Comments on "Is accurate recording of the ECG surface Laplacian feasible?" [with reply]

The authors comment on the work of D.B. Geselowitz and J. Ferrara (ibid., vol. 47, p. 1138, 2000) who concluded that the signal-to-noise ratio (SNR) of the ECG surface Laplacian (SL) is very low. Averaged over all recording channels on the anterolateral chest and averaged over all six subjects, when the noise level is estimated during the window of the P-R segment, our experimental results indicated that the potential ECG, regular Laplacian ECG, and diagonal Laplacian ECG have SNR of 105.70/spl plusmn/22.49, 33.73/spl plusmn/9.86, and 46.83/spl plusmn/12.84, respectively. This is in contrast with the magnitude of SNR reported in the text of the above paper, which was too low becaase of their choice of a residue window containing noise as well as signal. Geselowitz and Ferrara reply to these comments.