Summary: The present study was undertaken to deter mine the apparent value for the volume of distribution of water to be used in the dynamic/integral technique for generating functional CBF images. A value of 0.86 re sulted in only a minor loss of accuracy compared to the more accurate (but time-consuming) dynamic only anal ysis, which incorporated the regionally fitted estimates of the volume of distribution of water. In contrast to the Recently, a new method has been described for the measurement of CBF using inhalation of C1502 and positron emission tomography (PET) (Lam mertsma et aI., 1990). This method involves two separate steps in generating reproducible functional CBF images, which are required in the analysis of activation or stimulation studies.First, the dynamic PET frames collected during and after inhalation of C1502 are analyzed using a nonlinear regression technique to obtain accurate estimates of delay and dispersion of the arterial blood curve (Lammertsma et al., 1989). Second, us ing these values of delay and dispersion, a pixel-by pixel integral analysis is then applied to the inhala tion (buildup) phase.The accuracy of functional CBF images is some what compromised by the need to assume, in the final integral analysis, a value for the volume of distribution of water (V d) in brain tissue relative to that in blood. It has been shown that the value of Vd