Is accurate recording of the ECG surface Laplacian feasible?

Experimental and model studies were performed to measure the electrocardiographic surface Laplacian using a rectangular finite difference approximation. The experimental approach used ten normal subjects with two sites on the torso. Electrode spacing was 2 cm. The surface Laplacian is theoretically independent of rotation of the electrode array. The data showed considerable variation with rotation. Model studies employed a realistic 23-dipole source. A spherical volume conductor showed invariance with rotation, as anticipated theoretically. A realistic torso, however, showed variation with rotation, although not as severe as that measured. A separate experimental study considered the signal-to-noise ratio (SNR) for the surface Laplacian. The average SNR was 3.3 and 2.5 at the two sites. These results raise serious questions about the practical ability to measure the surface Laplacian on the torso.