Adaptive logic network compared with backpropagation network in automated detection of ischemia from resting ECG

摘要

The aim of this study was to introduce an adaptive logic network (ALN) computing method for detection of myocardial ischemia in standard 12-lead resting electrocardiogram (ECG) and to compare this method with the commonly used backpropagation network method and linear discriminant analysis techniques. Of 1,367 study subjects, 733 were women and 634 were men aged 65 and above. Ischemic episodes were identified by 24-hour ambulatory Holter ECG recordings. Accuracy was 67% using ALN, 56% using backpropagation and 65% using discriminant analysis. ALN offers a slightly higher accuracy and shows several advantages over the other two.