Functional connectivity-based sub-network changes related to Alzheimer’s disease progression based on group cohesive parcellation of rsfMRI

Motivation: Brain modeling based on functional connectivity is a powerful tool for characterizing brain networks, yet the search for AD related network biomarkers has yielded few consistent results. Goal(s): We show how cohesion-focused network analysis of intrinsic sub-networks correlate with AD disease progression. Approach: Several node-based and global graph-theoretic measures were calculated, and were examined across several intrinsic sub-networks. Results: Two network metrics, cohesion and small-worldness, showed progressive changes from cognitively normal (CN) to mildly cognitive impaired (MCI) to AD. The fraction of rich club nodes decreased with disease progression, and the distributions of these nodes transferred from the DMN to VAN and SN. Impact: The data demonstrate increasing disorganization within the brain with potential compensatory reorganization in other sub-networks.