Dissertations, Theses, and Capstone Projects

Date of Degree


Document Type


Degree Name



Cognitive Neuroscience


David Johnson

Committee Members

Tatiana Emmanouil

Tony Ro

Subject Categories

Cognitive Science | Social and Behavioral Sciences


reconsolidation update, reconsolidation interference, post-retrieval extinction, threat memory, fear memory, meta-analysis


Threat memories are created when an aversive stimulus is paired with other stimuli such that the associated stimuli elicit the same reaction as the unpleasant one. Anxiety disorders are usually treated with exposure to aversive stimuli. Emotional memories can be altered through a process called reconsolidation interference. Consolidated memories are dynamic: reminders of previous experiences prompt the reactivation of a previously-encoded memory; destabilization of the memory opens it to update. While a memory is updating, it is labile. If extinction training occurs during reconsolidation (i.e., shortly after reactivation), the conditioned stimulus will no longer elicit the conditioned response. Thus, previously-encoded threat memories can be weakened through the process of reconsolidation. Maximizing the long-term positive effects of reconsolidation interference by delivering extinction training during the reconsolidation window has the potential to revolutionize the treatment of anxiety disorders. Yet, the results of studies that behaviorally manipulate reconsolidation have proven difficult to replicate. Kredlow and colleagues’ (2016) comprehensive meta-analysis on aversive and appetitive memory reconsolidation in rodents and humans showed a small-to-moderate effect for the interference of human threat-memory reconsolidation (g = 0.40, n = 16, p < .01). The current paper is a verification and expansion of Kredlow and colleagues’ (2016) analyses on post-reactivation extinction of human threat memories. The analysis verifying the original study failed to replicate the overall effect (g = 0.3144, n=16, p = 0.0646), though moderators were similar in terms of impact on effect and statistical significance; specifically, the number of acquisition trials and inclusion of expectancy-of-aversive-outcome ratings were significant moderators on effect. With a larger number of studies (n = 26), the expanded analysis yielded a smaller, albeit significant (p = 0.0138), overall effect (g = 0.2923). Across all three analyses, use of between-subject design and threat-irrelevant stimuli were significant moderators. The expanded study failed to show that the number of acquisition trials and inclusion of expectancy ratings were significant moderators. A post-hoc power analysis indicated that all included studies were underpowered by virtue of sample sizes; thus, the effect may not be accurately captured by studies with small sample sizes. Further, a trim-and-fill analysis for publication bias revealed moderate heterogeneity among the included effect sizes and suggested that six effect sizes were missing from the expanded sample.

This work is embargoed and will be available for download on Monday, June 02, 2025

Graduate Center users:
To read this work, log in to your GC ILL account and place a thesis request.

Non-GC Users:
See the GC’s lending policies to learn more.