Our understanding of body ownership largely relies on the so-called Rubber Hand Illusion (RHI). In this paradigm, synchronous stroking of the real and the rubber hands leads to an illusion of ownership of the rubber hand provided that it is physically, anatomically, and spatially plausible. Self-attribution of an artificial hand also occurs during visuomotor synchrony. In particular, participants experience ownership over a virtual or a rubber hand when the visual feedback of self-initiated movements follows the trajectory of the instantiated motor commands, such as in the Virtual Hand Illusion (VHI) or the moving Rubber Hand Illusion (mRHI). Evidence yields that both when the cues are triggered externally (RHI) and when they result from voluntary actions (VHI and mRHI), the experience of ownership is established through bottom-up integration and top-down prediction of proximodistal cues (visuotactile or visuomotor) within the peripersonal space. It seems, however, that depending on whether the sensory signals are externally (RHI) or self-generated (VHI and mRHI), the top-down expectation signals are qualitatively different. On the one hand, in the RHI the sensory correlations are modulated by top-down influences which constitute empirically induced priors related to the internal (generative) model of the body. On the other hand, in the VHI and mRHI body ownership is actively shaped by processes which allow for continuous comparison between the expected and the actual sensory consequences of the actions. Ample research demonstrates that the differential processing of the predicted and the reafferent information is addressed by the central nervous system via an internal (forward) model or corollary discharge. Indeed, results from the VHI and mRHI suggest that, in action-contexts, the mechanism underlying body ownership could be similar to the forward model. Crucially, forward models integrate across all self-generated sensory signals including not only proximodistal (i.e., visuotactile or visuomotor) but also purely distal sensory cues (i.e., visuoauditory). Thus, if body ownership results from a consistency of a forward model, it will be affected by the (in)congruency of purely distal cues provided that they inform about action-consequences and are relevant to a goal-oriented task. Specifically, they constitute a corrective error signal. Here, we explicitly addressed this question. To test our hypothesis, we devised an embodied virtual reality-based motor task where action outcomes were signaled by distinct auditory cues. By manipulating the cues with respect to their spatial, temporal and semantic congruency, we show that purely distal (visuoauditory) feedback which violates predictions about action outcomes compromises both performance and body ownership. These results demonstrate, for the first time, that body ownership is influenced by not only externally and self-generated cues which pertain to the body within the peripersonal space but also those arising outside of the body. Hence, during goal-oriented tasks body ownership may result from the consistency of forward models.
Read full abstract