Rapid Encoding of New Memories by Individual Neurons in the Human Brain

Engrams of Fast Learning

Fast learning designates the behavioral and neuronal mechanisms underlying the acquisition of a long-term memory trace after a unique and brief experience. As such it is opposed to incremental, slower reinforcement or procedural learning requiring repetitive training. This learning process, found in most animal species, exists in a large spectrum of natural behaviors, such as one-shot associative, spatial, or perceptual learning, and is a core principle of human episodic memory. We review here the neuronal and synaptic long-term changes associated with fast learning in mammals and discuss some hypotheses related to their underlying mechanisms. We first describe the variety of behavioral paradigms used to test fast learning memories: those preferentially involve a single and brief (from few hundred milliseconds to few minutes) exposures to salient stimuli, sufficient to trigger a long-lasting memory trace and new adaptive responses. We then focus on neuronal activity patterns observed during fast learning and the emergence of long-term selective responses, before documenting the physiological correlates of fast learning. In the search for the engrams of fast learning, a growing body of evidence highlights long-term changes in gene expression, structural, intrinsic, and synaptic plasticities. Finally, we discuss the potential role of the sparse and bursting nature of neuronal activity observed during the fast learning, especially in the induction plasticity mechanisms leading to the rapid establishment of long-term synaptic modifications. We conclude with more theoretical perspectives on network dynamics that could enable fast learning, with an overview of some theoretical approaches in cognitive neuroscience and artificial intelligence.

The Architecture of Human Memory: Insights from Human Single-Neuron Recordings

Deciphering the mechanisms of human memory is a central goal of neuroscience, both from the point of view of the fundamental biology of memory and for its translational relevance. Here, we review some contributions that recordings from neurons in humans implanted with electrodes for clinical purposes have made toward this goal. Recordings from the medial temporal lobe, including the hippocampus, reveal the existence of two classes of cells: those encoding highly selective and invariant representations of abstract concepts, and memory-selective cells whose activity is related to familiarity and episodic retrieval. Insights derived from observing these cells in behaving humans include that semantic representations are activated before episodic representations, that memory content and memory strength are segregated, and that the activity of both types of cells is related to subjective awareness as expected from a substrate for declarative memory. Visually selective cells can remain persistently active for several seconds, thereby revealing a cellular substrate for working memory in humans. An overarching insight is that the neural code of human memory is interpretable at the single-neuron level. Jointly, intracranial recording studies are starting to reveal aspects of the building blocks of human memory at the single-cell level. This work establishes a bridge to cellular-level work in animals on the one hand, and the extensive literature on noninvasive imaging in humans on the other hand. More broadly, this work is a step toward a detailed mechanistic understanding of human memory that is needed to develop therapies for human memory disorders.

Spatiotemporal pattern of brain electrical activity related to immediate and delayed episodic memory retrieval

In the present study we used the event-related brain potentials (ERP) technique and eLORETA (exact low-resolution electromagnetic tomography) method in order to characterize and compare the performance and the spatiotemporal pattern of the brain electrical activity related to the immediate episodic retrieval of information (words) that is being learned relative to delayed episodic retrieval twenty-minutes later. For this purpose, 16 young participants carried out an old/new word recognition task with source memory (word colour). The task included an immediate memory phase (with three study-test blocks) followed (20 min later) by a delayed memory phase with one test block. The behavioural data showed progressive learning and consolidation of the information (old words) during the immediate memory phase. The ERP data to correctly identified old words for which the colour was subsequently recollected (H/H) compared to the correctly rejected new words (CR) showed: (1) a significant more positive-going potential in the 500-675 ms post-stimulus interval (parietal old/new effect, related to recollection), and (2) a more negative-going potential in the 950-1850 ms interval (LPN effect, related to retrieval and post-retrieval processes). The eLORETA data also revealed that the successful recognition of old words (and probably retrieval of their colour) was accompanied by activation of (1) left medial temporal (parahippocampal gyrus) and parietal regions involved in the recollection in both memory phases, and (2) prefrontal regions and the superior temporal gyrus (in the immediate and delayed memory phases respectively) involved in monitoring, evaluating and maintaining the retrieval products. These findings indicate that episodic memory retrieval depends on a network involving medial temporal lobe and frontal, parietal and temporal neocortical structures. That network was involved in immediate and delayed memory retrieval and during the course of memory consolidation, with greater activation of some nodes (mobilization of more processing resources) for the delayed respect to the immediate retrieval condition.

Emergence of Stable Synaptic Clusters on Dendrites Through Synaptic Rewiring

The connectivity structure of neuronal networks in cortex is highly dynamic. This ongoing cortical rewiring is assumed to serve important functions for learning and memory. We analyze in this article a model for the self-organization of synaptic inputs onto dendritic branches of pyramidal cells. The model combines a generic stochastic rewiring principle with a simple synaptic plasticity rule that depends on local dendritic activity. In computer simulations, we find that this synaptic rewiring model leads to synaptic clustering, that is, temporally correlated inputs become locally clustered on dendritic branches. This empirical finding is backed up by a theoretical analysis which shows that rewiring in our model favors network configurations with synaptic clustering. We propose that synaptic clustering plays an important role in the organization of computation and memory in cortical circuits: we find that synaptic clustering through the proposed rewiring mechanism can serve as a mechanism to protect memories from subsequent modifications on a medium time scale. Rewiring of synaptic connections onto specific dendritic branches may thus counteract the general problem of catastrophic forgetting in neural networks.

