Anuloma Viloma Timer
The most expensive version of slow breathing did not clearly beat the cheapest one.
That is the result of a four-week randomised comparison published in Scientific Reports in May 2026. Eighty-eight adults were split three ways. One group breathed at their personal resonance frequency, the individually tuned rate (typically somewhere between 4.5 and 6.5 breaths per minute) at which a given person’s heart rate variability swings widest. A second group breathed at a fixed 0.1 Hz, which is six breaths per minute for everybody. A third group went on a waitlist. Both breathing groups improved HRV against the waitlist. The personalised group did not clearly outperform the fixed group on the primary endpoint.
It is a quiet finding, and it is the most practically useful sentence in this week’s reading. Not because slow breathing “works” (the waitlist comparison tells you that much), but because the trial asked a harder question: does the tuned, lab-calibrated version work better than the plain one? The answer was: not clearly. The value of the result lives entirely in what it was compared against.
That turns out to be the thread running through every paper that crossed the desk this week.
The question the skeptics are asking
Mainstream press has started asking whether any of this holds up. AOL ran a piece this month under the headline “Everyone’s Downloading Breathwork Apps for Neurowellness in 2026. Are They Actually Worth It?” It is a fair question, and the honest answer is not yes or no. It is: compared to what?
This week’s podcast, Biological Levers: New 2026 Science on Nasal Breathing and Heart Rate, walks through the mechanisms behind these papers (the baroreflex, the vagal brake, the pressure of a hummed exhale). This piece is about the fine print underneath them. Sort the week’s studies by their control groups, from weakest comparison to strongest, and a clearer picture falls out than any single paper gives you.
Rung one: compared to itself
The Bhramari study (PubMed 42521250, published July 2026) measured twenty adults before and after a single five-minute session of humming-bee breath. Low-frequency HRV power rose from about 300 to 4,843 ms². RMSSD, a cleaner marker of parasympathetic activity, rose from 29.95 to 45.83 ms. The EEG showed less delta and theta activity and more beta and gamma, the signature of an alert brain rather than a drowsy one.
Those are large, fast shifts. What the design can tell you is that the mechanism is real and easy to trigger. What it cannot tell you is whether the hum did anything a slow, long exhale without a hum would not have done, because no comparison practice was reported. A before-and-after study is a demonstration, not a contest.
Rung two: compared to nothing added
A step up are trials with a passive control, where the comparison group gets no added practice. The strongest example on the desk this week is a systematic review and meta-analysis from University Hospital Essen, published in August 2025, pooling six randomised controlled trials of pranayama in 517 patients with post-traumatic stress disorder, depression, or mixed non-psychotic disorders. Symptom severity after the intervention fell significantly versus passive controls, in both the intention-to-treat sample (everyone enrolled, whether they finished or not) and the per-protocol sample (only those who completed).
Surviving intention-to-treat is the harder test, and it passed. But passive controls produce the largest effect estimates in any field, because the practice is being compared with the absence of anything. The fair reading is “a real signal that belongs alongside treatment”, not “a treatment”.
The same frame applies to the one alternate-nostril trial with a clinical outcome this week. A randomised controlled trial in Primary Health Care Research and Development (Cambridge University Press) assigned people with migraine to nadi shodhana or a control group, and the practice group saw reductions in both the severity and the frequency of attacks. A separate 2026 paper, indexed on PMC, studied pranayama in recently diagnosed migraine with cardiac autonomic function as a measured outcome, which is the right instrument for asking why. Promising, specific, and still a comparison with doing nothing extra.
Rung three: compared to what you would do anyway
This is where the week’s evidence gets more interesting. A randomised trial in Medicine and Science in Sports and Exercise (July 2026) took 62 tactical-athlete cadets (44 men, 18 women) to maximal exertion, then had them recover with box breathing, cyclic sighing, or ordinary spontaneous breathing. Both structured techniques brought back high-frequency HRV, the parasympathetic recovery signal, faster than spontaneous breathing. The authors describe it as a “field-ready intervention”.
The control arm here is not a waitlist. It is the thing every person does after a hard effort: pant until the heart settles on its own. Beating that means the structure itself is doing work. And note the detail that matters most: two quite different shapes of breath, an even held box and a double-inhale sigh, both won. For this outcome, deliberately slowing and structuring the breath mattered more than which structure was used.
