Mapping Selective Gray Matter Variations in Autism Spectrum Disorder via Bayes fACtor mOdeliNg

Structural magnetic resonance imaging techniques such as voxel-based morphometry (VBM) have identified heterogeneous gray matter variations in individuals diagnosed with autism spectrum disorder (ASD) compared to typically developing controls. However, it remains unclear whether and to what extent disorder-selective brain variations occur in this spectrum. This research gap, along with the observation that other clear-cut clinical conditions exhibit similar neuroanatomical features, poses a significant challenge in translating neuroimaging findings to clinical practice. In our study, we adopted a novel meta-analytic reverse inference approach (i.e., Bayes fACtor mOdeliNg) to assess the presence of selective neuroanatomical patterns associated with ASD. We analyzed VBM data from the BrainMap and MEDLINE databases, encompassing 849 published experiments. These experiments covered 132 psychiatric and neurological disorders, involving more than 40,000 subjects and 16,572 recorded foci of gray matter variation. Our analysis identified specific clusters of variation in the parietal, occipital, and cerebellar areas, showing a selectivity value of 90% or higher in ASD. These findings not only enhance our understanding of ASD pathophysiology but also highlight potential targets for future neuroimaging-based interventions in clinical settings.