A two-arm randomized controlled trial (RCT) was conducted among undergraduate students at a public university in the northeastern U.S. between February 2024 and October 2025. Participants were recruited using multiple strategies while maintaining privacy and confidentiality. Study information was disseminated through university-wide and classroom announcements, and recruitment flyers were posted on on-campus bulletin boards accessible to students. Interested students contacted the research team and completed an initial screening to assess eligibility prior to enrollment.

Eligible participants were full-time undergraduate students aged ≥18 years who scored ≤13 on the Revised Life Orientation Test (LOT-R), had no significant visual or hearing impairments, reported no history of severe motion sickness, and had no prior experience with the VR games used in the control condition. Exclusion criteria included inability to provide informed consent, current participation in psychosocial therapy, or current use of antidepressant or antipsychotic medications within the past 28 days, in order to ensure that observed effects could be attributed to the PCBT-based VR intervention rather than to concurrent mental health treatments. A prior history of psychotherapy or psychiatric diagnosis, in the absence of ongoing treatment, did not preclude participation.

During all VR sessions, adverse events (e.g., cybersickness, dizziness, nausea) were actively monitored using structured verbal symptom checks and continuous observation. Participants were informed that they could stop the session at any time if they experienced discomfort. No adverse events were reported or observed throughout the study.

A priori power analysis was conducted using G*Power 3.1. Assuming a medium effect size (r = 0.30) between optimism and mental health variables (Lai et al., 2024), α = 0.05, and 80% power, at least 84 students were required for screening. For the intervention phase, a repeated-measures ANOVA with an effect size of f = 0.30 required 24 participants (12 per group). For four repeated assessments, 18 participants (9 per group) were sufficient to detect within–between interactions.

Baseline sociodemographic information and psychological assessments were collected using Qualtrics surveys at four time points: pre-intervention (T1), post-intervention (T2), 3-month follow-up (T3), and 6-month follow-up (T4). All participants provided informed consent electronically before beginning the survey. Trained assessors blinded to group allocation conducted the outcome assessments. Each participant received a unique identifier to maintain confidentiality and link longitudinal data. All data were stored on secure, password-protected university servers with restricted access.

All analyses were conducted using SPSS 28.0 (IBM Corp.) and the R package lme4 for mixed-effects modeling (Douglas et al., 2015). Descriptive statistics (means, standard deviations, and frequencies) were used to summarize demographic characteristics and baseline psychological measures. Group differences at baseline were examined using Mann–Whitney U tests for continuous variables and chi-square tests for categorical variables.

Longitudinal intervention effects were examined using linear mixed-effects models that incorporated all repeated measurements (T1, T2, T3, and T4). Mixed-effects modeling was selected to maximize statistical efficiency and retain participants with partial follow-up data. In this model structure, the main effect of the intervention (coded as 0 for all control group observations and the experimental group at T1 (baseline); 1 for the experimental group T2, T3, and T4) estimates the immediate level change from baseline to post-intervention. The month variable and its interaction with the intervention group then capture the subsequent longitudinal trajectories (slopes) during the follow-up period. This approach allows for a clear differentiation between the instantaneous treatment effect and the rate of change over time. To distinguish the immediate impact of the intervention from the long-term maintenance of effects, time was treated as a numerical variable representing months since the intervention’s completion. Specifically, baseline (T1) and immediate post-intervention (T2) were both coded as month = 0, while follow-up assessments at 3 months (T3) and 6 months (T4) were coded as month = 3 and month = 6, respectively.

The model included intervention group (PCBT-based VR vs. control), time, and the interaction (group × time) as fixed effects. The interaction tested whether longitudinal change over time differed between groups. Age, gender, and academic period (semester vs. non-semester) were included as covariates. A random intercept for each participant accounted for within-subject correlations due to repeated measures. Model coefficients (β), standard errors (SE), and p-values were reported for all psychological outcomes, including optimism, stress, anxiety, and depression. Statistical significance was set at p < 0.05, with trend-level effects defined as 0.05 ≤ p < 0.10. Given the exploratory nature of this pilot study, no correction for multiple comparisons was applied.

