You remember a summer evening by a lake. The dock, the sunset, and the still water feel clear. Years later, a red canoe appears in the scene. Was it always there, or did a photograph, a conversation, or a later visit add it?
Memory feels like playback, but the brain does not keep a perfect recording. Remembering is an active reconstruction. Under some conditions, retrieving a memory can make it temporarily open to change before it becomes stable again. Researchers call this process reconsolidation.
Memory is not playback
When you remember an event, the brain rebuilds it from stored details, general knowledge, emotion, and the present situation. This is useful. A flexible memory system can connect old experience with new facts and help behaviour adapt when the world changes.
The same flexibility creates errors. Details from another event can enter the reconstruction. A question can change which features become prominent. Later information can compete with, cover, or become linked to what was learned first.
These effects do not all prove reconsolidation. Memory can change through ordinary new learning, interference, source confusion, suggestion, or forgetting. Reconsolidation is a narrower claim: after a consolidated memory is reactivated, it becomes unstable enough to require a new period of stabilization.
Consolidation and reconsolidation
A newly formed memory is initially fragile. Cellular changes stabilize it over time, while longer processes reorganize how it is represented across brain systems. This is consolidation.
For much of the twentieth century, researchers often treated a consolidated memory as permanently fixed. It could become harder to retrieve, but the stored trace itself was assumed to be protected. Earlier experiments had raised doubts, yet the modern reconsolidation field accelerated in 2000.
Karim Nader, Glenn Schafe, and Joseph LeDoux trained rats to associate a tone with a mild shock. After the fear memory had become stable, a tone reactivated it. The researchers then infused a protein-synthesis inhibitor into the amygdala. On a later test, the rats showed much less learned fear. The same drug did not have that effect when given without reactivation.
The interpretation was striking: retrieval had returned the established memory to a vulnerable state, and new protein synthesis was needed to store it again. This did not mean the animals consciously relived a detailed event, or that a drug literally deleted a complete memory. The experiment measured a learned fear response in a controlled animal model.
What happens to human event memories?
Human episodic memory concerns events situated in a particular context, such as what happened at a birthday or what objects appeared in a room. Researchers cannot use the invasive methods used in many animal experiments, so they rely on carefully timed reminders and new learning.
In a 2007 study, Almut Hupbach and colleagues asked college students to learn a list of objects. The next day, one group received a subtle reminder of the first session before learning a second list. Another group learned the second list without that reminder. On the third day, reminded participants were more likely to insert objects from the second list when recalling the first.
The effect was delayed. When researchers tested the original list immediately after the second list, the same integration did not appear. That time dependence was consistent with a process that required restabilization rather than simple confusion during the test.
Jason Chan and Jessica LaPaglia reported another human result in 2013. Participants watched an episode of a television program. Some later answered questions that reactivated specific details before hearing an audio summary containing altered information. On a final test, reactivated details were more vulnerable to the later misinformation.
Yet this pattern is not automatic. A later study by Jillian Wilkes and colleagues did not find that reactivating a witnessed event consistently increased impairment from subsequent misinformation. Differences in reminder structure, final test, prediction error, and memory strength may help explain why studies diverge.
A 2017 meta-analysis examined human episodic-memory experiments rather than relying on one demonstration. It found that reactivation followed by interference could change later retrieval, and that new learning after reactivation reliably increased intrusions into an older memory. Effects varied considerably across methods and tended to be larger for remote memories and narrative material.
Why surprise may open the memory
Simply recalling a familiar fact may retrieve it without destabilizing it. A leading idea is that reconsolidation depends on prediction error, a mismatch between what a memory leads the brain to expect and what actually happens.
Imagine hearing the opening notes of a song that always continues in the same way. If it stops unexpectedly, the mismatch signals that the old model may need updating. The brain has a reason to make the memory flexible.
Dieuwke Sevenster, Tom Beckers, and Merel Kindt studied this boundary in human fear learning. Their experiments found that the amount and timing of unexpected information helped determine whether retrieval was followed by reconsolidation, ordinary retrieval, or new learning. Too little mismatch may leave the memory stable. A large change may create a separate memory instead of updating the old one.
Prediction error is useful evidence, but it is not a unique signature. Surprise also drives attention and new learning. Researchers must still show that an effect depends on reactivation, occurs within an appropriate time window, and selectively changes the earlier memory.
The debate over changing fear memories
Fear conditioning provides a simple way to test memory updating. A neutral image or sound is paired with an unpleasant event. Later, skin conductance or another physiological response indicates whether the learned threat association returns.
In 2010, Daniela Schiller and colleagues reported that extinction training given shortly after a reminder prevented several forms of fear return. Extinction means repeatedly presenting the cue without the expected unpleasant outcome. When this occurred within the proposed reconsolidation window, the old fear response appeared to be durably updated rather than temporarily suppressed.
The result attracted attention because it suggested a drug-free way to alter emotional learning. It also attracted scrutiny. Follow-up studies produced mixed outcomes, and the original paper later received an addendum clarifying how participants were excluded. Other experiments found no evidence that reminder-before-extinction prevented a threat memory from returning.
This does not erase the entire reconsolidation field. It shows that translating a laboratory mechanism into a reliable human procedure is difficult. Fear memories vary in age, strength, context, and structure. Small changes in the reminder can decide whether the old association becomes unstable.
What changes when a memory is updated?
Several different outcomes can look like a changed memory.
- Content updating: New details become integrated with an older event.
- Strengthening: Retrieval followed by successful restabilization can make later recall stronger.
