The Inhibitory Learning Approach to Exposure and Response Prevention

It should be noted that all of the following procedures are still currently being researched. While there is evidence to suggest that they could potentially be helpful in reducing OCD symptoms, they are not concretely proven to do so. These should be utilized as a last resort when all of the evidence-based treatment methods for OCD have already been accessed. For more information on evidence-based treatment methods for OCD, please click here.

By Jonathan Abramowitz, PhD

Jonathan Abramowitz, PhD is a clinical psychologist and Professor at the University of North Carolina at Chapel Hill. His clinical expertise and research focus is OCD and other anxiety-related disorders.

This article was initially published in the Fall 2018 edition of the OCD Newsletter.

Over the past several years, there have been some interesting developments in how we think about and use exposure and response prevention (ERP) treatment for OCD. These developments are based on research into how the brain learns and remembers information related to fear. These new developments have many of us in the OCD field excited about the potential to further improve the effects of ERP and open up new opportunities for people who have not been able to benefit from this treatment in the past. The term for this new approach to ERP treatment is the inhibitory learning model. Clinicians and people with OCD alike seem to get the most out of ERP when they really understand how it works. There’s something about knowing why we use a particular technique in therapy that helps you work harder at it and use it most effectively. Therefore, my goal in this article is to explain what the “inhibitory learning model” means, describe its implications for how we think about and implement ERP, and why it’s important.

What is the Inhibitory Learning Model?

To put it simply, the inhibitory learning model is a way to understand and optimize how ERP reduces (or extinguishes) obsessional fear. Traditionally, we have thought that ERP (which involves systematically confronting feared situations and resisting compulsive rituals) works by helping people “unlearn” or “erase” obsessional fears from memory. We also believed this to happen because of habituation — the process by which anxiety naturally declines over time during and between exposure sessions. Lastly, we assumed habituation to be an indicator that exposure was “working.” In other words, if anxiety declines during ERP, then we thought you were getting over your obsessional fear.

Unfortunately, none of these traditional views of ERP are consistently supported by research or clinical experience. For one thing, many people who do ERP experience habituation of anxiety during and between sessions, and yet do not improve; or they have a good response to ERP at first, but then relapse at some point later on. In addition, some peoples’ OCD symptoms improve with ERP in the absence of habituation. All of this means that, although habituation is common during ERP (indeed, habituation is a natural process), it is not as straightforward as we once thought.

Another important development that research has shown is that ERP does not cause an obsessional fear to be “unlearned” or “erased.” Instead, ERP teaches new information about safety so that following successful exposure, a feared stimulus has both its original fear-based meaning (“door knobs are dangerous”) as well as the new safety meaning (“door knobs are generally safe”). In order to be optimally effective, ERP needs to help people learn safety in such a way that it is strong enough to block out (or inhibit) the original fear — and this is where the term inhibitory learning gets its name. Fear extinction is maximized when a person uses ERP to learn new safety information that inhibits existing obsessional fear. The goal of ERP is therefore to learn safety in a way that permanently inhibits obsessional fear.

Here is another way to think about this. Imagine obsessional fear as a path that running water could follow, like a brook or a stream. Over time, the running water has cut this path pretty deep (think about how obsessional fears become stronger and stronger over the years). When you do ERP, you face your fears and learn that obsessional triggers and intrusive thoughts are safe, and this is like cutting a new path for the water to follow. But this new path needs to be cut even deeper than the older path so that it diverts the flowing water in a new direction. The inhibitory learning model provides us with ways to do ERP that help to cut this new path as deeply as possible — like the Grand Canyon!

 Elements of Inhibitory Learning

  • Focusing on Anxiety Tolerance instead of Habituation
  • Disconfirming Expectations
  • The Element of Surprise
  • Combining Fear Cues
  • Variety

Desirable Difficulties: The Theme of Inhibitory Learning

Research shows that the more effort we have to put into learning something, the better we remember what we’ve learned. The inhibitory learning model applies this science to ERP and provides strategies for making exposure practices more challenging in order to improve the chances of permanent inhibitory learning, lasting fear extinction, and long-term improvement in OCD symptoms. In our clinic, we refer to these strategies as “desirable difficulties” because they make doing ERP more effortful and challenging in the moment, but are likely to lead to more desirable outcomes in the long-term. In the next sections, I describe some of the subtle and not-so-subtle ways we import “desirable difficulties” into ERP to maximize inhibitory learning.

Focusing on Anxiety Tolerance instead of Habituation: The aim of ERP is to provide the person with new knowledge that (a) obsessional fears are less probable or severe than predicted, (b) anxiety and obsessional thoughts themselves are safe and tolerable, and (c) compulsive rituals are not necessary for safety or to tolerate anxiety. Thus, rather than aiming for the decline of anxiety (habituation) during exposure, the inhibitory learning approach to ERP teaches people how to be open-minded toward experiencing anxiety and fear when these experiences inevitably show up. Indeed, fear and anxiety (and other emotions in OCD such as disgust or guilt) are universal and even adaptive experiences, not something that need to be “fixed” or gotten rid of. Most importantly, even if they can be unwanted, intense, and distressing, these emotions and thoughts are safe. From an inhibitory learning perspective, fear extinction (and long-term improvement in OCD) depends not only on learning that feared stimuli are safe, but that it is also safe to experience the emotional response that is triggered by these stimuli.

