Dipolar electrocardiotopographic imaging of instantaneous vectorcardiographic QRS vectors represents the ventricular activation front as a bounded area on the spherical image surface. Fuzzy set treatment of the elements of this area allows for the recognition of five classes of ventricular activation: I, abnormal; II, abnormal with normal component; III, normal with abnormal component; IV, marginally normal; and V, normal. In a group of 71 patients with chest pain and various degrees of coronary artery stenosis classes II, III, and IV were most frequently observed in patients with one- or two-vessel disease. The frequency distributions of classes II and III throughout the QRS were bimodal with peaks at 20 and 60 ms after QRS onset, while the frequency distribution of class IV was unimodal with a peak at 30 ms after QRS onset. The advantage of this method is the smoothing of threshold values on a quantitative basis.