Materials and methods: Six male subjects (age: 33±11) completed [11C]OMAR PET studies at baseline and 5 h after one single oral dose of AVE1625 at 4 dosing levels (5–60 mg) with arterial blood sampled. Total distribution volume of baseline (VT0) and post-treatment (VT1) scans were obtained for 14 brain regions using plasma reference graphical analysis (PRGA; Logan et al., 1990). Because of the lack of a reference region for CB1 receptors, non-displaceable distribution volume (VND) was obtained as x-intercept of plot of VT0 minus VT1 (=y) versus VT0 in individual subjects. Receptor occupancy (OR) was calculated for regions as follows: OR=(VT0−VT1)/ (VT1− VTND). Plots of OR (=y) versus plasma drug concentration (Cp) were fitted with the Langmuir equation: OR=Omax·Cp/(ED50+Cp), where Omax and ED50 are maximal occupancy and the Cp that caused 50% of Omax. Mean absolute deviation (MAD) between observed and modelpredicted data were calculate as an index of goodness of fit. Simulated PET data were generated to estimate accuracies of Omax and ED50 estimates given by the Langmuir equation in relation to accuracy of BPND which is not established for [11C]OMAR, although it can be calculated as BPND=VT /VND−1. Conditions for the simulation study were as follows: True Omax: 80–100%; True ED50: 20 ng/ml; biological inter-subject variations: 5% for OR and 7% for BPND; accuracy of BPND estimates: 3–15%; observed Cp; and 1000 realizations per a set of input values.