Gray Matter Functional Connectivity Networks by Integrating White Matter Signals: From Method, Properties and Applications

Resting-state functional connectivity (FC) analysis has predominantly focused on gray matter (GM), overlooking the potential functional contributions of white matter (WM). However, emerging evidence suggests that WM BOLD signals may actively shape large-scale brain networks. In this study, we introduce a novel GM-WM-GM connectivity framework that explicitly integrates WM signals to enhance our understanding of functional communication pathways. Using six independent datasets, we establish the test-retest reliability, topological characteristics, and clinical relevance of the GM-WM-GM network. Our findings reveal that GM-WM-GM connectivity demonstrates robust short- and long-term reliability, comparable to traditional GM-GM networks, while capturing unique network features. Graph-theoretical analyses confirm that GM-WM-GM networks exhibit small-world properties, modularity, and distinct hub distributions, emphasizing WM’s active role in functional architecture. Comparative analyses with GM-GM connectivity highlight increased sensitivity to inter-individual variability and functional coupling patterns mediated by WM pathways. Furthermore, we identify significant age-related connectivity changes, characterized by linear declines and nonlinear trajectories, with peak connectivity observed in early adulthood. In a clinical cohort, individuals with autism spectrum disorder (ASD) exhibit hyperconnectivity in GM-WM-GM networks, correlating with symptom severity, underscoring the model’s diagnostic potential. Additionally, GM-WM-GM connectivity predicts cognitive performance, particularly in language and reasoning tasks, demonstrating its behavioral relevance. Collectively, these findings provide compelling evidence that WM-mediated FC contributes to functional brain organization and individual variability. The GM-WM-GM framework offers a more comprehensive perspective on neural communication, bridging the gap between structural and functional connectivity, and holds promise for advancing neuroimaging biomarkers in aging and neuropsychiatric disorders.