Possible effects of regional wall motion abnormalities on apparent regional myocardial tracer concentrations on emission tomographic images were evaluated in six open chest dogs. Each dog was studied twice: In Run 1, 13N ammonia and microspheres were injected during a 6 min coronary occlusion, and serial images acquired by positron emission tomography during occlusion and reperfusion. In Run 2, 1 h later, 13N ammonia and microspheres were reinjected at control, and serial images recorded at control, during a repeat 6 min coronary occlusion, and after reperfusion. Segmental function was monitored with ultrasonic crystals, and 13N tissue concentrations determined in vivo from the tomographic images and postmortem by well counting. In Run 1, fractional shortening in ischemic segments fell by 89 +/- 16% SD from control. The ischemic versus control segment ratio for 13N activity concentrations averaged 0.29 +/- 0.08 and for microspheres 0.20 +/- 0.15. In Run 2 the ischemic versus control segment ratio was at control 0.77 +/- 0.12 for 13N tissue activity and 0.85 +/- 0.07 for microspheres. Fractional shortening fell during occlusion by 131 +/- 29% from control, returned to control early, and fell again by 11 +/- 16% late during reperfusion. These changes were paralleled by changes in apparent regional 13N tissue concentrations of the prelabeled myocardium. Compared with control, they were 37 +/- 9% lower during occlusion and rose to 94 +/- 20% early and to 89 +/- 16% at control late during reperfusion. In vitro determined tissue concentration ratios of ischemic to control myocardium were similar for 13N and microsphere activity (0.83 and 0.85), which ruled out loss of 13N ammonia from tissue during occlusion or reperfusion. Our results indicate that regional wall motion abnormalities cause artifactual segmental defects in tracer concentrations on emission tomographic images of the heart, which must be considered for qualitative and quantitative analysis of regional tracer tissue concentrations.