Overlapping memory replay during sleep builds cognitive schemata

Reactivation of schema representation in lateral occipital cortex supports successful memory encoding

Schemas provide a scaffold onto which we can integrate new memories. Previous research has investigated the brain activity and connectivity underlying schema-related memory formation. However, how schemas are represented and reactivated in the brain, in order to enhance memory, remains unclear. To address this issue, we used an object–location spatial schema that was learned over multiple sessions, combined with similarity analyses of neural representations, to investigate the reactivation of schema representations of object–location memories when a new object–scene association is learned. In addition, we investigated how this reactivation affects subsequent memory performance under different strengths of schemas. We found that reactivation of a schema representation in the lateral occipital cortex (LOC) during object–scene encoding affected subsequent associative memory performance only in the schema-consistent condition and increased the functional connectivity between the LOC and the parahippocampal place area. Taken together, our findings provide new insight into how schema acts as a scaffold to support the integration of novel information into existing cortical networks and suggest a neural basis for schema-induced rapid cortical learning.

Sleep-like unsupervised replay reduces catastrophic forgetting in artificial neural networks

Artificial neural networks are known to suffer from catastrophic forgetting: when learning multiple tasks sequentially, they perform well on the most recent task at the expense of previously learned tasks. In the brain, sleep is known to play an important role in incremental learning by replaying recent and old conflicting memory traces. Here we tested the hypothesis that implementing a sleep-like phase in artificial neural networks can protect old memories during new training and alleviate catastrophic forgetting. Sleep was implemented as off-line training with local unsupervised Hebbian plasticity rules and noisy input. In an incremental learning framework, sleep was able to recover old tasks that were otherwise forgotten. Previously learned memories were replayed spontaneously during sleep, forming unique representations for each class of inputs. Representational sparseness and neuronal activity corresponding to the old tasks increased while new task related activity decreased. The study suggests that spontaneous replay simulating sleep-like dynamics can alleviate catastrophic forgetting in artificial neural networks.

No evidence for a preferential role of sleep in episodic memory abstraction

Substantial evidence suggests that sleep has a role in declarative memory consolidation. An influential notion holds that such sleep-related memory consolidation is associated with a process of abstraction. The neural underpinnings of this putative process are thought to involve a hippocampo-neocortical dialogue. Specifically, the idea is that, during sleep, the statistical contingencies across episodes are re-coded to a less hippocampus-dependent format, while at the same time losing configural information. Two previous studies from our lab, however, failed to show a preferential role of sleep in either episodic memory decontextualisation or the formation of abstract knowledge across episodic exemplars. Rather these processes occurred over sleep and wake time alike. Here, we present two experiments that replicate and extend these previous studies and exclude some alternative interpretations. The combined data show that sleep has no preferential function in this respect. Rather, hippocampus-dependent memories are generalised to an equal extent across both wake and sleep time. The one point on which sleep outperforms wake is actually the preservation of episodic detail of memories stored prior to sleep.

Learning new meanings for known L2 words: Long‐term semantic representation is updated to integrate new information after consolidation

Previous research about learning new meanings for known words in second language (L2) has found that semantic relatedness, i.e., congruency, between new and existing meanings benefits encoding and explicit memory of new meanings, and reduces instant interference on accessing existing meanings. However, they did not take the memory consolidation process into account. Thus, integration of new meaning into long-term semantic memory, update of existing meaning representation, and the impact of semantic relatedness between new and existing meanings in this process remain unclear. The present study used the event-related potential (ERP) technique to explore these questions. We asked Chinese students to learn English known words' subdominant meanings variedly related to existing meanings and probed semantic representations with EEG recorded in primed lexical decision tasks four times before and after consolidation. We found that new meaning needs to go through offline consolidation to get integrated. Semantic relatedness/congruency boosted new meaning integration, not by directly expediting it during encoding or preliminary offline consolidation, but by promoting the update of existing meaning representation first, which presumably paved the way for better incorporation of new meaning in the long run. The whole pattern of results implies that long-term semantic representation of existing meaning is updated to integrate related new meaning after consolidation, which not only draws a clearer picture of L2 ambiguous word acquisition but also bears broader implications for research on memory updating.

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.

Contributions of memory and brain development to the bioregulation of naps and nap transitions in early childhood

The transition from multiple sleep bouts each day to a single overnight sleep bout (i.e., nap transition) is a universal process in human development. Naps are important during infancy and early childhood as they enhance learning through memory consolidation. However, a normal part of development is the transition out of naps. Understanding nap transitions is essential in order to maximize early learning and promote positive long-term cognitive outcomes. Here, we propose a novel hypothesis regarding the cognitive, physiological, and neural changes that accompany nap transitions. Specifically, we posit that maturation of the hippocampal-dependent memory network results in more efficient memory storage, which reduces the buildup of homeostatic sleep pressure across the cortex (as reflected by slow-wave activity), and eventually, contributes to nap transitions. This hypothesis synthesizes evidence of bioregulatory mechanisms underlying nap transitions and sheds new light on an important window of change in development. This framework can be used to evaluate multiple untested predictions from the field of sleep science and ultimately, yield science-based guidelines and policies regarding napping in childcare and early education settings.

Targeted memory reactivation during sleep can induce forgetting of overlapping memories

Memory reactivation during sleep can shape new memories into a long-term form. Reactivation of memories can be induced via the delivery of auditory cues during sleep. Although this targeted memory reactivation (TMR) approach can strengthen newly acquired memories, research has tended to focus on single associative memories. It is less clear how TMR affects retention for overlapping associative memories. This is critical, given that repeated retrieval of overlapping associations during wake can lead to forgetting, a phenomenon known as retrieval-induced forgetting (RIF). We asked whether a similar pattern of forgetting occurs when TMR is used to cue reactivation of overlapping pairwise associations during sleep. Participants learned overlapping pairs—learned separately, interleaved with other unrelated pairs. During sleep, we cued a subset of overlapping pairs using TMR. While TMR increased retention for the first encoded pairs, memory decreased for the second encoded pairs. This pattern of retention was only present for pairs not tested prior to sleep. The results suggest that TMR can lead to forgetting, an effect similar to RIF during wake. However, this effect did not extend to memories that had been strengthened via retrieval prior to sleep. We therefore provide evidence for a reactivation-induced forgetting effect during sleep.