Twenty-eight of the 115 invited students enrolled in the study, yielding a recruitment rate of 24.35%. All participants (100%) completed the pre-intervention (T1) and post-intervention (T2) assessments. At the 3-month follow-up (T3), 27 participants (96.43%) completed the assessment, and 25 participants (89.29%) completed the 6-month follow-up (T4) (Figure 2).

A total of 28 students participated in the study, with 13 assigned to the control group and 15 to the experimental group (Table 1). The mean age was 20.54 years (SD = 1.39) in the control group and 19.33 years (SD = 1.35) in the experimental group. Most participants identified as female (69.23% in the control group; 60.00% in the experimental group). No significant group differences emerged for gender (p = 0.61). The experimental group was slightly younger than the control group (p = 0.03); all models were adjusted for age. Racial/ethnic distributions were similar across groups, with African American, Afro-Caribbean, and Non-Hispanic White students comprising the largest categories. No significant between-group differences were observed (p = 0.72).

Participants represented diverse academic majors and class years, with no significant differences between groups in either distribution (both p = 0.86). Most students reported no prior psychotherapy, psychiatric medication use, or psychiatric diagnoses, with no significant differences between groups (all p > 0.23). Participants who reported a history of psychotherapy, psychiatric medication use, or psychiatric diagnosis were not receiving active treatment at the time of enrollment and therefore met the study’s eligibility criteria. Baseline optimism scores did not differ significantly between groups (p = 0.89).

Short-term intervention effects were evaluated using the main effect of intervention in the mixed-effects model, reflecting group differences observed during the early post-intervention period. The model indicated a significant short-term increase in optimism among participants in the PCBT-based VR group (β = 2.02, SE = 0.77, p = 0.01) and a significant reduction in depressive symptoms (β = −3.04, SE = 1.15, p = 0.01). Anxiety showed a trend toward short-term reduction (β = −1.55, SE = 0.84, p = 0.07), although this effect did not reach statistical significance, and stress levels did not significantly change during this period (β = −0.75, SE = 1.27, p = 0.56) (Table 2). Collectively, these findings suggest that the PCBT-based VR intervention produced notable early improvements in optimism and depression, with smaller or nonsignificant short-term effects on anxiety and stress.

Long-term effects were examined using the mixed-effects model by assessing the month × intervention interaction term. As shown in Table 3, the month × intervention interaction was not significant (β = −0.32, SE = 0.20, p = 0.12), indicating that longitudinal change in optimism did not significantly differ between groups. Optimism scores increased over time for participants in both groups (β = 0.38, SE = 0.16, p = 0.02). Consistent with this pattern, interaction effects were also nonsignificant for anxiety (β = 0.33, SE = 0.21, p = 0.13), depression (β = 0.51, SE = 0.30, p = 0.09), and stress (β = 0.53, SE = 0.34, p = 0.13) (Table 4). Taken together, these findings indicate that the PCBT-based VR intervention was associated with early post-intervention improvements in optimism. While these initial gains were maintained across the follow-up period, the long-term trajectories did not significantly differ between the intervention and control groups.

