The multiinjection approach was used to study in vivo interactions between alpha4beta2(*) nicotinic acetylcholine receptors and 2-[(18)F]fluoro-A-85380 in baboons. The ligand kinetics was modeled by the usual nonlinear compartment model composed of three compartments (arterial plasma, free and specifically bound ligand in tissue). Arterial blood samples were collected to generate a metabolite-corrected plasma input function. The experimental protocol, which consisted of three injections of labeled or unlabeled ligand, was aiming at identifying all parameters in one experiment. Various parameters, including B'(max) (the binding sites density) and K(d)V(R) (the apparent in vivo affinity of 2-[(18)F]fluoro-A-85380) could then be estimated in thalamus and in several receptor-poor regions. B'(max) estimate was 3.0+/-0.3 pmol/mL in thalamus, and ranged from 0.25 to 1.58 pmol/mL in extrathalamic regions. Although K(d)V(R) could be precisely estimated, the association and dissociation rate constants k(on)/V(R) and k(off) could not be identified separately. A second protocol was then used to estimate k(off) more precisely in the thalamus. Having estimated all model parameters, we performed simulations of 2-[(18)F]fluoro-A-85380 kinetics to test equilibrium hypotheses underlying simplified approaches. These showed that a pseudo-equilibrium is quickly reached between the free and bound compartments, a favorable situation to apply Logan graphical analysis. In contrast, the pseudo-equilibrium between the plasma and free compartments is only reached after several hours. The ratio of radioligand concentration in these two compartments then overestimates the true equilibrium value, an unfavorable situation to estimate distribution volumes from late images after a bolus injection.