Feature extraction for movement
disorders of neurological patients based on EMG signals Department of Sensory Motor and
Systems
Abstract In this study, we extracted feature parameter for movement disorders of neurological patients using a wrist movement. Specially, based on EMG signals, we captured feature patterns of movement disorders for Parkinson’s patients and cerebellar patients from the motor commands level. As an experimental task, we asked subjects to perform smooth pursuit wrist movement, in which the subjects follow a smoothly moving target with a cursor. We recorded movement of the wrist joint and EMG signals of four wrist prime movers with surface electrodes. The participants included eight patients with cerebellar diseases, four patients with Parkinson’s disease and eight normal controls. We succeeded to extract two feature parameters from the EMG signals of the four wrist prime movers, Variability of Total Contraction (VTC) and Directionality of Muscle Activity (DMA), which characterize the pathological patterns of muscle activities for the neurological disorders. We found that these feature parameters, if combined appropriately, are useful to characterize complex patterns of muscle activities in a way easy to be recognized visually. In other words, the high-dimensional parameter space is also useful to evaluate effects of a medical treatment as a shift toward or away from the normal control in the parameter space. Consequently, it is expected that our proposed methods will be useful for a navigation system of medical treatments or rehabilitation based on Information Technology (IT) in the future. Keywords: Movement Disorders, Feature
Extraction, EMG Signal, Motor Commands. +: Corresponding author: Jongho Lee Setagaya-ku,
Tokyo 156-8506, Japan, Tel: +81-3-6834-2343 |