The purpose of this study is to evaluate how well bipolar and intramural potentials are able to classify the state of the myocardium (normal or infarcted) proximal to the electrode recording sites. Nine mongrel dogs with anterior myocardial infarcts were used to generate a database for the study. Classification of the myocardium as normal or infarcted was attempted from potentials recorded in and around the infarcted region using electrodes within plunge needles. In addition to the potentials, the database contains the locations of the plunge needles, the grossly visible borders of the infarcts, and the epicardium and endocardium of both ventricles. Initially, the analysis of six dogs suggested that the peak amplitude and the maximum slope of the intramural QRS complex Were the empirical variables best suited for classifying the state of the myocardium. A two-variable, unbiased, cross validation technique was employed to obtain an unbiased, two-dimensional discriminant function that classified both normal and infarcted myocardium with 90 percent accuracy using bipolar recordings and 85 percent accuracy using unipolar recordings. In three additional dogs, histologic instead of gross examination was used to determine if the electrodes were in normal or infarcted myocardium. The same two variables were again combined in cross validation yielding 90 percent accuracy in classifying normal myocardium and 95 percent accuracy in infarcted myocardium using bipolar recordings. Copyright © 1985 by The Institute of Electrical and Electronics Engineers, Inc.