Learning and recall in a dynamic theory of coordination patterns

A dynamic theory of learning and recall of coordination patterns is developed in the context of relative timing skills. Characterizing the coordination patterns in such skills by the collective variable, relative phase, we choose a model system in which the intrinsic pattern dynamics as well as the influence of environmental and memorized information are well understood from previous experimental and theoretical work. To describe learning we endow memorized information with dynamics which is determined by a phenomenological strategy. Similarly, additional degrees of freedom must be introduced to understand recall. As such recall variables we choose the relative strengths with which each memorized pattern acts on the pattern dynamics and model their dynamics phenomenologically. The resulting dynamical system that resembles models used in pattern recognition theory is shown to adequately describe the learning and recall processes. Moreover, due to the operational character of the theory, several predictions emerge that are open to experimental test. In particular, we show under which conditions phase transitions occur in the dynamics of the coordination patterns during learning and during recall. Considering different time scales and their relations we demonstrate how these phase transitions can be identified and observed. Other predictions include the influence of the intrinsic pattern dynamics on the recall process and the existence of history and hysteresis effects in recall. We discuss different forms of forgetting and differentiation of memorized information. The results show how a new theoretical view of learning and recall as change of behavioral dynamics can lead to a different understanding of these processes by providing testable predictions.     

A Computational Predictor of Human Episodic Memory Based on a Theta Phase Precession Network

In the rodent hippocampus, a phase precession phenomena of place cell firing with the local field potential (LFP) theta is called “theta phase precession” and is considered to contribute to memory formation with spike time dependent plasticity (STDP). On the other hand, in the primate hippocampus, the existence of theta phase precession is unclear. Our computational studies have demonstrated that theta phase precession dynamics could contribute to primate–hippocampal dependent memory formation, such as object–place association memory. In this paper, we evaluate human theta phase precession by using a theory–experiment combined analysis. Human memory recall of object–place associations was analyzed by an individual hippocampal network simulated by theta phase precession dynamics of human eye movement and EEG data during memory encoding. It was found that the computational recall of the resultant network is significantly correlated with human memory recall performance, while other computational predictors without theta phase precession are not significantly correlated with subsequent memory recall. Moreover the correlation is larger than the correlation between human recall and traditional experimental predictors. These results indicate that theta phase precession dynamics are necessary for the better prediction of human recall performance with eye movement and EEG data. In this analysis, theta phase precession dynamics appear useful for the extraction of memory-dependent components from the spatio–temporal pattern of eye movement and EEG data as an associative network. Theta phase precession may be a common neural dynamic between rodents and humans for the formation of environmental memories.     

Network Evolution Induced by Asynchronous Stimuli through Spike-Timing-Dependent Plasticity

In sensory neural system, external asynchronous stimuli play an important role in perceptual learning, associative memory and map development. However, the organization of structure and dynamics of neural networks induced by external asynchronous stimuli are not well understood. Spike-timing-dependent plasticity (STDP) is a typical synaptic plasticity that has been extensively found in the sensory systems and that has received much theoretical attention. This synaptic plasticity is highly sensitive to correlations between pre- and postsynaptic firings. Thus, STDP is expected to play an important role in response to external asynchronous stimuli, which can induce segregative pre- and postsynaptic firings. In this paper, we study the impact of external asynchronous stimuli on the organization of structure and dynamics of neural networks through STDP. We construct a two-dimensional spatial neural network model with local connectivity and sparseness, and use external currents to stimulate alternately on different spatial layers. The adopted external currents imposed alternately on spatial layers can be here regarded as external asynchronous stimuli. Through extensive numerical simulations, we focus on the effects of stimulus number and inter-stimulus timing on synaptic connecting weights and the property of propagation dynamics in the resulting network structure. Interestingly, the resulting feedforward structure induced by stimulus-dependent asynchronous firings and its propagation dynamics reflect both the underlying property of STDP. The results imply a possible important role of STDP in generating feedforward structure and collective propagation activity required for experience-dependent map plasticity in developing in vivo sensory pathways and cortices. The relevance of the results to cue-triggered recall of learned temporal sequences, an important cognitive function, is briefly discussed as well. Furthermore, this finding suggests a potential application for examining STDP by measuring neural population activity in a cultured neural network.     

