Progressive muscle relaxation: the sleep technique with the best numbers

Of all the bedtime relaxation techniques, the one with the most trial evidence behind it is also the least fashionable: tense a muscle group, hold it, let it go, move on. No app required, no subscription, invented in the 1920s. It kept turning up at the top when we went through this research, which was not what we expected, and it makes PMR the quiet overachiever of the sleepmaxxing list: a century older than the trend, with better numbers than almost anything on it.

The headline number

A 2026 systematic review and meta-analysis pooled 31 randomized controlled trials covering 2,277 adults. Progressive muscle relaxation improved sleep quality with a standardized mean difference of −1.74, and a Pittsburgh Sleep Quality Index difference of −3.79 points. It also reduced anxiety and improved quality of life (Donato 2026).

For scale: the Cochrane review of bedtime music found a PSQI difference of −2.79 across 10 trials, and that’s generally treated as a solid result. On its face, PMR beats it.

Now the asterisk

That number needs handling with tongs, for two reasons.

The heterogeneity is enormous. I² was 92.1% for sleep quality and 94.2% for the PSQI. I² estimates how much of the variation between studies reflects genuine differences rather than chance; above 75% is conventionally called considerable. Values in the low nineties mean the 31 trials were not really measuring the same thing in the same way, and the pooled average papers over a lot of disagreement. The authors ran sensitivity analyses and report the direction held, but a tidy headline number from a very untidy evidence base deserves suspicion.

The populations vary, and the authors say so. Their closing line is that PMR’s effects “vary across populations and settings”. This literature is built largely in clinical settings rather than on ordinary poor sleepers, and people whose sleep is wrecked by pain, anxiety, and a ward at 3am have more room to improve than a generally healthy person having a rough fortnight. The effect in your bedroom is probably smaller.

So the honest version is: a large body of trials consistently points the same direction, the magnitude is uncertain, and the technique is free and carries no known risk. That combination is rarer than it sounds.

The finding that argues the other way

One result cuts against this, and it’s a good one. A network meta-analysis of 241 trials and 31,452 people separated out which components of insomnia therapy do the work. Sleep restriction, stimulus control, and cognitive restructuring helped. Relaxation was the single component the authors flagged as potentially counterproductive (Furukawa 2024).

Two caveats keep that from being fatal. The estimate was uncertain: its confidence interval still crossed no-effect, and the authors wrote “potentially.” And it was assessing relaxation as a component of insomnia therapy, judged by whether adding it improved a treatment package. That’s a different question from whether PMR helps a stressed person wind down.

Still, the proposed mechanism deserves respect: relaxation at bedtime can encourage you to stay in bed longer while awake, working directly against the components that do help. Which gives you a concrete rule, below.

Why it might work

The reasoning is old and plausible. Edmund Jacobson’s premise in the 1920s was that anxiety and muscular tension are the same event viewed from two directions, so you can work backwards — release the muscle and the mind follows. The modern framing is that PMR down-regulates both muscular tension and autonomic hyper-arousal, which is exactly the state that keeps a tired person awake.

There’s a second, less-discussed advantage. Unlike visualization or hypnotic suggestion, PMR doesn’t require you to believe anything. Your shoulder doesn’t care whether you’re skeptical. As we cover in the roundup of relaxation techniques, how suggestible you are strongly predicts which methods work on you, and mechanical techniques like this one are the ones that work regardless.

How to do it

The classic protocol runs bottom to top, tensing each group for about five seconds and releasing for around twenty, breathing out as you let go.

1. Feet and calves: curl your toes, hold, release.
2. Thighs and glutes: squeeze, hold, release.
3. Stomach and lower back: tighten, hold, release.
4. Hands and forearms: make fists, hold, release.
5. Upper arms and shoulders: shrug toward your ears, hold, release.
6. Neck and jaw: the one most people are unknowingly clenching all evening.
7. Face: scrunch everything, hold, release.

The whole pass takes ten to fifteen minutes. Two points that get missed: the release is the active ingredient, not the tension, so let go abruptly rather than easing off; and tense to about three-quarters effort, never to the point of cramping. Skip any area that’s injured.

And the rule that follows from Furukawa: do it once. If you finish the pass and you’re still awake and irritated, get up and do something dull in dim light rather than starting a second round. Lying in bed running relaxation drills at someone is precisely the failure mode that meta-analysis was worried about.

If you want it guided

Talking yourself through it works fine, but the counting and sequencing is one more thing to stay alert for, which rather defeats the purpose. Slumber has guided muscle-relaxation and breathing exercises alongside its 2,200+ sleep meditations and stories, so the pacing isn’t your job.

Said plainly, though: this is the one technique on the list that needs nothing from us. A free recording, or just the sequence above from memory, is the whole intervention.

---

Sources: Donato et al. 2026 · Furukawa et al. 2024 · Jespersen et al. 2022. The whole list, graded. This is general information, not medical advice — persistent insomnia is worth taking to a doctor.