Five Minutes Before You Speak

Sixty-six students come in, sixty-two of them fasted. They give a baseline — heart rate, an anxiety questionnaire, and saliva, drooled into a vial — and then rest forty minutes, to let everything settle to where it normally sits.

Then they are told they have five minutes to write a speech arguing they are the best candidate for a job, to be delivered over video to two judges who are dressed for it, take notes, and do not react. The judges are a recording. Nobody mentions that.

Before speaking, each of them spends five minutes doing one of three things. A third breathe in a square — four in, four held, four out, four held — off a timer on a screen. A third take one deep three-second inhale through the nose and a six-second exhale through pursed lips, once every thirty seconds, breathing normally in between. A third are told to breathe however they breathe.

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

The control group got stressed, which the study had to show before anything else counted. Heart rate up 4.4 beats. Anxiety up. Salivary alpha-amylase up 27.7.

The two breathing groups did not go up. Not less — not at all. Heart rate went down in both, 6.4 beats and 5.3. Anxiety flat in one, slightly negative in the other. Amylase moved 1.9 and −1.2 against the control's 27.7.

One caution on the anxiety number, which the paper is straight about and most write-ups are not. Compare the groups against each other at peak stress and they are indistinguishable. The control group happened to start out less anxious, so what the statistics caught was three different trajectories, not three different levels. The breathers did not end up calmer than the controls. They ended up where they began.

Three things did not move, and the paper says so plainly. Cortisol fell in every group along the ordinary daily curve; the task, broken up by breathing breaks, was never continuous enough to provoke the slow hormonal response. Heart rate variability showed nothing. And the cognitive scores were the same across all three groups — the breathing made nobody sharper and nobody duller.

That cognitive result is worth sitting with, if you have ever heard breathwork sold as a performance enhancer.

A fourth marker, secretory IgA, went up with the stressor in everybody, and the breathing did not touch that either.

A bigger version of the same question was run at Stanford three years earlier, over four weeks instead of five minutes. A hundred and eight people, five minutes a day, randomized to mindfulness meditation, cyclic sighing, box breathing, or cyclic hyperventilation with retention. All four improved mood. Breathwork beat meditation on positive affect. And cyclic sighing — the one with the long exhale — gave both the biggest daily lift in mood and the biggest drop in resting breathing rate, the two of them linked: the more someone's rate fell, the more their mood rose. It grew with days on the protocol instead of wearing off.

Here is the part nobody covering these studies mentions.

Box breathing is not in the old texts. Not the Haṭha Yoga Pradīpikā (हठप्रदीपिका), not the Gheraṇḍa Saṃhitā (घेरण्डसंहिता), not Patañjali (पतञ्जलि), not the yoga Upaniṣads. The Sanskrit name it travels under — samavṛtti (समवृत्ति), equal movement — looks like a twentieth-century coinage, most visibly in Iyengar's Light on Prāṇāyāma of 1981, which defines it as making all four parts of the breath equal. It is built on vṛtti (वृत्ति) — the word Patañjali uses for the movement of the breath, though Iyengar does not say so. It came into the study through the military and the police, who call it tactical breathing.

What the texts prescribe is lopsided. The Gheraṇḍa Saṃhitā counts sixteen in, sixty-four held, thirty-two out — that being its middling grade; it gives twelve for the lowest and twenty for the best, and the shape stays the same. One to four to two. The Yogatattva Upaniṣad (योगतत्त्वोपनिषत्) gives the same three numbers. The exhale is twice the inhale and the hold is four times it. The Haṭha Yoga Pradīpikā, which spends a chapter on prāṇāyāma (प्राणायाम), gives no ratio at all — hold yathāśakti (यथाशक्ति), as long as you are able.

So the lab found the square worked and the long exhale worked, about equally, over five minutes. The texts never taught the square. And the texts that do give numbers weight the exhale — which is also the shape that won the four-week study.

That is not a vindication. It is one narrow place where the old instruction and the new data land on the same thing.

If you want to use this. Five minutes, before the hard thing:

1. Sit. Breathe through the nose.
2. Inhale for three counts.
3. Exhale for six counts, through pursed lips.
4. Breathe normally for about thirty seconds.
5. Repeat from step 2 for five minutes.
6. Do one more minute after the hard thing is over.

Do not hold the breath if holding makes you anxious. The retention is not what moved the numbers.

Five minutes, once, changed a physiological response. Five minutes a day for a month changed mood, and kept changing it. Neither study is large.