To evaluate the feasibility of detecting denervated myocardium in the infarcted canine heart, the distribution of sympathetic nerve endings using I-123 metaiodobenzylguanidine (MIBG) was compared with the distribution of perfusion using thallium-201, with the aid of color-coded computer functional map in 16 dogs. Twelve dogs underwent myocardial infarction by injection of vinyl latex into the left anterior descending coronary artery (transmural myocardial infarction, n = 6), or ligation of the left anterior descending coronary artery (nontransmural myocardial infarction, n = 6). Four dogs served as sham-operated controls. Image patterns were compared with tissue norepinephrine content and with histofluorescence microscopic findings in biopsy specimens. Hearts with transmural infarction showed zones of absent MIBG and thallium, indicating scar. Adjacent and distal regions showed reduced MIBG but normal thallium uptake, indicating viable but denervated myocardium. Denervation distal to infarction was confirmed by reduced norepinephrine content and absence of nerve fluorescence. Nontransmural myocardial infarction showed zones of wall thinning with decreased thallium uptake and a greater reduction or absence of MIBG localized to the region of the infarct, with minimal extension of denervation beyond the infarct. Norepinephrine content was significantly reduced in the infarct zone, and nerve fluorescence was absent. These findings suggest that 1) MIBG imaging can detect viable and perfused but denervated myocardium after infarction; and 2) as opposed to the distal denervation produced by transmural infarction, nontransmural infarction may lead to regional ischemic damage of sympathetic nerves, but may spare subepicardial nerve trunks that course through the region of infarction to provide a source of innervation to distal areas of myocardium.