Generalization through the recurrent interaction of episodic memories: a model of the hippocampal system

Exploring the role of dimensionality transformation in episodic memory

Episodic memory must accomplish two adversarial goals: encoding and storing a multitude of experiences without exceeding the finite neuronal structure of the brain, and recalling memories in vivid detail. Dimensionality reduction and expansion ('dimensionality transformation') enable the brain to meet these demands. Reduction compresses sensory input into simplified, storable codes, while expansion reconstructs vivid details. Although these processes are essential to memory, their neural mechanisms for episodic memory remain unclear. Drawing on recent insights from cognitive psychology, systems neuroscience, and neuroanatomy, we propose two accounts of how dimensionality transformation occurs in the brain: structurally (via corticohippocampal pathways) and functionally (through neural oscillations). By examining cross-species evidence, we highlight neural mechanisms that may support episodic memory and identify crucial questions for future research.

Integration of conditioned threat with pre-existing memories

How does negative affect spread through existing memories? Whereas many studies have investigated generalization of learned threat responses across perceptual and semantic dimensions, little attention has been given to the possibility that Pavlovian threat responses may spread beyond what is directly learned to previously encoded memories that overlap in content. Here, we increased the demand on associative memory in a modified sensory preconditioning task to investigate this. First, participants encoded 40 unique episodes, each consisting of two neutral stimuli. On the following day, one of each pair was newly associated with either an aversive or a neutral stimulus. Another day later, both stimuli of the original memories were found to trigger enhanced pupil dilation if one was indirectly linked to an aversive stimulus. This effect was independent of whether the associations encoded on day 1 were accurately retained on the day of testing, and confined to trials on which the indirectly associated stimulus was consciously brought to mind, suggesting the formation of a link that directly connects preconditioned stimuli to subsequently learned aversive outcomes. The present study demonstrates that the human defensive system is remarkably adept at quickly anticipating threat based on information acquired over separate events, and gives a first glimpse into the associative structures that enable this ability.

Memory updating and the structure of event representations

People form memories of specific events and use those memories to make predictions about similar new experiences. Living in a dynamic environment presents a challenge: How does one represent valid prior events in memory while encoding new experiences when things change? There is evidence for two seemingly contradictory classes of mechanism: One differentiates outdated event features by making them less similar or less accessible than updated event features. The other integrates updated features of new events with outdated memories, and the relationship between them, into a structured representation. Integrative encoding may occur when changed events trigger inaccurate predictions based on remembered prior events. We propose that this promotes subsequent recollection of events and their order, enabling adaptation to environmental changes.

The effects of episodic context on memory integration

Information encountered in different events, such as people and objects, can be interlinked in memory. Such memory integration supports novel inferences about the world. This study investigates the role of episodic context on memory integration in two experiments using an associative inference task. Participants encoded events with overlapping (AB and BC) and non-overlapping associations (XY) presented in the same or different episodic contexts. Inference performance across events (AC) was tested in the absence (Experiment 1) and in the presence (Experiment 2) of the encoding context. Our data show that inferences across events encoded in the same context were more accurate, faster, and made with greater confidence compared to those encoded in different contexts. However, this effect was observed only when the context was presented during testing, suggesting that context enhances associative inferences by facilitating retrieval of events associated with that context. These findings demonstrate that revisiting the encoding context promotes memory integration by providing privileged access to contextually associated memory traces and facilitating their flexible recombination to form novel inferences.

Electrophysiological dynamics of salience, default mode, and frontoparietal networks during episodic memory formation and recall revealed through multi-experiment iEEG replication

Dynamic interactions between large-scale brain networks underpin human cognitive processes, but their electrophysiological mechanisms remain elusive. The triple network model, encompassing the salience network (SN), default mode network (DMN), and frontoparietal network (FPN), provides a framework for understanding these interactions. We analyzed intracranial electroencephalography (EEG) recordings from 177 participants across four diverse episodic memory experiments, each involving encoding as well as recall phases. Phase transfer entropy analysis revealed consistently higher directed information flow from the anterior insula (AI), a key SN node, to both DMN and FPN nodes. This directed influence was significantly stronger during memory tasks compared to resting state, highlighting the AI's task-specific role in coordinating large-scale network interactions. This pattern persisted across externally driven memory encoding and internally governed free recall. Control analyses using the inferior frontal gyrus (IFG) showed an inverse pattern, with DMN and FPN exerting higher influence on IFG, underscoring the AI's unique role. We observed task-specific suppression of high-gamma power in the posterior cingulate cortex/precuneus node of the DMN during memory encoding, but not recall. Crucially, these results were replicated across all four experiments spanning verbal and spatial memory domains with high Bayes replication factors. Our findings advance understanding of how coordinated neural network interactions support memory processes, highlighting the AI's critical role in orchestrating large-scale brain network dynamics during both memory encoding and retrieval. By elucidating the electrophysiological basis of triple network interactions in episodic memory, our study provides insights into neural circuit dynamics underlying memory function and offer a framework for investigating network disruptions in memory-related disorders.

Eliminating episodic memory?

In Tulving's initial characterization, episodic memory was one of multiple memory systems. It was postulated, in pursuit of explanatory depth, as displaying proprietary operations, representations and substrates such as to explain a range of cognitive, behavioural and experiential phenomena. Yet the subsequent development of this research programme has, paradoxically, introduced surprising doubts about the nature, and indeed existence, of episodic memory. On dominant versions of the 'common system' view, on which a single simulation system underlies both remembering and imagining, there are no processes unique to memory to support robust generalizations with inductive potential. Eliminativism about episodic memory seems to follow from the claim that it has no dedicated neurocognitive system of its own. After identifying this under-noticed threat, we push back against modern eliminativists by surveying recent evidence that still indicates specialized mechanisms, computations and representations that are distinctly mnemic in character. We argue that contemporary realists about episodic memory can retain lessons of the common system approach while resisting the further move to eliminativism. This article is part of the theme issue 'Elements of episodic memory: lessons from 40 years of research'.

Neuroplasticity Meets Artificial Intelligence: A Hippocampus-Inspired Approach to the Stability-Plasticity Dilemma

The stability-plasticity dilemma remains a critical challenge in developing artificial intelligence (AI) systems capable of continuous learning. This perspective paper presents a novel approach by drawing inspiration from the mammalian hippocampus-cortex system. We elucidate how this biological system's ability to balance rapid learning with long-term memory retention can inspire novel AI architectures. Our analysis focuses on key mechanisms, including complementary learning systems and memory consolidation, with emphasis on recent discoveries about sharp-wave ripples and barrages of action potentials. We propose innovative AI designs incorporating dual learning rates, offline consolidation, and dynamic plasticity modulation. This interdisciplinary approach offers a framework for more adaptive AI systems while providing insights into biological learning. We present testable predictions and discuss potential implementations and implications of these biologically inspired principles. By bridging neuroscience and AI, our perspective aims to catalyze advancements in both fields, potentially revolutionizing AI capabilities while deepening our understanding of neural processes.

