Apps, M. A. J. & Tsakiris, M. (2014). The free-energy self: a predictive coding account of self-recognition. Neuroscience & Biobehavioral Reviews 41, 85–97. DOI 10.1016/j.neubiorev.2013.01.029. Online 15 February 2013; PMC3848896.
Key takeaways
From the reading, NOTE-991.
- The self as the best explanation. One’s own body is the object that, when “touched, moved, threatened, or acted upon in any way”, evokes events in the other senses. It is learned as a high-level multimodal prior, and “the notion that there is a ‘self’ is the most parsimonious and accurate explanation for sensory inputs”.
- One account of three illusions. Synchronous visuo-tactile or auditory-vocal stimulation produces surprise in one sense that a multimodal node explains away by updating which hand, face or voice is “me”. Earlier accounts (visual capture, updated reference frames) dealt with each illusion separately.
- Ownership is a competition between models. The illusion raises the probability that the rubber hand is “me” and lowers it for the real hand, which the authors read in reports that the real hand cools and its histamine rises during the illusion. The choice between the two models is compared to binocular rivalry.
- Context is a prior. A non-corporeal object, a hand of another skin tone or in an incongruent place, priming, and culture modulate ownership and self-recognition before any stimulation.
- No self-specific circuit. Self-stimuli engage unimodal areas that process all faces, bodies and voices, and multimodal areas (TPJ, anterior insula, IPS, IFG) that are not self-specific. The account needs no “self” network, only integration.
- Untested by its own account. “There is, as yet, little direct evidence that predictions and prediction errors related to self-stimuli are processed anywhere in the brain”, and the theory is “hard to falsify using standard neuroscientific methods”.
Standing in the record
Filed on 2026-10-19 at the owner’s request, to fill the record’s gap on the rubber hand illusion and bodily self-consciousness, and read in full the same day (NOTE-991) from the authors’ accepted manuscript in PubMed Central. It carries no instruction for machine-learning practice.
Active. It is the bridge the owner wanted: it takes the multisensory body-ownership findings that Metzinger’s “Self models” (LIT-450, NOTE-372, C6) cites and gives them a mechanism in the free-energy terms of Friston et al. (LIT-1181), which it cites. It names interoception among the self’s inputs but does not model it, so its tie to the record’s interoceptive-inference holdings (LIT-542, LIT-135) is thinner than the request assumed. It is a theoretical review: it reports no data and no model, and its case is that the account fits the findings, not that the findings require it. Its relations to the record are set out in the NOTE.
Filed alongside, in the 2026-10-19 batch on the rubber hand illusion and bodily self-consciousness: Botvinick & Cohen 1998, the rubber hand illusion (LIT-1277); Lenggenhager et al. 2007, the full-body illusion (LIT-1279); Blanke & Metzinger 2009, minimal phenomenal selfhood (LIT-1274); Seth 2013, interoceptive inference (LIT-1263).
Filed from this work on 2026-10-20: THEORY-769, on ownership as a graded belief about the common cause of multisensory input.