Detection of arrhythmias by different methods of neural network and Fuzzy algorithm has been investigated intensively in recent years. Electrical restitution curve (ERC) and some other ventricular electro-physiological properties which are discussing matter in prediction of cardiac arrhythmias in medical sciences were presented to an adaptive neuro-fuzzy inference system (ANFIS). These features were obtained from action potential signal which was recorded from isolated rabbit heart. This algorithm can detect steep part of this curve which is an indicator of ventricular fibrillation (VF) induction with high accuracy of 96.6%. Then in second step these features applied for detection of VF or stimulation block. The result shows that this curve can't predict inducibility of VF individually and it should be considered with other properties such as conduction velocity and memory effects.

Full text: IJEIR_860_Final.pdf