# Exercise and Sleep: What a Review of Trials Found

Published 2026-11-18, updated 2026-11-18. Online at https://thenuttracker.com/blog/exercise-and-sleep-quality-what-trial-reviews-found

Two reviews of randomized trials found people in exercise groups reported better sleep on questionnaires. Wearable-measured sleep moved much less, and one review found no clear change.

Workouts and sleep are two of the things people log most, so it's fair to ask what the trials say about how they connect. The answer depends on how sleep was counted. This post looks at two reviews that pooled randomized trials, what each one measured, and why the gap between "I slept better" and "the wearable agrees" matters. More posts like this live on [the blog](/blog).

## What did the reviews pool?

Two meta-analyses pooled adult trials in which one group exercised and another did not, and both found better scores on sleep questionnaires.

The larger one, a [2021 review in Frontiers in Psychiatry](https://pmc.ncbi.nlm.nih.gov/articles/PMC8215288/), covered 22 randomized trials with 1,806 adults. It left out people with other diseases and required programs of at least two months. The older one, a [2018 review in PeerJ](https://pmc.ncbi.nlm.nih.gov/articles/PMC6045928/), had 17 trials in its qualitative synthesis and pooled nine of them, with 557 participants.

| | 2021 review | 2018 review |
|:-|:-|:-|
| Trials pooled | 22 (14 for the main questionnaire) | 9 (6 for the main questionnaire) |
| Questionnaire used | PSQI | PSQI |
| PSQI difference, exercise vs comparison | 2.19 points lower | 2.87 points lower |
| Reported certainty | Not graded in the text we read | Low-quality evidence |

Lower PSQI means better reported sleep. A drop of two to three points on that scale is the headline both reviews report. The 2021 review also found lower scores on the Insomnia Severity Index (1.52 points) and the Epworth Sleepiness Scale (2.55 points), with fewer trials behind each.

## Does measured sleep change as much as reported sleep?

Much less. In the 2021 review, six studies used actigraphy, a wrist device that infers sleep from movement. Only one result reached statistical significance: sleep efficiency, up by about 1.2 percentage points. Time to fall asleep, total sleep time and time awake after sleep onset all leaned the favorable way, but their confidence intervals included zero. The authors concluded that exercise "primarily improved subjective sleep quality rather than physiological sleep quality."

The 2018 review saw less still. Across four trials with sleep lab or actigraphy data, sleep efficiency was 0.56 percentage points lower in the exercise group, with an interval running from 3.42 lower to 2.31 higher. Falling asleep took 1.9 minutes longer and total sleep time was 4.32 minutes longer, again with wide intervals. The authors state that exercise was not associated with improved sleep efficiency.

None of this shows the questionnaire gains are imaginary. It shows the devices didn't clearly pick up the same change. Wearables carry their own error, which is the subject of [how well sleep trackers match lab measures](/blog/sleep-tracker-accuracy-what-validation-studies-found).

## Does the type or length of exercise matter?

In the 2021 review, not by much for type. Physical exercise such as walking, cycling or aerobic classes lowered PSQI by 2.22 points across 6 trials. Mind-body exercise such as yoga, tai chi or qigong lowered it by 2.17 across 9 trials. The gap between the two groups was not significant (p = 0.95). Sessions ran from one to five times a week and lasted from 30 minutes to 2 hours.

Program length showed a difference: 2.95 points for programs of three months or less and 1.42 for longer ones. The authors point out that the exercise types differed between studies, so this does not establish that shorter programs work better. The 2018 review's included trials were mostly walking programs, lasting from a single session up to six months.

## What about evening workouts?

The two trial reviews did not analyze time of day, so they cannot answer it. The [Sleep Foundation](https://www.sleepfoundation.org/physical-activity/exercise-and-sleep) describes the traditional advice as avoiding intense exercise in the three hours before bed, notes that some studies found no negative effect, and says vigorous workouts in the hour before bed can affect sleep efficiency and total sleep time. It also describes one survey in which most people who exercised at 8 p.m. or later reported falling asleep quickly.

That's a mixed picture, and a survey of habits is a weaker design than a trial. Good reason to compare your own evenings rather than adopt a rule.

## Why hold these results loosely?

Because exercise trials can't hide who is exercising. The 2021 authors write that blinding participants was not possible, so every included study carried a high risk of performance bias, and self-reported outcomes cannot be blinded either. Only 12 of the 22 trials described adequate random sequence generation, and 5 described allocation concealment.

The 2018 review lists similar limits: most trials had a high or unclear risk of selection bias, the total sample fell short of the 400 participants its authors wanted, and only four studies reported adverse effects at all. Someone who has just started a program may rate their sleep differently, and a trial like this has no easy way to separate that from the sleep itself.

## Tracking it yourself

If you want to see your own pattern, log two things on the same scale: whether you worked out that day, and a sleep rating the next morning. A few weeks is enough to see whether the two lines move together or you're mostly reading noise. Keep the rating question the same each day, since the reviews show how much the wording of a self-report can matter.

[The Nut Tracker](/) keeps entries with a rating, tags and a note, then shows them in stats and a calendar, all stored on your phone with no account. Tag a workout, rate the night, and look at the weeks side by side. If you are curious how neighboring habits show up in logs, there is also [exercise and sex drive](/blog/exercise-and-sex-drive-what-studies-say) and [sleep and sex drive](/blog/sleep-and-sex-drive-what-the-evidence-shows). If sleep is a worry for you, that is a conversation for a doctor.

## Bottom line

In reviews of randomized trials, people in exercise groups reported better sleep, while measured sleep moved little and one review found no clear change. Treat the questionnaire gain as real but unconfirmed by the devices. More like this is under [Body and Mood](/blog/category/body-and-mood).

## Questions

### Does exercise improve sleep?

In two meta-analyses of randomized trials, people in exercise groups scored better on sleep questionnaires. Measured sleep (actigraphy or polysomnography) showed little or no clear change, apart from a small sleep efficiency difference in the larger review.

### What is the PSQI?

The Pittsburgh Sleep Quality Index, a questionnaire about sleep over the past month. Lower scores mean better reported sleep, so a lower score in the exercise group is the favorable result.

### Is exercise right before bed bad for sleep?

The Sleep Foundation says the traditional advice is to avoid intense exercise in the three hours before bed, that some studies found no negative effect, and that vigorous workouts in the last hour can affect sleep efficiency and total sleep time. The two trial reviews did not analyze time of day.

### Why do the questionnaire and the wearable disagree?

The reviews did not settle that. Participants know whether they are exercising, so a self-report can shift for reasons the trial cannot separate from sleep itself. Wearable measures have their own error.

## Sources

- [Frontiers in Psychiatry: Effects of Exercise on Sleep Quality and Insomnia in Adults, a Systematic Review and Meta-Analysis of Randomized Controlled Trials (2021)](https://pmc.ncbi.nlm.nih.gov/articles/PMC8215288/)
- [PeerJ: Exercise can improve sleep quality, a systematic review and meta-analysis (2018)](https://pmc.ncbi.nlm.nih.gov/articles/PMC6045928/)
- [Sleep Foundation: Exercise and Sleep](https://www.sleepfoundation.org/physical-activity/exercise-and-sleep)
