Beyond Recall: How Memory Shapes Tomorrow

Abstract

Memory serves a crucial role beyond mere storage or recollection of the past; it is profoundly future-oriented. This theoretical review synthesizes peer-reviewed studies, including neuroimaging (fMRI, PET) and behavioral data, to examine two fundamental features of the human memory system. Firstly, memory storage is not centralized but rather distributed, i.e., memory traces are represented by distributed neural patterns across brain regions and distributed throughout the brain. Secondly, memory is not fixed but it is constructed; rather than a literal recall of the past, the brain builds upon previous beliefs and experiences to reconstruct the experience. These features render the brain susceptible to various errors, such as source memory failure (falsely attributing the source of a particular memory) and gist-based distortions (failing to recall distinctive item-specific information). This paper argues that this error-prone nature of memory is not an unfortunate result of an imperfect human brain, but rather a fundamental feature of the human memory system that allows it to prepare for the future. Drawing upon the Constructive Episodic Simulation Hypothesis—as well as the functional MRI (fMRI) and the positron emission tomography (PET) data in support of it—the paper elucidates how the distributed and constructed nature of memory enables us to “pre-experience” the future. Adaptive forgetting complements this by prioritizing the retention of useful information. Together, these mechanisms allow memory to guide decision-making and shape future cognition, with implications for education and clinical interventions.

Introduction

The Distributed Nature of Memory

Memory storage in the hippocampus does not follow a simple accumulation process. Instead, during memory consolidation—the process of stabilizing and distributing memory traces across brain regions—experiences are fragmented into neural patterns and dispersed across areas like the neocortex for long-term storage¹. This process involves strengthening synapses, forming patterns of neural activity that encode memories². Over time, the hippocampus gradually spreads memory to other brain regions, such as the neocortex, for long-term storage¹. For better understanding, see Figure 1, where newly encoded memory is fragmented into smaller patterns, is dispersed across the brain, and strengthens its connection, the synapse.

Theories of memory build on this distributed framework. Donald Hebb’s principle—“neurons that fire together, wire together”—explains how partial activation retrieves entire memories⁴. Howard Eichenbaum’s “memory trace” concept describes memories as distributed neural patterns⁵., while Lynn Nadel’s multiple trace theory posits that each retrieval creates a new trace, linking dispersed fragments¹. These theories highlight memory’s flexibility, setting the stage for its future-oriented role.

The Challenges of The Distributed Nature of Memory

This distributed storage, while efficient, introduces challenges in reconstructing coherent memories. The dispersed nature of memory requires the brain to exert a great deal of energy and solve a series of significant challenges to recreate a past experience. That is, because memory is distributed through the brain, a variety of distinct features that constitute an episode need to be tied together¹. Most simply put, the process looks like the following. First, multiple regions and neural networks in our brain involved with the specific past experience are coordinated together. Then, the brain integrates information from a variety of sensory modalities, such as visual, auditory, and emotional information, in order to form a coherent memory⁶. For example, when we remember an event, like our birthday party, our brain integrates information from different sources such as the people we interacted with, the food we ate, and the emotions we experienced. This process requires a significant amount of cognitive resources, including attention, working memory, and other executive functions^6,7. These tasks are significant challenges; there would have been no need to bind together constituent elements of a past memory if it was stored cohesively rather than in a dispersed manner.

After the binding process, there is another challenge. The memory system needs to keep these episodes separate from one another⁸. This may actually be a greater feat than the foregoing as it requires the brain to make fine distinctions between similar experiences and create

non-overlapping distinct representations of each experience in the memory system. The process of pattern separation also requires a great deal of cognitive resources and can be particularly challenging in situations where memories are very similar or where there is a high degree of interference between memories⁹.

Our brain frequently fails in the process, causing numerous errors. First, failure to link constituent elements of an episode gives rise to a problem often referred to as ‘source memory failure,’ which leads to falsely attributing the source of a particular memory, or even failing to remember when or how that particular memory was acquired⁸. Failure in the second task is also critical. When episodes overlap too much but are not encoded with intervals, we may end up remembering overlapping details but fail to recall distinctive item-specific information. These are the ‘gist-based distortions’ that have been noted since Bartlett^2,10.

