles were selected that reported performance-based measures of social cognition in both SSDs and ASD samples. Selected articles also had to be published in English and use International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, DSM-IV, or more recent diagnostic criteria.
Data Extraction and Synthesis
This systematic review and meta-analysis was conducted according to Preferred Reporting Items for Systematic Reviews and Meta-analyses and Meta-analysis of Observational Studies in Epidemiology reporting guidelines, including data extraction and quality assessment using a modified version of the Newcastle-Ottawa Scale. Data were pooled using a random-effects model.
Main Outcomes and Measures
Effect sizes were calculated as Hedges g (SSDs vs ASD). The primary outcomes were performance on emotion processing tasks, theory of mind (ToM) tasks, and the Reading the Mind in the Eyes Test (RMET) in SSDs compared with ASD. Meta-regressions were performed for age difference, publication year, quality assessment scores, and antipsychotic medication use.
Results
Of the 4175 screened articles, 36 studies directly comparing social cognitive performance in individuals with SSDs vs ASD were included in the qualitative analysis (n = 1212 for SSDs groups and n = 1109 for ASD groups), and 33 studies were included in the quantitative analyses (n = 1113 for SSDs groups and n = 1015 for ASD groups). Most study participants were male (number of studies [k] = 36, 72% [878 of 1212] in SSDs groups and 82% [907 of 1109] in ASD groups), and age (k = 35) was older in SSDs groups (mean [SD], 28.4 [9.5] years) than in ASD groups (mean [SD], 23.3 [7.6] years). Included studies highlighted the prevalence of small, male-predominant samples and a paucity of cross-disorder clinical measures. The meta-analyses revealed no statistically significant differences between SSDs and ASD on emotion processing measures (k = 15; g = 0.12 [95% CI, –0.07 to 0.30]; P = .21; I2 = 51.0%; 1 outlier excluded), ToM measures (k = 17; g = –0.01 [95% CI, –0.21 to 0.19]; P = .92; I2 = 56.5%; 1 outlier excluded), or the RMET (k = 13; g = 0.25 [95% CI, –0.04 to 0.53]; P = .10; I2 = 75.3%). However, SSDs vs ASD performance differences between studies were statistically significantly heterogeneous, which was only minimally explained by potential moderators.
Conclusions and Relevance
In this analysis, similar levels of social cognitive impairment were present, on average, in individuals with SSDs and ASD. Cross-disorder studies of social cognition, including larger samples, consensus batteries, and consistent reporting of measures, as well as data across multiple levels of analysis, are needed to help identify subgroups within and across disorders that may be more homogeneous in etiology and treatment response.
Schizophrenia spectrum disorders (SSDs) and autism spectrum disorder (ASD) both feature social cognitive deficits,1,2 which contribute to disability and poor functional outcome,3,4 and have limited treatment options. Overlapping clinical symptoms in SSDs and ASD, and social impairments in particular, have long been recognized. Highly heterogeneous clinical presentation is also characteristic within and across these disorders.5-7 Although historically few studies have included both diagnostic groups, cross-disorder research on social cognition in people with SSDs and ASD is increasing because of a shift toward transdiagnostic research,8 use of the Research Domain Criteria framework,9 and the prioritization of improving functional outcome.10,11
Social cognition can be divided into subprocesses, including emotion processing and theory of mind (ToM [also known as mentalizing]), subserved by partially dissociable neural networks.12,13 Meta-analyses have provided evidence for impaired emotion processing and ToM in people with SSDs14,15 and those with ASD16,17 vs typically developing individuals. The onset of ASD begins within the early years of life, during the development of lower-level social cognitive processing (eg, emotion processing), whereas the onset of SSDs begins during late adolescence, when higher-level social cognitive processing (eg, ToM) is still developing.6 Despite these developmental onset differences, similar levels of ToM impairment have been shown in both clinical groups in 2 meta-analyses18,19 of studies that included people with SSDs compared with studies that included people with ASD. However, methods and sample characteristics across studies have been highly variable, making cross-disorder comparisons difficult. Investigations directly comparing social cognition in both conditions are less common, with mixed results regarding relative levels of impairment.
