Breathwork techniques: what the evidence actually shows

Part of our nervous system regulation guide cluster. If you already know which technique you want and just need to match it to the state you’re in, start here instead.

Where box breathing goes wrong

You get to the second hold and your chest tightens.

Four counts in, fine. Four counts holding, and somewhere around count three the thought arrives: I am not getting enough air. Which is the exact sensation you opened the app to escape. So you abandon it, and file the result under personal failure, because a technique used by special forces should not be defeated by you sitting on the edge of your bed.

That reading is wrong, and it is worth correcting early, because it is the single most common way people bounce off breathwork.

Breath-holding produces air hunger. Air hunger is one of the physical signatures of a panic response. For someone whose anxiety already runs through the body, a held breath is not a neutral pause. It is a rehearsal of the thing they are afraid of. Therapists say this out loud in anxiety forums, and people who tried box breathing say it back in almost identical words: it felt like mild suffocation, so they stopped.

Meanwhile the technique the trial singled out has no holds in it at all, and most people have never heard its name.

What breathwork actually is

Breathwork is deliberately changing the pattern of your breathing to change how you feel. That is the whole category. The pattern is the intervention.

This is what separates it from meditation, and the distinction is more useful than it sounds. In most meditation you leave the breath alone and change your relationship to it: you notice it, you let attention rest on it, you come back when you wander. In breathwork you change the breath itself and let the body respond. One works on attention. The other works on physiology.

The two overlap in practice, and plenty of people do both in one sitting. But if you have tried meditation and found the instruction to simply watch your breath maddening, breathwork is a different proposition rather than a gentler version of the same one. It gives you something to do.

The four techniques, and where the names came from

Four patterns account for nearly all of the breathwork you will meet online. They are not variations on a theme. The shapes are genuinely different, and the differences are the reason they feel so unlike each other.

Read the shapes and the pattern jumps out. Box breathing is a square: equal in, hold, out, hold. The 4-7-8 pattern is dominated by its hold and its long exhale. Cyclic sighing is two inhales stacked, then an exhale several times longer, at whatever length is comfortable. Cyclic hyperventilation is a run of fast full breaths followed by a hold on empty.

Two things follow from the shapes alone. If holds are what make you anxious, two of these four are built around a hold and one is built around a very long one. And if your instinct is that the longer exhale is doing the work, the evidence is on your side.

The trial that compared them

In 2023, a team at Stanford published a randomised controlled trial in Cell Reports Medicine that did something we could not find another example of. Rather than testing one breathing technique against doing nothing, it tested three of them against each other, and against mindfulness meditation.

The design is worth stating precisely, because almost everything written about this study rounds it off.

The headline result: breathwork produced greater improvement in mood and a greater reduction in respiratory rate than mindfulness meditation, and cyclic sighing, the exhale-dominant one, was the arm the authors singled out on both of those measures.

Notice what the result is not. It is not that meditation failed. Both breathwork and meditation reduced negative emotion, including state anxiety. The finding is comparative and fairly narrow: on mood and on respiratory rate, five minutes of daily exhale-emphasised breathing outperformed five minutes of daily mindfulness in this sample, over this month.

That is a genuinely useful thing to know. It is also a much smaller claim than the one that reached the internet.

What that trial did not test

Here is where the popular version and the paper come apart, and it matters, because the gap is where people end up disappointed.

4-7-8 was not in the study. The three breathing arms were cyclic sighing, box breathing and cyclic hyperventilation. Any article that ranks 4-7-8 against box breathing on the strength of “the Stanford study” is citing a trial that did not include it.

A rescue breath was not what was measured. The intervention was five minutes a day for a month. The version that spread, a couple of sighs to abort a spike of stress in thirty seconds, is a different thing being asked of the same pattern. It may well work. We could not locate a trial of the single-instance physiological sigh as an acute intervention, and that absence is not evidence either way. What it means is that the acute claim borrows credibility from a study of a daily practice.

The sample excluded a lot of people. For safety, the trial screened out anyone reporting moderate to severe psychiatric or medical conditions, naming heart disease, glaucoma, seizure history, pregnancy, psychosis, suicidality, bipolar disorder and substance use disorders. That is sensible trial design. It also means the result does not straightforwardly transfer to the readers most likely to go looking for a breathing technique.