In the book “The Seven Sins of Memory: How the Mind Forgets and Remembers”, Daniel Schacter classifies the numerous errors, illusions, and distortions of memory into the following seven ‘sins’: transience, absent-mindedness, blocking, misattribution, suggestibility, bias, and persistence¹⁰. These sins from forgetting to confusing reality are constituents of our daily lives. Because memory is saved in a dispersed fashion, we end up with so many errors. To restate, these errors aren’t just found in individuals with deficient brains^8,9,11. They are the products of a system that needs each episode to be unpacked so that it can be repackaged for another purpose in the future. All this may seem puzzling from an evolutionary perspective: Why hasn’t the human brain evolved to record the past more accurately? Why adopt an error-prone system? Evidence suggests these errors serve adaptive functions, such as prioritizing survival-relevant information¹².

The constructive episodic simulation hypothesis provides an answer to this puzzle. Building on the distributed nature of memory, it posits that if the goal of memory was solely to record the past accurately, it wouldn’t be constructed this way. Our memory system is designed as such—in full recognition of the potential for error—because it needs to play an adaptive function^13,14. We may think of memory as a simple storage functionality, but memory serves another, perhaps more important role. Memory is essential to our decision-making in the present and future. This distributed aspect of memory is essential for our actions oriented toward the future. We would not be able to efficiently piece together past experiences when making decisions for the future if they weren’t encoded in fragments and distributed throughout the brain. In other words, memory’s error-prone design, co-opted for flexibility, enables future simulation, as supported by studies on animal memory showing similar fragmentation¹⁵. This is why one can perhaps say that memory is hypothesized to be future-oriented, supported by neuroimaging and behavioral evidence.

The Constructive Episodic Simulation Hypothesis

Foundations of CESH

Bartlett and others have consistently pointed out that memory is constructive. Bartlett argued that memory is influenced by a person’s prior knowledge, beliefs, and expectations and that one uses these factors to fill in gaps in their memory to create a coherent story or narrative^2,10. For example, in one of his famous studies, Bartlett asked participants to read a Native American folk tale and then recall it from memory several times over a period of weeks or months. He discovered that as time went on, the participants’ recall of the story became distorted; they adapted to fit pre-existing knowledge and cultural expectations. In particular, participants tended to omit or distort details that did not fit into their existing cultural schema. Also, the famous “War of the Ghost” experiment showed that memory is not just a factual recording of what has occurred, but that we make “effort after meaning.” Memory is shaped by an individual’s prior experience, current beliefs, and future expectations, and these schemas are drawn upon to complete a narrative. Rather than a literal recall of the past, memory builds on these prior beliefs and experiences to reconstruct the experience¹⁶. Tulving is another influential scholar who pointed to the constructed nature of episodic memory. According to Tulving, when we recall past events, we do not just retrieve stored information about the event, but rather actively fill in gaps with our own knowledge, expectations, and assumptions¹⁷. Bartlett and Tulving’s work in the 1960s and 70s has been extremely influential and many scholars studying memory have built on their framework to examine the various ways in which memory is constructed^13,14. 

While the fact that our memory is constructive is widely accepted in the literature, many have debated the cause behind it. The Constructive Episodic Simulation Hypothesis is one such attempt at providing an answer. It is a hypothesis about the “origins” of episodic memory, an attempt to explain why memory is constructed as such^14,18,19. According to the hypothesis, our memory is structured in a constructive fashion to be utilized in the future. While many conceive memory to be primarily a storage of the past, the hypothesis claims that memory plays a critical future-oriented function. As Schacter and Addis put it, “an important function of a constructive episodic memory is to allow individuals to simulate or imagine future episodes, happenings and scenarios”¹¹. Memory is essentially future-oriented, and this direction requires memory to be constructive.

What this means is that the brain flexibly extracts and recombines elements of previous experience for future simulation. This happens through the challenging two-step process of binding and separating noted above. As explained earlier, the brain coordinates multiple neural networks, combines specific elements to form a coherent episode, and keeps the units separate for use in future mental tasks. Since the future is not a literal replay of the past, flexibly using memories from previous experiences is necessary for future simulation. In other words, this whole process allows us to simulate, imagine, or ‘pre-experience’ events that have never occurred in the exact form in which we imagine them¹⁹.

