Predication of motor recovery using kinematic and kinetic measures

The objective of this study was to characterize the time-course of changes in motor impairment after stroke, and to use the kinematic and kinetic measures of elbow voluntary movement at 1 month to predict the ensuing recovery patterns over 1 year. The arm motor impairment was assessed using the Fugl-Meyer Assessment (FMA) of the upper extremity at 1, 2, 3, 6, and 12 months after stroke. Several kinematic features of voluntary rapid elbow extension were quantified by measuring the movement trajectory and its derivatives. The "growth mixture” and "logistic regression" models were used to characterize recovery patterns of FMA over 1 year and to predict these patterns, based on the kinematic and kinetic measures at 1 month. We observed two major distinct recovery classes. Class 1 subjects started with low-level FMA score and then increased significantly before tapering off gradually. Conversely, class 2 subjects started with a high-level FMA score but remained invariant or increased slightly. Our results showed that the kinematic and kinetic measures at 1 month were a significant predictor of the motor recovery. These findings demonstrate that the logistical regression class membership may enable us to accurately predict arm impairment recovery during the first year using kinematic and kinetic measures at the early stage of stroke.