Removing ECG artifacts from the EMG: A comparison between combining empirical-mode decomposition and independent component analysis and other filtering methods

Abstract

Surface electromyography (EMG) is used for rehabilitation and clinical treatment for muscle disease. However, these recordings are often critically contaminated by cardiac artifact and many methods are applied to EMG in order to remove the artifacts from the EMG signals. We applied to both simulation and real EMG data a recently developed method of a combination of ensemble empirical mode decomposition and independent component analysis (EEMD+ICA), and compared its performance with that of other previously developed filtering methods. Relative root-mean-square errors (RRMSE) and correlations between the cleaned EMG and ECG contaminated EMG were calculated to evaluate the performance. The EMD based single channel technique showed better performance compared to the cubic smoothing spline and high-pass-filter (HPF) method for varied amplitude without a reference signal. Therefore, if the reference signal is not present, the combined EEMD and ICA procedure prove to be a reliable and efficient tool for removing ECG artifact from surface EMG.