Distinct ensemble codes in hippocampal areas CA3 and CA1

The hippocampal code for space in Mongolian gerbils

Large parts of our knowledge about the physiology of the hippocampus in the intact brain are derived from studies in rats and mice. While many of those findings fit well to the limited data available from humans and primates, there are also marked differences, for example, in hippocampal oscillation frequencies and in the persistence of theta oscillations. To test whether the distinct sensory specializations of the visual and auditory system of primates play a key role in explaining these differences, we recorded basic hippocampal physiological properties in Mongolian gerbils, a rodent species with high visual acuity, and good low-frequency hearing, similar to humans. We found that gerbils show only minor differences to rats regarding hippocampal place field activity, theta properties (frequency, persistence, phase precession, theta compression), and sharp wave ripple events. The only major difference between rats and gerbils was a considerably higher degree of head direction selectivity of gerbil place fields, which may be explained by their visual system being able to better resolve distant cues. Thus, differences in sensory specializations between rodent species only affect hippocampal circuit dynamics to a minor extent, which implies that differences to other mammalian lineages, such as bats and primates, cannot be solely explained by specialization in the auditory or visual system.

Influence of pharmacological and epigenetic factors to suppress neurotrophic factors and enhance neural plasticity in stress and mood disorders

Stress-induced major depression and mood disorders are characterized by behavioural abnormalities and psychiatric illness, leading to disability and immature mortality worldwide. Neurobiological mechanisms of stress and mood disorders are discussed considering recent findings, and challenges to enhance pharmacological effects of antidepressant, and mood stabilizers. Pharmacological enhancement of ketamine and scopolamine regulates depression at the molecular level, increasing synaptic plasticity in prefrontal regions. Blood-derived neurotrophic factors facilitate mood-deficit symptoms. Epigenetic factors maintain stress-resilience in hippocampal region. Regulation of neurotrophic factors blockades stress, and enhances neuronal survival though it paralyzes limbic regions. Molecular agents and neurotrophic factors also control behavioral and synaptic plasticity in addiction and stress disorders. Future research on neuronal dynamics and cellular actions can be directed to obtain the etiology of synaptic dysregulation in mood disorder and stress. For the first time, the current review contributes to the literature of synaptic plasticity representing the role of epigenetic mechanisms and glucocorticoid receptors to predict depression and anxiety in clinical conditions.

α4βδ GABAA Receptors Trigger Synaptic Pruning and Reduce Dendritic Length of Female Mouse CA3 Hippocampal Pyramidal Cells at Puberty

Synaptic pruning during adolescence is critical for optimal cognition. The CA3 hippocampus contains unique spine types and plays a pivotal role in pattern separation and seizure generation, where sex differences exist, but adolescent pruning has only been studied in the male. Thus, for the present study we assessed pruning of specific spine types in the CA3 hippocampus during adolescence and investigated a possible mechanism in the female mouse. To this end, we used Golgi-impregnated brains from pubertal (∼PND 35, assessed by vaginal opening) and post-pubertal (PND 56) mice. Spine density was assessed from z-stack (0.1-μm steps) images taken using a Nikon DS-U3 camera through a Nikon Eclipse Ci-L microscope and analyzed with NIS Elements. Spine density decreased significantly (P < 0.05) during adolescence, with 50-60% decreases in mushroom and stubby spine-types (P < 0.05, ∼PND35 vs. PND56) in non-proestrous mice. This was associated with decreases in kalirin-7, a spine protein which stabilizes the cytoskeleton and is required for spine maintenance. Because our previous findings suggest that pubertal increases in α4βδ GABAA receptors (GABARs) trigger pruning in CA1, we investigated their role in CA3. α4 expression in CA3 hippocampus increased 4-fold at puberty (P < 0.05), assessed by immunostaining and verified electrophysiologically by an increased response to gaboxadol (100 nM), which is selective for α4βδ. Knock-out of α4 prevented the pubertal decrease in kalirin-7 and synaptic pruning and also increased the dendritic length, demonstrating a functional link. These data suggest that pubertal α4βδ GABARs alter dendritic morphology and trigger pruning in female CA3 hippocampus.

Dendritic mechanisms of hippocampal place field formation

Place cells in the hippocampus are thought to form a cognitive map of space and a memory of places. How this map forms when animals are exposed to novel environments has been the subject of a great deal of research. Numerous technical advances over the past decade greatly increased our understanding of the precise mechanisms underlying place field formation. In particular, it is now possible to connect cellular and circuit mechanisms of integration, firing, and plasticity discovered in brain slices, to processes taking place in vivo as animals learn and encode novel environments. Here, we focus on recent results and describe the dendritic mechanisms most likely responsible for the formation of place fields. We also discuss key open questions that are likely to be answered in the coming years.

Transient hippocampal CA1 lesions in humans impair pattern separation performance

Day-to-day life involves the perception of events that resemble one another. For the sufficient encoding and retrieval of similar information, the hippocampus provides two essential computational processes. Pattern separation refers to the differentiation of overlapping memory representations, whereas pattern completion reactivates memories based on noisy or degraded input. Evidence from human and rodent studies suggest that pattern separation specifically relies on neuronal ensemble activity in hippocampal subnetworks in the dentate gyrus and CA3. Although a role for CA1 in pattern separation has been shown in animal models, its contribution in the human hippocampus remains elusive. In order to elucidate the contribution of CA1 neurons to pattern separation, we examined 14 patients with an acute transient global amnesia (TGA), a rare self-limiting dysfunction of the hippocampal system showing specific lesions to CA1. Patients' pattern separation performance was tested during the acute amnestic phase and follow-up using an established mnemonic similarity test. Patients in the acute phase showed a profound deficit in pattern separation (p < .05) as well as recognition memory (p < .001) that recovered during follow-up. Specifically, patients tested in a later stage of the amnesia were less impaired in pattern separation than in recognition memory. Considering the time dependency of lesion-associated hippocampal deficits in early and late acute stages of the TGA, we showed that the pattern separation function recovered significantly earlier than recognition memory. Our results provide causal evidence that hippocampal CA1 neurons are critical to pattern separation performance in humans.

