**Background:** Cognitive impairment (CI) is highly prevalent in bipolar disorder (BD), affecting approximately 30% of patients across at least two cognitive domains. CI involves deficits in attention, verbal fluency, memory, and executive function, and is a barrier to clinical, personal, and social recovery. Traditional cognitive remediation (CR) programs have shown preliminary but limited evidence in BD due to small samples and high dropout rates. Virtual reality (VR)-based CR offers ecological, engaging scenarios that may improve generalization of trained skills. To date, no studies have examined fully immersive VR-based CR specifically for BD.
**Methods:** This was a randomized-controlled, two-arm, cross-over feasibility trial conducted at the Consultation and Psychosomatic Psychiatry Center of the University Hospital of Cagliari, Italy. Fifty participants meeting DSM-IV criteria for BD were randomized (25 per arm). The experimental group (A) received a fully immersive VR-based CR intervention (CEREBRUM software on Oculus Go) as an add-on to conventional treatment for 3 months (24 sessions of 45 min, twice weekly). The control group (B) received conventional treatment only (waiting list). After 3 months, group A underwent a 1-month interval followed by conventional treatment, while group B crossed over to receive the VR-CR intervention. Due to dropouts, the final analysis included 39 participants in the experimental group and 25 in the control group. Cognitive functions were assessed using the Rey Figure Test, Matrix test, Rey's Words Test (immediate and delayed recall), Test of the Tale, Forward and Backward Digit Span, Trail Making Test (A and B), Phonological and Semantic Verbal Fluency tests, Digital Symbol Substitution Test, Stroop Test, Frontal Assessment Battery (FAB), and Cognitive Estimates Test (CET). Personal and social functioning was evaluated using SF-12, TAS-20, SAS, PHQ-9, HoNOS, and BRIAN. Repeated-measures ANOVA with Bonferroni correction was used to compare pre- and post-intervention outcomes between groups.
**Key Results:** For the primary feasibility outcome, 119 subjects were contacted, 62 enrolled, and 50 randomized. The dropout rate was approximately 20% (7 of 25 in group A, 4 of 25 in group B after crossover), primarily due to work commitments or distance from the health service. No participants dropped out while on the waiting list. Satisfaction was high: 48.7% rated the intervention as "excellent," 28.2% as "great," and 23.1% as "good." Side effects were minimal—76.9% reported no side effects during the first session, and 87.2% reported none at the end of the intervention. For cognitive outcomes, statistically significant time×group interactions favoring the VR-CR group were found for: attention (Matrix test, p=0.002; Rey's Words immediate recall, p=0.019), memory (Rey's Words delayed recall, p=0.003), verbal function (Verbal Semantic Test, p=0.010), and executive function (Cognitive Estimates Test, p=0.003). For personal and social functioning, significant improvements were observed in depressive symptoms (PHQ-9, p=0.030), biological rhythms regulation (BRIAN, p=0.029), and emotional awareness (TAS-20, p=0.007).
**Clinical Implications:** This feasibility trial provides preliminary evidence that a fully immersive VR-based CR program is acceptable, well-tolerated, and associated with improvements in multiple cognitive domains and functional outcomes in people with BD. The 20% dropout rate compares favorably to the approximately 30% dropout typically seen in traditional CR interventions for BD, suggesting VR may enhance engagement. The cross-over design ensured all participants received the intervention, an ethical advantage. However, the small sample size (limited further by early enrollment cessation due to COVID-19), lack of blinding, and completer-only analysis are important limitations. These findings support the need for larger, adequately powered randomized controlled trials to confirm efficacy and establish VR-CR as a standard rehabilitation option for cognitive impairment in bipolar disorder.