Brain morphology network alterations in adolescents with autism spectrum disorder: a sex-stratified study

One of the most persistent questions in autism research is why males are more consistently diagnosed than females. Neuroimaging studies have sought to understand this disparity by examining sex differences, primarily through functional and structural connectivity. However, much less is known about how brain networks are organized in autism from a morphological perspective, and how this organization may help explain its sex-related characteristics. This study aimed to elucidate sex-specific alterations in morphology-based connectivity among adolescents with autism spectrum disorder (ASD).

T1-weighted MRI scans from 337 individuals (181 with ASD; 162 females in total), aged 8-18 years, were preprocessed using FreeSurfer to extract five morphological features-thickness, surface area, sulcal depth, mean curvature, and gray matter volume-across 68 cortical regions. Connectivity was estimated via Kullback-Leibler divergence between region-wise multivariate feature distributions. Network-Based Statistics was applied within sex-stratified models to identify clusters of significantly altered connections in ASD relative to controls (individual edge threshold p = 0.005; FWE-corrected p = 0.05; 5,000 permutations). An exploratory post hoc analysis using partial correlations further associated identified sex-specific network profiles with behavior and comorbidity.

ASD males showed significantly increased connectivity in a cluster centered on the fusiform gyrus, along with the medial orbitofrontal, entorhinal, and parahippocampal cortices. This pattern was positively correlated with restricted and repetitive behaviors (r = 0.22, 95% CI [0.01, 0.41]). In females with ASD, increased connectivity was found in a sub-network implicating mainly the entorhinal cortex, followed by the inferior parietal lobule, and lateral occipital cortex. Notably, the fusiform gyrus was disrupted yet in the opposite direction, exhibiting decreased connectivity with the superior temporal sulcus. These alterations were negatively associated with communication skills (r = -0.25, 95% CI [-0.45, -0.03]). No overlap between male- and female-specific profiles was found, highlighting their distinctiveness.

Our cross-sectional sample does not allow for neurodevelopmental inference. Sex-stratified modeling precluded sex-by-diagnosis testing.

Males and females with ASD show non-overlapping profiles of altered morphology-based connectivity, each associated with different behavioral traits, which may contribute to better characterize sex differences in autism.