The role of mPFC and MTL neurons in human choice under goal-conflict

Resolving approach-avoidance conflicts relies on encoding motivation outcomes and learning from past experiences. Accumulating evidence points to the role of the Medial Temporal Lobe (MTL) and Medial Prefrontal Cortex (mPFC) in these processes, but their differential contributions have not been convincingly deciphered in humans. We detect 310 neurons from mPFC and MTL from patients with epilepsy undergoing intracranial recordings and participating in a goal-conflict task where rewards and punishments could be controlled or not. mPFC neurons are more selective to punishments than rewards when controlled. However, only MTL firing following punishment is linked to a lower probability for subsequent approach behavior. mPFC response to punishment precedes a similar MTL response and affects subsequent behavior via an interaction with MTL firing. We thus propose a model where approach-avoidance conflict resolution in humans depends on outcome value tagging in mPFC neurons influencing encoding of such value in MTL to affect subsequent choice.

Brain computation by assemblies of neurons

Our expanding understanding of the brain at the level of neurons and synapses, and the level of cognitive phenomena such as language, leaves a formidable gap between these two scales. Here we introduce a computational system which promises to bridge this gap: the Assembly Calculus. It encompasses operations on assemblies of neurons, such as project, associate, and merge, which appear to be implicated in cognitive phenomena, and can be shown, analytically as well as through simulations, to be plausibly realizable at the level of neurons and synapses. We demonstrate the reach of this system by proposing a brain architecture for syntactic processing in the production of language, compatible with recent experimental results.

Assemblies are large populations of neurons believed to imprint memories, concepts, words, and other cognitive information. We identify a repertoire of operations on assemblies. These operations correspond to properties of assemblies observed in experiments, and can be shown, analytically and through simulations, to be realizable by generic, randomly connected populations of neurons with Hebbian plasticity and inhibition. Assemblies and their operations constitute a computational model of the brain which we call the Assembly Calculus, occupying a level of detail intermediate between the level of spiking neurons and synapses and that of the whole brain. The resulting computational system can be shown, under assumptions, to be, in principle, capable of carrying out arbitrary computations. We hypothesize that something like it may underlie higher human cognitive functions such as reasoning, planning, and language. In particular, we propose a plausible brain architecture based on assemblies for implementing the syntactic processing of language in cortex, which is consistent with recent experimental results.

A Model for Structured Information Representation in Neural Networks of the Brain

Humans can reason at an abstract level and structure information into abstract categories, but the underlying neural processes have remained unknown. Recent experimental data provide the hint that this is likely to involve specific subareas of the brain from which structural information can be decoded. Based on this data, we introduce the concept of assembly projections, a general principle for attaching structural information to content in generic networks of spiking neurons. According to the assembly projections principle, structure-encoding assemblies emerge and are dynamically attached to content representations through Hebbian plasticity mechanisms. This model provides the basis for explaining a number of experimental data and provides a basis for modeling abstract computational operations of the brain.

Single Neuron Coding of Identity in the Human Hippocampal Formation

Experimental findings show the ubiquitous presence of graded responses and tuning curves in the neocortex, particularly in visual areas [1-15]. Among these, inferotemporal-cortex (IT) neurons respond to complex visual stimuli, but differences in the neurons' responses can be used to distinguish the stimuli eliciting the responses [8, 9, 16-18]. The IT projects directly to the medial temporal lobe (MTL) [19], where neurons respond selectively to different pictures of specific persons and even to their written and spoken names [20-22]. However, it is not clear whether this is done through a graded coding, as in the neocortex, or a truly invariant code, in which the response-eliciting stimuli cannot be distinguished from each other. To address this issue, we recorded single neurons during the repeated presentation of different stimuli (pictures and written and spoken names) corresponding to the same persons. Using statistical tests and a decoding approach, we found that only in a minority of cases can the different pictures of a given person be distinguished from the neurons' responses and that in a larger proportion of cases, the responses to the pictures were different to the ones to the written and spoken names. We argue that MTL neurons tend to lack a representation of sensory features (particularly within a sensory modality), which can be advantageous for the memory function attributed to this area [23-25], and that a full representation of memories is given by a combination of mostly invariant coding in the MTL with a representation of sensory features in the neocortex.

Conscious Processing and the Global Neuronal Workspace Hypothesis

We review the central tenets and neuroanatomical basis of the global neuronal workspace (GNW) hypothesis, which attempts to account for the main scientific observations regarding the elementary mechanisms of conscious processing in the human brain. The GNW hypothesis proposes that, in the conscious state, a non-linear network ignition associated with recurrent processing amplifies and sustains a neural representation, allowing the corresponding information to be globally accessed by local processors. We examine this hypothesis in light of recent data that contrast brain activity evoked by either conscious or non-conscious contents, as well as during conscious or non-conscious states, particularly general anesthesia. We also discuss the relationship between the intertwined concepts of conscious processing, attention, and working memory.

Dentate gyrus circuits for encoding, retrieval and discrimination of episodic memories

The dentate gyrus (DG) has a key role in hippocampal memory formation. Intriguingly, DG lesions impair many, but not all, hippocampus-dependent mnemonic functions, indicating that the rest of the hippocampus (CA1-CA3) can operate autonomously under certain conditions. An extensive body of theoretical work has proposed how the architectural elements and various cell types of the DG may underlie its function in cognition. Recent studies recorded and manipulated the activity of different neuron types in the DG during memory tasks and have provided exciting new insights into the mechanisms of DG computational processes, particularly for the encoding, retrieval and discrimination of similar memories. Here, we review these DG-dependent mnemonic functions in light of the new findings and explore mechanistic links between the cellular and network properties of, and the computations performed by, the DG.

