EEG-based Motion Task for Healthy Subjects Using Time Domain Feature Extraction: A Preliminary Study for Finding Parameter for Stroke Rehabilitation Monitoring

Abstract

Nowadays, Stroke has been the second most cause of deaths in the world after Ischaemic heart disease. Rehabilitation of stroke patients after the attack is still the most effective way of restoring the patients to normal. However, most of the rehabilitation methods are done manually. In most of stroke rehabilitation programs, the evaluation procedures are still done using visual observation by clinicians. Considering that background, this study is the preliminary stage in preparing stroke rehabilitation monitoring by using EEG. Since EEG has been used widely for studying the human motion and human control especially in the neural system, applying EEG for stroke rehabilitation monitoring and evaluation would be a great solution because the assessment of the rehabilitation progress can be quantified in a better way. Eleven healthy subjects performing specific motion tasks: baseline (no motion), finger motion, grasping and elbow-flexion, the EEG is then recorded and extracted. Statistical parameters were calculated to get the EEG pattern such as mean and mean absolute value (MAV). From the data analysis, we found that during motion, the value of MAV was tended to decrease in low beta bands. We also found that the maximum amplitude of relaxing or no motion (MAR) is higher than the maximum amplitude of the movement (MAM) in the low beta band both C3 and C4 channel.