This paper discusses the basic principles of drug/P-glycoprotein (P-gp) interaction, focusing on the methodology and design of positron emission tomography (PET) studies investigating P-gp function. The requirements of a good PET P-gp radiotracer are also evaluated. (R)-[(11)C]verapamil is used as an example, as this drug is the most common tracer for P-gp studies, but [(11)C]loperamide, [(11)C]desmethyl-loperamide and other compounds are also mentioned. The article also discusses the various study designs that can be used for PET drug disposition studies, such as administration of the inhibitor before or after the radiolabeled drug (tracer) and the use of bolus injections or infusions. Concepts such as the unbound partition coefficient (K(p,uu)) and the volume of distribution of unbound drug in brain (V(u,brain)), which are not easily measured directly with PET, can be used to describe the impact of protein binding and non-specific binding on drug distribution in brain tissue. It is concluded that new imaging probes will be required if the role of PET in studies of the interactions of drugs with efflux transporters is to expand.