Context-dependent representations of objects and space in the primate hippocampus during virtual navigation

The hippocampus is implicated in associative memory and spatial navigation. To investigate how these functions are mixed in the hippocampus, we recorded from single hippocampal neurons in macaque monkeys navigating a virtual maze during a foraging task and a context-object associative memory task. During both tasks, single neurons encoded information about spatial position; a linear classifier also decoded position. However, the population code for space did not generalize across tasks, particularly where stimuli relevant to the associative memory task appeared. Single-neuron and population-level analyses revealed that cross-task changes were due to selectivity for nonspatial features of the associative memory task when they were visually available (perceptual coding) and following their disappearance (mnemonic coding). Our results show that neurons in the primate hippocampus nonlinearly mix information about space and nonspatial elements of the environment in a task-dependent manner; this efficient code flexibly represents unique perceptual experiences and correspondent memories.

Modeling of Brain-Like Concept Coding with Adulthood Neurogenesis in the Dentate Gyrus

Mammalian brains respond to new concepts via a type of neural coding termed “concept coding.” During concept coding, the dentate gyrus (DG) plays a vital role in pattern separation and pattern integration of concepts because it is a brain region with substantial neurogenesis in adult mammals. Although concept coding properties of the brain have been extensively studied by experimental work, modeling of the process to guide both further experimental studies and applications such as natural language processing is scarce. To model brain-like concept coding, we built a spiking neural network inspired by adulthood neurogenesis in the DG. Our model suggests that neurogenesis may facilitate integration of closely related concepts and separation of less relevant concepts. Such pattern agrees with the previous experimental observations in classification tasks and place cells in the hippocampus. Therefore, our simulation provides insight for future experimental studies on the neural coding difference between perception and cognition. By presenting 14 contexts each containing 4 concepts to the network, we found that neural responses of the DG changed dynamically as the context repetition time increased and were eventually consistent with the category organization of humans. Thus, our work provides a new framework of word representation for the construction of brain-like knowledge map.

High-Dimensional Brain: A Tool for Encoding and Rapid Learning of Memories by Single Neurons

Codifying memories is one of the fundamental problems of modern Neuroscience. The functional mechanisms behind this phenomenon remain largely unknown. Experimental evidence suggests that some of the memory functions are performed by stratified brain structures such as the hippocampus. In this particular case, single neurons in the CA1 region receive a highly multidimensional input from the CA3 area, which is a hub for information processing. We thus assess the implication of the abundance of neuronal signalling routes converging onto single cells on the information processing. We show that single neurons can selectively detect and learn arbitrary information items, given that they operate in high dimensions. The argument is based on stochastic separation theorems and the concentration of measure phenomena. We demonstrate that a simple enough functional neuronal model is capable of explaining: (i) the extreme selectivity of single neurons to the information content, (ii) simultaneous separation of several uncorrelated stimuli or informational items from a large set, and (iii) dynamic learning of new items by associating them with already "known" ones. These results constitute a basis for organization of complex memories in ensembles of single neurons. Moreover, they show that no a priori assumptions on the structural organization of neuronal ensembles are necessary for explaining basic concepts of static and dynamic memories.

The Role of Hierarchical Dynamical Functions in Coding for Episodic Memory and Cognition

Behavioral research in human verbal memory function led to the initial definition of episodic memory and semantic memory. A complete model of the neural mechanisms of episodic memory must include the capacity to encode and mentally reconstruct everything that humans can recall from their experience. This article proposes new model features necessary to address the complexity of episodic memory encoding and recall in the context of broader cognition and the functional properties of neurons that could contribute to this broader scope of memory. Many episodic memory models represent individual snapshots of the world with a sequence of vectors, but a full model must represent complex functions encoding and retrieving the relations between multiple stimulus features across space and time on multiple hierarchical scales. Episodic memory involves not only the space and time of an agent experiencing events within an episode but also features shown in neurophysiological data such as coding of speed, direction, boundaries, and objects. Episodic memory includes not only a spatio-temporal trajectory of a single agent but also segments of spatio-temporal trajectories for other agents and objects encountered in the environment consistent with data on encoding the position and angle of sensory features of objects and boundaries. We will discuss potential interactions of episodic memory circuits in the hippocampus and entorhinal cortex with distributed neocortical circuits that must represent all features of human cognition.

Testing Models of Human Declarative Memory at the Single-Neuron Level

Deciphering the mechanisms of declarative memory is a major goal of neuroscience. While much theoretical progress has been made, it has proven difficult to experimentally verify key predictions of some foundational models of memory. Recently, single-neuron recordings in human patients have started to provide direct experimental verification of some theories, including mnemonic evidence accumulation, balance-of-evidence for confidence judgments, sparse coding, contextual reinstatement, and the ventral tegmental area (VTA)-hippocampus loop model. Here, we summarize the cell types that have been described in the medial temporal lobe and posterior parietal cortex, discuss their properties, and reflect on how these findings inform theoretical work. This body of work exemplifies the scientific power of a synergistic combination of modeling and human single-neuron recordings to advance cognitive neuroscience.

