Embodiment of Virtual Body and Extremities With Movement Control in Reaching Tasks Using Virtual Reality

Abstract

The study explores embodiment in virtual reality (VR), investigating its effects on task performance across various body representations. Previous research highlighted the impact of embodiment on motor performance in VR, emphasizing the importance of body representation. While prior studies mainly focused on full-body avatars, this study aims to understand the perceived embodiment in hand and foot-reaching tasks using different body representations. Thirty-two participants engaged in reaching tasks within VR environments, employing a 4 × 3 within-subjects design. Results indicated significant differences in embodiment across body representation conditions, with notable effects on task performance. These findings underscore the importance of tailoring virtual environments to enhance embodiment and consider task-specific demands to optimize performance outcomes. Personalizing VR experiences based on body representations could effectively enhance user engagement and satisfaction, offering valuable insights for future VR applications in various domains.

Introduction

Embodiment in virtual reality (VR) refers to taking the perspective of another person or avatar, which is a virtual character driven by human behavior (Yu et al., 2021) and is often enabled through a virtual representation of the body (Kilteni et al., 2013). Having the sense of embodiment has been shown to enhance motor performance during locomotion training (Berg et al., 2023). While previous locomotion studies have employed full-body avatars (Dewez et al., 2020; Saint-Aubert et al., 2023), research has shown that full-body avatar representation is not a prerequisite for inducing the sense of embodiment in virtual environment (Škola & Liarokapis, 2018). While it is commonplace to show part of the virtual body in VR applications by tracking user’s head and hands, there is a push for tracking the rest of the body, for example, the torso and the legs (Hirao et al., 2020; Parger et al., 2018; Steed et al., 2016). Previous research has shown that better range of motion of the virtual body helped improve user experience of embodiment (Cha et al., 2021), and representations of virtual body and extremities have been used to increase embodiment (Parger et al., 2018), thus to influence task performance (Porssut et al., 2022; Škola et al., 2019). In general, there are many methods for self-avatar representation in VR based on the coverage of the body. This study aims to investigate the perceived level of embodiment over different virtual representations of the body in hand reaching and foot reaching tasks in VR. Furthermore, the effects of perceived embodiment on task performance were also investigated.

Methods

Informed consent was obtained from 32 participants for this study (n = 32, age range: 19–54 years, mean age: 28.31 ± 7.68, 18 female, 14 male). The virtual scenes were displayed using a head-mounted display (HMD) (VIVE Pro, HTC Corporation, Taiwan) and the VIVE trackers affixed to participants’ wrists and ankles for motion tracking. The reaching task in the virtual environment was based on the whack-a-mole game, aiming to reach for moles (score +1 if hit mole) and avoid bombs (score −1 if hit bomb) using either hand reaching on a vertical board, foot reaching on a horizontal board, or both.

A 4 × 3 (body representation condition x reaching task type) within-subjects study design was employed. The four virtual representations of body included: full-body avatar, half-body avatar, representation with only hammers, and absence of any virtual representation. As for the reaching tasks, participants engaged in three different task types: hand reaching, foot reaching, and simultaneous hand and foot reaching. Each trial in the study involved participants engaging in one of three reaching task types within one of four body representation conditions (12 trials = 3 × 4). All participants completed a total of 12 trials with a randomized order. In each trial, the participant performed the whack-a-mole task for 1 min. The outcome measures included sense of embodiment, and task performance (score) for each trial.

Statistical analyses such as ANOVA, were conducted to compare the sense of embodiment and task performance for different conditions. The statistical significance level was set at 0.05.

Results

There was a statistically significant difference in the sense of embodiment among the four body representation conditions (F [3, 28] = 5.06, p = .002). Furthermore, significant main effects of embodiment (F [1, 24] = 5.73, p = .017) and task type (F [2, 24] = 5.07, p < .001) on task performance were observed.

Discussion

1. Significant differences observed in embodiment across different virtual body representations. By understanding how different body representations impact embodiment, it is possible to tailor the setup of virtual environments to enhance embodiment.

2. The significant main effects of embodiment on task performance, coupled with the observed main effects of task type, highlighted the multifaceted nature of movement control process within virtual environments. This suggested that the level of embodiment experienced by users and the nature of the task being performed both played crucial roles in determining performance outcomes. Consequently, strategies aimed at enhancing embodiment and considering task-specific demands may synergistically improve task performance outcomes within virtual environments.

3. Personalizing virtual experiences based on body representation conditions could be used to enhance embodiment and task performance outcomes, aiming for user engagement and satisfaction.