Why the Nose

Most of what gets taught about the nostrils is half right, and the half that is wrong is the half everyone repeats.

Start with what the texts say, because they are more careful than their students. Every prāṇāyāma in the Haṭha Yoga Pradīpikā (हठप्रदीपिका) that names a route names the nose. Three of the eight — bhrāmarī, mūrcchā, plāvinī — name no route at all. There is no verse saying do not breathe through the mouth — it is never a prohibition, only an instruction repeated so consistently that the exceptions give it away. (Pancham Sinh's 1915 English does read "should not be expelled through the mouth" at 2.54, and that sentence gets quoted as a rule. It is his gloss of घ्राणेनैव. The particle is exclusive, not prohibitive, and it sits inside a verse that tells you to inhale through the mouth.) In ujjāyī (उज्जायी): मुखं संयम्य (mukhaṃ saṃyamya), closing the mouth, draw the breath through both nāḍīs — and the same phrase again at the head of bhastrikā. In sūrya-bhedana (सूर्यभेदन), in through the right nāḍī and out through the left. In nāḍī-śodhana (नाडीशोधन) — the text's own word is नाडीशुद्धि, nāḍīśuddhi — in through candra (चन्द्र) and out through sūrya (सूर्य), and back again.

The two techniques that do use the mouth are the cooling ones, and both send the exhale back out through the nose. Sītkārī (सीत्कारी): make the sīt sound in the mouth and exhale घ्राणेनैव (ghrāṇenaiva) — through the nose only. Śītalī (शीतली): draw the air in with the tongue — the verse says only that, the curling comes from the commentaries — then exhale slowly through the two nostril openings.

Why they cared has an answer now. It is not the answer usually given.

Seven people with drug-resistant epilepsy had electrodes put into their brains for surgical planning — amygdala and hippocampus in all seven, piriform cortex in five. An eighth had amygdala electrodes and was recorded while doing the behavioural task. This is essentially the only way anyone records human limbic activity directly, and it exists only because these people needed the surgery.

While they breathed, the recordings rose and fell with the breath. In the piriform cortex the slow signal itself tracked respiration. In the amygdala and the hippocampus what tracked it was delta-band power, rising on the inhale. Then three of them were asked to breathe through the mouth instead, and the locking dissipated.

So they went looking for it in behaviour. Sixty-two healthy people judged expressions on faces, split across three groups. The twenty-one breathing through the nose identified fearful faces faster when the face came up during an inhale than during an exhale — a difference of about twenty-eight milliseconds. Surprised faces, nothing — the effect was specific. Breathing through the mouth it vanished. In another group, memory retrieval was better on the inhale, again nasally and not orally.

The mechanism is presumed olfactory, and the presumption is worth naming as one. Zelano's own wording is that nasal airflow induces the oscillation, which then propagates. The sensor half comes from elsewhere: olfactory neurons in mice respond to pressure as well as to odour, which is the likeliest way the airflow gets converted into a rhythm. Mouth breathing moves the same air into the same lungs and skips it.

That is the solid part. Now the part that is real but smaller than advertised.

The nose alternates. A German physician named Kayser described it in 1895 — one nostril carries most of the airflow while erectile tissue in the wall of the other swells and partly closes it, and hours later they trade. It is under autonomic control.

Measured properly it is less tidy than the story. Thirty-three people recorded for twenty-four hours: the left-dominant phase averaged 2.63 hours and the right 2.17. The finding worth keeping is that the cycle ran about two hours awake and four and a half asleep, and that slower breathing went with a stronger cycle.

The rhinology literature, in a 2018 review, puts numbers on the rest of it. Around 70 to 80% of adults have a nasal cycle, but a genuinely regular alternation shows up in only 21 to 39% of people, and a phase runs anywhere from thirty minutes to six hours. Everybody swapping every ninety minutes is not what the measurements say.

And now the part that did not survive.

In 1983 a study reported that the dominant nostril goes with greater EEG activity in the opposite hemisphere — left nostril, right brain. That one was purely observational, on the spontaneous cycle. The claim people actually use came four years later, from the same lab: force the breath through one nostril and you raise the other side, demonstrated in five untrained subjects. Nearly every modern claim that left-nostril breathing is calming and right-nostril breathing is activating traces back to those two papers.

It has not held up. A 2022 EEG study in thirty people found the expected contralateral bias "was not supported." What it measured was spectral power falling in both hemispheres during both kinds of nostril breathing — symmetric, with right-nostril breathing leaving relatively more than left. Not a side-to-side switch. A high-density EEG study in 2025 found no lateralization of power at all. It found a difference in functional connectivity — and that ran ipsilateral, the opposite way round from the folk claim. A 2024 paper states its result in the title: the lateralization of the nasal cycle is not reflected in olfactory bulb volumes or cerebral activations. Recent studies summarising the field call the direction unsettled — contralateral in most, ipsilateral or absent in others.

The texts never made the hemisphere claim anyway. They speak of iḍā and piṅgalā, candra and sūrya, moon and sun, and they say to alternate until the nāḍīs are clean after three months. That is a claim about a practice, not about neuroanatomy, and it does not get better by being dressed in the wrong century's science.

What is left standing is enough. The sinuses make nitric oxide; that the nose then delivers some of it to the lungs rests on small studies with partial responders, so hold it loosely. Nasal airflow drives an oscillation in the limbic system that mouth breathing does not. The nostrils alternate under autonomic control, and the cycle slows in sleep.

Close the mouth. Breathe through the nose, including in the two practices that use the mouth on the way in. The texts have said so for five and a half centuries, and the reason turns out to be that the nose is not a hole.