Partial reinforcement and conditioned taste aversion: No evidence for resistance to extinction

Abstract

The partial reinforcement extinction effect (PREE) is the observation that, following training in which a response is followed by reward on only a subset of trials, the response is more resistant to extinction following the total removal of reward than it is after training in which reward is presented on all trials. The PREE is almost ubiquitous in instrumental conditioning procedures but only inconsistently observed in Pavlovian conditioning. In his classic review of animal learning, Mackintosh attributes the bulk of the PREE to generalisation decrement relating to the fact that partial reinforcement typically ensures that acquisition of responding has taken place in conditions similar to that of extinction (e.g., in the absence of the reinforcer). We report here that extinction of a conditioned taste aversion is not retarded by partial reinforcement in terms of either consumption of the taste or hedonic reactions to it (assessed through the analysis of licking microstructure). These results are consistent with Mackintosh’s analysis of the PREE and the way in which it might differ between instrumental and Pavlovian conditioning.