Quantitative evaluation of cerebellar ataxia based on EMG signals

Abstract

In this paper, we propose a new method to make a quantitative evaluation of motor commands based on the EMG signals. Especially, we tried to resolve a remaining issue in our previous study: relatively poor estimation of the wrist joint torque from the EMG signals for extreme movements, such as jerky movements of the cerebellar patients. Subjects included eight cerebellar patients and eight normal controls. As an experimental task, we asked the subjects to perform step-tracking wrist movements with a manipulandum, and recorded movements of the wrist joint and activities of the four wrist prime movers with surface electrodes. For improved estimation of the wrist joint torque from the EMG signals, we developed a new approximating method, by considering the moment arms of the four muscles during the movement. As a result, our improved method proved higher correlations between the kinematics torque and the estimated torque with the four EMG signals for both the cerebellar patients and the normal controls than the previous method. In fact, we confirmed the effectiveness of our improved methods, identifying the causal abnormality of muscle activities for the cerebellar ataxia with high accuracy.