An automated technique for identification and analysis of activation fronts in a two-dimensional electrogram array.

Cardiac activation sequences are normally determined by (i) the detection and timing of local activations in cardiac electrograms, (ii) the grouping together of activations in different electrodes that are generated by the same activation fronts, and (iii) the construction by interpolation of isochronal maps showing the pathways of the activation fronts. This process is typically carried out by manual or semiautomated methods. These methods are usually adequate for stable, repeatable rhythms in normal hearts. However, in situations in which the electrograms are distorted, as in those recorded from abnormal myocardium, or the mapped rhythms are rapidly changing, as in ventricular fibrillation, they are tedious and time-consuming and yield results that are subjective and not repeatable from one investigator to another. Therefore, we developed a computer-based method for automating the identification and analysis of activation fronts recorded from a large array of electrodes. The electrodes are closely spaced (1 mm) so that interpolation is not required. Electrodes are identified as recording an activation when the temporal derivative of the potential is more negative than a user-specified value. Activations occurring less than a user-specified distance apart in time and space are identified as part of the same activation front. Characteristics of the activation fronts, such as their number, size, and the presence of reentry or collision, are then quantified. The differences between the results obtained by this automated method and those obtained by four human investigators was no greater than the differences in results among the four investigators themselves. Because the method is automated and algorithmic, it is both rapid and repeatable.