Schemas provide a scaffold for neocortical integration of new memories over time

Memory transformation is increasingly acknowledged in theoretical accounts of systems consolidation, yet how memory quality and neural representation change over time and how schemas influence this process remains unclear. We examined the behavioral quality and neural representation of schema-congruent and incongruent object-scene pairs retrieved across 10-minutes and 72-hours using fMRI. When a congruent schema was available, memory became coarser over time, aided by post-encoding coupling between the anterior hippocampus and medial prefrontal cortex (mPFC). Only schema-congruent representations were integrated in the mPFC over time, and were organized according to schematic context. In the hippocampus, pattern similarity changed across 72-hours such that the posterior hippocampus represented specific details and the anterior hippocampus represented the general context of specific memories, irrespective of congruency. Our findings suggest schemas are used as a scaffold to facilitate neocortical integration of congruent information, and illustrate evolution in hippocampal organization of detailed contextual memory over time.

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.

Sleep targets highly connected global and local nodes to aid consolidation of learned graph networks

Much of our long-term knowledge is organised in complex networks. Sleep is thought to be critical for abstracting knowledge and enhancing important item memory for long-term retention. Thus, sleep should aid the development of memory for networks and the abstraction of their structure for efficient storage. However, this remains unknown because past sleep studies have focused on discrete items. Here we explored the impact of sleep (night-sleep/day-wake within-subject paradigm with 25 male participants) on memory for graph-networks where some items were important due to dense local connections (degree centrality) or, independently, important due to greater global connections (closeness/betweenness centrality). A network of 27 planets (nodes) sparsely interconnected by 36 teleporters (edges) was learned via discrete associations without explicit indication of any network structure. Despite equivalent exposure to all connections in the network, we found that memory for the links between items with high local connectivity or high global connectivity were better retained after sleep. These results highlight that sleep has the capacity for strengthening both global and local structure from the world and abstracting over multiple experiences to efficiently form internal networks of knowledge.

Two distinct ways to form long-term object recognition memory during sleep and wakefulness

Memory consolidation is promoted by sleep. However, there is also evidence for consolidation into long-term memory during wakefulness via processes that preferentially affect nonhippocampal representations. We compared, in rats, the effects of 2-h postencoding periods of sleep and wakefulness on the formation of long-term memory for objects and their associated environmental contexts. We employed a novel-object recognition (NOR) task, using object exploration and exploratory rearing as behavioral indicators of these memories. Remote recall testing (after 1 wk) confirmed significant long-term NOR memory under both conditions, with NOR memory after sleep predicted by the occurrence of EEG spindle-slow oscillation coupling. Rats in the sleep group decreased their exploratory rearing at recall testing, revealing successful recall of the environmental context. By contrast, rats that stayed awake after encoding showed equally high levels of rearing upon remote testing as during encoding, indicating that context memory was lost. Disruption of hippocampal function during the postencoding interval (by muscimol administration) suppressed long-term NOR memory together with context memory formation when animals slept, but enhanced NOR memory when they were awake during this interval. Testing remote recall in a context different from that during encoding impaired NOR memory in the sleep condition, while exploratory rearing was increased. By contrast, NOR memory in the wake rats was preserved and actually superior to that after sleep. Our findings indicate two distinct modes of long-term memory formation: Sleep consolidation is hippocampus dependent and implicates event-context binding, whereas wake consolidation is impaired by hippocampal activation and strengthens context-independent representations.

Flexibility of memory for future-oriented cognition

Memories of daily experiences contain incidental details unique to each experience as well as common latent patterns shared with others. Neural representations focusing on the latter aspect can be reinstated by similar new experiences even though their perceptual features do not match the original experiences perfectly. Such flexible memory use allows for faster learning and better decision-making in novel situations. Here, I review evidence from rodent and primate electrophysiological studies to discuss how memory flexibility is implemented in the spiking activity of neuronal ensembles. These findings uncovered innate and learned coding properties and their potential refinement during sleep that support flexible integration and application of memories for better future adaptation.

Selective engram coreactivation in idling brain inspires implicit learning

Passive priming of prior knowledge to assimilate ongoing experiences underlies advanced cognitive processing. However, the necessary neural dynamics of memory assimilation remains elusive. Uninstructed brain could also show boosted creativity, particularly after idling states, yet it remains unclear whether the idling brain can spontaneously spark relevant knowledge assimilations. We established a paradigm that links/separates context-dependent memories according to geometrical similarities. Mice exploring one of four contexts 1 d before undergoing contextual fear conditioning in a square context showed a gradual fear transfer to preexposed geometrically relevant contexts the next day, but not after 15 min. Anterior cingulate cortex neurons representing relevant, rather than distinct, memories were significantly coreactivated during postconditioning sleep only, before their selective integration the next day during testing. Disrupting sleep coreactivations prevented assimilation while preserving recent memory consolidation. Thus, assimilating pertinent memories during sleep through coreactivation of their respective engrams represents the neural underpinnings of sleep-triggered implicit cortical learning.

Effects of Information Load on Schema and Episodic Memory Formation

The formation of semantic memories is assumed to result from the abstraction of general, schema-like knowledge across multiple experiences, while at the same time, episodic details from individual experiences are forgotten. Against this backdrop, our study examined the effects of information load (high vs. low) during encoding on the formation of episodic and schema memory using an elaborated version of an object-place recognition (OPR) task in rats. The task allowed for the abstraction of a spatial rule across four (low information load) or eight (high information load) encoding episodes (spaced apart by a 20 min interval) in which the rats could freely explore two objects in an open field arena. After this encoding phase, animals were left undisturbed for 24 h and then tested either for the expression of schema memory, i.e., for the spatial rule, or memory for an individual encoding episode. Rats in the high information load condition exhibited a more robust schema memory for the spatial rule than in the low information load condition. In contrast, rats in the low load condition showed more robust memory for individual learning episodes than in the high information load condition. Our findings of opposing effects might point to an information-load-dependent competitive relationship between processes of schema and episodic memory formation, although other explanations are possible.

Sleep bolsters schematically incongruent memories

Our ability to recall memories is improved when sleep follows learning, suggesting that sleep facilitates memory consolidation. A number of factors are thought to influence the impact of sleep on newly learned information, such as whether or not we rehearse that information (e.g. via restudy or retrieval practice), or the extent to which the information is consistent with our pre-existing schematic knowledge. In this pre-registered, online study, we examined the effects of both rehearsal and schematic congruency on overnight consolidation. Participants learned noun-colour pairings (e.g. elephant-red) and rated each pairing as plausible or implausible before completing a baseline memory assessment. Afterwards, participants engaged in a period of restudy or retrieval practice for the pairings, and then entered a 12 h retention interval of overnight sleep or daytime wakefulness. Follow-up assessments were completed immediately after sleep or wake, and again 24 h after learning. Our data indicated that overnight consolidation was amplified for restudied relative to retested noun-colour pairings, but only when sleep occurred soon after learning. Furthermore, whereas plausible (i.e. schematically congruent) pairings were generally better remembered than implausible (i.e. schematically incongruent) pairings, the benefits of sleep were stronger for implausible relative to plausible memories. These findings challenge the notion that schema-conformant memories are preferentially strengthened during post-learning sleep.

