Effect of noise sources on the averaged PQRST morphology

We analysed noise interference on the PQRST waveform of the electro- and magnetocardiogram as recovered by an average technique. Three signals (adult A, neonatal N, and fetal F) were corrupted with breathing, electromyographic and power-line noise at different levels of the QRS amplitude. Using the correlation coefficient r, averaged PQRST patterns from corrupted signals were compared with corresponding uncorrupted templates. Lower r values were obtained at higher noise levels; this effect was more critical for the F signal. Noise sources caused different distortion; e.g. at 0.8 level and 50 averaged beats, r remained high for breathing (0.996 A, 0.995 N, 0.972 F); lower for electromyography (0.990 A, 0.972 N, 0.905 F), and even lower for power-line (0.931 A , 0.852 N, 0.574 F). Thus, to perform morphological analysis of the PQRST waveform the source, noise level and signal´s nature should be considered for defining an optimal number of beats to be averaged