Parcellation of functional subregions using an improved affinity propagation clustering algorithm based on structure‐function coupling

LEARNING POINTS Integrating structural and functional connectivity via a nonlinear SFC framework improves the biological plausibility and robustness of primary visual cortex (V1) subregion parcellation compared with unimodal methods. The dorsal subregion of V1 (Dor‐V1) exhibits selective functional abnormalities in individuals with autism spectrum disorder (ASD), which are significantly correlated with the severity of core social symptoms. Dor‐V1 dysfunction may serve as a promising neuroimaging biomarker for patient subtyping and predicting treatment response in ASD, highlighting the clinical translational value of this fine‐grained subregion atlas.