Through synapses to spatial memory maps via a topological model

Various neurophysiological and cognitive functions are based on transferring information between spiking neurons via a complex system of synaptic connections. In particular, the capacity of presynaptic inputs to influence the postsynaptic outputs–the efficacy of the synapses–plays a principal role in all aspects of hippocampal neurophysiology. However, a direct link between the information processed at the level of individual synapses and the animal’s ability to form memories at the organismal level has not yet been fully understood. Here, we investigate the effect of synaptic transmission probabilities on the ability of the hippocampal place cell ensembles to produce a cognitive map of the environment. Using methods from algebraic topology, we find that weakening synaptic connections increase spatial learning times, produce topological defects in the large-scale representation of the ambient space and restrict the range of parameters for which place cell ensembles are capable of producing a map with correct topological structure. On the other hand, the results indicate a possibility of compensatory phenomena, namely that spatial learning deficiencies may be mitigated through enhancement of neuronal activity.

Using computational models to predict in vivo synaptic inputs to interneuron specific 3 (IS3) cells of CA1 hippocampus that also allow their recruitment during rhythmic states

Brain coding strategies are enabled by the balance of synaptic inputs that individual neurons receive as determined by the networks in which they reside. Inhibitory cell types contribute to brain function in distinct ways but recording from specific, inhibitory cell types during behaviour to determine their contributions is highly challenging. In particular, the in vivo activities of vasoactive intestinal peptide-expressing interneuron specific 3 (IS3) cells in the hippocampus that only target other inhibitory cells are unknown at present. We perform a massive, computational exploration of possible synaptic inputs to IS3 cells using multi-compartment models and optimized synaptic parameters. We find that asynchronous, in vivo-like states that are sensitive to additional theta-timed inputs (8 Hz) exist when excitatory and inhibitory synaptic conductances are approximately equally balanced and with low numbers of activated synapses receiving correlated inputs. Specifically, under these balanced conditions, the input resistance is larger with higher mean spike firing rates relative to other activated synaptic conditions investigated. Incoming theta-timed inputs result in strongly increased spectral power relative to baseline. Thus, using a generally applicable computational approach we predict the existence and features of background, balanced states in hippocampal circuits.

Hippocampal Dentate Gyrus Atrophy Predicts Pattern Separation Impairment in Patients with LGI1 Encephalitis

Day-to-day life involves the perception of events that resemble one another. For the sufficient encoding and correct retrieval of similar information, the hippocampus provides two essential cognitive processes. Pattern separation refers to the differentiation of similar input information, whereas pattern completion reactivates memory representations based on noisy or degraded stimuli. It has been shown that pattern separation specifically relies on the hippocampal dentate gyrus (DG), whereas pattern completion is performed within CA3 networks. Lesions to these hippocampal networks emerging in the course of neurological disorders may thus affect both processes. In anti-leucine-rich, glioma-inactivated 1 (LGI1) encephalitis it has been shown in animal models and human imaging studies that hippocampal DG and CA3 are preferentially involved in the pathophysiology process. Thus, in order to elucidate the structure-function relationship and contribution of hippocampal subfields to pattern separation, we examined patients (n = 15, age range: 36-77 years) with the rare LGI1 encephalitis showing lesions to hippocampal subfields. Patients were tested 3.53 ± 0.65 years after the acute phase of the disease. Structural sequelae were determined by hippocampal subfield volumetry for the DG, CA1, and CA2/3. Patients showed an overall memory deficit including a significant reduction in pattern separation performance (p = 0.016). In volumetry, we found a global hippocampal volume reduction. The deficits in pattern separation performance were best predicted by the DG (p = 0.029), whereas CA1 was highly predictive of recognition memory deficits (p < 0.001). These results corroborate the framework of a regional specialization of hippocampal functions involved in cognitive processing.

Episodic recognition memory and the hippocampus in Parkinson's disease: A review

Parkinson's disease is a progressive neurodegenerative disorder of aging. The hallmark pathophysiology includes the development of neuronal Lewy bodies in the substantia nigra of the midbrain with subsequent loss of dopaminergic neurons. These neuronal losses lead to the characteristic motor symptoms of bradykinesia, rigidity, and rest tremor. In addition to these cardinal motor symptoms patients with PD experience a wide range of non-motor symptoms, the most important being cognitive impairments that in many circumstances lead to dementia. People with PD experience a wide range of cognitive impairments; in this review we will focus on memory impairment in PD and specifically episodic memory, which are memories of day-to-day events of life. Importantly, these memory impairments severely impact the lives of patients and caregivers alike. Traditionally episodic memory is considered to be markedly dependent on the hippocampus; therefore, it is important to understand the exact nature of PD episodic memory deficits in relation to hippocampal function and dysfunction. In this review, we discuss an aspect of episodic memory called recognition memory and its subcomponents called recollection and familiarity. Recognition memory is believed to be impaired in PD; thus, we discuss what aspects of the hippocampus are expected to be deficient in function as they relate to these recognition memory impairments. In addition to the hippocampus as a whole, we will discuss the role of hippocampal subfields in recognition memory impairments.

Assembly Responses of Hippocampal CA1 Place Cells Predict Learned Behavior in Goal-Directed Spatial Tasks on the Radial Eight-Arm Maze

Hippocampus is needed for both spatial working and reference memories. Here, using a radial eight-arm maze, we examined how the combined demand on these memories influenced CA1 place cell assemblies while reference memories were partially updated. This was contrasted with control tasks requiring only working memory or the update of reference memory. Reference memory update led to the reward-directed place field shifts at newly rewarded arms and to the gradual strengthening of firing in passes between newly rewarded arms but not between those passes that included a familiar-rewarded arm. At the maze center, transient network synchronization periods preferentially replayed trajectories of the next chosen arm in reference memory tasks but the previously visited arm in the working memory task. Hence, reference memory demand was uniquely associated with a gradual, goal novelty-related reorganization of place cell assemblies and with trajectory replay that reflected the animal's decision of which arm to visit next.