This study showed that the PCBT-based VR intervention produced a substantial early advantage in optimism relative to the control condition, and this difference was maintained over time. Although both groups improved at a similar rate during follow-up, the experimental group maintained a higher level of optimism throughout the observation period, suggesting the potential of structured, technology-based interventions to enhance optimism among college students. Although some prior research suggested that PCBT interventions may not yield statistically significant changes in optimism among patients with major depressive disorder, they still showed higher mean optimism scores compared to baseline (Geschwind et al., 2019). This suggests that while PCBT-based VR approaches may not fully eliminate pessimistic thinking in clinical populations, they can promote meaningful gains in optimism, which is a known protective factor for mental health and may reduce vulnerability to stress and depressive symptoms in non-clinical student samples. This finding suggests that optimism increased over time in both groups, supporting the observation that participants’ optimism scores rose gradually across months following the intervention. Similarly, Geschwind et al. (2019) reported improvements in optimism scores after treatment, with participants reaching levels comparable to normative population averages. However, the negative intervention-by-time interaction was not statistically significant (p = 0.12), indicating that long-term optimism trajectories did not differ significantly between groups over the follow-up period. This pattern reflects broader evidence indicating that, without reinforcement or booster sessions, the effects of psychological interventions often attenuate. Consistent with this, meta-analyses of positive psychology interventions have found that initial gains in optimism tend to diminish unless interventions are maintained or repeated (Bolier et al., 2013). The PCBT-based VR intervention produced a significant short-term improvement in optimism among students with low baseline optimism, consistent with prior work showing that PPIs and PCBT can enhance positive expectations and cognitive appraisal (Bolier et al., 2013; Sin and Lyubomirsky, 2009; Lyubomirsky and Layous, 2013). VR immersion may further support these effects by enhancing attentional focus and emotional engagement (Li Pira et al., 2023; Xu et al., 2024; Tarrant et al., 2018). Although optimism remained higher in the intervention group through follow-up, the nonsignificant month × intervention interaction indicates that the rate of long-term change was similar between groups. This suggests that early gains were sustained but did not continue to increase in the absence of reinforcement (Geschwind et al., 2019; Lyubomirsky and Layous, 2013).

After the intervention, the experimental group experienced significant reductions in depression and trend-level reductions in anxiety compared to the control group in the short term; however, the difference was not statistically significant. The finding was aligned with previous research showing that PCBT interventions can effectively reduce depressive symptoms in patients with major depressive disorder (Xu et al., 2024; Freeman et al., 2017; Geschwind et al., 2019). Additionally, online PPIs have been shown to decrease depressive symptoms among Chinese college students (Auyeung and Mo, 2019). University students who participated in PPIs also reported significantly lower levels of depression (Tarrats-Pons et al., 2025). Moreover, some meta-analysis studies further support these findings, indicating that PPIs reliably lead to significant reductions in depression among adults aged 18 and older (Bolier et al., 2013; Sin and Lyubomirsky, 2009; Lim and Tierney, 2023). Furthermore, the decrease in anxiety scores among undergraduate students after this intervention aligns with existing studies that have shown that an interactive virtual environment and VR exposure therapies can reduce anxiety in clinical patients (Freeman et al., 2017; Felnhofer et al., 2019). These findings suggest that combining PCBT principles with immersive VR delivery offers a promising approach for addressing depressive and anxiety symptoms in university settings by integrating evidence-based cognitive restructuring techniques with the engagement and accessibility of technology-based interventions.

The intervention also produced significant short-term reductions in depressive symptoms and a trend toward reduced anxiety, aligning with evidence that PCBT and VR-based approaches can improve emotional regulation and reduce negative affect (Xu et al., 2024; Freeman et al., 2017; Geschwind et al., 2019). These short-term benefits mirror findings from other PPIs implemented among university students (Bolier et al., 2013; Sin and Lyubomirsky, 2009; Auyeung and Mo, 2019). However, consistent with the optimism results, long-term slopes did not differ between groups, suggesting that the intervention’s mental health effects were strongest during the active treatment phase and did not diverge from the control condition over time.

Taken together, the results indicate that the PCBT-based VR intervention is most effective for producing short-term cognitive and emotional benefits, which were maintained but not amplified across 6 months. This pattern is consistent with the broader PPI and PCBT literature, which often reports strong immediate gains that level off without continued practice or booster sessions (Geschwind et al., 2019; Lyubomirsky and Layous, 2013). Additionally, natural academic and environmental fluctuations common in college populations may influence long-term psychological trajectories (American College Health Association, 2024), underscoring the value of incorporating ongoing support to sustain intervention effects.

