Your chest is tight, your jaw is tight, and the thing you are dreading starts in four minutes. Someone tells you to breathe, which you were already doing.
Here is the short version. One breathing pattern, five minutes a day, lifted people's mood more than meditation did over a month. One minute of it, used in a moment that was genuinely stressful, lowered anxiety compared with doing nothing. What nobody has shown is that any technique calms you faster than another, and the six-breaths-a-minute trick you see everywhere has the thinnest evidence of all.
How do you do it?
- Breathe in through your nose.
- At the top of that breath, take one more short sip of air through your nose.
- Let it out slowly through your mouth until your lungs feel empty. Take longer going out than coming in.
- Keep going for about five minutes.
Do three rounds now, before you read on. A long exhale slows the heart, though the trial measured only that this works, not why.
Does five minutes of breathing actually do anything?
Yes, but two questions get muddled together. Over a month, one pattern won. In the moment, nothing has beaten anything else. This section is the month.
108 people spent five minutes a day for a month on one of four things: three breathing patterns, or meditation. The pattern above won (Cell Reports Medicine, 2023), in a Stanford trial run by David Spiegel and Andrew Huberman.
Two things improved more than with meditation: day-to-day mood, and how slowly people breathed at rest. Now the part most articles get wrong. Anxiety in the moment fell in every group, meditators included, and no group beat any other on it. The longer-standing kind of anxiety did not move in anyone. Nor did heart rate, or the beat-to-beat variation your watch calls heart rate variability.
It was a small study: most of the 108 were psychology students at one university, from home in 2020.
Other labs have since looked at the fast version. German researchers had people breathe for one minute during genuinely stressful moments in real life, then measured them five minutes later. Anxiety was lower than when they did nothing, though reaction times were slower too (47 people, 2026). Researchers at Sussex and Oxford put 12 trials together and found a small but real drop in stress. Their own warning is worth repeating: the excitement about breathing has run ahead of the evidence.
What about box breathing, the one people call the Navy SEAL method?
Box breathing is equal counts: in for four, hold four, out for four, hold four.
The SEAL name traces to books and interviews by one former SEAL, not to any published study, and the US Navy's own handout cites no research and never mentions SEALs. Soldiers do train a version, though: German researchers put 100 trainee soldiers in a shooting simulator, and the group using the breathing shot better (2024).
Box breathing was one of the four things Stanford compared, and it lost to cyclic sighing, which has no holds at all. The closest test came this year, when researchers added a single hold to slow breathing rather than box breathing's two: it shifted some body measurements, but nobody felt calmer (32 people, 2026). Use it if counting helps. Do not expect the holds to add anything.
What about breathing six times a minute?
It has the thinnest evidence here.
Breathe in for five seconds and out for five seconds and your heart rate, breathing and blood pressure begin rising and falling in step. That much is real and well measured. You will see this called "coherence," which sounds like a scientific term. It is not one: the word comes from a single lab, funded by the company that sells the device.
That is not the same as going down, and it is where this falls apart. Across 22 trials and 17,214 people, machine-paced slow breathing did not lower blood pressure, and the researchers called breathing without a device an open question rather than a settled one. An earlier review found a small drop, then found it disappeared once the trials paid for by the device maker were removed.
The step you see everywhere, picturing something you appreciate while you breathe, is not in that trial either. It comes from the device company's manual. We repeated it for months and have taken it out; corrections are logged in our editorial standards. The trial paced people to their own heart rhythm, not a flat six a minute.
What calms you down in seconds?
Cold water on the face sets off the diving reflex, the one every mammal has, and your heart slows. Across the human studies, the heart starts slowing within about ten seconds. What was tested is a face held in cold water while holding the breath, not a splash at the sink. A splash is a lighter dose of the same trigger, and the warnings below cover both.
This one is not fast at all, but it has the strongest hard number. Forty adults did 20 sessions of belly breathing over eight weeks and ended with less of the main stress hormone in their blood than the group that did nothing. A machine also paced them at about four breaths a minute, slower than the six above and a different protocol from it.
Who should not do this?
Slow breathing is about as safe as anything gets. The exceptions are specific, so read them once.
Start with who the Stanford trial left out, because that is who "safe for most people" quietly excludes. The researchers turned away anyone with heart disease, glaucoma, a history of seizures, pregnancy, a serious mental health condition or a substance use problem. They were being careful rather than reacting to anything that went wrong, and the caution mostly belongs to the fast, forceful kind of breathing they were also testing. If you are on that list, ask your doctor.
Cold water on the face is the one trick here with a documented heart risk. Some people are born with a fault in the heart's electrical timing, called long QT syndrome, and never find out. Doctors unmask it by having the person put their face in cold water, because it reliably provokes the problem. Among 54 healthy teenagers who tried it, eight had extra or skipped heartbeats, and the heart reacted more strongly to cold on the face than on the whole body (2025). Those authors call the overall risk low for healthy people. The sensible reading is a filter, not panic: if anyone in your family has had an unexplained blackout or a sudden cardiac death, skip this one and talk to a doctor.
Never do breath holds or fast breathing in or near water. People have drowned exactly this way, because fast breathing removes the urge to breathe without adding oxygen (US CDC). That is why the Wim Hof method, rapid breathing followed by breath holds, has no place in a bath, a pool or the sea. It is not what Stanford tested either, and the same caution covers cold water generally.
And if slowing down makes you feel worse rather than better, you are not doing it wrong. Relaxation makes some people more anxious instead of less, most often people who already live with anxiety (2019). Open your eyes, look around the room, and stop.
What should you do tomorrow?
Do not build a practice. Do one thing.
Pick a moment tomorrow that is already stressful and already in your calendar. Just before it starts, do five double inhales with long slow exhales. About ninety seconds. Then walk in.
Do that for three days before deciding anything. If it helps, stretch it to five minutes a day for a month, the length of the trial. If it does nothing, you are out four and a half minutes in total, and now you know.
Since meditation keeps coming up: breathing did more for mood in the Stanford month. That trial says nothing about anything else, and a separate set of five trials this year found meditation beat slow breathing for pain. Different tools, not rivals, and here is how I fit both into an ordinary day.
