Multimodal neuroimaging correlates of NREM and REM sleep: EEG spectral power in aging

Abstract

AbstractBackgroundSleep disturbances have been associated with cognitive decline and Alzheimer’s disease (AD), with growing efforts to understand how these changes may relate to AD pathophysiology. However, we lack a comprehensive overview of the associations between objectively‐measured sleep oscillations and brain changes in aging so that efficient treatment strategies could be developed. Using an unbiased, fast, and powerful quantitative (spectral) analysis of sleep EEG rhythms, we assessed the relationships between both non‐rapid eye movement (NREM) and REM sleep oscillations, and amyloid burden, gray matter (GM) volume and perfusion in cognitively unimpaired older adults (CUOA).Method125 CUOA from the Age‐Well randomized controlled trial underwent a neuropsychological assessment, an ambulatory polysomnography recording and a structural MRI. A Florbetapir‐PET scan was also performed with early and late acquisitions to measure brain perfusion and amyloid burden respectively. Normalized EEG spectral power values were obtained using the ASEEGA® algorithm for delta (0.1‐4 Hz), theta (4‐8 Hz), alpha (8‐12 Hz) and beta (16‐50 Hz) EEG frequency bands, over the Cz‐Pz derivation (Berthomier et al., 2007). Voxel‐wise multiple regressions were then conducted between EEG spectral power values and neuroimaging data (i.e., GM volume, perfusion and amyloid burden), separately for NREM and REM sleep, controlling for age, gender, education, the apnea‐hypopnea index, sleep medication use, and the ApoE4 status.ResultLower NREM sleep delta power, reflecting a decrease in slow wave activity, was related to decreased GM volume and perfusion in fronto‐cingulate areas, but not to amyloid deposition. By contrast, decreased REM sleep theta power was associated with i) greater amyloid deposition in frontal, cingulate, parietal and temporal areas; ii) lower GM volume in frontal, temporal, occipital and insular cortices, and iii) increased perfusion in frontal and parietal regions.ConclusionBoth NREM and REM sleep alterations were associated with structural and functional changes in regions sensitive to aging and AD pathophysiology. Importantly, REM sleep changes were associated with greater amyloid deposition. This may be due to early changes in cholinergic structures, acetylcholine playing a neuroprotective role against amyloid. Thus older individuals with REM sleep alterations may be more at risk of belonging to the Alzheimer’s continuum.