Dr Laura Bradfield

School of Life Sciences, UTS

Title: Contextual regulation of reinstatement involving choice

Abstract: Relapse to substance use disorder, overeating, or other maladaptive actions following abstinence and/or therapy is commonly modelled in animals, however, the majority of these studies employ procedures that are either Pavlovian, or involve only a single instrumental response. Therefore, the mechanisms of relapse in situations involving choice between multiple actions and outcomes are less well-understood. To address this question, we investigated the contextual modulation of outcome-selective reinstatement. In Experiment 1, rats were trained to press a left and a right lever for a pellet and a sucrose outcome, respectively (counterbalanced) in context A. Extinction on both levers then took place in either context A or B, followed by testing in context A or B. This rendered 4 groups in total: AAA, AAB, ABA, and ABB. On test, animals received ‘free’ deliveries of pellets or sucrose and their subsequent lever presses recorded. Group AAA reinstated selectively on the lever that had previously earned the presented outcome (e.g. pellet presentation reinstated pressing on the pellet lever, sucrose on the sucrose lever). Surprisingly, animals in group ABB also demonstrated intact outcome-selective reinstatement (reinstated > nonreinstated), suggesting that outcome-response (O-R) contingency knowledge had transferred across contexts. In contrast, animals in groups AAB and ABA responded equally on both levers (i.e. reinstated = nonreinstated), suggesting that extinction learning, unlike O-r learning, was context-dependent. Experiment 2 was conducted identically, except that rats received two sessions of extinction and were tested one day later rather than immediately. This time, all groups demonstrated evidence of intact outcome-selective reinstatement, regardless of context. Together, these findings support the notion that any contextual modulation of instrumental responding involving choice is transient, and occurs immediately after new learning but not when that learning is well-established.