Further improvement of classical criteria for differentiation of wide-QRS tachycardia in SVT and VT using artificial neural network techniques

A number of criteria for differentiation of wide-QRS tachycardias have been published. Recently criteria were published based on a sequential analysis of four parameters (P. Brugada et al., Circulation, vol. 83, no. 5, p. 1649-59, 1991). Based on the same ECGs as used to derive these rules, an artificial neural network was trained, and tested using a separate test set. With the help of a model, based on this network, the parameters used in the criteria mentioned above could be improved. These improvements were validated using an automatic learning system based on inductive methods.<
>