Late adolescence and young adulthood (ages 18-25 years) represent a high-risk window for the onset of mental health disorders, with a large proportion of lifetime conditions first emerging by the mid-20s [1-4]. Depression and anxiety in this developmental period are shaped by interacting social, psychological, and biological processes and remain a substantial public health concern [5-7].

However, conventional interventions such as counseling and health education often struggle to sustain participation, and this limitation extends to many digital mental health tools. Even when mental health apps are widely available, long-term engagement remains low, suggesting that motivational design is a persistent implementation challenge [8-10].

The convergence of immersive technologies and psychologically informed intervention design offers a promising alternative. Virtual reality (VR) can deliver embodied and engaging exergaming experiences [11-16], music can support emotional regulation and social bonding [17-19], and socially shared or synchronized activity may reduce loneliness and enhance connectedness [20-28]. Although VR-based mental health interventions have shown encouraging results for distress reduction [22,29-32], these components have often been studied in isolation. To address this gap, this trial is informed by self-determination theory (SDT), which proposes that autonomy, competence, and relatedness are basic psychological needs linked to intrinsic motivation and well-being [33,34]. Importantly, this intervention is designed as a bundled SDT-informed intervention rather than an experimental test of each need in isolation. The study therefore evaluates the effectiveness of a multicomponent social VR musical exergame and explores whether changes in need satisfaction are associated with symptom change; these exploratory analyses are hypothesis generating and do not permit causal conclusions about individual SDT mechanisms.

This study will use a 3-arm, parallel-group, single-center randomized controlled trial design. Assessments will be conducted at baseline (T0), immediately after the intervention (T1, week 6), and at a 1-month follow-up (T2, week 10) to evaluate short-term maintenance of effects. The primary planned comparison is the experimental group vs the active control group, which isolates the added value of the social layer over a matched solo VR intervention. Comparisons involving the waitlist control are secondary. This protocol was prepared in accordance with the SPIRIT (Standard Protocol Items: Recommendations for Interventional Trials) guidelines.

Participants will be recruited from local universities, vocational colleges, and community health centers in Changsha, China.

Recruitment strategies will include posters and flyers at universities, vocational colleges, and community centers; online announcements; and referrals from collaborating educational and community services.

Following screening and informed consent, eligible participants will complete baseline assessment (T0), undergo randomization, receive the 6-week intervention or waitlist condition, complete the postintervention assessment at week 6 (T1), and complete the follow-up assessment at week 10 (T2; Table 1).

A total of 110 young adults aged 18 to 25 years will be recruited. The eligibility range was narrowed to 18 to 25 years to improve developmental homogeneity and align the sample with late adolescence and emerging adulthood, a period marked by elevated risk for depression and anxiety [1-4]. The participant eligibility criteria are shown in Textbox 1. Participants will be recruited through advertisements at local universities, vocational colleges, and community health centers.

An a priori power analysis was conducted using G*Power. On the basis of meta-analyses of related VR-based mental health interventions [30,31], a medium effect size (f=0.25) was assumed. To achieve 80% power at an alpha level of .05 for the primary group-by-time comparison, a required sample of 90 participants was indicated. To account for an anticipated attrition rate of 20% to 30%, consistent with rates reported in VR intervention literature, the recruitment target is set at 110 participants. A sensitivity analysis indicates that with 30% attrition (leaving approximately 77 participants), statistical power for the primary comparison would be reduced to approximately 70%, which is acknowledged as a limitation. To minimize attrition, the study will implement flexible scheduling, session reminders, brief mid-study check-ins, and graduated VR exposure in the first session to identify and manage cybersickness early. If observed attrition exceeds 30%, this will be reported transparently and considered in the interpretation of results. This sample size determination was made for the primary clinical outcomes and was not specifically powered for mediation or moderation analyses; those analyses will therefore be treated as exploratory.

A computer-generated random number sequence with variable block sizes will be used to allocate participants in a 1:1:1 ratio. Randomization will be stratified by gender and primary presenting concern (depression vs anxiety) to ensure balanced group allocation on these prognostically important variables. The allocation sequence will be concealed from research staff responsible for recruitment. Because of the nature of the intervention, participants and session facilitators cannot be blinded. Outcome data will be collected primarily through self-report instruments administered through a secure platform, and analyses will be conducted on deidentified group codes until the primary models are finalized.