Key aspects of memory construction that enable future simulation include its flexibility to recombine episodic elements and its reliance on schematic reconstruction. Specific mechanisms identified in this review are the two-step process of binding and separating, where neural networks integrate sensory and emotional fragments into coherent episodes while maintaining distinct representations, and the organization of memories into gist clusters, which encode semantic meaning for efficient recombination. These processes, supported by overlapping Default Mode Network activation during recall and future thinking, allow the brain to simulate novel future scenarios by drawing on past experiences.

Neural Evidence

Scholars have described this constructive process through the concept of ‘gist.’ Gist, or semantic clustering, is a method used by the brain to effectively manage all memory fragments. Gist clusters are organized neural patterns encoding semantic meaning. In fact, memory is not disseminated randomly, but rather in the form of organized clusters that consist of similar blocks of ‘gist’ or meaning^10,18. The brain flexibly extracts elements from each ‘gist’ cluster and recombines them to simulate a future. Going back to the birthday party example, in order to remember the episode, our brain needs to integrate fragments from different gist groups that each represent the people we interacted with, the place the party was held, the food we ate, and the emotions we experienced. On the other hand, to simulate or ‘pre-experience’ the next year’s birthday party, we use the same gist groups but different kernels from the group. Say the party was held in a swimming pool last year. The next birthday party may take place in a public park or at home, somewhere other than the swimming pool visited the previous year.

The Constructive Episodic Simulation Hypothesis has been supported by a growing body of research. Functional MRI studies show overlapping brain activity in the Default Mode Network (DMN)—a network of brain regions including the hippocampus, medial prefrontal cortex, and others—during memory retrieval and future thinking¹⁸. Behavioral experiments further demonstrate that memory errors facilitate flexible future planning¹⁹. The DMN supports memory, imagination, and mind-wandering, but its role in episodic simulation is distinct due to specific activation patterns¹⁸. The particular brain region, Default Mode Network, consists of the medial temporal lobe including the hippocampus, anterior cingulate cortex, medial prefrontal cortex, posterior cingulate cortex, and inferior parietal cortex (see Figure 3). Functional Magnetic Resonance Imaging (fMRI) and Positron Emission Tomography (PET) have observed overlap in the brain activity of default network regions when remembering past events and imagining future events¹⁸. Transcranial magnetic stimulation (TMS) studies provide causal evidence, showing that disrupting DMN regions impairs future thinking²¹. The default network is the very region that carries out the challenging two-step process of binding and separating.

This hypothesis, while compelling, is not universally accepted. Critics argue it relies heavily on correlational data, and alternative theories, like reality monitoring, suggest memory prioritizes distinguishing real from imagined events²². However, CESH is favored due to stronger evidence from fMRI overlap and behavioral studies^18,19. Hybrid models acknowledge memory’s multiple functions, such as social bonding, alongside future simulation²³.

Integrating CESH and Adaptive Forgetting

Beyond CESH, adaptive forgetting further illustrates memory’s future-oriented role. Indeed, the Constructive Episodic Memory Hypothesis is just a hypothesis. While many researchers from Schacter and Addis to Watson and McDermott have offered data in support of this view, including fMRI and behavioral evidence^18,19. the hypothesis is not fully embraced by the entire community. There is more work that needs to be done for the hypothesis to be accepted. However, this hypothesis isn’t the only proof of the future-oriented mind²⁴. For this paper, we point to one in particular: studies on adaptive forgetting.

Since the brain is merely finite, it is required to effectively manage memory storage. For instance, the numerous physics formulas we’ve memorized last semester for an exam are probably forgotten or mostly left as fractions. On the contrary, the four fundamental arithmetic operations—addition, subtraction, multiplication, and division—are used daily and remain in our brains. Adaptive forgetting involves complex mechanisms beyond simple retention or loss of information. Adaptive forgetting is the brain’s mechanism for retaining information about the past when it is likely to be useful in the future and forgetting when it is not. Information about a past experience is assessed as useful when it is frequently used in daily life or when it allows us to anticipate what may happen¹⁶. For example, forgetting non-recurring emotional events, like a one-time embarrassment, frees cognitive resources, while retaining trauma-related memories may be adaptive or harmful depending on context²⁵. Older memories that are relatively not recalled or reused so frequently are considered unuseful and become forgotten. However, when an old memory is necessary for anticipating the future, it will remain in the brain waiting to be extracted and recombined.