Spatial and temporal coding in single neurons

Convergence between cells which differ in both spatial and temporal properties create higher order neurons with response properties that are distinctly different from those of the input neurons. The spatial properties of target neurons are not necessarily cosinetuned. In addition, unlike the independence between spatial and temporal properties in cosine-tuned afferent neurons, higher-order target cells generally exhibit a dependence of temporal dynamics on spatial properties. The response properties of target neurons receiving spatio-temporal convergence (STC) from tonic and phasic-tonic or phasic afferents is investigated here by considering a general case where the dynamic input is represented by a fractional, leaky, derivative transfer function. It is shown that, at frequencies below the corner frequency of the dynamic input, the temporal properties of target neurons can be described by leaky differentiators having time constants that are a function of spatial direction. Thus, STC target neurons exhibit tonic temporal response properties during stimulation along some spatial directions (having small time constants) and phasic properties along other directions (having large time constants). Specifically, target neurons encode the complete derivative of the stimulus along certain spatial directions. Thus, STC acts as a directionally specific high-pass filter and produces complete derivatives from fractional, leaky derivative afferent signals. In addition, spatio-temporal transformations can generate novel temporal dynamics in the central nervous system. These observations suggest that spatio-temporal computations might constitute an alternative to parallel, independent spatial and temporal channels.     

An oscillatory neural network model that demonstrates the benefits of multisensory learning

Since the world consists of objects that stimulate multiple senses, it is advantageous for a vertebrate to integrate all the sensory information available. However, the precise mechanisms governing the temporal dynamics of multisensory processing are not well understood. We develop a computational modeling approach to investigate these mechanisms. We present an oscillatory neural network model for multisensory learning based on sparse spatio-temporal encoding. Recently published results in cognitive science show that multisensory integration produces greater and more efficient learning. We apply our computational model to qualitatively replicate these results. We vary learning protocols and system dynamics, and measure the rate at which our model learns to distinguish superposed presentations of multisensory objects. We show that the use of multiple channels accelerates learning and recall by up to 80%. When a sensory channel becomes disabled, the performance degradation is less than that experienced during the presentation of non-congruent stimuli. This research furthers our understanding of fundamental brain processes, paving the way for multiple advances including the building of machines with more human-like capabilities.     

Place Cells,Grid Cells,and Memory

The hippocampal system is critical for storage and retrieval of declarative memories, including memories for locations and events that take place at those locations. Spatial memories place high demands on capacity. Memories must be distinct to be recalled without interference and encoding must be fast. Recent studies have indicated that hippocampal networks allow for fast storage of large quantities of uncorrelated spatial information. The aim of the this article is to review and discuss some of this work, taking as a starting point the discovery of multiple functionally specialized cell types of the hippocampal–entorhinal circuit, such as place, grid, and border cells. We will show that grid cells provide the hippocampus with a metric, as well as a putative mechanism for decorrelation of representations, that the formation of environment-specific place maps depends on mechanisms for long-term plasticity in the hippocampus, and that long-term spatiotemporal memory storage may depend on offline consolidation processes related to sharp-wave ripple activity in the hippocampus. The multitude of representations generated through interactions between a variety of functionally specialized cell types in the entorhinal–hippocampal circuit may be at the heart of the mechanism for declarative memory formation.The scientific study of human memory started with Herman Ebbinghaus, who initiated the quantitative investigation of associative memory processes as they take place (Ebbinghaus 1885). Ebbinghaus described the conditions that influence memory formation and he determined several basic principles of encoding and recall, such as the law of frequency and the effect of time on forgetting. With Ebbinghaus, higher mental functions were brought to the laboratory. In parallel with the human learning tradition that Ebbinghaus started, a new generation of experimental psychologists described the laws of associative learning in animals. With behaviorists like Pavlov, Watson, Hull, Skinner, and Tolman, a rigorous program for identifying the laws of animal learning was initiated. By the middle of the 20th century, a language for associative learning processes had been developed, and many of the fundamental relationships between environment and behavior had been described. What was completely missing, though, was an understanding of the neural activity underlying the formation of the memory. The behaviorists had deliberately shied away from physiological explanations because of the intangible nature of neural activity at that time.Then the climate began to change. Karl Lashley had shown that lesions in the cerebral cortex had predictable effects on behavior in animals (Lashley 1929, 1950), and Donald Hebb introduced concepts and ideas to account for complex brain functions at the neural circuit level, many of which have retained a place in modern neuroscience (Hebb 1949). Both Lashley and Hebb searched for the engram, but they found no specific locus for it. A significant turning point was reached when Scoville and Milner (1957) reported severe loss of memory in an epileptic patient, patient H.M., after bilateral surgical removal of the hippocampal formation and the surrounding medial temporal lobe areas. “After operation this young man could no longer recognize the hospital staff nor find his way to the bathroom, and he seemed to recall nothing of the day-to-day events of his hospital life.” This tragic misfortune inspired decades of research on the function of the hippocampus in memory. H.M.’s memory impairment could be reproduced in memory tasks in animals and studies of H.M., as well as laboratory animals, pointed to a critical role for the hippocampus in declarative memory—memory, which, in humans, can be consciously recalled and declared, such as memories of experiences and facts (Milner et al. 1968; Mishkin 1978; Cohen and Squire 1980; Squire 1992; Corkin 2002). What was missing from these early studies, however, was a way to address the neuronal mechanisms that led information to be stored as memory.The aim of this article is to show how studies of hippocampal neuronal activity during the past few decades have brought us to a point at which a mechanistic basis of memory formation is beginning to surface. An early landmark in this series of investigations was the discovery of place cells, cells that fire selectively at one or few locations in the environment. At first, these cells seemed to be part of the animal’s instantaneous representation of location, independent of memory, but gradually, over the course of several decades, it has become clear that place cells express current as well as past and future locations. In many ways, place cells can be used as readouts of the memories that are stored in the hippocampus. More recent work has also shown that place cells are part of a wider network of spatially modulated neurons, including grid, border, and head direction cells, each with distinct roles in the representation of space and spatial memory. In this article, we shall discuss potential mechanisms by which these cell types, particularly place and grid cells, in conjunction with synaptic plasticity, may form the basis of a mammalian system for fast high-capacity declarative memory.     

