Equilibrium models are derived and applied to in vivo binding of spiperone in the rat brain. The models express the concentration of the ligand in the striatum and frontal cortex as a function of the accumulation in the cerebellum. The models differ with respect to the description of specific binding. Nonlinear regression analysis shows that the in vivo specific binding of 3H-labeled spiperone in the frontal cortex (mainly serotonergic) can be described by a noninteracting sites model, whereas the specific binding in the striatum (mainly dopaminergic) can best be described by models that lead to sigmoid saturation curves. These results were tested and partly confirmed by determining the region-of-interest/cerebellar radioactivity ratio of 11C-labeled N-methylspiperone, with and without pretreatment with haloperidol. The estimated Bmax was 32 fmol/mg wet tissue in the frontal cortex and approximately 90 fmol/mg wet tissue in the striatum. The free plus nonspecific binding of spiperone was similar in the frontal cortex but lower in the striatum than in the cerebellum. The occurrence of sigmoidicity can be best explained by the existence of high-affinity/low-capacity sites in the cerebellum rather than mutual interactions of striatal sites. The consequence of the present analysis for positron emission tomography is that the striatal/cerebellar activity ratio is not an accurate parameter of specific binding features at tracer doses of spiperone or N-methylspiperone.