Evaluation of Multipolar Effects in the High-fidelity Standard Electrocardiogram by Means of Factor Analysis.

MANY STUDIES have rather clearly demonstrated the inadequacy of the simple dipolar concept to account for the total elec-trocardiographic information available on the body surface.'-6 For example, Okada, Langner, and Briller have attempted extraction of the dipolar component from the precordial leads by cancellation with synthetic dipolar leads derived from vectorcardiograms. An unac-counted-for residual was found.3 By means of closely spaced multiple lead recordings, subsequent workers have constructed isopo-tential maps demonstrating, not only dipolar pattern, but islands of apparent multipolar activity as well.1 2, 4, 5 Factor analysis provides still another means by which multipolar contribution may be detected .5 ,7 8 Mathematically, factor analysis compares each of a population of waveforms with every other waveform in the group, retains the several bits of unique information, and casts out the redundant information as it reappears in lead after lead. Previously, when this process of factor analysis was applied to total-surface electrocardiograms recorded from 153 to 160 closely spaced sites completely encompassing the thorax, at least seven or eight unique waveforms or principal factors were found. In brief, all of the nonredundant information present in the many QRS complexes could be expressed concisely in the form of seven or eight completely different waveforms.5 Scher has indicated that finding not more than three principal factors when this process is applied to a population of QRS complexes is compatible with a dipolar generator. Such compatiblity, however, is not proof of a dipolar generator, since a number of three-function generators other than a single dipole is possible .8 In addition, the emergence of more than three principal factors is not even compatible with a single dipole and immediately demands a more complex generator for explanation. One knows at the outset that when the process of factor analysis is applied to the six frontal plane leads of the standard electrocar-two significant factors or waveforms should emerge. This follows from the fact that we are dealing with only three sampling points, the right arm, the left arm, and the left leg, which describe a plane and from the information in any two of these leads the information in a third may be predicted. In the contrasting case of the multiple-lead total-surface electro-cardiogram obtained from man, it was necessary to employ eight principal factors of unique waveform to reconstruct the original QRS complexes. In this instance when the information in only the first three factors was employed …