Only 1 meta-analysis20 has examined social cognition across studies directly comparing SSDs and ASD; it found similar ToM performance but greater emotion processing impairment in ASD vs SSDs. Since then, more than 10 studies directly comparing social cognitive performance in SSDs and ASD have been published. Including these new studies more than doubles the sample size and allows for the examination of additional moderators, such as differences in age and antipsychotic treatment, to examine their contribution to the heterogeneity of findings. Formal assessment of study quality, detection of outliers, and sensitivity analyses to evaluate robustness are also needed, which were not included in the previous meta-analysis.20
Herein, we aimed to investigate how deficits in emotion processing and ToM diverge or overlap between individuals with SSDs and ASD by conducting a comprehensive, updated systematic review and meta-analysis of studies directly comparing social cognitive performance in SSDs and ASD. Given differences in age at onset, variation in the use of antipsychotic medication, and changes in diagnostic criteria and research practices over time, we evaluated the implications of age difference, publication year, study quality assessment scores, and antipsychotic medication use as potential moderators for group differences in social cognition using meta-regressions. We hypothesized that effect sizes would be heterogeneous, with a large degree of overlap in emotion processing and ToM performance across SSDs and ASD. Based on meta-analyses to date and evidence of early-onset lower-level social cognitive challenges in ASD,17 we also hypothesized that people with ASD would be more impaired overall on emotion processing than those with SSDs but that similar levels of ToM impairment would be observed.
This systematic review and meta-analysis was conducted according to Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA)21 and Meta-analysis of Observational Studies in Epidemiology (MOOSE)22 guidelines, including data extraction and quality assessment using a modified version of the Newcastle-Ottawa Scale (NOS). Data were pooled using a random-effects model.
After consultation with a reference librarian, literature searches were conducted in MEDLINE, PsycInfo, Embase, and Web of Science (eMethods in the Supplement) to identify original research articles published from database inception until July 26, 2020, that reported performance-based measures of social cognition in both SSDs and ASD samples. Backward and forward citation searches were also conducted for all studies that met inclusion criteria, as well as relevant reviews and meta-analyses.5,6,18,23,24
Eligibility criteria for study inclusion were as follows: (1) a group with SSDs (that could include first-episode psychosis) and (2) a group with ASD according to International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10),25 DSM-IV,26 or DSM-51 definitions (details are provided in the eMethods in the Supplement), (3) reported data from a performance-based measure of social cognition, and (4) original research article. Additional eligibility criteria implemented during the full-text review stage were (5) articles published in English and (6) diagnoses based on ICD-10, DSM-IV, or more recent criteria. Only studies that used performance-based measures (rather than questionnaire or self-report measures) were included because of their greater validity as an index of social cognitive abilities.
References were managed using EndNote Online (Thomson Reuters). Titles and abstracts of nonduplicate publications were screened twice independently (by L.D.O., I.M.-E., and L.G.). Full texts of the potentially eligible studies were then assessed for inclusion by 2 reviewers (L.D.O. and I.M.-E.). Any uncertainties regarding eligibility (n = 135) were reconciled among the reviewers (eMethods in the Supplement).
Data including social cognitive outcomes and participant demographics (eMethods in the Supplement) were extracted from included articles by 1 of 2 reviewers (L.D.O. and I.M.-E.) and subsequently cross-checked by the other. When SSDs and ASD group means and SDs were not provided, they were calculated27-31 or extracted from plots32-34 where possible (eMethods in the Supplement). Calculations and plot extractions were performed twice independently.
A modified version of the NOS35 was used to assess the quality and risk of bias of included articles (eg, group comparability) by either L.D.O. or I.M-E. (score range, 1-10 from low to high quality) (eTable 1 in the Supplement). Uncertainties were discussed between reviewers (L.D.O. and I.M-E.) until consensus was reached.
Data were analyzed in RStudio, version 1.1.447 (R Foundation for Statistical Computing) using the metafor package.36 Effect sizes were calculated as standardized mean difference (SSDs vs ASD Hedges g).37 The primary outcomes were performance on emotion processing tasks, ToM tasks, and the RMET (Reading the Mind in the Eyes Test) in SSDs compared with ASD. Random-effects meta-analyses were conducted using inverse-variance weighting and the restricted maximum likelihood estimator.38 For each meta-analysis (emotion processing, ToM, and RMET), heterogeneity of the pooled effect sizes was assessed using the Cochran Q test. Tau2 (τ2) was calculated to estimate between-study variance. The I2 statistic was used to quantify the percentage of total variation across studies (ie, variation in effect sizes for differences between SSDs vs ASD) that was because of true between-study heterogeneity vs chance; 25%, 50%, and 75% reflect low, moderate, and high heterogeneity, respectively.39 Outliers and impactful studies were detected using studentized residuals greater than 3 (in absolute value) and leave-one-out and combinatorial meta-analyses (eMethods in the Supplement).40,41 Normality was checked for the final models using Q-Q plots. Publication bias was also assessed for each meta-analysis via visual inspection of funnel plots and Egger regression tests for funnel plot asymmetry.42 Statistical tests were performed with a significance level of .05 (2-sided).