Encoding of long-term associations through neural unitization in the human medial temporal lobe

Besides decades of research showing the role of the medial temporal lobe (MTL) in memory and the encoding of associations, the neural substrates underlying these functions remain unknown. We identified single neurons in the human MTL that responded to multiple and, in most cases, associated stimuli. We observed that most of these neurons exhibit no differences in their spike and local field potential (LFP) activity associated with the individual response-eliciting stimuli. In addition, LFP responses in the theta band preceded single neuron responses by ~70 ms, with the single trial phase providing fine tuning of the spike response onset. We postulate that the finding of similar neuronal responses to associated items provides a simple and flexible way of encoding memories in the human MTL, increasing the effective capacity for memory storage and successful retrieval.

In this work, the authors recorded single neurons and field potentials from the human medial temporal lobe (MTL) and show indistinguishable responses to associated stimuli. This coding mechanism provides a simple and flexible way of encoding memories in the human MTL.

Respiration Modulates Olfactory Memory Consolidation in Humans

In mammals respiratory-locked hippocampal rhythms are implicated in the scaffolding and transfer of information between sensory and memory networks. These oscillations are entrained by nasal respiration and driven by the olfactory bulb. They then travel to the piriform cortex where they propagate further downstream to the hippocampus and modulate neural processes critical for memory formation. In humans, bypassing nasal airflow through mouth-breathing abolishes these rhythms and impacts encoding as well as recognition processes thereby reducing memory performance. It has been hypothesized that similar behavior should be observed for the consolidation process, the stage between encoding and recognition, were memory is reactivated and strengthened. However, direct evidence for such an effect is lacking in human and nonhuman animals. Here we tested this hypothesis by examining the effect of respiration on consolidation of episodic odor memory. In two separate sessions, female and male participants encoded odors followed by a 1 h awake resting consolidation phase where they either breathed solely through their nose or mouth. Immediately after the consolidation phase, memory for odors was tested. Recognition memory significantly increased during nasal respiration compared with mouth respiration during consolidation. These results provide the first evidence that respiration directly impacts consolidation of episodic events, and lends further support to the notion that core cognitive functions are modulated by the respiratory cycle.SIGNIFICANCE STATEMENT Memories pass through three main stages in their development: encoding, consolidation, and retrieval. Growing evidence from animal and human studies suggests that respiration plays an important role in the behavioral and neural mechanisms associated with encoding and recognition. Specifically nasal, but not mouth, respiration entrains neural oscillations that enhance encoding and recognition processes. We demonstrate that respiration also affects the consolidation stage. Breathing through the nose compared with the mouth during consolidation enhances recognition memory. This demonstrates, first, that nasal respiration is important during the critical period were memories are reactivated and strengthened. Second, it suggests that the neural mechanisms responsible may emerge from nasal respiration.

The unreasonable effectiveness of small neural ensembles in high-dimensional brain

Complexity is an indisputable, well-known, and broadly accepted feature of the brain. Despite the apparently obvious and widely-spread consensus on the brain complexity, sprouts of the single neuron revolution emerged in neuroscience in the 1970s. They brought many unexpected discoveries, including grandmother or concept cells and sparse coding of information in the brain. In machine learning for a long time, the famous curse of dimensionality seemed to be an unsolvable problem. Nevertheless, the idea of the blessing of dimensionality becomes gradually more and more popular. Ensembles of non-interacting or weakly interacting simple units prove to be an effective tool for solving essentially multidimensional and apparently incomprehensible problems. This approach is especially useful for one-shot (non-iterative) correction of errors in large legacy artificial intelligence systems and when the complete re-training is impossible or too expensive. These simplicity revolutions in the era of complexity have deep fundamental reasons grounded in geometry of multidimensional data spaces. To explore and understand these reasons we revisit the background ideas of statistical physics. In the course of the 20th century they were developed into the concentration of measure theory. The Gibbs equivalence of ensembles with further generalizations shows that the data in high-dimensional spaces are concentrated near shells of smaller dimension. New stochastic separation theorems reveal the fine structure of the data clouds. We review and analyse biological, physical, and mathematical problems at the core of the fundamental question: how can high-dimensional brain organise reliable and fast learning in high-dimensional world of data by simple tools? To meet this challenge, we outline and setup a framework based on statistical physics of data. Two critical applications are reviewed to exemplify the approach: one-shot correction of errors in intellectual systems and emergence of static and associative memories in ensembles of single neurons. Error correctors should be simple; not damage the existing skills of the system; allow fast non-iterative learning and correction of new mistakes without destroying the previous fixes. All these demands can be satisfied by new tools based on the concentration of measure phenomena and stochastic separation theory. We show how a simple enough functional neuronal model is capable of explaining: i) the extreme selectivity of single neurons to the information content of high-dimensional data, ii) simultaneous separation of several uncorrelated informational items from a large set of stimuli, and iii) dynamic learning of new items by associating them with already "known" ones. These results constitute a basis for organisation of complex memories in ensembles of single neurons.

