Accuracy of inferred heart potentials is limited by errors in the geometry of the torso model. To quantify effects of torso-model construction, we segmented images of the Visible Human Male to create full-body, torso- and heart-model surfaces. All body surfaces contained the same 163 electrode sites. Relative errors of forward-problem solutions compared to the full-body results at the electrodes were 0.024, 0.026, and 0.027 for models with cumulative removal of arms and legs, head and pelvis, and shoulders, respectively. Relative errors of inverse solutions near the peak of the QRS complex with 1% noise in body-surface potentials increased from 0.34 for the full body, to 0.40, 0.42, and 0.42 in the reduced-surface models. Results suggest that removal of arms and legs is the most critical step in model construction and that surface reduction may increase relative errors of inverse solutions by about 20%.

Keywords: inverse electrocardiology, epicardial potentials, relative error, Tikhonov regularization, torso models