Geometric transformation of cognitive maps for generalization across hippocampal-prefrontal circuits

Cell-type-specific manifold analysis discloses independent geometric transformations in the hippocampal spatial code

Integrating analyses of genetically defined cell types with population-level approaches remains poorly explored. We investigated this question by focusing on hippocampal spatial maps and the contribution of two genetically defined pyramidal cell types in the deep and superficial CA1 sublayers. Using single- and dual-color miniscope imaging in mice running along a linear track, we found that population activity from these cells exhibited three-dimensional ring manifolds that encoded the animal position and running direction. Despite shared topology, sublayer-specific manifolds displayed distinct geometric features. Manipulating track orientation revealed rotational and translational changes in manifolds from deep cells, contrasting with more stable representations by superficial cells. These transformations were not observed in manifolds derived from the entire CA1 population. Instead, cell-type-specific chemogenetic silencing of either sublayer revealed independent geometric codes. Our results show how genetically specified subpopulations may underpin parallel spatial maps that can be manipulated independently.

Three systems of circuit formation: assembly, updating and tuning

Understanding the relationship between genotype and neuronal circuit phenotype necessitates an integrated view of genetics, development, plasticity and learning. Challenging the prevailing notion that emphasizes learning and plasticity as primary drivers of circuit assembly, in this Perspective, we delineate a tripartite framework to clarify the respective roles that learning and plasticity might have in this process. In the first part of the framework, which we term System One, neural circuits are established purely through genetically driven algorithms, in which spike timing-dependent plasticity serves no instructive role. We propose that these circuits equip the animal with sufficient skill and knowledge to successfully engage the world. Next, System Two is governed by rare but critical 'single-shot learning' events, which occur in response to survival situations and prompt rapid synaptic reconfiguration. Such events serve as crucial updates to the existing hardwired knowledge base of an organism. Finally, System Three is characterized by a perpetual state of synaptic recalibration, involving continual plasticity for circuit stabilization and fine-tuning. By outlining the definitions and roles of these three core systems, our framework aims to resolve existing ambiguities related to and enrich our understanding of neural circuit formation.

Learning produces an orthogonalized state machine in the hippocampus

Cognitive maps confer animals with flexible intelligence by representing spatial, temporal and abstract relationships that can be used to shape thought, planning and behaviour. Cognitive maps have been observed in the hippocampus1, but their algorithmic form and learning mechanisms remain obscure. Here we used large-scale, longitudinal two-photon calcium imaging to record activity from thousands of neurons in the CA1 region of the hippocampus while mice learned to efficiently collect rewards from two subtly different linear tracks in virtual reality. Throughout learning, both animal behaviour and hippocampal neural activity progressed through multiple stages, gradually revealing improved task representation that mirrored improved behavioural efficiency. The learning process involved progressive decorrelations in initially similar hippocampal neural activity within and across tracks, ultimately resulting in orthogonalized representations resembling a state machine capturing the inherent structure of the task. This decorrelation process was driven by individual neurons acquiring task-state-specific responses (that is, 'state cells'). Although various standard artificial neural networks did not naturally capture these dynamics, the clone-structured causal graph, a hidden Markov model variant, uniquely reproduced both the final orthogonalized states and the learning trajectory seen in animals. The observed cellular and population dynamics constrain the mechanisms underlying cognitive map formation in the hippocampus, pointing to hidden state inference as a fundamental computational principle, with implications for both biological and artificial intelligence.

Conjoint generalized and trajectory-specific coding of task structure by prefrontal neurons

Neurons in the medial prefrontal cortex (mPFC) are spatially tuned. Trajectory-specific firing with distinct spatial tuning on different paths to reward sites as well as generalized spatial tuning with similar responses on separate trajectories have been described. However, it is unclear whether such distinct populations contribute differently to the encoding of task space. Here, we find coexisting populations of neurons with trajectory-specific and generalized tuning profiles in an olfaction-guided spatial memory task in mice. Neurons with generalized representation show stable spatial tuning within and across days, allow accurate predictions of the animal's position, and preferentially emerge upon task learning. In contrast, cells with trajectory-specific spatial tuning display dynamically changing tuning functions, are less informative about the current position, and can be identified at a larger proportion early in task learning. These results highlight a role for neurons with generalized tuning in the efficient and stable representation of task space.