Navigating Memorability Landscapes: Hyperbolic Geometry Reveals Hierarchical Structures in Object Concept Memory

Why are some object concepts (e.g., birds, cars, vegetables, etc.) more memorable than others? Prior studies have suggested that features (e.g., color, animacy, etc.) and typicality (e.g., robin vs. penguin) of object images influences the likelihood of being remembered. However, a complete understanding of object memorability remains elusive. In this study, we examine whether the geometric relationship between object concepts explains differences in their memorability. Specifically, we hypothesize that image concepts will be geometrically arranged in hierarchical structures and that memorability will be explained by a concept's depth in these hierarchical trees. To test this hypothesis, we construct a Hyperbolic representation space of object concepts (N=1,854) from the THINGS database (Hebart et al., 2019), which consists of naturalistic images of concrete objects, and a space of 49 feature dimensions derived from data-driven models. Using ALBATROSS (Stier, A. J., Giusti, C., & Berman, M. G., In prep), a stochastic topological data analysis technique that detects underlying structures of data, we demonstrate that Hyperbolic geometry efficiently captures the hierarchical organization of object concepts above and beyond a traditional Euclidean geometry and that hierarchical organization is related to memorability. We find that concepts closer to the center of the representational space are more prototypical and also more memorable. Importantly, Hyperbolic distances are more predictive of memorability and prototypicality than Euclidean distances, suggesting that concept memorability and typicality are organized hierarchically. Taken together, our work presents a novel hierarchical representational structure of object concepts that explains memorability and typicality.

Electrophysiological dynamics of salience, default mode, and frontoparietal networks during episodic memory formation and recall: A multi-experiment iEEG replication

Dynamic interactions between large-scale brain networks underpin human cognitive processes, but their electrophysiological mechanisms remain elusive. The triple network model, encompassing the salience (SN), default mode (DMN), and frontoparietal (FPN) networks, provides a framework for understanding these interactions. We analyzed intracranial EEG recordings from 177 participants across four diverse episodic memory experiments, each involving encoding as well as recall phases. Phase transfer entropy analysis revealed consistently higher directed information flow from the anterior insula (AI), a key SN node, to both DMN and FPN nodes. This directed influence was significantly stronger during memory tasks compared to resting-state, highlighting the AI's task-specific role in coordinating large-scale network interactions. This pattern persisted across externally-driven memory encoding and internally-governed free recall. Control analyses using the inferior frontal gyrus (IFG) showed an inverse pattern, with DMN and FPN exerting higher influence on IFG, underscoring the AI's unique role. We observed task-specific suppression of high-gamma power in the posterior cingulate cortex/precuneus node of the DMN during memory encoding, but not recall. Crucially, these results were replicated across all four experiments spanning verbal and spatial memory domains with high Bayes replication factors. Our findings advance understanding of how coordinated neural network interactions support memory processes, highlighting the AI's critical role in orchestrating large-scale brain network dynamics during both memory encoding and retrieval. By elucidating the electrophysiological basis of triple network interactions in episodic memory, our study provides insights into neural circuit dynamics underlying memory function and offer a framework for investigating network disruptions in memory-related disorders.

A geometrical solution underlies general neural principle for serial ordering

A general mathematical description of how the brain sequentially encodes knowledge remains elusive. We propose a linear solution for serial learning tasks, based on the concept of mixed selectivity in high-dimensional neural state spaces. In our framework, neural representations of items in a sequence are projected along a "geometric" mental line learned through classical conditioning. The model successfully solves serial position tasks and explains behaviors observed in humans and animals during transitive inference tasks amidst noisy sensory input and stochastic neural activity. This approach extends to recurrent neural networks performing motor decision tasks, where the same geometric mental line correlates with motor plans and modulates network activity according to the symbolic distance between items. Serial ordering is thus predicted to emerge as a monotonic mapping between sensory input and behavioral output, highlighting a possible pivotal role for motor-related associative cortices in transitive inference tasks.

Interhemispheric CA1 projections support spatial cognition and are affected in a mouse model of the 22q11.2 deletion syndrome

Untangling the hippocampus connectivity is critical for understanding the mechanisms supporting learning and memory. However, the function of interhemispheric connections between hippocampal formations is still poorly understood. So far, two major hippocampal commissural projections have been characterized in rodents. Mossy cells from the hilus of the dentate gyrus project to the inner molecular layer of the contralateral dentate gyrus and CA3 and CA2 pyramidal neuron axonal collaterals to contralateral CA3, CA2 and CA1. In contrary, little is known about commissural projection from the CA1 region. Here, we show that CA1 pyramidal neurons from the dorsal hippocampus project to contralateral dorsal CA1 as well as dorsal subiculum. We further demonstrate that the interhemispheric projection from CA1 to dorsal subiculum supports spatial memory and spatial working memory in WT mice, two cognitive functions impaired in male mice from the Df16(A) +/- model of 22q11.2 deletion syndrome (22q11.2DS) associated with schizophrenia. Investigation of the CA1 interhemispheric projections in Df16(A) +/- mice revealed that these projections are disrupted with male mutants showing stronger anatomical defects compared to females. Overall, our results characterize a novel interhemispheric projection from dCA1 to dorsal subiculum and suggest that dysregulation of this projection may contribute to the cognitive deficits associated with the 22q11.2DS.

A mathematical theory of relational generalization in transitive inference

Humans and animals routinely infer relations between different items or events and generalize these relations to novel combinations of items. This allows them to respond appropriately to radically novel circumstances and is fundamental to advanced cognition. However, how learning systems (including the brain) can implement the necessary inductive biases has been unclear. We investigated transitive inference (TI), a classic relational task paradigm in which subjects must learn a relation ([Formula: see text] and [Formula: see text]) and generalize it to new combinations of items ([Formula: see text]). Through mathematical analysis, we found that a broad range of biologically relevant learning models (e.g. gradient flow or ridge regression) perform TI successfully and recapitulate signature behavioral patterns long observed in living subjects. First, we found that models with item-wise additive representations automatically encode transitive relations. Second, for more general representations, a single scalar "conjunctivity factor" determines model behavior on TI and, further, the principle of norm minimization (a standard statistical inductive bias) enables models with fixed, partly conjunctive representations to generalize transitively. Finally, neural networks in the "rich regime," which enables representation learning and improves generalization on many tasks, unexpectedly show poor generalization and anomalous behavior on TI. We find that such networks implement a form of norm minimization (over hidden weights) that yields a local encoding mechanism lacking transitivity. Our findings show how minimal statistical learning principles give rise to a classical relational inductive bias (transitivity), explain empirically observed behaviors, and establish a formal approach to understanding the neural basis of relational abstraction.

Age differences in generalization, memory specificity, and their overnight fate in childhood

Memory enables generalization to new situations, and memory specificity that preserves individual episodes. This study investigated generalization, memory specificity, and their overnight fate in 141 4- to 8-year-olds (computerized memory game; 71 females, tested 2020-2021 in Germany). The results replicated age effects in generalization and memory specificity, and a contingency of generalization on object conceptual properties and interobject semantic proximity. Age effects were stronger in generalization than in memory specificity, and generalization was more closely linked to the explicit regularity knowledge in older than in younger children. After an overnight delay, older children retained more generalized and specific memories and showed greater gains but only in generalization. These findings reveal distinct age differences in generalization and memory specificity across childhood.

