Functional magnetic resonance imaging (fMRI) has enabled insight into the human brain that was not possible 30 years ago. However, there were high expectations that fMRI would yield information about the neurobiological basis of autism spectrum disorder (ASD) that generally have not been met. While some shortcomings are related to methodological differences across laboratories relating to reproducibility, in large part, the difficulties can be attributed to artifacts in data caused by motion. Motion-related problems are not limited to the study of ASD but impact virtually all areas of human MRI-based imaging.

The short-term goal of this proposal is to eliminate head motion in MRI scans at the time of acquisition by developing new, patient-specific hardware that physically prevents the head from moving in the scanner. This will lead to developing improved neuroimaging techniques that will facilitate our longitudinal investigations of the neural bases of ASD, including aspects related to restricted interests. The majority of individuals with ASD have a restricted interest that often interferes with daily functioning, and for a subset of individuals, the content of the interest shifts over time. It is unknown how or why restricted interests develop, why they switch over time and what the neural representations of the evolution of specific interests are in individuals with ASD. Our longer-term goal is to longitudinally track changes in representations and functional connectivity patterns associated with changing restricted interests to identify candidate regions and mechanisms responsible for the phenotype in ASD.

Eliminating head motion during MRI scanning would have a far-reaching impact on the study of the connectivity and function of the human brain in typical and atypical development. In the longer term, understanding the neural representation of restricted interests will provide much-needed insight into the mechanisms of this aspect of the ASD phenotype that is not possible with animal models.