Active suppression prevents the return of threat memory in humans

Intrusive traumatic memories can be reactivated by everyday cues, motivating research into how to prevent the return of aversive associative memories. Prior approaches include extinction (repeated unreinforced presentation of conditioned stimuli), cognitive reappraisal, and reconsolidation-based interventions. However, aversive memories are often encoded in specific contexts, linked to multiple sensory cues, and consolidated long before attempts at modification, leading many methods to be cue-specific. Work in declarative memory shows that conscious mnemonic control—particularly direct suppression in the think/no-think paradigm—can impair later recall, sometimes in a cue-independent manner and via mechanisms distinct from simple retrieval failure. The present study asks whether such mnemonic control principles apply to consolidated associative threat memories. Specifically, the authors tested whether active, conscious suppression during extinction reduces the later return of fear irrespective of the triggering CS, whereas thought diversion (associating CS with a neutral outcome) would yield only cue-specific reductions in fear. They further hypothesized that individual differences in thought-control ability would predict the effect of suppression, but not diversion.

The paper reviews strategies for modifying aversive associative memories: extinction, top-down reappraisal, and reconsolidation-targeted interventions (both pharmacological and behavioral). Reconsolidation updating can disrupt a reactivated CS–US trace but tends to be cue-specific, leaving other US-linked traces intact. In declarative memory, the think/no-think paradigm demonstrates that direct suppression reduces recall and can show cue-independent effects measured with independent probes, implicating executive control mechanisms and distinct neurobiology. Interference-based methods (e.g., thought diversion or substitution) can impair retrieval by engaging competing associations. Prior work links suppression-induced forgetting to prefrontal control (including DLPFC engagement) and shows deficits in clinical groups (e.g., PTSD, anxiety), suggesting relevance for pathological intrusive memories. The authors note uncertainty about whether these mechanisms extend to consolidated fear memories, which are typically more robust and contextually embedded.

Design: Two experiments used a double-cue Pavlovian fear conditioning and extinction-reinstatement-test protocol across two days. Day 1: acquisition to two CS+ (CS1+, CS2+) and one CS- (neutral). Day 2: extinction with either (Experiment 1) active suppression applied to CS1+ (CS2+ served as suppression control) or (Experiment 2) thought diversion applied to CS1+ (CS2+ served as diversion control), followed by reinstatement and test. Reinstatement occurred 30 min after extinction; testing followed shortly after (timeline indicates an 18 s gap). Participants: Healthy university students recruited from Peking University. The main experiments included Experiment 1 (active suppression; n = 28) and Experiment 2 (thought diversion; n = 27). Participants provided informed consent and received compensation. Ethical approval was obtained. Stimuli and US: Three colored squares (yellow, red, blue) served as CS1+, CS2+, and CS-. During acquisition, CS1+ and CS2+ were paired with a mild electric shock (US) on a 75% partial reinforcement schedule; CS- was never paired. Shock delivered to right wrist via isolated current source; intensity individualized (“uncomfortable but not painful,” up to 10 V), duration 200 ms at 50 Hz. Memory control manipulations:

• Active suppression (Experiment 1): During extinction, for CS1+ trials, participants were instructed to actively expel from mind what might have followed the CS (i.e., suppress US-related thoughts). CS2+ and CS- served as controls without suppression instructions.
• Thought diversion (Experiment 2): During extinction, for CS1+ trials, participants associated the CS with neutral/soothing natural scenes (mental-relief conditioning). CS2+ and CS- served as controls without diversion. Physiological measures: Skin conductance responses (SCRs) recorded from the left hand (index and middle fingers) using Biopac MP160 BioNomadix; analyzed with AcqKnowledge software. SCR amplitude was computed as peak-to-trough within 0.5–5.5 s after CS onset; values < 0.02 μS were set to zero. Only non-reinforced trials were included in primary analyses. Questionnaires: Thought Control Ability Questionnaire (TCAQ) assessed perceived ability to control intrusive thoughts (Likert-type ratings; high internal consistency reported). Procedure:
• Day 1 acquisition: Pseudorandom trial order; trial structure ~4 s CS with 10–12 s ITI.
• Day 2 extinction: As above with manipulation-specific instructions applied to CS1+.
• Reinstatement: Unsignaled US presentations.
• Test: SCRs to CS1+, CS2+, and CS- measured to assess fear recovery (fear recovery index computed as SCR difference between first test trial and last extinction trial). Statistics: Mixed-model ANOVAs tested acquisition (CS+ vs. CS-), extinction (last-trial equivalence across conditions), and fear recovery (test vs. last extinction). Key effects included within-subject factors for CS and phase, and between-subject factor for experiment (suppression vs. diversion). Associations between TCAQ and fear recovery were examined. Bayes factors were also reported for some analyses.

