How Solid Is Breathing Science? Real, Young, and Nobody Can Patent a Breath

One famous breathing study, and only 27 people in the arm that made it famous

The double-inhale, long-exhale that swept through wellness feeds a few years ago - the "physiological sigh" - has a real study behind it, which is more than most breathing advice can say. It came out of Stanford, ran in a Cell Press journal, and found that five minutes a day of cyclic sighing lifted mood and calmed the body - the strongest all-round profile in the trial, across mood, negative feelings and breathing rate. It did not sweep the board: on anxiety, mindfulness actually did slightly better, −3.95 versus −3.85 points on a standard scale. ^[1]11Balban 2023, cyclic sighing trial Here is the part that rarely travels with the headline: the arm that made the technique famous analysed twenty-seven people. Not two thousand. Not four hundred. Twenty-seven.

The number that travels is the big one. The paper enrolled 108 people, randomised across three breathing techniques and a meditation control; the cyclic-sighing arm that carried the finding analysed 27. ^[1]11Balban 2023, per-arm N=27 The big number makes the better headline - but the size of that single arm, not the enrolment total, is what tells you how solid one technique's result is. This is the exact seam where the hype outruns the evidence: a careful five-minute study, amplified by press coverage, ^[2]22Stanford Medicine press feature became a household certainty passed around as if it were pharma-grade proof.

That is not a knock on the study, which is a perfectly respectable small trial, and the fuller read on it lives in our cyclic sighing article. It is a knock on what the rest of us did to it. Twenty-seven is not a scandal. It is what a science looks like when nobody on Earth can make money proving it works.

223 studies behind HRV, single small trials behind the rest - still unproven on mood

Widen the lens from that one study and breathing science sorts into two clean piles. Where the outcome is something a machine can measure, it is already solid: slow breathing at around six breaths a minute reliably raises heart-rate variability - a marker of how flexibly your heart tracks your nervous system - a result pooled across 223 studies read off a monitor, not a questionnaire. ^[3]33Laborde 2022, 223 studies

Now the other pile. The controlled trials are recent and small: more than half of the studies in the field's own breathwork meta-analysis were completed from 2020 onward, ^[4]44Fincham 2023 meta-analysis and in the wider yoga literature - most of it breathing-based - 87% of randomised trials appeared only after the year 2000. ^[5]55Cramer 2014 The samples run in the tens; the single biggest trial the field has ever mounted enrolled 400 people, against the roughly 2,000 that statisticians want before they call a result settled. ^[6]66Guyatt 2011, GRADE imprecision

Set that against what breathing science has not shown. The mood lift from cyclic sighing - the headline effect - rests on that one study ^[1]11Balban 2023 and has not been independently reproduced. The biggest, best-built trial in the whole field - 400 people doing coherent breathing - found no special edge for that exact pace over a sham pace engineered to feel just as real. ^[7]77Fincham 2023 RCT, N=400 Box breathing, the "Navy SEAL" technique, was one arm in that same Stanford study, which was too small to prove any technique beat another. ^[1]11Balban 2023, box-breathing arm None of that means breathing does nothing: measured against doing nothing at all, breathwork still helps - a small but real pooled effect of about g = −0.35 on stress and mood. ^[4]44Fincham 2023 meta, g=-0.35 It means the strong, specific, replicated claims are rarer than the marketing implies. Thin is not empty. It is thin.

Nobody can patent a breath, so the big trials never come

Why does a field with real signal stay stuck at 27-person arms and a 400-person ceiling? Not laziness - the structural absence of money. A new drug earns a multi-million-dollar pivotal trial because a patent lets whoever paid for it recoup the cost. A breathing pattern has no patent. You cannot own "inhale twice, exhale slowly." No patent means no company stands to make its money back, so no company funds the big trial, so the big trial never happens. ^[8]88Gøtzsche 2018

The bill falls to public science, and the public purse for this corner is small. The one U.S. government centre that funds this kind of work runs on about $170 million a year - and that budget covers acupuncture, massage, meditation, tai chi, yoga and supplements as well as breathing, of which breathing is a sliver. ^[9]99NCCIH annual budget That centre is roughly a third of one percent of the National Institutes of Health's ~$47-48 billion, ^[10]1010NIH budget overview which in turn sits beside something like $96 billion a year in U.S. pharmaceutical research and development. ^[11]1111PhRMA R&D spending None of these are matched comparisons - breathing has no line item to set against a drug - only the scale, stepped from one public budget to the next. The shape of the problem is plain.