Disrupting dorsal hippocampus impairs category learning in rats

Categorization requires a balance of mechanisms that can generalize across common features and discriminate against specific details. A growing literature suggests that the hippocampus may accomplish these mechanisms by using fundamental mechanisms like pattern separation, pattern completion, and memory integration. Here, we assessed the role of the rodent dorsal hippocampus (HPC) in category learning by combining inhibitory DREADDs (Designer Receptors Exclusively Activated by Designer Drugs) and simulations using a neural network model. Using touchscreens, we trained rats to categorize distributions of visual stimuli containing black and white gratings that varied along two continuous dimensions. Inactivating the dorsal HPC impaired category learning and generalization, suggesting that the rodent HPC plays an important role during categorization. Hippocampal inactivation had no effect on a control discrimination task that used identical trial procedures as the categorization tasks, suggesting that the impairments were specific to categorization. Model simulations were conducted with variants of a neural network to assess the impact of selective deficits on category learning. The hippocampal inactivation groups were best explained by a model that injected random noise into the computation that compared the similarity between category stimuli and existing memory representations. This model is akin to a deficit in mechanisms of pattern completion, which retrieves similar memory representations using partial information.

No evidence that arousal affects reactivated memories

Memory for inherently neutral elements of emotional events is often enhanced on delayed tests - an effect that has been attributed to noradrenergic arousal. Reactivation of a memory is thought to return its corresponding neural ensemble to a state that is similar to when it was originally experienced. Therefore, we hypothesized that neutral elements of memories, too, can be enhanced through reactivation concurrent with heightened arousal. Participants (n = 94) visited the lab for three sessions. During the first session, they encoded 120 neutral memories consisting of an object presented in unique context images. In session two, the 80 objects were reactivated by presenting their corresponding context images, 40 of which were immediately followed by an arousal-inducing shock. Finally, recognition memory for all objects was tested. It was found that memory for reactivated objects was enhanced, but even though the shocks elicited elevations in arousal as indexed by skin conductance, there was no difference between memory of objects reactivated with and without heightened arousal. We thus conclude that arousal, when isolated from other cognitive and affective variables that might impact memory, has no enhancing effect on reactivated memories.

The role of morphology in novel word learning: a registered report

The majority of the new words that we learn every day as adults are morphologically complex; yet, we do not know much about the role of morphology in novel word learning. In this study, we tackle this issue by comparing the learning of: (i) suffixed novel words (e.g. flibness); (ii) novel words that end in non-morphological, but frequent letter chunks (e.g. fliban); and (iii) novel words with non-morphological, low-frequency endings (e.g. flibov). Words are learned incidentally through sentence reading, while the participants' eye movements are monitored. We show that morphology has a facilitatory role compared with the other two types of novel words, both during learning and in a post-learning recognition memory task. We also showed that participants attributed meaning to word parts (if flibness is a state of happiness, then flib must mean happy), but this process was not specifically triggered by the presence of a suffix (flib must also mean happy in fliban and flibov), thus suggesting that the brain tends to assume similar meanings for similar words and word parts.

Emergent neural dynamics and geometry for generalization in a transitive inference task

Relational cognition-the ability to infer relationships that generalize to novel combinations of objects-is fundamental to human and animal intelligence. Despite this importance, it remains unclear how relational cognition is implemented in the brain due in part to a lack of hypotheses and predictions at the levels of collective neural activity and behavior. Here we discovered, analyzed, and experimentally tested neural networks (NNs) that perform transitive inference (TI), a classic relational task (if A > B and B > C, then A > C). We found NNs that (i) generalized perfectly, despite lacking overt transitive structure prior to training, (ii) generalized when the task required working memory (WM), a capacity thought to be essential to inference in the brain, (iii) emergently expressed behaviors long observed in living subjects, in addition to a novel order-dependent behavior, and (iv) expressed different task solutions yielding alternative behavioral and neural predictions. Further, in a large-scale experiment, we found that human subjects performing WM-based TI showed behavior inconsistent with a class of NNs that characteristically expressed an intuitive task solution. These findings provide neural insights into a classical relational ability, with wider implications for how the brain realizes relational cognition.

Differential effects of bilateral hippocampal CA3 damage on the implicit learning and recognition of complex event sequences

Learning regularities in the environment is a fundament of human cognition, which is supported by a network of brain regions that include the hippocampus. In two experiments, we assessed the effects of selective bilateral damage to human hippocampal subregion CA3, which was associated with autobiographical episodic amnesia extending ~50 years prior to the damage, on the ability to recognize complex, deterministic event sequences presented either in a spatial or a non-spatial configuration. In contrast to findings from related paradigms, modalities, and homologue species, hippocampal damage did not preclude recognition memory for an event sequence studied and tested at four spatial locations, whereas recognition memory for an event sequence presented at a single location was at chance. In two additional experiments, recognition memory for novel single-items was intact, whereas the ability to recognize novel single-items in a different location from that presented at study was at chance. The results are at variance with a general role of the hippocampus in the learning and recognition of complex event sequences based on non-adjacent spatial and temporal dependencies. We discuss the impact of the results on established theoretical accounts of the hippocampal contributions to implicit sequence learning and episodic memory.

Can the prosocial benefits of episodic simulation transfer to different people and situational contexts?

Previous research has found that episodic simulation of events of helping others can effectively enhance intentions to help the same person involved and the identical situational context as the imagined scenarios. This 'prosocial simulation effect' is argued to reflect, at least in part, associative memory mechanisms whereby the simulation is reactivated when in the same situation as that imagined. However, to date, no study has examined systematically whether this 'prosocial simulation effect' can be transferred to response scenarios involving different people and/or situational contexts to the imagined scenarios, and if so, whether the degree of overlap with the imagined helping episode modulated the transfer effect. Across two experiments, we systematically varied the overlap of the simulated and response scenarios, both in terms of the persons in need and/or the situational contexts, and whether would influence the magnitude of prosocial simulation effect. Results from both experiments showed that the prosocial simulation effect can be transferred to response scenarios involving different people and situational contexts to the simulated scenarios. However, this finding was primarily driven by response scenarios that had a high degree of overlap to the simulated scenarios. The application of our findings to the practical implementation of simulation to promote prosociality in the real world is discussed.

What is the role of the hippocampus and parahippocampal gyrus in the persistence of tinnitus?

