What kind of exercise, and how much, actually helps you sleep better?
How Much Exercise Improves Sleep After 60? 62 Trials Looked at Type and Dose
· By Anthony Falcone
Verdict: Keep it. Yes, exercise helps: across 62 randomised trials in adults over 60, it improved self-rated sleep quality by a moderate-to-large amount. Combining cardio with strength training ranked best, and the benefit peaked at a moderate weekly dose rather than rising forever.
What the researchers did
Sleep problems are common after 60: the authors cite estimates that they affect more than half of older adults. Exercise is often recommended, but earlier reviews disagreed on which kind of exercise works best, and how much you need was unclear. This review set out to answer both.
The authors searched four databases up to April 2025 and kept only randomised controlled trials in people over 60 that reported sleep quality. They found 62 trials with 5,005 participants, average age 68. Most used questionnaires such as the Pittsburgh Sleep Quality Index, which asks people how well they slept over the past month.
They sorted the exercise into nine types, including aerobic exercise, resistance (strength) training, the two combined, walking, yoga, tai chi and qigong. To compare doses across such different activities, they converted every programme into MET-minutes per week: the intensity of the activity (in METs, multiples of resting energy use) times the minutes done. The WHO activity guideline works out to roughly 600–1,200 MET-minutes a week.
Beyond simply asking whether exercise helps, they used a network meta-analysis to rank the exercise types against each other, and a dose–response model to find the weekly amount with the biggest effect. They also worked out the smallest change people would actually notice, and the dose needed to reach it.
What they found
Overall, exercise improved sleep quality compared with no exercise or usual care, with a standardised effect of 0.68. By the usual yardstick, 0.5 is a moderate effect and 0.8 a large one, so this sits in between. The trials varied a lot, though (heterogeneity of 80%).
Combined aerobic and strength training ranked first, with an effect of 1.71, followed by aerobic exercise alone (0.96) and strength training alone (0.85). Yoga, qigong, virtual-reality exercise and walking also helped (0.77, 0.74, 0.67 and 0.66), and tai chi least (0.45).
The relationship with dose was a curve, not a straight line. The effective range ran from about 320 to 1,400 MET-minutes a week, with the best overall dose at 990. A noticeable improvement needed around 660 MET-minutes a week. For the combined cardio-plus-strength programmes, the best dose was about 1,000 MET-minutes a week, which the authors translate into roughly three 42-minute or five 47-minute sessions.
Time mattered too. A noticeable improvement showed up after about 5 weeks, and the effect peaked at about 15 weeks before levelling off. Moderate-intensity exercise beat light exercise, although light exercise still produced a noticeable change. People who slept worst at the start improved the most.
When the authors repeated the analysis using only trials in healthy older adults, the overall effect was almost identical (0.70), and removing the four trials at high risk of bias didn't change the picture either.
What 990 MET-minutes looks like
MET-minutes are just intensity times time, so you can do the arithmetic yourself. Moderate activity such as brisk walking or cycling is around 3–6 METs. At 6 METs, 990 ÷ 6 = 165 minutes a week, and the 660 threshold for a noticeable change is 660 ÷ 6 = 110 minutes. At a lighter 3 METs it takes twice as long: 330 and 220 minutes a week. The authors' own table gives the same figures.
In plain terms, the best results came from roughly the amount of activity already in the WHO guideline, and going far beyond it did not add more benefit for sleep.
What this means for you
- If you want better sleep, train regularly and combine cardio with strength work: about three 42-minute or five 47-minute sessions a week was the best-ranked approach.
- You don't need extreme volumes. The benefit peaked around 990 MET-minutes a week (about 165 minutes of moderate activity) and didn't keep rising beyond that.
- Give it time: expect a noticeable change after about 5 weeks and the full effect after about 15.
- If you have a heart condition, joint problems or another medical condition, talk to your doctor before starting a new training programme.
How much to trust it
- The authors' own: MET values are standard estimates that don't account for each person's body, age or sex; the trials varied widely (heterogeneity of 80%); only English-language studies were included; and they call for larger, more uniform trials with standard measurements.
- Sleep was mostly self-rated with questionnaires. The review didn't measure deep sleep or REM with sleep recordings, so it says nothing direct about sleep stages.
- Many trials were small: 19 of the 62 had between 10 and 19 participants, and just over half (33) studied people with a diagnosed disease, although the healthy-only analysis gave a similar result.
- All participants were over 60. Younger adults may respond differently, and the authors rated the overall quality of the evidence as moderate.
- The paper contains a few apparent typing errors in its results (for example, a sign that doesn't match its interval), so we've left out the numbers those affect.
The study
Xiong Z, Yuan Y, Qiu B, Yang Y, et al. Optimal exercise type and dose to improve sleep quality in older adults: a systematic review and network meta-analysis. BMC Geriatrics, 2025. doi.org/10.1186/s12877-025-06607-z
Our Deep Sleep Protocol covers the rest of the night: when to see a doctor, the CBT-I rules sleep clinics use, a two-week sleep diary, an evening light and screen plan, and the honest evidence on magnesium and melatonin. See the protocol