Adverse events will be monitored throughout the trial. Participants will be asked about dizziness, nausea, eyestrain, disorientation, social distress, and symptom worsening at each VR session and at formal assessment points. The Simulator Sickness Questionnaire (SSQ) will be administered at the end of every VR session to track habituation trajectories and detect differential cybersickness across groups; SSQ scores will be examined as a potential covariate in sensitivity analyses. Sessions may be paused or discontinued if participants experience clinically meaningful discomfort. All adverse events and withdrawals will be recorded.

Predefined clinical thresholds will be used to identify participants requiring immediate attention. A BDI-II total score exceeding 28 (indicating severe depression) or any endorsement of suicidal ideation on BDI-II item 9 (score ≥1) at any assessment point will trigger immediate clinical review. Similarly, a BAI total score exceeding 25 (indicating severe anxiety) will prompt review. Self-report assessments will be scored in real time at each formal assessment point, with automatic flagging of scores exceeding these thresholds. Session facilitators will also conduct a brief verbal check-in before and after each VR session to monitor participant well-being.

When a clinical threshold is met, the session facilitator will immediately notify the principal investigator, who will conduct or arrange a clinical risk assessment within 24 hours. Participants meeting clinical concern criteria will be referred to the psychological counseling center at Hunan Traditional Chinese Medical College for nonurgent concerns or to the psychiatric emergency department at the nearest affiliated hospital for acute crises. The principal investigator (GZW) will serve as the designated safety officer responsible for overseeing all safety monitoring procedures and referral decisions. Participants who are referred for clinical care may continue in the trial if deemed safe by the assessing clinician or may be withdrawn if continued participation poses a risk to their well-being.

Data collection will be conducted via a secure online survey platform and directly from the VR software. All primary analyses will follow the intention-to-treat principle and will include all randomized participants with available data.

All data will be entered into a secure electronic database with verification procedures. Identifiable information will be stored separately from outcome data on encrypted institutional servers, accessible only to authorized members of the research team.

For the primary outcomes, linear mixed-effects models will be used to examine group, time, and group×time effects on BDI-II and BAI scores. Separate models will be fitted for depression and anxiety outcomes, with participant-level random intercepts to account for repeated measures over time. Gender and primary presenting concern (the stratification variables) will be included as covariates, and baseline symptom severity will be considered in model specification as appropriate. The primary planned contrasts will compare the experimental group with the active control group at T1 and T2; comparisons involving the waitlist control group will be secondary and should be interpreted with caution, given the expectancy and novelty confounds inherent in waitlist comparisons. Post hoc pairwise comparisons will be adjusted for multiple testing using a Holm procedure. Secondary outcomes will be analyzed using analogous mixed-model approaches. Cybersickness (SSQ scores) will be examined as a potential covariate in sensitivity analyses to assess whether differential simulator sickness across groups confounds the primary results. Missing data assumptions will be examined, with maximum likelihood estimation in mixed models as the primary approach and multiple imputation and per-protocol analyses used as sensitivity analyses where appropriate. Exploratory analyses will examine whether changes in Basic Psychological Need Satisfaction Scale scores are associated with changes in BDI-II and BAI scores and whether music-movement synchronization moderates selected outcomes; these analyses will be interpreted as hypothesis generating rather than confirmatory causal tests.

This study was approved by the ethics committee of Hunan Traditional Chinese Medical College on September 8, 2025 (YXLL202509006). The trial was prospectively registered at ClinicalTrials.gov on March 15, 2026 (NCT07482852), prior to participant enrollment. Freely given, written informed consent will be obtained from all participants before enrollment. Participation will be voluntary, and participants may withdraw at any time without penalty. Identifiable information will be stored separately from study data on encrypted institutional servers and will be accessible only to authorized members of the research team. Analyses and publications will use deidentified data only, and no personally identifying information will be reported. Any reimbursement or compensation will be provided only as approved by the ethics committee and described in the participant information sheet.