In this paper, the authors describe how robotic measurements may be used a surrogate marker of motor function—a “biomarker” of stroke recovery in those with hemiparesis after stroke. The concept is that a robotic measurement may have more precision in measurement, such that early differences in functional recovery can be detected. This may decrease the enrollment numbers needed to demonstrate effectiveness in clinical trials of motor recovery and neuroprotection. The authors compared robot assisted measurements in kinematics and kinetics (RMK2) to standard clinical rating scales such as the NIHSS, mRS, Fugel-Meyer and Motor Power.
RMK2 measurements are actually a composite of 35 measurements of arm average speed, peak speed, ability to move against resistance, among others. A complicated technique using neural networks to build non-linear models of the clinical scalses was used for comparison. The study found that in stroke patients (evaluated at multiple time points up to 90 days), robotic measurements correlated with standard rating scales with high correlation coefficients (r2 of approximately 0.6). The authors predicted that the standardized effect size would be increased by 1.47, allowing for greater sensitivity in measuring recovery.
The authors point out that the population used in this study was highly selected, and utility in hemiplegic patients is unknown. The handedness of subjects in the paper was also not clear. A majority of the subjects did not complete all visits. Taking these limitations into account, there is potential for robotic measurements to become a standard of motor outcome in clinical trials. The promise of saving money and completing trials sooner, fuels the progress of robotic technology. We look forward to the future of robots as biomarkers and neuroprotection in stroke!