The role of prediction error and memory destabilization in extinction of cued-fear within the reconsolidation window

Abstract

Extinction of a cued-fear memory within the reconsolidation window has been proposed to prevent fear reacquisition by reconsolidation interference. This 'retrieval-extinction' procedure has received interest for its therapeutic potential to reduce the impact of fear memories on behavior. To fully exploit its therapeutic potential, it is critical to understand the mechanisms that underlie the 'retrieval-extinction' effect. If the effect depends upon reconsolidation of the original memory, then it would be predicted that destabilization, induced by prediction error, would be critical for observing the effect. Here, the dependency of the retrieval-extinction effect on memory destabilization or prediction error was investigated in pavlovian cued-fear conditioned adult male rats. The requirement for memory destabilization, and thus reconsolidation, for the retrieval-extinction effect was subsequently investigated using region-specific pharmacological blockade of dopamine D1-receptors. Intra-basolateral amygdala antagonism of dopamine D1-receptors did not prevent the reacquisition of fear associated with the retrieval-extinction procedure. The requirement for prediction error was assessed by using a reinforced or non-reinforced memory retrieval trial before extinction, compared to a no-retrieval, extinction-only control. Both the reinforced (no prediction error) and non-reinforced retrieval sessions led to a decrease in fear reacquisition, suggesting that engagement of prediction error does not influence the occurrence of retrieval-extinction. Together, these data suggest that retrieval-extinction does not require memory destabilization, since behavioral or pharmacological interventions that prevent destabilization did not disrupt any capacity to attenuate fear.

Fig. 1

a Rats were divided into four groups, two receiving vehicle (Veh) infusion prior to either memory retrieval (Ret) or returned to the home cage (NoRet), and two receiving the D1R antagonist Schering-23390 (Sch) before retrieval or returned to home cage. b All groups acquired fear conditioning across CS-US presentations to an equal extent. c There was no difference in freezing between the two Ret groups, suggesting acute Sch infusion into the BLA had no effect on freezing expression itself. d All groups extinguished responding to the CS over time. e Following extinction all groups had reduced fearful behavior to the CS to an equivalent level. At the end of the CS the US was presented again for fear reacquisition (ReAcq). f The effects of reactivation (Ret) and/or Sch on the long term memory (LTM) of fear were tested 24 h later. Only the groups pre-treated with Sch appeared to have a retrieval-extinction effect on fear reacquisition. **p < 0.001

Fig. 2

a Schema of experimental protocol. Animals were divided into three groups after training: no retrieval controls (NoRet), reinforced retrieval (Ret+), or non-reinforced retrieval (Ret−). b All groups acquired fear conditioning to a similar level across three CS-US pairings during training. c The Ret− and Ret+ groups had similar freezing levels during memory retrieval (Ret), for the Ret+ group this CS presentation terminated with the US. d All animals were presented with 19 CSs to extinguish the fear memory. e Following extinction all groups had reduced freezing to the CS to an equivalent level across groups. During the reacquisition (ReAcq) session the CS was presented again but coterminous with a foot shock to all groups. f The NoRet animals reacquired to the greatest extent, significantly more than the Ret + group

Fig. 3

a Schema of experimental protocol. Rats were divided into three groups. NoRet and Ret− receiving 8 US-CS pairings (4 per session across 2 days), whereas the Ret + group received 7 US-CS pairings across training (4 in first session and 3 in the second). b All groups acquired fear conditioning to a similar level across pairings during training. The NoRet and Ret group received 8 pairings, whilst Ret+ received 7, to equate for CS-US pairing exposure across days. c The Ret− and Ret+ groups had similar freezing levels during memory reactivation (Ret). d All groups extinguished fear over the session. NoRet and Ret+ groups were presented with 19 CSs whilst Ret− were presented with 18 CSs to equate for CS exposure across days. e Following extinction all groups had reduced fearful behavior to the CS to an equivalent level. During the reacquisition (ReAcq) session the CS was presented again coterminous with a footshock to all groups. f The NoRet animals reacquired freezing to the greatest extent, significantly more than the Ret− group and the Ret+ group

Fig. 4

a Rats were divided into three groups after training. b The prospective groups acquired fear conditioning to the same extent. c The Ret groups were re-exposed to an unreinforced presentation of the CS, and the Cxt group was placed in the experimental context for the same duration of time. d The Ext groups (Cxt-Ext and Ret-Ext) extinguished across the session whereas the Ret-Cxt group were placed in the experimental context for the same duration of time. e The groups that underwent extinction presented low levels of freezing the following day, in contrast to the Ret-Cxt group. At the end of the session all groups received a foot shock coterminous with the CS presentation. f The Ret-Ext group reacquired freezing to a lower extent than the Cxt-Ext or Ret-Cxt groups, indicating that CS presentation, and not just context exposure, is needed for the effect