False (bogus) physiological feedback paradigm

Valins (1966), “Cognitive effects of false heart-rate feedback.” Participants’ beliefs about their own cardiac responses are manipulated with fabricated feedback; the question is whether believed bodily change alone shifts emotional evaluation.

Two readings that recur in the wiki

• Friedman (2010) — the “apogee” of 1960s cognitive views: false-feedback effects implied actual ANS feedback was incidental to feeling (“primarily a result of cognitive factors and not physiological ones”). Friedman notes this is not irreconcilable with the somatic-marker-hypothesis: somatosensory cortex may simulate bodily feedback (an “as-if” loop), making the effect physiological after all.
• Seth (2013) — false feedback is predicted by interoceptive-inference: Gray et al. (2007) found false cardiac feedback increased right-AIC responses and the emotional salience attributed to previously unthreatening stimuli. In predictive-coding terms, injected (false) interoceptive evidence updates the generative model and thus the felt emotion — top-down and bottom-up are two directions of one loop, not “cognitive vs physiological.”

The same paradigm thus flips meaning depending on framework: evidence against the body’s role under 1960s cognitivism, evidence for an integrated predictive body-model under Seth. See feedforward-vs-predictive-interoception.

The paradigm turned on a physiological variable (Iodice et al. 2019)

Every use above manipulates an emotional judgment — how salient, how attractive, how threatening a stimulus feels. Iodice et al. (2019) point the same tool at a homeostatic/effort variable, which is what makes it an interoception study rather than an emotion study, and the source of the interoceptive-illusion construct.

Cyclists heard synthesised heartbeat sounds they believed were their own; when the sound was faster than their real heart rate, the same physical work felt harder (RPE rose). Three controls the classic false-feedback studies never all ran together:

• No entrainment: actual heart rate did not change with the feedback (BF = 0.014), so the raised effort is a misperception, not the heart obeying the sound.
• The sound is inert: congruent feedback produced the same ratings as no feedback, so hearing a heartbeat does not bias reports — only incongruent feedback does.
• Asymmetry: slower feedback did not lower perceived effort (Bayesian support for the null). This rules out stimulus-following, beat-counting and salience/attention accounts, all of which predict a symmetric effect, and points instead to a risk-averse prior that resists downward revision.

For this method page the lesson is that false feedback is not only a probe of appraisal but a probe of physiological estimation — a clean handle on interoceptive-inference that holds the body fixed and moves the percept. Note also that Iodice et al.’s a priori prediction — low interoceptive sensibility (MAIA) → stronger illusion, following Tsakiris’s RHI work — failed, so the paradigm’s individual-difference moderators remain open.