δ13C values of some succulent plants from Madagascar

Summary ¹³C values were determined in 20 succulents from Madagascar. The values were indicative of Crassulacean Acid Metabolism in 10 species of the Didiereaceae, 4 species of the Euphorbiaceae, 2 species of the Crassulaceae and 1 species of the Cucurbitaceae. The Didiereaceae and Euphorbiaceae studied are major components of a high biomass xerophytic flora in the semi-arid southwest and south of Madagascar. Three species of the Euphorbiaceae with succulent stems and non-succulent leaves, which were cultivated outdoors in the Tananarive Botanic Garden, showed C₃ like ¹³C values for both leaves and stems. ¹³C values of leaf and stem material from a similar species, collected in the south of Madagascar, indicated Crassulacean Acid Metabolism.Abbreviations CAM Crassulacean Acid Metabolism     

Protein Kinase C in Rat Brain Is Altered by Developmental Lead Exposure

The absence of learning-related redistribution of hippocampal protein kinase C (PKC) has been correlated with impairment of learning performance induced by developmental lead (Pb) exposure. This study was designed to examine whether the properties of brain PKC are altered by chronic Pb exposure during development. Two-tenth percent Pb acetate was administered to pregnant and lactating dams and then administered to weanlings in drinking water until postnatal day (PN) 56. Effects of Pb on translocation of PKC were studied in brain slices prepared from hippocampus. When the slices were treated with 0.33 M phorbol-12, 13-dibutyrate (PDBu) for 15 min, a significant increase in PKC activity was observed in the membrane fraction of hippocampal slices from Pb-exposed rats, suggesting that chronic Pb exposure potentiates PDBu-activated PKC translocation. Data obtained from saturation binding assays in the frontal cortices of Pb-exposed rats showed a decrease in the dissociation constant (K_D) in both membrane and cytosolic PKC. A decrease in the total binding sites (B_max) of [³H]PDBu binding was only observed in membrane PKC. Furthermore, developmental Pb exposure decreased PKC-, but not PKC-, -II, and - in the membrane fraction of the hippocampus and the frontal cortex. These results indicate that chronic Pb exposure during development increases phorbol ester binding affinity, enhances phorbol ester-induced translocation of PKC, and down-regulates membrane PKC, mainly PKC-.     

Spatial learning and action planning in a prefrontal cortical network model

The interplay between hippocampus and prefrontal cortex (PFC) is fundamental to spatial cognition. Complementing hippocampal place coding, prefrontal representations provide more abstract and hierarchically organized memories suitable for decision making. We model a prefrontal network mediating distributed information processing for spatial learning and action planning. Specific connectivity and synaptic adaptation principles shape the recurrent dynamics of the network arranged in cortical minicolumns. We show how the PFC columnar organization is suitable for learning sparse topological-metrical representations from redundant hippocampal inputs. The recurrent nature of the network supports multilevel spatial processing, allowing structural features of the environment to be encoded. An activation diffusion mechanism spreads the neural activity through the column population leading to trajectory planning. The model provides a functional framework for interpreting the activity of PFC neurons recorded during navigation tasks. We illustrate the link from single unit activity to behavioral responses. The results suggest plausible neural mechanisms subserving the cognitive "insight" capability originally attributed to rodents by Tolman & Honzik. Our time course analysis of neural responses shows how the interaction between hippocampus and PFC can yield the encoding of manifold information pertinent to spatial planning, including prospective coding and distance-to-goal correlates.     

