Fig 1.
Experimental procedure and design.
a) Experimental design. Participants first observed the virtual body for 60 seconds (i.e., visual exposure) and then experienced the multisensory body illusion. After two minutes of illusion induction, a virtual knife appeared and stabbed the fake virtual body. The skin conductance level (SCL) was collected. After each visual exposure (i.e., T0) and body illusion (i.e., T1), participants performed a Body Localization Task and completed the embodiment questionnaire. b) Body Illusion type and Location conditions. Each illusion (vRHI or 1pp-FBI) was presented in two possible location conditions: aligned (i.e., the avatar was aligned with the real body) and misaligned (the avatar’s left hand or legs were shifted towards the body’s midline). c) Body illusion procedure. Frames extracted from the VR environment depicting the visual-tactile stimulation for the vRHI (upper panel) and 1ppFBI (lower panel). During the illusion, participants received tactile stimulation on their left hand or leg while observing the avatar being synchronously touched in the corresponding area. SCL = Skin Conductance Level; vRHI = virtual Rubber Hand Illusion; 1pp-FBI = first-person perspective Full Body Illusion.
Fig 2.
The figure shows the results of the Bayesian regression on the averaged answers to Embodiment (a) and Disembodiment (b) statements. Error bars indicate 95% Credible Interval limits.
Table 1.
Results of the Bayesian regression ran on the embodiment responses.
Table 2.
Results of the Bayesian regression ran on the disembodiment responses.
Fig 3.
Results of Bayesian regression on the proprioceptive drift.
The figure shows the results of the Bayesian regression on the Proprioceptive Drift (Real–Estimate position), depending on Illusion, Time, and Location; Error bars indicate 95% Credible Interval limits.
Table 3.
Results of the Bayesian regression ran on the proprioceptive drift.
Fig 4.
Results of additional Bayesian regression for the vRHI.
a) Results of the Bayesian regression on Question 6 depending on the Order of Sessions, Time, and Location; b) Results of the Bayesian regression on Proprioceptive Drift depending on the Order of Sessions, Time, and Location; Error bars indicate 95% Credible Interval limits. Given that the cumulative distribution provided the most accurate fit for the data, we recode the variable. We transformed the original scale from 1 to 7, thus creating an integer variable without zero.