Studies show that adaptive forgetting prioritizes relevant memory fragments, enabling flexible recombination for future simulations²⁶. This complements CESH by enhancing memory’s efficiency for future-oriented tasks. While there is more work to be done, numerous other studies are pointing in this direction. By integrating CESH and adaptive forgetting, this review illuminates memory’s essential role in anticipating and shaping future outcomes, advancing our understanding of human cognition’s adaptive capacity.

Conclusion and Future Directions

This review aimed to demonstrate that memory is fundamentally future-oriented, leveraging the Constructive Episodic Simulation Hypothesis (CESH) and adaptive forgetting to elucidate how distributed and constructive memory processes enable future simulation. By synthesizing neuroimaging (fMRI, PET) and behavioral evidence, this paper advances CESH by integrating adaptive forgetting as a complementary mechanism, offering a novel perspective on how memory’s fragmented structure and flexible recombination via gist clusters guide decision-making and adaptation.

Several limitations may affect the interpretation and generalizability of these findings. The reliance on correlational neuroimaging data, such as fMRI and PET, restricts causal inferences about memory’s future-oriented functions. Most studies focus on healthy adult populations, potentially limiting applicability to clinical groups with memory disorders like amnesia or dementia. Additionally, cultural and individual differences in memory processes, such as varying reliance on gist-based strategies, are underexplored, and the emphasis on CESH may undervalue alternative memory theories, restricting the findings’ scope across diverse contexts.

Future research should validate CESH through longitudinal fMRI studies and transcranial magnetic stimulation experiments to establish causality, advancing cognitive neuroscience. Tailoring curricula to leverage gist-based learning could enhance retention, while CESH-informed therapies may improve future planning in amnesia patients. Understanding memory’s role in decision-making could inform interventions for memory-related disorders like dementia, fostering interdisciplinary applications. By enabling flexible simulation of future scenarios, memory’s future-oriented design drives human cognition, shaping decision-making and resilience across contexts.

Methods

This theoretical review synthesized peer-reviewed studies through a systematic literature search to explore memory’s distributed and constructive nature, the Constructive Episodic Simulation Hypothesis (CESH), and adaptive forgetting. The search was conducted in databases including PubMed and PsycINFO, using keywords such as “constructive episodic simulation,” “memory consolidation,” “default mode network,” “adaptive forgetting,” and “future-oriented memory.” Filters included English-language articles published between 1933 and 2023. Studies were selected based on inclusion criteria requiring peer-reviewed publications with neuroimaging (fMRI, PET) or behavioral evidence relevant to CESH or memory’s future-oriented functions. Exclusion criteria eliminated non-peer-reviewed sources, case studies, or studies lacking empirical data. Data were extracted on study authors, publication year, research design, key findings, and relevance to CESH or adaptive forgetting. A narrative synthesis approach was used to organize findings thematically, focusing on how distributed memory and constructive processes support future simulation. Quality was assessed by prioritizing studies from high-impact journals and those with robust methodologies (e.g., controlled experiments, large sample sizes). This synthesis ensured a comprehensive evaluation of evidence supporting memory’s adaptive, future-oriented role.