It was not powered to rank the three breathing techniques against each other. The authors say so directly: the per-group samples were small, which limited the power to compare individual breathwork groups with one another, and the study was powered to compare breathwork combined against meditation. So “cyclic sighing beats box breathing” is not a finding this trial can carry. What it supports is that breathwork as a group beat meditation, and that cyclic sighing was the arm singled out on the two measures where that gap showed.

The box breathing arm was not 4-4-4-4. Participants took a CO2 tolerance test first and had their phase lengths set from the result, practising 3-4, 5-6 or 8-10 seconds a side depending on how long it took them to empty their lungs. Anyone counting an even four because “that is what the Stanford study tested” is doing something the study did not test.

Roughly one in ten had a bad time. Ninety percent of participants reported positive experiences with the exercises. Ten percent reported negative ones. That figure almost never survives into the coverage, and it belongs in any honest account, because it tells you that finding a technique unpleasant puts you in ordinary company rather than in a category of one.

None of this makes the study weak. It is a careful exploratory trial that answered a specific question, and its authors are candid about its limits. They say plainly that it was not pre-registered, and that they registered it retrospectively while planning a larger confirmatory trial that will be. The problem is entirely downstream, in the retelling.

Why the double inhale does anything

The second sip of air looks like an affectation. It is the functional part.

Sighing is not a habit, it is a reflex. Your body already does it spontaneously every few minutes, and the reason is mechanical: the small air sacs in your lungs gradually collapse during quiet shallow breathing, and a sigh reinflates them. Li and colleagues at Stanford and UCLA described this in Nature in 2016 while mapping the brainstem circuit that generates sighs, work carried out in mice. The circuit findings are murine. The housekeeping function of the sigh is general respiratory physiology.

So the double inhale is not two breaths for emphasis. The first fills the lungs the ordinary way. The second, shorter and sharper, pushes air into regions that had partly closed, which restores surface area for gas exchange. The long exhale that follows can then offload more carbon dioxide than it otherwise would.

The long exhale is doing something too, and it is the part that generalises: extending the out-breath relative to the in-breath is the common thread running through nearly every calming pattern, which is why it also turns up in our guide to what actually lowers cortisol.

It also explains why cyclic sighing is the easiest of the four to do badly and still benefit from. There is no count to get wrong. Someone in an anxiety thread described it to a stranger without using a single technical word, telling them to breathe in through the nose, take a little more air on top, then let it out slowly through the mouth, with no counting necessary. That is the whole technique.

The dose nobody puts on the poster

The most useful number in the Stanford paper is not which technique won. It is how long people did it for.

Set that against how breathwork is usually encountered. You read about a technique, you try it once while stressed, nothing dramatic happens, and you conclude it does not work on you. People describe exactly this arc: giving it sixty seconds, or three days, then stopping. One person who eventually got somewhere with it said the thing they had been doing wrong all along was stopping too early, and that setting a timer for five minutes was what changed it.

There is a second pattern worth knowing about, which is that the turn tends to come at around three weeks. In breathwork communities, people describe the first fortnight as effortful and somewhere in week three it stops being a task and becomes something they want. That is consistent with a month-long trial finding accumulating benefits, though it is a report from practitioners rather than a measured finding, and I would not want to dress it up as one.

The practical consequence is unglamorous. Breathwork is closer to a daily five-minute thing you keep doing than to a technique you deploy. Which means the hard part is not learning the pattern. It is remembering to do it on the ordinary days, when nothing is wrong and there is no obvious reason to.

Practise it when you are calm

This is the most valuable thing in this article, and it arrived from a therapist by way of an anxiety forum rather than from a paper.

If the only time you ever do a breathing exercise is when you are already at nine out of ten, your brain will learn the pairing. The technique stops being a way out of the state and becomes a signal that you are in it. People report this happening: the count that once helped begins to produce a flicker of dread on its own, because it now belongs to the worst moments of their week.

Practising while calm inverts that. You are building an association between the pattern and a settled body, on purpose, so that the pattern carries some of that settledness with it when you eventually reach for it under pressure. The people in those threads who got somewhere describe the same sequence: learn it when nothing is wrong, use it in mildly stressful moments, and only then expect anything of it in genuinely hard ones.

