Maximizing Exposure Therapy: An Inhibitory Learning Approach Summary and Analysis by Dr. Katy Manetta
One of the most seminal articles in the field of Exposure and Response Prevention to date is the 2014 article by Craske et al. The article, entitled Maximizing Exposure Therapy: An Inhibitory Learning Approach, was authored by Michelle G. Craske, Michael Treanor, Chris Conway, Tomislav Zbozinek,and Bram Vervliet.
Within the clinical practice of Exposure and Response Prevention (ERP), this article represents one of the most important pieces of research to date. Any clinician who treats anxiety disorders and/or Obsessive Compulsive Disorder (OCD) would be well-served to familiarize themselves with this seminal piece of literature.
Exposure therapy is known as the most effective therapy approach for treating anxiety disorders and OCD. However, studies suggest that some individuals do not achieve significant remission of their symptoms during exposure therapy or have symptoms that return shortly after treatment.
Various models have been proposed to explain the mechanisms for change that underlie successful exposure therapy. Evidence from prior studies strongly supports the inhibitory model of extinction. However, prior to the Craske et al article, there had been little discussion of how to apply this model in clinical practice. Craske reviews the inhibitory learning model of extinction as a mechanism of change during ERP. Her paper compares the application of the inhibitory learning model to other models proposed in prior research. Additionally, it provides specific guidance on how to apply this model successfully in ways that distinguish it from the standard “fear habituation” and “belief disconfirmation” approaches to fear extinction.
Inhibitory Learning Model of Extinction To understand the strategies proposed in the Craske et al article, one must first have a basic grasp of the inhibitory model of learning. An easy way to understand this model of learning is to use the example that Dr. Craske herself references in a wonderful interview with Dr. Jacqueline Persons. Dr. Craske describes the inhibitory learning model through use of a common phobia—a fear of dogs. An individual’s fear of dogs generally develops following a negative experience wherein an individual is either bitten or attacked by a dog. That individual may subsequently become fearful when they once again find themselves confronted by a dog who is close enough to reflect a potential threat. This response is referred to as an excitatory response, meaning that subsequent exposures to the original stimulus produce the psychological andphysiological symptoms of arousal, fear and anxiety. The basic principal behind exposure therapy is that through subsequent exposures to dogs who are calm and docile, the learned excitatory response is ultimately extinguished. However, what the inhibitory learning model uniquely posits is that during these subsequent exposure experiences, the original excitatory response is not erased, but rather a new secondary learning experience develops. The individual still retains the original excitatory response that occurred during the initial fearful experience, but now has also developed a new inhibitory response that competes with the former excitatory response.
An important thing to note about the process of creating inhibitory learning experiences to counteract the excitatory experiences is to recognize that the excitatory responses are far more powerful than the inhibitory ones. As such, when clients initially go through these inhibitory learning experiences, their new responses are actually quite fragile and are easily shaken through additional negative experiences. Therefore, the Inhibitory learning experiences are vulnerable to reverting back to the original learned response. As we approach exposure therapy through the lens of the inhibitory learning model, our goal is to encourage the inhibitory response in such a way as to develop a very strong inhibitory relationship that is unlikely to revert back to its original fearful excitatory state.
Deficits in Inhibition and Anxiety Disorders Craske’s paper reviews research which demonstrates that certain individuals have a propensity for anxiety and OCD and that their development of these disorders and their failure in ERP treatment are likely the result of deficits in the processes of inhibitory learning. Therefore, patients with anxiety and OCD, particularly those who initially fail to benefit from ERP, are likely to have deficits in inhibitory learning. Furthermore, given the fragility of the learned response, Craske illustrates the importance of designing exposures to maximize the inhibitory learning experience. Her paper provides practical clinical examples to demonstrate the application of the inhibitory learning model to optimize exposure therapy for patients with anxiety or OCD. The paper also reviews specific strategies to maximize the impact of the intervention based on the inhibitory learning model.