Exploring replay

Animals face uncertainty about their environments due to initial ignorance or subsequent changes. They therefore need to explore. However, the algorithmic structure of exploratory choices in the brain still remains largely elusive. Artificial agents face the same problem, and a venerable idea in reinforcement learning is that they can plan appropriate exploratory choices offline, during the equivalent of quiet wakefulness or sleep. Although offline processing in humans and other animals, in the form of hippocampal replay and preplay, has recently been the subject of highly informative modelling, existing methods only apply to known environments. Thus, they cannot predict exploratory replay choices during learning and/or behaviour in the face of uncertainty. Here, we extend an influential theory of hippocampal replay and examine its potential role in approximately optimal exploration, deriving testable predictions for the patterns of exploratory replay choices in a paradigmatic spatial navigation task. Our modelling provides a normative interpretation of the available experimental data suggestive of exploratory replay. Furthermore, we highlight the importance of sequence replay, and license a range of new experimental paradigms that should further our understanding of offline processing.

Prefrontal cortex synchronization with the hippocampus and parietal cortex is strategy-dependent during spatial learning

During spatial learning, subjects progressively adjust their navigation strategies as they acquire experience. The medial prefrontal cortex (mPFC) supports this operation, for which it may integrate information from distributed networks, such as the hippocampus (HPC) and the posterior parietal cortex (PPC). However, the mechanism underlying the prefrontal coordination with HPC and PPC during spatial learning is poorly understood. Here we show that during navigation trials, mice displayed two sequential behavioral stages: searching and exploration. Exclusively during searching, mice gradually increased their efficiency by transitioning from non-spatial to spatial strategies. When mice used spatial strategies specifically in searching stage, hippocampal and parietal oscillations synchronized gamma oscillations (60-100 Hz) and neuronal firing in the mPFC. This coincided with an increase in the incidence of gamma and task-stage-related changes in firing patterns in the mPFC. These findings relate the goal-directed organization of behavior during spatial learning to transient task-related prefrontal large-scale synchronization.

Sleep microstructure organizes memory replay

Recently acquired memories are reactivated in the hippocampus during sleep, an initial step for their consolidation1-3. This process is concomitant with the hippocampal reactivation of previous memories4-6, posing the problem of how to prevent interference between older and recent, initially labile, memory traces. Theoretical work has suggested that consolidating multiple memories while minimizing interference can be achieved by randomly interleaving their reactivation7-10. An alternative is that a temporal microstructure of sleep can promote the reactivation of different types of memories during specific substates. Here, to test these two hypotheses, we developed a method to simultaneously record large hippocampal ensembles and monitor sleep dynamics through pupillometry in naturally sleeping mice. Oscillatory pupil fluctuations revealed a previously unknown microstructure of non-REM sleep-associated memory processes. We found that memory replay of recent experiences dominated in sharp-wave ripples during contracted pupil substates of non-REM sleep, whereas replay of previous memories preferentially occurred during dilated pupil substates. Selective closed-loop disruption of sharp-wave ripples during contracted pupil non-REM sleep impaired the recall of recent memories, whereas the same manipulation during dilated pupil substates had no behavioural effect. Stronger extrinsic excitatory inputs characterized the contracted pupil substate, whereas higher recruitment of local inhibition was prominent during dilated pupil substates. Thus, the microstructure of non-REM sleep organizes memory replay, with previous versus new memories being temporally segregated in different substates and supported by local and input-driven mechanisms, respectively. Our results suggest that the brain can multiplex distinct cognitive processes during sleep to facilitate continuous learning without interference.

