Contribution
The report gives a simple, repeatable procedure that makes people feel a rubber hand as their own, and shows that the procedure works only when what is seen and what is felt are synchronous. It adds a behavioural measure, the drift of the felt position of the real hand toward the rubber one, that tracks how long the illusion was reported. It turns the question of what makes a body part feel like mine into an experimental one.
Key insight
Seeing a touch and feeling one at the same moment is enough, on its own, to move where touch is felt, where one’s hand is felt to be, and which hand is felt to be one’s own. Ownership of a body part is not fixed by the body: it follows the correlation among the senses, and when the senses conflict the system resolves the conflict by moving the felt hand.
Assumptions
- The questionnaire measures the experience. The illusion is established by self-report: an open-ended description and agreement on a seven-step visual-analogue scale with nine statements “presented in a random order” (Fig. 1). Three describe the predicted effects and six serve as controls.
- Intermanual reaching measures position sense. “With eyes closed, the right index finger was drawn along a straight edge below the table until it was judged to be in alignment with the index finger of the left hand”. Displacement is the difference between the means of three reaches before and three after exposure.
- Reported prevalence is a usable predictor. Figure 2’s x-axis is “the percentage of the 30-min viewing period during which the ownership illusion was experienced”. How subjects reported this over 30 minutes is not stated.
- A brief communication. Ten subjects are named for the first experiment. The size of the second experiment’s group and of the asynchronous control group is not given in the text, and neither are the statistical tests behind the P-values. Figure 2 plots ten points (my count from the rendered page).
Key results
All on p. 756.
- Setup. Each subject’s left arm rested on a small table, hidden by a standing screen; “a life-sized rubber model of a left hand and arm was placed on the table directly in front of the subject”. Two small paintbrushes stroked the rubber hand and the hidden hand, “synchronising the timing of the brushing as closely as possible”.
- The illusion (Fig. 1). After ten minutes, subjects “seemed to feel the touch not of the hidden brush but that of the viewed brush, as if the rubber hand had sensed the touch”. The three underlined, predicted statements drew affirmative responses: “It seemed as if I were feeling the touch of the paintbrush in the location where I saw the rubber hand touched”, “It seemed as though the touch I felt was caused by the paintbrush touching the rubber hand” and “I felt as if the rubber hand were my hand” (“P<0.002 for underlined questions, P<0.018 after correcting for multiple comparisons”). On the rendered figure the first has every response at “agree strongly”, and the other two have means near ++ with no response below +. The six control statements (the real hand drifting, a second left hand, touch felt between the hands, the real hand turning “rubbery”, the rubber hand drifting or coming to resemble one’s own) do not reach significance. Five have means on the disagree side; the last, that the rubber hand “began to resemble my own (real) hand”, sits slightly above neutral. Their ranges are wide.
- Ownership (final paragraph). Subjects who referred touch to the rubber hand “also consistently reported, in both sections of the questionnaire, experiencing the rubber hand as belonging to themselves”. “Eight of ten subjects spontaneously employed terms of ownership”, for example “I found myself looking at the dummy hand thinking it was actually my own.”
- Proprioceptive drift (Fig. 2). After a prolonged (30-minute) exposure, “Subjects’ reaches after experiencing the illusion were displaced rightward toward the rubber hand, the magnitude of this displacement varying significantly in proportion to the reported duration of the illusion”. The fitted line is y = 0.09x − 1.4 (cm against % of 30 min), “r2 0.47, p<0.03”.
- Asynchronous control. “A small asynchrony” in the brushing gave “low prevalence of the illusion (mean 7% of the exposure period compared to 42% for the synchronously brushed group; P<0.01)” and no drift toward the rubber hand: “mean displacement 13 mm away from the rubber hand, compared with 23 mm towards it in the experimental group; P<.04”.
- The model. “We hypothesized that this illusion involves a constraint-satisfaction process operating between vision, touch and proprioception — a process structured by the correlations normally holding among these modalities.” A connectionist model “suggested that the illusion’s spurious reconciliation of visual and tactile inputs relies upon a distortion of position sense”; the drift experiment tested that prediction. The model is not described or cited. Its layer of bimodal units is compared with premotor cells in monkeys that respond to touch on a body region and to a seen object approaching it (Graziano and Gross 1994).
- Conclusion. “While the rubber hand illusion does not tell us precisely what ingredient might make only certain forms of intermodal correlation relevant to the self, it does show that intermodal matching can be sufficient for self-attribution.”
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Synchronous visuo-tactile stimulation makes touch be felt at a seen rubber hand | moderate | ten subjects, questionnaire, Fig. 1; strong agreement on the first item |
| C2 | The same stimulation makes the rubber hand be felt as one’s own | moderate | Fig. 1 third item; 8 of 10 used ownership terms unprompted |
| C3 | The illusion displaces felt position of the real hand toward the rubber hand, in proportion to its reported duration | moderate | 30-minute exposure, regression r² 0.47, p<0.03, about ten points, Fig. 2 |
| C4 | Asynchrony between seen and felt touch abolishes the illusion and the drift | moderate | control group of unstated size; 7% against 42%, P<0.01; −13 mm against +23 mm, P<.04 |
| C5 | The illusion arises from constraint satisfaction among vision, touch and proprioception, structured by their usual correlations | weak | a connectionist model mentioned and not shown; C3 is its one tested prediction |
| C6 | Intermodal matching can be sufficient for self-attribution of a body part | moderate as an inference from C2 and C4 | the authors’ conclusion; it concerns a hand, not the whole body or the self |
Concepts
- Intersensory bias — the class of effects in which one sense shifts what another reports; the authors place the illusion in it, citing prism adaptation (Hay, Pick and Ikeda 1965), Tastevin (1937) and others.
