Multimodal Data Acquisition for the Assessment of Cerebellar Ataxia via Ballistic Tracking

Cerebellar ataxia (CA) results from injury to or disease of the cerebellum. It describes the resulting motor dysfunction, characterised by inaccuracy, incoordination and delay in initiation of movement, tremor, and imbalance. Assessment of ataxia to diagnose and monitor progress is by clinical observance of the performance of standard motor tasks. An accurate instrumented measurement of CA would therefore be of great interest. This study was aimed at assessing upper-limb ataxia during ballistic tracking of a computer-generated target in individuals with CA and controls using motion measures obtained from a Kinect camera and a wearable motioncaptured device. A set of features derived from these motion measurements were used to develop a method for objective quantification of CA. Difference between ataxic and non-ataxic movements can be readily be observed in features from both devices (p = 0.008) and their values associated with a standard clinical scale (rho = 0.80, p < 0.001). The combination of multimodal features improved the ability to distinguish between CA subjects and controls and to measure the severity of upper limb ataxia.