Separate meta-analyses were conducted for emotion processing and ToM performance. A third meta-analysis was performed for the RMET43 because it was the most commonly used measure (number of studies [k] = 13) and the literature is mixed regarding whether it assesses emotion processing or ToM abilities or both.43-45 We defined emotion processing tasks as those assessing emotion labeling or matching (eg, emotion recognition), largely thought to involve less complex, lower-level processing, whereas ToM (or mentalizing) is thought to involve more complex, higher-level processing, including perspective taking and mental or emotional state inference.12,13,45 Social cognitive outcomes were categorized by L.D.O. and verified by coauthors (Table 1 and eMethods in the Supplement).
Meta-regressions were performed for age difference, publication year, quality assessment scores, and antipsychotic medication use. The associations of age difference (SSDs vs ASD Hedges g) and publication year were examined using multivariable meta-regressions for each meta-analysis (eMethods in the Supplement). Exploratory univariable meta-regressions were also conducted to examine the associations of quality assessment scores and medication use (SSDs vs ASD mean proportion of participants taking antipsychotic medication) for emotion processing and ToM, as well as stimulus type (verbal and visual) for ToM (emotion processing was all visual).
The main meta-analyses were rerun after excluding studies that required calculations or plot extraction for the quantitative analyses27-29,31-34 or those with only youth participants (age range, 10-20 years)33,46,47 to ensure that these samples were not unduly altering results. Meta-analyses were also rerun after excluding articles with SSDs groups composed only of individuals with first-episode psychosis,48 because this designation does not constitute a formal SSDs diagnosis,1 and articles with SSDs groups composed only of those with schizotypal or schizoid personality disorder,49 because this classification can be considered both part of the schizophrenia spectrum and a personality disorder.
Study Selection and Characteristics
Of the 4175 screened articles, the literature search (Figure 1) resulted in 36 studies directly comparing social cognitive performance in individuals with SSDs vs ASD being included in the qualitative analysis (n = 1212 for SSDs groups and n = 1109 for ASD groups) and 33 studies being included in the quantitative analyses (n = 1113 for SSDs groups and n = 1015 for ASD groups). Eleven authors were contacted for further information (eTable 2 in the Supplement). Across included studies (Table 1),27-34,46-73 sample sizes of the SSDs and ASD groups tended to be small (a group with <25 participants in 22 studies; range, 10-113). Most study participants were male (k = 36, 72% [878 of 1212] in SSDs groups and 82% [907 of 1109] in ASD groups), and age (k = 35) was older in SSDs groups (mean [SD], 28.4 [9.5] years) than in ASD groups (mean [SD], 23.3 [7.6] years). IQ data reported (k = 31) were inconsistent (eg, full-scale, verbal, nonverbal, and unspecified), precluding data pooling. Antipsychotic medication use status was reported for both groups in 22 articles,27,28,30,32,34,47,48,50-58,60,62,67,70,72,73 10 of which reported chlorpromazine equivalents for both.28,30,50-57 All studies used ICD-10, DSM-IV, or DSM-5 criteria for SSDs and ASD diagnoses.
Social Cognitive and Additional Measures
Among the studies included in the meta-analyses, 16 studies28,31,34,46,48,51,52,54,57,60,63,67-69,72,73 used performance-based emotion processing measures, 18 studies27,29,32,33,46-50,60,61,64,66,68-72 included ToM measures, and 13 studies32,48,49,51,55,56,59,62,65,66,69,70,72 used the RMET (Table 1). Three articles30,53,58 were excluded from the quantitative analyses (details are provided in the eMethods in the Supplement).
Reporting of measures of nonsocial cognition, clinical symptoms, everyday or adaptive functioning, and neuroimaging was uncommon and inconsistent within and between studies (eMethods in the Supplement). Nine studies27-29,34,46,56,59,60,74 included a measure of nonsocial cognition, but these varied widely. Clinical scores or subscores provided from each study varied; only 2 studies61,62 reported measures of schizophrenia and autistic symptoms in both groups. Measures of everyday functioning were also rare (k = 5).29,60,62-64 Ten articles28,30,50,53-57,60,62 reported neuroimaging data in conjunction with performance-based social cognition.