- Referred touch — touch felt at a location other than the stimulated one, here at the seen rubber hand.
- Self-attribution / bodily self-identification — experiencing a body part as belonging to oneself. The paper does not use “ownership” as a technical term; it says “terms of ownership” and, in Fig. 2’s caption, “ownership illusion”.
- Intermodal perceptual correlation — the proposal, credited to Lewis and Brooks-Gunn (1979) and Bahrick and Watson (1985), that the body is told apart from other objects by its “participation in specific forms of intermodal perceptual correlation”.
Connections
The authors place the illusion beside Ramachandran’s mirror-box work, in which amputees who watched their intact arm touched in a mirror felt the touch in the phantom, and beside bimodal premotor neurons in monkeys (Graziano and Gross). The developmental literature on infants’ detection of contingency between their own movement and what they see (Bahrick and Watson) supplies the idea that intermodal correlation marks the body as self. The paper builds on these by citation only; none was read here.
Bearing on the record
- LIT-450 and NOTE-372 (Metzinger). NOTE-372’s C6, that body ownership “can be manipulated by multisensory conflict”, rests on Lenggenhager et al. (2007) and Blanke and Metzinger (2009), neither held. This is the primary result that line of work extends from a hand to the whole body, and it supports the manipulation half of C6 directly. It says nothing for the other half, that global body ownership is “the most simple form of self-awareness”: Botvinick and Cohen claim only sufficiency of intermodal matching for attributing a hand. Metzinger’s transparency fits the report’s phenomenology, since subjects felt the touch in the rubber hand rather than inferring it, but that reading is mine; the paper does not discuss it. NOTE-372’s C7, the congenital phantom, gets nothing from it, though the authors cite Ramachandran’s phantom-limb mirror work as “closely related”.
- THEORY-064 (Proposed). Its promotion criterion already says that “more body-ownership illusions … show that the self-model can be manipulated but not that it is all there is”. This report is the first such illusion in the record read at first hand. It strengthens the premise THEORY-064 grants and leaves the contested half untouched. The précis (LIT-568, NOTE-460) treats “mineness”, the sense of ownership, as what binds contents into the self-model; the rubber hand shows that mineness can be extended to an object by correlated input.
- THEORY-458 (Proposed), LIT-453 (NOTE-377) and LIT-1009 (NOTE-827). Gallagher and Haggard separate the sense of ownership from the sense of agency, and Haggard notes that involuntary movement brings ownership without agency. The rubber hand is a further dissociation in the same direction, mine to point out: ownership of a hand that does not move at all, induced by passive stimulation alone. It does not test THEORY-458’s claim that ownership depends on sensory feedback while agency depends on a forward model, but it is consistent with the ownership half. Nothing here bears on THEORY-322, which is about agency alone.
- LIT-1181 (NOTE-960) and the predictive-processing holdings (LIT-542, LIT-135). Friston et al.’s C4 has multisensory weighting for reaching follow from optimized precisions, in simulation. Botvinick and Cohen’s “constraint-satisfaction process” reconciling vision, touch and proprioception is a pre-Bayesian statement of the same kind of account, and the drift is its measurable trace. The paper uses no probabilistic language, and the mapping is the record’s, not the authors’. Seth’s and Barrett’s interoceptive models (LIT-135, LIT-542) concern the body from inside; this paper concerns only exteroceptive senses and proprioception.
- LIT-190 (Winkielman et al., Proposed) proposes that cross-modal coherence feeds a unified experience; the rubber hand is a case in which cross-modal coherence changes what is experienced as one’s own body. The pairing is mine.
- QUESTION-048. The report pairs a self-report measure with a behavioural one (drift) that covaries with it, which is the kind of convergence the question asks about. But the drift was measured after a longer exposure than the questionnaire, so the two were not taken on the same trials.
- No instruction for machine-learning practice.
Limitations
- Ten subjects, one laboratory, one page. No effect sizes beyond the regression, no named tests, and the control group’s size is not given.
- The questionnaire’s predicted items were known to the experimenters, and subjects saw the rubber hand brushed in a design that makes the intended answer guessable. The six control items, five with means on the disagree side, guard against blanket agreement but not against suggestion specific to the touch and ownership items. The authors do not discuss demand characteristics.
- The drift experiment changes the exposure (30 minutes against ten) and measures prevalence by a procedure not described, so its relation to the questionnaire result is indirect.
- The connectionist model that motivated the drift prediction is neither shown nor cited, so C5 cannot be assessed from the paper.
- In Figure 1’s caption the corrected threshold (P<0.018) is larger than the uncorrected one (P<0.002); this reads as two different statistics given without explanation.
- The conclusion concerns attribution of a hand. It does not show that whole-body ownership, self-location or the self works the same way.
Open questions
- Is the drift a measure of ownership or only of visuo-proprioceptive recalibration? The paper ties it to reported duration but does not separate the two.
- What makes “only certain forms of intermodal correlation relevant to the self”? The authors leave this open in their own words. Does the rubber hand need to look like a hand, be placed where a hand could be, or only be touched in synchrony?
- Is the self-report of ownership independent of the referred touch, or a description of it? The questionnaire items were affirmed together.