Understanding Differences in Social Learning
Abstract
Previous research has shown that individuals with autism spectrum disorder (ASD) appear to learn from social and non-social rewards at different rates compared to typically developing individuals. Several hypotheses have been developed to explain these differences, including the social motivation hypothesis, the weak central coherence hypothesis and hypotheses related to probabilistic learning ability. However, in all cases, the literature shows only mixed support for these ideas. This dissertation focuses on identifying which assumptions from these hypotheses replicate and what replication successes and failures mean for the study of autism-spectrum traits within the general population.
This work takes a “spectrum” approach to autism that assumes ASD-related traits occur on a scale continuum. It therefore is designed to test the central predictions of each of these hypotheses amongst participants sampled from the general population. The use of general population samples confers the considerable advantage of allowing adequate statistical power for hypothesis tests. In addition to these hypotheses, this dissertation explores how social behavior and interaction outcomes relate to ASD-traits and task outcomes.
Interestingly, results ran contrary to many of the previous findings in the literature. Despite evidence of associations within the general population and ASD-traits, I failed to find clear associations between ASD-traits and predictions made by the Social Motivation Hypothesis, the Weak Central Coherence Hypothesis or hypotheses related to probabilistic learning ability. Despite these results, data on real social behavior and social outcomes did vary as a function of ASD-relevant traits. Specifically, the interaction partners of individuals who reported higher levels of ASD-traits experienced them as less likable and reported worse interaction quality. Additionally, individuals reporting higher levels of ASD-related traits were less expressive than those reporting fewer traits.
Overall, while predictions about ASD-traits and cognitive/motivational processes did not appear to replicate within the general population, ASD-traits do appear to be related to real-life social behavior and interaction outcomes associated. Together, these findings document subtle social behavior differences associated with ASD traits in the absence of social cognitive differences and suggest that major theories of autism may not sufficiently explain the causes of altered social behavior in those with autism-spectrum conditions.