Training dose as a decisive factor for discrimination of a drug mixture in rats

The impact of training dose on the characteristics of a discrimination maintained by a mixture of two dissimilar drugs has been investigated in order to refine this approach to the study of drug interactions. Three groups of rats (n = 10) were trained to discriminate mixtures of (+)-amphetamine (0.2-0.8mg/kg) plus pentobarbitone (5-20mg/kg) from saline, in a two-lever operant procedure with food reinforcement, with the ratio of the doses held constant (amphetamine: pentobarbitone, 1:25). Discriminations were acquired to an accuracy of 90-97%. There was full generalisation to amphetamine alone, but only in rats trained with mixtures of the smaller doses of the single drugs. There was partial generalisation when either apomorphine (50%) or nicotine (63%) was administered alone, and the magnitude of these responses was inversely related to the dose of mixture used for training. Doses of pentobarbitone half of those used for training produced little discriminative response when administered alone to rats trained with the two smallest doses of the mixture; the same doses of pentobarbitone increased responses to amphetamine or apomorphine in a more than additive manner. Strikingly, some doses of apomorphine and pentobarbitone that did not generalise when administered separately, produced full generalisation when administered together, but only in rats trained with the smaller doses of the mixture. In contrast, pentobarbitone did not enhance generalisation to nicotine in any group. It was concluded that, on the one hand, patterns of generalisation to single drugs followed an orderly pattern resembling those for discriminations established with single drugs. On the other hand, there was a complex pattern of generalization from one mixture to another; thus, altering the doses of drugs used for training markedly influenced discriminations of an abused drug mixture, but no simple rules to predict the influence of training dose have been ascertained.