A cellular basis for mapping behavioural structure

To flexibly adapt to new situations, our brains must understand the regularities in the world, as well as those in our own patterns of behaviour. A wealth of findings is beginning to reveal the algorithms that we use to map the outside world1-6. However, the biological algorithms that map the complex structured behaviours that we compose to reach our goals remain unknown. Here we reveal a neuronal implementation of an algorithm for mapping abstract behavioural structure and transferring it to new scenarios. We trained mice on many tasks that shared a common structure (organizing a sequence of goals) but differed in the specific goal locations. The mice discovered the underlying task structure, enabling zero-shot inferences on the first trial of new tasks. The activity of most neurons in the medial frontal cortex tiled progress to goal, akin to how place cells map physical space. These 'goal-progress cells' generalized, stretching and compressing their tiling to accommodate different goal distances. By contrast, progress along the overall sequence of goals was not encoded explicitly. Instead, a subset of goal-progress cells was further tuned such that individual neurons fired with a fixed task lag from a particular behavioural step. Together, these cells acted as task-structured memory buffers, implementing an algorithm that instantaneously encoded the entire sequence of future behavioural steps, and whose dynamics automatically computed the appropriate action at each step. These dynamics mirrored the abstract task structure both on-task and during offline sleep. Our findings suggest that schemata of complex behavioural structures can be generated by sculpting progress-to-goal tuning into task-structured buffers of individual behavioural steps.

Maintenance and transformation of representational formats during working memory prioritization

Visual working memory depends on both material-specific brain areas in the ventral visual stream (VVS) that support the maintenance of stimulus representations and on regions in the prefrontal cortex (PFC) that control these representations. How executive control prioritizes working memory contents and whether this affects their representational formats remains an open question, however. Here, we analyzed intracranial EEG (iEEG) recordings in epilepsy patients with electrodes in VVS and PFC who performed a multi-item working memory task involving a retro-cue. We employed Representational Similarity Analysis (RSA) with various Deep Neural Network (DNN) architectures to investigate the representational format of prioritized VWM content. While recurrent DNN representations matched PFC representations in the beta band (15-29 Hz) following the retro-cue, they corresponded to VVS representations in a lower frequency range (3-14 Hz) towards the end of the maintenance period. Our findings highlight the distinct coding schemes and representational formats of prioritized content in VVS and PFC.

Hippocampal neuronal activity is aligned with action plans

Neurons in the hippocampus are correlated with different variables, including space, time, sensory cues, rewards, and actions, where the extent of tuning depends on ongoing task demands. However, it remains uncertain whether such diverse tuning corresponds to distinct functions within the hippocampal network or if a more generic computation can account for these observations. To disentangle the contribution of externally driven cues versus internal computation, we developed a task in mice where space, auditory tones, rewards, and context were juxtaposed with changing relevance. High-density electrophysiological recordings revealed that neurons were tuned to each of these modalities. By comparing movement paths and action sequences, we observed that external variables had limited direct influence on hippocampal firing. Instead, spiking was influenced by online action plans modulated by goal uncertainty. Our results suggest that internally generated cell assembly sequences are selected and updated by action plans toward deliberate goals. The apparent tuning of hippocampal neuronal spiking to different sensory modalities might emerge due to alignment to the afforded action progression within a task rather than representation of external cues.

Space is a latent sequence: A theory of the hippocampus

Fascinating phenomena such as landmark vector cells and splitter cells are frequently discovered in the hippocampus. Without a unifying principle, each experiment seemingly uncovers new anomalies or coding types. Here, we provide a unifying principle that the mental representation of space is an emergent property of latent higher-order sequence learning. Treating space as a sequence resolves numerous phenomena and suggests that the place field mapping methodology that interprets sequential neuronal responses in Euclidean terms might itself be a source of anomalies. Our model, clone-structured causal graph (CSCG), employs higher-order graph scaffolding to learn latent representations by mapping aliased egocentric sensory inputs to unique contexts. Learning to compress sequential and episodic experiences using CSCGs yields allocentric cognitive maps that are suitable for planning, introspection, consolidation, and abstraction. By explicating the role of Euclidean place field mapping and demonstrating how latent sequential representations unify myriad observed phenomena, our work positions the hippocampus in a sequence-centric paradigm, challenging the prevailing space-centric view.

Abstract representations emerge in human hippocampal neurons during inference

Humans have the remarkable cognitive capacity to rapidly adapt to changing environments. Central to this capacity is the ability to form high-level, abstract representations that take advantage of regularities in the world to support generalization1. However, little is known about how these representations are encoded in populations of neurons, how they emerge through learning and how they relate to behaviour2,3. Here we characterized the representational geometry of populations of neurons (single units) recorded in the hippocampus, amygdala, medial frontal cortex and ventral temporal cortex of neurosurgical patients performing an inferential reasoning task. We found that only the neural representations formed in the hippocampus simultaneously encode several task variables in an abstract, or disentangled, format. This representational geometry is uniquely observed after patients learn to perform inference, and consists of disentangled directly observable and discovered latent task variables. Learning to perform inference by trial and error or through verbal instructions led to the formation of hippocampal representations with similar geometric properties. The observed relation between representational format and inference behaviour suggests that abstract and disentangled representational geometries are important for complex cognition.

