Reactivation of threat conditioning memory in humans: disentangling the effects on emotional memory and cognitive biases.

Learning to detect and respond to threats is fundamental for survival and is often modeled through threat conditioning (TC) paradigms. While these paradigms reliably produce implicit memories that elicit physiological and behavioral responses to conditioned stimuli (CS), less is explored about how TC influences cognitive and emotional biases, particularly those implicated in anxiety disorders, such as threat overestimation and negative stimulus representation. In this study, we investigated the dynamic interaction between the reactivation of the implicit threat memory and these cognitive biases using a validated TC paradigm in humans. In Experiment 1, participants underwent TC on Day 1, followed by a memory reactivation session (incomplete reminder: one unreinforced CS+) and a highly demanding working memory (HWM) task, used as an amnesic manipulation, or a control condition on Day 2. On Day 3, memory retention was tested using a simplified, single-trial protocol (one CS+, one CS-, and one neutral CS), followed by tasks assessing threat valuation and representation. Results indicated that the HWM task administered post-reactivation significantly reduced skin conductance responses (SCRs) and attenuated cognitive biases, without altering expectancy of the unconditioned stimulus (US). In Experiment 2, we evaluated the effect of varying reactivation frequency (none, one, or two reminders) on implicit memory and cognitive biases. While repeated reactivations generalized the conditioned response to other stimuli, cognitive and emotional biases remained stable, suggesting a dissociation between memory generalization and evaluative processing. These findings demonstrate that implicit threat memories can be selectively modified through post-reactivation interventions, affecting both physiological and cognitive-emotional domains. Importantly, the distinct effects of memory reactivation and reconsolidation on physiological versus cognitive outcomes support the existence of temporally and functionally dissociable mechanisms. This research highlights the need to consider cognitive biases alongside physiological responses when evaluating memory-based interventions and offers novel insight into mechanisms underlying anxiety maintenance and treatment.