Analysis of Abnormal Brain Networks in Autism Based on fMRI

Abstract

Autism spectrum disorder (ASD) is characterized by widespread neurodevelopmental impairments, with emotional and behavioral deficits linked to abnormal local brain function and functional connectivity. Investigating whether the intrinsic functional brain network topology in ASD is altered is therefore of critical diagnostic importance. To identify aberrant brain networks in ASD, this study constructed functional networks based on 90 brain regions defined by an automated anatomical atlas, followed by network-based statistics (NBS) and graph-theoretical analyses. The findings were further validated using support vector machine (SVM) classification between ASD and neurotypical participants. NBS revealed that abnormalities in functional connectivity in ASD were predominantly located within the default mode and sensorimotor networks, involving regions in the temporal and frontal lobes, basal ganglia, and parts of the limbic system. Graph-theoretical analysis further indicated that topological alterations in ASD likely stem from dysregulated connectivity within these subnetworks, demonstrating significantly impaired efficiency in both information transfer and integration compared to controls—a finding potentially linked to characteristic ASD behavioral profiles. Using these aberrant connections as biological features, a systematic evaluation of multiple classifiers identified SVM as optimal. Through cross-validation and parameter tuning, classification achieved an accuracy of 84.86% on the NYU dataset. To further validate the generalizability of the identified network features, corresponding functional connections were extracted from the larger UM dataset (sample size >80), yielding a classification accuracy of 72.64%, confirming the clinical relevance of the detected network abnormalities. These results highlight abnormal topological organization in functional brain networks in ASD and provide novel insights into the neural mechanisms underlying communication deficits and repetitive behaviors in autism.