Analyzing four study cohorts spanning from infancy to adulthood, we compared DTI-derived diffusion metrics as well as connectome Edge Density between subjects with Autism Spectrum Disorders (ASD) and neurotypical controls. Additionally, we explored the performance of several machine learning algorithms applied to tract-based values for prediction of ASD. We found age- and ASD-related alterations in white matter microstructure and connectome in both voxel-wise and tract-based analyses that show how ASD-associated abnormalities emerge and change over time. Our machine learning analysis evaluated several different approaches and identified a model that achieved 0.75 AUC in the prediction of an ASD diagnosis.
This abstract and the presentation materials are available to members only; a login is required.