Introduction: Virtual Reality (VR) has been found useful for numerous rehabilitation applications, but has some intrinsic constraints such as the need for a visuospatial transformation when guiding movements. Augmented Reality (AR) is a new approach to human-computer interaction that enables patients to interact directly with virtual objects. The current study compared AR and VR in a stroke rehabilitation setting.
METHODS: The Fruit Ninja game simulates a rehab setting by having subjects perform repeated goal-directed wrist/hand reaching tasks. Subjects held a cup-shaped color-marker in the paretic hand, then reached for a virtual fruit target that sliced in 2 when reached. This game was implemented in both AR and VR settings, with identical movement demands across the two. The target plus real-time visual feedback on hand movements were provided by a computer monitor in VR, and by a projection onto a tabletop in AR. After undergoing baseline assessments (arm motor Fugl-Meyer scale (FMA) and Box and Blocks (B&B)), 10 patients with hemiparetic stroke >6 mo prior and age >18 yr played three 1-min rounds each of the AR and VR games; 4 other subjects who were unable to hold the color-marker object were excluded from current analysis.
RESULTS: Of the 10 patients, age = 59±10 yr (mean±SD), FMA score = 57±11 (range 31-66), Hand/Wrist FMA subscore = 22±3 (range 15-24), and B&B score = 41±13 (range 16-58). When playing the exact same Fruit Ninja game, all 10 patients scored significantly (p<0.0001) higher in the AR setting (60±9 targets, range 48-78) as compared to the VR setting (48±8 targets, range 37-64 setting. Also, AR scores were stronger correlates of FM Hand/Wrist (rho=0.68, p<0.04) and B&B scores (rho=0.70, p<0.03) than were VR scores.
CONCLUSIONS: This study shows promising results with use of Augmented Reality in a patient-computer interface. Results also suggest advantages as compared to use of a Virtual Reality approach, possibly due to the fact that moving the hand requires a visuospatial transform in the VR setting but not in the AR setting. Compared to VR, AR scores were higher and correlated better with clinical scores, suggesting great potential for the use of Augmented Reality in a patient-computer interface during stroke rehabilitation.