Study on detection algorithm of fatal arrhythmia using multiple regression analysis

The implantable cardioverter-defibrillator (ICD) is an effective therapeutic device for rescuing patients with cardiac diseases from death caused by life-threatening arrhythmias. For development of the ICD, it is important to accurately distinguish among normal sinus rhythm, ventricular tachycardia (VT), ventricular fibrillation, and supraventricular tachycardia (SVT). Thus, in this study, we have proposed a multiple regression model based on 14 indices extracted from two-dimensional statistics of intracardiac electrocardiograms to detect four kinds of cardiac rhythms as accurately and quickly as possible. The experimental results showed that the proposed method had a sensitivity of 0.97 for detecting SVT and a specificity of 0.99 for detecting VT, which were improved respectively from 0.83 and 0.85 obtained from the previous method, and that early detection within about 1.6 seconds was attained.