A taste aversion model of drug discrimination learning: training drug and condition influence rate of learning, sensitivity and drug specificity.

A model of drug discrimination based on a lithium chloride (LiCl) flavour aversion was described and examined. Mildly thirsty rats were presented daily with 4 ml of a distinctly flavoured solution which was followed on 50% of the days by an IP injection of LiCl. Prior to the flavour presentation, the rats were injection SC with saline or a training drug (0.04 mg/kg fentanyl or 20 mg/kg pentobarbital) to signal whether LiCl would follow. Almost all rats eventually exhibited stable behaviour that involved drinking most or all of the fluid when it was not to be followed by LiCl and little or no drinking when it was. Such discrimination occurred regardless of whether drug predicted LiCl (learned-discomfort) or predicted no LiCl (learned-safety). However, with fentanyl there were clear differences between rats trained with drug under learned-safety and under learned-discomfort conditions for 1) the rate of acquisition of stable performance as a function of LiCl dose, 2) generalization of the training dose to a test dose that was lower, and 3) elicitation of fentanyl responses by pentobarbital. These findings, together with indications that such effects did not always occur with pentobarbital as the training drug, were discussed from theoretical and practical perspectives.