Implicit Learning Deficits in Autism: A Neurocomputational Account

Experiments using the Serial Response Time Task (SRTT) have suggested that implicit learning is impaired in people with autism. Implicit learning is learning that occurs without explicit awareness of the knowledge being acquired. Researchers have suggested that poor implicit learning could be a major factor in other profound learning difficulties exhibited by people with autism. In this report, we use a neurocomputational model of the SRTT to show that disturbed interactions between the mesolimbic dopamine (DA) system and the prefrontal cortex (PFC) may underlie the implicit learning problems observed in autism. This model is shown to fit reaction time data from the literature for both individuals with autism and normally developing controls. This work expands on a previous body of research showing that abnormal DA/PFC interactions can account for a disparate collection of autistic behavioral patterns, suggesting that a common neurological mechanism might produce executive dysfunction, stimulus overselectivity, and impaired implicit learning in this population.