Evaluation of a Double Threshold Algorithm to Detect Electromyographic Activity in the Healthy and Paretic Tibialis Anterior Muscle

摘要

In this study, the performance of a double-threshold algorithm for detecting electromyographic activity (EMG) in healthy and paretic Tibialis Anterior muscle was evaluated. Detection of EMG activity in this muscle is of interest since it could be used to control a neuroprosthesis to correct foot drop. The performance of the algorithm was assessed by means of the sensitivity and specificity in detecting the interval of EMG activity and the temporal error in detecting the beginning of the EMG activity. Indications by experts were used as standard reference. The results showed that changing the value of one of the thresholds modified significantly the performance of the algorithm. Furthermore, it was possible to achieve a specificity of 0,97, a temporal error lower than 13 ms, and a sensitivity of 0,8 for low values of the threshold. The present results suggest a future application to detect EMG activity, which could be used to control a neuroprosthesis to correct foot drop.