Theta Activity During Encoding Interacts With NREM Sleep Oscillations to Predict Memory Generalization

Relatively little is known regarding the interaction between encoding-related neural activity and sleep-based memory consolidation. One suggestion is that a function of encoding-related theta power may be to “tag” memories for subsequent processing during sleep. This study aimed to extend previous work on the relationships between sleep spindles, slow oscillation-spindle coupling, and task-related theta activity with a combined Deese-Roediger-McDermott (DRM) and nap paradigm. This allowed us to examine the influence of task- and sleep-related oscillatory activity on the recognition of both encoded list words and associative theme words. Thirty-three participants (29 females, mean age = 23.2 years) learned and recognised DRM lists separated by either a 2 h wake or sleep period. Mixed-effects modelling revealed the sleep condition endorsed more associative theme words and fewer list words in comparison to the wake group. Encoding-related theta power was also found to influence sleep spindle density, and this interaction was predictive of memory outcomes. The influence of encoding-related theta was specific to sleep spindle density, and did not appear to influence the strength of slow oscillation-spindle coupling as it relates to memory outcomes. The finding of interactions between wakeful and sleep oscillatory-related activity in promoting memory and learning has important implications for theoretical models of sleep-based memory consolidation.

A sleep schedule incorporating naps benefits the transformation of hierarchical knowledge

Study Objectives

The learning brain establishes schemas (knowledge structures) that benefit subsequent learning. We investigated how sleep and having a schema might benefit initial learning followed by rearranged and expanded memoranda. We concurrently examined the contributions of sleep spindles and slow-wave sleep to learning outcomes.

Methods

Fifty-three adolescents were randomly assigned to an 8 h Nap schedule (6.5 h nocturnal sleep with a 90-minute daytime nap) or an 8 h No-Nap, nocturnal-only sleep schedule. The study spanned 14 nights, simulating successive school weeks. We utilized a transitive inference task involving hierarchically ordered faces. Initial learning to set up the schema was followed by rearrangement of the hierarchy (accommodation) and hierarchy expansion (assimilation). The expanded sequence was restudied. Recall of hierarchical knowledge was tested after initial learning and at multiple points for all subsequent phases. As a control, both groups underwent a No-schema condition where the hierarchy was introduced and modified without opportunity to set up a schema. Electroencephalography accompanied the multiple sleep opportunities.

Results

There were main effects of Nap schedule and Schema condition evidenced by superior recall of initial learning, reordered and expanded memoranda. Improved recall was consistently associated with higher fast spindle density but not slow-wave measures. This was true for both nocturnal sleep and daytime naps.

Conclusion

A sleep schedule incorporating regular nap opportunities compared to one that only had nocturnal sleep benefited building of robust and flexible schemas, facilitating recall of the subsequently rearranged and expanded structured knowledge. These benefits appear to be strongly associated with fast spindles.

Clinical Trial registration

NCT04044885 (https://clinicaltrials.gov/ct2/show/NCT04044885).

Learning cortical representations through perturbed and adversarial dreaming

Humans and other animals learn to extract general concepts from sensory experience without extensive teaching. This ability is thought to be facilitated by offline states like sleep where previous experiences are systemically replayed. However, the characteristic creative nature of dreams suggests that learning semantic representations may go beyond merely replaying previous experiences. We support this hypothesis by implementing a cortical architecture inspired by generative adversarial networks (GANs). Learning in our model is organized across three different global brain states mimicking wakefulness, non-rapid eye movement (NREM), and REM sleep, optimizing different, but complementary, objective functions. We train the model on standard datasets of natural images and evaluate the quality of the learned representations. Our results suggest that generating new, virtual sensory inputs via adversarial dreaming during REM sleep is essential for extracting semantic concepts, while replaying episodic memories via perturbed dreaming during NREM sleep improves the robustness of latent representations. The model provides a new computational perspective on sleep states, memory replay, and dreams, and suggests a cortical implementation of GANs.

Narratives bridge the divide between distant events in episodic memory

Many studies suggest that information about past experience, or episodic memory, is divided into discrete units called "events." Yet we can often remember experiences that span multiple events. Events that occur in close succession might simply be linked because of their proximity to one another, but we can also build links between events that occur farther apart in time. Intuitively, some kind of organizing principle should enable temporally distant events to become bridged in memory. We tested the hypothesis that episodic memory exhibits a narrative-level organization, enabling temporally distant events to be better remembered if they form a coherent narrative. Furthermore, we tested whether post-encoding memory consolidation is necessary to integrate temporally distant events. In three experiments, participants learned and subsequently recalled events from fictional stories, in which pairs of temporally distant events involving side characters ("sideplots") either formed one coherent narrative or two unrelated narratives. Across participants, we varied whether recall was assessed immediately after learning, or after a delay: 24 hours, 12 hours between morning and evening ("wake"), or 12 hours between evening and morning ("sleep"). Participants recalled more information about coherent than unrelated narrative events, in most delay conditions, including immediate recall and wake conditions, suggesting that post-encoding consolidation was not necessary to integrate temporally distant events into a larger narrative. Furthermore, post hoc modeling across experiments suggested that narrative coherence facilitated recall over and above any effects of sentence-level semantic similarity. This reliable memory benefit for coherent narrative events supports theoretical accounts which propose that narratives provide a high-level architecture for episodic memory.

Sleep Facilitates Extraction of Temporal Regularities With Varying Timescales

Evidence suggests that memory consolidation is facilitated by sleep, both through the strengthening of existing memories and by extracting regularities embedded in those memories. We previously observed that one sleep stage, Slow-Wave sleep (SWS), is particularly involved in the extraction of temporal regularities. We suggested that this attribute can naturally stem from the time-compressed memory replay known to occur in the hippocampus during SWS. A prediction coming out of this “temporal scaffolding” hypothesis is that sleep would be especially influential on extraction of temporal regularities when the time gap between the events constituting the regularities is shortish. In this study, we tested this prediction. Eighty-three participants performed a cognitive task in which hidden temporal regularities of varying time gaps were embedded. Detecting these regularities could significantly improve performance. Participants performed the task in two sessions with an interval filled with either wake or sleep in between. We found that sleep improved performance across all time gaps and that the longer the gap had been, the smaller was the improvement across both sleep and wake. No interaction between sleep and gap size was observed; however, unlike sleeping participants, awake participants did not exhibit any further performance improvement for the long gaps following the interval. In addition, across all participants, performance for the long gaps was associated with the development of conscious awareness to the regularities. We discuss these results in light of the temporal scaffolding hypothesis and suggest future directions to further elucidate the mechanisms involved.