The hippocampus and parahippocampal gyrus have been implicated as part of a tinnitus network by a number of studies. These structures are usually considered in the context of a "limbic system," a concept typically invoked to explain the emotional response to tinnitus. Despite this common framing, it is not apparent from current literature that this is necessarily the main functional role of these structures in persistent tinnitus. Here, we highlight a different role that encompasses their most commonly implicated functional position within the brain-that is, as a memory system. We consider tinnitus as an auditory object that is held in memory, which may be made persistent by associated activity from the hippocampus and parahippocampal gyrus. Evidence from animal and human studies implicating these structures in tinnitus is reviewed and used as an anchor for this hypothesis. We highlight the potential for the hippocampus/parahippocampal gyrus to facilitate maintenance of the memory of the tinnitus percept via communication with auditory cortex, rather than (or in addition to) mediating emotional responses to this percept.

The effect of hippocampal subfield damage on rapid temporal integration through statistical learning and associative inference

INTRODUCTION

The hippocampus (HPC) supports integration of information across time, often indexed by associative inference (AI) and statistical learning (SL) tasks. In AI, an indirect association between stimuli that never appeared together is inferred, whereas SL involves learning item relationships by extracting regularities across experiences. A recent model of hippocampal function (Schapiro et al., 2017) proposes that the HPC can support temporal integration in both paradigms through its two distinct pathways.

METHODS

We tested this models' predictions in four patients with varying degrees of bilateral HPC damage and matched healthy controls, with two patients with complementary damage to either the monosynaptic or trisynaptic pathway. During AI, participants studied overlapping paired associates (AB, BC) and their memory was tested for premise pairs (AB) and for inferred pairs (AC). During SL, participants passively viewed a continuous picture sequence that contained an underlying structure of triplets that later had to be recognized.

RESULTS

Binomial distributions were used to calculate above chance performance at the individual level. For AI, patients with focal HPC damage were impaired at inference but could correctly infer pairs above chance once premise pair acquisition was equated to controls; however, the patient with HPC and cortical damage showed severe impairment at recalling premise and inferred pairs, regardless of accounting for premise pair performance. For SL, none of the patients performed above chance, but notably neither did most controls.

CONCLUSIONS

Associative inference of indirect relationships can be intact with HPC damage to either hippocampal pathways or the HPC more broadly, provided premise pairs can first be formed. Inference may remain intact through residual HPC tissue supporting premise pair acquisition, and/or through extra-hippocampal structures supporting inference at retrieval. Clear conclusions about hippocampal contributions to SL are precluded by low performance in controls, which we caution is not dissimilar to previous amnesic studies using the same task. This complicates interpretations of studies claiming necessity of hippocampal contributions to SL and warrants the use of a common and reliable task before conclusions can be drawn.

Examining memory linking and generalization using scFLARE2, a temporally precise neuronal activity tagging system

How memories are organized in the brain influences whether they are remembered discretely versus linked with other experiences or whether generalized information is applied to entirely novel situations. Here, we used scFLARE2 (single-chain fast light- and activity-regulated expression 2), a temporally precise tagging system, to manipulate mouse lateral amygdala neurons active during one of two 3 min threat experiences occurring close (3 h) or further apart (27 h) in time. Silencing scFLARE2-tagged neurons showed that two threat experiences occurring at distal times are dis-allocated to orthogonal engram ensembles and remembered discretely, whereas the same two threat experiences occurring in close temporal proximity are linked via co-allocation to overlapping engram ensembles. Moreover, we found that co-allocation mediates memory generalization applied to a completely novel stimulus. These results indicate that endogenous temporal evolution of engram ensemble neuronal excitability determines how memories are organized and remembered and that this would not be possible using conventional immediate-early gene-based tagging methods.

A neural network model of hippocampal contributions to category learning

In addition to its critical role in encoding individual episodes, the hippocampus is capable of extracting regularities across experiences. This ability is central to category learning, and a growing literature indicates that the hippocampus indeed makes important contributions to this form of learning. Using a neural network model that mirrors the anatomy of the hippocampus, we investigated the mechanisms by which the hippocampus may support novel category learning. We simulated three category learning paradigms and evaluated the network's ability to categorize and recognize specific exemplars in each. We found that the trisynaptic pathway within the hippocampus-connecting entorhinal cortex to dentate gyrus, CA3, and CA1-was critical for remembering exemplar-specific information, reflecting the rapid binding and pattern separation capabilities of this circuit. The monosynaptic pathway from entorhinal cortex to CA1, in contrast, specialized in detecting the regularities that define category structure across exemplars, supported by the use of distributed representations and a relatively slower learning rate. Together, the simulations provide an account of how the hippocampus and its constituent pathways support novel category learning.

A registered report testing the effect of sleep on Deese-Roediger-McDermott false memory: greater lure and veridical recall but fewer intrusions after sleep

Human memory is known to be supported by sleep. However, less is known about the effect of sleep on false memory, where people incorrectly remember events that never occurred. In the laboratory, false memories are often induced via the Deese-Roediger-McDermott (DRM) paradigm where participants are presented with wordlists comprising semantically related words such as nurse, hospital and sick (studied words). Subsequently, participants are likely to falsely remember that a related lure word such as doctor was presented. Multiple studies have examined whether these false memories are influenced by sleep, with contradictory results. A recent meta-analysis suggests that sleep may increase DRM false memory when short lists are used. We tested this in a registered report (N = 488) with a 2 (Interval: Immediate versus 12 h delay) × 2 (Test Time: 9:00 versus 21:00) between-participant DRM experiment, using short DRM lists (N = 8 words/list) and free recall as the memory test. We found an unexpected time-of-day effect such that completing free recall in the evening led to more intrusions (neither studied nor lure words). Above and beyond this time-of-day effect, the Sleep participants produced fewer intrusions than their Wake counterparts. When this was statistically controlled for, the Sleep participants falsely produced more critical lures. They also correctly recalled more studied words (regardless of intrusions). Exploratory analysis showed that these findings cannot be attributed to differences in output bias, as indexed by the number of total responses. Our overall results cannot be fully captured by existing sleep-specific theories of false memory, but help to define the role of sleep in two more general theories (Fuzzy-Trace and Activation/Monitoring theories) and suggest that sleep may benefit gist abstraction/spreading activation on one hand and memory suppression/source monitoring on the other.

Mental representation of equivalence and order

With mental models based on relational information, the present research shows that the semantics expressed by the relation can determine the structural properties of the constructed model. In particular, we demonstrate a reversal of the classical, well-replicated symbolic distance effect (SDE), as a function of relational semantics. The classical SDE shows that responses are more accurate, and faster, the wider the distance between queried elements on a mentally constructed rank order. We replicate this effect in a study using a relation that expresses a rank hierarchy ("older than," Experiment 4). In contrast, we obtain a clear reversal of the same effect for accuracy data when the relation expresses a number of equivalence classes ("is from the same city," Experiments 1-3). In Experiment 3, we find clear evidence of a reversed SDE for accuracy and latency in the above standard condition, and flat curves of means, across pair distances, for accuracy and latency in a condition that makes equivalence classes salient from the beginning. We discuss these findings in the context of a process model of equivalence class formation based on learned piecemeal information.

