Understanding sleep restriction therapy: a network intervention analysis of symptom- and process-level change

Sleep restriction therapy (SRT) is a core component of cognitive behavioral therapy for insomnia and an effective stand-alone treatment. Although a theoretical model exists, few studies have empirically investigated how SRT works. In this network study, we map how changes in insomnia symptoms and proposed treatment processes unfold during SRT.

We performed a secondary analysis of a randomized controlled trial (N = 147) in which adults with insomnia were allocated to telephone-guided SRT or a sleep diary control group. Using network intervention analysis, we estimated weekly network series: (1) with symptoms based on items from the insomnia severity index (ISI), and (2) with process measures of perpetuating behaviors, the circadian rhythm, sleep pressure, and arousal.

Symptom networks showed that SRT was associated with reduced difficulty initiating sleep and maintaining sleep (ISI-1, ISI-2, weeks 1-6), reduced early morning awakenings (ISI-3, weeks 1-2), and reduced sleep dissatisfaction (ISI-4, weeks 2-4, 6), along with a temporary increase in daytime interference (ISI-5, weeks 1-3). Process networks showed that SRT was first associated with reduced time in bed (TIB) (weeks 1-6), reduced bedtime variability (weeks 1-6), and reduced risetime variability (weeks 1-4), followed by reductions in sleep onset latency (weeks 3, 5, 6) and pre-sleep arousal (weeks 3-6).

Symptom findings suggest that SRT immediately and directly targets nighttime complaints, while daytime complaints improve later. Process findings support SRT's theoretical model, suggesting that SRT initiates change by restricting and regularizing TIB, followed by increasing sleep pressure and reducing arousal. Statement of Significance Insomnia is a prevalent sleep disorder, and sleep restriction is among the most effective behavioral interventions. We demonstrate how sleep restriction affects specific insomnia complaints and processes that have been theoretically proposed to underlie SRT's effects. By capturing treatment-induced change in systems of symptoms and processes over time, this study moves beyond pre-to-post comparisons, contributing to an understanding of what sleep restriction targets, when, and in what sequence. These insights may inform expectations of the treatment trajectory and guide further studies on causal links between process and symptom change.