Constructive episodic simulation in dreams

Memories of the past help us adaptively respond to similar situations in the future. Originally described by Schacter & Addis in 2007, the "constructive episodic simulation" hypothesis proposes that waking thought combines fragments of various past episodes into imagined simulations of events that may occur in the future. This same framework may be useful for understanding the function of dreaming. N = 48 college students were asked to identify waking life sources for a total of N = 469 dreams. Participants frequently traced dreams to at least one past or future episodic source (53.5% and 25.7% of dreams, respectively). Individual dreams were very often traced to multiple waking sources (43.9% of all dreams with content), with fragments of past memory incorporated into scenarios that anticipated future events. Waking-life dream sources are described in terms of their phenomenology and distribution across time and sleep stage, providing new evidence that dreams not only reflect the past, but also utilize memory in simulating potential futures.

Effects of Sleep on Language and Motor Consolidation: Evidence of Domain General and Specific Mechanisms

The current study explores the effects of time and sleep on the consolidation of a novel language learning task containing both item-specific knowledge and the extraction of grammatical regularities. We also compare consolidation effects in language and motor sequence learning tasks, to ask whether consolidation mechanisms are domain general. Young adults learned to apply plural inflections to novel words based on morphophonological rules embedded in the input, and learned to type a motor sequence using a keyboard. Participants were randomly assigned into one of two groups, practicing each task during either the morning or evening hours. Both groups were retested 12 and 24 hours post-training. Performance on frequent trained items in the language task stabilized only following sleep, consistent with a hippocampal mechanism for item-specific learning. However, regularity extraction, indicated by generalization to untrained items in the linguistic task, as well as performance on motor sequence learning, improved 24 hours post-training, irrespective of the timing of sleep. This consolidation process is consistent with a frontostriatal skill-learning mechanism, common across the language and motor domains. This conclusion is further reinforced by cross-domain correlations at the individual level between improvement across 24 hours in the motor task and in the low-frequency trained items in the linguistic task, which involve regularity extraction. Taken together, our results at the group and individual levels suggest that some aspects of consolidation are shared across the motor and language domains, and more specifically, between motor sequence learning and grammar learning.

No difference between slow oscillation up- and down-state cueing for memory consolidation during sleep

The beneficial effects of sleep for memory consolidation are assumed to rely on the reactivation of memories in conjunction with the coordinated interplay of sleep rhythms like slow oscillations and spindles. Specifically, slow oscillations are assumed to provide the temporal frame for spindles to occur in the slow oscillations up-states, enabling a redistribution of reactivated information within hippocampal-neocortical networks for long-term storage. Memory reactivation can also be triggered externally by presenting learning-associated cues (like odours or sounds) during sleep, but it is presently unclear whether there is an optimal time-window for the presentation of such cues in relation to the phase of the slow oscillations. In the present within-subject comparison, participants (n = 16) learnt word-pairs visually presented with auditory cues of the first syllable. These syllables were subsequently used for real-time cueing either in the up- or down-state of endogenous slow oscillations. Contrary to our hypothesis, we found differences in memory performance neither between up- and down-state cueing, nor between word-pairs that were cued versus uncued. In the up-state cueing condition, higher amounts of rapid eye movement sleep were associated with better memory for cued contents, whereas higher amounts of slow-wave sleep were associated with better memory for uncued contents. Evoked response analyses revealed signs of cue processing in both conditions. Interestingly, both up- and down-state cueing evoked a similar spindle response with the induced slow oscillations up-state at ~1000 ms post-cue. We speculate that our cueing procedure triggered generalised reactivation processes that facilitated the consolidation of both cued and uncued memories irrespective of the slow oscillation phase.

Sleep influences neural representations of true and false memories: An event-related potential study

Episodic memory is reconstructive and is thus prone to false memory formation. Although false memories are proposed to develop via associative processes, the nature of their neural representations, and the effect of sleep on false memory processing is currently unclear. The present research employed the Deese-Roediger-McDermott (DRM) paradigm and a daytime nap to determine whether semantic false memories and true memories could be differentiated using event-related potentials (ERPs). We also sought to illuminate the role of sleep in memory formation and learning. Healthy participants (N = 34, 28F, mean age = 23.23, range = 18-33) completed the learning phase of the DRM task followed by an immediate and a delayed recognition phase. The two recognition phases were separated by either a 2hr daytime nap or an equivalent wake period. Linear mixed modelling of effects at delayed recognition revealed larger LPC amplitudes for true memories in contrast to false memories for those in the wake group, and larger P300 amplitudes for false compared to true memories across sleep and wake groups. Larger LPC amplitudes for true memories were associated with enhanced true memory recognition following sleep, whilst larger P300 amplitudes were associated with similar true and false memory recognition rates. These findings are argued to reflect sleep's ability to promote memory generalisation associated with pattern completion, whilst also enhancing true memory recognition when memory traces have a strong episodic basis (linked to pattern separation). The present research suggests that true and false memories have differing neural profiles and are reflective of adaptive memory processes.

Sleep and incubation: Using problem reactivation during sleep to study forgetting fixation and unconscious processing during sleep incubation

When people are stuck on a problem, they sometimes benefit from an incubation period -a break from working on the problem. Anecdotes and empirical evidence suggest that sleeping during incubation is useful, but the mechanisms remain poorly understood. We examined how targeted memory reactivation during sleep, which boosts next-day solving, relates to forgetting fixation, a well-supported explanation of awake incubation. In evening sessions, participants attempted puzzles, while a unique sound cue played during each puzzle. Half the time, puzzles included fixating information reinforcing an incorrect representation. Later, during deep sleep, sounds associated with half of participants' previously unsolved puzzles were presented. The sounds should strengthen puzzle memories and reduce forgetting of the fixating information. In morning solving, overnight cueing reliably interacted with fixating information: participants solved numerically more cued than uncued puzzles, but only when puzzles included fixating information. These results suggest that additional processing occurred beyond simple fixation forgetting.