Differentiation of Human Medial Prefrontal Cortex Activity Underlies Long-Term Resistance to Forgetting in Memory

It is well known that distributing study events over time leads to better memory over long time scales, compared with massing study events together. One explanation for such long-term resistance to forgetting is that distributed study leads to neural differentiation in memory, which supports retrieval of past experiences by disambiguating highly similar memory representations. Neuroanatomical models of episodic memory retrieval propose that the hippocampus and medial prefrontal cortex (MPFC) work together to enable retrieval of behaviorally appropriate memories. However, it is not known how representations in these regions jointly support resistance to forgetting long after initial learning. Using fMRI, we measured differentiation in retrieved memory representations following an extended delay in male and female human participants. After 1 week, word-object associations were better remembered if studied across 2 d (overnight), allowing associations to be learned in distinct temporal contexts, compared with learning within a single day (same day). MPFC retrieval patterns showed differentiation for overnight relative to same day memories, whereas hippocampal patterns reflected associative retrieval success. Overnight memory differentiation in MPFC was higher for associative than item memories and higher than differentiation assessed over a brain-wide set of retrieval-active voxels. The memory-related difference in MPFC pattern differentiation correlated with memory success for overnight learning and with hippocampal-MPFC functional connectivity. These results show that learning information across days leads to differentiated MPFC memory representations, reducing forgetting after 1 week, and suggest this arises from persistent interactions between MPFC and hippocampus.SIGNIFICANCE STATEMENT Neural activity in both the hippocampus and medial prefrontal cortex (MPFC) has been linked to memory-related representations, but prior work has not examined how these representations support episodic memory retrieval over extended time scales that are characteristic of everyday retrieval. We show that differentiation in MPFC activity 1 week after encoding is higher for retrieved information learned across 2 d compared with within a single day. In hippocampus, differentiation was greater for detailed memory retrieval but was not influenced by whether information had been learned over 1 or 2 d. Differentiation in MPFC predicted behavioral robustness to forgetting and was correlated with hippocampal-MPFC connectivity. The results suggest that context-based differentiation supports robust long-term memory via persistent MPFC-hippocampal interactions.

Modeling Contextual Modulation of Memory Associations in the Hippocampus

We present a computational model of how memories can be contextually acquired and recalled in the hippocampus. Our adaptive contextual memory model comprises the lateral entorhinal cortex (LEC), the dentate gyrus (DG) and areas CA3 and CA1 in the hippocampus, and assumes external inputs about context that originate in the prefrontal cortex (PFC). Specifically, we propose that there is a top-down bias on the excitability of cells in the DG of the hippocampus that recruits a sub-population of cells to differentiate contexts, independent of experienced stimuli, expanding the "pattern separation" role typically attributed to the DG. It has been demonstrated in rats that if PFC is inactivated, both acquisition and recall of memory associations are impaired. However, PFC inactivation during acquisition of one set of memory associations surprisingly leads to subsequent facilitation of the acquisition of a conflicting set of memory associations in the same context under normal PFC operation. We provide here the first computational and algorithmic account of how the absence or presence of the top-down contextual biases on the excitability of DG cells during different learning phases of these experiments explains these data. Our model simulates PFC inactivation as the loss of inhibitory control on DG, which leads to full or partial activation of DG cells related to conflicting memory associations previously acquired in different contexts. This causes context-inappropriate memory traces to become active in the CA3 recurrent network and thereby the output CA1 area within the hippocampus. We show that these incongruous memory patterns proactively interfere with and slow the acquisition of new memory associations. Further, we demonstrate that pattern completion within CA3 in response to a partial cue for the recall of previously acquired memories is also impaired by PFC inactivation for the same reason. Pre-training the model with interfering memories in contexts different from those used in the experiments, simulating a lifetime of experiences, was crucial to reproduce the rat behavioral data. Finally, we made several testable predictions based on the model that suggest future experiments to deepen our understanding of brain-wide memory processes.

Hippocampal place cell encoding of sloping terrain

Effective navigation relies on knowledge of one's environment. A challenge to effective navigation is accounting for the time and energy costs of routes. Irregular terrain in ecological environments poses a difficult navigational problem as organisms ought to avoid effortful slopes to minimize travel costs. Route planning and navigation have previously been shown to involve hippocampal place cells and their ability to encode and store information about an organism's environment. However, little is known about how place cells may encode the slope of space and associated energy costs as experiments are traditionally carried out in flat, horizontal environments. We set out to investigate how dorsal-CA1 place cells in rats encode systematic changes to the slope of an environment by tilting a shuttle box from flat to 15 ° and 25 ° while minimizing external cue change. Overall, place cell encoding of tilted space was as robust as their encoding of flat ground as measured by traditional place cell metrics such as firing rates, spatial information, coherence, and field size. A large majority of place cells did, however, respond to slope by undergoing partial, complex remapping when the environment was shifted from one tilt angle to another. The propensity for place cells to remap did not, however, depend on the vertical distance the field shifted. Changes in slope also altered the temporal coding of information as measured by the rate of theta phase precession of place cell spikes, which decreased with increasing tilt angles. Together these observations indicate that place cells are sensitive to relatively small changes in terrain slope and that terrain slope may be an important source of information for organizing place cell ensembles. The terrain slope information encoded by place cells could be utilized by efferent regions to determine energetically advantageous routes to goal locations.

Hippocampal Place Fields Maintain a Coherent and Flexible Map across Long Timescales

To provide a substrate for remembering where in space events have occurred, place cells must reliably encode the same positions across long timescales. However, in many cases, place cells exhibit instability by randomly reorganizing their place fields between experiences, challenging this premise. Recent evidence suggests that, in some cases, instability could also arise from coherent rotations of place fields, as well as from random reorganization. To investigate this possibility, we performed in vivo calcium imaging in dorsal hippocampal region CA1 of freely moving mice while they explored two arenas with different geometry and visual cues across 8 days. The two arenas were rotated randomly between sessions and then connected, allowing us to probe how cue rotations, the integration of new information about the environment, and the passage of time concurrently influenced the spatial coherence of place fields. We found that spatially coherent rotations of place-field maps in the same arena predominated, persisting up to 6 days later, and that they frequently rotated in a manner that did not match that of the arena rotation. Furthermore, place-field maps were flexible, as mice frequently employed a similar, coherent configuration of place fields to represent each arena despite their differing geometry and eventual connection. These results highlight the ability of the hippocampus to retain consistent relationships between cells across long timescales and suggest that, in many cases, apparent instability might result from a coherent rotation of place fields.

Storage fidelity for sequence memory in the hippocampal circuit

Episodic memories have been suggested to be represented by neuronal sequences, which are stored and retrieved from the hippocampal circuit. A special difficulty is that realistic neuronal sequences are strongly correlated with each other since computational memory models generally perform poorly when correlated patterns are stored. Here, we study in a computational model under which conditions the hippocampal circuit can perform this function robustly. During memory encoding, CA3 sequences in our model are driven by intrinsic dynamics, entorhinal inputs, or a combination of both. These CA3 sequences are hetero-associated with the input sequences, so that the network can retrieve entire sequences based on a single cue pattern. We find that overall memory performance depends on two factors: the robustness of sequence retrieval from CA3 and the circuit's ability to perform pattern completion through the feedforward connectivity, including CA3, CA1 and EC. The two factors, in turn, depend on the relative contribution of the external inputs and recurrent drive on CA3 activity. In conclusion, memory performance in our network model critically depends on the network architecture and dynamics in CA3.

