Domain-general versus domain-specific learning mechanisms: Neurochemical mechanisms and relevance to autism

The theory that various features of autism spectrum disorders (ASD) can be explained by differences in the learning (or “predictive coding”) process is growing in popularity. However, extant studies have focused on the domain of sensory perception, i.e., learning what to expect in the visual or auditory domains. It is thus unclear whether such models are restricted to the perceptual domain, or whether they are outlining differences in domain- general learning processes. Consequently, how such theories can explain the social and motor features of ASD is currently unclear. The first part of the current thesis asks whether autistic adults exhibit differences, compared to non-autistic adults, with respect to social learning and motor learning. The second part of this thesis focuses in detail on one of these learning types - social learning. Here I investigate the neurochemical mechanisms that underpin social learning and ask whether they are dissociable from the neurochemical mechanisms that underpin learning from one’s own individual experience (individual learning). In integrating these results with the wider literature, I reflect upon the broader question of whether there are common domain-general learning mechanisms, or domain (e.g., social, motor, individual) specific learning “modules”. Together the studies presented in this thesis implicate the dopaminergic neurotransmitter system in both social and individual learning. Results support the view that there are domain-general neurochemical mechanisms that support various types of learning. These results do not, however, support the view that autistic adults exhibit differences in these domain-general learning processes. That is, our empirical work showed no differences in either social or motor learning when comparing autistic and non-autistic adults. These results do not add support for impaired predictive coding as a core deficit that can explain social and motor atypicalities in autism, but rather force us to think more critically about what overarching conclusions can be drawn from studies of predictive coding in autism within the perception domain.