The generative grammar of the brain: a critique of internally generated representations.

The past decade of progress in neurobiology has uncovered important organizational principles for network preconfiguration and neuronal selection that suggest a generative grammar exists in the brain. In this Perspective, I discuss the competence of the hippocampal neural network to generically express temporally compressed sequences of neuronal firing that represent novel experiences, which is envisioned as a form of generative neural syntax supporting a neurobiological perspective on brain function. I compare this neural competence with the hippocampal network performance that represents specific experiences with higher fidelity after new learning during replay, which is envisioned as a form of neural semantic that supports a complementary neuropsychological perspective. I also demonstrate how the syntax of network competence emerges a priori during early postnatal life and is followed by the later development of network performance that enables rapid encoding and memory consolidation. Thus, I propose that this generative grammar of the brain is essential for internally generated representations, which are crucial for the cognitive processes underlying learning and memory, prospection, and inference, which ultimately underlie our reason and representation of the world.

The Ontogeny of Hippocampus-Dependent Memories.

The formation of memories that contain information about the specific time and place of acquisition, which are commonly referred to as "autobiographical" or "episodic" memories, critically relies on the hippocampus and on a series of interconnected structures located in the medial temporal lobe of the mammalian brain. The observation that adults retain very few of these memories from the first years of their life has fueled a long-standing debate on whether infants can make the types of memories that in adults are processed by the hippocampus-dependent memory system, and whether the hippocampus is involved in learning and memory processes early in life. Recent evidence shows that, even at a time when its circuitry is not yet mature, the infant hippocampus is able to produce long-lasting memories. However, the ability to acquire and store such memories relies on molecular pathways and network-based activity dynamics different from the adult system, which mature with age. The mechanisms underlying the formation of hippocampus-dependent memories during infancy, and the role that experience exerts in promoting the maturation of the hippocampus-dependent memory system, remain to be understood. In this review, we discuss recent advances in our understanding of the ontogeny and the biological correlates of hippocampus-dependent memories.

Spectral forecast: A general purpose prediction model as an alternative to classical neural networks.

Here, we describe a general-purpose prediction model. Our approach requires three matrices of equal size and uses two equations to determine the behavior against two possible outcomes. We use an example based on photon-pixel coupling data to show that in humans, this solution can indicate the predisposition to disease. An implementation of this model is made available in the supplementary material.

Preconfigured patterns are the primary driver of offline multi-neuronal sequence replay.

Spontaneous neuronal ensemble activity in the hippocampus is believed to result from a combination of preconfigured internally generated dynamics and the unique patterns of activity driven by recent experience. Previous research has established that preconfigured sequential neuronal patterns (i.e., preplay) contribute to the expression of future place cell sequences, which in turn contribute to the sequential neuronal patterns expressed post-experience (i.e., replay). The relative contribution of preconfigured and of experience-related factors to replay and to overall sequential activity during post-run sleep is believed to be highly biased toward the recent run experience, despite never being tested directly. Here, we use multi-neuronal sequence analysis unbiased by firing rate to compute and directly compare the contributions of internally generated and of recent experience-driven factors to the sequential neuronal activity in post-run sleep in naïve adult rats. We find that multi-neuronal sequences during post-run sleep are dominantly contributed by the pre-run preconfigured patterns and to a much smaller extent by the place cell sequences and associated awake rest multi-neuronal sequences experienced during de novo run session, which are weakly and similarly correlated with pre- and post-run sleep multi-neuronal sequences. These findings indicate a robust default internal organization of the hippocampal network into sequential neuronal ensembles that withstands a de novo spatial experience and suggest that integration of novel information during de novo experience leading to lasting changes in sequential network patterns is much more subtle than previously assumed.

Preplay of future place cell sequences by hippocampal cellular assemblies.

During spatial exploration, hippocampal neurons show a sequential firing pattern in which individual neurons fire specifically at particular locations along the animal's trajectory (place cells). According to the dominant model of hippocampal cell assembly activity, place cell firing order is established for the first time during exploration, to encode the spatial experience, and is subsequently replayed during rest or slow-wave sleep for consolidation of the encoded experience. Here we report that temporal sequences of firing of place cells expressed during a novel spatial experience occurred on a significant number of occasions during the resting or sleeping period preceding the experience. This phenomenon, which is called preplay, occurred in disjunction with sequences of replay of a familiar experience. These results suggest that internal neuronal dynamics during resting or sleep organize hippocampal cellular assemblies into temporal sequences that contribute to the encoding of a related novel experience occurring in the future.

