Identification of cardiac arrhythmias using a damped exponential-predominant frequency algorithm

The authors describe a new approach for the discrimination of ventricular fibrillation (VF) from ventricular tachycardia (VT) based on a damped exponential (DE) modeling algorithm. Classification is achieved by simple linear separation applied to a novel feature, dubbed predominant frequency (PF), derived from the DE model. Tests conducted using 70 episodes drawn from the MIT-BIH malignant arrhythmia database produced total predictive accuracy exceeding 98%.