The purpose of this investigation was to correlate magnetic resonance (MR) perfusion measurements with absolute regional cerebral blood flow (rCBF) in a rat model of focal ischemia. The MR perfusion measurements were made using dynamic first-pass bolus tracking of a susceptibility contrast agent, whereas rCBF was measured using radioactive microspheres. Two simple MR perfusion parameters, the maximum change in R2* (m delta R2*) and time delay to m delta R2* (t delta R2*), were derived from the signal intensity versus time curves on a pixel-to-pixel basis, without applying curve-fitting procedures or tracer kinetic theory. In each hemisphere, m delta R2* and t delta R2* were compared with the rCBF measurements in four selected regions of interest. Sixteen MR bolus tracking series were performed in 12 rats with occlusion of the middle cerebral artery. In all of the individual series there was a significant correlation (.0001 < or = p < or = .02) between m delta R2* and the microsphere rCBF measurements, with correlation coefficients ranging from .784 to .983. Pooling the m delta R2* data resulted in a correlation coefficient of .809 (p = .0001). There was a nonlinear correlation between the t delta R2* and rCBF. For both parameters there was considerable variation between different measurements regarding both the slope of the regression line and its intercept with the y-axis. Our results justify the use of m delta R2* as a relative measure of perfusion during acute cerebral ischemia. Because of the interindividual variation, calibration of MR perfusion measurements for the estimation of absolute flow values must be considered unreliable. The t delta R2* may have physiological relevance as a marker of collateral flow.