Be careful what that shape is, though. There is no published figure that says "breathing is underfunded by X," and anyone who quotes one has invented it. The honest claim is structural: no owner, no sponsor, no big trial - by design. Which quietly turns "the evidence is thin" from a complaint into a forecast. Of course it is thin. The funding structure guarantees it stays thin. Waiting for breathing to be proven to drug-trial standards is waiting for a bus that, on the route it is on, is simply never going to arrive.

We have been here before: jogging and home aerobics spread late, too

If that feels unfair, it is at least familiar. Cheap, effective, unpatentable self-care has always spread late, and mostly when a product carried it. Organised morning exercise feels timeless; it is barely a century old. Japan's national radio calisthenics - millions of people stretching in unison to a broadcast - began in 1928, dreamed up by a life-insurance bureau that had lifted the idea from a U.S. insurer's radio health broadcasts three years earlier. ^[12]1212Radio calisthenics, 1928 Jogging, as a thing ordinary adults simply do, was essentially invented in 1966 by a track coach's paperback; within a decade, tens of millions were doing it. ^[13]1313Jogging history, 1966 Home aerobics arrived on Jane Fonda's 1982 videotape, which outsold every other VHS in America for six straight years. ^[14]1414Jane Fonda's Workout, 1982

Notice the common thread. Each wave was free to practice, good for the body, and known for years - and each one waited for a book, a broadcast, or a tape to carry it into ordinary life. Where there was no product, there was no push. That is the same no-market drag that keeps the trials small, playing out in culture instead of in labs. It is not a reason for despair. It is a reason to expect slow adoption - and to build the thing that carries the practice.

Some of it is as proven as science gets - start with walking

Walk more, live longer is one of the best-supported findings in all of lifestyle medicine. Pool fifteen long-running studies that followed 47,471 adults and the pattern is clean: the more you walk, the lower your risk of dying over the study window, with the benefit levelling off somewhere around 6,000 to 8,000 steps a day for adults over 60 and 8,000 to 10,000 for the under-60s. ^[15]1515Paluch 2022, 15 cohorts A separate 2025 review across many outcomes landed on about 7,000 steps a day as a target worth aiming for, and found that even 4,000 beats 2,000. ^[16]1616Ding 2025, dose-response None of this makes us evidence-sceptics: when the case is genuinely strong, we say so flatly, and for some cheap bodily self-care it is about as strong as science ever gets.

The honest caveat rides along: these are observational studies, not randomised trials, so part of the effect is simply that healthier people walk more. The exact step count is a target, not a dose. But the direction is certain, replicated, and enormous - and that is what earns the top of our scale. It is also exactly why the scale means anything. When all we can muster for a technique is a 27-person single study, we do not lend it walking's certainty. Same ruler, read honestly, different marks.

How we grade a science this thin: an honest call, with the receipt shown

Every breathing technique in the Library carries a one-word grade for one specific job - Confident, Practical, Promising, Curious or Doubtful - and every grade ships with its receipt: how many people were really in that arm, what they were compared against, whether anyone has repeated it, and what it did not show. It is our honest call, shown with its working - not a number faking a precision this field does not have. The definitions, and why a plain word beats a score, live in their own read: how Brizzy grades, and what a confidence really is.