Threat impairs flexible use of a cognitive map

Goal-directed behavior requires adaptive systems that respond to environmental demands. In the absence of threat (or presence of reward), individuals can explore many behavioral trajectories, effectively interrogating the environment across multiple dimensions. This leads to flexible, relational memory encoding and retrieval. In the presence of danger, motivation shifts to an imperative state characterized by a narrow focus of attention on threatening information. This impairs flexible, relational memory. We test how these motivational shifts (Murty & Adcock, 2017) affect behavioral flexibility in an ecologically valid setting. Participants learned the structure of maze-like environments and navigated to the location of objects in both safe and threatening contexts. The latter contained a predator that could 'capture' participants, leading to electric shock. After learning, the path to some objects was unpredictably blocked. forcing a detour for which one route was significantly shorter. We predicted that threat would push participants toward an imperative state, leading to less efficient and less flexible navigation. Threat caused participants to take longer paths to goal objects and less efficient detours when obstacles were encountered. Threat-related impairments in detour navigation persisted after controlling for non-detour navigation performance. and non-detour navigation was not a reliable predictor of detour navigation, This suggests a specific impairment in flexible navigation during detours, an impairment unlikely to be explained by more general processes like predator avoidance or divided attention that may be present during non-detour navigation. These results provide ecologically valid evidence that dynamic, observable threats reduce flexible use of cognitive maps to guide behavior.

Abstract cognitive maps of social network structure aid adaptive inference

Social navigation-such as anticipating where gossip may spread, or identifying which acquaintances can help land a job-relies on knowing how people are connected within their larger social communities. Problematically, for most social networks, the space of possible relationships is too vast to observe and memorize. Indeed, people's knowledge of these social relations is well known to be biased and error-prone. Here, we reveal that these biased representations reflect a fundamental computation that abstracts over individual relationships to enable principled inferences about unseen relationships. We propose a theory of network representation that explains how people learn inferential cognitive maps of social relations from direct observation, what kinds of knowledge structures emerge as a consequence, and why it can be beneficial to encode systematic biases into social cognitive maps. Leveraging simulations, laboratory experiments, and "field data" from a real-world network, we find that people abstract observations of direct relations (e.g., friends) into inferences of multistep relations (e.g., friends-of-friends). This multistep abstraction mechanism enables people to discover and represent complex social network structure, affording adaptive inferences across a variety of contexts, including friendship, trust, and advice-giving. Moreover, this multistep abstraction mechanism unifies a variety of otherwise puzzling empirical observations about social behavior. Our proposal generalizes the theory of cognitive maps to the fundamental computational problem of social inference, presenting a powerful framework for understanding the workings of a predictive mind operating within a complex social world.

Memory consolidation: Building influence over the entorhinal cortex

Memory consolidation involves interactions between the hippocampus and other cortical areas. A new study identifies neurons in the medial entorhinal cortex that over learning increase their coordination with hippocampal replay events, suggesting a route for consolidation of spatial memories.

Conceptual relatedness promotes memory generalization at the cost of detailed recollection

An adaptive memory system is one that allows us to both retrieve detailed memories as well as generalize knowledge about our past, the latter termed memory generalization and is useful for making inferences about new situations. Research has indicated that memory generalization relies on forming knowledge structures by integrating experiences with shared encountered elements. Whether memory generalization occurs more readily when  experiences also have elements that share established (conceptual) information is less clear. It is also unclear if engaging in memory generalization during learning comes at the cost of retrieving detailed memories, the other function of episodic memory. To address these two knowledge gaps, we paired a modified version of the acquired equivalence task with a recognition memory test. Across three experiments, participants first learned a series of overlapping object-scene pairs (A-X, B-X and A-Y) in which half of the overlapping pairs contained conceptually-related objects (e.g., A-pencil; B-scissors; conceptual condition) and the other half contained unrelated objects (neutral condition). Participants ability to generalize to new overlapping object-scene pairs (B-Y) as well as not-learned but semantically-related objects was measured. Finally, participants completed a recognition memory test that included the encoded objects, perceptually similar lures or new foil objects. Across all experiments, we found higher rates of generalization but reduced detailed memory (indexed by increased false alarms to lure objects) for information learned in the conceptual than neutral condition. These results suggest the presence of conceptual knowledge biases an individual towards a generalization function of memory, which comes at the expense of detailed recollection.

Integration of event experiences to build relational knowledge in the human brain

We investigated how the human brain integrates experiences of specific events to build general knowledge about typical event structure. We examined an episodic memory area important for temporal relations, anterior-lateral entorhinal cortex, and a semantic memory area important for action concepts, middle temporal gyrus, to understand how and when these areas contribute to these processes. Participants underwent functional magnetic resonance imaging while learning and recalling temporal relations among novel events over two sessions 1 week apart. Across distinct contexts, individual temporal relations among events could either be consistent or inconsistent with each other. Within each context, during the recall phase, we measured associative coding as the difference of multivoxel correlations among related vs unrelated pairs of events. Neural regions that form integrative representations should exhibit stronger associative coding in the consistent than the inconsistent contexts. We found evidence of integrative representations that emerged quickly in anterior-lateral entorhinal cortex (at session 1), and only subsequently in middle temporal gyrus, which showed a significant change across sessions. A complementary pattern of findings was seen with signatures during learning. This suggests that integrative representations are established early in anterior-lateral entorhinal cortex and may be a pathway to the later emergence of semantic knowledge in middle temporal gyrus.

Interpolated retrieval retroactively increases recall and promotes cross-episode memory interdependence

Retrieving existing memories before new learning can lead to retroactive facilitation. Three experiments examined whether interpolated retrieval is associated with retroactive facilitation and memory interdependence that reflects integrative encoding. Participants studied two lists of cue-response word pairs that repeated across lists (A-B, A-B), appeared in list 1 (A-B, -), or included the same cues with changed responses in each list (A-B, A-C). For A-B, A-C pairs, the tasks interpolated between lists required recall of list 1 (B) responses (with or without feedback) or restudy of complete list 1 (A-B) pairs. In list 2, participants only studied pairs (experiment 1) or studied pairs, attempted to detect changed (C) responses, and attempted to recall list 1 responses for detected changes (experiments 2 and 3). On a final cued recall test, participants attempted to recall list 1 responses, indicated whether responses changed between lists, and if so, attempted to recall list 2 responses. Interpolated retrieval was associated with subsequent retroactive facilitation and greater memory interdependence for B and C responses. These correlational findings are compatible with the view that retrieval retroactively facilitates memories, promotes coactivation of existing memories and new learning, and enables integrative encoding that veridically binds information across episodes.

Organizing memories for generalization in complementary learning systems

Memorization and generalization are complementary cognitive processes that jointly promote adaptive behavior. For example, animals should memorize safe routes to specific water sources and generalize from these memories to discover environmental features that predict new ones. These functions depend on systems consolidation mechanisms that construct neocortical memory traces from hippocampal precursors, but why systems consolidation only applies to a subset of hippocampal memories is unclear. Here we introduce a new neural network formalization of systems consolidation that reveals an overlooked tension-unregulated neocortical memory transfer can cause overfitting and harm generalization in an unpredictable world. We resolve this tension by postulating that memories only consolidate when it aids generalization. This framework accounts for partial hippocampal-cortical memory transfer and provides a normative principle for reconceptualizing numerous observations in the field. Generalization-optimized systems consolidation thus provides new insight into how adaptive behavior benefits from complementary learning systems specialized for memorization and generalization.

