The effect of sleep restriction therapy for insomnia on sleep pressure and arousal: a randomized controlled mechanistic trial.

Study Objectives: Sleep restriction therapy (SRT) effectively treats insomnia but mechanisms are poorly understood. Theoretical models suggest that potentiation of sleep pressure and reduction of arousal are key mechanisms of action. To our knowledge, this has never been directly tested. We designed a randomized controlled trial with embedded mechanistic measurement to investigate if SRT causally modifies multidimensional assessments of sleep pressure and arousal.

Methods: Participants aged 25-55 who met DSM-5 diagnostic criteria for insomnia disorder were randomized to four weeks of SRT or time in bed regularization (TBR), a control intervention that involves prescription of a regular but not reduced time in bed. Sleep pressure was assessed through daily diary appraisal of morning and evening sleepiness, weekly Epworth sleepiness scale (ESS) scores, psychomotor vigilance, and non-rapid eye movement (NREM) delta power (0.75-4.5 Hz) from ambulatory polysomnographic recordings. Arousal was assessed through daily diary appraisal of cognitive arousal, the pre-sleep arousal scale (PSAS), and NREM beta power (15-32 Hz). Outcomes were assessed at baseline (2-week period prior to randomization), during the intervention phase (1-4 weeks post-randomization), and at 12-week follow-up. We performed intention-to-treat analyses using linear mixed models. For continuous daily measures, the treatment period was split into early (weeks 1-2) and late (weeks 3-4) treatment.

Results: Fifty-six participants (39 females, mean age = 40.78 ± 9.08) were assigned to SRT (n = 27) or TBR (n = 29). The SRT group showed enhanced sleep pressure relative to TBR, reflected in (1) enhanced sleepiness in the evening during early (d = 1.17) and late treatment (d = 0.92), and in the morning during early treatment (d = 0.47); (2) higher daytime sleepiness on the ESS at weeks-1 and -2 (d = 0.54, d = 0.45); and (3) reduced psychomotor vigilance at week-1 (d = 0.34). The SRT group also showed reduced arousal relative to TBR, reflected in lower levels of daily-monitored cognitive arousal during early treatment (d = 0.53) and decreased PSAS total score at week-4 and week-12 (ds ≥ 0.39). Power spectral analysis of all night NREM sleep revealed an increase in relative, but not absolute, EEG delta power at week-1 and week-4 (ds ≥ 0.52) and a decrease of relative EEG beta power at week-4 (d = 0.11).

Conclusion: For the first time, we show that SRT increases sleep pressure and decreases arousal during acute implementation, providing support for mechanism-of-action.