Why Memories Change: The Neuroscience of Rewriting the Past

Summary: A new review explains how episodic memories form, persist, and change over time, showing why recollections often differ from the original event. Rather than fixed files, memories consist of multiple components—some immediately accessible, others latent until triggered by environmental cues. When recalled, these components mix with our general knowledge, past experiences, and current situation to produce an updated version of the original episode. These insights clarify common memory distortions and have practical implications for mental health, education, and legal settings where accuracy matters.

Episodic memory—the kind that lets us relive personal events like a birthday or vacation—is not a static record. This review, led by researchers at the University of East Anglia in collaboration with the University of Texas at Dallas, synthesizes evidence from psychology, neuroscience, philosophy, and animal studies to present a cohesive view of memory as a dynamic, reconstructive process.

Key Facts

  • Dynamic memories: Episodic memories are continually updated; they are not stored as perfect, immutable copies.
  • Trigger-based recall: Memory traces can remain latent and become conscious only when activated by cues in the environment.
  • Real-world impact: The reshaping of memory affects mental health treatment, educational approaches to learning and retention, and the reliability of eyewitness testimony.

Source: University of East Anglia

What the review found

The authors argue that a memory representation must be causally linked to a past event to qualify as a memory, but the recalled content often contains additional elements. Retrieving an episodic memory typically yields a conscious representation made up of: (1) information reinstated from the original event, (2) semantic or schematic knowledge, and (3) details shaped by the current context. Over repeated retrievals, especially of older memories, re-encoding processes can create a chain of changes that links the present recollection back to the original experience but also distances it from that event.

Lead author Professor Louis Renoult of UEA’s School of Psychology explains that even memories tied to real events can differ each time they are accessed. “When we recall the past, the memory we retrieve may not be a perfect copy,” he notes. “It can incorporate general knowledge, other experiences, or cues from the present situation. Over time, those reconstructions can reshape the memory.”

How the research was conducted

The review examined nearly 200 studies across psychology and neuroscience, as well as relevant philosophical work and experiments using animal models. By integrating perspectives from multiple disciplines, the team aimed to clarify unresolved questions and propose directions for new empirical research.

A central focus was the neural substrate of memory: the hippocampus and its role in forming and organizing episodic memory traces. The authors highlight the reinstatement framework, which offers a mechanism linking the original experience, the memory trace that encoded it, and the later episodic recollection. In this view, hippocampal engrams encode patterns of neocortical activity that, when reactivated, constitute the neural representation of an episodic memory.

The review emphasizes that many memory traces can remain latent—present in the brain but not consciously experienced—until a relevant cue triggers their activation. Once activated, these traces are combined with other information to create the active recollection, which explains why memories are flexible and susceptible to modification.

Key Questions Answered

Q: Why do memories change over time?

A: Each retrieval blends original memory traces with semantic knowledge and the present context, producing a revised version that can differ from the initial event.

Q: What determines whether something is a “real” memory?

A: To qualify as a memory, a representation should be causally connected to an actual past event, even if the recalled content has been altered or supplemented.

Q: Why is this research important for everyday life?

A: Recognizing memory as dynamic informs therapeutic interventions in mental health, guides evidence-based learning strategies, and highlights the limits of eyewitness testimony in legal contexts.

Editorial Notes

  • This article was edited by a Neuroscience News editor.
  • The journal paper was reviewed in full.
  • Additional context was provided by the editorial staff.

About this memory and neuroscience research news

Author: Lisa Horton
Source: University of East Anglia
Contact: Lisa Horton – University of East Anglia
Image: The image is credited to Neuroscience News

Original Research: Open access. “The cognitive neuroscience of memory representations” by Louis Renoult et al., Neuroscience & Biobehavioral Reviews. DOI: 10.1016/j.neubiorev.2025.106417


Abstract

The cognitive neuroscience of memory representations

This paper examines memory from a representational, cognitive neuroscience perspective to address ongoing empirical and theoretical challenges. Focusing on episodic memory, the authors distinguish active from latent representations and cognitive from neural forms. They adopt a causal approach: a memory representation must be linked to a past event to count as a memory. Retrieved episodic information, however, may only partially determine the content of an active representation, which typically combines reinstated details with semantic, schematic, and situational information. For remote events, repeated re-encoding likely creates a causal chain extending from the original experience to the currently accessible trace. The reinstatement framework provides a mechanistic account of how experience, encoding, and episodic recollection are linked, highlighting the role of hippocampal engrams in encoding neocortical patterns that underlie episodic memory. The paper concludes by discussing mechanisms through which memories can be modified and thereby diverge from the episodes that produced them.