Increasing knowledge of the genome sequences of several organisms and the development of genome-wide, high-throughput screening techniques for gene expression are likely to generate a vast amount of data aimed at elucidating the molecular mechanisms of addiction. These findings are likely to have potential for future addiction pharmacotherapies. However, it is important to employ animal models that dissociate the molecular and cellular consequences of the direct pharmacological effects of addictive drugs from those that result from the cognitive processes associated with self-administration of these drugs. In this article, we suggest that the short-term and long-term neuroadaptive effects of addictive drugs in the brain depend crucially on the drug-exposure paradigm used [i.e. passive (non-contingent) drug exposure and active (contingent) self-administration]. This has important ramifications for future molecular and cellular studies of drug addiction.