Exploring Network Topology-Based Methods to Differentiate Healthy and Alzheimer's Cohorts: An EEG-Based Study

Functional connectivity and network topology dif-ferences between 23 Alzheimer's disease (AD) patients and 23 healthy subjects were studied using resting state electroen-cephalography (EEG) data. FC across different pairs of EEG channels were obtained using two different metrics: the phase lag index (PLI) and the phase locking value (PLV). MSTs were generated using PLI-weighted networks. We also explored the hierarchical clustering organization of these MSTs. Finally, we employed MST topology network metrics to classify Alzheimer's and healthy groups using Support Vector Machine (SVM), Decision Tree (DT), and Random Forest (RF) classifiers. Our analysis revealed frequency -dependent functional connectivity patterns. Specifically, in the alpha 1 band, Alzheimer's patients exhibited lower network topology metrics, whole-brain PLI, and regional PLI compared to the healthy group. MST analysis further indicated a loss of network organization, particularly in posterior regions, in AD patients. Our results suggest that DT and RF classifiers are most suitable for distinguishing between AD and healthy cohorts using MST network topology measures.