Estimation of cardiac conduction velocities using small data sets

Cardiac conduction velocities can provide information about both the speed and angle of a propagating electrical wavefront. Catheter-based vector mapping may improve the visualization of cardiac arrhythmias, but catheters can only provide limited data. We address the problem of estimating conduction speed and angle from small data sets that are suitable for recording from a catheter device. We estimated cardiac conduction velocities using data subsets of 4-7 electrograms, and then compared the estimates to a larger reference grid. From six swine hearts, we studied 137 ventricular beats and 17,756 velocity vectors. Angle errors were 0.4/spl deg//spl plusmn/16/spl deg/ and speed errors were 5%/spl plusmn/33%. This initial data suggests that the signal processing required for catheter-based vector mapping is feasible.