Surface EMG classification using moving approximate entropy

Moving approximate entropy has been proposed as a new method to extract information from the surface electromyographic signal. Twenty subjects performed wrist flexion/extension, isometric contraction and co-contraction while electromyographic signals were recorded with surface electrodes. A moving data window of 200 values was applied to the data (moving approximate entropy). The results show that there is regularity in an EMG signal at the beginning and end of a muscle contraction with low regularity during the middle part.