Abstract Technological progress has facilitated the integration of VR in sectors like E-commerce and design, and improves how products are represented by virtual prototypes (VPs) using head-mounted displays (HMDs), allowing their qualities to be understood even when no physical prototype is available. Despite VR having become well-established as an efficient tool to evaluate VPs, factors like physical activity, presentation type and product design may influence consumer perceptions. This research work presents two case studies to analyse the impact of presentation type (real vs. VR) and product design (high- vs. low-end product) on perceiving a treadmill. The Semantic Differential technique was used to evaluate product attributes, and was supplemented with measurements of trust, perceived fatigue, satisfaction with the physical experience, cybersickness and presence. The results showed no differences for how the product was perceived between visual settings, but differences appeared between different designs, which suggests that VR is a valid option for evaluating products. Nevertheless, special attention needs to be paid to the impact of this technology on user well-being, because some factors like eye strain may increase in complex settings and may affect the assessment of some product attributes. Finally, no differences were found in perceived fatigue, but satisfaction with the experience was higher in the VR environment with the high-end product, suggesting that the design appeal may positively impact the experience in virtual environments.

Abstract Traditional psychiatric assessments often depend on self-reports, which may be biased, while real-world evaluations pose safety and feasibility challenges. Specific phobias are underdiagnosed and often comorbid, with affected individuals at elevated risk for severe outcomes. This study aimed to integrate the DSM-5 diagnostic framework, Virtual Reality (VR), and Machine Learning (ML) to identify multimodal features most predictive of phobia severity. Ninety-four adults completed 100 independent trials across four VR scenarios (neutral, Cynophobia, Astraphobia, combined), during which behavioral, physiological, demographic, and self-reported measures were recorded alongside DSM-5 severity ratings. Feature selection and ML analyses revealed that Cynophobia severity was best predicted by a multimodal subset comprising age, neutral-scenario task completion time, Cynophobic virtual distance, oxygen-level variations, and DSM-5 Astraphobia severity. The next most predictive subset comprised sense of presence and DSM-5 Astraphobia severity. The Naïve Bayes classifier achieved robust performance across all features, underscoring the complementary value of multimodal inputs. These findings suggest a potential comorbidity between Cynophobia and Astraphobia and demonstrate the diagnostic relevance of integrating VR-based behavioral and physiological measures with DSM-5 criteria. The study contributes an integrated and scalable framework for enhancing the objectivity and reliability of phobia assessment and highlights future potential for VR–AI applications in clinical diagnostics. Graphical abstract

Abstract This systematic scoping review mapped the empirical literature on Virtual Reality Exposure-Based Therapy (VRET) delivered via commercially available head-mounted displays for adult anxiety-related disorders, to characterize study targets, methods, and gaps. A comprehensive database search yielded 1097 records. Publications were excluded if anxiety was not a measured outcome; if exposure was paired with other techniques (e.g., relaxation or additional therapies) in a way that prevented evaluation of exposure as a stand-alone component; or if virtual reality was delivered via methods other than head-mounted displays. Thirty studies met inclusion criteria and were charted for synthesis. The 30 included studies examined phobias (n = 11), Obsessive-Compulsive Disorder (n = 4), public speaking anxiety and social anxiety (n = 13), and other anxiety presentations (n = 2; social physique anxiety; MRI anxiety). Most studies reported pre- to post-intervention reductions in anxiety symptoms, but study designs and outcome measures varied substantially. Acceptability and engagement were generally favorable when assessed, but measurement approaches were inconsistent and adverse effects were not uniformly reported. Methodological heterogeneity and limited replication constrained cross-study comparability. The current evidence base indicates growing application of HMD-based VRET across multiple anxiety-related targets and suggests potential clinical benefit in many studies; however, heterogeneity and small samples limit the strength of inferences regarding comparative outcomes. Future research would benefit from standardized reporting of intervention parameters and equipment, consistent measurement of acceptability, and adequately powered comparative designs with longer follow-up to clarify where HMD-based VRET is most feasible and beneficial.

Abstract Virtual reality (VR) has become more affordable and is increasingly used for therapeutic and educational purposes where it is important to transfer learned experiences to the real world. But is it worth switching from conventional 2D displays to VR to strengthen such learning effects? Here, we investigated how well incidentally learned emotional and neutral scenes, encoded either in VR or as a conventional 2D presentation (encoding context VR, 2D), could be recognized in congruent or incongruent presentation contexts (retrieval context VR, 2D). Using a 3 × 2 × 2 mixed design, 60 participants viewed (and incidentally encoded) emotional and neutral scenes in VR and as 2D video. After 24 h, 30 participants performed an unannounced old-new judgment task in 2D, while the others completed it in VR. 2D-encoded scenes were recognized better in the congruent 2D than in the incongruent VR context. Context congruency effects were less pronounced for VR-encoding. On average, participants reliably recognized VR-encoded scenes in the incongruent 2D retrieval context. Participants retrieved emotional scenes better and more confidently than neutral ones, independent of encoding context. However, scenes experienced in VR were more likely to be perceived as emotional than when the same scenes were viewed as 2D video. Taken together, the presentation mode of a scene (2D, VR) served as a contextual mnemonic aid, mainly for 2D-encoded scenes. Overall, VR-encoded scenes were less dependent on context congruency effects, suggesting a reduced sensitivity to the retrieval context rather than a direct transfer advantage. Finally, VR-encoding increased the likelihood that the content would be perceived as emotional, suggesting its potential as a useful educational and therapeutic tool.