Sleep and electroencephalography biomarkers in preclinical Alzheimer’s Disease

Abstract

AbstractBackgroundAccumulating evidence supports a bi‐directional relationship between sleep and Alzheimer’s Disease, including sleep disruptions in preclinical Alzheimer’s Disease (pAD), prior to onset of overt cognitive symptoms. Subsequently, sleep phenotypic variables and electroencephalography (EEG) features are attractive noninvasive biomarkers for pAD. The Biomarkers of Alzheimer’s Disease in Sleep and EEG (BASE) study seeks to develop such sleep‐EEG biomarkers for pAD.MethodCommunity‐dwelling participants (n = 52, mean age: 70.8 ± 4.5 years) underwent cognitive testing including Montreal Cognitive Assessment (MOCA, general cognition), Craft Story 21 (CS21, memory), and Trails A and B (visuospatial, executive); performance was z‐scored by age, sex, and education. Objective measures of sleep quality were obtained using actigraphy over 4‐14 (mean 9.1) days at home: wake time after sleep onset (WASO), number of awakenings (NoA), and sleep efficiency. In‐lab overnight polysomnography was performed, and EEG features were extracted during NREM sleep from F3, F4, C3, and C4 channels: relative slow and fast delta power (RSDP, RFDP), spindle density, and slow wave density. The following morning (9‐10AM), CSF was obtained by fasted lumbar puncture, and assessed for amyloid‐b‐42 and tau; pAD was determined by tau/amyloid‐b‐42 ratio. Correlations between cognitive performance and EEG /sleep variables were performed using linear models, adjusted for age, sex, education, and sleep apnea status. Interaction between pAD status and correlations were assessed.ResultThere were 36 biomarker‐negative healthy controls (HC) and 16 with pAD; groups did not significantly differ demographically. RSDP and WASO were negatively associated with CS21 (RSDP: p = 0.040; WASO: p = 0.034). RFDP and NoA interacted with pAD status for MOCA (RFDP: p = 0.010; NoA: p = 0.026), and RFDP interacted with pAD status for CS21 (p = 0.017).ConclusionWe assessed objective sleep quality and sleep‐EEG features against cognitive performance, and any interaction with pAD status. Worse sleep quality was associated with worse memory performance. Among the EEG features assessed, RFDP was strongly negatively correlated with memory performance in those with pAD, suggesting neurophysiological aberrations of NREM sleep in this cohort.