Graph-theoretical analysis of resting-state EEG networks differentiates Alzheimer's disease and frontotemporal...

Alzheimer's disease (AD) and frontotemporal dementia (FTD) are neurodegenerative diseases with overlapping symptoms, complicating diagnosis. EEG-derived brain connectivity metrics, based on network neuroscience, can quantify brain network organization, but comparisons between AD and FTD using standardized EEG datasets are limited.

We analyzed a publicly available EEG dataset consisting of 36 AD patients, 23 FTD patients, and 29 healthy controls (HCs). Resting-state eyes-closed and eyes-open EEGs were analyzed across delta, theta, alpha, and beta bands. Phase-locking values (PLV) estimated functional connectivity between 19 electrodes, and graph-theory metrics were derived using the Brain Connectivity Toolbox. Group differences were assessed using ANOVAs with FDR correction, followed by Tukey tests.

AD patients showed reduced global efficiency and small-worldness, especially in the alpha and beta bands under eyes-closed conditions, indicating decreased integration. FTD patients exhibited localized network disruptions in frontal and central regions, particularly reduced node degree and local efficiency at F3/F4 and Pz electrodes, suggesting region-specific dysfunction. These differences were more prominent in the eyes-closed state.

EEG graph-theory analysis revealed distinct network alterations in AD and FTD. AD showed impaired global integration and loss of small-world architecture, while FTD demonstrated region-specific disruptions. These findings suggest that EEG graph metrics may serve as cost-effective biomarkers for differentiating dementia subtypes and understanding disease-specific network dysfunction.