Social Cognitive Profiles of Adults with Autism and Schizophrenia

Background: Severe social dysfunction is the principal impediment reducing quality of life and social outcomes in adults with autism spectrum disorder (ASD) and schizophrenia (SCZ). Investigations of the mechanisms underlying social dysfunction in SCZ and ASD have focused on impaired social cognition, and recent genetic, behavioral, and neuroimaging data suggest substantial overlap in the social cognitive phenotypes in the two disorders.

Objectives: No study to date has systematically examined social cognitive profiles in ASD and SCZ to elucidate patterns of similarity and distinction. Delineating patterns within matched samples on a common set of standardized tasks assessing multiple domains of social cognition is imperative for identifying areas of convergence and divergence that can inform treatment efforts for each disorder.

Methods: Adults with ASD (n=79), SCZ (n=95), and typically-developing (TD) individuals (n=128) completed eight tasks spanning 4 domains of social cognition: A) attribution style/bias was evaluated with the Ambiguous Intentions and Hostility Questionnaire (AIHQ); B) emotion recognition was assessed with the Bell Lysaker Emotion Recognition Task and the Penn Emotion Recognition Test; C) social perception was measured with the Relationships Across Domains test; d) mental state attribution was assessed with the Reading the Mind in the Eyes Test, The Awareness of Social Inferences Test, Part III, (TASIT), and the Hinting Task. Participants also completed the Trustworthiness Task, which does not fall cleanly into the four domains.

Results:  The groups were comparable on IQ (group mean range: 102.07–105.30; p=.11), but differed on gender (90% male in ASD, 80% in TD, 74% in SCZ; p=.048), age (group mean range: 23.61-26.97; p<.01), and ethnicity (90% Caucasian in ASD, 76% TD, 64% SCZ; p<.01). Gender, age, and ethnicity were covaried in analyses.

A MANOVA of the group differences was significant, F(16,564)=7.10, p<.001, and follow up one-way ANOVAs indicated significant differences between the groups on all tests (p’s<.04). TD individuals scored significantly higher than both clinical groups on emotion recognition, social perception, and mental state attribution tasks (p’s<.001). TD individuals rated faces significantly more trustworthy than SCZ (p=.04), but not ASD individuals (p=.53). On the AIHQ, TD individuals trended towards significance compared to the SCZ group (p=.08), with the SCZ group rating more bias to blame others, but did not differ compared to the ASD group (p=.99)

When directly compared, the ASD and SCZ groups performed significantly different on TASIT (p=.01) and approached significance on the AIHQ (p=.06). The ASD group was less accurate in detecting lies and sarcasm, and individuals with SCZ showed a stronger attribution bias to blame others for negative outcomes. The clinical groups did not differ from each other on the other social cognitive tasks (p’s>.50).

Conclusions: Although individuals with ASD and SCZ share significant social cognitive impairments, they also exhibit some disorder-specific distinctions, including a greater hostility bias in SCZ and reduced ability to detect lies and sarcasm in ASD. These may contribute to social dysfunction in unique ways, and treatment efforts may benefit from targeting these different profiles to improve quality of life and outcomes for adults with these disorders.