The storage and recall of memories in the hippocampo-cortical system

A quantitative computational theory of the operation of the hippocampus as an episodic memory system is described. The CA3 system operates as a single attractor or autoassociation network (1) to enable rapid one-trial associations between any spatial location (place in rodents or spatial view in primates) and an object or reward and (2) to provide for completion of the whole memory during recall from any part. The theory is extended to associations between time and object or reward to implement temporal order memory, which is also important in episodic memory. The dentate gyrus performs pattern separation by competitive learning to create sparse representations producing, for example, neurons with place-like fields from entorhinal cortex grid cells. The dentate granule cells generate, by the very small number of mossy fibre connections to CA3, a randomizing pattern separation effect that is important during learning but not recall and that separates out the patterns represented by CA3 firing as being very different from each other. This is optimal for an unstructured episodic memory system in which each memory must be kept distinct from other memories. The direct perforant path input to CA3 is quantitatively appropriate for providing the cue for recall in CA3 but not for learning. The CA1 recodes information from CA3 to set up associatively learned backprojections to the neocortex to allow the subsequent retrieval of information to the neocortex, giving a quantitative account of the large number of hippocampo-neocortical and neocortical-neocortical backprojections. Tests of the theory including hippocampal subregion analyses and hippocampal NMDA receptor knockouts are described and support the theory.

Knowledge Transfer Between Artificial Intelligence Systems

We consider the fundamental question: how a legacy "student" Artificial Intelligent (AI) system could learn from a legacy "teacher" AI system or a human expert without re-training and, most importantly, without requiring significant computational resources. Here "learning" is broadly understood as an ability of one system to mimic responses of the other to an incoming stimulation and vice-versa. We call such learning an Artificial Intelligence knowledge transfer. We show that if internal variables of the "student" Artificial Intelligent system have the structure of an n-dimensional topological vector space and n is sufficiently high then, with probability close to one, the required knowledge transfer can be implemented by simple cascades of linear functionals. In particular, for n sufficiently large, with probability close to one, the "student" system can successfully and non-iteratively learn k ≪ n new examples from the "teacher" (or correct the same number of mistakes) at the cost of two additional inner products. The concept is illustrated with an example of knowledge transfer from one pre-trained convolutional neural network to another.

Synaptic plasticity in human cortical circuits: cellular mechanisms of learning and memory in the human brain?

Synaptic plasticity is the cellular basis of learning and memory, but to what extent this holds for the adult human brain is not known. To study synaptic plasticity in human neuronal circuits poses a huge challenge, since live human neurons and synapses are not readily accessible. Despite this, various lines of research have provided insights in properties of adult human synapses and their plasticity both in vitro and in vivo, with some unexpected surprises. We first discuss the experimental approaches to study activity-dependent plasticity of adult human synapses, and then highlight rules and mechanisms of Hebbian spike timing-dependent plasticity (STDP) found in these synapses. Finally, we conclude with thoughts on how these synaptic principles can underlie human learning and memory.

Blessing of dimensionality: mathematical foundations of the statistical physics of data

The concentrations of measure phenomena were discovered as the mathematical background to statistical mechanics at the end of the nineteenth/beginning of the twentieth century and have been explored in mathematics ever since. At the beginning of the twenty-first century, it became clear that the proper utilization of these phenomena in machine learning might transform the curse of dimensionality into the blessing of dimensionality This paper summarizes recently discovered phenomena of measure concentration which drastically simplify some machine learning problems in high dimension, and allow us to correct legacy artificial intelligence systems. The classical concentration of measure theorems state that i.i.d. random points are concentrated in a thin layer near a surface (a sphere or equators of a sphere, an average or median-level set of energy or another Lipschitz function, etc.). The new stochastic separation theorems describe the thin structure of these thin layers: the random points are not only concentrated in a thin layer but are all linearly separable from the rest of the set, even for exponentially large random sets. The linear functionals for separation of points can be selected in the form of the linear Fisher's discriminant. All artificial intelligence systems make errors. Non-destructive correction requires separation of the situations (samples) with errors from the samples corresponding to correct behaviour by a simple and robust classifier. The stochastic separation theorems provide us with such classifiers and determine a non-iterative (one-shot) procedure for their construction.This article is part of the theme issue 'Hilbert's sixth problem'.

The Sync/deSync Model: How a Synchronized Hippocampus and a Desynchronized Neocortex Code Memories

Neural oscillations are important for memory formation in the brain. The desynchronization of alpha (10 Hz) oscillations in the neocortex has been shown to predict successful memory encoding and retrieval. However, when engaging in learning, it has been found that the hippocampus synchronizes in theta (4 Hz) oscillations, and that learning is dependent on the phase of theta. This inconsistency as to whether synchronization is "good" for memory formation leads to confusion over which oscillations we should expect to see and where during learning paradigm experiments. This paper seeks to respond to this inconsistency by presenting a neural network model of how a well functioning learning system could exhibit both of these phenomena, i.e., desynchronization of alpha and synchronization of theta during successful memory encoding.We present a spiking neural network (the Sync/deSync model) of the neocortical and hippocampal system. The simulated hippocampus learns through an adapted spike-time dependent plasticity rule, in which weight change is modulated by the phase of an extrinsically generated theta oscillation. Additionally, a global passive weight decay is incorporated, which is also modulated by theta phase. In this way, the Sync/deSync model exhibits theta phase-dependent long-term potentiation and long-term depression. We simulated a learning paradigm experiment and compared the oscillatory dynamics of our model with those observed in single-cell and scalp-EEG studies of the medial temporal lobe. Our Sync/deSync model suggests that both the desynchronization of neocortical alpha and the synchronization of hippocampal theta are necessary for successful memory encoding and retrieval.SIGNIFICANCE STATEMENT A fundamental question is the role of rhythmic activation of neurons, i.e., how and why their firing oscillates between high and low rates. A particularly important question is how oscillatory dynamics between the neocortex and hippocampus support memory formation. We present a spiking neural-network model of such memory formation, with the central ideas that (1) in neocortex, neurons need to break out of an alpha oscillation to represent a stimulus (i.e., alpha desynchronizes), whereas (2) in hippocampus, the firing of neurons at theta facilitates formation of memories (i.e., theta synchronizes). Accordingly, successful memory formation is marked by reduced neocortical alpha and increased hippocampal theta. This pattern has been observed experimentally and gives our model its name-the Sync/deSync model.

