Inhibitory Learning Model of Exposure Therapy
research · Craske, M. G., et al. (2008, 2014) (2008)
Exposure therapy works by creating new safety memories that inhibit old fear associations, not by erasing or habituating the original fear.
Core Concepts
The Problem
Traditional habituation models of exposure therapy failed to explain why fear often returns after treatment, even when habituation had occurred.
The Claim
Exposure works primarily through expectancy violation: the client expects a terrible outcome, it doesn't happen, and the brain stores a new inhibitory memory that competes with the original fear memory.
Key Evidence
- •Clinical trials showing that exposure outcomes correlate more strongly with expectancy violation than with within-session habituation
- •Neurobiological evidence that extinction involves new learning in the ventromedial prefrontal cortex, which inhibits the amygdala's fear output
- •Greater context generality from random-vs-blocked exposure schedules supports an inhibitory learning account
Practical Implication
To break avoidance loops, one must willingly face feared situations and stay with the discomfort until the feared outcome fails to materialize. The new safety learning then competes with the old pattern, allowing flexibility.
Nuance & Limits
The new inhibitory memory must be strong enough to outcompete the old fear memory; contextual cues, task variety, and therapeutic alliance all influence strength. There is also risk that stress or novel contexts can temporarily release the fear if the new memory isn't robustly encoded.
Source Material
Citation Density
Growing body of clinical and neuroscience research, forming the dominant theoretical framework for exposure therapy today
Related Ideas
Inhibitory learning is a direct application of predictive processing: expectancy violation is prediction error, and the new safety memory is a competing internal model.
Gaps
- ⚠ How to maximize the strength and generalization of the new inhibitory memory in brief clinical settings
- ⚠ Individual differences in fear learning and extinction are still poorly understood
Discuss Further
Open this concept in an AI assistant for deeper discussion, critique, or exploration.