A neural network to differentiate wide QRS tachycardias

The development of a neural network designed to perform the differentiation of wide QRS tachycardias in superventricular tachycardia (SVT) vs. ventricular tachycardia (VT) is described. Artificial neural networks are formed by a large number of simulated artificial neurons, interconnected much like the brain´s neurons. Rather than following programmed rules like in classical programs or expert systems, they learn by example. In this way a causal relation between input and diagnostic statement does not have to be known. An implementation example is given to show that the system is at least as sensitive and specific in differentiating SVT from VT as the manually defined rules