Overtraining and Reversal Discrimination Learning in Rats

Abstract

In a series of three experiments, the effect of overtraining on the reversal discrimination learning of rats was studied. Experiment 1 also investigated the effects of type of task and number of irrelevant dimensions, while type of task and reward magnitude were varied in Experiment 2, and number of original training trials and reward magnitude in Experiment 3. In the first two experiments, 100 overtraining trials were found to interfere with reversal learning, although not significantly in Experiment 1. Increasing the number of irrelevant dimensions retarded reversal learning as did small rewards and more difficult original discriminations. In Experiment 3, 200 overtraining trials were found to facilitate reversal learning, the magnitude of the effect falling just short of the .05 level of significance. In addition, reward magnitude and number of training trials were found to interact in a manner predicted by the incentive model of Theios and Blosser.