Five studies28,32,34,66,68 received low (<5 points), 18 studies29-31,33,46,49,52,53,58-61,63-65,69,70,73 received moderate (5-7 points), and 13 studies27,47,48,50,51,54-57,62,67,71,72 received high (8-10 points) scores on the modified NOS (Table 1). Independent validation of diagnoses was most often unreported, and additional points were lost fairly uniformly among the remaining criteria.
Quantitative findings included emotion processing, ToM, and RMET analyses. Detailed results from the main meta-analyses and sensitivity analyses are listed in Table 2.
There was no statistically significant between-group difference in emotion processing, although a suggestion toward better performance in the SSDs vs ASD groups was found (k = 16; g = 0.21 [95% CI, −0.03 to 0.44]; P = .08) (eFigure 1A in the Supplement). Effect sizes were statistically significantly heterogeneous (I2 = 71.8%). One impactful outlier was detected using studentized residuals and the leave-one-out (eTable 3 in the Supplement) and combinatorial (Figure 2A) meta-analyses.31,41 After removal of 1 outlier, the between-group difference remained non–statistically significant and decreased in size (k = 15; g = 0.12 [95% CI, −0.07 to 0.30]; P = .21) (Figure 3A),28,34,46,48,51,52,54,57,60,63,67-69,72,73 and heterogeneity was reduced (I2 = 51.0%). Funnel plot visualization (eFigure 2A in the Supplement) and Egger regression test revealed no evidence of publication bias (Table 2).
Moderator analyses demonstrated no statistically significant associations of age difference (z = −0.27; P = .79) or publication year (z = −1.58; P = .11) with emotion processing effect sizes (k = 15; P = .26; R2 = 8.96%). Exploratory moderator analyses also revealed no statistically significant association of quality assessment scores (k = 15; z = 0.05; P = .96; R2 = 0%) or antipsychotic medication use (k = 11; z = 0.50; P = .61; R2 = 0%). The overall emotion processing effect size remained non–statistically significant in sensitivity analyses (Table 2 and eResults in the Supplement).
There was no statistically significant between-group difference in ToM performance (k = 18; g = −0.13 [95% CI, −0.45 to 0.18]; P = .42) (eFigure 1B in the Supplement). Effect sizes were highly heterogeneous (I2 = 83.0%), and 1 impactful outlier was identified (Figure 2B and eTable 4 in the Supplement).46 The between-group difference remained non–statistically significant after removal of 1 outlier (k = 17; g = −0.01 [95% CI, −0.21 to 0.19]; P = .92) (Figure 3B),28,34,46,48,51,52,54,57,60,63,67-69,72,73 and heterogeneity decreased but remained statistically significant (I2 = 56.5%). There was no evidence of publication bias after outlier removal based on funnel plot inspection (eFigure 2B in the Supplement) and Egger regression test (Table 2).
There were no statistically significant associations of age difference (z = 0.59; P = .55) or publication year (z = −0.16; P = .88) with ToM effect sizes (k = 17; P = .84; R2 = 0%). Exploratory meta-regressions also revealed no statistically significant association of quality assessment scores (k = 17; z = 0.22; P = .83; R2 = 0%) or stimulus type (k = 14; z = −0.54; P = .59; R2 = 0%). The between-group difference in the proportion of participants taking antipsychotic medication was statistically significantly associated with ToM effect sizes, accounting for a substantial amount of heterogeneity (k = 8; z = −1.99; P = .047; R2 = 46.40%) (eFigure 3 in the Supplement). As the proportion of individuals with SSDs vs ASD taking antipsychotic medication increased, ToM performance of the SSDs vs ASD groups worsened. The overall ToM effect size remained non–statistically significant in sensitivity analyses (Table 2 and eResults in the Supplement).
Reading the Mind in the Eyes Test
The RMET meta-analysis revealed no statistically significant difference in performance between groups, although SSDs groups performed slightly better than ASD groups (k = 13; g = 0.25 [95% CI, −0.04 to 0.53]; P = .10) (Figure 3C).32,48,49,51,55,56,59,62,65,66,69,70,72 Effect sizes were statistically significantly and highly heterogeneous (I2 = 75.3%), although no outliers were identified (Figure 2C and eTable 5 in the Supplement). Funnel plot visualization (eFigure 2C in the Supplement) and Egger regression test revealed no evidence of publication bias (Table 2).
