Five Minutes Before You Speak

Here is the experiment. Sixty-six university students arrive fasted. They rest ten minutes, then give a baseline: heart rate, an anxiety questionnaire, and a saliva sample, drooled into a vial.

Then they are told they have five minutes to prepare a speech arguing that they are the best candidate for a job, to be delivered to a panel of judges over video. The judges are professionally dressed, take notes, and do not react. They are also a recording, though nobody says so.

Before speaking, each student spends five minutes doing one of three things. A third do box breathing: four seconds in, four held, four out, four held, paced by a timer on a screen. A third do prolonged exhalation: one deep three-second inhale through the nose, then a six-second exhale through pursed lips, once every thirty seconds, breathing normally in between. A third are told to breathe however they normally breathe.

Then the speech. Then one more minute of their assigned technique. Then a Stroop task and rapid mental arithmetic. Then saliva, heart rate, and anxiety again — immediately, and twenty-five minutes later.

What moved

The control group got stressed, which is the first thing the study had to show. Their heart rate rose 4.4 beats per minute. Their state anxiety rose. Their salivary alpha-amylase — an enzyme that tracks sympathetic activation within minutes — rose 27.7 U/mL.

The two breathing groups did not rise. Not "rose less" — did not rise. Heart rate fell in both: 6.4 beats per minute in the box group, 5.3 in the prolonged-exhalation group. Anxiety was flat in the box group and slightly negative in the other. Alpha-amylase moved 1.9 and −1.2 against the control's 27.7. After the speech, the control group's amylase was 60 U/mL above the box group's.

Three things did not move, and the paper is straightforward about all three. Cortisol declined in every group, on the ordinary daily curve — the modified task, with its breathing breaks, was not continuous enough to provoke the slower hormonal response. Heart rate variability showed nothing. And cognitive performance was identical across groups: the breathing neither sharpened nor dulled anyone.

The larger version

A study from Stanford two years earlier ran the same question over four weeks instead of five minutes.

A hundred and eight people, randomized to five minutes a day for a month of one of four practices: mindfulness meditation; cyclic sighing (a double inhale through the nose, the second one topping up the first, then a long slow exhale through the mouth); box breathing; or cyclic hyperventilation with breath retention. Everything remote, with a wrist strap logging heart rate, heart rate variability, respiratory rate, and sleep.

All four improved mood and reduced anxiety. But breathwork beat meditation on positive affect, and cyclic sighing — the extended-exhale one — produced both the largest daily improvement in mood and the largest fall in respiratory rate. The two were linked: the more a person's resting breathing rate dropped, the more their mood rose. And the effect compounded with days on the protocol rather than fading.

Where the texts disagree

Now the interesting part, which the coverage of these studies always misses.

Box breathing has no classical source. Not in the Haṭha Yoga Pradīpikā, not in the Gheraṇḍa Saṃhitā, not in Patañjali or the yoga Upaniṣads. The Sanskrit name it circulates under — samavṛtti, "equal movement" — appears to be a twentieth-century coinage, most visibly in Iyengar's Light on Prāṇāyāma (1981), built out of Patañjali's word vṛtti. The technique reached the study through the US military and police, where it is called tactical breathing.

What the old texts prescribe instead is lopsided. The Gheraṇḍa Saṃhitā counts sixteen for the inhale, sixty-four for the retention, thirty-two for the exhale — one to four to two. The Yogatattva Upaniṣad gives the same three numbers. The exhale is twice the inhale; the hold is four times it.

The Haṭha Yoga Pradīpikā, which teaches prāṇāyāma for a whole chapter, gives no ratio at all. Its instruction for retention is yathā-śakti — as long as you are able.

So set the two claims beside each other. The lab found that equal counts worked and that a long exhale worked, roughly as well as each other, over five minutes. The texts never taught equal counts, and every text that does give numbers makes the exhale twice the inhale — which is also the shape that won the four-week study, where cyclic sighing beat both box breathing and meditation.

That is not a vindication. It is a small, specific convergence: where the old texts and the new data actually overlap, both land on the extended exhale.

The practical reading

Five minutes of deliberate breathing before something hard prevents the spike. It does not make you sharper — the arithmetic and the Stroop were unchanged, and anyone selling breathwork as a cognitive enhancer should read that table. What it buys is a nervous system that does not run away while you are working.

For the pattern: the evidence is thinner for the equal count than its popularity suggests, and the exhale-weighted form has both the older pedigree and the better four-week result. If you are choosing one, weight the exhale.

And note the dose. Five minutes, once, measurably changed a physiological response. Five minutes a day for a month changed mood, and the change grew. Neither study is large. Both point where the practice already goes.