Methods for measuring rCBF with PET are based upon a one-compartment model which assumes that flow is constant during the scan period. This assumption may not hold in several experimental situations, however, such as after a pharmacologic intervention, during and after a seizure, if there is habituation during activation studies, or if arterial pC02 varies. We used simulations to determine the flow values that would be obtained in practice when the model is used under such non-steady state conditions. Conditions simulated included a step change in flow and linear or exponential drifts. Calculations were performed for two widely-used flow methods, the autoradiographic method and a rapid least squares technique. Simulations showed that when flow changes, the calculated rCBF reflects the flow at the start of the scan, rather than the mean flow during the scanning period. For example, with a step change in rCBF of plus/minus50% at 20 sec into the scan, estimated rCBF is within plus/minus4-6% of baseline flow for both flow methods. Prior to manuscript preparation, further simulations are being performed for semiquantitative 0-15 water scans, when tissue counts alone are used to map rCBF changes. These findings will help in the interpretation of rCBF data, and will assist in the design of activation experiments, because of the demonstration that the early scan period determines the observed rCBF changes.