Moderator analyses yielded no statistically significant associations of age difference (z = 0.93; P = .35) or publication year (z = 1.36; P = .17) with RMET performance (k = 13; P = .25; R2 = 3.60%). Exploratory meta-regressions also revealed no statistically significant association of quality assessment scores (k = 13; z = 1.60; P = .11; R2 = 8.55%) or antipsychotic medication use (k = 7; z = 1.11; P = .27; R2 = 0.33%). The overall RMET effect size remained non–statistically significant in sensitivity analyses (Table 2 and eResults in the Supplement).
This systematic review and meta-analysis included studies that directly compared individuals with SSDs vs ASD on performance-based social cognitive measures to identify overlapping and divergent deficits. Findings herein suggest that similar levels of social cognitive impairment may be present in SSDs and ASD across emotion processing and ToM domains. However, heterogeneity of effect sizes was apparent and was only minimally explained by the moderators explored. Included studies highlighted the prevalence of small, male-predominant samples and a paucity of cross-disorder clinical measures.
Across the main meta-analyses comparing emotion processing, ToM, and RMET performance in individuals with SSDs vs ASD, no statistically significant group differences were identified. This finding builds on previous meta-analyses showing similar impairment levels in those with SSDs and ASD vs typically developing individuals on ToM tasks18,19 and the RMET,18 providing further evidence based on direct SSDs vs ASD comparisons. The ToM and RMET findings herein align with results from the only other meta-analysis20 to date examining direct SSDs vs ASD comparisons; the present study provides additional support for minimal differences among individuals with SSDs vs ASD and includes 9 and 5 additional studies, respectively. The previous meta-analysis20 found that participants with SSDs outperformed those with ASD on emotion processing tasks; however, heterogeneity was high, only 8 studies were included, and assessment of impactful studies was not reported. With twice the number of studies, we found no statistically significant group differences in emotion processing performance. Results herein suggest that a consistent pattern of differences in social cognitive performance between SSDs and ASD is not apparent across studies to date. However, many sample sizes were small, and the moderate to high levels of effect size heterogeneity indicate that differences in social cognitive performance between SSDs and ASD varied among studies, warranting further investigation. This heterogeneity represents variability in SSDs vs ASD performance differences rather than within-disorder performance variability.
Age difference between SSDs vs ASD and publication year (partially reflecting changes in clinical practices and diagnostic criteria over time) did not explain between-study variability observed herein in emotion processing, ToM, or RMET effect sizes. The previous meta-analysis20 found better emotion processing and RMET performance in SSDs vs ASD among studies with younger participants across groups. It may be that age difference is less relevant than whether participants are generally younger or older across diagnoses because of relative differences in neurodevelopment and pathoplastic progression (eg, earlier emerging challenges in ASD).6 The exploratory moderator analyses herein revealed that a greater proportion of people with SSDs vs ASD taking antipsychotic medication was associated with poorer ToM performance in SSDs vs ASD, accounting for a moderate amount of between-study heterogeneity. This finding may suggest that antipsychotic medication use negatively impacts ToM performance or could reflect worse performance in those receiving antipsychotic treatment potentially because of greater illness severity. Although there is some meta-analytic evidence for moderating associations of antipsychotic medication use with emotion processing performance in SSDs,15 other meta-regressions have found no association between antipsychotic medication use in SSDs and emotion processing, ToM, or RMET performance.18,75 Additional work that includes complete reporting of medication use status is needed to investigate how antipsychotic treatment and other medication use impact social cognition across SSDs and ASD.
Both SSDs and ASD are highly heterogeneous disorders, and differences in nonsocial cognition, clinical symptoms, and everyday functioning may also have contributed to between-study variability observed herein. However, reporting of such metrics was uncommon across included studies, and measures used to assess these domains were inconsistent both within and between groups. Social cognition has been consistently positively associated with nonsocial cognition and functional outcome in SSDs3 and ASD4 and inversely associated with negative symptoms in SSDs.45,76 Meta-regressions15,75 have demonstrated a negative association between emotion processing performance and negative symptoms in SSDs. Illness duration may also be an important factor. Studies that concurrently include assessments tapping these variables in both SSDs and ASD are necessary to elucidate their association and facilitate transdiagnostic comparisons.
