Separation of the Electromyographic from the Electrocardiographic Signals and Vice Versa. A Topical Review of the Dynamic Procedure

Abstract

Electrocardiographic (ECG) and electromyographic (EMG) signals are inevitably and simultaneously recorded from the same electrodes and are respectively useful signal and noise in electrocardiography, and vice versa in electromyography. The frequency domains of the two signals overlap, making it difficult to filter the noise without distortion of the useful signal. An original ‘dynamic method’ for separation of the two signals was created. In a series of publications that began in 1999 with filtering of EMG noise from ECG signal, we have described the method and have made a number of improvements such as noise analysis and automatic on/off triggering in presence/absence of noise, online application, and tuning the parameters, to fulfill the last filtering recommendations of the American Heart Association. No matter if the Dynamic procedure is to be used in electrocardiography or in electromyography, the method contains the following: (i) Evaluation of the frequency bands of the ECG signal; (ii) filtering (suppression) of the EMG signal by dynamic change of the size of the filtering window for maximal preservation of the morphology of the ECG waves. The cutoff frequency is individual for any signal sample and varies from 13 Hz at the linear segments of the ECG signal, trough 25 Hz for the T-waves of high amplitude, and up to 400 Hz for the QRS-complexes; (iii) EMG signal separation by subtraction of the filtered ECG signal from ECG + EMG initial signal. With the current review, we are attempting to summarize all done over the years on the Dynamic procedure.