Experimental models of focal cerebral ischemia have provided important data on early circulatory and biochemical changes, but typically their correspondence with metabolic and hemodynamic findings in stroke patients has been poor. To fill the gap between experimental studies at early time points and rather late clinical studies, we repeatedly measured CBF, CMRO2, oxygen extraction fraction (OEF), cerebral blood volume (CBV), and CMRglc in six cats before and up to 24 h after permanent middle cerebral artery (MCA) occlusion (MCAO), using the 15O steady state and [18F]fluorodeoxy-glucose methods and a high-resolution positron emission tomography (PET) scanner. Likewise, three sham-operated control cats were studied during the same period. Final infarct size was determined on serial histologic sections. In the areas of final glucose metabolic depression that were slightly larger than the histologic infarcts, mean CBF dropped to approximately 40% of control values immediately on arterial occlusion. If further decreased to < 20% during the course of the experiment. This progressive ischemia was most conspicuous in border zones. CMRO2 fell to a lesser degree (55%), eventually reaching approximately 25% of its control level. At early stages, OEF increased mainly in the center of ischemia. With time, areas of increased OEF moved from the center to the periphery of the MCA territory. Concurrently, progressive secondary decreases in OEF in conjunction with further reductions of CBF and CMRO2 indicated the development of central necrosis. The findings are highly suggestive of a dynamic penumbra. In five cats with complete MCA infarcts, CBF decreased and OEF increased in the contralateral hemisphere after 24 h, suggesting whole-brain damage. This effect may be explained by the widespread brain edema found histologically in addition to the nonspecific CBF reductions and OEF elevations observed also in the sham-operated controls after 1 day in the experimental condition. In one cat, cortical OEF increased only transiently. Normal CMRO2 and CMRglc were eventually restored, and the final infarct was small. This study demonstrates that acute regional pathophysiologic changes can be repeatedly assessed by multivariate PET in cats. Viable tissue can be detected up to several hours after MCA occlusion, and the transition of misery-perfused regions into necrosis or preserved tissue can be followed over time. The present results support the concept of a dynamic penumbra, in which for up to 24 h tissue damage spreads progressively from the center to the periphery of ischemia. Sequential high-resolution PET provides insight into the dynamics of regional pathophysiology and may thus further the development of rational therapeutic strategies.