Neural Firing Patterns Are More Schematic and Less Sensitive to Changes in Background Visual Scenes in the Subiculum than in the Hippocampus

Literature suggests that the hippocampus is central to processing visual scenes to remember contextual information, but the roles of its downstream structure, subiculum, remain unknown. Here, single units were recorded simultaneously in the dorsal CA1 and subiculum while male rats made spatial choices using visual scenes as cues in a T-maze. The firing fields of subicular neurons were schematically organized following the task structure, largely divided into pre-choice and post-choice epochs, whereas those of CA1 cells were more punctate and bound to specific locations. When the rats were tested with highly familiar scenes, neurons in the CA1 and subiculum were indistinguishable in coding the task-related information (e.g., scene, choice) through rate remapping. However, when the familiar scenes were blurred parametrically, the neurons in the CA1 responded sensitively to the novelty in task demand and changed its representations parametrically following the physical changes of the stimuli, whereas these functional characteristics were absent in the subiculum. These results suggest that the unique function of the hippocampus is to acquire contextual representations in association with discrete positions in space, especially when facing new and ambiguous scenes, whereas the subiculum may translate the position-bound visual contextual information of the hippocampus into schematic codes once learning is established.SIGNIFICANCE STATEMENT Although the potential functional significance has been recognized for decades for the subiculum, its exact roles in a goal-directed memory task still remain elusive. In the current study, we present experimental evidence that may indicate that the neural population in the subiculum could translate the location-bound spatial representations of the hippocampus into more schematic representations of task demands. Our findings also imply that the visual scene-based codes conveyed by the hippocampus and subiculum may be identical in a well learned task, whereas the hippocampus may be more specialized in representing altered visual scenes than the subiculum.

Time-resolved neural reinstatement and pattern separation during memory decisions in human hippocampus

Mnemonic decision-making has long been hypothesized to rely on hippocampal dynamics that bias memory processing toward the formation of new memories or the retrieval of old ones. Successful memory encoding may be best optimized by pattern separation, whereby two highly similar experiences can be represented by underlying neural populations in an orthogonal manner. By contrast, successful memory retrieval is thought to be supported by a recovery of the same neural pattern laid down during encoding. Here we examined how hippocampal pattern completion and separation emerge over time during memory decisions. We measured electrocorticography activity in the human hippocampus and posterior occipitotemporal cortex (OTC) while participants performed continuous recognition of items that were new, repeated (old), or highly similar to a prior item (similar). During retrieval decisions of old items, both regions exhibited significant reinstatement of multivariate high-frequency activity (HFA) associated with encoding. Further, the extent of reinstatement of encoding patterns during retrieval was correlated with the strength (HFA power) of hippocampal encoding. Evidence for encoding pattern reinstatement was also seen in OTC on trials requiring fine-grained discrimination of similar items. By contrast, hippocampal activity showed evidence for pattern separation during these trials. Together, these results underscore the critical role of the hippocampus in supporting both reinstatement of overlapping information and separation of similar events.

Memory Allocation: Mechanisms and Function

Memories for events are thought to be represented in sparse, distributed neuronal ensembles (or engrams). In this article, we review how neurons are chosen to become part of a particular engram, via a process of neuronal allocation. Experiments in rodents indicate that eligible neurons compete for allocation to a given engram, with more excitable neurons winning this competition. Moreover, fluctuations in neuronal excitability determine how engrams interact, promoting either memory integration (via coallocation to overlapping engrams) or separation (via disallocation to nonoverlapping engrams). In parallel with rodent studies, recent findings in humans verify the importance of this memory integration process for linking memories that occur close in time or share related content. A deeper understanding of allocation promises to provide insights into the logic underlying how knowledge is normally organized in the brain and the disorders in which this process has gone awry.

The cognitive nuances of surprising events: exposure to unexpected stimuli elicits firing variations in neurons of the dorsal CA1 hippocampus

The ability to recognize novel situations is among the most fascinating and vital of the brain functions. A hypothesis posits that encoding of novelty is prompted by failures in expectancy, according to computation matching incoming information with stored events. Thus, unexpected changes in context are detected within the hippocampus and transferred to downstream structures, eliciting the arousal of the dopamine system. Nevertheless, the precise locus of detection is a matter of debate. The dorsal CA1 hippocampus (dCA1) appears as an ideal candidate for operating a mismatch computation and discriminating the occurrence of diverse stimuli within the same environment. In this study, we sought to determine dCA1 neuronal firing during the experience of novel stimuli embedded in familiar contexts. We performed population recordings while head-fixed mice navigated virtual environments. Three stimuli were employed, namely a novel pattern of visual cues, an odor, and a reward with enhanced valence. The encounter of unexpected events elicited profound variations in dCA1 that were assessed both as opposite rate directions and altered network connectivity. When experienced in sequence, novel stimuli elicited specific responses that often exhibited cross-sensitization. Short-latency, event-triggered responses were in accordance with the detection of novelty being computed within dCA1. We postulate that firing variations trigger neuronal disinhibition, and constitute a fundamental mechanism in the processing of unexpected events and in learning. Elucidating the mechanisms underlying detection and computation of novelty might help in understanding hippocampal-dependent cognitive dysfunctions associated with neuropathologies and psychiatric conditions.

Development of mnemonic discrimination during childhood

The present study examined mnemonic discrimination in 5- and 6-yr-old children, 8- and 9-yr-old children, 11- and 12-yr-old children, and young adults. Participants incidentally encoded pictorial stimuli and subsequently judged whether targets (i.e., repeated stimuli), lures (i.e., mnemonically related stimuli), and foils (i.e., novel stimuli) were old, similar, or new. Compared to older age groups, younger children were more likely to (1) incorrectly identify lures as “old” (rather than “similar”) and (2) fail to recognize lures altogether, especially when lures were more mnemonically distinct from targets. These results suggest age-related improvements in pattern separation and pattern completion during childhood.