Dataset of human medial temporal lobe single neuron activity during declarative memory encoding and recognition

We present a dataset of 1,576 single neurons recorded from the human amygdala and hippocampus in 65 sessions from 42 patients undergoing intracranial monitoring for localization of epileptic seizures. Subjects performed a recognition memory task with pictures as stimuli. Subjects were asked to identify whether they had seen a particular image the first time ('new') or second time ('old') on a 1-6 confidence scale. This comprehensive dataset includes the spike times of all neurons and their extracellular waveforms, behavior, electrode locations determined from post-operative MRI scans, demographics, and the stimuli shown. As technical validation, we provide spike sorting quality metrics and assessment of tuning of cells to verify the presence of visually-and memory selective cells. We also provide analysis code that reproduces key scientific findings published previously on a smaller version of this dataset. Together, this large dataset will facilitate the investigation of the neural mechanism of declarative memory by providing a substantial number of hard to obtain human single-neuron recordings during a well characterized behavioral task.

Coding of episodic memory in the human hippocampus

Neurocomputational models have long posited that episodic memories in the human hippocampus are represented by sparse, stimulus-specific neural codes. A concomitant proposal is that when sparse-distributed neural assemblies become active, they suppress the activity of competing neurons (neural sharpening). We investigated episodic memory coding in the hippocampus and amygdala by measuring single-neuron responses from 20 epilepsy patients (12 female) undergoing intracranial monitoring while they completed a continuous recognition memory task. In the left hippocampus, the distribution of single-neuron activity indicated that only a small fraction of neurons exhibited strong responding to a given repeated word and that each repeated word elicited strong responding in a different small fraction of neurons. This finding reflects sparse distributed coding. The remaining large fraction of neurons exhibited a concurrent reduction in firing rates relative to novel words. The observed pattern accords with longstanding predictions that have previously received scant support from single-cell recordings from human hippocampus.

Cognitive representations of spatial location

Spatial memory has fascinated psychologists ever since the discipline began, but a series of findings beginning in the middle of last century propelled its study into the domain of neuroscience and helped bring about the cognitive revolution in psychology. Starting with the discovery that the hippocampus plays a central role in memory, particularly spatial memory, studies of the mammalian hippocampus and related regions over the latter half of the century slowly uncovered an extensive neural system involved in processing place, head direction, objects, speed and other spatially informative parameters. Meanwhile, the concurrent discovery of hippocampal synaptic plasticity allowed theoreticians and experimentalists to collaborate in linking spatial perception and memory, and genetic techniques developed towards the end of the century opened the door to circuit dissections of these processes. Building on these discoveries, spatial cognition and episodic memory may be the first cognitive competences understood across all levels from molecules to behaviour.

An online adaptive screening procedure for selective neuronal responses

BACKGROUND

A common problem in neurophysiology is to identify stimuli that elicit neuronal responses in a given brain region. Particularly in situations where electrode positions are fixed, this can be a time-consuming task that requires presentation of a large number of stimuli. Such a screening for response-eliciting stimuli is employed, e.g., as a standard procedure to identify 'concept cells' in the human medial temporal lobe.

NEW METHOD

Our new method evaluates neuronal responses to stimuli online during a screening session, which allows us to successively exclude stimuli that do not evoke a response. Using this method, we can screen a larger number of stimuli which in turn increases the chances of finding responsive neurons and renders time-consuming offline analysis unnecessary.

RESULTS

Our method enabled us to present 30% more stimuli in the same period of time with additional presentations of the most promising candidate stimuli. Our online method ran smoothly on a standard computer and network.

COMPARISON WITH AN EXISTING METHOD

To analyze how our online screening procedure performs in comparison to an established offline method, we used the Wave_Clus software package. We did not observe any major drawbacks in our method, but a much higher efficiency and analysis speed.

CONCLUSIONS

By transitioning from a traditional offline screening procedure to our new online method, we substantially increased the number of visual stimuli presented in a given time period. This allows to identify more response-eliciting stimuli, which forms the basis to better address a great number of questions in cognitive neuroscience.

Latching dynamics in neural networks with synaptic depression

Prediction is the ability of the brain to quickly activate a target concept in response to a related stimulus (prime). Experiments point to the existence of an overlap between the populations of the neurons coding for different stimuli, and other experiments show that prime-target relations arise in the process of long term memory formation. The classical modelling paradigm is that long term memories correspond to stable steady states of a Hopfield network with Hebbian connectivity. Experiments show that short term synaptic depression plays an important role in the processing of memories. This leads naturally to a computational model of priming, called latching dynamics; a stable state (prime) can become unstable and the system may converge to another transiently stable steady state (target). Hopfield network models of latching dynamics have been studied by means of numerical simulation, however the conditions for the existence of this dynamics have not been elucidated. In this work we use a combination of analytic and numerical approaches to confirm that latching dynamics can exist in the context of a symmetric Hebbian learning rule, however lacks robustness and imposes a number of biologically unrealistic restrictions on the model. In particular our work shows that the symmetry of the Hebbian rule is not an obstruction to the existence of latching dynamics, however fine tuning of the parameters of the model is needed.

