B. Hedén, M. Ohlsson, R. Rittner, O. Pahlm, W.K. Haisty Jr., C. Peterson and L. Edenbrandt
Agreement between Artificial Neural Networks and Human Expert for the Electrocardiographic Diagnosis of Healed Myocardial Infarction
Journal of the American College of Cardiology 28, 1012-1016 (1996)

An artificial neural network that is used as a classifier can, under certain circumstances, indicate a Bayesian probability for a correct classification. The purpose of this study was to compare probability estimates of myocardial infarction from a neural network to those of an experienced electrocardiographer. The study was based on 1664 electrocardiograms from 351 healthy volunteers and 1313 patients with a history of chest pain. The results show that a feed-forward neural network can be trained to diagnose healed anterior myocardial infarction at high levels of specificity and sensitivity, in good agreement with an expert electrocardiographer.

LU TP 95-09