Resting-state EEG predicts cognitive decline in a neuropathologically diagnosed longitudinal community autopsied cohort

Alexander Choi; Nan Zhang; Charles H Adler; Thomas G Beach; Holly A Shill; Erika Driver-Dunckley; Shyamal Mehta; Christine Belden; Alireza Atri; Marwan N Sabbagh; John N Caviness

doi:10.1016/j.parkreldis.2024.107120

Resting-state EEG predicts cognitive decline in a neuropathologically diagnosed longitudinal community autopsied cohort

Parkinsonism Relat Disord. 2024 Aug 31:128:107120. doi: 10.1016/j.parkreldis.2024.107120. Online ahead of print.

Affiliations

¹ Banner Sun Health Research Institute, 10515 W Santa Fe Dr, Sun City, AZ, 85351, USA. Electronic address: alexander.h.choi@kp.org.
² Department of Neurology, Mayo Clinic College of Medicine, Mayo Clinic Arizona, 13400 E Shea Blvd, Scottsdale, AZ, 85259, USA.
³ Banner Sun Health Research Institute, 10515 W Santa Fe Dr, Sun City, AZ, 85351, USA.
⁴ Barrow Neurological Institute, 240 W. Thomas Suite 301, Phoenix, AZ, 85013, USA.

PMID: 39236511
DOI: 10.1016/j.parkreldis.2024.107120

Abstract

Objective: To assess correlative strengths of quantitative electroencephalography (qEEG) and visual rating scale EEG features on cognitive outcomes in only autopsied cases from the Arizona Study of Neurodegenerative Disorders (AZSAND). We hypothesized that autopsy proven Parkinson Disease will show distinct EEG features from Alzheimer's Disease prior to dementia (mild cognitive impairment).

Background: Cognitive decline is debilitating across neurodegenerative diseases. Resting-state EEG analysis, including spectral power across frequency bins (qEEG), has shown significant associations with neurodegenerative disease classification and cognitive status, with autopsy confirmed diagnosis relatively lacking.

Methods: Biannual EEG was analyzed from autopsied cases in AZSAND who had at least one rsEEG (>1 min eyes closed±eyes open). Analysis included global relative spectral power and a previously described visual rating scale (VRS). Linear mixed regression was performed for neuropsychological assessment and testing within 2 years of death (n = 236, 594 EEG exams) in a mixed linear regression model.

Results: The cohort included cases with final clinicopathologic diagnoses of Parkinson's disease (n = 73), Alzheimer disease (n = 65), and tauopathy not otherwise specified (n = 56). A VRS score of 3 diffuse or frequent generalized slowing) over the study duration was associated with an increase in consensus diagnosis cognitive worsening at 4.9 (3.1) years (HR 2.02, CI 1.05-3.87). Increases in global theta power% and VRS were the most consistently associated with large regression coefficients inversely with cognitive performance measures.

Conclusion: Resting-state EEG analysis was meaningfully related to cognitive performance measures in a community-based autopsy cohort. EEG deserves further study and use as a cognitive biomarker.

Keywords: Alzheimer disease; Clinical-pathologic cohort; Cognitive trajectories; Neurodegenerative longitudinal outcomes; Parkinson disease; Quantitative electroencephalography; qEEG.