Stochastic separation theorems

The problem of non-iterative one-shot and non-destructive correction of unavoidable mistakes arises in all Artificial Intelligence applications in the real world. Its solution requires robust separation of samples with errors from samples where the system works properly. We demonstrate that in (moderately) high dimension this separation could be achieved with probability close to one by linear discriminants. Based on fundamental properties of measure concentration, we show that for M<aexp(bn) random M-element sets in Rⁿ are linearly separable with probability p, p>1-ϑ, where 1>ϑ>0 is a given small constant. Exact values of a,b>0 depend on the probability distribution that determines how the random M-element sets are drawn, and on the constant ϑ. These stochastic separation theorems provide a new instrument for the development, analysis, and assessment of machine learning methods and algorithms in high dimension. Theoretical statements are illustrated with numerical examples.

Plasticity of hippocampal memories in humans

The human hippocampus is a brain region that supports episodic and spatial memory. Recent experiments have drawn on animal research and computational modelling to reveal how the unique computations and representations of the hippocampus support episodic and spatial memory. Invasive electrophysiological recordings and non-invasive functional brain imaging have provided evidence for the rapid formation of hippocampal representations, as well as the ability of the hippocampus to both pattern-separate and pattern-complete input from the neocortex. Further, recent evidence has shown that hippocampal representations are in constant flux, undergoing a continual process of strengthening, weakening and altering. This research offers a glimpse into the highly plastic and flexible nature of the human hippocampal system in relation to episodic memory.

Sleep Enhances Recognition Memory for Conspecifics as Bound into Spatial Context

Social memory refers to the fundamental ability of social species to recognize their conspecifics in quite different contexts. Sleep has been shown to benefit consolidation, especially of hippocampus-dependent episodic memory whereas effects of sleep on social memory are less well studied. Here, we examined the effect of sleep on memory for conspecifics in rats. To discriminate interactions between the consolidation of social memory and of spatial context during sleep, adult Long Evans rats performed on a social discrimination task in a radial arm maze. The Learning phase comprised three 10-min sampling sessions in which the rats explored a juvenile rat presented at a different arm of the maze in each session. Then the rats were allowed to sleep (n = 18) or stayed awake (n = 18) for 120 min. During the following 10-min Test phase, the familiar juvenile rat (of the Learning phase) was presented along with a novel juvenile rat, each rat at an opposite arm of the maze. Significant social recognition memory, as indicated by preferential exploration of the novel over the familiar conspecific, occurred only after post-learning sleep, but not after wakefulness. Sleep, compared with wakefulness, significantly enhanced social recognition during the first minute of the Test phase. However, memory expression depended on the spatial configuration: Significant social recognition memory emerged only after sleep when the rat encountered the novel conspecific at a place different from that of the familiar juvenile in the last sampling session before sleep. Though unspecific retrieval-related effects cannot entirely be excluded, our findings suggest that sleep, rather than independently enhancing social and spatial aspects of memory, consolidates social memory by acting on an episodic representation that binds the memory of the conspecific together with the spatial context in which it was recently encountered.

Scene-selective coding by single neurons in the human parahippocampal cortex

Imaging, electrophysiological, and lesion studies have shown a relationship between the parahippocampal cortex (PHC) and the processing of spatial scenes. Our present knowledge of PHC, however, is restricted to the macroscopic properties and dynamics of bulk tissue; the behavior and selectivity of single parahippocampal neurons remains largely unknown. In this study, we analyzed responses from 630 parahippocampal neurons in 24 neurosurgical patients during visual stimulus presentation. We found a spatially clustered subpopulation of scene-selective units with an associated event-related field potential. These units form a population code that is more distributed for scenes than for other stimulus categories, and less sparse than elsewhere in the medial temporal lobe. Our electrophysiological findings provide insight into how individual units give rise to the population response observed with functional imaging in the parahippocampal place area.

Memory: Organization and Control

A major goal of memory research is to understand how cognitive processes in memory are supported at the level of brain systems and network representations. Especially promising in this direction are new findings in humans and animals that converge in indicating a key role for the hippocampus in the systematic organization of memories. New findings also indicate that the prefrontal cortex may play an equally important role in the active control of memory organization during both encoding and retrieval. Observations about the dialog between the hippocampus and prefrontal cortex provide new insights into the operation of the larger brain system that serves memory.

Long-term coding of personal and universal associations underlying the memory web in the human brain

Neurons in the medial temporal lobe (MTL), a critical area for declarative memory, have been shown to change their tuning in associative learning tasks. Yet, it is unclear how durable these neuronal representations are and if they outlast the execution of the task. To address this issue, we studied the responses of MTL neurons in neurosurgical patients to known concepts (people and places). Using association scores provided by the patients and a web-based metric, here we show that whenever MTL neurons respond to more than one concept, these concepts are typically related. Furthermore, the degree of association between concepts could be successfully predicted based on the neurons' response patterns. These results provide evidence for a long-term involvement of MTL neurons in the representation of durable associations, a hallmark of human declarative memory.

Neurons in the medial temporal lobe change their firing patterns as people learn to pair items together, yet it is unclear if this pairing lasts. Here, authors find that single medial temporal lobe neurons in humans tend to respond similarly to items that are closely conceptually related.