Distinct preplay of multiple novel spatial experiences in the rat.

The activity of ensembles of hippocampal place cells represents a hallmark of an animal's spatial experience. The neuronal mechanisms that enable the rapid expression of novel place cell sequences are not entirely understood. Here we report that during sleep or rest, distinct sets of hippocampal temporal sequences in the rat preplay multiple corresponding novel spatial experiences with high specificity. These findings suggest that the place cell sequence of a novel spatial experience is determined, in part, by an online selection of a subset of cellular firing sequences from a larger repertoire of preexisting temporal firing sequences in the hippocampal cellular assembly network that become rapidly bound to the novel experience. We estimate that for the given context, the recorded hippocampal network activity has the capacity to preplay an extended repertoire of at least 15 future spatial experiences of similar distinctiveness and complexity.

Geometric experience sculpts the development and dynamics of hippocampal sequential cell assemblies.

Euclidean space is the fabric of the world we live in. Whether and how geometric experience shapes our spatial-temporal representations of the world remained unknown. We deprived rats of experience with crucial features of Euclidean geometry by rearing them inside translucent spheres, and compared activity of large hippocampal neuronal ensembles during navigation and sleep with that of cuboid cage-reared controls. Sphere-rearing from birth permitted emergence of accurate neuronal ensemble spatial codes and preconfigured and plastic time-compressed neuronal sequences. However, sphere-rearing led to diminished individual place cell tuning, similar neuronal mapping of different track ends/corners, and impaired neuronal pattern separation and plasticity of multiple linear track experiences, partly driven by reduced preconfigured network repertoires. Subsequent experience with multiple linear environments over four days largely reversed these effects, substantiating the role of geometric experience on hippocampal neural development. Thus, early-life experience with Euclidean geometry enriches the hippocampal repertoire of preconfigured neuronal patterns selected toward unique representation and discrimination of multiple linear environments.

Orientation selectivity enhances context generalization and generative predictive coding in the hippocampus.

We learn and remember multiple new experiences throughout the day. The neural principles enabling continuous rapid learning and formation of distinct representations of numerous sequential experiences without major interference are not understood. To understand this process, here we interrogated ensembles of hippocampal place cells as rats explored 15 novel linear environments interleaved with sleep sessions over continuous 16 h periods. Remarkably, we found that a population of place cells were selective to environment orientation and topology. This orientation selectivity property biased the network-level discrimination and re/mapping between multiple environments. Novel environmental representations emerged rapidly as more generic, but repeated experience within the environments subsequently enhanced their discriminability. Generalization of prior experience with different environments consequently improved network predictability of future novel environmental representations via strengthened generative predictive codes. These coding schemes reveal a high-capacity, high-efficiency neuronal framework for rapid representation of numerous sequential experiences with optimal discrimination-generalization balance and reduced interference.

Temporal encoding of place sequences by hippocampal cell assemblies.

Both episodic memory and spatial navigation require temporal encoding of the relationships between events or locations. In a linear maze, ordered spatial distances between sequential locations were represented by the temporal relations of hippocampal place cell pairs within cycles of theta oscillation in a compressed manner. Such correlations could arise due to spike "phase precession" of independent neurons driven by common theta pacemaker or as a result of temporal coordination among specific hippocampal cell assemblies. We found that temporal correlation between place cell pairs was stronger than predicted by a pacemaker drive of independent neurons, indicating a critical role for synaptic interactions and precise timing within and across cell assemblies in place sequence representation. CA1 and CA3 ensembles, identifying spatial locations, were active preferentially on opposite phases of theta cycles. These observations suggest that interleaving CA3 neuronal sequences bind CA1 assemblies representing overlapping past, present, and future locations into single episodes.

Organization of cell assemblies in the hippocampus.

Neurons can produce action potentials with high temporal precision. A fundamental issue is whether, and how, this capability is used in information processing. According to the 'cell assembly' hypothesis, transient synchrony of anatomically distributed groups of neurons underlies processing of both external sensory input and internal cognitive mechanisms. Accordingly, neuron populations should be arranged into groups whose synchrony exceeds that predicted by common modulation by sensory input. Here we find that the spike times of hippocampal pyramidal cells can be predicted more accurately by using the spike times of simultaneously recorded neurons in addition to the animals location in space. This improvement remained when the spatial prediction was refined with a spatially dependent theta phase modulation. The time window in which spike times are best predicted from simultaneous peer activity is 10-30 ms, suggesting that cell assemblies are synchronized at this timescale. Because this temporal window matches the membrane time constant of pyramidal neurons, the period of the hippocampal gamma oscillation and the time window for synaptic plasticity, we propose that cooperative activity at this timescale is optimal for information transmission and storage in cortical circuits.

