|Zhang ISBI 2016 Paper||757 KB|
Autism Spectrum Disorder (ASD) has been suggested to associate with alterations
in brain connectivity. In this study, we focus on a fiber clustering tractography segmentation
strategy to observe white matter connectivity alterations in ASD. Compared to another
popular parcellation-based approach for tractography segmentation based on cortical
regions, we hypothesized that the clustering-based method could provide a more
anatomically correspondent division of white matter. We applied this strategy to conduct a population-based group statistical analysis for the automated prediction of ASD. We obtained a maximum classification accuracy of 81.33% be- tween ASDs and controls, compared to the results of 78.00% from the parcellation-based method.