Development of an algorithm for detection of fatal cardiac arrhythmia for implantable cardioverter-defibrillator using a self-organizing map

In this study, we have introduced the pattern classifier using the self-organizing map (SOM) for detecting fatal cardiac arrhythmia in implantable cardioverter-defibrillators (ICDs). The SOM has learned patterns of sinus rhythm, ventricular fibrillation and ventricular tachycardia with the feature vectors extracted from electrocardiogram and right ventricular volume measured during an arrhythmia induction experiment of a dog. After learning, neurons of the SOM were labeled by using the k-Nearest Neighbor method. It was shown that the accuracy of the proposed method was higher than other competitive methods applied to the same test data