The findings suggest that VR-delivered PCBT may be a practical and engaging approach for supporting psychological wellbeing in college students. The immersive nature of VR can enhance user engagement and attentional focus, offering a structured environment for practicing positive cognitive and emotional skills (Xu et al., 2024; Freeman et al., 2017; Geschwind et al., 2019). These features may be especially useful for students who prefer technology-based mental health resources or who face barriers to participating in traditional services. The intervention’s ability to produce early improvements that were maintained across several months also highlights its potential as a scalable, low-stigma tool for strengthening positive cognitive resources such as optimism. However, because long-term trajectories did not differ between groups, future VR-based PPIs may benefit from incorporating booster sessions or continued practice opportunities to help sustain and strengthen initial gains.

Several limitations should be considered when interpreting these findings. First, the study was conducted with a small sample from a single public university, which may limit the generalizability of the results. Second, all outcomes were assessed using self-report measures, which may be influenced by social desirability or limited self-awareness (FvdM, 2008; Paulhus and Vazire, 2007). Third, although participants were randomly assigned, the groups differed in age at baseline, which could have influenced outcomes despite statistical adjustment. The significant baseline age difference between groups represents an additional limitation. Participants in the intervention group were, on average, younger than those in the control group, which may have influenced engagement with the VR technology. Younger participants may be more accustomed to immersive digital environments and interactive technologies, potentially leading to greater comfort, immersion, or responsiveness to VR-based interventions. Although age was included as a covariate in the analyses, the small sample size limits conclusions regarding age-related effects. Future studies should aim to more closely match groups on age or examine age as a potential moderator of engagement and intervention efficacy.

Fourth, the intervention period coincided with typical academic fluctuations, which may naturally affect stress, mood, and optimism among undergraduate students (American College Health Association, 2024; Lin et al., 2020). In addition, the VR-based intervention was conceptualized and implemented as a research-oriented, wellness-focused intervention rather than a clinically regulated therapeutic product intended for diagnostic or treatment decision-making. Formal regulatory classification, such as consideration under Software as a Medical Device (SaMD) frameworks, was beyond the scope of this pilot study and was not addressed. This limits conclusions regarding its clinical deployment, regulatory pathway, and scalability beyond research or wellness settings.

Although attrition was low, any loss to follow-up may introduce bias if participants who withdrew differed from those who completed all assessments. Finally, although students with severe psychiatric conditions were excluded, a small number of participants reported a prior history of psychotherapy or psychiatric treatment. While this approach enhanced ecological validity, it may limit the generalizability of the findings and influence the interpretation of intervention effects. Future studies with larger samples should apply stricter inclusion criteria or conduct subgroup analyses to better examine whether prior treatment history moderates intervention outcomes. Taken together, these limitations highlight the need for larger, multisite studies that incorporate objective measures, evaluate regulatory considerations, and further clarify the clinical and translational positioning of VR-based mental health interventions.

Future research should build on this pilot study by recruiting larger and more diverse samples across multiple university settings to enhance generalizability. Expanding the intervention to include booster sessions or sustained practice components may help strengthen and maintain long-term effects, consistent with evidence that ongoing engagement is often necessary to preserve gains in positive psychology and CBT-based programs (Geschwind et al., 2019; Lyubomirsky and Layous, 2013). Incorporating qualitative methods, such as interviews or open-ended feedback, could provide deeper insight into participants’ experiences, perceived benefits, and barriers to engagement. Additionally, examining objective indicators—such as physiological stress markers or behavioral measures—may complement self-reported data and reduce the influence of reporting biases (FvdM, 2008; Paulhus and Vazire, 2007). Finally, comparing VR-delivered PCBT with alternative digital or in-person PPIs may help clarify which delivery formats are most effective, accessible, and feasible for college populations.