Breathing for sleep.

Slow breathing before sleep has a small, real evidence base. In 14 self-reported insomniacs, Tsai et al. (2015) found that when they practised slow paced breathing — six breaths a minute — for 20 minutes before bed, polysomnography recorded shorter sleep onset latency, fewer awakenings and higher sleep efficiency. In a randomised trial of 64 healthy young adults, 15 minutes of slow-paced breathing every evening for 30 days improved self-rated sleep quality and raised overnight cardiac vagal activity, compared with spending the same time on social media.

Neither study is large, and neither tested the protocol most people reach for. Here is what has been tested, what hasn’t, and what to do when the first night feels like nothing.

The canonical protocol: 4-7-8

Lying in bed, lights off or low. Inhale through the nose for four seconds. Hold for seven. Exhale through the mouth (lips slightly pursed) for eight. Four cycles is the standard dose the first week.

Andrew Weil, MD, popularised the protocol, drawing on yogic pranayama. Its rationale is the long exhale, and there is lab support for exhale bias: Van Diest et al. (2014) had 30 people breathe at 6 or 12 breaths a minute with either a short or a long exhale. The long exhale produced higher self-reported relaxation and stress reduction, and — at the slow rate only — more high-frequency HRV.

What 4-7-8 itself lacks is a sleep study. The closest trial we found, Vierra et al. (2022), measured the immediate effect of three sets in 43 young adults: heart rate and systolic blood pressure fell, and high-frequency HRV rose in those who weren’t sleep-deprived. Nobody’s sleep was measured. 4-7-8 earns its place here as the easiest long-exhale protocol to remember, not as the best-tested one.

If the first night feels like nothing, that isn’t a verdict. The randomised trial above ran for 30 evenings; judge it over weeks, not cycles.

If 4-7-8 feels too intense

Some people find the seven-second hold uncomfortable, particularly in early sessions. Two alternatives drop the hold.

Extended exhale 4/8. Inhale four, exhale eight. No holds. The same long exhale without the retention. Six to eight cycles is a reasonable dose.

Extended exhale 4/12.Inhale four, exhale twelve. We found no study showing that stretching the exhale past 1:2 does more for sleep. The case for it is practical: with no hold, it’s easy to sustain for the 15 to 20 minutes the studies above used.

What about resonance breathing (6 BPM)?

Here the evidence cuts against the usual advice, including the advice on this page. The best sleep data above is for slow paced breathing — six breaths a minute in Tsai’s study, which is the resonance range — not for 4-7-8. The case for the long-exhale protocols is Van Diest’s finding that a long exhale feels more relaxing. That is reasonable, but it is not a sleep outcome.

So if you already do resonance breathing during the day and want to keep the same rhythm before bed, you are doing the thing that has actually been tested. If you prefer 4-7-8 or an extended exhale, use them. The honest recommendation is whichever slow protocol you will do every night.

Don’t do this before bed

• Box breathing.We found no study of box breathing and sleep, in either direction. Its equal phases (4-4-4-4) give up the long exhale the rest of this page leans on. It’s a composure tool; save it for the morning or before a meeting.
• Wim Hof or any other hyperventilation-and-retention protocol. In Kox et al. (2014), volunteers trained in the method’s techniques — cyclic hyperventilation followed by breath retention among them — showed profoundly increased plasma epinephrine (adrenaline) while practising. That is the opposite of what you want at bedtime.

When breathing alone isn’t enough

If you’re reliably awake for 30+ minutes despite a consistent breath practice, the breathing protocol is not the lever. Check the inputs it can’t override first:

• Afternoon caffeine. In Drake et al. (2013), 400 mg taken six hours before bed still disrupted sleep compared with placebo. The authors recommend no substantial caffeine for at least six hours before bedtime.
• Evening alcohol. A review of every known study in healthy volunteers (Ebrahim et al., 2013) found that at all doses alcohol shortens the time to fall asleep, then disrupts the second half of the night; at moderate and high doses it cuts total REM sleep.
• Light-emitting screens. Reading on a light-emitting eReader in the hours before bed, rather than a printed book, meant taking longer to fall asleep, less melatonin and a later circadian clock (Chang et al., 2015).
• Irregular sleep timing. Among 61 students tracked for 30 days, the most irregular sleepers had their melatonin onset just after midnight; the most regular, around 9:30pm (Phillips et al., 2017). An association, not a trial.

Slow breathing won’t outweigh any of those. But don’t mistake this list for a treatment either: for chronic insomnia, the American Academy of Sleep Medicine suggests clinicians not use sleep hygiene on its own.

For chronic insomnia

Breathwork is not a treatment for chronic insomnia. The first-line treatment is cognitive behavioral therapy for insomnia (CBT-I), and it is the only strong recommendation in the AASM’s 2021 guideline.

Against sleep medication, a systematic review by Mitchell et al. (2012) found low-to-moderate-grade evidence that CBT-I works better in the long term, while very-low-grade evidence favoured benzodiazepines in the short term.

The same AASM guideline gives relaxation therapy only a conditional recommendation. A nightly breathing practice deserves no more weight than that: a reasonable addition, the wrong primary tool.