Abstract
OBJECTIVE: Cerebellar ataxia impairs coordination and balance, reducing quality of life. Conventional clinical scales, including the Scale for the Assessment and Rating of Ataxia (SARA) and the International Cooperative Ataxia Rating Scale (ICARS), are widely used to assess ataxia but are limited by subjectivity and inter-rater variability. Therefore, we aimed to develop a virtual reality-based system to objectively and quantitatively assess upper limb ataxia. METHODS: A "virtual nose-finger test" was implemented using a head-mounted display, in which participants performed repetitive reaching tasks. Six parameters were measured: four spatial (subtracted length, trajectory ratio, terminal trajectory length, and maximum overshoot distance) and two temporal (required time and movement speed). These parameters were compared across groups, correlated with clinical scales, and analyzed for diagnostic accuracy using receiver operating characteristic curves. Motor adaptation was assessed using parameter changes across trials. RESULTS: Ninety-five participants were recruited: 39 with cerebellar ataxia, 30 controls, and 26 with Parkinsonian disorders. Participants with ataxia exhibited significantly greater spatial deviations and temporal variability than other groups did. Trajectory ratio, required time, and movement speed variability coefficient significantly correlated with clinical ataxia scores. The system demonstrated high diagnostic accuracy from the receiver operating characteristic analyses, and participants with ataxia showed different motor adaptations by compensating for spatial errors through reduced movement speed. INTERPRETATION: This virtual reality-based system enables objective, quantitative, portable, and ambulatory-independent evaluation of upper limb ataxia, enhancing its feasibility in clinical and research settings and its potential as a biomarker for cerebellar ataxia.