Studies have demonstrated that patients with Q-wave infarctions on the electrocardiogram (ECG) frequently have nontransmural scar formation, whereas non-Q-wave infarctions may have transmural scars. The precise pathophysiologic substrate that underlies Q waves remains unclear. Magnetic resonance imaging (MRI) is the preferred technique to evaluate patients who have infarction because information can be obtained on function, contractile reserve (viability), and scar tissue. Consecutive patients (n = 69) who had coronary artery disease and a history of myocardial infarction underwent MRI; the protocol included MRI at rest, small-dose dobutamine MRI, and contrast-enhanced MRI. Parameters included left ventricular ejection fraction, left ventricular volumes, end-diastolic wall thickness and contractile reserve in the infarct region, transmurality and spatial extent of scar tissue, total scar score, and quantified percent left ventricular scar tissue. MRI data were related to the presence/absence of Q waves on the ECG. Q waves were present in 39 patients (57%). Univariate analysis identified transmurality, spatial extent, total scar score, and quantified percent scar tissue as predictors of Q waves. Multivariate analysis demonstrated that quantified percent scar tissue was the single best predictor of Q waves on the ECG. A cut-off value of 17% infarcted tissue of the left ventricle yielded a sensitivity and specificity of 90% to predict the presence/absence of Q waves. When quantified percent scar tissue was removed from the model, spatial extent of infarction was the best predictor. Thus, Q waves on the ECG correlate best with quantified percent scar tissue on contrast-enhanced MR images.