A Genetics-First Approach to Shared Risk in Neurodevelopmental Disorders

PROJECT SUMMARY Autism spectrum disorder (ASD) has a complex genetic architecture including multiple rare variants of large effect, some of which are also associated with schizophrenia (SCZ). This genetic heterogeneity mirrors phenotypic heterogeneity, as core ASD behaviors of social impairment and restricted/repetitive behavior differ in severity and

presentation across individuals. Further complicating the phenotype, ASD is associated with impairment across the Research Domain Criteria (RDoC) Cognitive Systems domain and co-occurring psychopathology. This limits effective care. Preliminary work I and others have conducted shows the value of a “genetics-first” approach to

addressing this problem, but few studies have examined rare copy number (CNVs) and single gene variants (SNVs and indels) associated with both ASD and SCZ. This proposal capitalizes on the Simons Powering Autism Research for Knowledge (SPARK) cohort, with genetic data available for more than 50,000 autistic people, to

evaluate cognitive systems and psychopathology in autistic people who carry rare variants with shared SCZ risk. Variants shared by ASD and SCZ have the potential to identify one or more subgroups within ASD. SCZ is defined by psychosis but, like ASD, is associated with social difficulties. It is also associated with RDoC Cognitive

Systems impairment within the cognitive control and perceptual inference constructs. These constructs are heritable and associated with specific neural circuits, and they can be computationally modeled. In SCZ, cognitive control deficits are marked and associated with disorder severity, and perceptual inference deficits

involve strongly weighted prior expectations of sensory evidence that may contribute to psychosis, in which expectation distorts perceived reality. In ASD, studies do not support a characteristic profile of impairment in cognitive control, but rather a variable impairment in executive function. Perceptual inference in ASD has been

hypothesized to involve weakly weighted prior expectations, giving rise to stereotyped behaviors and resistance to change, but studies have variously shown weak, strong, or unchanged weighting of priors. This K23 proposal tests the overall hypothesis that autistic people who carry genetic variants associated with both ASD and SCZ

(ASD+SCZ variants) define a “SCZ-like” subgroup with impaired cognitive control; stronger perceptual bias towards prior expectations; and a greater burden of co-occurring psychopathology, with these effects 1) moderated by contributions from common variation and type of rare variant and 2) worsening over time. I will

remotely assess cognition and psychopathology in 400 participants with ASD (200 with ASD+SCZ variants, 200 without) at two timepoints across multiple levels of analysis, including task paradigms, self-report, and parent- report. This study represents a unique opportunity to examine connections among genes, cognitive

endophenotypes, and psychopathology in a genetically constrained ASD sample, opening opportunities for later studies of underlying circuits and emerging endophenotypes as diagnostic predictors across development.