Assessment and monitoring of recovery of spatial neglect within a virtual environment

Spatial neglect has proven to be a significant factor limiting the success of the rehabilitation process following stroke. Current tests for neglect however have several substantial drawbacks, which often lead to a misdiagnosis of less severe cases. Further, while asymmetries of performance have been reported in the past along independent spatial dimensions, current tests are mostly limited to the horizontal dimension and do not reflect the reality of a three-dimensional world. We have previously demonstrated the feasibility of virtual reality tools for detailed assessments of attentional deficits [1]. We now provide further evidence for the sensitivity of our Virtual Environment for Spatial Neglect Assessment (VESNA) application for assessment of neglect as well as for monitoring recovery of patients. Seven stroke patients with neglect, nine stroke patients without neglect and nine age-matched healthy controls were tested on a target-detection task. Subjects were exposed to a three-dimensional virtual scene and were instructed to press a response button when they detected a target appearing within the scene. Percent of correct detection and reaction time to initiate a button press were calculated. Our results indicated significant differences between neglect patients and control subjects. All neglect patients exhibited asymmetries of performance, where their mean reaction time and detection accuracy systematically varied across space. This asymmetry was not a harsh transition but instead showed a gradual reduction of attention across the space. Importantly, while these results indicated an obvious spatial neglect for all seven neglect patients, their performance on the standard paper-and-pencil tests, administered at the day of testing, was less conclusive. A follow-up study with two of the neglect patients (10 months following the initial testing) revealed an obvious recovery pattern, showing a reduction of the spatial bias over time. By contrast, the paper-and-pencil tests showed no obvious change. Thus, we demonstrated that our paradigm