Computational Models of Dopamine Dysfunction in Autism Spectrum Disorders

Background:   Despite the diverse behavioral and neurobiological profile exhibited by individuals with autism, dopamine, a neurochemical with widespread influence throughout cortex, is conjectured to play a central role in  many aspects of the autism phenotype.  Abnormal neurogenesis, increased seizure prevalence, motor problems, stereotypies, and difficulties learning to follow eye gaze are all associated with autism, and, importantly, are also influenced by dopamine.  Initial computational modeling results demonstrate how dopamine dysfunction may explain aspects of executive dysfunction, overselectivity, and weak central coherence (WCC).  In these models, the prefrontal cortex (PFC) supports top down attentional modulation based on current goals and task constraints. Dopamine interacts with the PFC supporting both a mechanism capable of toggling maintenance currents within the PFC, effectively switching the attentional target, and a teaching signal allowing the system to learn the proper timing of attentional updating.  Dopamine dysfunction is postulated to result in overly perseverative attention, leading to excessively restricted access to parts of an object or situation, possibly accounting for many behavioral aspects of the disorder.
Objectives: To demonstrate the viability of viewing dopamine dysfunction as central to the disorder, unifying many formally disparate areas of theorizing in autism.
Methods:   Computational models, inspired and constrained by actual neuroscientific findings, are employed to investigate aspects of overselectivity, executive dysfunction, and WCC.  Each model was manipulated to investigate our hypothesis of overly perseverative top-down attentional influence of the PFC, driven by perturbed dopamine / PFC interactions, on behavior.
Results: Initial modeling results capture performance of both people with autism and controls on tasks of overselectivity, executive functioning, and WCC.
Conclusions: Dopamine’s widespread influence and strong ties to the autism phenotype make it an intriguing candidate for a neural underpinning of autism.  Our initial modeling results provide additional support, warranting further investigation into dopamine's role in autistic behavior.