Abstract—With the advent of new approaches to upper extremity recovery after stroke and spinal cord injury, the quantitative evaluation of hand function has become a crucial component of outcome evaluation. Recently we developed a workstation, the ReJoyce (Rehabilitation Joystick for Computer Exercise) on which subjects perform a variety of movement tasks while playing computer games. An important feature of the system is the ReJoyce Automated Hand Function Test (RAHFT). In this study we compared and validated the RAHFT against two widely-used clinical tests, the Action Research Arm Test (ARAT) [1][2] and the Fugl-Meyer Assessment (FMA) [3]. All three tests were performed in 34 separate sessions in 13 tetraplegic individuals. Principal component and regression analyses revealed that both the ARAT and the RAHFT correlated well with the first principle component fitted to the scores of the three tests. The FMA was less well correlated. These data help validate the RAHFT as a quantitative, automated alternative to the ARAT and FMA. The RAHFT is the first comprehensive test of hand function that does not depend on human judgment. I. INTRODUCTION growing body of evidence suggests that intense, task- oriented rehabilitation improves functional outcomes [4]. In recent years several groups have explored ways of automating rehabilitation [5] [6] [7], primarily to reduce costs, and more recently, to allow a continuation of tele- supervised rehabilitation exercises in patients’ homes after they leave the clinical environment [8] [9]. A The exercise tasks performed by patients in the conventional rehabilitation setting tend to be repetitive and boring. In order to maintain patient interest there is an increasing trend to incorporate computer games into treatment protocols [8]. This has been accelerated by the world-wide adoption of the Nintendo Wii into rehabilitation clinics [11] [12]. The Wii allows users to play computer games by moving a hand-held motion sensor. It is useful for practising unloaded motions of the shoulder and elbow but it has no sensors of hand function so even if its motion signals were available for analysis, they would not form a sufficient basis for a quantitative test of arm and hand function. The ReJoyce (Rehabilitation Joystick for Computer Exercise) provides the requisite signals. It is a passive workstation April 1st, 2011. This work was supported in part by International Spinal Research Trust (ISRT), Canadian Foundation for Innovation (CFI), Alberta Heritage Foundation for Medical Research (AHFMR) and the Canadian Institutes of Health Research (CIHR) J. Kowalczewski is with the Centre for Neuroscience at University of Alberta (phone: 780-492-1616; e-mail: jan@angeltear.com). E. Ravid is with the Centre for Neuroscience at University of Alberta(e- mail: ravid@ualberta.ca). A. Prochazka is with the Centre for Neuroscience at University of Alberta(e-mail: arthur.prochazka@ualberta.ca). comprising a segmented arm that presents the user with a variety of spring-loaded manipulanda [15]. Each manipulandum is instrumented with one or more sensors, whose signals are fed to a computer. The signals are analyzed with custom software to control computer games and to run the “ReJoyce Automated Hand Function Test” (RAHFT). For an automated hand function test such as the RAHFT to become clinically accepted, it must be shown to be substantially equivalent to existing clinical tests. The equipment should be affordable and simple to use, with software that provides audiovisual prompts during the test and quantitative results in the form of spreadsheets and printed reports. The purpose of this study was to examine the RAHFT and see how well it correlated with two conventional, clinically accepted hand function tests. II.METHODS A. Participants : Thirteen people aged between 24 and 56 with tetraplegia resulting from a C5-C6 SCI participated in this study, as part of a broader project [15]. B. Procedures : Three hand function tests were performed in randomized order in single sessions that took place at 2-weekly intervals during the 6-week treatment periods. The three tests were the Fugl-Myer Assessment (FMA), the Action Research Arm Test (ARAT), and the ReJoyce Arm-hand Function Test (RAHFT), described in detail below. The FMA and ARAT tests were videotaped and scored by a blinded, independent rater. C. Apparatus : Each of the six manipulanda in the ReJoyce workstation was designed to represent a task commonly encountered in daily life. The spring-loaded arm had 4 degrees of freedom of movement and was instrumented with rotational potentiometers about each joint. The spring loading of the arm provided some elastic resistance to movement and ensured that the manipulanda returned to a neutral position when they were released. The manipulanda included a pair of horizontal handles that could be rotated about their long axis, a vertical peg that could be lifted, a gripper the size of a pop can that could be squeezed and a spring-loaded door knob with an exposable key-like element, both of which could be independently rotated. The easiest task to perform was to grasp one or both of the horizontal handles. The gripper, the door knob, key and peg, were located above the Fully-Automated Test of Upper-extremity Function J. Kowalczewski, E. Ravid, A. Prochazka 978-1-4244-4122-8/11/$26.00 ©2011 IEEE 7332 33rd Annual International Conference of the IEEE EMBS Boston, Massachusetts USA, August 30 - September 3, 2011