Fig. 1: Schematic overview of the study framework.

A Development of the gray-white matter heterogeneous fusion network (GWM-HFN). This phase initiated with the extraction of gray matter (GM) and white matter (WM) signals from preprocessed rs-fMRI datasets to create a GM-WM connectivity matrix B. Following this, the values were standardized row-wise to generate a Z-score matrix, highlighting the relative interaction profile of each region within its connectivity framework. Ultimately, the covariance matrix C, which was obtained from Z, functioned as a GM FC matrix mediated by WM, also referred to as the GWM-HFN. B Analysis of GWM-HFN Characteristics. The research evaluated the test-retest reliability of the GWM-HFN network alongside its topological characteristics using graph theoretical metrics. Additionally, comparative assessments between GWM-HFN and conventional GM-GM connectivity networks were carried out. C Utilization of GWM-HFN. The practical relevance of the GWM-HFN was investigated across three areas. First, examining its age-related trends, which included both linear and quadratic effects; second, evaluating its use in clinical settings; and third, applying partial least squares (PLS) regression to assess its predictive ability for cognitive and behavioral outcomes. Icon elements in this figure were sourced from iSlide (islide.cc) and Freepik.com.