Experience in the development and application of a method for extracting cardiographic signals from noisy records is reported. The method, here called average transient computing (ATC), is based on modifications of well-known average response computing techniques. ECG complexes of several successive cardiac cycles are coherently summed so that random noise in the record tends to cancel itself, and the ratio of the ECG signal to the noise improves in proportion to the square root of the number of beats summed. The result is a "cleaned up" ECG suitable for clinical interpretation or quantitative measurement. Proper use of the increasingly available small devices for average transient computing requires knowledge and management of errors inherent in the process. Sophisticated triggering methods with anticipation and delay techniques are necessary to avoid averaged records which have little relevance to the original cardiographic function. A coincidence-logic pattern triggering method is described for the prevention of distortion of the averaged signal by artefacts, such as extrasystoles, and intentional smoothing of the ECG is carried out with electronic analog systems. It is emphasized that the averaged record obtained by this computing method is a limited statistic that serves as a starting point for empirical interpretation or for more detailed quantitative analysis of information contained in the exercise electrocardiogram.
Bibliographical noteFunding Information:
This study was supported in part by research grants from the Medical Research Council of Canada (MA 1441). the United States Public Health Service (H-03088 and HE-04697). and the Minnesota Heart Association. Received for publication June 17, 1964. *Department of Physiology and Biophysics, Dalhousie University, Halifax, Nova Scotia, Canada. **Laboratory of Physiological Hygiene, University of Minnesota School of Public Health, Stadium Gate 27, Minneapo-lis, Minn.. 5.5455.