Using network dynamic fMRI for detection of epileptogenic foci

Background

Epilepsy is one of the most prevalent neurological disorders. It remains medically intractable for about one-third of patients with focal epilepsy, for whom precise localization of the epileptogenic zone responsible for seizure initiation may be critical for successful surgery. Existing fMRI literature points to widespread network disturbances in functional connectivity. Per previous scalp and intracranial EEG studies and consistent with excessive local synchronization during interictal discharges, we hypothesized that, relative to same regions in healthy controls, epileptogenic foci would exhibit less chaotic dynamics, identifiable via entropic analyses of resting state fMRI time series.

Methods

In order to first validate this hypothesis on a cohort of patients with known ground truth, here we test individuals with well-defined epileptogenic foci (left mesial temporal lobe epilepsy). We analyzed voxel-wise resting-state fMRI time-series using the autocorrelation function (ACF), an entropic measure of regulation and feedback, and performed follow-up seed-to-voxel functional connectivity analysis. Disruptions in connectivity of the region exhibiting abnormal dynamics were examined in relation to duration of epilepsy and patients’ cognitive performance using a delayed verbal memory recall task.

Results

ACF analysis revealed constrained (less chaotic) functional dynamics in left temporal lobe epilepsy patients, primarily localized to ipsilateral temporal pole, proximal to presumed focal points. Autocorrelation decay rates differentiated, with 100 % accuracy, between patients and healthy controls on a subject-by-subject basis within a leave-one-subject out classification framework. Regions identified via ACF analysis formed a less efficient network in patients, as compared to controls. Constrained dynamics were linked with locally increased and long-range decreased connectivity that, in turn, correlated significantly with impaired memory (local left temporal connectivity) and epilepsy duration (left temporal – posterior cingulate cortex connectivity).

Conclusions

Our current results suggest that data driven functional MRI methods that target network dynamics hold promise in providing clinically valuable tools for identification of epileptic regions.

     

The analytical characteristics of the dynamics of interneuronal functional connections during conditioned-reflex activity]

The dynamics of functional relations between neurons was studied in the frontal cortex of dogs performing reversal conditioning task. To reveal the functionally relevant relationships between the temporal patterns of correlated firing and behavioral events, we developed an original processing technique. The technique included the following procedures: a) isolation of the "coupled spikes" (CS) from simultaneously recorded impulse trains: b) search for the temporal patterns of correlated firings and their classification by clustering single trials with similar temporal distribution of CS; c) assessment of behavioral significance of the identified patterns by evaluation of the probabilities of coincidence of behavioral events and different CS patterns. Significant correlations between impulse trains were revealed in 38 neuronal pairs of 456 analyzed. The effects of change in behavioral context on the CS dynamics during the task performance were found in 87% of neuronal pairs with correlated activity. In 17 pairs the behavioral conditions were identified, under which potentially connected neurons fired independently during all the periods of the behavioral task. The potentialities of the advanced processing technique are discussed. We suggest that this analysis can provide useful information about the temporal distribution of correlated firings under conditions of nonstereotyped behavior, when an animal reacts in the dynamically organized experimental context.     

Hippocampal Damage Disrupts Eyeblink Conditioning in Mice Lacking Glutamate Receptor Subunit δ2

Cerebellar long-term depression (LTD) at the parallel fiber-Purkinje cell synapses has been proposed to be a neural substrate for classical eyeblink conditioning. Mutant mice lacking the glutamate receptor subunit 2 (GluR2), in which the cerebellar LTD is disrupted, exhibited a severe impairment in the delay eyeblink conditioning with a temporal overlap of CS and US. However, they learned normally trace and delay conditioning without CS-US overlap, suggesting a learning mechanism which does not require the cerebellar LTD.In the present study, we tested possible involvement of the hippocampus in this cerebellar LTD-independent learning. We examined effects of scopolamine and hippocampal lesion on the delay conditioning without CS-US overlap. TheGluR2 mutant mice that received scopolamine or aspiration of the dorsalhippocampus together with its overlying cortex exhibited a severe impairment in learning, while the control mutant mice that received saline or aspiration of the overlying cortex learned normally. In contrast, wild-type mice that received either treatment learned as normally as the control wild-type mice. These results suggest that the hippocampus is essential in the cerebellar LTD-independent learning in the GluR2 mutant mice, indicating a newrole of hippocampus in the paradigm with a short trace interval.     

