Can You Really Trust What You Remember? Understanding Memory Reconsolidation
A clinical research review prepared for the Rachmanas editorial library.
Rachmanas Counseling 
Psychological counseling, at Rachmanas or anywhere else, often rests on an unspoken assumption, that a memory recalled with confidence and emotional intensity is also an accurate one.
Since 2000, behavioral neuroscience has complicated that assumption. When an established memory is reactivated, it appears to re-enter a temporarily labile state before being restabilized, a process now called reconsolidation (Nader, Schafe, & LeDoux, 2000).
This review summarizes the animal and human evidence for reconsolidation, traces its uneven translation from rodent fear conditioning toward clinical application (Schiller et al., 2010; Chalkia, Van Oudenhove, & Beckers, 2020), and situates the phenomenon within a broader account of memory as reconstructive rather than reproductive (Schacter, 1999).
It closes by connecting this evidence to two situations common in clinical practice: partners who recall the same disagreement in incompatible ways, and theoretical models that treat reconsolidation as a mechanism of therapeutic change (Lane, Ryan, Nadel, & Greenberg, 2015).
Throughout, findings that are well replicated are distinguished from claims that remain contested.
The Development
A memory recalled in detail feels self-authenticating. The scene returns with its original charge, the words that were used, the tone of voice, the sequence of what was said and by whom. Confidence and vividness are typically treated, by the person remembering and by the people around them, as evidence of accuracy.
For most of the twentieth century, cognitive psychology offered a model of memory consistent with that intuition. Consolidation theory held that once a memory had stabilized, typically within hours to days of encoding, it became a fixed trace that could be retrieved but not further modified except through new learning (McGaugh, 2000).
Retrieval, on this account, functioned something like opening a sealed file. The record could be incomplete or degraded, but the act of reading it did not change its contents.
A series of experiments beginning in 2000 overturned that assumption for at least one well studied form of memory. What began as a finding about fear conditioning in rats has since grown into one of the more contested and productive research programs in behavioral neuroscience, one with direct implications for how clinicians think about the memories their clients bring into the room.
The Discovery of Reconsolidation
Nader, Schafe, and LeDoux (2000) trained rats to associate a tone with a mild foot shock, producing a stable, consolidated fear memory. Under the standard consolidation model, infusing a protein synthesis inhibitor into the amygdala at this point should have had no effect, since consolidation was already complete.
That is not what the researchers found. When the inhibitor anisomycin was infused into the lateral and basal amygdala shortly after the rats were briefly re-exposed to the tone, the animals lost their fear response on later testing.
The same infusion given without that reminder left the memory intact, and infusion delayed until six hours after the reminder also left the memory intact.
The pattern indicated that a consolidated memory, once retrieved, re-enters a state that again depends on new protein synthesis before it restabilizes. Nader and colleagues termed this second stabilization process reconsolidation.
Nader’s own follow-up review noted that the effect had already been replicated and extended across a range of species, from rodents to the garden slug Limax, and across both appetitive and aversive learning, although not every attempted replication succeeded (Nader, 2003).
That mixture of successful replication and occasional failure has remained characteristic of the field ever since.
From Rodent Fear Memory to Human Recollection
Reconsolidation was discovered using a controlled paradigm involving a single, simple association. Its relevance to the far more layered, narrative memories that structure human relationships is a separate empirical question, and the evidence here is considerably less settled.
The most influential attempt to translate the finding into human, non-pharmacological terms came from Schiller et al. (2010). Participants were conditioned to fear a visual stimulus, then given a brief reminder of that stimulus ten minutes before undergoing standard extinction training.
In the original report, this group showed no return of the fear response at follow-up testing conducted as much as a year later, while a group that underwent extinction alone, or extinction six hours after the reminder, did show its return.
The authors interpreted this as evidence that the reminder had opened a reconsolidation window in which the old fear association was updated with new, non-threatening information, rather than merely suppressed through the separate process of extinction learning.
The study was cited widely as evidence that reconsolidation could be harnessed clinically through timing alone. It has also proven difficult to reproduce.
