Body Surface Potential Mapping Based on Cylindrical Regression

A new statistical method is proposed for evaluating the electric potentials of the human heart measured on the body surface (body surface potential mapping). The method is based on the representation of the measured values in terms of a cylindrical regression model, where the regression is set up horizontally by means of a trigonometric polynomial and vertically by means of an ordinary one. The special experimental equipment required for the optimal exploitation of the model is briefly described. The estimated potential distribution can be graphically displayed in three dimensions; the polynomial degree can be chosen in such a way as to resolve all statistical detail desired. The statistical optimality of the new approximation method is established. The method requires relatively little computational effort and, thus, is suitable for implementation on small computers.