Responses of neurons in the medial temporal lobe during encoding and recognition of face-scene pairs

Associations between co-occurring stimuli are formed in the medial temporal lobe (MTL). Here, we recorded from 508 single and multi-units in the MTL while participants learned and retrieved associations between unfamiliar faces and unfamiliar scenes. Participant's memories for the face-scene pairs were later tested using cued recall and recognition tests. The results show that neurons in the parahippocampal cortex are most likely to respond with changes from baseline firing to these stimuli during both encoding and recognition, and this region showed the greatest proportion of cells showing differential responses depending on the phase of the task. Furthermore, we found that cells in the parahippocampal cortex that responded during both encoding and recognition were more likely to show decreases from baseline firing than cells that were only recruited during recognition, which were more likely to show increases in firing. Since all stimuli shown during recognition were familiar to the patients, these findings suggest that with familiarity, cell responses become more sharply tuned. No neurons in this region, however, were found to be affected by recombining face/scene pairs. Neurons in other MTL regions, particularly the hippocampus, were sensitive to stimulus configurations. Thus, the results support the idea that neurons in the parahippocampal cortex code for features of stimuli and neurons in the hippocampus are more likely to represent their specific configurations.

Multivoxel Object Representations in Adult Human Visual Cortex Are Flexible: An Associative Learning Study

Learning associations between co-occurring events enables us to extract structure from our environment. Medial-temporal lobe structures are critical for associative learning. However, the role of the ventral visual pathway (VVP) in associative learning is not clear. Do multivoxel object representations in the VVP reflect newly formed associations? We show that VVP multivoxel representations become more similar to each other after human participants learn arbitrary new associations between pairs of unrelated objects (faces, houses, cars, chairs). Participants were scanned before and after 15 days of associative learning. To evaluate how object representations changed, a classifier was trained on discriminating two nonassociated categories (e.g., faces/houses) and tested on discriminating their paired associates (e.g., cars/chairs). Because the associations were arbitrary and counterbalanced across participants, there was initially no particular reason for this cross-classification decision to tend toward either alternative. Nonetheless, after learning, cross-classification performance increased in the VVP (but not hippocampus), on average by 3.3%, with some voxels showing increases of up to 10%. For example, a chair multivoxel representation that initially resembled neither face nor house representations was, after learning, classified as more similar to that of faces for participants who associated chairs with faces and to that of houses for participants who associated chairs with houses. Additionally, learning produced long-lasting perceptual consequences. In a behavioral priming experiment performed several months later, the change in cross-classification performance was correlated with the degree of priming. Thus, VVP multivoxel representations are not static but become more similar to each other after associative learning.

Memory and Space: Towards an Understanding of the Cognitive Map

More than 50 years of research have led to the general agreement that the hippocampus contributes to memory, but there has been a major schism among theories of hippocampal function over this time. Some researchers argue that the hippocampus plays a broad role in episodic and declarative memory, whereas others argue for a specific role in the creation of spatial cognitive maps and navigation. Although both views have merit, neither provides a complete account of hippocampal function. Guided by recent reviews that attempt to bridge between these views, here we suggest that reconciliation can be accomplished by exploring hippocampal function from the perspective of Tolman's (1948) original conception of a cognitive map as organizing experience and guiding behavior across all domains of cognition. We emphasize recent studies in animals and humans showing that hippocampal networks support a broad range of domains of cognitive maps, that these networks organize specific experiences within the contextually relevant map, and that network activity patterns reflect behavior guided through cognitive maps. These results are consistent with a framework that bridges theories of hippocampal function by conceptualizing the hippocampus as organizing incoming information within the context of a multidimensional cognitive map of spatial, temporal, and associational context.

SIGNIFICANCE STATEMENT

Research of hippocampal function is dominated by two major views. The spatial view argues that the hippocampus tracks routes through space, whereas the memory view suggests a broad role in declarative memory. Both views rely on considerable evidence, but neither provides a complete account of hippocampal function. Here we review evidence that, in addition to spatial context, the hippocampus encodes a wide variety of information about temporal and situational context, about the systematic organization of events in abstract space, and about routes through maps of cognition and space. We argue that these findings cross the boundaries of the memory and spatial views and offer new insights into hippocampal function as a system supporting a broad range of cognitive maps.

Neuronal codes for visual perception and memory

In this review, I describe and contrast the representation of stimuli in visual cortical areas and in the medial temporal lobe (MTL). While cortex is characterized by a distributed and implicit coding that is optimal for recognition and storage of semantic information, the MTL shows a much sparser and explicit coding of specific concepts that is ideal for episodic memory. I will describe the main characteristics of the coding in the MTL by the so-called concept cells and will then propose a model of the formation and recall of episodic memory based on partially overlapping assemblies.

Learning of anticipatory responses in single neurons of the human medial temporal lobe

Neuronal processes underlying the formation of new associations in the human brain are not yet well understood. Here human participants, implanted with depth electrodes in the brain, learned arbitrary associations between images presented in an ordered, predictable sequence. During learning we recorded from medial temporal lobe (MTL) neurons that responded to at least one of the pictures in the sequence (the preferred stimulus). We report that as a result of learning, single MTL neurons show asymmetric shifts in activity and start firing earlier in the sequence in anticipation of their preferred stimulus. These effects appear relatively early in learning, after only 11 exposures to the stimulus sequence. The anticipatory neuronal responses emerge while the subjects became faster in reporting the next item in the sequence. These results demonstrate flexible representations that could support learning of new associations between stimuli in a sequence, in single neurons in the human MTL.