The Effect of Extinction Cues on Response Recovery: A Meta-Analysis

• Published in: Psykhe Vol. 31; no. especial; pp. 1 - 12
• Main Authors: Alfaro, Felipe, San Martín, Consuelo, Laborda, Mario A., Míguez, Gonzalo
• Format: Journal Article
• Language: English; Portuguese
• Published: Pontificia Universidad Católica de Chile 2022
• Subjects: extinction cue; meta-analysis; pavlovian conditioning; PSYCHOLOGY, MULTIDISCIPLINARY; response recovery; Pavlovian conditioning; response recovery; clave de extinción; extinction cue; meta-analysis; metaanálisis; condicionamiento pavloviano; recuperación de respuesta
• ISSN: 0718-2228; 0718-2228
• DOI: 10.7764/psykhe.2021.38063

A cue presented during Pavlovian extinction may help the recovery of the extinction memory, for which is called an extinction cue (EC). Pavlovian conditioning has been useful as a model of different behavior disorders and extinction as a model for their treatment. Extinguished responses may be recovered under different circumstances, akin to a relapse. Hence, it is important to strengthen extinction memory retrieval. There is contradictory evidence of the effectiveness of ECs to this end. There is also little information about the magnitude of response recovery prevention when using ECs. The magnitude of the ECs effect on response recovery was analyzed by a meta-analysis that considered possible sources of variance in the EC effect. The included studies were gathered mainly through scientific database search engines. Selection criteria included experiments that used a Pavlovian extinction and recovery procedure with an EC test. Effect size was calculated for each relevant experiment. Thirty-seven studies were included. These analyses showed that there is a robust effect of an extinction cue in reducing response recovery, d = 0.71, 95% CI [0.58, 0.85]. This effect is higher when a spontaneous recovery procedure is used and when the experiment is done with non-human animals. Interestingly, the type of control group did not affect the effect size. These results are robust under different statistical analyses, although a publication bias was detected.