What Learning Systems do Intelligent Agents Need? Complementary Learning Systems Theory Updated

We update complementary learning systems (CLS) theory, which holds that intelligent agents must possess two learning systems, instantiated in mammalians in neocortex and hippocampus. The first gradually acquires structured knowledge representations while the second quickly learns the specifics of individual experiences. We broaden the role of replay of hippocampal memories in the theory, noting that replay allows goal-dependent weighting of experience statistics. We also address recent challenges to the theory and extend it by showing that recurrent activation of hippocampal traces can support some forms of generalization and that neocortical learning can be rapid for information that is consistent with known structure. Finally, we note the relevance of the theory to the design of artificial intelligent agents, highlighting connections between neuroscience and machine learning.

Pattern separation and completion of distinct axonal inputs transmitted via micro-tunnels between co-cultured hippocampal dentate, CA3, CA1 and entorhinal cortex networks

OBJECTIVE

Functions ascribed to the hippocampal sub-regions for encoding episodic memories include the separation of activity patterns propagated from the entorhinal cortex (EC) into the dentate gyrus (DG) and pattern completion in CA3 region. Since a direct assessment of these functions is lacking at the level of specific axonal inputs, our goal is to directly measure the separation and completion of distinct axonal inputs in engineered pairs of hippocampal sub-regional circuits.

APPROACH

We co-cultured EC-DG, DG-CA3, CA3-CA1 or CA1-EC neurons in a two-chamber PDMS device over a micro-electrode array (MEA60), inter-connected via distinct axons that grow through the micro-tunnels between the compartments. Taking advantage of the axonal accessibility, we quantified pattern separation and completion of the evoked activity transmitted through the tunnels from source into target well. Since pattern separation can be inferred when inputs are more correlated than outputs, we first compared the correlations among axonal inputs with those of target somata outputs. We then compared, in an analog approach, the distributions of correlation distances between rate patterns of the axonal inputs inside the tunnels with those of the somata outputs evoked in the target well. Finally, in a digital approach, we measured the spatial population distances between binary patterns of the same axonal inputs and somata outputs.

MAIN RESULTS

We found the strongest separation of the propagated axonal inputs when EC was axonally connected to DG, with a decline in separation to CA3 and to CA1 for both rate and digital approaches. Furthermore, the digital approach showed stronger pattern completion in CA3, then CA1 and EC.

SIGNIFICANCE

To the best of our knowledge, these are the first direct measures of pattern separation and completion for axonal transmission to the somata target outputs at the rate and digital population levels in each of four stages of the EC-DG-CA3-CA1 circuit.

Hippocampal representations as a function of time, subregion, and brain state

How does the hippocampus represent interrelated experiences in memory? We review prominent yet seemingly contradictory theoretical perspectives, which propose that the hippocampus distorts experiential representations to either emphasize their distinctiveness or highlight common elements. These fundamentally different kinds of memory representations may be instantiated in the brain via conjunctive separated codes and adaptively differentiated codes on the one hand, or integrated relational codes on the other. After reviewing empirical support for these different coding schemes within the hippocampus, we outline two organizing principles which may explain the conflicting findings in the literature. First focusing on where the memories are formed and stored, we argue that distinct hippocampal regions represent experiences at multiple levels of abstraction and may transmit them to distinct cortical networks. Then focusing on when memories are formed, we identify several factors that can open and maintain specialized time windows, during which the very same hippocampal network is biased toward one coding scheme over the others. Specifically, we discuss evidence for (1) excitability-mediated integration windows, maintained by persistently elevated CREB levels following encoding of a specific memory, (2) fleeting cholinergically-mediated windows favoring memory separation, and (3) sustained dopaminergically-mediated windows favoring memory integration. By presenting a broad overview of different hippocampal coding schemes across species, we hope to inspire future empirical and modeling research to consider how factors surrounding memory formation shape the representations in which they are stored.

Spatial information is preferentially processed by the distal part of CA3: implication for memory retrieval

For the past decades, CA3 was considered as a single functional entity. However, strong differences between the proximal (close to the dentate gyrus) and the distal (close to CA2) parts of CA3 in terms of connectivity patterns, gene expression and electrophysiological properties suggest that it is not the case. We recently showed that proximal CA3 (together with distal CA1) preferentially deals with non-spatial information [1]. In contrast to proximal CA3, distal CA3 mainly receives and predominantly projects to spatially tuned areas. Here, we tested if distal CA3 preferentially processes spatial information, which would suggest a segregation of the spatial information along the proximodistal axis of CA3. We used a high-resolution imaging technique based on the detection of the expression of the immediate-early gene Arc, commonly used to map activity in the medial temporal lobe. We showed that distal CA3 is strongly recruited in a newly designed delayed nonmatching-to-location task with high memory demands in rats, while proximal CA3 is not. These results indicate a functional segregation of CA3 that mirrors the one reported in CA1, and suggest the existence of a distal CA3- proximal CA1 spatial subnetwork. These findings bring further evidence for the existence of 'specialized' spatial and non-spatial subnetworks segregated along the proximodistal axis of the hippocampus and put forward the 'segregated' view of information processing in the hippocampus as a reasonable alternative to the well-accepted 'integrated' view, according to which spatial and non-spatial information are systematically integrated in the hippocampus to form episodic memory.

Hippocampal ripples down-regulate synapses

The specific effects of sleep on synaptic plasticity remain unclear. We report that mouse hippocampal sharp-wave ripple oscillations serve as intrinsic events that trigger long-lasting synaptic depression. Silencing of sharp-wave ripples during slow-wave states prevented the spontaneous down-regulation of net synaptic weights and impaired the learning of new memories. The synaptic down-regulation was dependent on the N-methyl-d-aspartate receptor and selective for a specific input pathway. Thus, our findings are consistent with the role of slow-wave states in refining memory engrams by reducing recent memory-irrelevant neuronal activity and suggest a previously unrecognized function for sharp-wave ripples.