An open invitation to the labs: we can carry your intervention

If you run a lab studying any breathing technique, the two things that keep your studies small are two things we can hand you off the shelf: standardised delivery, and reach that does not depend on a lab visit. Brizzy is a web app that runs in any browser, on any device, with nothing to install, and it can be configured to your protocol's exact timing, ratios, dose, and feedback. Concretely, that means:

• the technique itself, built to your spec - inhale and exhale lengths, holds, pacing curve, session length, and the dose schedule across days;
• a matched active comparator - a sham or alternative pacer is trivially configurable, so your control arm can feel as real as your treatment arm;
• participant-facing guided sessions in the browser, with adherence logged automatically;
• recruitment-friendly delivery - because it runs in any browser with nothing to install and no clinic visit, people can take part from home, which lowers the practical barrier to enrolling and makes a geographically spread sample feasible (no promise on how many - only on the friction removed);
• participant dashboards, and clean data exported to you for your own analysis, with reporting support if you want it.

That matched-comparator piece matters more than it reads. Breathing cannot be blinded - a participant always knows whether they are breathing slowly or normally - so a fair trial has to build a believable sham pattern from scratch, which is a research project in itself. ^[17]1717Pitre 2023, unblinded inflation Configuring one is a checkbox for us. A believable sham - the kind the 400-person coherent-breathing trial needed so its paced control felt as real as the real thing - is exactly the research project we take off your plate. ^[7]77Fincham 2023 RCT, sham pacer

What we are not: we are not a lab. Ethics, IRB approval, and authorship stay entirely with you. We do not run your trial, own your findings, or provide medical oversight. You publish the trial; we carry the intervention. That division is the whole point - it is the honest form of a collaboration, not a quiet claim on your science.

And the field has a clear to-do list, the kind our own graded articles keep flagging: independent replications of the studies everyone already cites; head-to-head comparisons between techniques instead of technique-versus-nothing; dose-response - how much, how often, for how long; and simply bigger, better-instrumented samples than a field with no commercial sponsor has ever been able to afford. None of that needs a new molecule. It needs delivery standardised and participants reached - which is the one thing a digital platform is actually good at. If that is your work, tell us what you are testing. We will build the breath.

FAQ

Is breathing science fake, then?

No - it is strong where the outcome is objective, like heart-rate variability, and thin where it is subjective, like mood. Where a machine reads the result, slow breathing raising HRV is pooled across 223 studies off a monitor. ^[3]33Laborde 2022, 223 studies Where it rests on how people say they feel - mood, sleep, focus - the trials are small and often unrepeated. The field is real and early, not fake.

Why is the evidence so thin?

Because nobody can patent a breath, so no company can recoup a multi-million-dollar trial, and the big trials are left to public funding. The entire U.S. public budget for complementary health - breathing included - is about $170 million a year ^[9]99NCCIH annual budget - roughly a third of one percent of the NIH's ~$47-48 billion, ^[10]1010NIH budget overview and a rounding error beside the ~$96 billion in yearly pharmaceutical R&D. ^[11]1111PhRMA R&D spending The thinness is structural, not a sign the practices fail.

Should I wait for better evidence before trying a breathing technique?

For a five-minute, no-cost, low-risk practice, trying it is the rational move, not waiting - because the math is not a drug's. If we are wrong about a breath, the cost is five wasted minutes, not a side effect, and waiting for drug-trial certainty means waiting for trials the funding structure will never pay for. Start with the best-evidenced one and watch how your own body answers.

I run a study - what exactly does Brizzy provide?

A configurable delivery and measurement platform, plus the reach that comes with running in a browser: your technique built to exact timing, ratios, dose and feedback; a matched sham or comparator pacer; guided sessions any participant can complete remotely, with no lab visit and adherence logged automatically; participant dashboards; and your data exported for analysis. We are not a lab - ethics, IRB and authorship stay with you. You publish the trial; we carry the intervention. Tell us what you are testing.

Do you pay labs, or charge them?

Neither is the headline: we provide the platform as the collaboration, and the exact terms depend on the study. We would rather talk them through than pretend a one-size answer exists - the offer is genuine, and the arrangement is per collaboration. Reach us.