Hippocampal-prefrontal communication subspaces align with behavioral and network patterns in a spatial memory task

Rhythmic network states have been theorized to facilitate communication between brain regions, but how these oscillations influence communication subspaces, i.e, the low-dimensional neural activity patterns that mediate inter-regional communication, and in turn how subspaces impact behavior remains unclear. Using a spatial memory task in rats, we simultaneously recorded ensembles from hippocampal CA1 and the prefrontal cortex (PFC) to address this question. We found that task behaviors best aligned with low-dimensional, shared subspaces between these regions, rather than local activity in either region. Critically, both network oscillations and speed modulated the structure and performance of this communication subspace. Contrary to expectations, theta coherence did not better predict CA1-PFC shared activity, while theta power played a more significant role. To understand the communication space, we visualized shared CA1-PFC communication geometry using manifold techniques and found ring-like structures. We hypothesize that these shared activity manifolds are utilized to mediate the task behavior. These findings suggest that memory-guided behaviors are driven by shared CA1-PFC interactions that are dynamically modulated by oscillatory states, offering a novel perspective on the interplay between rhythms and behaviorally relevant neural communication.

Prefrontal cortical ripples mediate top-down suppression of hippocampal reactivation during sleep memory consolidation

Consolidation of initially encoded hippocampal representations in the neocortex through reactivation is crucial for long-term memory formation and is facilitated by the coordination of hippocampal sharp-wave ripples (SWRs) with cortical slow and spindle oscillations during non-REM sleep. Recent evidence suggests that high-frequency cortical ripples can also coordinate with hippocampal SWRs in support of consolidation; however, the contribution of cortical ripples to reactivation remains unclear. We used high-density, continuous recordings in the hippocampus (area CA1) and prefrontal cortex (PFC) over the course of spatial learning and show that independent PFC ripples dissociated from SWRs are prevalent in NREM sleep and predominantly suppress hippocampal activity. PFC ripples paradoxically mediate top-down suppression of hippocampal reactivation rather than coordination, and this suppression is stronger for assemblies that are reactivated during coordinated CA1-PFC ripples for consolidation of recent experiences. Further, we show non-canonical, serial coordination of independent cortical ripples with slow and spindle oscillations, which are known signatures of memory consolidation. These results establish a role for prefrontal cortical ripples in top-down regulation of behaviorally relevant hippocampal representations during consolidation.

Dynamic prediction of goal location by coordinated representation of prefrontal-hippocampal theta sequences

Prefrontal (PFC) and hippocampal (HPC) sequences of neuronal firing modulated by theta rhythms could represent upcoming choices during spatial memory-guided decision-making. How the PFC-HPC network dynamically coordinates theta sequences to predict specific goal locations and how it is interrupted in memory impairments induced by amyloid beta (Aβ) remain unclear. Here, we detected theta sequences of firing activities of PFC neurons and HPC place cells during goal-directed spatial memory tasks. We found that PFC ensembles exhibited predictive representation of the specific goal location since the starting phase of memory retrieval, earlier than the hippocampus. High predictive accuracy of PFC theta sequences existed during successful memory retrieval and positively correlated with memory performance. Coordinated PFC-HPC sequences showed PFC-dominant prediction of goal locations during successful memory retrieval. Furthermore, we found that theta sequences of both regions still existed under Aβ accumulation, whereas their predictive representation of goal locations was weakened with disrupted spatial representation of HPC place cells and PFC neurons. These findings highlight the essential role of coordinated PFC-HPC sequences in successful memory retrieval of a precise goal location.

