A model study of the sensitivity of body surface potential distribution to variations of electrode placement.

The effect of electrode displacement as one of the sources of reproducibility errors in body surface potential maps was studied using a realistic computer model of the cardiac electric field. A uniform dipole layer model of the cardiac generator and a realistic geometry of the torso, heart, and lungs was adopted for the simulation of surface potentials during ventricular activation. The effect of systematic electrode displacement in terms of longitudinal shifts and variations of longitudinal size (height) of the mapped area was studied. The map reconstruction error of three different limited lead systems and the variability of maps measured on all points of the mapping grid, as well as maps reconstructed from limited lead systems, were investigated and quantified. A mean relative error of map reconstruction of less than 3.5% was found for longitudinal shifts from -4.4 to +1.7 cm, and for longitudinal size changes from 65 to 108% of the initial area. For vertical displacements of electrodes between the limits of +/- 2.0 cm for full grid maps and +/- 1.4 cm if limited lead systems were used, the mean relative error of the maps remained under 5%.