Disconfirming Expectations: People with OCD tend to over-predict negative outcomes, such as becoming sick, being responsible for harm, acting immorally, and not being able to tolerate anxiety or uncertainty. Within the inhibitory learning framework, ERP is designed to teach what “needs to be learned” to disconfirm these feared outcomes. For example, someone with obsessions about harm may predict that if they are alone with a baby for 30 minutes, they will lose control and harm the child. Thus, an ERP exercise is designed to directly disconfirm this prediction, perhaps by handling an infant while alone for 45 minutes and keeping track of whether any harmful acts are committed. To facilitate extinction learning, each exposure practice is focused on determining whether or not the expected negative outcome occurred (rather than waiting for anxiety to habituate). In the case of feared outcomes that won’t occur until the distant future (such as “I will go to hell if I have a bad thought”), ERP can be focused on determining whether the uncertainty about this future outcome was “manageable” or not. The end of an ERP session when conducted from an inhibitory learning perspective is determined by conditions that disconfirm expectations (“Look at that, I didn’t kill the baby even after 2 hours alone,” or “Wow, I guess I can manage uncertainty”), not when anxiety subsides (although it often does naturally).

Surprise: Research indicates that inhibitory learning and extinction are strongest when there is a large mismatch between feared expectations (“I will get sick from touching a toilet” or “Someone will break into the house unless I check again”) and actual outcome (not becoming sick or no break-ins). In other words, when we are surprised by the outcome of an exposure session it tends to stick in our minds. To this end, we try to set up exposures so that people are “pleasantly surprised” by the non-occurrence of their feared outcome (or their ability to manage obsessional thoughts, anxiety, and uncertainty) by deliberately orchestrating opportunities for feared outcomes that are unlikely or impossible (e.g., causing “bad luck” by writing certain numbers), or at least tolerable (e.g., feeling uncertain). We capitalize on the element of surprise by encouraging people to conduct exposures to feared situations at a greater level of intensity, duration, or frequency than the person believes would be “safe” or “tolerable” This helps solidify learned safety in a way that best inhibits obsessional fear.

Combining Fear Cues: Another inhibitory learning strategy is to combine multiple fear cues during ERP. For example, this might involve helping someone who fears molesting a child to (a) conduct imaginal exposure to thoughts of sexually touching the child, then (b) in vivo exposure to giving the child a hug or having him sit on your lap, and finally (c) engage in combined exposure to hugging the child while imagining sexually touching her. When a feared outcome fails to occur despite the presence of multiple fear cues, it creates strong learned safety to inhibit existing obsessional fear.

Variety: Optimal fear extinction also requires that safety learning occur in a variety of contexts. This is because if safety is learned in one situation, it may not necessarily be remembered in a different situation. Accordingly, we deliberately conduct exposure tasks with varied stimuli (the toilet, floor, sink), in different locations (e.g., the restroom at home versus a public restroom), in different interpersonal contexts (with the therapist versus alone), and under different emotional conditions (when calm versus when already anxious).

Research Evidence for Inhibitory Learning

The clinical strategies I’ve described here are based on a large body of experimental research on fear extinction in humans and in animals. It remains to be seen how well results from animal research really translate to understanding humans. Moreover, we are just starting to conduct research on inhibitory learning in OCD specifically since most of the existing research is on other types of anxiety problems, such as panic attacks and phobias. Therefore, although the inhibitory learning recommendations for ERP are derived from a sound theoretical model and preliminary research, we need additional studies to demonstrate that this approach gives us better results than the traditional (habituation-based) approach to ERP.


For over 50 years, ERP has been the best and most cost-effective treatment available for OCD. Still, not everyone responds optimally to this intervention. So, when advances come along that can be applied to potentially make this treatment better, it is well worth considering them. Progress in how we understand the neuroscience of extinction learning provides exciting new directions for ERP. Yet at this point, we still have much to learn regarding the precise methods to optimize inhibitory learning and how much of an improvement this is relative to traditional ERP. At this point, I do not believe it is time for a complete paradigm shift toward the inhibitory learning model of ERP, but I do believe that it is worth trying out the strategies described in this article in instances where someone has not had sufficient improvement from traditional ERP. I also hope that the years to come will see further studies to help us learn more about this new way of thinking about OCD treatment.

Suggested Readings:

Abramowitz, J. S. (2018). Getting over OCD (2nd edition). New York: Guilford Press

Abramowitz, J. S., & Arch, J. J. (2014). Strategies for improving long-term outcomes in cognitive behavioral therapy for obsessive-compulsive disorder: insights from learning theory. Cognitive and Behavioral Practice21(1), 20-31.

Arch, J. J., & Abramowitz, J. S. (2015). Exposure therapy for obsessive–compulsive disorder: An optimizing inhibitory learning approach. Journal of Obsessive-Compulsive and Related Disorders6, 174-182.

Craske, M. G., Treanor, M., Conway, C. C., Zbozinek, T., & Vervliet, B. (2014). Maximizing exposure therapy: an inhibitory learning approach. Behaviour research and therapy58, 10-23.

Jacoby, R. J., & Abramowitz, J. S. (2016). Inhibitory learning approaches to exposure therapy: A critical review and translation to obsessive-compulsive disorder. Clinical psychology review49, 28-40.