Spreading the reduction of fear: A systematic review on generalization of extinction learning in human fear conditioning

Extinction learning involves repeatedly presenting a conditioned stimulus (CS) that previously signaled a threat, but now occurs without the expected threat. This process is core to exposure-based treatments, a key treatment for anxiety-related disorders. Enhancing the generalization of extinction learning is crucial for improving treatment outcomes, as it helps reduce fear across a range of stimuli resembling the original fear evoking CS. This systematic review covers three generalization of extinction learning models: 1) generalization of CS extinction learning, examining how extinction learning to the CS generalizes to novel generalization stimuli (GSs), 2) generalization of GS extinction learning, assessing how extinction learning to a GS generalizes to other novel GSs or the original CS, and 3) generalization of multiple stimuli extinction learning, where extinction learning involves multiple GSs (and sometimes the CS) and its effect on other novel stimuli. Most findings indicate that extinction learning to the CS effectively generalizes to other stimuli. However, extinction learning to a GS or multiple GSs hardly generalizes to other novel GSs or the original CS. Since real-life exposure therapy rarely reproduces the CS, GS extinction learning may better reflect clinical practice; poor generalization of GS(s) extinction learning may constitute another pathway of return of fear. This review also highlights various factors that either hinder or promote the generalization of extinction learning across all three models. Future research should focus on identifying more strategies to enhance the generalization of extinction learning, which can help inform exposure-based treatments.