Individual gray-white matter functional connection predicts tau spread and cognitive decline in Alzheimer's disease.

PURPOSE: Alzheimer's disease is characterized by progressive accumulation of hyperphosphorylated tau protein, which propagates in a prion-like manner along connected neuronal pathways. However, it remains unclear whether functional connectivity between gray and white matter (FCGW) can predict tau spread. This study aimed to determine the association between FCGW and tau deposition and to evaluate its value in predicting longitudinal tau spread. METHODS: We integrated resting-state fMRI with cross-sectional and longitudinal tau-PET data from two independent cohorts. We assessed baseline associations between FCGW and tau deposition and then constructed an individual-level spreading model to predict longitudinal tau accumulation. RESULTS: In both cohorts, FCGW showed a positive correlation with tau deposition. Model-simulated white-matter tau deposition was associated with clinical scales and predicted cognitive decline. The spreading model, which incorporated baseline tau-PET and the top 10% of gray and white matter, yielded the highest predictive performance for future tau accumulation. CONCLUSION: FCGW captures key network pathways underlying tau spread in AD and improves prediction of future tau accumulation. These findings highlight the importance of FCGW in understanding tau propagation and support development of network-targeted therapeutic strategies.