Cell assemblies, sequences and temporal coding in the hippocampus.

Like social networks, neurons in the brain are organized in neuronal ensembles that constrain and at the same time enrich the role and temporal precision of activity of individual neurons. Changes in coordinated firing across cortical neurons as well as selective changes in timing and sequential order across neurons that are important for encoding of novel information have collectively been known as ensemble temporal coding. Here we review recent findings on the role of online and offline temporal coding within sequential cell assemblies from the rodent hippocampus thought be important for memory encoding and consolidation and for spatial navigation. We propose that temporal coding in the rodent hippocampus represented as plasticity in replay activity relies primarily on subtle and selective changes in coordinated firing within the microstructure of individual cell assembly organization during sleep.

A consensus statement: defining terms for reactivation analysis.

During a two-day Royal Society meeting entitled 'Memory reactivation: replaying events past, present and future' held at Chicheley Hall in May, 2019, we discussed and defined a set of terms for investigating and reporting in memory reactivations to facilitate a common language and thus simplifying comparison of results. Here, we present the results of the discussion and supply a set of terms such as reactivation and replay, for which the authors have reached a common consensus. This article is part of the Theo Murphy meeting issue 'Memory reactivation: replaying events past, present and future'.

Strengthened Temporal Coordination within Pre-existing Sequential Cell Assemblies Supports Trajectory Replay.

A central goal in learning and memory research is to reveal the neural substrates underlying episodic memory formation. The hallmark of sequential spatial trajectory learning, a model of episodic memory, has remained equivocal, with proposals ranging from de novo creation of compressed sequential replay from blank slate networks to selection of pre-existing compressed preplay sequences. Here, we show that increased millisecond-timescale activation of cell assemblies expressed during de novo sequential experience and increased neuronal firing rate correlations can explain the difference between post-experience trajectory replay and robust preplay. This increased activation results from an improved neuronal tuning to specific cell assemblies, higher recruitment of experience-tuned neurons into pre-existing cell assemblies, and increased recruitment of cell assemblies in replay. In contrast, changes in overall neuronal and cell assembly temporal order within extended sequences do not account for sequential trajectory learning. We propose the coordinated strengthening of cell assemblies played sequentially on robust pre-existing temporal frameworks could support rapid formation of episodic-like memory.

Place representation within hippocampal networks is modified by long-term potentiation.

In the brain, information is encoded by the firing patterns of neuronal ensembles and the strength of synaptic connections between individual neurons. We report here that representation of the environment by "place" cells is altered by changing synaptic weights within hippocampal networks. Long-term potentiation (LTP) of intrinsic hippocampal pathways abolished existing place fields, created new place fields, and rearranged the temporal relationship within the affected population. The effect of LTP on neuron discharge was rate and context dependent. The LTP-induced "remapping" occurred without affecting the global firing rate of the network. The findings support the view that learned place representation can be accomplished by LTP-like synaptic plasticity within intrahippocampal networks.

Generative Predictive Codes by Multiplexed Hippocampal Neuronal Tuplets.

Rapid internal representations are continuously formed based on single experiential episodes in space and time, but the neuronal ensemble mechanisms enabling rapid encoding without constraining the capacity for multiple distinct representations are unknown. We developed a probabilistic statistical model of hippocampal spontaneous sequential activity and revealed existence of an internal model of generative predictive codes for the regularities of multiple future novel spatial sequences. During navigation, the inferred difference between external stimuli and the internal model was encoded by emergence of intrinsic-unlikely, novel functional connections, which updated the model by preferentially potentiating post-experience. This internal model and these predictive codes depended on neuronal organization into inferred modules of short, high-repeat sequential neuronal "tuplets" operating as "neuro-codons." We propose that flexible multiplexing of neuronal tuplets into repertoires of extended sequences vastly expands the capacity of hippocampal predictive codes, which could initiate top-down hierarchical cortical loops for spatial and mental navigation and rapid learning.

