Arrhythmia diagnosis with discrimination of rhythm origin and measurement of heart-rate variation

Abstract

The authors developed an arrhythmia diagnosis system which distinguishes supra-ventricular events from ventricular ones. The system uses a power spectrum of QRS complex to discriminate the origin of a rhythm, and then classifies 12 types of arrhythmias based on the heart-rate end its variations. The system was evaluated with human ECG data. Two kinds of origin, supra-ventricular and ventricular, were distinguished with more than 98% sensitivity and specificity. The discrimination was conducted with a simple processing of weighted summation of two power spectrum inputs (4 and 8 Hz). In arrhythmia classification, there were 14 misclassified beats among 2234 beats, but no clinically fatal errors were found.