Early life stress determines the effects of glucocorticoids and stress on hippocampal function: Electrophysiological and behavioral evidence respectively

Exposure to early-life adversity may program brain function to prepare individuals for adaptation to matching environmental contexts. In this study we tested this hypothesis in more detail by examining the effects of early-life stress - induced by raising offspring with limited nesting and bedding material from postnatal days 2-9 - in various behavioral tasks and on synaptic function in adult mice. Early-life stress impaired adult performance in the hippocampal dependent low-arousing object-in-context recognition memory task. This effect was absent when animals were exposed to a single stressor before training. Early-life stress did not alter high-arousing context and auditory fear conditioning. Early-life stress-induced behavioral modifications were not associated with alterations in the dendritic architecture of hippocampal CA1 pyramidal neurons or principal neurons of the basolateral amygdala. However, early-life stress reduced the ratio of NMDA to AMPA receptor-mediated excitatory postsynaptic currents and glutamate release probability specifically in hippocampal CA1 neurons, but not in the basolateral amygdala. These ex vivo effects in the hippocampus were abolished by acute glucocorticoid treatment. Our findings support that early-life stress can hamper object-in-context learning via pre- and postsynaptic mechanisms that affect hippocampal function but these effects are counteracted by acute stress or elevated glucocorticoid levels.

Impairment of neural coordination in hippocampal neuronal ensembles after a psychotomimetic dose of dizocilpine

The discoordination hypothesis of schizophrenia posits discoordination of neural activity as the central mechanism that underlies some psychotic symptoms (including 'hallmark' cognitive symptoms) of schizophrenia. To test this proposition, we studied the activity of hippocampal neurons in urethane anesthetized Long Evans rats after 0.15mg/kg dizocilpine (MK-801), an N-Methyl-d-aspartate (NMDA) glutamate receptor antagonist, which can cause psychotic symptoms in humans and cognitive control impairments in animals. We observed that MK-801 altered the temporal coordination, but not rate, of neuronal firing. Coactivation between neurons increased, driven primarily by increased coincident firing of cell pairs that did not originally fire together before MK-801 injection. Increased pairwise coactivation manifested as disorganized discharge on the level of neuronal ensembles, which in turn could lead to disorganization in information processing. Disorganization of neuronal activity after a psychotomimetic dose of MK-801 supports the discoordination hypothesis of psychosis.

Integrating new findings and examining clinical applications of pattern separation

Pattern separation, the ability to independently represent and store similar experiences, is a crucial facet of episodic memory. Growing evidence suggests that the hippocampus possesses unique circuitry that is computationally capable of resolving mnemonic interference by using pattern separation. In this Review, we discuss recent advances in the understanding of this process and evaluate the caveats and limitations of linking across animal and human studies. We summarize clinical and translational studies using methods that are sensitive to pattern separation impairments, an approach that stems from the fact that the hippocampus is a major site of disruption in many brain disorders. We critically evaluate the assumptions that guide fundamental and translational studies in this area. Finally, we suggest guidelines for future research and offer ways to overcome potential interpretational challenges to increase the utility of pattern separation as a construct that can further understanding of both memory processes and brain disease.

Entorhinal fast-spiking speed cells project to the hippocampus

Our location in space is represented by a spectrum of space and direction-responsive cell types in medial entorhinal cortex and hippocampus. Many cells in these areas respond also to running speed. The presence of local speed-tuned cells is considered a requirement for position to be encoded in a self-motion–dependent manner; however, whether and how speed-responsive cells in entorhinal cortex and hippocampus are functionally connected have not been determined. The present study shows that a large proportion of entorhinal speed cells are fast-spiking with properties similar to those of GABAergic interneurons and that outputs from a subset of these cells, particularly the parvalbumin-expressing subset, form a component of the medial entorhinal input to the hippocampus.

The mammalian positioning system contains a variety of functionally specialized cells in the medial entorhinal cortex (MEC) and the hippocampus. In order for cells in these systems to dynamically update representations in a way that reflects ongoing movement in the environment, they must be able to read out the current speed of the animal. Speed is encoded by speed-responsive cells in both MEC and hippocampus, but the relationship between the two populations has not been determined. We show here that many entorhinal speed cells are fast-spiking putative GABAergic neurons. Using retrograde viral labeling from the hippocampus, we find that a subset of these fast-spiking MEC speed cells project directly to hippocampal areas. This projection contains parvalbumin (PV) but not somatostatin (SOM)-immunopositive cells. The data point to PV-expressing GABAergic projection neurons in MEC as a source for widespread speed modulation and temporal synchronization in entorhinal–hippocampal circuits for place representation.

Behavior-driven arc expression is reduced in all ventral hippocampal subfields compared to CA1, CA3, and dentate gyrus in rat dorsal hippocampus

Anatomical connectivity and lesion studies reveal distinct functional heterogeneity along the dorsal-ventral axis of the hippocampus. The immediate early gene Arc is known to be involved in neural plasticity and memory and can be used as a marker for cell activity that occurs, for example, when hippocampal place cells fire. We report here, that Arc is expressed in a greater proportion of cells in dorsal CA1, CA3, and dentate gyrus (DG), following spatial behavioral experiences compared to ventral hippocampal subregions (dorsal CA1 = 33%; ventral CA1 = 13%; dorsal CA3 = 23%; ventral CA3 = 8%; and dorsal DG = 2.5%; ventral DG = 1.2%). The technique used here to obtain estimates of numbers of behavior-driven cells across the dorsal-ventral axis, however, corresponds quite well with samples from available single unit recording studies. Several explanations for the two- to-threefold reduction in spatial behavior-driven cell activity in the ventral hippocampus can be offered. These include anatomical connectivity differences, differential gain of the self-motion signals that appear to alter the scale of place fields and the proportion of active cells, and possibly variations in the neuronal responses to non-spatial information within the hippocampus along its dorso-ventral axis.

Hippocampal Remapping and Its Entorhinal Origin

The activity of hippocampal cell ensembles is an accurate predictor of the position of an animal in its surrounding space. One key property of hippocampal cell ensembles is their ability to change in response to alterations in the surrounding environment, a phenomenon called remapping. In this review article, we present evidence for the distinct types of hippocampal remapping. The progressive divergence over time of cell ensembles active in different environments and the transition dynamics between pre-established maps are discussed. Finally, we review recent work demonstrating that hippocampal remapping can be triggered by neurons located in the entorhinal cortex.

Linking neuronal structure to function in rodent hippocampus: a methodological prospective

Since the discovery of place cells, hippocampus-dependent spatial navigation has proven to be an ideal model system for resolving the relationship between neural coding and behavior. Electrical recordings from the hippocampal formation in freely moving animals have revealed a rich repertoire of spatial firing patterns and have enormously advanced our understanding of the neural principles of spatial representation. However, limited progress has been achieved in resolving the underlying cellular mechanisms. This is partially attributable to the inability of standard recording techniques to link neuronal structure to function directly. In this review, we summarize recent efforts aimed at filling this gap. We also highlight the development of methodologies that allow functional measurements from identified neuronal elements in behaving rodents. Recent progress in the dentate gyrus serves as a showcase to reveal the potential of such methodologies and the necessity of resolving structure-function relationships in order to access the cellular mechanisms of hippocampal circuit computations.