Impaired Tuning of Neural Ensembles and the Pathophysiology of Schizophrenia: A Translational and Computational Neuroscience Perspective.

The functional optimization of neural ensembles is central to human higher cognitive functions. When the functions through which neural activity is tuned fail to develop or break down, symptoms and cognitive impairments arise. This review considers ways in which disturbances in the balance of excitation and inhibition might develop and be expressed in cortical networks in association with schizophrenia. This presentation is framed within a developmental perspective that begins with disturbances in glutamate synaptic development in utero. It considers developmental correlates and consequences, including compensatory mechanisms that increase intrinsic excitability or reduce inhibitory tone. It also considers the possibility that these homeostatic increases in excitability have potential negative functional and structural consequences. These negative functional consequences of disinhibition may include reduced working memory-related cortical activity associated with the downslope of the "inverted-U" input-output curve, impaired spatial tuning of neural activity and impaired sparse coding of information, and deficits in the temporal tuning of neural activity and its implication for neural codes. The review concludes by considering the functional significance of noisy activity for neural network function. The presentation draws on computational neuroscience and pharmacologic and genetic studies in animals and humans, particularly those involving N-methyl-D-aspartate glutamate receptor antagonists, to illustrate principles of network regulation that give rise to features of neural dysfunction associated with schizophrenia. While this presentation focuses on schizophrenia, the general principles outlined in the review may have broad implications for considering disturbances in the regulation of neural ensembles in psychiatric disorders.

Hippocampal pyramidal cell-interneuron spike transmission is frequency dependent and responsible for place modulation of interneuron discharge.

The interplay between principal cells and interneurons plays an important role in timing the activity of individual cells. We investigated the influence of single hippocampal CA1 pyramidal cells on putative interneurons. The activity of CA1 pyramidal cells was controlled intracellularly by current injection, and the activity of neighboring interneurons was recorded extracellularly in the urethane-anesthetized rat. Spike transmission probability between monosynaptically connected pyramidal cell-interneuron pairs was frequency dependent and highest between 5 and 25 Hz. In the awake animal, interneurons were found that had place-modulated firing rates, with place maps similar to their presynaptic pyramidal neuron. Thus, single pyramidal neurons can effectively determine the firing patterns of their interneuron targets.

Selection of preconfigured cell assemblies for representation of novel spatial experiences.

Internal representations about the external world can be driven by the external stimuli or can be internally generated in their absence. It has been a matter of debate whether novel stimuli from the external world are instructive over the brain network to create de novo representations or, alternatively, are selecting from existing pre-representations hosted in preconfigured brain networks. The hippocampus is a brain area necessary for normal internally generated spatial-temporal representations and its dysfunctions have resulted in anterograde amnesia, impaired imagining of new experiences, and hallucinations. The compressed temporal sequence of place cell activity in the rodent hippocampus serves as an animal model of internal representation of the external space. Based on our recent results on the phenomenon of novel place cell sequence preplay, we submit that the place cell sequence of a novel spatial experience is determined, in part, by a selection of a set of cellular firing sequences from a repertoire of existing temporal firing sequences in the hippocampal network. Conceptually, this indicates that novel stimuli from the external world select from their pre-representations rather than create de novo our internal representations of the world.

Development of schemas revealed by prior experience and NMDA receptor knock-out.

Prior experience accelerates acquisition of novel, related information through processes like assimilation into mental schemas, but the underlying neuronal mechanisms are poorly understood. We investigated the roles that prior experience and hippocampal CA3 N-Methyl-D-aspartate receptor (NMDAR)-dependent synaptic plasticity play in CA1 place cell sequence encoding and learning during novel spatial experiences. We found that specific representations of de novo experiences on linear environments were formed on a framework of pre configured network activity expressed in the preceding sleep and were rapidly, flexibly adjusted via NMDAR-dependent activity. This prior experience accelerated encoding of subsequent experiences on contiguous or isolated novel tracks, significantly decreasing their NMDAR-dependence. Similarly, de novo learning of an alternation task was facilitated by CA3 NMDARs; this experience accelerated subsequent learning of related tasks, independent of CA3 NMDARs, consistent with a schema-based learning. These results reveal the existence of distinct neuronal encoding schemes which could explain why hippocampal dysfunction results in anterograde amnesia while sparing recollection of old, schema-based memories. DOI: http://dx.doi.org/10.7554/eLife.01326.001.