Abstract
INTRODUCTION: Sleep misperception, where individuals report discrepancy between subjective and objective sleep, is common in insomnia. This study aimed to identify sleep electroencephalographic (EEG) features in insomnia and controls using a large dataset to elucidate the neural mechanisms underlying sleep misperception in insomnia. METHOD: The Sleep Heart Health Study (SHHS) dataset with one-night of polysomnography was used. Under-estimated and over-estimated sleep perception was identified using the 5th percentile of the mean total sleep time (TST) for control and insomnia groups. EEG spectral power for sleep stages and sleep cycles were calculated. Analysis included slow oscillation (SO; 0.5-1.25 Hz) and sleep spindle (12–15 Hz) activity, and slow wave dynamics defined as the decline rate in delta power (1-4.5 Hz) from the first to last NREM sleep period. RESULT: Data from 2024 controls (60 years old) and 284 insomnia (62 years old) were analysed. Insomnia had shorter TST (352.8 vs. 389.3 mins, p<0.001) and longer WASO (62 vs. 51.1 mins, p<0.001) and longer sleep latency (27 vs 20.8 mins, p<0.001) than controls. Insomnia showed lower NREM delta activity but higher theta, alpha, sigma, and beta activity. Insomnia Under-estimators (n=22) had lower SO activity (p=0.03) and attenuated slow wave dynamics (p=0.01) compared to Control Under-estimators (n=87). There was no significant difference in NREM EEG activity or SWA dynamics between Over-estimator groups. CONCLUSION: Deficits in NREM slow oscillation activity and altered slow wave dynamics across the night suggests that an abnormal sleep homeostatic process may underlie sleep misperception in insomnia.