Bo-Yong Park

Autism spectrum disorder is associated with widespread alterations in brain connectivity, but how structural and functional networks are jointly reorganized remains unclear. Here, we show that autism is associated with altered structure-function coupling, particularly in the hierarchical organization of sensory and association systems. Using brain imaging data from 91 individuals with autism and 67 typically developing controls, we model functional connectivity from structural networks using Riemannian optimization and examine communication across multiple steps. In the resulting simulated networks, hierarchical organization is comparable between groups at lower communication steps but becomes less differentiated in autism as communication extends across multiple steps. These alterations are associated with excitation/inhibition alterations, neurotransmitter distributions, gene expression, and multisensory processing. We also identify atypical connectivity involving the thalamus and putamen. Importantly, greater hierarchical distortion predicts more severe autistic symptoms, linking macroscale network organization with neurobiology and behavior. These findings highlight hierarchical structure-function coupling as a potential framework for understanding biological and behavioral characteristics in autism.

Human brain function emerges from the interplay between recurrent local activity and distributed inter-regional interactions. However, a biologically interpretable framework for quantifying this balance between the two factors at the whole-brain level remains lacking. Here, we extended an established biophysical model to quantify inter-regional excitation and newly introduced the recurrent ratio (R-ratio), a measure of the relative balance between intra- and inter-regional excitation. Dynamical analyses showed that an optimal R-ratio supports a trade-off between network stability and flexibility. Applying this framework to healthy aging and Alzheimer’s disease revealed progressively increased R-ratios with advancing age and disease severity. In healthy individuals, higher R-ratios were associated with age-related alterations in brain morphology, molecular pathology, and cognitive function, whereas Alzheimer’s disease was characterized by increased R-ratios accompanied by reduced dynamical persistence. Together, these findings establish the R-ratio as a biologically interpretable marker of whole-brain excitation balance and demonstrate its utility for linking biophysical mechanisms with large-scale brain dynamics, aging, and neurodegeneration.