Neural knowledge assembly in humans and neural networks

Human understanding of the world can change rapidly when new information comes to light, such as when a plot twist occurs in a work of fiction. This flexible "knowledge assembly" requires few-shot reorganization of neural codes for relations among objects and events. However, existing computational theories are largely silent about how this could occur. Here, participants learned a transitive ordering among novel objects within two distinct contexts before exposure to new knowledge that revealed how they were linked. Blood-oxygen-level-dependent (BOLD) signals in dorsal frontoparietal cortical areas revealed that objects were rapidly and dramatically rearranged on the neural manifold after minimal exposure to linking information. We then adapt online stochastic gradient descent to permit similar rapid knowledge assembly in a neural network model.

Generalization and false memory in acquired equivalence

Memory allows us to remember specific events but also combine information across events to infer new information. New inferences are thought to stem from integrating memories of related events during encoding but can be also generated on-demand, based on separate memories of individual events. Integrative encoding has been argued as dominant in the acquired equivalence paradigm, where people have a tendency to assume that when two faces share one preference, they also share another. A downside may be a loss of source memory, where inferred preferences are mistaken for observed ones. Here, we tested these predictions of the integrative encoding hypothesis across five datasets collected using variations of the acquired equivalence paradigm. Results showed a statistically reliable but numerically small tendency to generalize preferences across faces, with stronger evidence for on-demand inferences at retrieval rather than spontaneous integration during encoding. A newly included explicit source memory test showed that participants differentiated learned from inferred preferences to a high degree, irrespective of whether they generalized preferences across faces. False memory was however increased in participants who made generalization decisions faster, which could be consistent with integrative encoding and/or source monitoring frameworks. The results suggest that generalization in acquired equivalence may result from integrated representations that facilitate new inferences at the expense of source memory, but also demonstrate that on-demand retrieval-based processes may play a larger role in this paradigm than previously thought. Finally, the results indicate that reaction times may be more sensitive than performance as a means to assess representations underlying behavior. More broadly, the study informs current theories of generalization and knowledge representation and provides new insights into how memory biases decisions.

The binding structure of event elements in episodic memory and the role of animacy

Experienced events consist of several elements which need to be bound together in memory to represent the event in a coherent manner. Given such bindings, the retrieval of one event element should be related to the successful retrieval of another element of the same event, thus leading to a stochastic dependency of the retrieval of event elements. The way in which bindings are structured is not yet clearly established and only few moderators of the binding of event elements have been identified. We present results from three experiments aiming to distinguish between an integrated binding structure, in which event elements are bound into a unitary representation, and a hierarchical binding structure, in which event elements are preferentially bound to specific types of elements. Experiments 2 and 3 were additionally designed to identify animacy, an entity's property of being alive, as a potential moderator of the binding of event elements. We also offer a new approach for modelling dependencies of the retrieval of event elements which mitigates some limitations of previous approaches. Consistent with previous literature, we found dependencies of the retrieval of event elements if all of an event's constituent associations were shown. We found mixed evidence for integrated or hierarchical binding structures but found dependency of the retrieval of event elements to be sensitive to the presence of animacy in an event. The results suggest that binding structures may vary depending on moderators such as animacy or event structure awareness. Theoretical implications and directions for future research are discussed.

Flexible reuse of cortico-hippocampal representations during encoding and recall of naturalistic events

Although every life event is unique, there are considerable commonalities across events. However, little is known about whether or how the brain flexibly represents information about different event components at encoding and during remembering. Here, we show that different cortico-hippocampal networks systematically represent specific components of events depicted in videos, both during online experience and during episodic memory retrieval. Regions of an Anterior Temporal Network represented information about people, generalizing across contexts, whereas regions of a Posterior Medial Network represented context information, generalizing across people. Medial prefrontal cortex generalized across videos depicting the same event schema, whereas the hippocampus maintained event-specific representations. Similar effects were seen in real-time and recall, suggesting reuse of event components across overlapping episodic memories. These representational profiles together provide a computationally optimal strategy to scaffold memory for different high-level event components, allowing efficient reuse for event comprehension, recollection, and imagination.

Brain reserve contributes to distinguishing preclinical Alzheimer's stages 1 and 2

BACKGROUND

In preclinical Alzheimer's disease, it is unclear why some individuals with amyloid pathologic change are asymptomatic (stage 1), whereas others experience subjective cognitive decline (SCD, stage 2). Here, we examined the association of stage 1 vs. stage 2 with structural brain reserve in memory-related brain regions.

METHODS

We tested whether the volumes of hippocampal subfields and parahippocampal regions were larger in individuals at stage 1 compared to asymptomatic amyloid-negative older adults (healthy controls, HCs). We also tested whether individuals with stage 2 would show the opposite pattern, namely smaller brain volumes than in amyloid-negative individuals with SCD. Participants with cerebrospinal fluid (CSF) biomarker data and bilateral volumetric MRI data from the observational, multi-centric DZNE-Longitudinal Cognitive Impairment and Dementia Study (DELCODE) study were included. The sample comprised 95 amyloid-negative and 26 amyloid-positive asymptomatic participants as well as 104 amyloid-negative and 47 amyloid-positive individuals with SCD. Volumes were based on high-resolution T2-weighted images and automatic segmentation with manual correction according to a recently established high-resolution segmentation protocol.

RESULTS

In asymptomatic individuals, brain volumes of hippocampal subfields and of the parahippocampal cortex were numerically larger in stage 1 compared to HCs, whereas the opposite was the case in individuals with SCD. MANOVAs with volumes as dependent data and age, sex, years of education, and DELCODE site as covariates showed a significant interaction between diagnosis (asymptomatic versus SCD) and amyloid status (Aß42/40 negative versus positive) for hippocampal subfields. Post hoc paired comparisons taking into account the same covariates showed that dentate gyrus and CA1 volumes in SCD were significantly smaller in amyloid-positive than negative individuals. In contrast, CA1 volumes were significantly (p = 0.014) larger in stage 1 compared with HCs.

CONCLUSIONS

These data indicate that HCs and stages 1 and 2 do not correspond to linear brain volume reduction. Instead, stage 1 is associated with larger than expected volumes of hippocampal subfields in the face of amyloid pathology. This indicates a brain reserve mechanism in stage 1 that enables individuals with amyloid pathologic change to be cognitively normal and asymptomatic without subjective cognitive decline.

