Three cardio sessions a week reshape broken sleep

Sleep problems are usually judged by how long people sleep, but the structure of sleep matters just as much: how efficiently time in bed converts to sleep, how much of the night is spent awake after first dropping off, and how much deep slow-wave sleep the brain gets. Exercise is repeatedly proposed as a drug-free way to improve sleep, yet it comes in many forms, intensities and schedules. For adults with actual sleep disorders, it has been unclear which exercise prescription changes the architecture of sleep most, which is the question this review set out to settle.

The authors searched PubMed, Embase, EBSCO, Scopus, PsycINFO and Web of Science for randomised controlled trials published up to March 2025. Eighteen studies involving 1214 adults with sleep disorders were included. They ran conventional pooled analyses and a Bayesian network meta-analysis, which ranks different exercise modalities against each other even when trials never compared them directly. Subgroup analyses explored moderating factors such as programme length, weekly frequency and session duration, and evidence certainty and bias were formally assessed.

Exercise interventions improved sleep efficiency by a mean of 2.85, reduced wake after sleep onset by 10.16 and extended slow-wave sleep by 2.19. The dose pattern was specific: programmes of 8-12 weeks with 3 sessions a week of 45-60 minutes improved sleep efficiency and night-time waking most, while training 4 or more times a week was more effective at building slow-wave sleep, particularly in people with obstructive sleep apnoea. In the network ranking, moderate-intensity aerobic exercise came out as the most effective modality for sleep efficiency and wake after sleep onset.

This extends a familiar message, that exercise helps sleep, into harder territory: objectively measured sleep architecture in people with diagnosed disorders, rather than self-reported sleep quality in healthy adults. The finding that moderate aerobic work tops the ranking, rather than more intense options, is a useful counter to the assumption that harder training must mean better sleep. Because the modality comparisons are indirect, drawn across trials that differed in populations and designs, the rankings carry less certainty than the overall effect, and head-to-head trials remain the open gap.

For anyone whose sleep is genuinely disordered rather than just occasionally short, this is a reason to think regular movement changes the quality of sleep, not merely tiredness at bedtime. The pattern in these trials points towards moderation and consistency: an unremarkable schedule of moderate aerobic sessions, done a few times weekly for a couple of months, tracked with better sleep efficiency and less night-time waking. It is also a reason for people with sleep apnoea to note that more frequent training was linked with more deep sleep, though this is evidence of association within trials, not a prescription.