Frank lead vectorcardiograms (VCG) from four carefully selected patient subgroups (226 patients) were analyzed to develop optimal criteria for the diagnosis of anterior myocardial infarction. Specificity was evaluated using 100 healthy volunteers under age 30 and 80 patients with normal left ventriculogram and normal coronary arteriograms. Sensitivity was determined using 25 patients with evolutionary ST-T wave changes (V1-2), and LDH and CPK isoenzyme evidence of acute myocardial infarction; and 21 patients with anterior wall akinesia or dyskinesia and greater than 70% occlusion of the left anterior descending coronary artery. Patients with VCG evidence of bundle branch block, left or right ventricular hypertrophy were excluded. The criterion for the diagnosis of anterior myocardial infarction which was found to give the highest sensitivity with greater than or equal to 95% specificity was: initial anterior QRS forces must not exceed 0.1 mV in maximal anterior amplitude and also must not exceed 24 msec in duration. The performance of this proposed criterion was then tested using four similarly defined patient subgroups consisting of a total of 222 patients. The incidence of false positive diagnosis in these test subgroups was less than 1% with a sensitivity of greater than 95%. The overall performance of the proposed criterion was found to be significantly superior to both the widely accepted VCG and ECG criteria for anterior myocardial infarction. Thus, this quantitative criterion using both time and duration of initial anterior forces is both a highly specific and a sensitive indicator of anterior myocardial infarction.