Of the studies reviewed, those with neuroimaging alongside social cognitive measures had mixed findings and tended to include small samples and a single imaging modality. Neurobiological investigations into shared or distinct mechanisms underlying the common social cognitive deficits across SSDs and ASD are warranted. For example, hypermentalizing or hypomentalizing at a cognitive level or opposing patterns of brain activation or connectivity could both lead to impaired ToM performance,50 aligning with a diametrical hypothesis of SSDs and ASD.5 However, the degree of between-study heterogeneity may also align with evidence suggesting that subgroups with overlapping brain-behavior associations may exist transdiagnostically,60 reflecting multiple etiological pathways. Future work is needed to directly test these and alternative theories.7 Integrating data across multiple levels of analysis through approaches like network theory77 will provide additional insight into factors contributing to the emergence of psychiatric disorders, those associated with social cognition, and potential treatment targets.78
This study has some limitations. Few of the reviewed studies focused on adolescents, limiting generalizability of these findings. This limitation is particularly important given the different developmental trajectories of SSDs and ASD and of lower-level vs higher-level social cognition.6 Future work that includes people with co-occurring diagnoses of SSDs and ASD could be informative from a transdiagnostic perspective given the high prevalence of shared traits79 and co-occurrence of these conditions.80,81 Indeed, dimensional analyses of shared symptoms are becoming more common in investigations of the overlap between SSDs and ASD.82-84 Many existing social cognitive tasks may also lack discriminatory power between SSDs and ASD. Finally, although most studies reviewed herein implemented at least some practices to limit bias, a standardized approach would have improved opportunities for comparisons across studies.
Based on meta-analyses of the extant literature, similar levels of social cognitive impairment may be present in SSDs and ASD across emotion processing and ToM. These results highlight the need for cross-disorder studies of social cognition with larger samples, including adolescents, and consistent reporting of measures that may impact outcome. Integrating data spanning multiple levels of analysis across SSDs and ASD is a critical next step to identify associations that may delineate more homogeneous subgroups with similar etiology, treatment response, and phenotypic characteristics.
Accepted for Publication: September 27, 2020.
Published Online: December 8, 2020. doi:10.1001/jamapsychiatry.2020.3908
Corresponding Author: Stephanie H. Ameis, MD, Campbell Family Mental Health Research Institute, Centre for Addiction and Mental Health, 80 Workman Way, No. 5224, Toronto, ON M6J 1H4, Canada (stephanie.ameis@camh.ca).
Author Contributions: Drs Oliver and Ameis had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: All authors.
Acquisition, analysis, or interpretation of data: Oliver, Moxon-Emre, Lai, Grennan, Ameis.
Drafting of the manuscript: Oliver.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Oliver.
Obtained funding: Voineskos, Ameis.
Administrative, technical, or material support: Moxon-Emre, Grennan, Voineskos, Ameis.
Supervision: Lai, Voineskos, Ameis.
Conflict of Interest Disclosures: Drs Oliver and Moxon-Emre reported receiving funding from the Canadian Institutes of Health Research. Dr Lai reported receiving funding from the Ontario Brain Institute via the Province of Ontario Neurodevelopmental Disorders Network, the Canadian Institutes of Health Research, the Academic Scholars Award from the Department of Psychiatry of the University of Toronto, and the Centre for Addiction and Mental Health (CAMH) Foundation. Dr Voineskos reported receiving funding from the National Institute of Mental Health, the Canadian Institutes of Health Research, the Canada Foundation for Innovation, the CAMH Foundation, and the University of Toronto. Dr Ameis reported receiving funding from the National Institute of Mental Health, the Canadian Institutes of Health Research, the Academic Scholars Award from the Department of Psychiatry of the University of Toronto, and the CAMH Foundation. No other disclosures were reported.
Funding/Support: This work was supported by funding from the National Institute of Mental Health (grant R01MH114879; Dr Ameis).
Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Meeting Presentation: This paper was presented at the 59th Annual Meeting of the American College of Neuropsychopharmacology; December 8, 2020; virtual meeting.
Additional Contributions: We thank Sarah Bonato, MIS (reference and research librarian, CAMH Library), for assistance with the development of the search strategy. She was not compensated for her contributions.
1.American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 5th ed. American Psychiatric Association; 2013.
3.Halverson
TF, Orleans-Pobee
M, Merritt
C, Sheeran
P, Fett
AK, Penn
DL. Pathways to functional outcomes in schizophrenia spectrum disorders: meta-analysis of social cognitive and neurocognitive predictors.
Neurosci Biobehav Rev. 2019;105:212-219. doi:
10.1016/j.neubiorev.2019.07.020
PubMedGoogle ScholarCrossref 4.Sasson
NJ, Morrison
KE, Kelsven
S, Pinkham
AE. Social cognition as a predictor of functional and social skills in autistic adults without intellectual disability.
