On the Impact of Pathological Tremor Intensity on Noninvasive Characterization of Motor Unit Discharge Properties

Abstract

The impact of severity of pathological tremor on surface EMG decomposition was systematically assessed on eight essential tremor patients. The inertial data and surface EMG signals were concurrently recorded from wrist extensor and flexor muscles of both patients’ arms. The inertial recordings were segmented into different tremor cycles and the tremor amplitude was assessed in each tremor cycle. Surface EMG was decomposed by Convolution Kernel Compensation (CKC) technique in order to yield individual motor unit discharge patterns in each tremor cycle. Accuracy of EMG decomposition was assessed for each identified motor unit and was largely uncorrelated with tremor amplitude. In all the patients, the percentage of EMG energy identified by decomposition and the number of identified motor units were found to be positively correlated with tremor amplitude, though the correlation was relatively weak and not always significant. The results demonstrate that the CKC decomposition not only copes with moderate and severe tremor but also improves its performance with tremor intensity.