Abstract

After continuously applying spatially and temporally congruent visuo-tactile stimulation to external, body-like objects, healthy humans feel the fake part to be the origin of the touch sensations and feel it to be their own. Incongruent stimulation nullifies both illusory changes. This Rubber Hand Illusion (RHI) is the most used paradigm to experimentally investigate healthy humans’ bodily self-consciousness and the extension of body ownership to foreign objects [1]. Standard RHI setups are limited by physical constraints and suffer the disadvantage of imprecise temporo-spatial stimulus control. Further, little is known about the neural and electrophysiological mechanisms associated with the crossmodal conflict and the induced changes in body ownership during the RHI. Here we recorded 64-channel electroencephalography (EEG) to investigate the electrophysiological changes while exposed to a carefully controlled, bilateral RHI induced by haptics and virtual reality (VR) technology. Methods Participants (n=8) sat at a table with a chin rest. Arms were placed (palms up) underneath the table to rest on the legs. A virtual scene depicting the real-world experimental scene was displayed on an immersive Head Mounted Display (HMD). We employed a 2x2 factorial design (Object, Stroking). Participants saw either two arms or two non-body cylinders (Object) projecting from their shoulders onto a virtual table, seeing the virtual arms or cylinder in a position ~20cm above their real arms. Visuo-tactile stimulation was provided by four vibration motors affixed to the palms of the (unseen) left and right hands and by animation of corresponding virtual motors on the virtual hands. Two visuo-tactile modes were defined (Stroking): 1) synchronous: virtual visual motors were shown to vibrate in temporal and spatial synchrony with the tactile vibration motors and 2) asynchronous: virtual visual motors were shown to vibrate with a temporal delay (100±50ms) and a randomly selected spatial direction with respect to the tactile vibration motor pattern. Two additional baseline conditions were recorded (no visual animation of the virtual motors and no tactile stimulation; no visual animation of the virtual motors with tactile stimulation). Subjective experience of the illusion was gauged by a questionnaire. 64-channel EEG was sampled at 2048Hz. Bipolar electrooculograms were recorded for later artifact removal. EEG analysis consisted of breaking 30s of spontaneous EEG data per condition into epochs of 2s. Epochs contaminated with eye blink artifacts or transient changes in electrode-to-scalp conductance were removed. Spectral power changes were computed in three frequency bands: alpha (8-13 Hz), beta (14-30 Hz) and gamma (30-100 Hz). Statistical analysis was performed at the scalp level and included a suprathreshold cluster permutation test to control for Type I errors. The generators of the significant scalp maps were localized with an inverse solution (sLORETA) [2]. Results Repeated ANOVA analysis of the questionnaire indicated a significant interaction between synchrony and question (F=1.65, p=0.03), and a main synchrony effect (F=40.29, p<0.001). Questions of illusory touch (i.e. subjects felt the touch as if it were from where they saw it on the virtual hand) and ownership (i.e. subjects felt the virtual hand was their own hand) were significantly modulated by synchrony mode for the bodily visual objects (arms). Frequency analysis showed synchronous visuo-tactile stimulation leads to a suppression in mu-band power over bilateral sensorimotor cortex, but only for the body condition. This mu-band power suppression was selective insofar as it was not found in other cortical areas. Beta and gamma analysis did not reveal significant modulations. Conclusions Our data show strong bilateral changes in body ownership can be induced in a fully automated fashion using haptics and VR. Our data revealed suppression of 8-13 Hz oscillations at selected central scalp locations. This was further confirmed by frequency analysis of all recorded scalp electrodes revealing a large and significant cluster of adjacent scalp electrodes at bilateral fronto-parietal scalp regions when contrasting synchronous and asynchronous stroking of the arm, but not the control object. This pattern was localized to the right and left lateral premotor cortex extending posteriorly to the postcentral gyrus as well as smaller regions in medial prefrontal cortex and the right occipito-temporal cortex and is partly compatible with previous studies on limb ownership manipulation analyzing the BOLD signal without haptics and VR control [3]. We emphasize the potential in combining high-resolution EEG, immersive HMDs, and haptic technologies for researching bodily consciousness.

Details

Actions