Interoceptive illusion
The claim that the body you feel can be manipulated the way the body you see can — that the internal senses, like the external ones, build a percept by resolving conflicting evidence against priors, and can therefore be led into a stable, confident misperception. If interoception is inference rather than read-out, an interoceptive illusion is the prediction the framework most obviously makes: feed the generative model false evidence and the estimate moves, even though the body did not.
The construct enters the wiki with Iodice et al. (2019), who coined “illusion of effort” for their result — false (faster) cardiac feedback making the same cycling feel harder — and positioned it as the interoceptive extension of the rubber-hand and enfacement illusions (manos-tsakiris). But the phenomenon predates the name, and the value of the page is that it collects examples the wiki already held under other headings.
The family (loosely, most-to-least clearly interoceptive)
| illusion | conflict exploited | what is misperceived | wiki page |
|---|---|---|---|
| Illusion of effort | false heartbeat sound vs actual HR | physiological effort / exertion | iodice-2019-illusion-of-effort |
| Thermal grill | interleaved warm + cool bars | thermal pain (burning from merely cool) | thermal-grill-illusion |
| False-feedback appraisal | bogus HR feedback vs stimulus | emotional salience / attractiveness | false-feedback-paradigm |
| Cardio-visual RHI / full body | flashing light synced to (or against) heartbeat | body ownership | rubber-hand-illusion, experience-of-body-ownership |
These are not one mechanism. The thermal grill is a peripheral/spinal labelled-line disinhibition (“cold inhibits pain”, Craig); the illusion of effort and the false-feedback appraisal effect are central multisensory conflicts resolved against a prior; the cardio-visual illusions sit at the interoceptive/exteroceptive seam. What justifies one page is the shared logic: in each, a percept of the internal state is decoupled from the internal state itself by a manipulable conflict — which is exactly the handle predictive coding says perception offers.
Why it is a method as much as a concept
Iodice et al.’s framing — echoing the exteroceptive illusion literature — is that illusions are tools: a place where the normally-hidden inference process becomes visible and manipulable. Interoception is chronically hard to study because its signals are weak, private and confounded with the physiology producing them (the cardiodynamic confound haunts half the wiki). An illusion sidesteps this: it holds the body fixed and moves the percept, so any change in report is inference, not signal. This is the same argument control-based and anticipation paradigms make from other directions — routes to interoception that do not depend on the participant accurately detecting a faint visceral event.
The asymmetry, and what it says about priors
The single most theoretically loaded fact about the illusion of effort is that it is one-directional: faster feedback inflates perceived effort, slower feedback does not deflate it (Iodice et al., with Bayesian support for the null). A symmetric sensory-integration model does not predict this; a cost-sensitive prior does. Underestimating effort risks overexertion, so the interoceptive schema resists downward revision while accepting upward revision — precision-weighting shaped by the asymmetric metabolic cost of being wrong. This connects to allostasis (the estimate serves regulation, not veridicality) and to the general lesson on schema-guided-symptom-perception and feedforward-vs-predictive-interoception that a confident percept displaced from the signal is the signature of a prior, not of a weak sense organ — the same shape Van der Does found in heartbeat counting.
What it does not settle
That an illusion can be induced shows interoceptive perception is penetrable by prior/expectation — it does not by itself adjudicate the feed-forward vs predictive debate, because a read-out account with a top-down comparator (Craig) can also absorb “distorted interoceptive predictions.” Nor does the illusion’s magnitude yet track any individual difference: the a priori prediction that low interoceptive sensibility would strengthen it failed for the MAIA (Iodice et al.), leaving open whether objective accuracy would. So the concept is established as a phenomenon and a method; its mechanism and its moderators are not.