Beta-blocking Pavlov's Bells: Propranolol attenuates compound extinction in an error-dependent manner

One hundred years ago, Pavlov observed that omitting the reinforcer that was previously paired with a conditional stimulus resulted in a decrease in the conditional response evoked by that stimulus, a phenomenon labeled extinction. Notably, Pavlov found that extinction was not permanent and the conditional response could recover under a range of conditions. As extinction is the basis of many modern therapies, the aim of much current work is to identify ways to enhance its longevity. To this end, one strategy is compound extinction, where a combination of previously reinforced stimuli is presented together for the first time during extinction training. This treatment has been shown to enhance extinction learning, evidenced by reduced future spontaneous recovery. However, the mechanisms are not fully understood. Experiment 1 assessed whether the compound extinction effect is the result of increased expectation of reward generated by the compound of stimuli that is then violated, driving further learning, or more simply, whether the novelty of the compound is able to re-engage attention and thus promotes extinction without increasing prediction error. Experiment 2 tested whether the effects of the noradrenaline β-receptor antagonist propranolol, shown elsewhere to reduce the compound extinction effect, relate to prediction error or novelty. We found that, when equating the novelty of the stimulus compound, larger prediction error resulted in better extinction evidenced as reduced spontaneous recovery. Further, we found that propranolol reduced this effect suggesting that prediction error rather than novelty recruits noradrenergic signalling. Together our results identify behavioural and pharmacological strategies that can improve the long-term expression of extinction.