
PLoS ONE, Год журнала: 2024, Номер 19(8), С. e0308137 - e0308137
Опубликована: Авг. 8, 2024
Cognitive decline in Alzheimer's disease is associated with electroencephalographic (EEG) biosignatures even at early stages of mild cognitive impairment (MCI). The aim this work to provide a unified measure by aggregating from multiple EEG modalities and evaluate repeatability the composite an individual level. These included resting state (eyes-closed) two event-related potential (ERP) tasks on visual memory attention. We compared individuals MCI (n = 38) age-matched healthy controls HC 44). In EEG, group exhibited higher power Theta (3-7Hz) lower Beta (13-20Hz) frequency bands. both ERP tasks, reduced late positive (LPP), delayed component latency, slower reaction time, decreased response accuracy. Cluster-based permutation analysis revealed significant clusters difference between groups frequency-channel time-channel spaces. measures performance (12 total) were selected as predictors MCI. trained support vector machine (SVM) classifier achieving AUC 0.89, accuracy 77% cross-validation using all data. Split-data validation resulted (AUC 0.87, 76%) 0.75, 70%) testing data baseline follow-up visits, respectively. Classification scores visits correlated (r 0.72, p<0.001, ICC 0.84), supporting test-retest reliability biosignature. results utility EEG/ERP for prognostic testing, repeated assessments, tracking treatment outcomes limited duration clinical trials.
Язык: Английский