Autism Res. 2020;13(2):259-270. doi:
10.1002/aur.2195
PubMedGoogle ScholarCrossref 8.Cuthbert
BN. The RDoC framework: facilitating transition from
ICD/
DSM to dimensional approaches that integrate neuroscience and psychopathology.
World Psychiatry. 2014;13(1):28-35. doi:
10.1002/wps.20087
PubMedGoogle ScholarCrossref 11.Warren
Z, Taylor
JL, McPheeters
ML,
et al. Future research needs: interventions for adolescents and young adults with autism spectrum disorders: identification of future research needs from comparative effectiveness review No. 65. Agency for Healthcare Research and Quality; September 2012. Future Research Needs Papers No. 20. Accessed May 13, 2020.
https://www.ncbi.nlm.nih.gov/books/NBK121979/ 19.Bliksted
V, Ubukata
S, Koelkebeck
K. Discriminating autism spectrum disorders from schizophrenia by investigation of mental state attribution on an on-line mentalizing task: a review and meta-analysis.
Schizophr Res. 2016;171(1-3):16-26. doi:
10.1016/j.schres.2016.01.037
PubMedGoogle ScholarCrossref 20.Fernandes
JM, Cajão
R, Lopes
R, Jerónimo
R, Barahona-Corrêa
JB. Social cognition in schizophrenia and autism spectrum disorders: a systematic review and meta-analysis of direct comparisons.
Front Psychiatry. 2018;9:504. doi:
10.3389/fpsyt.2018.00504
PubMedGoogle ScholarCrossref 22.Stroup
DF, Berlin
JA, Morton
SC,
et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting: Meta-analysis of Observational Studies in Epidemiology (MOOSE) Group.
JAMA. 2000;283(15):2008-2012. doi:
10.1001/jama.283.15.2008
PubMedGoogle ScholarCrossref 24.Sugranyes
G, Kyriakopoulos
M, Corrigall
R, Taylor
E, Frangou
S. Autism spectrum disorders and schizophrenia: meta-analysis of the neural correlates of social cognition.
PLoS One. 2011;6(10):e25322. doi:
10.1371/journal.pone.0025322
PubMedGoogle Scholar 25.World Health Organization. International Statistical Classification of Diseases, Tenth Revision (ICD-10). World Health Organization; 1992.
26.American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th ed. American Psychiatric Association; 1994.
28.Hirata
K, Egashira
K, Harada
K,
et al. Differences in frontotemporal dysfunction during social and non-social cognition tasks between patients with autism spectrum disorder and schizophrenia.
Sci Rep. 2018;8(1):3014. doi:
10.1038/s41598-018-21379-w
PubMedGoogle ScholarCrossref 37.Hedges
LV, Olkin
I. Statistical Methods for Meta-Analysis. Academic Press; 1985.
43.Baron-Cohen
S, Wheelwright
S, Hill
J, Raste
Y, Plumb
I. The “Reading the Mind in the Eyes” Test revised version: a study with normal adults, and adults with Asperger syndrome or high-functioning autism.
J Child Psychol Psychiatry. 2001;42(2):241-251. doi:
10.1111/1469-7610.00715
PubMedGoogle ScholarCrossref 45.Oliver
LD, Haltigan
JD, Gold
JM,
et al. Lower- and higher-level social cognitive factors across individuals with schizophrenia spectrum disorders and healthy controls: relationship with neurocognition and functional outcome.
Schizophr Bull. 2019;45(3):629-638. doi:
10.1093/schbul/sby114
PubMedGoogle ScholarCrossref 46.Waris
P, Tani
P, Lindberg
N,
et al. Are there differences in neurocognition and social cognition among adolescents with schizophrenia, a pervasive developmental disorder, and both disorders?
Appl Neuropsychol Child. 2016;5(4):303-310. doi:
10.1080/21622965.2015.1064001
PubMedGoogle ScholarCrossref 48.Pepper
KL, Demetriou
EA, Park
SH,
et al. Autism, early psychosis, and social anxiety disorder: understanding the role of social cognition and its relationship to disability in young adults with disorders characterized by social impairments.
Transl Psychiatry. 2018;8(1):233. doi:
10.1038/s41398-018-0282-8
PubMedGoogle ScholarCrossref 49.Booules-Katri
TM, Pedreño
C, Navarro
JB, Pamias
M, Obiols
JE. Theory of mind (ToM) performance in high functioning autism (HFA) and schizotypal-schizoid personality disorders (SSPD) patients.
