Abstract
Operant renewal refers to the relapse of eliminated behavior as a function of context change. Two experiments were performed to assess operant renewal effects in rat subjects with different response location criteria. In both experiments, subjects were trained to respond to a variable-interval schedule on an array of five levers. For Experiment 1, in Context A, only responses on the middle lever were reinforced. For these subjects, response distributions formed a generalization gradient around the middle lever. For Experiment 2, responses on any lever could meet the reinforcement criteria. In this case, the response distribution across levers varied between subjects but was consistent within subjects across time. During both experiments, responding was next extinguished in Context B, which increased variability in response distributions. The reintroduction of Context A decreased variability levels. For most rats in Experiment 1 and some rats in Experiment 2, the pattern of responding during testing was more similar to the pattern during training than to the pattern at the end of extinction. The results suggest that renewed response distributions across locations varied as a function of the initial training conditions.