The study of autism as a distributed disorder

By Ralph-Axel Müller

Brain Development Imaging Laboratory, Department of Psychology, San Diego State University, San Diego, CA 92120; Department of Cognitive Science, University of California, San Diego, CA 92093

Key Words: Autism, brain imaging, connectivity, functional networks

ABSTRACT

Past autism research has often been dedicated to tracing the causes of the disorder to a localised neurological abnormality, a single functional network or a single cognitive-behavioural domain. In this review, I argue that autism is a ‘distributed disorder’ on various levels of study (genetic, neuroanatomical, neurofunctional, behavioural). ‘Localising’ models are therefore not promising. The large array of potential genetic risk factors suggests that multiple (or all) emerging functional brain networks are affected during early development. This is supported by widespread growth abnormalities throughout the brain. Interactions during development between affected functional networks and atypical experiential effects (associated with atypical behaviour) in children with autism further complicate the neurological bases of the disorder, resulting in an ‘exponentially distributed’ profile. Promising approaches to a better characterisation of neural endophenotypes in autism are provided by techniques investigating white matter and connectivity, such as MR spectroscopy, diffusion tensor imaging (DTI), and functional connectivity MRI. According to a recent hypothesis, the autistic brain is generally characterised by ‘underconnectivity’. However, not all findings are consistent with this view. The concepts and methodology of functional connectivity need to be refined and results need to be corroborated by anatomical studies (such as DTI tractography) before definitive conclusions can be drawn.