This article is about a recent research paper published in Neuron by Marianne Reddan, Tor Wager, and Daniela Schiller.

As we navigate the world we learn, through experience, what is dangerous and what is safe. This information is critical to our survival. It is equally adaptive, however, for us to “unlearn” these associations. For example, if you were in a traumatic car accident you may learn to fear riding in cars. This fear, though it may have a justifiable cause, may prevent you from fully participating in our society in a way that benefits you because cars are not inherently dangerous and are sometimes necessary for commuting and general transportation. How many we “unlearn” these fears that do not serve our organism?

The classical clinical approach to “unlearning” a fear association is exposure therapy. Exposure therapy is rooted in concepts related to Pavlovian conditioning. Through repeated exposures to the stimulus that you fear (i.e., cars) without its pairing with a dangerous consequence (i.e., the crash) the learned association between car and crash should become suppressed by a new memory– an extinction memory. In the laboratory we call this process extinction learning.

A common barrier to exposure therapy is physically re-exposing a person to a traumatic situation. It may be re-traumatizing to force a survivor of a car accident back into the passenger seat of an automotive. In some cases, it can be infeasible or impossible to re-expose someone to their trauma because it is situated in a specific place and time that could still be dangerous (i.e., war veterans and rape survivors). How then can we begin to learn to feel safe again in an unsafe world?

In this investigation we demonstrate that extinction learning can be successfully performed in one’s imagination. That is, imagining the threatening stimulus (i.e, the car or riding in a car) can alter the pairing between the threatening stimulus and the dangerous consequence in one’s  memory. To demonstrate this, participants first underwent auditory threat conditioning, where they learned that one tone (the CS+) was paired with an aversive shock 33% of the time (i.e., it signaled danger), and another tone, (the CS-) was never paired with shock (i.e., it signaled safety).  Participants were then randomly assigned to either (1) a “real” extinction session, where they heard both tones, repeatedly, without experiencing shocks; (2) an imagined extinction sesion, where they were cued to “replay the tones in their head” to the best of their ability, repeatedly, without experiencing shocks; or (3) a “no” extinction session, where they were cued to imagined two neutral sounds from nature (birds singing and rain falling), repeatedly, without experiencing shocks. The “no” extinction group served as a control for the intrinsic and relaxing properties of imagination.The threat memory was then reinstated in all subjects, that is, all participants were shocked four times without hearing any sounds, and then their threat recovery was examined by re-exposing all participants to the original threatening (CS+) and safety (CS-) tones.

Upon re-exposure, physiological responses to threat were reduced in the real and imagined extinction groups, to an equal degree. This reduction was significantly lower than threat responding in the no extinction control group. That is, participants who underwent real or imagined extinction training sweated less when they heard the threat eliciting tone again, than those who only imagined birds singing and raining falling. To examine the brain bases of this effect, we trained a whole brain classifier to identify a threat-related brain signature, and then tested it on each group during re-exposure. This pattern of brain activity can tell us, to a quantifiable degree, how likely it is that a person’s brain is representing threat. Threat-related brain responses were reduced in imagined and standard extinction groups relative to the no extinction control, and a person’s neural threat pattern expression was correlated with their physiological responses. These findings demonstrate that imagined exposures to threatening stimuli are effective in the reduction of both physiological and neural correlates of threat.

We also found that like real extinction training, imagined extinction recruits the ventromedial prefrontal cortex (vmPFC), a region known to be involved in emotion regulation. Furthermore, imagined extinction activates the auditory cortex, indicating that as people simulated the sounds, they activated regions of their brain which are critical to *hearing* sounds. Activation in the vmPFC, auditory cortex, and the amygdala, a brain region known to be critical to threat-learning, predicted the success of imagined extinction across individuals. This indicates that imagined extinction works via a mechanism similar to that of real extinction training. This indicates that imagination is an important clinical tool.

Your experience of the world is constructed in your brain. As you encounter the barrage of stimuli around you, you automatically evaluate it based on your prior experiences (i.e., Is this something I should attend to? Is this something I should approach?). In this way, your current experience is intertwined with both your past and your simulation of the future. In this study we showed that you can use this powerful tool — your imagination — to learn new things about the world around you. Through imagined exposures to a threatening sound, participants in this study learned that the threat they once expected to follow the sound will no longer come. This is an important finding for both basic scientists and clinicians because it demonstrates the utility of imagination in emotion regulation and learning.