Future-relevant memories are not selectively strengthened during sleep

Overnight consolidation processes are thought to operate in a selective manner, such that important (i.e. future-relevant) memories are strengthened ahead of irrelevant information. Using an online protocol, we sought to replicate the seminal finding that the memory benefits of sleep are enhanced when people expect a future test [Wilhelm et al., 2011]. Participants memorised verbal paired associates to a criterion of 60 percent (Experiment 1) or 40 percent correct (Experiment 2) before a 12-hour delay containing overnight sleep (sleep group) or daytime wakefulness (wake group). Critically, half of the participants were informed that they would be tested again the following day, whereas the other half were told that they would carry out a different set of tasks. We observed a robust memory benefit of overnight consolidation, with the sleep group outperforming the wake group in both experiments. However, knowledge of an upcoming test had no impact on sleep-associated consolidation in either experiment, suggesting that overnight memory processes were not enhanced for future-relevant information. These findings, together with other failed replication attempts, show that sleep does not provide selective support to memories that are deemed relevant for the future.

Stability of neural encoding moderates the contribution of sleep and repeated testing to memory consolidation

There is evidence suggesting that online consolidation during retrieval-mediated learning interacts with offline consolidation during subsequent sleep to transform memory. Here we investigate whether this interaction persists when retrieval-mediated learning follows post-training sleep and whether the direction of this interaction is conditioned by the quality of encoding resulting from manipulation of the amount of sleep on the previous night. The quality of encoding was determined by computing the degree of similarity between EEG-activity patterns across restudy of face pairs in two groups of young participants, one who slept the last 4 h of the pre-training night, and another who slept 8 h. The offline consolidation was assessed by computing the degree of coupling between slow oscillations (SOs) and spindles (SPs) during post-training sleep, while the online consolidation was evaluated by determining the degree of similarity between EEG-activity patterns recorded during the study phase and during repeated recognition of either the same face pair (i.e., specific similarity) or face pairs sharing sex and profession (i.e., categorical similarity) to evaluate differentiation and generalization, respectively. The study and recognition phases were separated by a night of normal sleep duration. Mixed-effects models revealed that the stability of neural encoding moderated the relationship between sleep- and retrieval-mediated consolidation processes over left frontal regions. For memories showing lower encoding stability, the enhanced SO-SP coupling was associated with increased reinstatement of category-specific encoding-related activity at the expense of content-specific activity, whilst the opposite occurred for memories showing greater encoding stability. Overall, these results suggest that offline consolidation during post-training sleep interacts with online consolidation during retrieval the next day to favor the reorganization of memory contents, by increasing specificity of stronger memories and generalization of the weaker ones.

Memory decay distinguishes subtypes of gist

Memories are thought to become more gist-like over time. Multiple related memories might form generalized memory representations, losing specific details but enhancing or retaining gist. The time course within which gist memory emerges, however, is the subject of less consensus. To address this question, we focused our design on four kinds of gist: inferential gist (relations extracted across non-contiguous events), statistical learning (regularities extracted from a series), summary gist (a theme abstracted from a temporally contiguous series of items), and category gist (characterization of a stimulus at a higher level in the semantic hierarchy). Seventy participants completed memory encoding tasks addressing these types of gist and corresponding retrieval tasks the same evening, the morning after, and one week later, as well as an MRI at a later time point. We found little evidence that gist slowly emerges over time or that gist traces are more resistant to forgetting than detail traces. Instead, we found that initial gist memory shortly after encoding was either retained over time or decayed. Inferential gist and statistical learning were retained over a week, whereas memory for category and summary gist decayed. We discuss several interpretations for differences between these two subtypes of gist. Individual differences in REM or slow-wave sleep and hippocampal volumes did not predict changes in memory for these four kinds of gist in a healthy young adult population.

Lecture start time and sleep characteristics: Analysis of daily diaries of undergraduate students from the LoST-Sleep project

OBJECTIVES

Emerging evidence shows that later high school start times are associated with increased sleep duration; however, little is known if this extends to the university setting. This study investigated associations of first lecture start times with sleep characteristics among university students.

DESIGN

Daily diaries.

SETTING

Lincoln, UK.

PARTICIPANTS

One hundred and fifty-five undergraduate students completed 7-night sleep diaries MEASUREMENTS: Of the plausible lecture-day diaries (Monday-to-Friday, expected N = 755 days), 567 days were lecture days (M = 3.8 lecture-days per student, SD = 1.1). The Consensus Sleep Diary was used to collect sleep characteristics. Two-level multilevel mixed effect generalized linear models were employed in the analyses.

RESULTS

Seventy-five percent of first lectures occurred before noon. Students reported short sleep (M = 7.0 hours, SD = 1.9) and fewer reported highest levels of sleep quality (42.8%) and restfulness (24.8%) when first lectures started at 09:00 or 09:30 compared to 10:00 or later. Every hour delay of first lecture start time was associated with 15.1 (95% confidence interval [CI]: 9.5; 20.7) minutes increase in sleep duration and higher odds of reporting the highest levels of sleep quality and restfulness. Focusing on attended lectures starting before noon, hourly delay of first lecture start time was associated with 37.4 (95% CI: 22.0; 52.8) minutes increased sleep duration. Bedtime, sleep time, and sleep onset latency were not significantly associated with first lecture start times.

CONCLUSION

This study found that undergraduate students had longer sleep and healthier sleep quality when university first lectures started later. The earliest lecture start time that afforded sufficient sleep duration for students was 10:00.

Resurrected memories: Sleep-dependent memory consolidation saves memories from competition induced by retrieval practice

Retrieval practice improves retention of tested information, and it can either impair or facilitate retention of untested information. Here, we investigated how semantic relatedness, episodic context, and sleep-dependent memory consolidation determine the effects of retrieval practice on retention of untested items. Participants studied lists of scene-word associations. Each scene was associated with two different words (“pairmates”) that were either semantically related or unrelated and either in the same (temporally close) or different lists (temporally far). In three experiments, retrieval practice of scene-word associations facilitated retention of unpracticed, temporally close pairmates and impaired retention of temporally far, semantically unrelated pairmates. Critically, retrieval practice impaired retention of temporally far, semantically related pairmates if participants were unable to sleep during the retention interval, but it facilitated retention of these items if participants were able to sleep. Our findings suggest that sleep extends the benefits of testing to related information learned in temporally separate episodes.

No benefit of auditory closed-loop stimulation on memory for semantically-incongruent associations

Auditory closed-loop stimulation has gained traction in recent years as a means of enhancing slow oscillatory activity and, consequently, sleep-associated memory consolidation. Previous studies on this topic have primarily focused on the consolidation of semantically-congruent associations. In this study, we investigated the effect of auditory closed-loop stimulation on the overnight retention of semantically-incongruent associations. Twelve healthy males (age: M = 20.06, SD = 2.02 years) participated in two experimental conditions (simulation and sham). In the stimulation condition, clicks were delivered in phase with slow oscillation up-states, whereas in the sham condition no auditory stimuli were applied. Corroborating earlier work, stimulation (vs. sham) enhanced the slow oscillation rhythm, phase-coupled spindle activity and slow oscillation power. However, there was no benefit of stimulation on overnight memory retention. These findings suggest that closed-loop stimulation does not benefit semantically-incongruent associations.

