Quantitative EEG signatures of power and functional connectivity alterations in Alzheimer’s disease and frontotemporal dementia

Dementia is a common neurodegenerative disease in the elderly, which affects the structural and functional connectivity of the brain. Recent studies indicate that electrophysiological measures, such as power spectral features and functional connectivity (FC), show promise for the diagnosis and classification of dementia. However, findings across studies remain inconsistent, and distinct electrophysiological patterns separating dementia subtypes, as well as Frontotemporal Dementia (FTD) and cognitively normal (CN) individuals, are not yet well established. This study focuses on spectral power and functional connectivity (FC) analyses of the Electroencephalography (EEG) frequency bands (delta, theta, alpha, beta, and gamma) in Alzheimer's Disease (AD) and FTD. A publicly available eyes-closed (EC), resting-state (RS) EEG dataset comprising 88 age-matched participants, 36 with AD, 29 CN, and 23 with FTD, was used in this study. Absolute power was computed using Welch's method, while FC within each frequency band was assessed using Inter-Site Phase Clustering (ISPC) and network-based statistics, edge and node strength. The global power analysis revealed a significantly higher alpha power in CN compared to both AD and FTD. Regional analysis revealed a significantly lower temporal and parietal alpha in AD relative to CN and a significantly lower occipital alpha and beta in both AD and FTD compared to CN. Topographical power analysis showed unique significant differences within lobes in delta, theta, alpha, and gamma bands in AD and FTD, with AD illustrating a relatively more heterogeneous power distribution than FTD. Furthermore, FC analysis indicated that compared with CN, AD exhibited significantly lower edge strength in delta, theta, beta, and gamma bands, while significantly lower node strength in delta, theta, and gamma bands. Likewise, compared with CN, FTD showed significantly lower edge and node strength in the delta and theta bands, while significantly higher in the beta band. Furthermore, when compared to FTD, AD revealed a significantly lower edge and node strength in the delta, beta, and gamma bands. In conclusion, AD was associated with widespread FC disruptions, while FTD retained partially preserved connectivity, with the temporal lobe more affected than the frontal lobe. These findings suggest that band power and FC alterations may serve as potential biomarkers for diagnosing and classifying dementia into AD and FTD.