References

1. Nadel, Lynn and Moris Moscovitch. Memory consolidation, retrograde amnesia and the hippocampal complex. Current Opinion in Neurobiology. 7, 217–227 (2002). [↩] [↩] [↩] [↩]
2. Bartlett, Fredric. REMEMBERING: A STUDY IN EXPERIMENTAL AND SOCIAL PSYCHOLOGY. British Journal of Educational Psychology. 3, 187–192 (1933). [↩] [↩] [↩]
3. McFarlan, Amanda. “How Sleep Helps Us Remember and Forget”. BrainPost. https://www.brainpost.co/weekly-brainpost/2021/1/26/how-sleep- helps-us-remember-and-forget (2023). [↩]
4. Hebb, Donald. The Organization of Behavior: A Neuropsychological Theory. Psychology Press. page 378 (1949). [↩]
5. Eichenbaum, Howard. Time cells in the hippocampus: a new dimension for mapping memories. Nature Reviews Neuroscience. 15, 732–744 (2014). [↩]
6. Davachi, Lila and Sarah DuBrow. How the hippocampus preserves order: the role of prediction and context. Trends in Cognitive Sciences. 19, 92–99 (2015). [↩] [↩]
7. Madore, Kevin and Daniel Schacter. An episodic specificity induction enhances means-end problem solving in young and older adults. Psychology and Aging. 29, 913–924 (2014). [↩]
8. Reyna, Valerie and Charles Brainerd. The Science of False Memory. Oxford University Press New York eBooks. page 559 (2005). [↩] [↩] [↩]
9. Jacoby, Larry and Kevin Whitehouse. An illusion of memory: False recognition influenced by unconscious perception. Journal of Experimental Psychology. 118, 126–135 (1989). [↩] [↩]
10. Reyna, Valerie and Charles Brainerd. Gist is the grist: Fuzzy-trace theory and the new intuitionism. Developmental Review. 10, 3–47 (1989). [↩] [↩] [↩] [↩]
11. Schacter, Daniel et al. The Future of Memory: Remembering, Imagining, and the Brain. Neuron. 76, 677–694 (2012). [↩] [↩]
12. Nairne, James S. Adaptive memory: Evolutionary constraints on remembering. Psychology of Learning and Motivation, 53, 1–32 (2008). [↩]
13. Schacter, Daniel. Adaptive constructive processes and the future of memory. American Psychologist. 67, 603–613(2013). [↩] [↩]
14. Schacter, Daniel and Dona Addis. The cognitive neuroscience of constructive memory: remembering the past and imagining the future. Philosophical Transactions of the Royal Society B. 362, 773–786 (2007). [↩] [↩] [↩]
15. Clayton, Nicola S. and Dickinson, Anthony. Episodic-like memory during cache recovery by scrub jays. Nature, 395, 272–274 (1998). [↩]
16. Buckner, Randy and Daniel Carroll. Self-projection and the brain. Trends in Cognitive Sciences. 11, 49–57 (2007). [↩] [↩]
17. Tulving, Endel. Elements of episodic memory. Oxford University Press. page 364 (1983). [↩]
18. Addis, Dona et al. Remembering the past and imagining the future: Common and distinct neural substrates during event construction and elaboration. Neuropsychologia. 45, 1363–1377 (2007). [↩] [↩] [↩] [↩] [↩] [↩] [↩]
19. Atance, Chistina and Daniela O’Neill. Episodic future thinking. Trends in Cognitive Sciences. 5, 533–539 (2001). [↩] [↩] [↩] [↩] [↩]
20. Schacter, Daniel. The seven sins of memory: how the mind forgets and remembers. Choice Reviews Online. 39, 39–2484 (2002). [↩] [↩]
21. Thakral, Preston P. et al. Transcranial magnetic stimulation to the angular gyrus modulates the temporal organization of autobiographical memory. Cerebral Cortex, 30, 3027–3038 (2020). [↩]
22. Johnson, Marcia K. Reality monitoring: Evidence from cognitive neuroscience. In W. Hirst (Ed.), Memory in the wild (pp. 3–28). Oxford University Press (1991). [↩]
23. Alea, Nicole and Bluck, Susan. Why are you telling me that? A conceptual model of the social function of autobiographical memory. Memory, 11, 165–178 (2003). [↩]
24. Benoit, Roland. “Our memory is not made for the past, but for the future”. Max Planck Institute for Human Cognitive and Brain Sciences. https://www.cbs.mpg.de/press-release/memory. (2016). [↩]
25. Hardt, Oliver et al. Decay happens: The role of active forgetting in memory. Trends in Cognitive Sciences, 17, 111–120 (2013). [↩]
26. Anderson, Michael C. and Schooler, Laurence J. The adaptive nature of memory. In E. Tulving & F. I. M. Craik (Eds.), The Oxford handbook of memory (pp. 557–570). Oxford University Press (2000). [↩]