The authors present 8 therapeutic strategies for enhancing inhibitory learning. These 8 strategies are described below:
1. Expectancy Violation The first strategy suggested by the authors is to “design exposures that maximally violate expectancies regarding the frequency or intensity of aversive outcomes”. Evidence presented in the article demonstrates that itis the degree of difference between what the client expects to happen and what actually happens that is predictive of successful exposure outcomes. The greater the violation of the clients’ expectations, the more powerful the exposure becomes. Clinicians must ask themselves what their clients need to learn in order to most violate their current expectations. In other words, exposures should be designed to focus on what the client needs to learn rather than by asking the client to “stay in the situation until fear declines” (which would have been the emphasis of a habituation-based model of exposure therapy). As an example—lets imagine that a socially anxious individual expects that they will experience rejection or humiliation in certain social situations. The exposure is then designed to help the client learn that the feared outcomes may not occur or that if they do occur, they are not as intolerable as expected. The length of time the individual stays in the exposure is not determined by fear reduction (as was the case with habituation-based models) but is determined by the length of time required for the expectancy to be violated.
In order to maximize learning, after exposures, clinicians should routinely ask their clients to explore what they learned. They should ask their clients to compare their expectations with what actually occurred, to identify any discrepancies between the prediction and outcome, and to report any surprises they experienced as a result of the exposure. This helps to reinforce and strengthen the inhibitory learning experience. Through this mental rehearsal of the events that transpired, memory is strengthened and consolidated.It is important to note that some of the traditional strategies used by clinicians to “prepare” their clients for exposure may actually interfere with extinction. For example, encouraging clients to estimate event probability prior to exposure (e.g. “I am unlikely to be humiliated since the party is all friends”) or to remind themselves of coping self-talk (e.g. “If people humiliate me, then they weren’t really my friends anyway”) may actually interfere with inhibitory learning. Rather, as clinicians, we should avoid any efforts at cognitive restructuring and encourage our clients to approach exposures with their existing predictions. Only after the exposure is complete should we use cognitive interventions to encourage clients to examine the degree of difference between prediction and outcome in order to solidify learning and memory.
2. Deepened Extinction Craske et al describe research which shows that combining multiple internal and external cues in various stages of extinction is likely to resultin a deepened extinction experience that reduces the likelihood that our client’s original fear response will return. One of the examples used to clarify this concept is the treatment of agoraphobia. Interoceptive exposure, such as caffeine consumption, may initially be used as a single internal trigger. However, it can also be combined with in vivo exposure to external triggers, such as shopping in a crowded mall. Combining various cues and triggers in the same exposure results in a deepened extinction experience for the client and reduces the likelihood that the client will revert back to the original excitatory response.
Deepening the extinction experience can also be as simple as including various alterations of the original feared exposure. In once again considering a fear of dogs, exposing the patient to barking dogs, small dogs, big dogs, leashed dogs, etc., can be combined to provide a deepened extinction experience for our clients.
3. Occasional Reinforced ExtinctionWhile we generally anticipate that our client’s predicted outcomes will differ from the actual exposure outcomes, research indicates that it is also a helpful experience if the client’s predictions occasionally turn out to be accurate. In other words, occasional aversive outcomes are helpful. The likely explanation for this is that the negative outcome experience reinforces original expectancies, thus allowing subsequent trials to result in more violations of the original expectations. For example, if an individual occasionally experiences a social rejection, this actually helps to reinforce learning because it provides them with yet another opportunity to expect the worst and have their expectancies violated anew. Another excellent example occurs with a fear of vomiting. Clients who fear vomiting may, on occasion, actually vomit. In this case, not only does this experience reinvigorate their original feared outcomes, but it may also serve to violate a different expectancy that underlies many clients’ fear of vomiting—the fear that they cannot cope with the experience of vomiting. Therefore, these occasional reinforcements can in many ways actually serve to improve outcomes.
