Investigating Techniques to Enhance the Sense of Embodiment in Immersive Virtual Reality

Participants complete multiple immersive virtual reality conditions across three study days. Conditions manipulate factors expected to influence embodiment. The order of conditions is randomized within each study day. Subjective embodiment is assessed after each condition using questionnaire, together with objective physiological measures.

Participants undergo multiple immersive virtual reality conditions designed to modulate embodiment. Conditions include synchronized visuo-tactile stimulation, heart-rate-synchronized visual biofeedback, varying degrees of active control of the avatar, and tasks with different levels of familiarity. Conditions are presented as short VR scenarios across three study days, with randomized order within each day.

Embodiment-related self-report assessed with the Virtual Embodiment Questionnaire (VEQ). Three subscale scores are computed: Ownership, Agency, and Perceived Change in Body Representation. Each subscale score is calculated as the mean of 4 items rated on a 7-point Likert scale (1 = strongly disagree to 7 = strongly agree). Higher scores indicate stronger perceived ownership of the virtual body (Ownership), greater sense of control over the virtual body/actions (Agency), and stronger perceived change in body representation during VR (Perceived Change in Body Representation).

Participants perform a VR reaction-time task in which brief light stimuli are presented at predefined locations: on the avatar's right upper limb, left upper limb, right lower limb, left lower limb, or in the virtual environment outside the avatar. Stimuli occur at randomized intervals (1-5 seconds). Participants respond as quickly as possible when a light appears. The outcome is the mean reaction time (milliseconds) computed separately for each stimulus location category. Location-specific averages are used to capture potential asymmetries relevant to the clinical groups (e.g., right/left differences in stroke due to hemiparesis, and upper vs lower limb differences in spinal cord injury due to predominant lower-limb impairment).

Electrodermal activity (EDA), a measure of sympathetic arousal, includes a tonic component (skin conductance level, SCL) and a phasic component (skin conductance responses, SCR). EDA is recorded during the VR sessions. For each VR condition, EDA metrics are computed relative to the baseline. The tonic component is summarized as the mean SCL during the condition. Phasic activity is summarized by SCR measures, including the number of SCR peaks and the mean SCR peak amplitude during the condition. Baseline-referenced values are computed as change from baseline for each metric.

Chest skin temperature will be analyzed as a slow physiological indicator that may vary with embodiment-related experiences in VR; based on prior findings, decreases in peripheral skin temperature have been reported with stronger embodiment. For each VR condition, skin temperature will be summarized as the mean value and condition values will be expressed relative to the baseline period (change from baseline).

Electrocardiogram (ECG) is recorded continuously during each study session; on Day 1 and Day 3, the ECG-derived heart rate is also used to provide real-time heart-rate visual biofeedback in VR. R-peaks are detected to derive the RR-interval time series, with artifact beat correction prior to analysis. Heart rate variability (HRV) is quantified for each VR condition and compared with baseline. Planned HRV metrics include standard time-domain indices (e.g., mean RR, SDNN, RMSSD) and frequency-domain indices (e.g., LF and HF power and LF/HF ratio). In the literature, HRV frequency components have been reported as sensitive to embodiment-related manipulations, with higher LF activity associated with lower embodiment and higher HF activity associated with moderate/high embodiment.

Electroencephalography (EEG) is recorded during the second study day using the eego mylab system with a 32-channel cap. EEG data will be pre-processed and analyzed in MATLAB using EEGLAB and custom scripts to remove artifacts and restrict the frequency content to the bands of interest. Power spectral density is computed and summarized within standard EEG bands (Delta, Theta, Alpha, and Beta). From band-power features, indices reflecting cognitive state are derived. For each VR condition, EEG-derived indices are compared with the preceding resting baseline segment to quantify condition-related deviations. In prior VR research, spectral features-particularly in the alpha and beta bands, including spatio-spectral patterns-have been investigated as potential correlates of embodiment-related processes.