Abstract
AbstractResting-state functional MRI (rs-fMRI) detects spontaneous low-frequency oscillations in the MRI signal at rest. When they occur simultaneously in distant brain regions, they define functional connectivity (FC) between these regions. While blood oxygen level-dependent (BOLD) fMRI serves as the most widely used contrast for rs-fMRI, its reliance on neurovascular coupling poses challenges in accurately reflecting neuronal activity, resulting in limited spatial and temporal specificity and reduced sensitivity in white matter regions. To overcome these limitations, apparent diffusion coefficient fMRI (ADC-fMRI) is emerging as a promising alternative. This approach captures neuronal activity by monitoring changes in ADC resulting from activity-driven neuromorphological alterations such as transient cell swelling. Using graph theory analysis of resting-state FC networks, this study confirms that ADC-fMRI mirrors the positive correlations observed in BOLD-fMRI in the gray-to-gray matter edges (GM-GM), while diverging significantly from BOLD-fMRI for white-to-white matter (WM-WM) connections. While comparable average clustering and average node strength were found for GM-GM connections, higher average clustering (p<10—3) and average node strength (p<10—3) for ADC-fMRI in WM-WM edges suggests that it captures different information to BOLD in the WM. In addition, a significantly higher FC similarity between subjects for ADC-fMRI (mean 0.70, 95% CI [0.68, 0.72]) than BOLD-fMRI (0.38 [0.31, 0.44]) in WM-WM connections suggests a higher reliability of ADC-fMRI in this brain tissue type, demonstrating its broader applicability across the entire brain and reduced sensitivity to physiological noise. Taken together, these results indicate a higher sensitivity and robustness of ADC-fMRI in the WM, and encourage its use, together with careful mitigation of vascular contributions, to further investigate WM functional connectivity.
Publisher
Cold Spring Harbor Laboratory