Optimal training parameters in the two-bar fixed-ratio drug discrimination task.

In a two-lever compartment, thirsty rats were trained to press one bar when drugged with phenobarbital and the other bar when undrugged using water as a reinforcer. Several different training procedures were employed in order to compare their effects on speed of acquisition and/or asymptotic accuracy of discrimination. Results were as follows: (1) Some shaping procedures allowed more rapid acquisition of discriminative control than others. The "traditional" shaping procedure was significantly less efficient than any others tested. (2) Several indices of degree of discriminability based on the speed of acquisition of discriminations were compared and evaluated. Some varied linearly with 1n dosage. (3) Variations in session duration from 5 to 60 minutes did not alter asymptotic accuracy. (4) Fixed ratio sizes ranging from FR-3 to FR-30 resulted in similar asymptotic accuracies. Overall, the results define alterations in the fixed-ratio training procedure that will make it somewhat easier to use. However, no procedures were found that fundamentally improved the properties of the paradigm.