Robotic assessment of sensorimotor impairment began in the mid 1990s as a means to address some of the issues regarding inter-rater reliability and the lack of precision associated with traditional measures of sensorimotor impairment. Robotic measures of postural control, reaction time, movement smoothness, and movement error associated with robotic assessment of the upper-limb fail to recognize the inherent spatial and geometric differences between stroke and control hand path trajectories. In this study we propose the application of a class of algorithms, Dynamic Time Warping, designed to quantify the spatial difference and skew between hand written characters and vocal waveforms as a means for identifying individuals exhibiting sensorimotor impairment. In order to achieve this 85 stroke subjects, and 54 age, gender, and handedness matched control subjects, underwent robotic assessment of the upper-limb. Subjects were identified as either stroke or control using a K Nearest Neighbour classifier with a Dynamic Time Warping distance metric. Classification accuracy, sensitivity, and specificity in excess of %80 percent was achieved.