Multitaper Time-Frequency Peak Detection based Sleep Analysis in longitudinal 5XFAD mice model

Abstract

Introduction: Sleep disturbance is common in Alzheimer’s disease (AD), but the characteristics of sleep disturbance remain unclear. Multitaper spectral analysis (MSA) is a novel method for investigating sleep. However, MSA-based sleep research in AD is lacking; hence we applied MSA to examine the sleep of AD. Methods: Electroencephalograms were recorded on 3-, 6-, and 10-month-old 5XFAD mice, and the time-frequency (TF) peaks were detected using MSA. We comparatively analyzed the TF peaks between genotypes and age groups. Results: The sigma TF peaks (~80%) were sleep spindles. MSA-based TF plot showed distinct patterns, agreeing with manual scoring. With AD progression, the characteristics of TF peaks coherently changed; shorter sigma TF peaks outnumbered longer ones; dark-period fast spindle TF peak density decreased significantly at both 6 and 10 months. Conclusion: Multitaper TF peaks might provide biomarkers for the progression of AD. Further investigations are warranted.