Test of an Inverse Electrocardiographic Solution Based on Accurately Determined Model Data

Encouraging progress has been made recently with a multiple-dipole type of inverse electrocardiographic procedure. A vitally important feature of the procedure is constraint of equivalent dipole moments to nonnegative values by means of quadratic programming. Application of the method to data derived from purely clinical sources has great heuristic value, but suffers from the lack of undisputed baseline information against which to compare the results of inverse solution. Accordingly, in this paper we attempt to evaluate the dependability of fixed-dipole array inversion by applying the method to a series of accurately modeled situations. Generally acceptable results are obtained with the dipoles arrayed in ``favorable'' configurations, with considerably less acceptable outcome likely under ``unfavorable'' conditions. Superposition of signal noise or perturbation of equivalent dipole locations tends to degrade the quality of solution, but does not entirely vitiate it. Study of one unfavorable array shows how a competitive situation might occur which could cause ``crosstalk'' between the interventricular septum and the free wall of the right ventricle. Despite the quantitative deficiencies which become apparent when multiple-dipole inversion is essayed with accurately modeled data, the procedure continues to offer promise as a useful means of extracting intrinsic electrophysiologic information from electrocardiographic signals.