Magic Mustang Tamer

Enhancing Learning Inhibition: my reading

Each paper in the “My Reading” series is an interpretation of a scientific paper on a topic of interest to animal trainers. This paper is presented in preparation for a larger work on cognitive models that deal with fear.

Enhancing Learning Inhibition

READ — Date: 5 March 2020 Reason: Review of Issues around fear training by exposure therapy

CITATION: 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.

KEYWORDS: inhibitory learning, expectancy violation, deepened extinction

QUESTIONS ASKED: How can we enhance inhibitory learning? What are the best practices for interventions?

METHOD/APPROACH: a review article

SIGNIFICANCE: This paper provides guidance for designing fear interventions using exposure therapy.

KEY CONCEPTS: Some individuals have deficits in their inhibitory learning ability and develop fears as a result. By teaching them to process inhibitory learning directly, they can have less fear or at least more realistic fears.

There are other mechanisms underlying exposure therapy, but the inhibitory learning model is a construct that provides heuristic and testable strategies to maximize efficacy. The problems with exposure therapy, which is basically repetitively approaching fear-provoking stimuli. A substantial number of individuals fail to show long term improvement after an intervention based on exposure therapy. The reasons for failure are 1) spontaneous recovery of conditioned fear after extinction as a function of time, 2) a change in context from intervention to subsequent encounters with the CS, 3) retraumatization by a reoccurrence of the unconditional aversive stimulus, 4) if the CS and US are paired again.

The inhibitory learning doesn’t erase learned fear, it provides additional information that the CS does not predict the US. At a neurochemical level, cortical influences inhibit the activity of the amygdala. Because the fear at the level of the amygdala is never actually erased, it can quickly spontaneously reoccur.

This paper provides 8 strategies to enhance inhibitory learning.

  1. Expectancy violation
  2. Deepened extinction
  3. Occasionally reinforced extinction
  4. Removal of safety signals
  5. Variability
  6. Retrieval Cues
  7. Multiple contexts
  8. Affect labeling

Expectancy Violation – Make the doom/gloom prediction fail in a surprising way. The CS should be very salient, as should the lack of an aversive US. Working up a fear hierarchy can take make the system too predictable. Sustain arousal, avoid habituation. Don’t try to eliminate the prediction itself, just don’t let the outcome match the prediction.

Deepen Extinction – If the prediction is based on multiple types of CS (all predicting that outcome), it helps to extinguish them separately first, then combine them so that none of them are very good predictors. Authors suggest also introducing a new cue as a CS in a different modality (in vivo, or interoceptive) just to extinguish it with the old cue.

Occasional Reinforced Extinction – This procedure can augment expectancy violation by allowing the aversive US to follow the CS. The research showed this might not alleviate fear but can attenuate the reacquisition of fear. Personally, I would not do this to an animal.

Removal of Safety Signals – This procedure has a history of mixed success. Safety signals can act as a crutch and introject predictability. They can be removed immediately or gradually removed in phases.

Stimulus Variability – The CS is made as variable as possible (duration, proximity, etc.) and the variations presented in random order rather than graded. This helps generalize the extinction to the CS in various intensities.

Retrieval Cues – Additional cues may help the individual to trigger the extinction association. These are associated after the extinction process is complete. They may interfere with long term success by adding elements of predictability.

Multiple Contexts – It seems obvious that repeating the extinction under different circumstances might generalize the extinction, but research has not supported this.

Affect Labeling – Teaching the individual to verbally identify the emotions being evoked is helpful for humans.

Preconsolidation – Another technique sometimes used is momentarily showing the CS about 30 minutes before the extinguishing training. The results are not clear.


  • Inhibitory learning: learning that the CS no longer predicts the aversive US, where cortical influences inhibit amygdala activity.
  •  Preconsolidation: pre-session brief exposure to the CS.
  • Affect labeling: naming the emotion being experienced.


  • Are equids capable of cortical influences reducing amygdala responses? (the small cortex – large amygdala problem)
  • Does desensitization in multiple contexts provide more fear reduction in equids?
  •  Do expectancy violations help or harm desensitization in equids?
  • Could an equid be trained to indicate its emotional status with something other than its ears?
  • Could the ears be utilized to train a conscious affect labeling behavior?
  • Is there evidence for the efficacy of the relaxation component of systematic desensitization?
  • Is spontaneous recovery of fear more likely in equids?
  • Does allowing escape ever diminish fear without consolidating escape as a habitual response to the fear stimulus?

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.