Chalkia, Van Oudenhove, and Beckers (2020) obtained the original data set for a planned direct replication and, in the process, identified inconsistencies in how participants had been excluded from the original analysis and in the way the statistics had been reported; their reanalysis did not reproduce the central finding.
The same research group subsequently ran an independent, pre-registered replication of the full procedure and again found no lasting reduction in fear responding (Chalkia et al., 2020). Schiller and colleagues have disputed the reanalysis and maintain that their original conclusions hold.
The status of the retrieval-extinction effect in humans remains, at the time of writing, an open and actively debated question.
This does not put reconsolidation itself in doubt. The underlying mechanism is well supported at the level of animal neurobiology. What it complicates is a single, widely publicized route to harnessing that mechanism behaviorally in people.
A comprehensive review of the human reconsolidation literature found a substantial number of results consistent with reconsolidation alongside a substantial number of inconsistent findings and plausible alternative explanations, and concluded that its presence cannot yet be conclusively inferred at the neurobiological level from human behavioral studies alone (Elsey, Van Ast, & Kindt, 2018).
That is the appropriately cautious summary of where the field currently stands.
Divergent Recall of Shared Events: A Reconstructive Account
Reconsolidation offers one route by which a memory can change with each retrieval. A longer established line of research explains why two people can retrieve the same event and arrive at genuinely different, internally coherent accounts, without either person being dishonest.
Schacter (1999) organized the routine ways that memory departs from an accurate record into seven categories. Three are directly relevant to relational conflict.
Bias describes the tendency for present mood, knowledge, and self-concept to color how the past is recalled, so that a person who currently feels unappreciated will tend to retrieve past incidents in ways consistent with that feeling, regardless of how the incident was experienced at the time.
Suggestibility describes the incorporation of information encountered after an event into the memory of the event itself, so that a later conversation or a friend’s interpretation can become woven into the recollection without the person’s awareness that this has occurred.
Misattribution describes confusion about the source of a memory, including mistaking something imagined, feared, or discussed for something that actually happened.
None of these processes requires reconsolidation to operate. Reconsolidation does, however, offer a plausible mechanism for how they take hold.
If a memory becomes labile each time it is retrieved and rehearsed, repeated retrieval under a particular emotional frame, rehearsing an argument while feeling wronged, for example, provides repeated opportunity for that frame to become incorporated into the stored memory itself.
Two partners who each rehearse the same disagreement from different emotional vantage points are not simply interpreting a shared, fixed record differently. Each act of retrieval may be modestly reshaping the record.
Reconsolidation as a Candidate Mechanism of Psychotherapeutic Change
The clinical interest in reconsolidation extends beyond explaining memory distortion. Lane, Ryan, Nadel, and Greenberg (2015), in a target article for Behavioral and Brain Sciences, proposed that reconsolidation supplies a common mechanism underlying change across otherwise dissimilar treatment models, including behavioral, cognitive-behavioral, emotion-focused, and psychodynamic therapies.
Their model specifies three components: an old emotional memory must be reactivated, a new emotional experience sufficiently different from what the old memory would predict must occur while the memory is labile, and the updated memory must then be stabilized through repeated practice across varied contexts.
The proposal is a theoretical synthesis rather than a demonstrated mechanism. As a Behavioral and Brain Sciences target article, it was published alongside dozens of peer commentaries, several of which questioned whether the model generated predictions distinct from existing accounts of therapeutic change, and whether emotional arousal was as necessary an ingredient as the authors proposed. The proposal remains useful for a specific reason.
It gives clinicians a falsifiable, mechanistic vocabulary for something therapists have long done by intuition: bringing a difficult memory into the room in a way that is emotionally alive rather than merely described, and then supplying an experience that does not confirm what the memory predicted.
Boundary Conditions and Open Questions
Reconsolidation does not appear to occur automatically every time a memory is retrieved. Several conditions have been proposed for when a reactivated memory does, and does not, destabilize. A mismatch between what the reactivating cue predicts and what actually happens, a prediction error, appears necessary in several paradigms, though not sufficient by itself.
