15.4 — Meditation in the Scanner: Proven and Oversold

This is the most directly testable thing in the volume, it has been tested extensively, and the results are mixed in a specific and informative way.

And the field has a history worth knowing: the early studies were mostly small, mostly conducted by people who meditated, mostly without adequate controls, and they produced dramatic results that have not all survived.

The methodological problem

Before the findings, the reason to be careful.

Blinding is impossible. A participant knows whether they spent eight weeks meditating.

Which means most studies are comparing meditation against doing nothing — and any activity done for eight weeks in a group with an enthusiastic instructor produces improvements.

The good studies use an active control: another structured group programme of equal duration, credibility and attention. And a consistent pattern in the literature is that effect sizes shrink substantially when an active control is used.

Add to that: small samples, self-selected participants, publication bias, and researchers with a personal commitment to the outcome.

Several large reviews have concluded that the evidence base is weaker than the popular presentation suggests, and that many studies had significant methodological problems.

This does not mean it does not work. It means the honest summary is narrower than what is usually claimed.

What holds up reasonably well

Four findings, and these are the ones that survive active controls and replication reasonably well.

Reduction in anxiety, depression and pain. Structured programmes — the eight-week format is the most studied — produce moderate improvements, comparable to other active psychological interventions. Not larger. Comparable.

Which is a real and useful finding: it is an effective treatment, roughly as effective as the alternatives, and the alternatives are also effective.

Attention effects. Improvements on sustained attention tasks are found reasonably consistently in trained meditators, though the transfer to everyday functioning is less clear.

Reduced physiological stress markers. Measurable, and they are also produced by exercise and relaxation training.

And the effects are dose-dependent. Long-term practitioners show larger and more consistent differences than eight-week participants — which is what you would expect if the practice is doing something, and which is also confounded by the fact that people who meditate for twenty years may differ from those who do not.

What is oversold

Five claims, stated plainly.

Large structural brain changes from short practice. Early studies reporting substantial changes in grey matter after eight weeks have not replicated reliably, and a well-conducted larger trial failed to find them. The brain does change with practice — that is not in doubt — and the specific dramatic claims from short interventions are not well supported.

Meditation as superior to other interventions. The comparison studies do not show this. It is one effective option among several.

Meditation as universally beneficial. See below.

Specific numerical claims — percentage reductions in this or that — which are usually drawn from single small studies and repeated until they sound established.

And claims about specific brain regions being "activated," which are frequently overinterpretations of imaging data. Volume V covers why functional imaging is much harder to interpret than the popular presentation suggests.

The adverse effects

And this is the part that was suppressed for a long time and is now being taken seriously.

Meditation can cause harm in a minority of people.

Reported effects include: anxiety, depersonalisation, intrusive memories including traumatic ones, sleep disturbance, and in rare cases episodes requiring clinical care.

Surveys of meditators report unpleasant experiences in a non-trivial minority, and studies of intensive retreat participants report higher rates.

And the traditions knew this. 6.9's five hindrances and the classical warnings about practice going wrong exist for a reason, and the traditional context included a teacher who knew the person and a graded approach. The modern context frequently includes an app.

The practical guidance: if you have a history of trauma, psychosis, dissociation or severe anxiety, take advice before beginning an intensive practice, and prefer a short, guided, grounded approach to a long silent retreat. Volume VI covers this properly.

What the traditions claimed that has not been tested

And an honest note about scope.

The traditions do not primarily claim stress reduction.

Their claims are: that the sense of a separate self can be seen through (6.7); that craving can be substantially reduced (6.5); that a permanent change in character is achievable (13.4); and that specific states of absorption exist and can be entered reliably (6.9).

Almost none of that is what has been tested. The research has largely measured what is easy to measure in eight weeks with undergraduates.

And the few studies of long-term practitioners — including work with monks who have tens of thousands of hours of practice — do find unusual and large effects, and the samples are tiny and self-selected.

Which means: the traditions' actual claims are largely open, and the popular claims are the ones that have been tested and partly deflated.

What this book concludes

Five points.

It works, moderately, for what the good studies measured. Comparable to other effective interventions, not superior.

The dramatic claims are not supported and are repeated constantly.

It can harm a minority and this should be said before somebody starts, not after.

The dose matters and almost nobody in the studies had a real dose.

And the reason to do it is not the scan.

5.6, 6.9 and 9.5 gave the actual reason: it trains the capacity to notice a state before it becomes a story (6.8's gap), and that capacity is the intervention point for everything in Part 8.

Whether that shows up as a change in grey matter is genuinely secondary.

The honest recommendation

Ten minutes, most days, for six months, with the format from 6.9.

And judge it by a specific test rather than by how it feels or by what a study says.

Are you noticing the reaction earlier than you used to?

That is the thing being trained, it is directly observable to you, and it is the only measure that matters for the purpose.

What to carry from this page

Effect sizes shrink when compared against another activity of equal credibility. Which is the single most useful thing to know about this literature and about most intervention literature.

It is one effective option among several, and that is a respectable finding rather than a disappointing one.

It can harm a minority, the traditions knew, and the modern delivery has dropped the safeguards.

The traditions' actual claims have mostly not been tested.

And the measure is whether you are catching the reaction earlier. Not the scan.

Next: 15.5 — how to recognise, in one paragraph, when physics is being used to sell something.