Does immersive VR improve diagnostic accuracy of post-stroke spatial neglect relative to conventional and digital tests?

ObjectiveAssessing post-stroke spatial neglect in immersive Virtual Reality (iVR) may improve diagnostic accuracy due to its ability to combine experimental control with increased ecological validity. This study investigated the diagnostic accuracy of an iVR assessment relative to conventional and non-iVR digital tests. Additionally, feasibility was evaluated. MethodStroke patients performed an iVR assessment, conventional pen-and-paper tests (3 Cancellation tests and a line bisection test), and non-iVR digital tests (Digital Cancellation and Posner test). A new Bayesian statistical method was developed to address the absence of a gold standard for neglect. Instead of conventional diagnostic cut-offs, we quantified the uncertainty of classifying patients as having neglect or not and incorporated this into estimating diagnostic accuracy. Feasibility was evaluated using recruitment and drop-out rates, sample characteristics, cybersickness and user experience. Results54 stroke patients and 56 healthy controls completed the study. Both iVR and non-iVR digital tests had higher diagnostic accuracies than pen-and-paper tests, although differences were not statistically significant due to high diagnostic uncertainty (resulting from inconsistencies across tests). For 6 patients, the iVR assessment and Posner test indicated neglect, while pen-and-paper tests did not. The iVR assessment was feasible for most patients, with low cybersickness and positive evaluations. ConclusionThe iVR assessment and Posner test identified spatial biases missed by conventional tests. Inconsistencies across tests highlight the complexity of neglect assessment, suggesting that future studies should explore crucial test characteristics needed for a precise and sensitive evaluation. Furthermore, results established good feasibility of iVR neglect assessment.