Physical activity, sedentary behaviour and interoception

An applied literature that arrives late to this wiki and immediately touches several of its oldest disputes. The claim in its strong form: interoception is not merely a capacity people have in different amounts but one that everyday behaviour modulates, upward with movement and downward with screens — which, if true, makes it a plausible public-health target rather than only a laboratory individual difference.

Two mechanisms, never yet separated

Everything in this literature is explained by one of two accounts, and Wallman-Jones et al. (2023) concede they cannot tell them apart:

Salience. Movement drives sympathetic activation, which makes visceral signals louder (Craig 2006). Easterbrook’s (1959) cue-utilization hypothesis adds that as arousal rises the number of usable cues narrows, so interoceptive sources crowd out exteroceptive ones. On this account interoception improves during exercise for the same reason a heartbeat is easier to count on a treadmill — the signal got bigger.

Attention. Goal-directed exercise directs attention to the body — pacing the breath, placing the limbs (Toner et al. 2016) — where automatized daily activity (walking upstairs, cleaning) has an externally-oriented focus. On this account the signal is unchanged and the spotlight moved.

The wiki should note how thoroughly these two are confounded in practice. Wallman-Jones et al.’s cleanest evidence for attention is that exercise beat daily-life physical activity (+4.48 vs +1.21 against a screen-time reference) — but exercise is also more intense, and intensity was not measured, so the same gap is exactly what salience predicts. Ekkekakis’s dual-mode theory of exercise, cited there, actually straddles the two: it proposes that attentional dominance shifts from cognitive to interoceptive cues around the anaerobic threshold, making attention a function of intensity rather than an alternative to it.

The salience account also has an uncomfortable relative on this wiki. Van der Does et al. (2000) showed that exercise manufactures accurate heartbeat perceivers above ~100 bpm who dissolve by ~95, and the wiki files that as the cardiodynamic confound — evidence that the heartbeat task measures signal amplitude rather than perceptual skill. This literature takes the same phenomenon and files it as a benefit. Both readings cannot be right without an account of when a louder signal is better perception and when it is merely a louder signal, and nobody has offered one. That tension is the most useful thing this page holds.

The chronic claim, and its embarrassment

The acute claim is that movement raises interoceptive contact in the moment. The chronic claim, from the group’s own theoretical review (Wallman-Jones, Perakakis, Tsakiris & Schmidt 2021), is that repeated exposure trains the pathway — physical activity as a naturally occurring form of interoceptive-exposure, with habituation altering anticipatory representations of bodily sensation — and that sedentary behaviour leaves an “untrained” pathway, degrading brain–body communication until perceived and actual bodily state diverge.

Supporting evidence, all held at one remove:

• Reduced heartbeat-evoked potentials in sedentary vs physically trained young adults (Perakakis et al. 2017 — a preprint).
• Attenuated anterior insula response to aversive inspiratory loading in elite athletes vs non-athletes (Paulus et al. 2012), read as more efficient adaptation to physiological perturbation.
• Negative correlations between BMI and interoceptive accuracy (Robinson et al. 2021 meta-analysis) — inactivity by proxy, not measured.
• Reduced resting-state insula connectivity with prolonged bedtime smartphone use (Paik et al. 2019) and altered cingulate connectivity in problematic smartphone users (Kwon et al. 2022).

The embarrassment is that the athlete finding is used twice, in opposite directions. Attenuated insular response in athletes is cited as showing that training improves interoceptive efficiency; then, when habitual physical activity turns out to predict lower self-reported interoception, the same attenuation is cited as showing that trained people habituate and stop noticing. A mechanism that explains a result and its reverse is not yet doing explanatory work. What the field needs, and does not have, is a criterion distinguishing efficient attenuation from decayed contact — the same problem in a different vocabulary as is-more-interoceptive-awareness-better.

The timescale reversal

The single most surprising result on this page: in Wallman-Jones et al., momentary physical activity raised self-reported interoception (B = 0.0025 per millig) while habitual weekly physical activity lowered it (B = −0.05, β = −.29), in one multilevel model on one sample. Their explanation is habituation: people accustomed to exercise-induced bodily change stop attending to it, freeing attention for the environment (Marshall et al. 2022).

The same shape appeared one level down in their own data — participants in the highest baseline-accuracy tertile showed decreases in self-reported interoception with acute activity where the lower two tertiles showed increases. Two independent expressions of “the more practised perceiver notices less,” which is at least internally consistent, and which is why the habituation account survives its circularity better than it might.

For what this does to the wiki’s construct pages, see state-vs-trait-interoception.

Screen time specifically

The claim that screen-based sedentary behaviour is distinctly bad for interoception — as opposed to sitting in general — rests on one study. In Wallman-Jones et al., both non-screen sedentary categories outscored screen time (non-screen social +1.13, non-screen solitary +0.67), which is the paper’s cleanest structural result because it dissociates posture from attention: the participants were sitting either way. Sleeping scored lowest of all (−1.73), which is unsurprising and mildly reassuring about the measure.

Read cautiously, it says attention captured by a screen is unavailable to the body — Pennebaker & Lightner’s (1980) competition of cues, a 45-year-old finding from exercise psychology arriving in interoception research. Read incautiously, it becomes the health recommendation the paper’s conclusion offers, on N = 70 self-report observations with no control questionnaire and no experimental manipulation. The wiki holds the first reading.

Where this connects

• interoceptive-exposure — if physical activity is a naturally occurring interoceptive exposure, the clinical literature’s deliberate version and the sport-science literature’s incidental one are the same intervention at different doses. Neither literature cites the other much.
• khalsa-2018-interoception-roadmap — the field’s consensus roadmap calls for interoceptive interventions; exercise is the cheapest candidate anyone has proposed and is essentially absent from that document.
• does-mindfulness-enhance-interoceptive-accuracy — the parallel case, and the cautionary one. Contemplative training also promised to raise interoceptive contact, delivered on attention and sensibility, and did not deliver on accuracy. Everything measured on this page so far is self-report. The accuracy version of the exercise claim has not been tested in daily life at all.