Individual Differences in Affective Social Perception in ASD: Neural, Behavioral, and Psychiatric Contributions

Background: The social motivation hypothesis posits that reduced social drive leads to inattention to people and consequent failure of developmental specialization in experience-driven brain systems for social perception. Abnormalities in face perception and in the action perception system are documented throughout the lifespan in ASD, but heterogeneity is poorly understood. The current work explores the influence of individual differences in neurocognitive characteristics and psychiatric symptomatology on social brain circuitry underlying emotion perception.

Objectives: To apply an innovative experimental paradigm to (a) examine electrophysiological markers of both emotional face processing and action perception and to (b) explore relationships among event-related potential (ERP) markers of social perception and neurocognitive and psychiatric characteristics. 

Methods: The paradigm employed a novel stimulus set of 210 unique 3D photorealistic face stimuli capable of producing movements consistent with human musculoskeletal structure. Children with ASD (N=27) and matched controls (N=35) viewed a 500ms static initial pose, which segued into 500ms facial movement of three types: (1) affective movement (fearful expression); (2) neutral movement (puffed cheeks); and (3) biologically impossible movement (upward dislocation of eyes and mouth). ERPs (reflecting stages of face processing) were time-locked to onset of static face stimuli, and oscillatory EEG power in the mu range (reflecting activation in the action-perception system) was extracted during periods of facial movement. Comorbid symptoms were measured with the Child Behavior Checklist (CBCL).

Results: An ERP index of face identification (N170) differentiated clinical groups, such that individuals with ASDs exhibited delayed N170s to faces [F(1,60) =18.3, p<.001]. A main effect of emotion on N170 latency [F(3,60) =4.419, p=.005] revealed that neutral faces elicited faster N170s than expressive facial. Bayesian structural equation models were applied to examine shared versus distinct latent sources of variability for ERPs and EEGs to faces, as well as an integrative model incorporating brain activity and behavior. These models revealed that early markers of face processing (P1) predicted later markers (N170; p = .023). However, in models incorporating diagnostic category and reported social problems from the CBCL, this relationship was attenuated; both social problems and diagnosis predicted N170 latency (ps = .082 and .004, respectively), and social problems uniquely predicted variability in the P1 (p = .034). Bivariate models investigating these effects revealed that, across groups, N170 latency was significantly associated with comorbid symptoms: social difficulties (r = .439, p = .001), aggression (r = .306, p = .019), externalizing behavior (r = .339, p =.009), affective problems (r = .363, p = .005), inattention and hyperactivity (r = .414, p = .001), cognitive problems (r = .465, p <.001), and obsessive-compulsive behaviors (r = .386, p= .003).

Conclusions: Neural responses differentiated diagnostic groups and explained transdiagnostic variability in comorbid symptoms. Results suggest observed heterogeneity in ERP indices of social perception may reflect individual differences in neurocognitive and psychiatric factors. These findings have significant implications for individualized treatment.