Selection of experience for memory by hippocampal sharp wave ripples

Experiences need to be tagged during learning for further consolidation. However, neurophysiological mechanisms that select experiences for lasting memory are not known. By combining large-scale neural recordings in mice with dimensionality reduction techniques, we observed that successive maze traversals were tracked by continuously drifting populations of neurons, providing neuronal signatures of both places visited and events encountered. When the brain state changed during reward consumption, sharp wave ripples (SPW-Rs) occurred on some trials, and their specific spike content decoded the trial blocks that surrounded them. During postexperience sleep, SPW-Rs continued to replay those trial blocks that were reactivated most frequently during waking SPW-Rs. Replay content of awake SPW-Rs may thus provide a neurophysiological tagging mechanism to select aspects of experience that are preserved and consolidated for future use.

A persistent prefrontal reference frame across time and task rules

Behavior can be remarkably consistent, even over extended time periods, yet whether this is reflected in stable or 'drifting' neuronal responses to task features remains controversial. Here, we find a persistently active ensemble of neurons in the medial prefrontal cortex (mPFC) of mice that reliably maintains trajectory-specific tuning over several weeks while performing an olfaction-guided spatial memory task. This task-specific reference frame is stabilized during learning, upon which repeatedly active neurons show little representational drift and maintain their trajectory-specific tuning across long pauses in task exposure and across repeated changes in cue-target location pairings. These data thus suggest a 'core ensemble' of prefrontal neurons forming a reference frame of task-relevant space for the performance of consistent behavior over extended periods of time.

Hippocampal sequences span experience relative to rewards

Hippocampal place cells fire in sequences that span spatial environments and non-spatial modalities, suggesting that hippocampal activity can anchor to the most behaviorally salient aspects of experience. As reward is a highly salient event, we hypothesized that sequences of hippocampal activity can anchor to rewards. To test this, we performed two-photon imaging of hippocampal CA1 neurons as mice navigated virtual environments with changing hidden reward locations. When the reward moved, the firing fields of a subpopulation of cells moved to the same relative position with respect to reward, constructing a sequence of reward-relative cells that spanned the entire task structure. The density of these reward-relative sequences increased with task experience as additional neurons were recruited to the reward-relative population. Conversely, a largely separate subpopulation maintained a spatially-based place code. These findings thus reveal separate hippocampal ensembles can flexibly encode multiple behaviorally salient reference frames, reflecting the structure of the experience.

Distinct manifold encoding of navigational information in the subiculum and hippocampus

The subiculum (SUB) plays a crucial role in spatial navigation and encodes navigational information differently from the hippocampal CA1 area. However, the representation of subicular population activity remains unknown. Here, we investigated the neuronal population activity recorded extracellularly from the CA1 and SUB of rats performing T-maze and open-field tasks. The trajectory of population activity in both areas was confined to low-dimensional neural manifolds homoeomorphic to external space. The manifolds conveyed position, speed, and future path information with higher decoding accuracy in the SUB than in the CA1. The manifolds exhibited common geometry across rats and regions for the CA1 and SUB and between tasks in the SUB. During post-task ripples in slow-wave sleep, population activity represented reward locations/events more frequently in the SUB than in CA1. Thus, the CA1 and SUB encode information distinctly into the neural manifolds that underlie navigational information processing during wakefulness and sleep.

Regional specialization manifests in the reliability of neural population codes

The brain has the remarkable ability to learn and guide the performance of complex tasks. Decades of lesion studies suggest that different brain regions perform specialized functions in support of complex behaviors1-3. Yet recent large-scale studies of neural activity reveal similar patterns of activity and encoding distributed widely throughout the brain4-6. How these distributed patterns of activity and encoding are compatible with regional specialization of brain function remains unclear. Two frontal brain regions, the dorsal medial prefrontal cortex (dmPFC) and orbitofrontal cortex (OFC), are a paradigm of this conundrum. In the setting complex behaviors, the dmPFC is necessary for choosing optimal actions2,7,8, whereas the OFC is necessary for waiting for3,9 and learning from2,7,9-12 the outcomes of those actions. Yet both dmPFC and OFC encode both choice- and outcome-related quantities13-20. Here we show that while ensembles of neurons in the dmPFC and OFC of rats encode similar elements of a cognitive task with similar patterns of activity, the two regions differ in when that coding is consistent across trials ("reliable"). In line with the known critical functions of each region, dmPFC activity is more reliable when animals are making choices and less reliable preceding outcomes, whereas OFC activity shows the opposite pattern. Our findings identify the dynamic reliability of neural population codes as a mechanism whereby different brain regions may support distinct cognitive functions despite exhibiting similar patterns of activity and encoding similar quantities.

