Project Summary Autism spectrum disorders (ASDs) are characterized by vast neural and behavioral phenotypic heterogeneity. However, little is yet known about how variability in these processes impacts functional outcomes; this, in turn, hampers development of highly effective and targeted interventions. As children encounter increasing social demands through school, these challenges become especially urgent. The current project aims to address these concerns through improved multi-level characterization of the processes contributing to “real world” social deficits in ASD. The principal investigator has previously shown that simultaneous study of socially- sensitive EEG/ERP and behaviorally-indexed cognitive variables can predict more than 50% of variance in key social outcomes (e.g. diagnostic severity). The current proposal extends this work by focusing on predictors of “real world” social functioning (e.g. friendship-making and prosocial behavior) that have been characterized in typically-developing (TD) children (e.g. emotion recognition, Theory of Mind), but whose relations have never been firmly established in ASD. These variables will be examined at both electrophysiological (perceptual and early cognitive) and behavioral levels. 160 youth with ASD and 100 TD youth, ages 11 – 17, will be assessed. They will complete lab-based batteries of EEG and behavioral tasks; parent and teacher report of social functioning, sociometrics in school, and observed prosocial behavior data will be collected as “real world” outcome indices. Predictors and “real world” outcomes will be correlated to see whether these long-presumed, yet rarely-tested relations are indeed evident. Then, advanced predictive modeling will be used to assess specific, optimal patterns of factors that best characterize the variance in social outcomes in ASD. These patterns will be cross-validated to maximize generalizability. Third, subgroups of individuals with ASD for whom these patterns are especially salient will be identified. This innovative approach will test bedrock assumptions of the field (i.e. the importance of neural and behavioral measures in functional outcomes) to uncover basic biopsychosocial variables underlying social deficits in ASD. It will also use contemporary statistical modeling approaches to – for the first time – identify much more precisely how putative predictors combine to affect social functioning, creating a more realistic and ecologically-valid picture of how internal factors “scale up” to affect the social world, and deriving more useful and specific treatment targets . Finally, this approach will help achieve the goal of identifying functionally meaningful subgroups within ASD, which will yield profound insights to guide investigations into discrete developmental processes by which social deficits (and capabilities) arise in ASD, and to whom precision intervention approaches may be matched. The proposed research addresses Objective 1 of the NIMH Stra...