Threshold based reduction of EMS stimulation artifacts in the electromyogram when stimulation intensity increases

To investigate muscular strain, it is possible to record an electromyogram (EMG) during electromyostimulation (EMS). However, stimulation artifacts make it difficult to examine evoked compound muscle action potentials (M-waves). The aim of this work was to algorithmically remove stimulation artifacts from the EMG signal when the stimulation intensity increases. For this purpose, EMG signals were recorded on four subjects who underwent a 30-second EMS of the right M. triceps surae. Afterwards we conducted offline signal processing to reduce stimulation artifacts. We implemented a two-stage threshold algorithm, adapted from O'Keeffe et al. (2001), to remove signal segments passing the thresholds in a defined sequence. Here, the thresholds in the first iteration are two constants, while the algorithm uses linearly or exponentially increasing thresholds in the second iteration. After empirically adjusting the threshold parameters, the stimulation artifacts were successfully reduced. This allows further investigations of the M-waves with respect to muscular fatigue.