Yoga as interoceptive training
Yoga occupies an awkward position in this wiki: it is by far the most widely practised thing that plausibly trains interoception, and it is among the least well characterized. The reason is structural. Yoga is simultaneously exercise and contemplative practice, and almost no study specifies the mixture.
The two hypotheses, and why they are testable
Exercise hypothesis. Postures move the body in ways comparable to other physical activity and raise heart rate into the lower bound of the American Heart Association’s cardiovascular recommendation (Cowen & Adams 2007). Physical activity reduces depressive and anxiety symptoms and stress hormone levels; a single 20-40 minute bout reduces anxiety for hours. On this account yoga’s mental-health benefits are exercise’s benefits, and the contemplative packaging is incidental.
Contemplative hypothesis. Breath regulation, mindful awareness of bodily signals, and cultivated tolerance for somatic sensation and its associated emotion deliver self-regulatory benefit beyond conventional exercise — the mechanism Farb et al. (2015) frame as a shift from active-inference toward perceptual-inference.
These are separable because real yoga traditions weight them differently, which makes an isomorphic two-arm trial possible. That trial is Eusebio et al. (2022), and its verdict is mixed in an instructive way.
What the one controlled test found
| outcome | interoceptive arm vs movement arm |
|---|---|
| self-reported interoception (maia) | no difference — both rose |
| symptom burden (PHQ-SADS) | no difference — both fell |
| sustained attention (sart accuracy) | interoceptive arm better (p = .006, corrected) |
| attentional lapses (tau), post-error slowing | no difference |
So the contemplative hypothesis is supported on a behavioural transfer measure and unsupported on every self-report outcome — including the self-report of interoception itself, which the interoceptive curriculum was explicitly designed to raise.
Two smaller precedents point the same way. Smith et al. (2011) found stress reduction in both a movement-focused and an interoception-focused programme, with anxiety and salivary cortisol reduction only in the interoceptive one, but participants self-selected. Schmalzl et al. (2018) found reduced stress and cortisol after both movement-focused and breath-focused yoga, with the anxiety-reduction/sustained-attention relationship appearing only in the breath-focused arm — the same shape as Eusebio et al.’s moderation finding, arriving four years earlier.
The finding this page should be organized around
Not “does interoceptive emphasis work better,” but for whom. Eusebio et al.’s two moderation results run in opposite directions across the arms: attentional benefit in the interoceptive arm scaled with interoceptive sensibility, while in the movement arm only the lowest-sensibility participants improved at all.
If that survives replication, the practical claim is not that one yoga is better. It is that interoceptive emphasis is the more inclusive option — it helps high-sensibility practitioners and does not appear to cost low-sensibility ones — while movement emphasis is contraindicated for people already disposed to attend to their bodies. That is an aptitude-treatment interaction, and it is the shape is-more-interoceptive-awareness-better has been groping toward with dose-response language that does not fit.
Held loosely: N = 34 analysed, one trial, one moderator measured with a composite MAIA that this wiki elsewhere argues should not be composited.
What is missing
No study in this wiki has compared yoga against non-yoga exercise matched for intensity, which is the direct test of the exercise hypothesis. No study has measured interoceptive accuracy before and after yoga of either kind. And the home-practice component that distinguishes traditional yoga from a gym class — prayer, formal meditation, philosophical study — has been absent from every trial here by design.