Wake slow waves and aperiodic activity during resting‐state EEG correlate with amyloid burden and hypometabolism

Abstract Background Spectral power of slow rhythms in resting‐state EEG increases along Alzheimer's disease (AD) continuum. Besides, recent studies have revealed 1) the importance of analyzing the aperiodic component of an EEG power spectrum and 2) the intrusions of sleep‐like slow waves identifiable in wake EEG of animals and young adults. Importantly, the occurrence of these wake slow waves is known i) to increase after sleep deprivation, ii) to be associated with markers of sleepiness, and iii) to predict behavioral errors at different tasks. Our objective was to investigate the links between AD biomarkers and 1) the aperiodic component of the EEG signal and 2) wake slow waves. Method 314 elderly individuals with memory complaints from the INSIGHT‐preAD cohort (76.5 ± 3.5 years, range: 70‐85 years) were recruited. Participants underwent a 5‐year follow‐up, including 1) high‐density EEG recordings (256 electrodes) at rest (2 sessions of 2 min) each year, 2) neuropsychological assessments each year, and 3) Florbetapir and 18 F‐FDG PET scans at baseline to quantify amyloid burden and hypometabolism. For each EEG recording, characteristics of aperiodic activity (intercept and slope) and wake slow waves (density, amplitude, slope, and frequency) were determined. For each neuroimaging modality, mean standardized uptake value ratios (SUVRs) were computed in AD relevant brain regions. Associations between AD biomarkers (amyloid and FDG SUVRs) and EEG metrics (characteristics of aperiodic activity and slow waves) were tested using linear regressions controlled for age, sex, and education level, after FDR correction. Longitudinal effects, including neuropsychological metrics, will be tested using linear mixed effect models. Result At baseline, high amyloid burden was positively associated with 1) the intercept and slope of aperiodic activity and 2) the density and frequency of slow waves. Besides, hypometabolism was associated with lower slow wave density. Conclusion These preliminary results suggest that aperiodic activity and wake slow waves could serve as early EEG indicators of certain AD biomarkers. This seems particularly promising as these metrics are quantifiable on a simple resting EEG and at a very early stage, even before the onset of cognitive alterations. Longitudinal and neuropsychological data will be analyzed in the next few months.