- Weakening: Interference during a vulnerable period can reduce later expression.
- Changed emotional response: A cue can remain recognizable while the learned fear response becomes smaller.
- Retrieval competition: A newer memory can block access to the original without changing the original trace.
Behaviour alone does not always reveal which mechanism occurred. A person may fail to report a detail because it was altered, because another detail competes with it, or because the test does not provide an effective cue. This is one reason researchers avoid saying a memory was erased unless the evidence rules out recovery in new contexts or later tests.
Does every recall rewrite the memory?
No. The popular phrase that every act of remembering rewrites the past is too strong.
Reactivation is necessary in many reconsolidation experiments, but it may not be sufficient. Strong or repeatedly rehearsed memories can resist destabilization. Older memories can sometimes be more open to interference, yet other studies find that strong training narrows the conditions under which change occurs. Reminder duration, context, expectation, sleep, stress, and the interval before new information all influence the result.
A recent 2025 review of reconsolidation timing concluded that retrieval can initiate overlapping molecular processes linked to change and persistence. Researchers still do not know the exact duration or boundaries of the window in different memory systems.
It is safer to say that remembering creates an opportunity for updating under particular conditions. Most everyday recall probably combines some stable structure with a reconstruction shaped by the present. Whether the stored representation itself enters reconsolidation cannot be inferred from the feeling of remembering.
Why confidence does not guarantee accuracy
A changed memory can still feel vivid and sincere. Confidence tells us how strongly a person endorses a recollection, not whether every detail matches the original event.
Research on the misinformation effect shows that misleading post-event information can influence later reports. In a 1978 series involving more than 1,200 participants, Elizabeth Loftus and colleagues showed how verbal information presented after an accident sequence could become integrated with visual memory. A large contemporary replication published in 2025 again found robust susceptibility to misinformation, while suggesting that demand effects and retrieval competition explain part of the pattern.
Misinformation research and reconsolidation research overlap, but they are not identical. A person can report misinformation because of social expectations, source confusion, or response strategy without the original memory being rewritten. Strong experiments compare reactivated and non-reactivated conditions and use delayed tests to isolate post-retrieval change.
What this means for identity
Autobiographical memory helps construct a continuous self. If memories are flexible, personal identity is not a perfect archive of earlier experience. It is an organized model that preserves some features, loses others, and updates meaning over time.
That flexibility is not merely a flaw. A system that never changed would keep treating outdated predictions as current. Reconsolidation may allow knowledge to remain useful as environments, relationships, and goals change.
The idea also requires restraint. Laboratory studies of object lists or conditioned fear do not show that a person can intentionally rewrite any life event. They do not provide a do-it-yourself method for changing traumatic memory. Clinical treatments can reduce distress or alter responses through several mechanisms, and reconsolidation is only one candidate explanation.
What researchers still need to learn
Future studies need reliable markers that show when a human memory has actually destabilized. Prediction error is promising, but it is not exclusive to reconsolidation. Brain imaging, physiology, and computational models may help distinguish updating of an old representation from formation of a competing new one.
Researchers also need larger preregistered studies that repeat the same protocols across laboratories. Mixed results are informative when methods are reported clearly. They reveal the boundaries of the phenomenon rather than turning one dramatic finding into a universal rule.
The evidence supports a dynamic view of memory. Recall can make old learning open to new information, but the brain controls that opening carefully. Memory must be stable enough to preserve experience and flexible enough to keep experience useful.
Common questions
Can remembering something create a false detail?
Yes. New information can become mixed with an older event, especially when sources are similar or the original detail is uncertain. That does not mean every error is caused by reconsolidation.
Does talking about a memory change it?
Conversation can emphasize details, introduce suggestions, and influence later reconstruction. Whether it destabilizes and reconsolidates the original trace depends on conditions that cannot be identified from conversation alone.
Can a traumatic memory be erased?
Current evidence does not support a reliable procedure that selectively erases a complex autobiographical memory. Treatments can reduce distress and change responses, but claims of guaranteed deletion go beyond the evidence.
Are vivid memories more accurate?
Vividness and confidence can accompany accurate memories, but neither guarantees accuracy. Emotion, rehearsal, and later information can strengthen the feeling of certainty.
Why would the brain make memories editable?
Updating lets old knowledge incorporate new conditions. A flexible memory can correct predictions and guide future behaviour when the world changes.
Sources and further reading
Nader, Schafe, and LeDoux, 2000, Nature. Foundational animal experiment on fear-memory reactivation and protein synthesis.
Hupbach and colleagues, 2007, Learning & Memory. Human episodic-memory updating after a subtle reminder.
Chan and LaPaglia, 2013, PNAS. Retrieval followed by new learning in human declarative memory.
Wilkes and colleagues, 2017, Journal of Memory and Language. Failure to find greater misinformation impairment after reactivation.
Scully, Napper, and Hupbach, 2017, Neurobiology of Learning and Memory. Meta-analysis of reactivation-induced episodic-memory change.
Schiller and colleagues, 2010, Nature. Human reminder-extinction study and long-term fear-return tests.
Sevenster, Beckers, and Kindt, 2013, Science. Prediction error as a boundary condition for human fear-memory reconsolidation.
Loftus, Miller, and Burns, 1978, Journal of Experimental Psychology. Post-event verbal information and visual memory.
Wiechert and colleagues, 2025, Psychological Science. Contemporary replication and analysis of the misinformation effect.
Bayer and colleagues, 2025, Neuroscience & Biobehavioral Reviews. Review of the timing and molecular boundaries of reconsolidation.
