Understanding how learned fear can be reduced is at the heart of treatments for anxiety disorders. Tremendous progress has been made in this regard through extinction training in which the aversive outcome is omitted. However, current progress almost entirely rests on this single paradigm, resulting in a very specialized knowledgebase at the behavioural and neural level of analysis. Here, we used a dual-paradigm approach to show that different methods that lead to reduction in learned fear in rats are dissociated in the cortex. We report that the infralimbic cortex has a very specific role in fear reduction that depends on the omission of aversive events but not on overexpectation. The orbitofrontal cortex, a structure generally overlooked in fear, is critical for downregulating fear when novel predictions about upcoming aversive events are generated, such as when fear is inflated or overexpected, but less so when an expected aversive event is omitted.

Understanding how learned fear can be reduced is at the heart of treatments for anxiety 11 disorders. Tremendous progress has been made in this regard through extinction training in 12 which an expected aversive outcome is omitted. However, current progress almost entirely rests 13 on this single paradigm, resulting in a very specialized knowledgebase at the behavioural and 14 neural level of analysis. Here, we used a paradigm-independent approach to show that different 15 methods that lead to reduction in learned fear are dissociated in the cortex. We report that the 16 infralimbic cortex has a very specific role in fear reduction that depends on the omission of 17 aversive events but not on overexpectation. The orbitofrontal cortex, a structure generally 18 overlooked in fear, is critical for downregulating fear when fear is inflated or overexpected, but 19 not when an aversive event is omitted. 20 3 Extinction learning has captivated behavioural and neural science for more than a century. It has done 21 so because it allows for the reduction of behaviours that were once adaptive but are no longer so, and 22 gives the therapist a handle to combat others that were never adaptive in the first place. The most-23 widely used method for supressing unwanted behaviour relies on the omission of the event that drives 24 this behaviour, that is, extinction driven by outcome omission. In the context of fear learning, 25 extinction by omission involves the dramatic reduction in fear-related behaviours typically observed 26 after presenting a previously established signal for an aversive event (i.e., a tone paired with shock; 27 tone→shock) in the absence of that event (tone presented alone; tone→nothing). Given its simplicity 28 and effectiveness in the treatment of anxiety disorders 1-6 , extinction by omission has received 29 significant attention in a quest to understand its underlying behavioural and neural mechanisms 7-15 . 30 Critically, although much progress has been made, this progress is limited to the case of outcome 31 omission, while another equally relevant form of extinction learning that also drives reduction in 32 unwanted behaviour, namely overexpectation, remains largely unexplored. This single-paradigm 33 approach is restrictive because at best it can oversimplify and at worst even misrepresent the function 34 of brain areas implicated in extinction learning. Here, we move beyond this paradigm-specific 35 approach and embarked on an investigation into how the brain learns from extinction using two 36 behavioural designs: extinction driven by outcome omission (described above) and extinction driven 37 by overexpectation (described below). 38In overexpectation, reduction in previously established fear responses ensue, strikingly, despite 39 continued delivery of the aversive event. This is possible because separately established signals of a 40 common aversive event (i.e., tone→shock; light→shock) can summate their fear-inducing properties 41 when encountered simultaneously (tone+light), triggeri...

No abstract