Sensation during Active Behaviors

A substantial portion of our sensory experience happens during active behaviors such as walking around or paying attention. How do sensory systems work during such behaviors? Neural processing in sensory systems can be shaped by behavior in multiple ways ranging from a modulation of responsiveness or sharpening of tuning to a dynamic change of response properties or functional connectivity. Here, we review recent findings on the modulation of sensory processing during active behaviors in different systems: insect vision, rodent thalamus, and rodent sensory cortices. We discuss the circuit-level mechanisms that might lead to these modulations and their potential role in sensory function. Finally, we highlight the open questions and future perspectives of this exciting new field.

Collective Behavior of Place and Non-place Neurons in the Hippocampal Network

Discussions of the hippocampus often focus on place cells, but many neurons are not place cells in any given environment. Here we describe the collective activity in such mixed populations, treating place and non-place cells on the same footing. We start with optical imaging experiments on CA1 in mice as they run along a virtual linear track and use maximum entropy methods to approximate the distribution of patterns of activity in the population, matching the correlations between pairs of cells but otherwise assuming as little structure as possible. We find that these simple models accurately predict the activity of each neuron from the state of all the other neurons in the network, regardless of how well that neuron codes for position. Our results suggest that understanding the neural activity may require not only knowledge of the external variables modulating it but also of the internal network state.

Sleep in Humans Stabilizes Pattern Separation Performance

Replay of hippocampal neural representations during sleep is thought to promote systems consolidation of declarative memory. How this reprocessing of memory during sleep affects the hippocampal representation itself, is unclear. Here we tested hippocampal stimulus processing (i.e., pattern separation) before and after periods of sleep and wakefulness in humans (female and male participants). Pattern separation deteriorated across the wake period but remained stable across sleep (p = 0.013) with this sleep-wake difference being most pronounced for stimuli with low similarity to targets (p = 0.006). Stimuli with the highest similarity showed a reversed pattern with reduced pattern separation performance after sleep (p = 0.038). Pattern separation performance was positively correlated with sleep spindle density, slow oscillation density, and theta power phase-locked to slow oscillations. Sleep, presumably by neural memory replay, shapes hippocampal representations and enhances computations of pattern separation to subsequent presentation of similar stimuli.SIGNIFICANCE STATEMENT The consolidation of hippocampus-dependent memories is causally related to reactivation during sleep of previously encoded representations. Here, we show that reactivation-based consolidation processes during sleep shape the hippocampal representation itself. We studied the effect of sleep and wakefulness on pattern separation (i.e., orthogonalization of similar representations) and completion performance (i.e., recall of a memory in light of noisy input) that are essential cognitive elements of encoding and retrieval of information by the hippocampus. Our results demonstrate that pattern separation was stabilized after sleep but diminished after wakefulness. We further showed that pattern separation was related to EEG oscillatory parameters of non-REM sleep serving as markers of sleep-dependent memory consolidation and hippocampal reactivation.

Spatial representation in the hippocampal formation: a history

Since the first place cell was recorded and the cognitive-map theory was subsequently formulated, investigation of spatial representation in the hippocampal formation has evolved in stages. Early studies sought to verify the spatial nature of place cell activity and determine its sensory origin. A new epoch started with the discovery of head direction cells and the realization of the importance of angular and linear movement-integration in generating spatial maps. A third epoch began when investigators turned their attention to the entorhinal cortex, which led to the discovery of grid cells and border cells. This review will show how ideas about integration of self-motion cues have shaped our understanding of spatial representation in hippocampal-entorhinal systems from the 1970s until today. It is now possible to investigate how specialized cell types of these systems work together, and spatial mapping may become one of the first cognitive functions to be understood in mechanistic detail.

Increased Prevalence of Calcium Transients across the Dendritic Arbor during Place Field Formation

Hippocampal place cell ensembles form a cognitive map of space during exposure to novel environments. However, surprisingly little evidence exists to support the idea that synaptic plasticity in place cells is involved in forming new place fields. Here we used high-resolution functional imaging to determine the signaling patterns in CA1 soma, dendrites, and axons associated with place field formation when mice are exposed to novel virtual environments. We found that putative local dendritic spikes often occur prior to somatic place field firing. Subsequently, the first occurrence of somatic place field firing was associated with widespread regenerative dendritic events, which decreased in prevalence with increased novel environment experience. This transient increase in regenerative events was likely facilitated by a reduction in dendritic inhibition. Since regenerative dendritic events can provide the depolarization necessary for Hebbian potentiation, these results suggest that activity-dependent synaptic plasticity underlies the formation of many CA1 place fields.

Memory integration constructs maps of space, time, and concepts

Recent evidence demonstrates that new events are learned in the context of their relationships to existing memories. Within the hippocampus and medial prefrontal cortex, related memories are represented by integrated codes that connect events experienced at different times and places. Integrated codes form the basis of spatial, temporal, and conceptual maps of experience. These maps represent information that goes beyond direct experience and support generalization behaviors that require knowledge be used in new ways. The degree to which an individual memory is integrated into a coherent map is determined by its spatial, temporal, and conceptual proximity to existing knowledge. Integration is observed over a wide range of scales, suggesting that memories contain information about both broad and fine-grained contexts.

Field repetition and local mapping in the hippocampus and the medial entorhinal cortex

Hippocampal place cells support spatial cognition and are thought to form the neural substrate of a global "cognitive map." A widely held view is that parts of the hippocampus also underlie the ability to separate patterns or to provide different neural codes for distinct environments. However, a number of studies have shown that in environments composed of multiple, repeating compartments, place cells and other spatially modulated neurons show the same activity in each local area. This repetition of firing fields may reflect pattern completion and may make it difficult for animals to distinguish similar local environments. In this review we 1) highlight some of the navigation difficulties encountered by humans in repetitive environments, 2) summarize literature demonstrating that place and grid cells represent local and not global space, and 3) attempt to explain the origin of these phenomena. We argue that the repetition of firing fields can be a useful tool for understanding the relationship between grid cells in the entorhinal cortex and place cells in the hippocampus, the spatial inputs shared by these cells, and the propagation of spatially related signals through these structures.

