Conventional functional connectivity of blood oxygenation level-dependent (BOLD) signals varies with Alzheimer's disease (AD) progression. However, it is unable to describe how white matter (WM) is engaged in brain networks. In this study, we introduced a novel communication connectivity metric, which was defined as the triple-wise correlation coefficient between BOLD signals from pairs of gray matter volume and white matter bundles, to investigate the change of signal transition through WM bundles. A total of 169 participants with longitudinal resting-state fMRI data from the ADNI dataset were included, which consisted of 44 cognitively normal (CN), 58 early mild cognitive impairment (EMCI), 45 late MCI (LMCI), and 22 AD. Cross-sectional analyses at baseline and longitudinal within-group comparisons were conducted to examine changes in pattern correlation coefficients (CC) between 2D graphs across the AD continuum. In the cross-sectional study, the 2D graph of the CN group showed moderate correlation with those of the EMCI and LMCI groups, whereas these associations generally declined in the AD dementia group. Bootstrapping test showed that the pattern CC for the right retrolenticular part of internal capsule (RLIC.R) and posterior corona radiata (PCR.R) significantly declined in the EMCI, LMCI, and AD groups for both cross-sectional and longitudinal studies. These results demonstrated that signal transmission in RLIC.R and PCR.R has great potential to be markers in the early diagnosis of AD and tracking the progression of AD. Communication connectivity based on rs-fMRI is a promising tool for investigating WM signal transmission alterations in AD.
© 2026. The Author(s).