We have validated ECG-gated emission tomography using technetium-99m methoxyisobutylisonitrile for the assessment of regional ventricular function by comparing it with cine magnetic resonance imaging (MRI). Gated tomography was performed at rest in 24 patients referred for myocardial perfusion imaging [17 males and seven females with a mean age of 58 years, nine of whom had had a previous myocardial infarction (MI)]. Scores were assigned to each of nine myocardial segments for wall motion and for thickening. Cine MRI was analysed in an identical fashion. Four out of 216 (2%) segments were uninterpretable by gated tomography because of inadequate tracer uptake. In eight patients without coronary artery disease (CAD), wall motion and thickening were normal by both methods. Gated tomography showed abnormal wall motion or thickening in all patients with previous MI and in five of seven patients with CAD but no prior MI. Association between wall motion and thickening was good (rs=0. 86). Overall, there was good agreement between gated tomography and MRI for both wall motion (178/212 segments, kappa=0.66) and wall thickening (184/212 segments, kappa=0.69). In segments with severely reduced perfusion, however, there was poorer agreement (kappa=0.31). Interobserver and intraobserver agreement was high (kappa from 0.61 to 0.78). Thus, in patients investigated for CAD, there is good overall agreement between gated tomography and MRI but the agreement is lower in segments with severe perfusion defects.