Abstract
In the rubber hand illusion (RHI), participants report ownership of a fake hand which is stroked in synchrony with their own concealed hand, with lower ownership ratings for asynchronous stroking. This effect has been historically attributed to multisensory integration. However, such a pattern of results is consistent with experimental demand characteristics, and an alternative theory proposes that the RHI is an effect of phenomenological control (the trait ability to change experience to satisfy goals). To date, support for a phenomenological control theory of the RHI has been limited to correlational evidence. Here we conduct a pre-registered causal test of the relative contribution of these two proposed mechanisms in RHI effects. 25 high phenomenological control participants (top 15%) were tested in an RHI procedure and the classic effects of greater response for synchrony than asynchrony were replicated. The procedure was then repeated following imaginative suggestions for a reversal of the classic effect (i.e. greater synchrony than asynchrony effects). Ratings of ownership and proprioceptive drift were modulated by suggestion. Effects for ownership and drift were reversed. Exploratory analysis showed the same pattern of results for referred touch ratings. Participant beliefs arising from demand characteristics dominated over multisensory mechanisms in RHI ownership of a fake hand and proprioceptive drift. Multisensory synchrony was not necessary to produce the classic effects, falsifying the theory that the illusion requires multisensory integration. Claims of an illusion beyond that attributable to demand characteristics and phenomenological control require experimental evidence from controlled experiments.