Objective: As virtual avatars become increasingly embedded in social, educational, and commercial platforms, the virtual avatars appearance realism (VAAR) has emerged as a critical factor influencing user perceptual interaction experience. This study explores the impact of VAAR on users’ perceptual interaction experiences. Method: We retrieved all relevant publications from the past decade (April 2015 to April 2025) across four major databases: Web of Science, Elsevier Science Direct, Springer Link, and Google Scholar. The analysis focused on three user perceptual interaction experience-related indicators: attractiveness, trustworthiness, and eeriness. A network meta-analysis (NMA) was conducted using Stata software, and the relative ranking of VAAR was determined based on the surface under the cumulative ranking curve. Results: A total of 13 studies with 2,343 participants were included in the analysis. According to the results of the NMA, high VAAR were perceived as more attractive and more trustworthy than those with medium or low VAAR. In contrast, medium VAAR elicited the strongest feelings of eeriness. Conclusion: This study highlights the subjective impact of VAAR on user interaction experience. These findings provide meaningful guidance for the future development and design of virtual avatars. Systematic review registration: The network meta-analysis (NMA) was preregistered on the Open Science Framework (OSF) and is publicly available at https://doi.org/10.17605/OSF.IO/5TRG6 .

Obstacle crossing is a fundamental, yet complex, motor task commonly observed in daily life and occupational settings. Nevertheless, accurately replicating real-world crossing strategies in a virtual reality environment (VRE) remains challenging, potentially limiting the broader application of virtual reality (VR) in these domains. In this context, this study compared individuals’ gait and lower-limb kinematics during their performance of obstacle crossing in a real environment (RE) and a VRE to determine whether the latter alters their crossing behavior. Participants completed obstacle-crossing tasks over three obstacles (low: 100 mm, middle: 200 mm, and high: 300 mm) in both environments. Our results indicated that the participants’ overall crossing speed was reduced in the VRE (low obstacle: −9.0 cm/s; middle obstacle: −6.0 cm/s; high obstacle: −2.0 cm/s). Specifically, we observed that in the VRE, the leading leg exhibited greater lower-limb flexion (low obstacle: +13.8°; middle obstacle: +11.1°; high obstacle: +8.2°), which resulted in a higher toe clearance (low obstacle: +76.4 mm; middle obstacle: +40.0 mm; high obstacle: +23.5 mm). In addition, the horizontal distance from the peak vertical position was extended (low obstacle: +93.1 mm; middle obstacle: +79.0 mm; high obstacle: +73.8 mm). For the trailing leg, a trend similar to that of the crossing leg was observed under the VRE low-obstacle condition; however, with increasing obstacle height, the VRE showed smaller lower-limb flexion (high obstacle: −12.7°) and toe clearance than the RE (middle obstacle: −68.2 mm; high obstacle: −115.2 mm). These findings provide insight into gait adaptation in virtual contexts and support the design of more effective VR-based interventions for rehabilitation and fall prevention.