Recently, neuroimaging procedures have been applied extensively to the acute-stage assessment of ischemic brain damage. Positron emission tomography studies of brain perfusion and oxygen metabolism performed within 18 hours of stroke onset both depict the amount of already irreversibly damaged tissue and accurately predict the subsequent neurologic course in a sizeable fraction of patients. The prognosis is poor in those patients with extensive irreversible damage and good when early spontaneous reperfusion has occurred before cortical damage has been established. The remaining patients, in whom ischemia is still ongoing and the neurological course is unpredictable, may constitute a subgroup more amenable to trials of therapy. New developments in diffusion-weighted magnetic resonance imaging suggest this noninvasive tool, like positron emission tomography, may be sensitive to failure of energy metabolism very soon after ischemia begins. Studies in rats suggest it may be possible to differentiate still salvageable tissue from irreversibly compromised tissue with this technique. Ultra-fast diffusion-weighted magnetic resonance imaging is feasible in humans and preliminary observations suggest the technique is as sensitive in humans as it is in rats. The usefulness in acute stroke management of other magnetic resonance techniques still in development, such as blood volume and perfusion imaging, deoxyhemoglobin-sensitive sequences, and H-1 or P-31 spectroscopy, remains to be assessed.