4. Removal of Safety SignalsA fourth strategy described by the authors for enhancing inhibitory learning is the removal of what are referred to in the ERP literature as “safety signals” or “safety behaviors”. Safety signals/behaviors can refer to almost anything that an individual uses to reduce the negative consequences of an exposureto a feared stimuli. Whether the feared consequences are perceived or actual, and whether they are physiological or psychological, safety signals/ behaviors are perceived to mitigate the feared consequences.Previous research has demonstrated that the use of safety signals or behaviors interfered with the learning that occurs during exposure therapy. Therefore, it has been common practice to instruct clients to refrain from the use of safety signals/behaviors during ERP. However, recently there has been research that contradicts the previous findings. For example, some researchers have found that the use of hygienic wipes following exposure to a perceived contaminant did not have a negative effect on outcomes compared to exposures performed without the perceived “safety” of the use of cleansing wipes. Craske et al propose that the likelihood that safety behaviors will interfere with extinction learning depends largely on how much the safety signal alters the predicted outcome. So if, for example, a client with panic disorder goes to a busy shopping mall but has a prescription bottle of Xanax with them, their predicted outcome might be that, “If I panic, I can take a Xanax to calm me down.” If the client takes the Xanax and it does calm then down, then it is evident that little learning has taken place. The client’s expectancy has not been violated and no learning occurs. However, if a client predicts that they will “pass out and run away screaming” if they “don’t wash their hands” after exposure to a public toilet, it is possible that the use of a hygienic wipe will not significantly alter the client’s expectancy. In other words, the impact of safety signals/behaviors will depend on the strength of its perceived impact on the likelihood of the feared outcome.
The authors do propose however that it is generally preferred to remove safety signals/behaviors and only retain them during the initial phase of treatment, especially if it is necessary to reduce treatment attrition.
5. Variability ERP is customarily performed in a hierarchical fashion, with clients proceeding from the least feared stimuli to the most feared in a predicable, ordered fashion. However, the authors of the Craske et al article make a strong case for restricting the orderly use of the hierarchy for the initial stages of treatment to encourage treatment compliance. Evidence is presented which demonstrates that clients benefit from a less orderly and predictable approach to treatment. They argue for the use of random order of exposure without regard to the fear hierarchy as well as the use of variability with respect to the number of exposures and the length of exposure. The researchers present evidence which demonstrates that thistype of variability actually increases the intensity of the exposure experience for the client and leads to better long term outcomes.
6. Retrieval Cues In the treatment of anxiety disorders, retrieval cues are typically items or locations that act as reminders of the learning that took place during the exposure experiences. This technique of using retrieval cues must be used with caution so as not to be mistakenly used as safety signals. For example, a therapist’s office that reminds the client of a powerful exposure, and the subsequent learning that took place, can be an excellent retrieval cue and can help to reinforce the client’s inhibitory learning. However, for a client who fears death from a panic attack, the therapists’s office can represent a safety signal if they believe themselves to be “in safe hands” with a trustworthy therapist who would call 911 if their client were in distress. As such, retrieval cues must be used with caution and only when sufficient evidence exists to support the use of a location or object as a successful retrieval cue.
7. Multiple Contexts The seventh strategy for enhancing outcomes presented by the authors is referred to as “contact renewal”. This involves conducting all types of exposures (in vivo, imagined, interoceptive) in multiple contexts, such as when alone, in different places, or at different times of the day. Although the results are somewhat inconsistent, the authors include this strategy as being of potential benefit to the process of extinguishing fear responses.
8. Reconsolidation The final strategy discussed in the Craske et al article for enhancing inhibitory learning involves memory reconsolidation. The authors report on research which suggests that new information presented at certain times in the consolidation of memory may be able to weaken an individual’s memory of fear. The resulting clinical recommendation is to introduce the original feared stimulus for a brief period 30 minutes prior to repeated trials of exposure. However, the authors caution that these recommendations are based on preliminary research findings and require additional exploration.
Therapeutic Strategy for Enhancing Inhibitory Regulation The authors of this article identify a strategy for enhancing inhibitory regulation which involves affect labeling. Research from social neuroscience is reviewed which suggests that labeling affect during exposure reduces activity in theamygdala (a part of the brain known to be associated with anxious responses). The authors also describe other research demonstrating improved outcomes through the use of affect labeling during exposures for simple phobias. These improved outcomes were both immediate and after a one-week follow up. The authors suggest that exposure outcomes can be enhanced by routinely asking clients to state their feelings during exposures.
Sample Questions The authors present numerous case studies which demonstrate the various strategies recommended by the findings in this article. They outline the following sample questions/prompts that they suggest be incorporated into clinical exposures:
BEFORE exposure: Goal - Identify specific exposure goal with client Ask, “What are you most worried will happen?” Ask, “On a scale of 0-100, how likely does this seem”
AFTER exposure: Ask, “Did what you were most worried about occur? Y or N?" Ask, “How do you know?” Ask, “What did you learn?”