Slow your breath and your meditation will deepen. You will have met this advice in some form, probably within your first week of practising. I have given it myself.

It contains two claims, and I had never properly separated them. The first is that slowing the breath changes your physiological state. The second is that the changed physiological state makes the awareness deeper.

The first is one of the better-evidenced things in this whole field. The second is close to untested, and the reason is interesting: not that anyone has looked and found nothing, but that nobody agrees what the second thing is or how you would measure it.

Artwork by Aditi.

There is no place where meditation happens

Before any of that, something worth getting straight, because it changes how you read every claim in this area.

There is no region of the brain that does meditation. There is not even one pattern that all meditation produces.

Fox and colleagues reviewed 78 functional neuroimaging studies of meditation, pooling 257 peak coordinates from 31 experiments and 527 participants. They compared seven styles. Four of them — focused attention, mantra recitation, open monitoring, and compassion or loving-kindness — showed reliably distinct patterns of activation and deactivation from one another. Three more showed differences that were only suggestive.

Some regions did come up across several techniques: the insula, the supplementary motor areas, the dorsal anterior cingulate, the frontopolar cortex. But the authors are explicit that this overlap is the exception rather than the rule, and their conclusion is that meditation should not be treated as one uniform state at all. It is a family of practices, and the brain patterns track what each practice is actually asking you to do.

That matters here because “deeper meditation” assumes a single dimension you can move along. Neurally, there may not be one.

What breathing actually reaches

Breathing is unusual among bodily functions in that it runs by itself and you can also take it over. That dual access is what makes it the standard handle.

The clearest evidence is autonomic. Laborde and colleagues pooled 223 studies of voluntary slow breathing in Neuroscience and Biobehavioral Reviews in 2022 — during the breathing itself, immediately after a single session, and after multi-session interventions. Vagally-mediated heart rate variability increased in all three. That is about as consistent as findings get in this literature, and it is why slow breathing turns up in so many protocols.

The second line is more surprising, and it is the one I find genuinely striking. Zelano and colleagues, in the Journal of Neuroscience in 2016, recorded directly from the brains of seven patients with electrodes implanted before epilepsy surgery, and found that the rhythm of breathing entrains oscillations in the piriform cortex, amygdala and hippocampus. They then ran behavioural experiments with around 60 people: fearful faces were identified faster when seen during inhalation than exhalation, and objects were remembered better when encountered during inhalation.

The part that matters for practice is what happened when they switched people to mouth breathing. Both effects faded. The entrainment is driven by airflow through the nose, not by the act of breathing as such.

So the tradition’s insistence on nasal breathing has something behind it, and the mechanism is not the one usually offered.

The part that holds up, and the part that does not

Here is the distinction I think the advice obscures.

The advice treats the two arrows as one. They are not the same kind of claim, and they do not have the same kind of evidence behind them.

You can be profoundly relaxed and not be meditating in any interesting sense. Most people reading this have fallen asleep in shavasana, which is maximal parasympathetic success and zero awareness. And it runs the other way too: clear, stable, strangely vivid attention sometimes arrives on an ordinary breath you never adjusted.

Physiological calm and depth of awareness are separate variables. The breathing literature measures the first beautifully. It has almost nothing to say about the second.

Why the second one is hard

The obstacle is measurement, which is the obstacle in most of this field.

I wrote a fortnight ago about a preprint I have not stopped thinking about. Lewis-Healey, Kringelbach, Canales-Johnson and Laukkonen recorded 33 experienced Transcendental Meditation practitioners, with a mean of nearly thirteen years of practice, against 33 matched controls. The meditators meditated for thirty minutes. The controls counted silently upwards in their heads for thirty minutes.

At the level of the brain measure the researchers were testing, there was no significant difference between the two groups in any frequency band.

What rescued the study was changing what they measured on the subjective side. Rather than a rating afterwards, participants traced how much effort they had felt, moment by moment, across the whole session. Reorganised around felt effortlessness instead of around the label “meditating”, a relationship appeared — and appeared more strongly in the practitioners.

I take two things from that. The state was real; the category was too coarse to find it. And depth, whatever it is, showed up as a subjective variable measured carefully over time, not as a physiological one. Nobody has run that design on breathing. Trace effortlessness minute by minute at six breaths per minute and at a natural rate, and you would have an answer. As far as I can find, it has not been done.

