Characterization of brain network features in alzheimer's disease based on complex network analysis

Abstract

Alzheimer's disease (AD) is a prevalent neurodegenerative disease in aging populations. Complex network analysis helps quantify brain features for AD diagnosis, yet existing studies use coarse brain atlases and fixed thresholds, failing to capture subtle topological changes along disease progression. This study used the fine-grained Zalesky_1024 atlas to optimize thresholds and characterize topological alterations across AD progression. Resting-state functional magnetic resonance imaging data from cognitively normal (CN), early mild cognitive impairment (EMCI), late mild cognitive impairment (LMCI) and AD groups were used to construct brain networks. Topological characteristics were then analyzed across thresholds to track progressive changes from CN→EMCI→LMCI→AD. Finally, support vector machine (SVM) was applied to explore the classification performance of topological characteristics. Results showed threshold-dependent topological alterations along the progression. SVM with betweenness centrality and degree achieved superior classification for EMCI versus AD and CN versus AD. Furthermore, the Zalesky_1024 template outperformed the traditional automated anatomical labeling 90 (AAL_90) template in terms of sensitivity, specificity, accuracy, and area under the curve for classification. These results reveal progressive brain topological degradation across the AD spectrum and supply imaging topological markers to advance AD staging and neuroimaging diagnostic tools.