Biomarkers Derived from Alterations in the Overlapping Community Structure of Resting Brain Functional Networks for Detection of Alzheimer’s Disease

This article was originally published here

Neurosciences. December 29, 2021: S0306-4522 (21) 00662-X. doi: 10.1016 / j.neuroscience.2021.12.031. Online ahead of print.


Recent studies show that the overlapping community structure is an important feature of the brain’s functional network. However, the overlapping alterations in community structure in patients with Alzheimer’s disease (AD) have not yet been examined. In this study, we investigate the overlapping community structure in AD using resting-state functional magnetic resonance imaging (f-rsMRI) data. Collective sparse symmetric non-negative matrix factorization (cssNMF) is adopted to detect the overlapping community structure. The experimental results on 28 patients with AD and 32 normal controls (NC) from the ADNI2 dataset show that the two groups show remarkable differences in terms of optimal number of communities, hierarchy of communities detected at different scales. , functional segregation of the network and nodal functioning. the diversity. In particular, the networks of the fronto-parietal and basal ganglia show significant differences between the two groups. A machine learning framework proposed in this article for the detection of AD achieved 76.7% accuracy using the detected community strengths of the frontal-parietal and basal lymph node networks only as input features. . These results provide new information on the understanding of pathological changes in the organization of the brain functional network of AD and show the potential of features related to the structure of the community for the detection of AD.

PMID: 34973385 | DOI: 10.1016 / j.neuroscience.2021.12.031

Ida M. Morgan