How do children with autism spectrum disorder form gist memory during sleep? A study of slow oscillation-spindle coupling

Sleep is assumed to support memory through an active systems consolidation process that does not only strengthen newly encoded representations but also facilitates the formation of more abstract gist memories. Studies in humans and rodents indicate a key role of the precise temporal coupling of sleep slow oscillations (SO) and spindles in this process. The present study aimed at bolstering these findings in typically developing (TD) children, and at dissecting particularities in SO-spindle coupling underlying signs of enhanced gist memory formation during sleep found in a foregoing study in children with autism spectrum disorder (ASD) without intellectual impairment. Sleep data from 19 boys with ASD and 20 TD boys (9-12 years) were analyzed. Children performed a picture-recognition task and the Deese-Roediger-McDermott (DRM) task before nocturnal sleep (encoding) and in the next morning (retrieval). Sleep-dependent benefits for visual-recognition memory were comparable between groups but were greater for gist abstraction (recall of DRM critical lure words) in ASD than TD children. Both groups showed a closely comparable SO-spindle coupling, with fast spindle activity nesting in SO-upstates, suggesting that a key mechanism of memory processing during sleep is fully functioning already at childhood. Picture-recognition at retrieval after sleep was positively correlated to frontocortical SO-fast-spindle coupling in TD children, and less in ASD children. Critical lure recall did not correlate with SO-spindle coupling in TD children but showed a negative correlation (r = -.64, p = .003) with parietal SO-fast-spindle coupling in ASD children, suggesting other mechanisms specifically conveying gist abstraction, that may even compete with SO-spindle coupling.

Temporally delayed linear modelling (TDLM) measures replay in both animals and humans

There are rich structures in off-task neural activity which are hypothesized to reflect fundamental computations across a broad spectrum of cognitive functions. Here, we develop an analysis toolkit - temporal delayed linear modelling (TDLM) - for analysing such activity. TDLM is a domain-general method for finding neural sequences that respect a pre-specified transition graph. It combines nonlinear classification and linear temporal modelling to test for statistical regularities in sequences of task-related reactivations. TDLM is developed on the non-invasive neuroimaging data and is designed to take care of confounds and maximize sequence detection ability. Notably, as a linear framework, TDLM can be easily extended, without loss of generality, to capture rodent replay in electrophysiology, including in continuous spaces, as well as addressing second-order inference questions, for example, its temporal and spatial varying pattern. We hope TDLM will advance a deeper understanding of neural computation and promote a richer convergence between animal and human neuroscience.

Experience replay is associated with efficient nonlocal learning

To make effective decisions, people need to consider the relationship between actions and outcomes. These are often separated by time and space. The neural mechanisms by which disjoint actions and outcomes are linked remain unknown. One promising hypothesis involves neural replay of nonlocal experience. Using a task that segregates direct from indirect value learning, combined with magnetoencephalography, we examined the role of neural replay in human nonlocal learning. After receipt of a reward, we found significant backward replay of nonlocal experience, with a 160-millisecond state-to-state time lag, which was linked to efficient learning of action values. Backward replay and behavioral evidence of nonlocal learning were more pronounced for experiences of greater benefit for future behavior. These findings support nonlocal replay as a neural mechanism for solving complex credit assignment problems during learning.

'Sleep-dependent' memory consolidation? Brief periods of post-training rest and sleep provide an equivalent benefit for both declarative and procedural memory

Sleep following learning facilitates the consolidation of memories. This effect has often been attributed to sleep-specific factors, such as the presence of sleep spindles or slow waves in the electroencephalogram (EEG). However, recent studies suggest that simply resting quietly while awake could confer a similar memory benefit. In the current study, we examined the effects of sleep, quiet rest, and active wakefulness on the consolidation of declarative and procedural memory. We hypothesized that sleep and eyes-closed quiet rest would both benefit memory compared with a period of active wakefulness. After completing a declarative and a procedural memory task, participants began a 30-min retention period with PSG (polysomnographic) monitoring, in which they either slept (n = 24), quietly rested with their eyes closed (n = 22), or completed a distractor task (n = 29). Following the retention period, participants were again tested on their memory for the two learning tasks. As hypothesized, sleep and quiet rest both led to better performance on the declarative and procedural memory tasks than did the distractor task. Moreover, the performance advantages conferred by rest were indistinguishable from those of sleep. These data suggest that neurobiology specific to sleep might not be necessary to induce the consolidation of memory, at least across very short retention intervals. Instead, offline memory consolidation may function opportunistically, occurring during either sleep or stimulus-free rest, provided a favorable neurobiological milieu and sufficient reduction of new encoding.

The overfitted brain: Dreams evolved to assist generalization

Understanding of the evolved biological function of sleep has advanced considerably in the past decade. However, no equivalent understanding of dreams has emerged. Contemporary neuroscientific theories often view dreams as epiphenomena, and many of the proposals for their biological function are contradicted by the phenomenology of dreams themselves. Now, the recent advent of deep neural networks (DNNs) has finally provided the novel conceptual framework within which to understand the evolved function of dreams. Notably, all DNNs face the issue of overfitting as they learn, which is when performance on one dataset increases but the network's performance fails to generalize (often measured by the divergence of performance on training versus testing datasets). This ubiquitous problem in DNNs is often solved by modelers via "noise injections" in the form of noisy or corrupted inputs. The goal of this paper is to argue that the brain faces a similar challenge of overfitting and that nightly dreams evolved to combat the brain's overfitting during its daily learning. That is, dreams are a biological mechanism for increasing generalizability via the creation of corrupted sensory inputs from stochastic activity across the hierarchy of neural structures. Sleep loss, specifically dream loss, leads to an overfitted brain that can still memorize and learn but fails to generalize appropriately. Herein this "overfitted brain hypothesis" is explicitly developed and then compared and contrasted with existing contemporary neuroscientific theories of dreams. Existing evidence for the hypothesis is surveyed within both neuroscience and deep learning, and a set of testable predictions is put forward that can be pursued both in vivo and in silico.

Sleep after learning aids the consolidation of factual knowledge, but not relearning

Study Objectives

Sleep strengthens and reorganizes declarative memories, but the extent to which these processes benefit subsequent relearning of the same material remains unknown. It is also unclear whether sleep-memory effects translate to educationally realistic learning tasks and improve long-term learning outcomes.

