Exploring Cognitive Functions Through Virtual Reality: a Systematic Review of Ecologically Valid Assessments.
TL;DR
Findings suggest that HMD‑VR environments can elicit more ecologically relevant cognitive behaviors and may capture performance features difficult to assess with traditional tools, however, the current evidence base is limited by methodological variability, restricted age diversity, and underreporting of core psychometric indices.
Abstract
This PRISMA‑guided systematic review synthesized evidence on the psychometric properties of immersive head‑mounted display (HMD) virtual reality (VR) tools developed for the ecological assessment of cognition across different developmental stages. The review targeted studies evaluating attention, inhibitory control, memory, and executive functions; published between January 1, 2021, and March 13, 2025, across PubMed, PsycINFO, Scopus, and Web of Science. Twenty‑one studies met the inclusion criteria. Convergent validity emerged as the most frequently reported property, appearing in 14 studies (67%) and showing predominantly low‑to‑moderate correlation coefficients (r = .29-0.66). Construct validity was reported less often (3 of 21; 14%) explaining 62%-80.18% of the variance, while test‑retest reliability data were available in only two studies (9.5%), showing overall significant correlation coefficients (r = .18-0.85). Discriminant validity (7 of 21; 33%) was adequate in 5 studies but limited in 2, which reported moderate cross-construct correlations. Predictive validity was assessed in five studies (24%), mainly for psychopathology‑related applications, though most relied on single‑rater parent questionnaires without behavioral or longitudinal validation, with evidence ranging from strong classifier performance (AUC = 0.92) to moderate predictive correlations (r = .38-0.43) and variable diagnostic classification rates. Child and adolescent investigations (9 of 21 studies) consistently demonstrated smaller sample sizes and less comprehensive validation protocols than adult‑focused research. Despite these heterogeneities, findings suggest that HMD‑VR environments can elicit more ecologically relevant cognitive behaviors and may capture performance features difficult to assess with traditional tools. However, the current evidence base is limited by methodological variability, restricted age diversity, and underreporting of core psychometric indices. Future research should adopt standardized validation protocols, include representative developmental samples, and integrate multi‑informant and longitudinal approaches to verify the ecological validity and clinical generalizability of HMD‑VR assessments.