Recognition and tracking of impulse patterns with delay adaptation in biology-inspired pulse processing neural net (BPN) hardware

The application of an electronic real time emulator for biology-inspired pulse processing neural networks (BPN) to recognition and temporal tracking of discrete impulse patterns via delay adaptation is demonstrated. The electronic emulation includes biologically plausible features, such as asynchronous impulses, membrane potentials and adaptive weights, as well as a mechanism to modify signal delays. The rule for the adaptation of impulse propagation delays is as follows: error neurons detect temporal differences between single impulses of other neurons and adjust corresponding signal delay parameters. In the application presented BPN adapts its time delays in order to form a finely tuned match with a given sequence of three discrete impulses. After learning, BPN is capable not only of highly selective recognition of the learned impulse pattern but also of tracking a gradually changing impulse pattern. Tracking is achieved by continuously re-adjusting the delay profile. Delay adaptation (rather than weight adaptation) appears to be the more effective mechanism for such applications.     

What can the hippocampal representation of environmental geometry tell us about Hebbian learning?

 The importance of the hippocampus in spatial representation is well established. It is suggested that the rodent hippocampal network should provide an optimal substrate for the study of unsupervised Hebbian learning. We focus on the firing characteristics of hippocampal place cells in morphologically different environments. A hard-wired quantitative geometric model of individual place fields is reviewed and presented as the framework in which to understand the additional effects of synaptic plasticity. Existent models employing Hebbian learning are also reviewed. New information is presented regarding the dynamics of place field plasticity over short and long time scales in experiments using barriers and differently shaped walled environments. It is argued that aspects of the temporal dynamics of stability and plasticity in the hippocampal place cell representation both indicate modifications to, and inform the nature of, the synaptic plasticity in place cell models. Our results identify a potential neural basis for long-term incidental learning of environments and provide strong constraints for the way the unsupervised learning in cell assemblies envisaged by Hebb might occur within the hippocampus. Received: 8 March 2002 / Accepted: 13 June 2002 Acknowledgements. This work was supported by the Medical Research Council of the United Kingdom. Correspondence to: C. Lever or N. Burgess (e-mail: colin.lever@ucl.ac.uk; n.burgess@ucl.ac.uk, Tel.: +44-20-76793388 or 1147, Fax: +44-20-76791306 or 1145)     

Shared‐storage clusters

Clusters represent a collection of interconnected computers that collaborate on executing an application and present themselves as one unified computing resource. They are becoming an important segment in the computer industry. The two main flavors of cluster architectures are sharedstorage and sharednothing. This article presents host and I/O implementation details, and performance tradeoffs that need to be enforced due to sharing data in sharedstorage clusters. Sharing data requires the need for global concurrency and coherency protocols to maintain consistency of the database, and enforce data consistency in the local nodes buffers, respectively. Various sharedstorage architectures will be investigated, including the Virtual Shared and SharedIntermediate Memory models. This article also presents few selected sharedstorage clusters, including the DEC VAXCluster, IBM parallel Sysplex and Compaq/Tandem ServerNet.     

The prebiotic chemistry of nucleotides

Diiminosuccinonitrile (DISN), formed by the oxidation of diaminomaleonitrile (DAMN), has been investigated as a potential prebiotic phosphorylating agent. DISN effects the cyclization of 3-adenosine monophosphate to adenosine 2, 3-cyclic phosphate in up to 39% yield. The mechanism of this reaction was investigated. The DISN-mediated phosphorylation of uridine to uridine monophosphate does not proceed efficiently in aqueous solution. The reaction of DISN with uridine-5-phosphate and uridine results in the formation of 2,2-anhydronucleotides and 2,2-anhydronucleosides respectively, and other reaction products resulting from an initial reaction at the 2- and 3-hydroxyl groups. The clay mineral catalysis of the cyclization of adenosine-3-phosphate was investigated using homoionic montmorillonites.     