Methods

Young adults learned factual knowledge in two learning sessions that were 12 h apart and separated by either nocturnal sleep (n = 26) or daytime wakefulness (n = 26). Memory before and after the retention interval was compared to assess the effect of sleep on consolidation, while memory before and after the second learning session was compared to assess relearning. A final test 1 week later assessed whether there was any long-term advantage to sleeping between two study sessions.

Results

Sleep significantly enhanced consolidation of factual knowledge (p = 0.01, d = 0.72), but groups did not differ in their capacity to relearn the materials (p = 0.72, d = 0.10). After 1 week, a numerical memory advantage remained for the sleep group but was no longer significant (p = 0.21, d = 0.35).

Conclusions

Reduced forgetting after sleep is a robust finding that extends to our ecologically valid learning task, but we found no evidence that sleep enhances relearning. Our findings can exclude a large effect of sleep on long-term memory after 1 week, but hint at a smaller effect, leaving open the possibility of practical benefits from organizing study sessions around nocturnal sleep. These findings highlight the importance of revisiting key sleep-memory effects to assess their relevance to long-term learning outcomes with naturalistic learning materials.

Teaching Dream Interpretation to Psychiatric Residents

Though the clinical and experimental literature suggests that discussing dreams with therapy patients can be beneficial in a variety of important ways, there appears to be a gap in educational opportunities for psychiatric residents regarding this process. To address this educational need at the authors' residency program, a class in psychodynamically oriented dream analysis was implemented, and data was collected in the form of learner surveys both before and after they took the course. The survey found that the level of importance placed on dream work, the comfort level in discussing dreams with patients, and the frequency with which dreams were discussed in sessions were all increased after taking the course. Our conclusion was that these preliminary results suggest that implementing a structured course on dream analysis may help to fill the educational gap.

Sleep strengthens integration of spatial memory systems

Spatial memory comprises different representational systems that are sensitive to different environmental cues, like proximal landmarks or local boundaries. Here we examined how sleep affects the formation of a spatial representation integrating landmark-referenced and boundary-referenced representations. To this end, participants (n = 42) were familiarized with an environment featuring both a proximal landmark and a local boundary. After nocturnal periods of sleep or wakefulness and another night of sleep, integration of the two representational systems was tested by testing the participant's flexibility to switch from landmark-based to boundary-based navigation in the environment, and vice versa. Results indicate a distinctly increased flexibility in relying on either landmarks or boundaries for navigation, when familiarization to the environment was followed by sleep rather than by wakefulness. A second control study (n = 45) did not reveal effects of sleep (vs. wakefulness) on navigation in environments featuring only landmarks or only boundaries. Thus, rather than strengthening isolated representational systems per se, sleep presumably through forming an integrative representation, enhances flexible coordination of representational subsystems.

The different roles of sleep on false memory formation between young and older adults

Aging is accompanied by an increase in the probability of false memory. However, what role sleep plays in the age effect in false memory is less understood. Our study utilized a simplified conjoint recognition (SCR)-based Deese-Roediger-McDermott (DRM) paradigm to investigate the role of sleep on false memory in young and older adults. The results showed that sleep effect in false memory was modulated by age, manifested as sleep increased young adults' falsely recognized critical lures, while it reduced older adults'. In addition, in a more fine-grained analysis, the results of multinomial processing tree (MPT) modeling further revealed that young adults were more likely to retrieve memory based on gist traces than older adults, and young adults were more susceptible to guess a probe as "old" than older adults in the sleep condition. Combined findings from the number and ratio of falsely recognized critical lures and the MPT modeling, the current study suggested that sleep might increase young adults' false memory via gist extraction, while it decreased older adults' false memory via verbatim trace consolidation. The study contributes to a comprehensive view on the age-by-sleep effect on false recognition, with the segregation of cognitive components of verbatim memory, gist memory, and response bias.

Sleep reduces the semantic coherence of memory recall: An application of latent semantic analysis to investigate memory reconstruction

Sleep is thought to help consolidate hippocampus-dependent memories by reactivating previously encoded neural representations, promoting both quantitative and qualitative changes in memory representations. However, the qualitative nature of changes to memory representations induced by sleep remains largely uncharacterized. In this study, we investigated how memories are reconstructed by hypothesizing that semantic coherence, defined as conceptual relatedness between statements of free-recall texts and quantified using latent semantic analysis (LSA), is affected by post-encoding sleep. Short naturalistic videos of events featuring six animals were presented to 115 participants who were randomly assigned to either 12- or 24-h delay groups featuring sleep or wakefulness. Participants' free-recall responses were analyzed to test for an effect of sleep on semantic coherence between adjacent statements, and overall. The presence of sleep reduced both forms of semantic coherence, compared to wakefulness. This change was robust and not due to shifts in conciseness or repetitiveness with sleep. These findings support the notion that sleep-dependent consolidation qualitatively changes the features of reconstructed memory representations by reducing semantic coherence.

Sleep Facilitates Problem Solving With No Additional Gain Through Targeted Memory Reactivation

According to the active systems consolidation theory, memories undergo reactivation during sleep that can give rise to qualitative changes of the representations. These changes may generate new knowledge such as gaining insight into solutions for problem solving. targeted memory reactivation (TMR) uses learning-associated cues, such as sounds or odors, which have been shown to improve memory consolidation when re-applied during sleep. Here we tested whether TMR during slow wave sleep (SWS) and/or rapid eye movement (REM) sleep increases problem solving. Young healthy volunteers participated in one of two experiments. Experiment 1 tested the effect of natural sleep on problem solving. Subjects were trained in a video game-based problem solving task until being presented with a non-solved challenge. Followed by a ~10-h incubation interval filled with nocturnal sleep or daytime wakefulness, subjects were tested on the problem solving challenge again. Experiment 2 tested the effect of TMR on problem solving, with subjects receiving auditory TMR either during SWS (SWSstim), REM sleep (REMstim), or wakefulness (Wakestim). In Experiment 1, sleep improved problem solving, with 62% of subjects from the Sleep group solving the problem compared to 24% of the Wake group. Subjects with higher amounts of SWS in the Sleep group had a higher chance to solve the problem. In Experiment 2, TMR did not change the sleep effect on problem solving: 56 and 58% of subjects from the SWSstim and REMstim groups solved the problem compared to 57% from the Wakestim group. These findings indicate that sleep, and particularly SWS, facilitates problem solving, whereas this effect is not further increased by TMR.

Sleep’s Role in Schema Learning and Creative Insights

Purpose of Review

A recent resurgence of interest in schema theory has influenced research on sleep-dependent memory consolidation and led to a new understanding of how schemata might be activated during sleep and play a role in the reorganisation of memories. This review is aimed at synthesising recent findings into a coherent narrative and draw overall conclusions.

