Abstract
Introduction
The baseline non-REM sleep EEG of individuals with insomnia has been found to display increased spectral power at frequencies >14Hz, which may reflect hyperarousal. There is some evidence in this population of reduced slow wave activity after total sleep deprivation (TSD), potentially indicating altered sleep homeostasis. We investigated non-REM sleep EEG spectra at baseline and after TSD in individuals with sleep-onset insomnia.
Methods
10 individuals with sleep-onset insomnia and 5 healthy controls (ages 22-40y, 11 females) completed a 5-day laboratory study with an adaptation night, baseline night, assignment to 38h TSD (n=5 insomnia, n=5 control) or equivalent non-TSD control (n=5 insomnia), and recovery night. Sleep periods were 10h (22:00-08:00) with digital polysomnography (250Hz; Nihon Kohden). Following artifact rejection, 5s subepochs of the non-REM (stages N2, N3) sleep EEG (C3-M2 derivation) in baseline and recovery nights were subjected to spectral analysis. Spectra (0.2Hz bins) were averaged over subepochs in 30s epochs. Repeated-measures ANOVA compared baseline spectra between insomnia and controls, and baseline-recovery difference spectra between TSD insomnia, non-TSD insomnia, and TSD controls.
Results
Average non-REM sleep amount was 5.9 at baseline, increasing by 1.1h after TSD, with no differences between groups (p≥0.20). At baseline, the insomnia group showed increased power in theta/alpha (~4–12Hz), reaching significance in the lower spindle range, compared to controls (p<0.05). As anticipated, no differences emerged between baseline and recovery nights in the non-TSD insomnia group. However, the TSD insomnia group showed increased delta (~1–3Hz) and theta/alpha (~6–10Hz) power (p<0.05) during recovery. Healthy controls showed expected power increases in delta and lower spindle range, and decreases in upper spindle range (~14–15Hz), after TSD (p<0.05).
Conclusion
Compared to healthy controls, individuals with sleep-onset insomnia showed increased non-REM sleep EEG power in the theta/alpha bands and low spindle frequency range, with further significant increases in theta/alpha in addition to delta power following TSD, despite small sample size. The increase in delta power following TSD was equivalent to that in healthy controls, suggesting no sleep homeostasis abnormality. Whether the elevated theta/alpha power may be related to hyperarousal is unclear.
Support (if any)
ONR grant N00014-13-C-0063
Associations between REM‐sleep EEG spectral power, the cholinergic basal forebrain volume and episodic memory in ageing
