Selection of the Number and Positions of Measuring Locations for Electrocardiography

Abstract
This paper considers how many and which locations on the body surface must be measured (by taking ECGs at these positions) to be able to determine consistently the total-body QRS surface-potential distribution as it varies in time. The answers to these questions have implications about the complexity of models of heart electrical activity in addition to their experimental value. An advantage of using the ability to compute the total-body potential distribution as a criterion of quality is that untestable assumptions about the nature of heart electrical activity are avoided. The accuracy of computed potential distributions with respect to corresponding experimental ones is specified by the mean-square difference between them. Acceptable maps had an average relative mean square error of less than 4 percent in the presence of about 1 percent noise, since inspection of the surface maps showed this to be the maximum error allowable for the same clinical or physiological interpretation of the surface maps analyzed.