Proximodistal Heterogeneity of Hippocampal CA3 Pyramidal Neuron Intrinsic Properties, Connectivity, and Reactivation during Memory Recall

The hippocampal CA3 region is classically viewed as a homogeneous autoassociative network critical for associative memory and pattern completion. However, recent evidence has demonstrated a striking heterogeneity along the transverse, or proximodistal, axis of CA3 in spatial encoding and memory. Here we report the presence of striking proximodistal gradients in intrinsic membrane properties and synaptic connectivity for dorsal CA3. A decreasing gradient of mossy fiber synaptic strength along the proximodistal axis is mirrored by an increasing gradient of direct synaptic excitation from entorhinal cortex. Furthermore, we uncovered a nonuniform pattern of reactivation of fear memory traces, with the most robust reactivation during memory retrieval occurring in mid-CA3 (CA3b), the region showing the strongest net recurrent excitation. Our results suggest that heterogeneity in both intrinsic properties and synaptic connectivity may contribute to the distinct spatial encoding and behavioral role of CA3 subregions along the proximodistal axis.

Theta sequences of grid cell populations can provide a movement-direction signal

It has been proposed that path integration in mammals is performed by the convergence of internally generated speed and directional inputs onto grid cells. Although this hypothesis has been supported by the discovery that head direction, speed, and grid cells are intermixed within entorhinal cortex and by the recent finding that head-direction inputs are necessary for grid firing, many details on how grid cells are generated have remained elusive. For example, analysis of recording data suggests that substituting head direction for movement direction accrues errors that preclude the formation of grid patterns. To address this discrepancy, we propose that the organization of grid networks makes it plausible that movement-direction signals are an output from grid cells and that temporally precise grid cell sequences provide a robust directional signal to other spatial and directional cell types.

Dedicated Hippocampal Inhibitory Networks for Locomotion and Immobility

Network activity is strongly tied to animal movement; however, hippocampal circuits selectively engaged during locomotion or immobility remain poorly characterized. Here we examined whether distinct locomotor states are encoded differentially in genetically defined classes of hippocampal interneurons. To characterize the relationship between interneuron activity and movement, we used in vivo, two-photon calcium imaging in CA1 of male and female mice, as animals performed a virtual-reality (VR) track running task. We found that activity in most somatostatin-expressing and parvalbumin-expressing interneurons positively correlated with locomotion. Surprisingly, nearly one in five somatostatin or one in seven parvalbumin interneurons were inhibited during locomotion and activated during periods of immobility. Anatomically, the somata of somatostatin immobility-activated neurons were smaller than those of movement-activated neurons. Furthermore, immobility-activated interneurons were distributed across cell layers, with somatostatin-expressing cells predominantly in stratum oriens and parvalbumin-expressing cells mostly in stratum pyramidale. Importantly, each cell's correlation between activity and movement was stable both over time and across VR environments. Our findings suggest that hippocampal interneuronal microcircuits are preferentially active during either movement or immobility periods. These inhibitory networks may regulate information flow in "labeled lines" within the hippocampus to process information during distinct behavioral states.SIGNIFICANCE STATEMENT The hippocampus is required for learning and memory. Movement controls network activity in the hippocampus but it's unclear how hippocampal neurons encode movement state. We investigated neural circuits active during locomotion and immobility and found interneurons were selectively active during movement or stopped periods, but not both. Each cell's response to locomotion was consistent across time and environments, suggesting there are separate dedicated circuits for processing information during locomotion and immobility. Understanding how the hippocampus switches between different network configurations may lead to therapeutic approaches to hippocampal-dependent dysfunctions, such as Alzheimer's disease or cognitive decline.

Benefits of flexible prioritization in working memory can arise without costs

Most recent models conceptualize working memory (WM) as a continuous resource, divided up according to task demands. When an increasing number of items need to be remembered, each item receives a smaller chunk of the memory resource. These models predict that the allocation of attention to high-priority WM items during the retention interval should be a zero-sum game: improvements in remembering cued items come at the expense of uncued items because resources are dynamically transferred from uncued to cued representations. The current study provides empirical data challenging this model. Four precision retrocueing WM experiments assessed cued and uncued items on every trial. This permitted a test for trade-off of the memory resource. We found no evidence for trade-offs in memory across trials. Moreover, robust improvements in WM performance for cued items came at little or no cost to uncued items that were probed afterward, thereby increasing the net capacity of WM relative to neutral cueing conditions. An alternative mechanism of prioritization proposes that cued items are transferred into a privileged state within a response-gating bottleneck, in which an item uniquely controls upcoming behavior. We found evidence consistent with this alternative. When an uncued item was probed first, report of its orientation was biased away from the cued orientation to be subsequently reported. We interpret this bias as competition for behavioral control in the output-driving bottleneck. Other items in WM did not bias each other, making this result difficult to explain with a shared resource model.

This study challenges the dominant model for how we remember and prioritize pieces of information over short intervals (working memory). The dominant view is that all items in working memory share a single resource, and that we can prioritize one item by redistributing resources in its favor. This view predicts that nonprioritized memories become lost or impoverished. By testing how well participants remember both prioritized and nonprioritized items, we show that this is not the case. Our findings suggest that memories can be prioritized flexibly without necessarily jeopardizing others that may still become relevant.

From grid cells to place cells with realistic field sizes

While grid cells in the medial entorhinal cortex (MEC) of rodents have multiple, regularly arranged firing fields, place cells in the cornu ammonis (CA) regions of the hippocampus mostly have single spatial firing fields. Since there are extensive projections from MEC to the CA regions, many models have suggested that a feedforward network can transform grid cell firing into robust place cell firing. However, these models generate place fields that are consistently too small compared to those recorded in experiments. Here, we argue that it is implausible that grid cell activity alone can be transformed into place cells with robust place fields of realistic size in a feedforward network. We propose two solutions to this problem. Firstly, weakly spatially modulated cells, which are abundant throughout EC, provide input to downstream place cells along with grid cells. This simple model reproduces many place cell characteristics as well as results from lesion studies. Secondly, the recurrent connections between place cells in the CA3 network generate robust and realistic place fields. Both mechanisms could work in parallel in the hippocampal formation and this redundancy might account for the robustness of place cell responses to a range of disruptions of the hippocampal circuitry.