Association Between Embodiment of a Muscular Avatar and Muscle Performance in Virtual Reality for Chronic Low Back Pain: Exploratory Pre-Post Study

Abstract Background Chronic low back pain (CLBP) is associated with impaired muscle performance, including reduced trunk strength, as well as disturbances in body perception. Virtual reality (VR) enables modulation of body representation, and the Proteus effect suggests that embodying a stronger avatar may influence perceived bodily capability and behavior. Although preliminary CLBP studies have reported short-term improvements in pain and beliefs, it remains unclear whether avatar-based embodiment can acutely modify muscle strength or relate to subjective embodiment. Objective This study aimed to examine immediate changes in muscle strength and pain after a VR-based squat task with a progressively muscular avatar in individuals with CLBP and explore associations between strength changes and subjective embodiment. Methods In this exploratory pre-post proof-of-concept study, 16 individuals with CLBP and 16 age- and sex-matched healthy controls performed a VR-based squat task with a progressively muscular avatar. Muscle strength (handgrip, trunk extension, and knee extension) was assessed before and after VR. Pain and fear were assessed in CLBP only. Subjective embodiment was assessed using the Virtual Embodiment Questionnaire (VEQ). Group × time effects were analyzed using repeated-measures ANOVA, and associations were analyzed using Spearman correlations with false discovery rate correction. All participants completed the VR-based task and assessments, and no missing data were observed in variables included in the analyses. Results Significant time effects were observed for handgrip strength ( P =.04), trunk extension strength ( P <.001), and knee extension strength ( P <.001), whereas no significant group × time interactions were observed for handgrip, trunk extension, or knee extension strength ( P =.75, P =.19, and P =.28, respectively). Trunk extension strength significantly increased in healthy controls (mean difference 0.04 kgf/kg, 95% CI 0.03 to 0.05; Bonferroni-adjusted P <.001) but not in individuals with CLBP (mean difference 0.02 kgf/kg, 95% CI 0.001-0.05; Bonferroni-adjusted P =.08). Knee extension strength increased in both the CLBP group (mean difference 0.06 kgf/kg, 95% CI 0.02-0.11; Bonferroni-adjusted P =.02) and healthy controls (mean difference 0.09 kgf/kg, 95% CI 0.06-0.12; Bonferroni-adjusted P <.001). In the CLBP group, pain intensity decreased during VR compared with pre-VR levels ( P =.02), whereas post-VR pain did not differ from pre-VR levels. Changes in trunk extension strength were positively correlated with VEQ change in the CLBP group (ρ=0.76, 95% CI 0.41-0.94; false discovery rate–adjusted P =.006). Conclusions Trunk and knee extension strength showed time-related increases without significant group × time interactions, and trunk extension strength change was associated with VEQ change in the CLBP group. These findings suggest that avatar-based modulation of body perception may relate to individual variability in motor output. However, without a neutral-avatar control condition, the observed changes cannot be attributed specifically to the progressive muscular transformation of the avatar. Further research is needed to clarify the role of avatar-based embodiment in exercise-based rehabilitation.