Neuronal tuning to threat exposure remains stable in the mouse prefrontal cortex over multiple days

Intense threat elicits action in the form of active and passive coping. The medial prefrontal cortex (mPFC) executes top-level control over the selection of threat coping strategies, but the dynamics of mPFC activity upon continuing threat encounters remain unexplored. Here, we used 1-photon calcium imaging in mice to probe the activity of prefrontal pyramidal cells during repeated exposure to intense threat in a tail suspension (TS) paradigm. A subset of prefrontal neurons displayed selective activation during TS, which was stably maintained over days. During threat, neurons showed specific tuning to active or passive coping. These responses were unrelated to general motion tuning and persisted over days. Moreover, the neural manifold traversed by low-dimensional population activity remained stable over subsequent days of TS exposure and was preserved across individuals. These data thus reveal a specific, temporally, and interindividually conserved repertoire of prefrontal tuning to behavioral responses under threat.

From cell types to population dynamics: Making hippocampal manifolds physiologically interpretable

The study of the hippocampal code is gaining momentum. While the physiological approach targets the contribution of individual cells as determined by genetic, biophysical and circuit factors, the field pushes for a population dynamic approach that considers the representation of behavioural variables by a large number of neurons. In this alternative framework, neuronal activity is projected into low-dimensional manifolds. These manifolds can reveal the structure of population representations, but their physiological interpretation is challenging. Here, we review the recent literature and propose that integrating information regarding behavioral traits, local field potential oscillations and cell-type-specificity into neural manifolds offers strategies to make them interpretable at the physiological level.

Abstract representations emerge in human hippocampal neurons during inference behavior

Humans have the remarkable cognitive capacity to rapidly adapt to changing environments. Central to this capacity is the ability to form high-level, abstract representations that take advantage of regularities in the world to support generalization 1 . However, little is known about how these representations are encoded in populations of neurons, how they emerge through learning, and how they relate to behavior 2,3 . Here we characterized the representational geometry of populations of neurons (single-units) recorded in the hippocampus, amygdala, medial frontal cortex, and ventral temporal cortex of neurosurgical patients who are performing an inferential reasoning task. We find that only the neural representations formed in the hippocampus simultaneously encode multiple task variables in an abstract, or disentangled, format. This representational geometry is uniquely observed after patients learn to perform inference, and consisted of disentangled directly observable and discovered latent task variables. Interestingly, learning to perform inference by trial and error or through verbal instructions led to the formation of hippocampal representations with similar geometric properties. The observed relation between representational format and inference behavior suggests that abstract/disentangled representational geometries are important for complex cognition.

Selection of experience for memory by hippocampal sharp wave ripples

A general wisdom is that experiences need to be tagged during learning for further consolidation. However, brain mechanisms that select experiences for lasting memory are not known. Combining large-scale neural recordings with a novel application of dimensionality reduction techniques, we observed that successive traversals in the maze were tracked by continuously drifting populations of neurons, providing neuronal signatures of both places visited and events encountered (trial number). When the brain state changed during reward consumption, sharp wave ripples (SPW-Rs) occurred on some trials and their unique spike content most often decoded the trial in which they occurred. In turn, during post-experience sleep, SPW-Rs continued to replay those trials that were reactivated most frequently during awake SPW-Rs. These findings suggest that replay content of awake SPW-Rs provides a tagging mechanism to select aspects of experience that are preserved and consolidated for future use.

Protocol for geometric transformation of cognitive maps for generalization across hippocampal-prefrontal circuits

Memory generalization is the ability to abstract knowledge from prior experiences and is critical for flexible behavior in novel situations. Here, we describe a protocol for simultaneous recording of hippocampal (area CA1)-prefrontal cortical neural ensembles in Long-Evans rats during task generalization across two distinct environments. We describe steps for building and assembling experimental apparatuses, animal preparation and surgery, and performing experiments. We then detail procedures for histology, data processing, and assessing population geometry using Uniform Manifold Approximation and Projection. For complete details on the use and execution of this protocol, please refer to Tang et al. (2023).1.