Representational integration and differentiation in the human hippocampus following goal-directed navigation

As we learn, dynamic memory processes build structured knowledge across our experiences. Such knowledge enables the formation of internal models of the world that we use to plan, make decisions, and act. Recent theorizing posits that mnemonic mechanisms of differentiation and integration - which at one level may seem to be at odds - both contribute to the emergence of structured knowledge. We tested this possibility using fMRI as human participants learned to navigate within local and global virtual environments over the course of 3 days. Pattern similarity analyses on entorhinal cortical and hippocampal patterns revealed evidence that differentiation and integration work concurrently to build local and global environmental representations, and that variability in integration relates to differences in navigation efficiency. These results offer new insights into the neural machinery and the underlying mechanisms that translate experiences into structured knowledge that allows us to navigate to achieve goals.

What about "space" is important for episodic memory?

Early cognitive neuroscientific research revealed that the hippocampus is crucial for spatial navigation in rodents, and for autobiographical episodic memory in humans. Researchers quickly linked these streams to propose that the human hippocampus supports memory through its role in representing space, and research on the link between spatial cognition and episodic memory in humans has proliferated over the past several decades. Different researchers apply the term "spatial" in a variety of contexts, however, and it remains unclear what aspect of space may be critical to memory. Similarly, "episodic" has been defined and tested in different ways. Naturalistic assessment of spatial memory and episodic memory (i.e., episodic autobiographical memory) is required to unify the scale and biological relevance in comparisons of spatial and mnemonic processing. Limitations regarding the translation of rodent to human research, human ontogeny, and inter-individual variability require greater consideration in the interpretation of this literature. In this review, we outline the aspects of space that are (and are not) commonly linked to episodic memory, and then we discuss these dimensions through the lens of individual differences in naturalistic autobiographical memory. Future studies should carefully consider which aspect(s) of space are being linked to memory within the context of naturalistic human cognition. This article is categorized under: Psychology > Memory.

A rubric for human-like agents and NeuroAI

Researchers across cognitive, neuro- and computer sciences increasingly reference 'human-like' artificial intelligence and 'neuroAI'. However, the scope and use of the terms are often inconsistent. Contributed research ranges widely from mimicking behaviour, to testing machine learning methods as neurally plausible hypotheses at the cellular or functional levels, or solving engineering problems. However, it cannot be assumed nor expected that progress on one of these three goals will automatically translate to progress in others. Here, a simple rubric is proposed to clarify the scope of individual contributions, grounded in their commitments to human-like behaviour, neural plausibility or benchmark/engineering/computer science goals. This is clarified using examples of weak and strong neuroAI and human-like agents, and discussing the generative, corroborate and corrective ways in which the three dimensions interact with one another. The author maintains that future progress in artificial intelligence will need strong interactions across the disciplines, with iterative feedback loops and meticulous validity tests-leading to both known and yet-unknown advances that may span decades to come. This article is part of a discussion meeting issue 'New approaches to 3D vision'.

Experience and maturation: The contribution of co-occurrence regularities in language to the development of semantic organization

With development knowledge becomes organized according to semantic links, including early-developing associative (e.g., juicy-apple) and gradually developing taxonomic links (e.g., apple-pear). Word co-occurrence regularities may foster these links: Associative links may form from direct co-occurrence (e.g., juicy-apple), and taxonomic links from shared co-occurrence (e.g., apple and pear co-occur with juicy). Four experiments (2017-2020) investigated this possibility with 4- to 8-year-olds (N = 148, 82 female) and adults (N = 116, 35 female) in a U.S. city with 58.6% White; 29.0% Black, and 5.8% Asian demographics. Results revealed earlier development of the abilities to form direct (ds > 0.536) than the abilities to form shared co-occurrence-based links (ds > 1.291). We argue that the asynchronous development of abilities to form co-occurrence-based links may explain developmental changes in semantic organization.

Face familiarity detection with complex synapses

Synaptic plasticity is a complex phenomenon involving multiple biochemical processes that operate on different timescales. Complexity can greatly increase memory capacity when the variables characterizing the synaptic dynamics have limited precision, as shown in simple memory retrieval problems involving random patterns. Here we turn to a real-world problem, face familiarity detection, and we show that synaptic complexity can be harnessed to store in memory a large number of faces that can be recognized at a later time. The number of recognizable faces grows almost linearly with the number of synapses and quadratically with the number of neurons. Complex synapses outperform simple ones characterized by a single variable, even when the total number of dynamical variables is matched. Complex and simple synapses have distinct signatures that are testable in experiments. Our results indicate that a system with complex synapses can be used in real-world tasks such as face familiarity detection.

A model of autonomous interactions between hippocampus and neocortex driving sleep-dependent memory consolidation

How do we build up our knowledge of the world over time? Many theories of memory formation and consolidation have posited that the hippocampus stores new information, then "teaches" this information to the neocortex over time, especially during sleep. But it is unclear, mechanistically, how this actually works-How are these systems able to interact during periods with virtually no environmental input to accomplish useful learning and shifts in representation? We provide a framework for thinking about this question, with neural network model simulations serving as demonstrations. The model is composed of hippocampus and neocortical areas, which replay memories and interact with one another completely autonomously during simulated sleep. Oscillations are leveraged to support error-driven learning that leads to useful changes in memory representation and behavior. The model has a non-rapid eye movement (NREM) sleep stage, where dynamics between the hippocampus and neocortex are tightly coupled, with the hippocampus helping neocortex to reinstate high-fidelity versions of new attractors, and a REM sleep stage, where neocortex is able to more freely explore existing attractors. We find that alternating between NREM and REM sleep stages, which alternately focuses the model's replay on recent and remote information, facilitates graceful continual learning. We thus provide an account of how the hippocampus and neocortex can interact without any external input during sleep to drive useful new cortical learning and to protect old knowledge as new information is integrated.

Memory as a scale of simulation depending on the trace distinctiveness

It has been proposed that representations emerge from a single memory system organized along a continuum of specificity. This continuum is assumed to reflect a scale between the simulation of overlapping and specific features of the traces, which depends on trace distinctiveness. More specifically, higher trace distinctiveness facilitates the simulation of trace-specific features, which increase the discriminability of traces and lead to the emergence of a more specific representation. In two experiments, participants were asked to identify match (low task discrimination demand) or mismatch (high task discrimination demand) associations between actions and characters that were visually either highly or lowly distinctive. The results of Experiment 1 show that in the high-distinctiveness context, performance was better when identifying a mismatch rather than a match, while the opposite was true in the low-distinctiveness context. The results of Experiment 2 show that using a dynamic visual noise to interfere with the participants' ability to simulate the features of the characters also reduced the benefit of the high-distinctiveness context for the mismatch trials (Experiment 2a and 2b) and increased the benefit of the low-distinctiveness context for the match trials (Experiment 2b). Taken together, these results suggest that the simulation of trace-specific features underlies the emergence of specific representations, which can be beneficial when the discrimination demand of the task is high and detrimental when this demand is low. Memory might therefore be viewed as a scale of simulation between overlapping and specific trace features.