Basis of the Relationship between Ash Content in the Flag Leaf and Carbon Isotope Discrimination in Kernels of Durum Wheat

The relationship between ash content and carbon isotope discrimination () was studied in durum wheat (Triticum durum Desf.) grown in a Mediterranean region (Northwest Syria) under three different water regimes (hereafter referred to as environments). In two of these environments, 144 genotypes were cultivated under rain-fed conditions. In the third environment, 125 genotypes were cultivated under irrigation. Ash content was measured in the flag leaf about 3 weeks after anthesis, whereas was analysed in mature kernels. Total transpiration of the photosynthetic tissues of the culm contributing, from heading to maturity, to the filling of kernels was also estimated. Leaf ash content, expressed either on dry matter or leaf area basis or as total ash per blade, correlated positively (p< 0.001) with in the three environments. However, this relationship was not the result of a positive correlation across genotypes between and tissue water content. Moreover, only a small part of the variation in across genotypes was explained by concomitant changes in ash content. When all genotypes across the three environments were plotted, and ash content followed a non-linear relationship (r ² = 74), with tending to a plateau as the ash content increased. However, for the set of genotypes and environments combined, total ash content per leaf blade was positively and linearly related (r ² = 0.76) with the accumulated culm transpiration. The non-linear nature of the relationship between ash content and is sustained by the fact that culm transpiration also showed a non-linear relationship with kernel . Therefore, differences in leaf ash content between environments, and to a lesser extent between genotypes, seem to be brought about by variations in accumulated transpiration during grain formation.     

Distribution of the extended family of protein kinase C isoenzymes in fetal organs of mice: an immunohistochemical study

Using isoenzyme-specific antisera, we have studied the distribution of protein kinase C isoforms in fetal mouse organs at the developmental age of 17 days. Two different sets of antibodies, produced by different manufacturers, were employed in this study. The specificity of the antisera was tested by immunoblotting experiments using whole fetal mouse extracts. Immunohistochemistry was carried out by means of an alkaline phosphatase-conjugated secondary antibody. Analysis of fetal mouse longitudinal cryostat sections stained with the antibodies demonstrated a distinct distribution of protein kinase C isoforms in the tissues. Protein kinase C- and C-I were present in all tissues examined, whereas the C-II isoform was absent in the lung and the liver. Protein kinase C- was identified in brain, spinal ganglia, and adrenal gland. The C- isoenzyme was abundantly expressed in spinal ganglia and in the smooth muscle cells of the bronchial wall. Antisera to C- and C- isoforms heavily stained liver, kidney, and spinal ganglia, whereas the C- isozyme was mainly detected in brain, stomach and kidney. Thus, protein kinase C-, C-I, C-II, C-, C- and C- were the isoforms present in many of the organs investigated. The two sets of antibodies gave slightly different results that might be ascribed to the different epitopes recognized by the antisera. One set of antisera was employed to investigate the distribution of the isoforms in selected organs from an earlier developmental age (15 days) and from adult animals. Both qualitative and quantitative differences were seen in comparison with the same organs from a 17-day fetus.     

Intraspecific length heterogeneity of the rDNA-IGR in Arabidopsis thaliana due to homologous recombination

A new heterogeneity of the rDNA spacer of Arabidopsis thaliana, resulting from variation in copy number of the so-called C repeat located downstream of the presumptive polymerase I promoter, is reported. Variation is shown to occur within and between ecotypes. PCR analysis and sequence comparison suggests that the observed length heterogeneity is due to homologous recombination.     

Dual Coding with STDP in a Spiking Recurrent Neural Network Model of the Hippocampus

The firing rate of single neurons in the mammalian hippocampus has been demonstrated to encode for a range of spatial and non-spatial stimuli. It has also been demonstrated that phase of firing, with respect to the theta oscillation that dominates the hippocampal EEG during stereotype learning behaviour, correlates with an animal''s spatial location. These findings have led to the hypothesis that the hippocampus operates using a dual (rate and temporal) coding system. To investigate the phenomenon of dual coding in the hippocampus, we examine a spiking recurrent network model with theta coded neural dynamics and an STDP rule that mediates rate-coded Hebbian learning when pre- and post-synaptic firing is stochastic. We demonstrate that this plasticity rule can generate both symmetric and asymmetric connections between neurons that fire at concurrent or successive theta phase, respectively, and subsequently produce both pattern completion and sequence prediction from partial cues. This unifies previously disparate auto- and hetero-associative network models of hippocampal function and provides them with a firmer basis in modern neurobiology. Furthermore, the encoding and reactivation of activity in mutually exciting Hebbian cell assemblies demonstrated here is believed to represent a fundamental mechanism of cognitive processing in the brain.