Prefrontal circuits split fear extinction into retrieval and positive affect
Exposure-based therapy remains the backbone of treatment for trauma-related neuropsychiatric disorders, yet its biology has always been more complicated than simple fear suppression. The extinction process does reduce conditioned freezing, but durable recovery depends on whether the brain can store a safer interpretation of the threat and keep that memory available when stress returns. In this JCI commentary, Quan-Son Eric Le and Luis E. Rosas-Vidal examine new work showing that extinction is not carried by a single prefrontal output stream. Instead, distinct medial prefrontal cortex populations appear to route fear regulation and positive affect through separate hypothalamic targets.
The study discussed here used activity-dependent neuronal labeling in a mouse model of fear extinction to tag extinction-engaged medial prefrontal cortex ensembles. From that starting point, investigators separated two pathways with different behavioral functions: one projecting from the mPFC to the zona incerta and another projecting from the mPFC to the supramamillary nucleus. The first pathway governed extinction retrieval. The second was tied to positive affect and to the persistence of extinction over time. In plain clinical terms, the brain is not merely learning to fear less; it is also learning to feel better while holding on to that lesson.
Fear can extinguish without ever becoming durable.
Distinct outputs from the same prefrontal region carry different behavioral jobs
The mPFC-to-zona incerta projection acted through SST-expressing neurons in the zona incerta and was required for retrieval of extinction memory, as measured by reduced freezing to the conditioned tone. That matters because retrieval failure is one of the most obvious reasons patients relapse after initial improvement. If the learned safety signal cannot be accessed at the right moment, the original fear response returns even when the threat is no longer present. The pathway therefore looks like a retrieval circuit, not a mood circuit.
By contrast, the mPFC-to-supramamillary nucleus projection mapped to SIRT1-expressing neurons and was linked to both positive affect and long-term persistence of extinction retrieval. The functional split is clinically compelling because affective state is not simply a bystander in psychotherapy. Patients who experience a sense of relief, mastery, or safety during treatment often tolerate repeated exposure better and may consolidate extinction more effectively. The biology described here gives that bedside observation a more specific neural framework: positive affect may strengthen the memory trace that keeps fear from returning.
The commentary also draws attention to the fact that the same cortical region can generate behaviorally distinct projections depending on which downstream node it engages. That arrangement fits with modern circuit neuroscience, where a region’s effect is determined less by its label than by its wiring. Medial prefrontal cortex activity is often discussed broadly in fear regulation, but broad labels can obscure the real therapeutic target. What matters here is not “the prefrontal cortex” in the abstract, but the exact route it takes to shape extinction retrieval or emotional valence.
SIRT1 and relapse vulnerability raise the stakes for circuit-specific therapy
One of the more clinically provocative findings discussed in the commentary is the role of SIRT1-expressing neurons in the supramamillary nucleus. The extinction-recruited mPFC-SuM projection was not only associated with positive affect; it also controlled vulnerability to fear relapse. Overexpression of SIRT1 reduced that relapse susceptibility. In a field where relapse after exposure therapy remains a stubborn problem, the implication is straightforward: there may be molecular leverage points inside extinction circuitry that influence whether therapeutic gains last.
This is not a generic “more extinction is better” story. The work instead suggests that extinction may require two separable ingredients: access to the safety memory and a favorable affective state that helps preserve it. A circuit that can retrieve extinction but cannot sustain positive affect may leave patients exposed to return of fear under stress. A circuit that supports positive affect without the correct retrieval network may not suppress the conditioned response at the critical moment. The interaction between the two may be what ultimately determines resilience.
SIRT1 has long been discussed in broader cellular stress and metabolic contexts, but here it is embedded in a very specific behavioral role. That specificity raises an important translational question: can neuromodulatory or molecular strategies be designed to strengthen only the extinction-supportive branch of the prefrontal-hypothalamic network without altering unrelated fear or reward circuits? The commentary stops short of any clinical claim, and for good reason. These data come from a mouse model, and any move toward patient care would require careful validation in humans, where trauma history, comorbid depression, medication exposure, and developmental stage all shape extinction learning.
Why a positive-affect pathway changes the way extinction therapy is understood
The concept that positive affect participates in fear extinction shifts the therapeutic frame. Exposure therapy is often presented as a process of repeated confrontation with threat cues until the threat loses power. Yet the new circuit model suggests that the brain may also need a counterbalancing emotional signal during learning, one that marks the environment as tolerable and safe. That is a more biologically rich explanation for why some patients respond well to exposure while others struggle despite repeated sessions.
It also gives clinicians a more nuanced way to think about relapse. A return of fear may not simply mean the old memory was never erased. It may mean the extinction memory was weakly consolidated, poorly retrieved, or not linked strongly enough to a positive internal state to withstand later stress. The mPFC-ZI and mPFC-SuM projections described in this work separate those possibilities into distinct circuit elements. For translational psychiatry, that separation is valuable because it points toward more precise intervention targets than global suppression of fear-related activity.
The commentary closes by framing these cortico-hypothalamic projections as potential targets for suppressing fear expression in PTSD and related disorders. The caution is obvious: optogenetic and labeled-ensemble findings in mice do not translate directly into a ready-made therapy. Even so, the work clarifies a major biological question. Fear extinction is not one process, but at least two—one that retrieves safety and another that preserves the emotional conditions under which safety memory survives. Future studies will need to determine whether these same pathway divisions exist in humans and whether they can be shaped without disrupting other prefrontal functions, especially in patients whose trauma responses are already entangled with depression, anhedonia, or chronic stress.