Electroencephalography (EEG) Arousal Analysis: An exploratory study of spectral slope dynamics around arousals in insomnia

Abstract

Background: Sleep is fundamental for cognitive performance and physical health, yet its disruption can lead to significant impairments. Insomnia, characterized by fragmented sleep and frequent nocturnal arousals—including disrupted and fragmented REM sleep—is associated with a chronic state of hyperarousal. Traditional sleep metrics often fail to capture the nuanced, moment-to-moment fluctuations in cortical activity that contribute to sleep instability. Recent advances propose the EEG spectral slope as a continuous marker of sleep depth and arousal, offering a novel window into neural dynamics around arousal events. Aim: This study investigates how the EEG spectral slope evolves immediately before and after arousals during sleep in individuals with insomnia compared to healthy controls, with special attention to the dynamics during REM sleep. Methods: Archival polysomnographic (PSG) data from 349 participants (235 with insomnia and 114 controls) were analyzed. EEG recordings from standard derivations (F3–M2, C3– M2, O1–M2) were segmented into pre- and post-arousal windows. Spectral slopes were computed over the 30–45 Hz frequency range using the multitaper method. To delve further into the analyses, a spectral slope evolution plot was generated to visualize and examine the dynamic changes in slope around arousal events. Finally, a series of repeated-measures ANOVAs and t-tests were conducted to examine temporal changes in slope and group differences, including an explorative REM-specific analysis. Results: Across channels, spectral slopes significantly flattened immediately following arousals, indicating a transient shift toward wake-like neural activity. While overall group differences were modest, the REM-specific analysis revealed that individuals with insomnia exhibited more negative pre-arousal slopes compared to controls—a difference that was mostly moderated by age. Discussion: These findings suggest that spectral slope dynamics provide a sensitive measure of moment-to-moment cortical excitability in sleep and underscore the potential of spectral slope analysis to elucidate the neural mechanisms underlying sleep fragmentation and REM sleep instability in insomnia.