Comparative psychometric analysis of vector and isochrone cardiac activation maps

Isochronal cardiac activation maps can be constructed from local activation times associated with spatial locations, and are frequently used to study cardiac arrhythmias. Cardiac velocity vector mapping has been proposed as an alternative method to study cardiac activation in both clinical and research environments. Velocity vectors inherently contain more information than scalar measures of latency, but it is unknown how vector maps and isochronal maps compare when they are used to identify patterns and features associated with arrhythmias. In order to quantitatively compare these two visualization methods, eight cardiologists were asked to complete forced-choice tasks in which they selected ablation sites based on synthetic vector or isochronal maps. Maps varied in arrhythmia complexity, number of vectors or activation times included, and errors in magnitude or angle for maps of velocity vectors. Quantitative comparison was achieved by using psychometric functions to characterize the learning curve and the total number of measurements needed in order to choose a correct ablation site. For simple arrhythmias, performance with vector maps was superior to isochronal maps. Subjects required fewer measurements, and learned more rapidly by studying vector maps. For more complex arrhythmias, there was no significant difference in performance between vector and isochronal maps. However, arrhythmia features were clearer with vector maps, even though this clarity did not necessarily change the ablation site choice. When errors were added to vector maps, performance was satisfactory for angle errors < 55 degrees, and speed errors did not affect performance.