Analysis of Signal-Averaged Electrocardiogram Performance for Body Surface Recordings

Abstract

To test the performance of signal averaging on body surface electrocardiograms (SAECG), a comparative analysis of four sources of perturbation, 1) uncorrelated noise, 2) beat alignment, 3) physiological variability and 4) respiratory movement was performed. The first two cases were assessed using a computer model of a ventricular beat. The other two cases were tested using high resolution body surface signals recorded from a torso tank (N=2) and patient data (N=4) respectively. In the first case, SAECG successfully removed a high level of noise made up of white Gaussian noise (WGN) with σ = 10 µV and 50 Hz noise with a signal to noise ratio (SNR) of 9 dB since the root mean square error of the noise (RMSE noise ) was 0.65 ± 0.01 µV and 1.30 ± 0.01 µV, respectively. The RMSE of the averaged QRS (RMSE SAQRS ) was slightly changed by physiological variability (RMSE SAQRS =4.18 ± 1.38 µV) when comparing the SA QRS resulting from the average of 100 different beats taken from the same recording. While SA QRS are distorted by respiration artefacts, the beats selected during the exhalation phase produced the least distortion to the SA QRS with a RMSE SAQRS = 16.28 ± 12.58 µV. To conclude, SAECG can efficiently de-noise signals in presence of uncorrelated noise without distorting the SA QRS . However, respiration motion introduces amplitude shift between SA QRS .