What are social autistic traits and how do they affect quality of life? A causal inference approach

BackgroundThe social interaction styles of autistic individuals are determined by neural differences across multiple brain systems, and these are not necessarily pathological in nature or entirely specific for autism. Interventions that aim to modulate social function to simulate more neurotypical behavior can be detrimental to self-acceptance and quality of life (QoL). Therefore, this cross-sectional study aimed to test contributions of six specific social dimensions to the autistic neurotype and then use causal inference to identify which of them have adverse effects on QoL. MethodsWe quantified individual differences in multiple dimensions of socially relevant functions (social anhedonia, theory of mind, affective empathy, social anxiety, prosopagnosia and alexithymia), as well as confounding constructs, using validated questionnaires in a sample of 366 adults with a wide range of autistic traits (ATs), including 78 with diagnosed autism. We first used Bayesian linear regression to identify the most likely predictive model of social ATs. Then, we constructed a causal model with a sufficient adjustment set to estimate the differential roles of the constructs in lowering QoL.ResultsWe found that 76% of variance in social ATs was explained by a combination of social anhedonia, theory of mind differences, social anxiety, alexithymia, and face blindness, while individual differences in affective empathy did not contribute. However, only alexithymia and social anhedonia were found to adversely affect QoL.ConclusionWe suggest that autistic social function can be explained by variation across multiple dimensions that are more specific, neurally and behaviorally, than the broad AT construct. There were autism/AT-related differences across most tested dimensions, but QoL was disproportionately affected by alexithymia and social anhedonia, suggesting that these transdiagnostic constructs are relevant markers of poor outcomes.