Assessment of upper and lower limb kinematics using the low-cost Microsoft Kinect: reliability and validity in people with acquired brain injury.

Abstract

INTRODUCTION: Three-dimensional motion analysis is the criterion-reference for evaluation of movement kinematics. Whilst, Optitrack is a new, more user-friendly system comprehensive gait and upper limb kinematic assessment is still costly, time-consuming and requires specialist training, limiting its routine clinical use. The Microsoft Kinect is a low-cost, three-dimensional, depth-sensing camera, originally developed for the Xbox. It is portable and cheap with reduced set-up and analysis burden. AIM: To establish the test-retest reliability and concurrent validity of the Kinect compared to three-dimensional motion analysis for quantifying upper and lower limb kinematics during gait in people with acquired brain injury. DESIGN: Observational study. METHODS: Twenty people with chronic acquired brain injury with gait abnormalities and upper limb associated reactions were recruited. All participants underwent gait analysis with an Optitrack three-dimensional motion analysis system and the Kinect on two testing occasions, one-week apart. Four trials at self-selected walking speed were recorded. mean, standard deviation, maximum, minimum and total joint range of motion were captured for the upper and lower limb joint axes. Correlation coefficients reported the association between the two systems for concurrent validity and between sessions for the Microsoft Kinect's test-retest reliability. RESULTS: The Kinect displayed excellent concurrent validity (r > 0.7) to the Optitrack for shoulder flexion range of motion, shoulder abduction mean and maximum, elbow flexion maximum, hip flexion maximum and all measures of knee flexion for the affected limbs. The Kinect displayed excellent test-retest reliability (r > 0.8) for all upper and lower limb joint axes. CONCLUSION: - The Kinect displays promising concurrent validity to three-dimensional motion analysis for key kinematic variables for upper limb associated reactions. - The Kinect is highly reliable. - The Kinect may provide a valuable clinical tool to simply and objectively measure gait. - Investigation into the Kinect's responsiveness is warranted. - The Kinect is unable to capture forearm, wrist and ankle joint data.