Correcting the hebbian mistake: Toward a fully error-driven hippocampus

The hippocampus plays a critical role in the rapid learning of new episodic memories. Many computational models propose that the hippocampus is an autoassociator that relies on Hebbian learning (i.e., "cells that fire together, wire together"). However, Hebbian learning is computationally suboptimal as it does not learn in a way that is driven toward, and limited by, the objective of achieving effective retrieval. Thus, Hebbian learning results in more interference and a lower overall capacity. Our previous computational models have utilized a powerful, biologically plausible form of error-driven learning in hippocampal CA1 and entorhinal cortex (EC) (functioning as a sparse autoencoder) by contrasting local activity states at different phases in the theta cycle. Based on specific neural data and a recent abstract computational model, we propose a new model called Theremin (Total Hippocampal ERror MINimization) that extends error-driven learning to area CA3-the mnemonic heart of the hippocampal system. In the model, CA3 responds to the EC monosynaptic input prior to the EC disynaptic input through dentate gyrus (DG), giving rise to a temporal difference between these two activation states, which drives error-driven learning in the EC→CA3 and CA3↔CA3 projections. In effect, DG serves as a teacher to CA3, correcting its patterns into more pattern-separated ones, thereby reducing interference. Results showed that Theremin, compared with our original Hebbian-based model, has significantly increased capacity and learning speed. The model makes several novel predictions that can be tested in future studies.

Hippocampal and medial prefrontal cortices encode structural task representations following progressive and interleaved training schedules

Memory generalisations may be underpinned by either encoding- or retrieval-based generalisation mechanisms and different training schedules may bias some learners to favour one of these mechanisms over the other. We used a transitive inference task to investigate whether generalisation is influenced by progressive vs randomly interleaved training, and overnight consolidation. On consecutive days, participants learnt pairwise discriminations from two transitive hierarchies before being tested during fMRI. Inference performance was consistently better following progressive training, and for pairs further apart in the transitive hierarchy. BOLD pattern similarity correlated with hierarchical distances in the left hippocampus (HIP) and medial prefrontal cortex (MPFC) following both training schedules. These results are consistent with the use of structural representations that directly encode hierarchical relationships between task features. However, such effects were only observed in the MPFC for recently learnt relationships. Furthermore, the MPFC appeared to maintain structural representations in participants who performed at chance on the inference task. We conclude that humans preferentially employ encoding-based mechanisms to store map-like relational codes that can be used for memory generalisation. These codes are expressed in the HIP and MPFC following both progressive and interleaved training but are not sufficient for accurate inference.

Self-derivation of new knowledge through memory integration varies as a function of prior knowledge

ABSTRACTThe current research was an investigation of the effects of prior, domain-relevant knowledge on self-derivation of new, integrated knowledge. Adults were presented with novel "stem" facts and tested for self-derivation of new knowledge through integration of the facts in memory. To examine the effects of prior knowledge on memory integration, we tested participants under three within-subject conditions, in which in advance of stem encoding, they were provided with either: (1) no domain relevant information (No Knowledge control), (2) generally relevant information about the domain (General), or (3) generally relevant information about the domain along with a stem fact that was directly necessary for self-derivation (General + Stem). Prior exposure to both General and General + Stem knowledge facilitated memory for the novel, explicitly-taught stem facts. Moreover, for both prior knowledge conditions, the amount of domain-relevant knowledge retained in memory was associated with trial-by-trial self-derivation success. Importantly, the type of prior knowledge modulated memory integration in different ways. Whereas General + Stem knowledge enhanced selective retrieval and integration of the stem facts, General knowledge supported learning of the individual stem facts, but not their integration with one another. Together, the findings indicate how malleable, domain-specific experience shapes encoding, integration, and flexible extension of new, related information.

Stress-induced changes of the cholinergic circuitry promote retrieval-based generalization of aversive memories

Generalization, the process of applying knowledge acquired in one context to other contexts, often drives the expression of similar behaviors in related situations. At the cellular level, generalization is thought to depend on the activity of overlapping neurons that represent shared features between contexts (general representations). Using contextual fear conditioning in mice, we demonstrate that generalization can also occur in response to stress and result from reactivation of specific, rather than general context representations. We found that generalization emerges during memory retrieval, along with stress-induced abnormalities of septohippocampal oscillatory activity and acetylcholine release, which are typically found in negative affective states. In hippocampal neurons that represent aversive memories and drive generalization, cholinergic septohippocampal afferents contributed to a unique reactivation pattern of cFos, Npas4, and repressor element-1 silencing transcription factor (REST). Together, these findings suggest that generalization can be triggered by perceptually dissimilar but valence-congruent memories of specific aversive experiences. Through promoting the reactivation of such memories and their interference with ongoing behavior, abnormal cholinergic signaling could underlie maladaptive cognitive and behavioral generalization linked to negative affective states.

Mnemonic construction and representation of temporal structure in the hippocampal formation

The hippocampal-entorhinal region supports memory for episodic details, such as temporal relations of sequential events, and mnemonic constructions combining experiences for inferential reasoning. However, it is unclear whether hippocampal event memories reflect temporal relations derived from mnemonic constructions, event order, or elapsing time, and whether these sequence representations generalize temporal relations across similar sequences. Here, participants mnemonically constructed times of events from multiple sequences using infrequent cues and their experience of passing time. After learning, event representations in the anterior hippocampus reflected temporal relations based on constructed times. Temporal relations were generalized across sequences, revealing distinct representational formats for events from the same or different sequences. Structural knowledge about time patterns, abstracted from different sequences, biased the construction of specific event times. These findings demonstrate that mnemonic construction and the generalization of relational knowledge combine in the hippocampus, consistent with the simulation of scenarios from episodic details and structural knowledge.

Schematic information influences memory and generalisation behaviour for schema-relevant and -irrelevant information

Schemas modulate memory performance for schema-congruent and -incongruent information. However, it is assumed they do not influence behaviour for information irrelevant to themselves. We assessed memory and generalisation behaviour for information related to an underlying pattern, where a schema could be extracted (schema-relevant), and information that was unrelated and therefore irrelevant to the extracted schema (schema-irrelevant). Using precision measures of long-term memory, where participants learnt associations between words and locations around a circle, we assessed memory and generalisation for schema-relevant and -irrelevant information. Words belonged to two semantic categories: human-made and natural. For one category, word-locations were clustered around one point on the circle (clustered condition), while the other category had word-locations randomly distributed (non-clustered condition). The presence of an underlying pattern in the clustered condition allows for the extraction of a schema that can support both memory and generalisation. At test, participants were presented with old (memory) and new (generalisation) words, requiring them to identify a remembered location or make a best guess. The presence of the clustered pattern modulated memory and generalisation. In the clustered condition, participants placed old and new words in locations consistent with the underlying pattern. In contrast, for the non-clustered condition, participants were less likely to place old and new non-clustered words in locations consistent with the clustered condition. Therefore, we provide evidence that the presence of schematic information modulates memory and generalisation for schema-relevant and -irrelevant information. Our results highlight the need to carefully construct appropriate schema-irrelevant control conditions such that behaviour in these conditions is not modulated by the presence of a schema. Theoretically, models of schema processing need to account for how the presence of schematic information can have consequences for information that is irrelevant to itself.