What Happens in the Brain During Meditation? Dhyana on fMRI

What the brain does during meditation, region by region

When researchers slide a meditator into an fMRI scanner, they are not measuring calm. They are measuring where blood, and therefore attention, goes. The largest synthesis to date — Caitlin Baten and colleagues at the Sacchet Lab, published in Neuroscience & Biobehavioral Reviews in 2026 — pooled 34 studies and roughly 700 participants across four families of practice: focused attention, open monitoring, mantra, and loving-kindness. The shared signature was consistent. Across every practice type, activity rose in the insula, the brain's map of the body's internal state; in the hippocampus, seat of memory and context; in the superior temporal regions; and in the supplementary motor area.

The pattern is quietly telling. These are not the regions of rapture or transcendence. They are the regions of noticing — of registering a breath, a sensation, the small fact that the mind has moved. The authors call their work the most robust characterization of the neural mechanisms of meditative states so far. What it characterizes is not an exotic state bolted onto an ordinary brain. It is the ordinary machinery of attention, turned on itself and held there.

Patañjali, compiling the Yoga Sūtra in the early centuries of the common era, defined dhyana in a single line: the uninterrupted flow of attention toward one object. Read as a practice instruction rather than a doctrine, it describes almost exactly what the scanner sees — the insula and hippocampus warming as attention is brought back, again and again, to a single point. This is the first thing to understand about what happens in the brain during meditation: nothing mystical lights up. The lighting up is the returning.

The default mode network: the self that goes quiet

The more striking half of the picture is what switches off. The default mode network is a set of midline structures — the medial prefrontal cortex and the posterior cingulate chief among them — that activate when the mind is left to its own devices: planning, remembering, rehearsing, narrating the small autobiography of me. It is the network of mind-wandering, and mind-wandering, by most measures, is where a good deal of human unhappiness lives.

In 2011, Judson Brewer and colleagues at Yale scanned experienced meditators alongside matched novices across three practices. The main nodes of the default mode network were relatively deactivated in the experienced group — and stayed coupled to the regions of cognitive control even at rest, as if the watching had become a trait rather than a passing state. The tradition's name for this is sakshi, the witness — the awareness that observes a thought without being carried off by it. What Brewer measured was a brain in which the witness had, to some degree, stayed on after the eyes opened.

Patañjali opens the Yoga Sūtra by defining yoga as citta-vṛtti-nirodha, the stilling of the turnings of the mind. The default mode network is, in a fair approximation, those turnings made visible on a screen — and the work of quieting the mind is, in neural terms, the work of loosening that network's grip. Two traditions, separated by eighteen centuries and a category of instrument, pointing at the same machinery.

Does meditation change your brain, or just calm it?

A calm afternoon changes brain activity; it does not change brain structure. The question that matters for a practice is whether the change outlasts the session — whether anything is left in the morning. The most-cited evidence that something is came in 2011, when Britta Hölzel and colleagues took structural scans of sixteen meditation-naïve adults before and after an eight-week mindfulness course. In that short window, gray-matter concentration increased in the left hippocampus, the posterior cingulate, the temporo-parietal junction, and the cerebellum — regions tied to learning, emotion regulation, and perspective-taking.

Eight weeks. The honest caveats apply, and they matter: a small sample, a single program, and structural MRI is a blunt instrument for a subtle claim. But the direction has held across more than a decade of replication and review. The brain treats sustained attention the way it treats any repeated demand — by reinforcing the tissue that serves it. This is why the lineages insisted on daily practice long before anyone could image a hippocampus. The structure follows the repetition; it does not arrive on the strength of a few good sittings.

Gray matter, brain waves, and what shifts over weeks

Below the level of structure, the brain's electrical rhythm shifts too. Meditation reliably raises alpha and theta power — the slower oscillations of relaxed, inward-turned attention — and seasoned practitioners show altered gamma synchrony during certain practices. These are not exotic readings; they are the same rhythms that slow as you drift toward sleep, here held deliberately while the mind stays awake. The brain waves during meditation are, in a sense, the body's own evidence that something has changed register. The 2015 review by Yi-Yuan Tang, Britta Hölzel, and Michael Posner, still the standard map of the field, gathered these strands into a developmental account: attention regulation comes first, then body awareness, then emotion regulation, and last a shift in the sense of self.

That order is worth sitting with, because it is not the order a beginner expects. It says the brain does not learn meditation all at once, and it does not begin with peace. It learns in the same sequence the traditions teach: first the breath, then the body, then the feelings, then the one who has been watching them the whole time. The neuroscience did not design that curriculum. It found it already there.

From dhyana to samadhi: what the scans are measuring

The Yoga Sūtra draws a line from dhyana to samadhi — from sustained attention on an object to the state in which the effort of attending falls away and only the object, or only awareness itself, remains. The 2026 meta-analysis found something that rhymes with that progression: distinct practice types produced distinct neural patterns, yet all shared a core of self-monitoring and awareness. The scanner cannot photograph samadhi. What it can show is the substrate — the insula reporting the body, the default mode network releasing its hold, the witness staying lit.

It would be a mistake to read the neuroscience as proof that the tradition was right, or as a replacement for it. The fMRI does not tell you what to do at five in the morning when the mind will not settle. It only confirms that the thing the practice asks of you — rest attention, notice you have wandered, return — is doing something measurable to the organ doing the asking. This is also where the distinction between dhyana and modern mindfulness earns its keep: the tradition was never only after stress relief, but after that final shift in the sense of self the review places last. The word came first. The map arrived late, and mostly agrees.