Rationale: If a psychoactive drug shares discriminative effects with one that maintains self-administration, it is often inferred that the test drug is likely to be self-administered and to have abuse liability. This presumed predictive relationship has not been studied directly, however.
Objective: To determine at the level of the individual subject (1) whether a novel drug dose that shares discriminative effects with a reinforcing drug dose also will serve as a reinforcer, and (2) whether the results of generalization tests for drugs pharmacologically similar to the training drug predict whether the test drugs will or will not be self-administered.
Methods: Baboons were trained to discriminate midazolam (0.32 mg/kg, IV) from saline and also under a schedule of IV drug reinforcement. At the beginning of a period of self-administration, the first self-injection was followed 10 min later by a drug discrimination test session. The baboon then had the opportunity to self-administer the same dose 24 h/day (3-h timeout after each injection). A second drug discrimination test followed the last self-injection of the condition.
Results: Zolpidem and imidazenil shared discriminative effects with midazolam. Zolpidem was reinforcing in all baboons, but imidazenil was not. Chlordiazepoxide partially shared discriminative effects with midazolam, and the rate of self-administration was low. Pentobarbital did not share discriminative effects with midazolam, but was reinforcing. For all drugs, some doses did not share discriminative effects with midazolam but were reinforcing. Generalization gradients from tests after the last self-injection were similar to those after the first self-injection.
Conclusions: The discriminative effect of a drug in relation to a training drug of the same pharmacological class is not isomorphic with its reinforcing effectiveness.