Virtual Humans in Virtual Reality Mental Health Research: Systematic Review
JMIR XR and Spatial Computing · 2025-10-08 · 2 · e75087-e75087
DOI: 10.2196/75087 ↗At a glance
Übersichtsarbeit in JMIR XR and Spatial Computing, published online on 2025-10-08. Written by Shu Wei and 2 co-authors. Institution: Department of Psychiatry, University of Oxford, Warneford Hospital, Warneford Ln, HeadingtonOxford, OX3 7JX, United Kingdom. Cited 5 times so far (Crossref); 143 references. The full text is freely available in PubMed Central.
Abstract (original language)
Background
Virtual reality (VR) is showing increasing promise for assessing, understanding, and treating mental health difficulties. Virtual humans (VHs) represent a key aspect within many VR mental health applications. While VHs can play diverse roles and display varied characteristics, their design and influence have rarely been the primary focus of mental health research.
Objective
We aimed to carry out a systematic review of how VHs in immersive VR have been used in applications for mental health, focusing on their roles and interaction types, and the human characteristics being tested.
Methods
Following the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines, we searched PubMed, MEDLINE, PsycINFO, Scopus, and Web of Science, using defined keyword combinations involving VR, VHs, and mental health. Eligible studies included peer-reviewed research using immersive VR with VHs in a mental health context, without restrictions on study design or population. We excluded nonimmersive VR, nonmental health applications, and papers without empirical data. Data were synthesized narratively, and a taxonomy to categorize VHs that we developed was used.
Results
A total of 79 studies met all eligibility criteria. VHs were most frequently applied in studies on social anxiety (n=18), eating disorders (n=18), and psychosis (n=15). They were primarily used as active social interaction partners (n=40), as part of virtual crowds (n=16), and as virtual bodies for participants (n=23). Explicit interactions dominated active partner studies, while implicit and passive or no interactions were prevalent in crowd and body studies. Over half of the studies (n=44) varied the VH characteristics, with body size and gender being the most common variables, and personality was explored in fewer studies (n=5). Only a limited number of studies provided detailed descriptions of VH appearance and behavior, with some including still images and videos.
Conclusions
VHs are versatile tools to be used within VR mental health applications, but their design features are inconsistently reported and insufficiently examined in relation to intervention outcomes. Evidence is limited by heterogeneity in study aims, designs, and populations, and by incomplete reporting of VH characteristics, which constrains replication and cross-study comparison. Standardized reporting and systematic investigations of VH design are needed to optimize their roles in evidence-based mental health applications.
Subjects
Author keywords
Authors and institutions
- Department of Psychiatry, University of Oxford, Warneford Hospital, Warneford Ln, HeadingtonOxford, OX3 7JX, United Kingdom.
- Department of Paediatrics, Yale School of Medicine, New Haven, CT, United States.
- Department of Experimental Psychology, University of Oxford, Oxford, United Kingdom.
- Oxford Health NHS Foundation Trust, Oxford, United Kingdom.
ORCID profiles: Shu Wei, Daniel Freeman, Aitor Rovira
Cite
APA
Wei, S., Freeman, D., & Rovira, A. (2025). Virtual Humans in Virtual Reality Mental Health Research: Systematic Review. JMIR XR and Spatial Computing, 2, e75087-e75087. https://doi.org/10.2196/75087
Vancouver
Wei S, Freeman D, Rovira A. Virtual Humans in Virtual Reality Mental Health Research: Systematic Review. JMIR XR and Spatial Computing. 2025;2:e75087-e75087. doi:10.2196/75087
BibTeX
@article{Wei2025,
author = {Wei, Shu and Freeman, Daniel and Rovira, Aitor},
title = {Virtual Humans in Virtual Reality Mental Health Research: Systematic Review},
journal = {JMIR XR and Spatial Computing},
year = {2025},
volume = {2},
pages = {e75087--e75087},
doi = {10.2196/75087}
}
EUVIMED