The prediction story, and what it is

There is a theoretical frame that makes the breathing question more interesting, and I want to introduce it carefully, because it is a frame rather than a finding.

The idea is that the brain is not a passive receiver of sensation. It continuously generates predictions about the causes of its sensory signals, compares them against what arrives, and updates. Karl Friston set out the most ambitious version of this in Nature Reviews Neuroscience in 2010, where it is described in terms of minimising variational free energy. The brain, on this account, is inferring the world rather than observing it — and it is inferring the body too, which is what interoception means.

Laukkonen and Slagter applied this to meditation in Neuroscience and Biobehavioral Reviews in 2021. Their argument is that deconstructive practices work by progressively disengaging anticipatory processing — that focused attention, open monitoring and non-dual practice sit on one continuum, each loosening more entrenched predictive habits, including the brain’s prediction of a self.

If that is right, part of why breath-focused practice works is unglamorous: the breath supplies a slow, regular, self-generated signal for attention to organise around. Not a sacred channel. A reliable clock.

Now the caveats, and they are not small ones. The free energy principle is not an explanation of conscious experience; it does not touch the hard problem, and Friston has not claimed it does. The meditation application is a hypothesis that fits the existing data, which is a much weaker thing than a hypothesis that has predicted new data. Laukkonen and Slagter say so themselves and call for the experiments. I find it the most useful frame available for thinking about why breathing is the handle. I would not cite it as a reason to believe anything.

One overclaim worth retiring

While I am here: meditation does not switch off the default mode network.

The finding people are reaching for is real. Brewer and colleagues, in PNAS in 2011, scanned experienced meditators against matched novices across concentration, loving-kindness and choiceless awareness, and found the core default mode regions — medial prefrontal and posterior cingulate cortex — relatively deactivated in the experienced group across all three. They also found stronger coupling between the posterior cingulate and control regions, both during meditation and at baseline.

That is a difference in activity and connectivity between groups. It is not an off switch, it was not the same across every practice, and a cross-sectional comparison of experienced meditators against novices cannot tell you which direction the causation runs. The sentence “meditation turns off the default mode network” has travelled much further than the study that is meant to support it.

What I actually do

Slow the breath, because the autonomic evidence is good and the cost is nothing. Through the nose, because the entrainment finding gives that a mechanism it did not have before. Longer out than in, which is the part of the folk advice the physiology supports best.

And then stop treating the calm as the achievement. It is the condition you set up, not the thing you came for. On the days the breath settles and nothing else does, the breathing still worked. It just was not the point.

I would still like someone to run the experiment.

Sources

1. Fox KCR, Dixon ML, Nijeboer S, Girn M, Floman JL, Lifshitz M, Ellamil M, Sedlmeier P, Christoff K. Functional neuroanatomy of meditation: a review and meta-analysis of 78 functional neuroimaging investigations. arXiv:1603.06342. arxiv.org
2. Laborde S, Allen MS, Borges U, Dosseville F, Hosang TJ, Iskra M, Mosley E, Salvotti C, Spolverato L, Zammit N, Javelle F. Effects of voluntary slow breathing on heart rate and heart rate variability: a systematic review and a meta-analysis. Neuroscience and Biobehavioral Reviews (2022), 138:104711. doi.org
3. Zelano C, Jiang H, Zhou G, Arora N, Schuele S, Rosenow J, Gottfried JA. Nasal respiration entrains human limbic oscillations and modulates cognitive function. Journal of Neuroscience (2016), 36(49):12448–12467. doi.org
4. Lewis-Healey E, Kringelbach M, Canales-Johnson A, Laukkonen R. Near-critical brain dynamics track effortlessness during meditation. bioRxiv preprint (2026). biorxiv.org
5. Friston K. The free-energy principle: a unified brain theory? Nature Reviews Neuroscience (2010), 11(2):127–138. doi.org
6. Laukkonen RE, Slagter HA. From many to (n)one: meditation and the plasticity of the predictive mind. Neuroscience and Biobehavioral Reviews (2021), 128:199–217. doi.org
7. Brewer JA, Worhunsky PD, Gray JR, Tang YY, Weber J, Kober H. Meditation experience is associated with differences in default mode network activity and connectivity. Proceedings of the National Academy of Sciences (2011), 108(50):20254–20259. doi.org