Robotic technologies provide objective, highly reliable tools for assessment of brain function following stroke. KINARM is an exoskeleton device that quantifies sensorimotor brain function using a visually guided reaching task among many other behavioral tasks. As further tasks are developed to more broadly assess different aspects of behavior using the robot, techniques and approaches are required to reduce the time it takes to complete each task. The present study investigates how the value of robot-measured parameters changes under alternative schemes that significantly reduce assessment time compared to the current assessment protocol for the visually guided reaching task. Results of the study are validated by addressing an important diagnostic question using an SVM classifier, showing that the alternative schemes provide nearly identical performance in terms of classification sensitivity, specificity and accuracy.