Box breathing is four counts in, four holding, four out, four holding again, repeated for a few minutes. You'll see it called a Navy SEAL technique. What it is underneath the branding: one way of getting your breathing rate under ten breaths a minute, which is the rate at which measurable things start happening.
Does it work? Yes, modestly, and the honest number is worth having. Pooled across twelve randomised trials and 785 adults, breathwork moved stress with an effect size of g = −0.35. That's small to moderate. It's also statistically solid, and most of those studies carry a moderate risk of bias by the review authors' own assessment. So: a good use of four minutes, and not a reset button.
The correction that matters more is about timing. Box breathing gets sold as an emergency measure for the moment you're already coming apart. It works better as a fixed thing you do before the hard morning, on a schedule, whether or not you feel bad yet.
What the counting is actually for
The square isn't magic. The rate is the point.
A systematic review of slow breathing defines the territory as anything under ten breaths a minute, and reports that below that rate the balance tips toward the parasympathetic side, mediated by the vagus nerve. Four counts a side puts you near four breaths a minute. You land inside that range without having to think about rates at all, which is the entire practical value of a counted pattern: it does the arithmetic for you while you're too tired to do arithmetic.
A meta-analysis of 223 studies found vagally-mediated heart rate variability goes up during voluntary slow breathing, immediately after a single session, and after a multi-session programme. Its authors describe the technique as low-tech, low-cost, with few adverse effects expected. That's about as enthusiastic as meta-analyses get.
Now the part most explainers leave out. One trial ran four groups for 28 days, five minutes a day: cyclic sighing with an extended exhale, box breathing, cyclic hyperventilation, and mindfulness meditation. The exhale-focused group came out ahead on mood. Box breathing was not the standout, and no group showed a significant change in heart rate variability at all.
So the defensible claim is narrower than the internet's. Slow breathing does something. The box pattern is a reliable way to breathe slowly. It is not the pattern with the best evidence behind it, and anyone telling you the square specifically regulates your nervous system is ahead of the data.
Why a rescue tool is the wrong idea
Here's the argument this piece is making, and you can disagree with it.
A technique you reach for only when you're already spiralling is a technique you'll be worst at exactly when you need it. Counting requires working memory. Holding an empty breath while your chest is tight is unpleasant. You'll do two cycles, decide it isn't working, and stop. Now you've also got personal evidence that it doesn't work, which is what you'll remember next time.
Move it earlier. Same four minutes, done at a fixed point in the morning, before anything has gone wrong. It stops being a rescue and becomes a marker: after this, the day starts. The practice is then something you already know how to do on the mornings you need it, instead of a thing you're attempting for the first time under load.
Fixed point means attached to something, not scheduled by intention. After you put the kettle on. Before you open the laptop. Sitting in the car before you go in. The specific anchor matters less than it being the same one every day, because that's what stops the decision from being made fresh each morning by a version of you who would rather not.
We build a breathing app, and this argument is the one assumption baked into it. There's no panic button in AffirmateMe. What there is: a session you configure in advance, whose length is calculated from the pattern and the number of cycles, so you can size it to the four minutes you actually have before the day starts. That's a design position, and it's the one this article is defending. A tool you set up while calm is a tool that still works when you aren't.
Running your first session without an app
Sit down. Breathe out fully first. Most people start by inhaling, and are already too full by the second cycle.
Then: in for four, hold for four, out for four, hold for four with your lungs empty. That last empty hold is the one people forget, and it's the reason there are four numbers in the name. Repeat for about four minutes, which is roughly fifteen cycles.
Two things worth knowing before you start.
- The counts are a suggestion, not a standard. If four seconds leaves you gasping by cycle three, use three. The rate matters; the specific number doesn't.
- Your jaw and shoulders will try to help. They can't. If you're clenching through the holds you're adding tension faster than the breathing removes it.
Air hunger during the empty hold means the pattern is too big for you today, not that you're failing at it. Shorten the hold, keep the rest, carry on.
The first session usually feels like nothing much. That's the normal result and not a sign it's the wrong technique for you: what the research measures is the effect of doing this daily for weeks, not the effect of one attempt. The month-long trial above ran five minutes a day for 28 days before it had anything to report.
When to use 4-7-8 instead
Four in, seven holding, eight out. It's the same idea with the weight moved to the exhale, and it's the better evening pattern.
The reason isn't that a long exhale has some special property. That claim is genuinely contested: an experiment and its replication found no difference in heart rate variability between an even ratio and a doubled exhale, and of nine earlier studies, three found no effect, one favoured even timing, one favoured longer inhales and four favoured longer exhales. That is not a settled question dressed up as one.
The reason is simpler. The exhale-focused group in that four-arm trial is the one that improved on mood. And a long exhale is difficult to rush, so the pattern paces itself when you're tired and inclined to speed up.
Both patterns, counted for you
AffirmateMe runs box as 4·4·4·4 and the relaxing technique as 4·7·8, with an affirmation shown on each cycle and optional spoken cues. A box session takes 3, 5, 7 or 10 affirmations at one to five cycles each; the relaxing one is deliberately shorter.
When this doesn't work
The hold is where the honest caveats live.
If you have a respiratory condition, breath retention is not a neutral instruction and you should ask someone who knows your case. In panic disorder the picture is murkier than you'd hope: a scoping review of sixteen studies in adults with diagnosed anxiety disorders reports contradictory findings on the role of hyperventilation-style breathing, and holding an empty breath while you're already frightened makes some people worse rather than better. If that's you, this article is not the right source.
None of this treats anything. Slow breathing is not therapy, it doesn't replace care, and four minutes will not touch a situation that needs changing rather than calming. The effect it does have is small, measured against control conditions, in studies the reviewers themselves flag for bias.
What it's good for is narrower and still worth having: a repeatable few minutes that reliably slows your breathing down, at a time you choose, on a day that hasn't gone wrong yet. The seven-minute morning is the routine we build around exactly that.
If you want the affirmation half of this, the line you say while you're counting and why the schedule matters more than the wording, that's in how to use affirmations. And if the time of day you actually need this is the end of one rather than the start, the pattern changes: see the five-minute wind-down.