Adding exercise to sleep counselling beat either one alone in a small trial

Exercise training can improve sleep. Sleep interventions can improve sleep. Both statements are supported — but whether the two have a synergistic effect is unknown, as the authors of a new trial put it [s1]. Doing both might produce more than doing either, or the two might simply overlap, each fixing the same portion of the problem.

A four-arm randomised clinical trial published in JAMA Network Open on February 16 set out to separate those possibilities [s1].

The design

The trial was single-blind, parallel, and four-armed, conducted between July and September 2024 [s1]. Participants were sedentary women of Chinese nationality aged 18 to 30 with poor sleep health at baseline, defined as a total Pittsburgh Sleep Quality Index (PSQI) score above 5 [s1].

The four arms ran for eight weeks [s1]:

• HICT alone — high-intensity circuit training, three body-weight sessions per week, performed in a laboratory
• SH alone — a sleep-health protocol delivered by trained research staff in a laboratory and online, consisting of individualised sleep counselling followed by information drawn from a smartphone-based digital cognitive behavioural therapy for insomnia app, plus weekly visits
• HICT-SH — both
• Control — participants continued their original lifestyle

Sleep was measured two ways: subjectively by PSQI, and objectively by actigraphy [s1]. Secondary outcomes included cardiometabolic markers such as cholesterol and adiponectin [s1].

The four-arm structure is what makes the trial informative. A two-arm trial of "both versus nothing" cannot tell you whether the combination adds anything. Only running each component separately, in the same trial, against the same control, can.

What happened

In total, 112 women participated, mean age 23.5 years (SD 3.2) [s1].

Against the control group, all three intervention groups improved on sleep efficiency, sleep latency, wake after sleep onset, activity counts, sleep duration, and total PSQI score [s1]. That is the unsurprising part.

The combination arm then beat the single-component arms on several objective measures [s1]:

• Sleep efficiency: HICT-SH exceeded SH alone by 2.75 percentage points (95% CI, 0.65–4.85; P = .004)
• Wake after sleep onset: HICT-SH beat HICT alone by 14.51 minutes (95% CI, 3.76–25.27; P = .002) and SH alone by 16.26 minutes (95% CI, 5.50–27.01; P < .001)
• Activity counts during sleep: HICT-SH beat SH alone by 7,980 counts (95% CI, 2,026–13,934; P = .003)

The authors conclude that the combined intervention improved objective and subjective sleep outcomes and cardiometabolic health more than each intervention individually [s1].

How much this can carry

Four constraints govern the reading.

Sample size. 112 participants across four arms is roughly 28 per group. Confidence intervals this wide — the wake-after-sleep-onset comparison spans 3.76 to 25.27 minutes — reflect that.

Population. Sedentary Chinese women aged 18 to 30 with poor sleep quality. Not people with diagnosed insomnia disorder; not men; not older adults; not people with sleep apnea, depression, or shift work. The baseline entry criterion was a PSQI above 5, which captures poor sleep quality broadly rather than a clinical sleep disorder.

Duration. Eight weeks, with the trial conducted over a two-month window in mid-2024 [s1]. No follow-up beyond the intervention period is reported in the abstract, so nothing here speaks to whether the gains persist once supervised circuit training stops.

Blinding. Single-blind. Participants necessarily knew whether they were exercising three times a week in a laboratory. That matters most for PSQI, which is self-reported, and least for actigraphy, which is not — and it is the actigraphy measures where the combination arm separated from the single-component arms.

Effect size in absolute terms. Fifteen or sixteen fewer minutes awake after falling asleep is a real but modest change. A 2.75-percentage-point gain in sleep efficiency is likewise on the order of ten to fifteen minutes across a night.

What is genuinely useful here

The trial answers a question that matters for how sleep interventions are designed rather than for what any individual should do. Behavioural sleep medicine tends to be delivered as a single package — usually some form of CBT-I — and exercise tends to be recommended alongside it as general advice rather than as a specified, dosed component. This trial suggests the two are at least partly additive rather than redundant, and it suggests the additive part shows up on objective rather than subjective measures.

An accompanying commentary in the same journal frames the finding in terms of precision behavioural interventions [s2].

What to watch

Replication in a clinical insomnia population, in men, and in older adults; and whether the difference survives a maintenance period after supervised training ends.

The trial was registered with the Chinese Clinical Trial Registry as ChiCTR2400086853 [s1].

This article describes a research finding and is not medical advice.

Sources

• [s1] Zhang B, Zheng C, Liao Q, et al. High-Intensity Circuit Training Plus Sleep Health Intervention for Sleep Improvement: A Randomized Clinical Trial. JAMA Network Open, published online 2026-02-16.
• [s2] Seixas AA, Chung D, Jean-Louis G. Synergistic Sleep Health and Exercise Interventions for Sleep and Cardiometabolic Outcomes—Advancing Precision Behavioral Interventions. JAMA Network Open, published online 2026-02-16.