The purpose of the present study was to evaluate the performance of various parametric reference tissue models for quantification of [11C]PIB studies. Several models with and without fixing the reference tissue efflux rate constant (k'(2)) were investigated using both simulations and clinical data. The following parametric methods were evaluated: receptor parametric mapping (basis function implementation of the simplified reference tissue model with and without fixed k'(2)), reference Logan, and several multi-linear reference tissue methods (again with and without fixed k'(2)). In addition, standardised uptake value ratios with cerebellum (SUV(r)) were evaluated. Simulations were used to assess the effects of variation in flow (R(1)), fractional blood volume (V(b)) and binding potential (BP(ND)) itself on precision and accuracy of parametric BP(ND). For clinical studies, most parametric methods showed comparable performance, with poorest results for SUV(r). Best performance was obtained for receptor parametric mapping (RPM2) and one of the multi-linear reference tissue models (MRTM2), both with fixed k'(2): BP(ND) outcome was less affected by noise and the images showed better contrast than other tested methods. RPM2 and MRTM2 also provided best accuracy and precision in the simulation studies and are therefore the methods of choice for parametric analysis of clinical [11C]PIB studies.