Recent Findings

Rapid consolidation of schematic memories has been shown to benefit from an interval containing sleep. These memories have shown reduced reliance on the hippocampus following consolidation in both humans and rodents. Using a variety of methodologies, notably including the DRM paradigm, it has been shown that activation of a schema can increase the rate of false memory as a result of activation of semantic associates during slow wave sleep (SWS). Memories making use of a schema have shown increased activity in the medial prefrontal cortex, which may reflect both the schematic activation itself and a cognitive control component selecting an appropriate schema to use. SWS seems to be involved in assimilation of new memories within existing semantic frameworks and in making memories more explicit, while REM sleep may be more associated with creating entirely novel associations while keeping memories implicit.

Summary

Sleep plays an important role in schematic memory consolidation, with more rapid consolidation, reduced hippocampal involvement, and increased prefrontal involvement as the key characteristics. Both SWS and REM sleep may have a role to play.

Reactivation during sleep with incomplete reminder cues rather than complete ones stabilizes long-term memory in humans

Reactivation by reminder cues labilizes memories during wakefulness, requiring reconsolidation to persist. In contrast, during sleep, cued reactivation seems to directly stabilize memories. In reconsolidation, incomplete reminders are more effective in reactivating memories than complete reminders by inducing a mismatch, i.e. a discrepancy between expected and actual events. Whether mismatch is likewise detected during sleep is unclear. Here we test whether cued reactivation during sleep is more effective for mismatch-inducing incomplete than complete reminders. We first establish that only incomplete but not complete reminders labilize memories during wakefulness. When complete or incomplete reminders are presented during 40-min sleep, both reminders are equally effective in stabilizing memories. However, when extending the retention interval for another 7 hours (following 40-min sleep), only incomplete but not complete reminders stabilize memories, regardless of the extension containing wakefulness or sleep. We propose that, during sleep, only incomplete reminders initiate long-term memory stabilization via mismatch detection.

Forcato et al. show that incomplete reminder cues rather than complete ones stabilize human memories during sleep. This study suggests that only incomplete reminders initiate long-term memory stabilization via mismatch detection during sleep.

The Tolman-Eichenbaum Machine: Unifying Space and Relational Memory through Generalization in the Hippocampal Formation

The hippocampal-entorhinal system is important for spatial and relational memory tasks. We formally link these domains, provide a mechanistic understanding of the hippocampal role in generalization, and offer unifying principles underlying many entorhinal and hippocampal cell types. We propose medial entorhinal cells form a basis describing structural knowledge, and hippocampal cells link this basis with sensory representations. Adopting these principles, we introduce the Tolman-Eichenbaum machine (TEM). After learning, TEM entorhinal cells display diverse properties resembling apparently bespoke spatial responses, such as grid, band, border, and object-vector cells. TEM hippocampal cells include place and landmark cells that remap between environments. Crucially, TEM also aligns with empirically recorded representations in complex non-spatial tasks. TEM also generates predictions that hippocampal remapping is not random as previously believed; rather, structural knowledge is preserved across environments. We confirm this structural transfer over remapping in simultaneously recorded place and grid cells.

A daytime nap combined with nighttime sleep promotes learning in toddlers

Napping after learning promotes consolidation of new information during infancy. Yet, whether naps play a similar role during toddlerhood, a stage when many children are beginning to transition away from napping, is less clear. In Experiment 1, we examined whether napping after learning promotes generalization of novel category exemplars 24 h later. Young children (N = 54, age range = 29-36 months) viewed three category exemplars in different contexts from each of three categories and remained awake (No-Nap condition) or napped (Nap condition) after encoding and were then tested 24 h later. Children who napped after learning showed superior generalization 24 h later relative to children who did not nap. In a Nap-Control condition tested 4 h after awakening from a nap, children performed at the same low level as in the No-Nap condition, indicating that generalization stemmed from an additional period of nighttime sleep and not simply from a nap or increased time. In Experiment 2, we examined whether nighttime sleep is sufficient for generalization if it occurs soon after learning. An additional group of children (N = 18) learned before bedtime and were tested 4 h after waking up the next day. Children did not generalize as well as those who had a nap combined with subsequent nighttime sleep. These findings suggest that naps, when combined with a period of nighttime sleep, might help toddlers to retain newly learned information and lead to delayed benefits in generalization.

Neuronal Computation Underlying Inferential Reasoning in Humans and Mice

Every day we make decisions critical for adaptation and survival. We repeat actions with known consequences. But we also draw on loosely related events to infer and imagine the outcome of entirely novel choices. These inferential decisions are thought to engage a number of brain regions; however, the underlying neuronal computation remains unknown. Here, we use a multi-day cross-species approach in humans and mice to report the functional anatomy and neuronal computation underlying inferential decisions. We show that during successful inference, the mammalian brain uses a hippocampal prospective code to forecast temporally structured learned associations. Moreover, during resting behavior, coactivation of hippocampal cells in sharp-wave/ripples represent inferred relationships that include reward, thereby "joining-the-dots" between events that have not been observed together but lead to profitable outcomes. Computing mnemonic links in this manner may provide an important mechanism to build a cognitive map that stretches beyond direct experience, thus supporting flexible behavior.

Overnight sleep benefits both neutral and negative direct associative and relational memory

Strong evidence suggests that sleep plays a role in memory consolidation, which involves both stabilizing memory into long-term storage as well as integrating new information into existing stores. The current study investigated consolidation, across a day of wakefulness or night of sleep, of emotional and neutral directly learned visual paired associates (A-B/B-C pairs) as well as formation of memory for relational pairs formed via overlapping learned components (A-C pairs). Participants learned 40 negative and 40 neutral face-object pairs followed by a baseline test in session 1 either in the morning or evening. They then spent a 12-hour retention period during which participants either went about their normal day or spent the night in the sleep lab. During session 2, participants completed a surprise test to assess their memory for relational pairs (A-C) as well as memory for direct associates (A-B/B-C). As hypothesized, the results demonstrated that a 12-hour retention period predominantly spent asleep, compared to awake, benefited memory for both relational and direct associative memory. However, contrary to the hypothesis that emotional salience would promote preferential consolidation, sleep appeared to benefit both negative and neutral information similarly for direct associative and relational memories, suggesting that sleep may interact with other factors affecting encoding (e.g., depth of encoding) to benefit direct and relational associative memory. As one of the few studies examining the role of nocturnal sleep and emotion on both direct and relational associative memory, our findings suggest key insights into how overnight sleep consolidates these different forms of memory.