Rung four: compared to its nearest neighbours
The most rigorous design for the question practitioners actually ask is a randomised crossover trial in the International Journal of Yoga Therapy (February 2026). Forty-five volunteers each did five separate fifteen-minute sessions on different days, in random order: right-nostril breathing, left-nostril breathing, alternate-nostril breathing (anuloma viloma, named as such), plain breath awareness, and quiet rest. Every participant served as their own control, with autonomic variables recorded before and after each session.
This is the design that can answer whether alternating the nostrils does something that watching your breath, or simply sitting still, does not. I have not seen its effect sizes reported in the sources I work from this week, so I am not going to tell you which arm came out ahead. That is the point of this whole piece: before you believe anyone’s summary of a head-to-head trial (including ours), read which arm beat which, and by how much.
A sideways rung: compared to the tracker
One more comparison hides in plain sight. A systematic review in Sleep Medicine Reviews (June 2026) pooled nine studies, 457 participants in all, of slow breathing at ten breaths per minute or fewer before bed. Self-reported sleep improved in seven of the nine. Objective measures from actigraphy and polysomnography were mixed, with four studies showing no significant change in total sleep time. The six studies that measured HRV found slow breathing influenced both the parasympathetic and sympathetic branches.
People reported sleeping better while the devices often disagreed. The comparator here is not another practice. It is the choice of instrument, and it can flip the headline.
What the ladder says, read top to bottom
Put the rungs together and four things are visible that no single paper states.
First, deliberately slowing the breath beats leaving it alone, and that holds against realistic comparisons (spontaneous breathing after maximal effort), not only against a waitlist.
Second, fine-tuning the pace beyond “about six a minute” has not been shown to add much. The best-controlled trial of personalisation did not find a clear advantage for it.
Third, for at least one recovery outcome, two different breath shapes worked about equally well against doing nothing deliberate. Structure and slowness carried the effect.
Fourth, the open question is how much a specific technique adds on top of slow, structured breathing. That is exactly what the crossover design is built to probe, and it is the question behind a registered trial at the Medical University of Bialystok (NCT07529379), which is testing whether benefits come from the six-breaths-per-minute rhythm itself or from the experience of doing any structured activity. It has no posted results. It is a question, not a finding.
For alternate nostril breathing specifically, that makes the honest case narrower and sturdier than the one usually made. Nothing this week shows nadi shodhana beating every other breath. What the evidence does support is that it is a slow, structured, nasal practice that lands naturally at the cadence the best-controlled trial validated, and it has two features that help you actually hold that cadence: a single open nostril narrows the airway so you cannot gulp, and the alternation gives your attention a job. Add the one clinical trial that named it specifically, in migraine, and that is a reasonable place to stand.
What this means for your next session
Stop tuning, start counting. Aim for roughly six breaths a minute. In alternate nostril breathing, count one inhale on one side and one exhale on the other as one breath: about four seconds in and six seconds out gives you a ten-second breath and six a minute. If that feels strained, go four and five and let it lengthen.
Give yourself a comparison. You cannot run a trial on yourself, but you can borrow the crossover idea. Over the next three evenings, do ten minutes of one thing each night: alternate nostril breathing at six a minute, then both-nostril breathing at the same pace, then ten minutes of just sitting. Before and after each, rate how settled you feel from one to ten. If you wear a watch that reports HRV, note that too, and read the two side by side, because the sleep review shows they will not always agree. It is one person, unblinded, so it is not evidence. It will teach you what “compared to what” feels like from the inside.
After hard effort, do not leave it to instinct. Take what air you need for the first few seconds, then close your mouth and slow the breath down through the nose. The cadet trial suggests the minute after a hard set is when structure pays off most.
If you are managing a condition, bring it to your clinician as an addition. The migraine trial and the Essen meta-analysis both compared pranayama with nothing extra. That is grounds for a conversation, not a swap.
If you want something to hold the six-a-minute pace and call the nostril switches so your attention can stay on the breath, Anuloma Viloma Timer is free on the App Store, no subscription necessary.
The question to keep
Next time a breathing claim crosses your feed with a big number attached, ask the one question this week’s papers answer best. Compared to what? Compared to nothing, a result is a promise. Compared to what you would have done anyway, it is useful. Compared to the fancier version, and still holding up, it is something you can do tonight with a nose and a timer.