• Successful acquisition: Both experiments showed higher SCRs to CS1+ and CS2+ than CS- on Day 1, indicating robust conditioning (e.g., mixed-model ANOVA main effect of trial F(2,477) = 5.311, P < 0.001; post-hoc tests P < 0.001 for CS+ > CS-).
• Comparable extinction: On Day 2, extinction reduced SCRs similarly across CS types and experiments; no significant experiment or CS main effects on the last extinction trial (e.g., experiment F(1,58) = 0.001, p = 0.996). A strong phase effect confirmed reduced fear from acquisition to extinction (F(1,58) = 96.807, P < 0.001, η² = 0.646).
• Differential impact on fear return after reinstatement: • Active suppression (Experiment 1) eliminated fear recovery following reinstatement for both CSs, indicating a cue-independent reduction in physiological threat responses. • Thought diversion (Experiment 2) reduced fear recovery only for the targeted CS1+ (cue-specific), with the non-diverted CS2+ showing typical reinstatement.
• Individual differences: Thought-control ability (TCAQ) moderated the suppression effect—participants with higher TCAQ exhibited less fear recovery after active suppression for both CSs. No reliable association between TCAQ and fear recovery was found in the thought diversion condition. Overall, active suppression produced generalized attenuation of reinstated fear, whereas thought diversion was effective only for the treated cue.

The study demonstrates that consciously suppressing aversive associations during extinction can broadly prevent the return of fear after reinstatement, whereas diverting thoughts to a neutral association yields only cue-specific benefits. These findings extend mnemonic control principles from declarative memory to consolidated associative threat memories, suggesting that executive control mechanisms can reduce the reactivation of aversive responses rather than merely disrupting retrieval. The modulation of fear return by thought-control ability in the suppression condition supports the role of individual differences in executive control over intrusive content. Potential neural substrates include interactions between dorsolateral prefrontal cortex (executive control) and regions central to fear regulation (vmPFC, amygdala). The rapid observation of effects relative to known reconsolidation timescales argues against classic protein-synthesis-dependent reconsolidation as the primary mechanism, indicating that suppression may engage distinct cognitive/neural pathways. Clinically, leveraging active suppression could inform interventions for persistent maladaptive memories in disorders marked by intrusive fear (e.g., PTSD, phobias).

Active suppression during extinction prevented reinstatement of physiological fear responses in a cue-independent manner, while thought diversion provided cue-specific attenuation. Thought-control ability predicted the efficacy of suppression but not diversion. These results suggest that declarative mnemonic control mechanisms can be harnessed to reduce the return of consolidated threat memories and may offer avenues for improving treatments for fear- and anxiety-related disorders. Future research should determine the durability of suppression-induced fear attenuation over longer intervals, delineate underlying neural circuitry, and assess generalizability across populations and clinical conditions.

• Duration of effect unknown: The study tested reinstatement approximately 30 minutes after extinction; the long-term persistence of suppression-induced fear reduction was not assessed.
• Mechanisms inferred, not measured: Neural mechanisms (e.g., DLPFC–vmPFC–amygdala interactions) were discussed but not directly tested with neuroimaging or neuromodulation.
• Sample characteristics: Participants were healthy university students, which may limit generalizability to clinical populations and broader demographics.
• Some procedural/statistical details suggest typographical inconsistencies; however, primary conclusions rest on convergent patterns across two experiments.