There is a clinical reason to be careful here too. A technique used purely to make a panic attack stop can become a safety behaviour, the thing you believe you must do or something terrible will follow. Safety behaviours tend to maintain anxiety rather than resolve it, which is why we do not frame any breathing pattern on this site as a way to stop a panic attack. It is not squeamishness. It is that the framing itself can work against you.

The honest version is smaller and more useful. A long exhale will not switch a panic attack off. It can keep one from escalating as fast, which several people describe as the thing that actually helped, and it gives you something to do with your body that is not fighting it.

When breathwork is the wrong tool

Sometimes the answer is that this is not your technique, and no amount of persistence will change it.

When your anxiety is already in your body. There is a split worth knowing about. If your anxiety runs mostly on external worries, turning attention inward to the breath can help pull you out of the spiral. If your anxiety is already interoceptive, built out of a monitored heartbeat, a tight chest, a scan for the next strange sensation, then instructing you to attend closely to your breathing adds fuel. This is the same territory as the window of tolerance, where the question is not which technique is best but how much input your system can take right now. Grounding through the outside world tends to work better for that pattern: the temperature of a surface, sounds in another room, the texture of what you are sitting on. People who discovered this split describe real relief at learning why the standard advice had never worked, rather than concluding again that they were doing it wrong.

When breathing goes manual and will not go back. A specific and genuinely distressing thing happens to some people after they are told to focus on their breathing. Breathing switches from automatic to conscious, and it stays there, sometimes for hours. Their own descriptions are mechanical and exact. They describe being stuck in manual mode, say it is exhausting, and report that it worsens at night when there is nothing else to attend to. It usually settles as attention moves elsewhere, and it settles faster with an external focus than with more breath practice. If it persists and is distressing, that is worth raising with a doctor rather than solving with a better technique.

When the instructions do not fit your body. In through the nose is standard advice, and it assumes a clear nose. If you are congested, or crying, that instruction is unavailable and no amount of willingness will supply it. Mouth breathing is a fine substitute for a practice you can actually do.

When counting is the stressor. For some people the count becomes the problem: another performance to get right, another thing to fail at. Cyclic sighing is the natural fallback, because it has no count. Two inhales and a long exhale, at whatever length is comfortable.

If you have tried several patterns and they all land badly, that is information rather than failure. Movement, cold water and grounding practices do similar regulatory work through different routes, and there is no prize for arriving via the breath.

Safety, plainly

Most breathwork is benign. The exceptions are specific and worth stating without drama.

Never do breath-holding or hyperventilation in or near water. This is the one genuinely serious risk in the category. Hyperventilating before a breath-hold suppresses the urge to breathe without adding meaningful oxygen, so a swimmer can lose consciousness underwater with no warning. Clinical references call this shallow water blackout, also termed hypoxic blackout, and define it as loss of consciousness during a submersion preceded by hyperventilation. Swimming ability is no protection against it. No breathing practice in a pool, a bath or the sea.

Cyclic hyperventilation is more demanding than the way it is usually described. Thirty fast breaths followed by a hold on empty, repeated in rounds, is a substantial physiological intervention, and people do lose consciousness doing it on dry land. Sitting or lying down, never standing, and never alone if you are new to it. When a practice is described as gentle and in fact involves thirty forced breaths and a breath-hold, repeated in rounds, trust the description of the mechanics over the adjective.

Lightheadedness is common and usually not alarming. It generally means you are moving more air than you need. Slow down, shorten the pattern, return to normal breathing. Tingling in the hands or around the mouth is the same phenomenon. If it does not settle within a few minutes of ordinary breathing, stop.

Some conditions warrant a conversation first. The Stanford trial excluded participants with heart disease, glaucoma, a history of seizures, pregnancy, psychosis, suicidality, bipolar disorder and substance use disorders. Those exclusions are a reasonable starting list for who should ask a clinician before beginning an intensive breathing practice. Slow, gentle, exhale-emphasised breathing is a much lower bar than breath retention or hyperventilation, and the two should not be lumped together.

If you are being treated for a respiratory or cardiac condition, your clinician’s advice outranks anything on this page.

Where to start

If you take one thing from all of this, take the shape rather than the brand.

Make the exhale longer than the inhale. That single change is what the arm the trial singled out has in common with the slow-breathing advice people arrive at independently when they abandon the counts. Everything else is refinement.