The strength and age of the original memory matter, with some studies suggesting that older or more strongly consolidated memories are harder to destabilize than recently formed ones.
The duration of reactivation matters as well: a reminder that is too brief may fail to destabilize the memory, while one that runs too long can trigger extinction learning instead, a separate process that suppresses a response without necessarily altering the underlying trace (Elsey et al., 2018).
These boundary conditions are not yet settled with precision, and different laboratories have reported conflicting results using superficially similar procedures. That inconsistency is itself informative. It suggests that reconsolidation, where it occurs in humans, is a narrower and more conditional phenomenon than the popular description of memory as endlessly rewritable would suggest.
Implications for Practice and for Self-Reflection
Two conclusions follow from this body of work, and they are worth keeping separate.
The first is mechanistic and still provisional. Certain memories, retrieved under certain conditions, may be genuinely open to modification through subsequent experience, and this may be part of what makes some forms of psychotherapy effective.
The strength of the evidence varies considerably by paradigm, from well established in rodent fear conditioning to actively disputed in the human retrieval-extinction literature.
Clinicians have good reason to treat the more ambitious clinical claims, that any painful memory can be reliably neutralized through a brief, well-timed intervention, with corresponding caution.
The second conclusion is psychological and considerably better established. Memory is a reconstructive process, shaped at each retrieval by current mood, subsequent information, and self-relevant belief (Schacter, 1999). A memory can be recalled with complete confidence and still be a partial or altered account of what occurred.
This has a practical corollary for anyone examining a painful or contested memory, in a therapy session or independently.
The strength of a recollection is not, by itself, evidence of its accuracy, and a memory’s meaning is usually more tractable than its literal content.
Before treating a remembered incident as a fixed verdict on a relationship or on oneself, it can help to ask what has been added to that memory since it happened, what current mood or belief the recollection now serves, and what would have to be true for a different, equally sincere account of the same event to also be correct.
References
- Chalkia, A., Schroyens, N., Leng, L., Vanhasbroeck, N., Zenses, A.-K., Van Oudenhove, L., & Beckers, T. (2020). No persistent attenuation of fear memories in humans: A registered replication of the reactivation-extinction effect. Cortex, 129, 496–509. https://doi.org/10.1016/j.cortex.2020.04.017
- Chalkia, A., Van Oudenhove, L., & Beckers, T. (2020). Preventing the return of fear in humans using reconsolidation update mechanisms: A verification report of Schiller et al. (2010). Cortex, 129, 510–525. https://doi.org/10.1016/j.cortex.2020.03.031
- Elsey, J. W. B., Van Ast, V. A., & Kindt, M. (2018). Human memory reconsolidation: A guiding framework and critical review of the evidence. Psychological Bulletin, 144(8), 797–848. https://doi.org/10.1037/bul0000152
- Lane, R. D., Ryan, L., Nadel, L., & Greenberg, L. (2015). Memory reconsolidation, emotional arousal, and the process of change in psychotherapy: New insights from brain science. Behavioral and Brain Sciences, 38, e1. https://doi.org/10.1017/S0140525X14000041
- McGaugh, J. L. (2000). Memory: A century of consolidation. Science, 287(5451), 248–251. https://doi.org/10.1126/science.287.5451.248
- Nader, K. (2003). Memory traces unbound. Trends in Neurosciences, 26(2), 65–72. https://doi.org/10.1016/S0166-2236(02)00042-5
- Nader, K., Schafe, G. E., & LeDoux, J. E. (2000). Fear memories require protein synthesis in the amygdala for reconsolidation after retrieval. Nature, 406(6797), 722–726. https://doi.org/10.1038/35021052
- Schacter, D. L. (1999). The seven sins of memory: Insights from psychology and cognitive neuroscience. American Psychologist, 54(3), 182–203. https://doi.org/10.1037/0003-066X.54.3.182
- Schiller, D., Monfils, M.-H., Raio, C. M., Johnson, D. C., LeDoux, J. E., & Phelps, E. A. (2010). Preventing the return of fear in humans using reconsolidation update mechanisms. Nature, 463(7277), 49–53. https://doi.org/10.1038/nature08637
