Source
Seth (2013), LIT-1263, Trends in Cognitive Sciences 17(11), read in full in NOTE-1000: “Interoception and the rubber hand” (p. 569), the cardio-visual rubber hand and Figure 3 (pp. 569–570), and Box 3 (p. 571). The studies the claim rests on were not read.
What was actually shown
Seth reports three results, all from cited studies:
- The cardio-visual hand. Suzuki et al. (2013), from Seth’s own group, built an augmented-reality hand whose colour pulsed toward red over 500 ms, either in or out of time with the participant’s heartbeat. Synchronous feedback “enhanced the experience of ownership of the virtual hand”, by questionnaire and by proprioceptive drift. Control questions showed no effect (pp. 569–570, Fig. 3).
- The cardio-visual body. Aspell et al. (2013), from Blanke’s group, applied the same feedback to a virtual body seen from behind. Synchrony “increased identification with the virtual body”, objectively and subjectively (p. 570).
- Sensitivity and susceptibility. “Participants with lower IS are more susceptible to the RHI” (Tsakiris et al. 2011), where interoceptive sensitivity is measured by heartbeat detection (p. 569).
Seth’s conclusion is that “statistical correlations between interoceptive (e.g., cardiac) and exteroceptive (e.g., visual) signals can lead to updating of predictive models of self-related signals … just as may happen for purely exteroceptive multisensory conflicts” (p. 570). The cardiac result could have come out otherwise: a pulse out of time with the heart could have done as well as one in time. As reported, it did not.
What this does not say
- It does not separate the heartbeat from any periodic signal. The asynchronous condition breaks the timing relation; it does not show that the signal must be interoceptive. A visual cue locked to a non-bodily rhythm at heart rate would separate the two. No such control is reported, and NOTE-1000 names this as the open question.
- It is second-hand. All three results are Seth’s report of studies not read here. One of the two cardio-visual studies is the author’s own, and the text gives no effect sizes.
- The cardiac cue is weak beside movement. In Suzuki et al., when finger movements were mapped to the virtual hand, visual–proprioceptive coherence “overshadowed the influence of cardiac (interoceptive) prediction errors” (p. 569). The claim is that interoceptive signals contribute, not that they dominate.
- It does not establish the precision reading. Seth reads low sensitivity as “possibly reflecting lower precision-weighting of interoceptive prediction errors”. The finding is equally a result about heartbeat-counting performance, and the article does not discuss the validity of heartbeat detection.
- It is not the emotion claim. Seth’s thesis that felt emotion is the content of interoceptive prediction (NOTE-1000, C1) is a separate hypothesis, which he says has no direct test. This THEORY carries only the ownership half.
- The autonomic results are not part of it. Cooling and histamine changes in the real hand are called “compatible with” active inference, not predicted by it. They are left out here.
Added on 2026-10-20
- Its evidence uses the design the suggestion papers question. Suzuki et al.’s cardio-visual hand compares synchronous with asynchronous feedback, the comparison Lush & Seth (2022, LIT-1267) say “offers no protection against demand characteristic effects”. Lush et al.’s supplement (LIT-1271) lists Suzuki et al. 2013 and Tsakiris et al. 2011 among the studies using such controls. Applying this to the cardiac result is the record’s inference. Seth co-authored the source, Lush et al. and the reply.
- A third variable is possible. Rubber hand reports track trait suggestibility (Roseboom & Lush 2022, LIT-1280), so the sensitivity–susceptibility relation could run through it. Nothing read tests this.
- Part of its evidence is drift. Suzuki et al.’s drift result cannot carry ownership on its own (Rohde et al. 2011, LIT-1281).
Connections
No relation is declared: no open question asks how body ownership arises.
- THEORY-769 is the exteroceptive form of the same inference. This THEORY adds the claim that interoceptive signals enter it.
- THEORY-765 is the visuo-tactile illusion whose susceptibility the sensitivity result predicts.
- The interoception cluster. It joins body ownership to the record’s holdings on interoception: LIT-388, LIT-628, LIT-542 and NOTE-434, and LIT-996 with THEORY-291. Before this filing they had no bearing on ownership.
- THEORY-045. Seth takes the prediction side of that THEORY’s open question about felt affect, but this THEORY does not rest on it.
- LIT-135, NOTE-169. This is the 2013 root of the interoceptive view in Seth’s later biological naturalism.
Note of 2026-10-29: what holds it back
Kept Proposed by the sixteenth status review. All three results are Seth’s second-hand report, and neither Suzuki et al. (2013) nor Tsakiris et al. (2011) is held. No condition locks the cue to a non-bodily rhythm at heart rate, expectations were not measured or matched, and the sensitivity–susceptibility relation has not been tested with trait suggestibility as a covariate. Nothing live rests on it: ARG-018 is Rejected.