A reasonable first fortnight looks like this. Cyclic sighing, because it has no holds and no counting: breathe in through your nose, add a second smaller inhale on top, then let it out slowly through your mouth. Five minutes. Most days. Deliberately at a moment when nothing is wrong, so you are teaching your body the association you want. If five minutes is too long to start, three is not a failure, it is just a smaller dose than the one that was studied.

Add box breathing later, if you like it, and drop it without ceremony if the holds do not agree with you. That is not you failing the technique. It is a symmetrical pattern with two breath-holds meeting a nervous system that reads held breath as a threat, and the sensible response is to use a different shape.

Common questions

What is breathwork?

Breathwork is deliberately changing your breathing pattern to change how you feel. It covers slow-breathing patterns like box breathing and 4-7-8, exhale-emphasised practices like cyclic sighing, and more intense methods built on hyperventilation and breath retention. The defining feature is that you alter the breath itself rather than simply observing it.

Is breathwork the same as meditation?

No, though they overlap and are often practised together. Breathwork changes the breathing pattern to produce a physiological shift. Breath-focused meditation leaves the breath alone and trains attention on it. In a 2023 Stanford trial that compared them directly, five minutes of daily breathwork improved mood and lowered respiratory rate more than five minutes of daily mindfulness meditation, while both reduced anxiety.

What is box breathing and how do you do it?

Box breathing uses four equal phases: inhale for four counts, hold for four, exhale for four, hold for four. It is widely attributed to military training, which explains its reputation more than its evidence. Its two breath-holds make it a poor first choice for people whose anxiety includes air hunger or chest tightness, since holding the breath can reproduce those exact sensations.

How does the 4-7-8 breathing technique work?

You inhale for four counts, hold for seven, and exhale for eight. The long exhale and the extended hold slow your breathing rate substantially. A 2022 study of 43 healthy adults aged 19 to 25 found measurable immediate effects: heart rate and systolic blood pressure dropped in both groups, and in the non-sleep-deprived group high-frequency heart rate variability rose while low and very-low frequency power fell. That study measured physiology straight after the exercise rather than sleep or anxiety outcomes over time, and 4-7-8 was not included in the Stanford comparison trial.

What is the physiological sigh?

A double inhale through the nose followed by a long exhale through the mouth. It copies a reflex your body already performs spontaneously every few minutes to reinflate collapsed air sacs in the lungs. The second, smaller inhale reopens regions that had partly closed, which is what allows the long exhale to offload more carbon dioxide. Practised repeatedly for five minutes a day, under the name cyclic sighing, it was the exhale-focused practice the 2023 Stanford trial singled out.

Is breathwork safe?

Slow, gentle, exhale-focused breathing is safe for most people. The real risks sit with breath retention and hyperventilation. Never practise either in or near water, since hyperventilating before a breath-hold can cause loss of consciousness underwater without warning. Practise intense methods sitting or lying down. Lightheadedness and tingling are common and usually resolve by returning to normal breathing. Around one in ten participants in the Stanford trial reported a negative experience with the exercises, so disliking a technique is unremarkable.

Why do breathing exercises make my anxiety worse?

Three common reasons. Breath-holding produces air hunger, which for many people is indistinguishable from a panic symptom. If your anxiety is already focused on bodily sensations, attending closely to your breathing adds to it rather than relieving it, and external grounding tends to work better. And if you only ever practise mid-panic, your brain pairs the technique with panic, so it starts to trigger the state it was meant to ease. Practising while calm, and choosing a pattern without holds, addresses all three.

Sources

• Yilmaz Balban, M., Neri, E., Kogon, M. M., Weed, L., Nouriani, B., Jo, B., Holl, G., Zeitzer, J. M., Spiegel, D., & Huberman, A. D. (2023). Brief structured respiration practices enhance mood and reduce physiological arousal. Cell Reports Medicine. PMC9873947
• Vierra, J., et al. (2022). Effects of sleep deprivation and 4-7-8 breathing control on heart rate variability, blood pressure, blood glucose, and endothelial function in healthy young adults. Physiological Reports. PMC9277512
• Li, P., Janczewski, W. A., Yackle, K., Kam, K., Pagliardini, S., Krasnow, M. A., & Feldman, J. L. (2016). The peptidergic control circuit for sighing. Nature, 530(7590), 293-297. PMID 26855425
• Bart, R. M., Murray, B. P., & Lau, H. (2026). Shallow water blackout. StatPearls. PMID 32119507