J Autism Dev Disord. 2019;49(8):3376-3386. doi:
10.1007/s10803-019-04058-1
PubMedGoogle ScholarCrossref 56.Schwarz
K, Moessnang
C, Schweiger
JI,
et al. Transdiagnostic prediction of affective, cognitive, and social function through brain reward anticipation in schizophrenia, bipolar disorder, major depression, and autism spectrum diagnoses.
Schizophr Bull. 2020;46(3):592-602. doi:
10.1093/schbul/sbz075PubMedGoogle ScholarCrossref 57.Martínez
A, Tobe
R, Dias
EC,
et al. Differential patterns of visual sensory alteration underlying face emotion recognition impairment and motion perception deficits in schizophrenia and autism spectrum disorder.
Biol Psychiatry. 2019;86(7):557-567. doi:
10.1016/j.biopsych.2019.05.016
PubMedGoogle ScholarCrossref 58.Morrison
KE, Pinkham
AE, Penn
DL, Kelsven
S, Ludwig
K, Sasson
NJ. Distinct profiles of social skill in adults with autism spectrum disorder and schizophrenia.
Autism Res. 2017;10(5):878-887. doi:
10.1002/aur.1734
PubMedGoogle ScholarCrossref 60.Stefanik
L, Erdman
L, Ameis
SH,
et al. Brain-behavior participant similarity networks among youth and emerging adults with schizophrenia spectrum, autism spectrum, or bipolar disorder and matched controls.
Neuropsychopharmacology. 2018;43(5):1180-1188. doi:
10.1038/npp.2017.274
PubMedGoogle ScholarCrossref 63.Kuo
SS, Wojtalik
JA, Mesholam-Gately
RI, Keshavan
MS, Eack
SM. Establishing a standard emotion processing battery for treatment evaluation in adults with autism spectrum disorder: evidence supporting the Mayer-Salovey-Caruso Emotion Intelligence Test (MSCEIT).
Psychiatry Res. 2019;278:116-124. doi:
10.1016/j.psychres.2019.05.011
PubMedGoogle ScholarCrossref 65.Couture
SM, Penn
DL, Losh
M, Adolphs
R, Hurley
R, Piven
J. Comparison of social cognitive functioning in schizophrenia and high functioning autism: more convergence than divergence.
Psychol Med. 2010;40(4):569-579. doi:
10.1017/S003329170999078XPubMedGoogle ScholarCrossref 69.Kandalaft
MR, Didehbani
N, Cullum
CM,
et al. The Wechsler ACS social perception subtest: a preliminary comparison with other measures of social cognition.
J Psychoeduc Assess. 2012;30(5):455-465. doi:
10.1177/0734282912436411Google ScholarCrossref 73.Sasson
NJ, Pinkham
AE, Weittenhiller
LP, Faso
DJ, Simpson
C. Context effects on facial affect recognition in schizophrenia and autism: behavioral and eye-tracking evidence.
Schizophr Bull. 2016;42(3):675-683. doi:
10.1093/schbul/sbv176PubMedGoogle ScholarCrossref 78.Galderisi
S, Rucci
P, Mucci
A,
et al; Italian Network for Research on Psychoses. The interplay among psychopathology, personal resources, context-related factors and real-life functioning in schizophrenia: stability in relationships after 4 years and differences in network structure between recovered and non-recovered patients.
World Psychiatry. 2020;19(1):81-91. doi:
10.1002/wps.20700
PubMedGoogle ScholarCrossref 82.Trevisan
DA, Foss-Feig
JH, Naples
AJ, Srihari
V, Anticevic
A, McPartland
JC. Autism spectrum disorder and schizophrenia are better differentiated by positive symptoms than negative symptoms.
Front Psychiatry. 2020;11:548. doi:
10.3389/fpsyt.2020.00548PubMedGoogle ScholarCrossref 83.Ziermans
TB, Schirmbeck
F, Oosterwijk
F, Geurts
HM, de Haan
L; Genetic Risk and Outcome of Psychosis (GROUP) Investigators. Autistic traits in psychotic disorders: prevalence, familial risk, and impact on social functioning.
Psychol Med. 2020;1-10. doi:
10.1017/S0033291720000458PubMedGoogle Scholar 84.Larson
FV, Wagner
AP, Chisholm
K, Reniers
RLEP, Wood
SJ. Adding a dimension to the dichotomy: affective processes are implicated in the relationship between autistic and schizotypal traits.
Front Psychiatry. 2020;11:712. doi:
10.3389/fpsyt.2020.00712PubMedGoogle ScholarCrossref