AUTHOR=Zhao Yuqing , Cai Jiayu , Song Jian , Shi Haoran , Kong Weicheng , Li Xinlei , Wei Wei , Xue Xiehua 

TITLE=Peak alpha frequency and alpha power spectral density as vulnerability markers of cognitive impairment in Parkinson’s disease: an exploratory EEG study

JOURNAL=Frontiers in Neuroscience

VOLUME=Volume 19 - 2025

YEAR=2025

URL=https://www.frontiersin.org/journals/neuroscience/articles/10.3389/fnins.2025.1575815

DOI=10.3389/fnins.2025.1575815

ISSN=1662-453X

ABSTRACT=BackgroundCognitive impairment substantially impacts quality of life in Parkinson’s disease (PD), yet current biomarker frameworks lack sensitivity for detecting early-stage cognitive decline. While peak alpha frequency (PAF) and alpha power spectral density (PSD) have emerged as potential electrophysiological markers, prior studies primarily focused on global cortical measures, neglecting region-specific variations that may better reflect the heterogeneous nature of PD-related cognitive impairment (PDCOG). To address this gap, we conducted the first multiregional comparative analysis of PAF and alpha PSD between PDCOG and PD with normal cognition patients (PDNC).MethodsData from 76 participants (44 PD, 32 healthy controls) at The Affiliated Rehabilitation Hospital of Fujian University of Traditional Chinese Medicine (March–July 2024) were analyzed. PAF and alpha PSD were computed across brain regions; cognitive function was assessed via MoCA.ResultsGlobal PAF was reduced in PD vs. controls (p < 0.05) and correlated with cognition. PDCOG showed lower alpha PSD in parieto-occipital/posterior temporal regions (P3, P4, O1, T5, T6, PZ) vs. PDNC (p < 0.05), with these regions positively correlating with MoCA scores. ROC analysis identified P3, PZ, and T6 alpha PSD as optimal discriminators (AUC: 0.77–0.758). Executive function inversely correlated with alpha PSD in right posterior temporal/left occipital regions.ConclusionPAF differentiates PD from controls and links to global cognition, while regional alpha PSD (notably P3, PZ, T6) effectively distinguishes PDCOG from PDNC. These findings underscore regional QEEG’s utility in PD cognitive assessment, though sensitivity limitations warrant optimization.