Bayesian computation emerges in generic cortical microcircuits through spike-timing-dependent plasticity

A robust event-driven approach to always-on object recognition

We propose a neuromimetic architecture capable of always-on pattern recognition, i.e. at any time during processing. To achieve this, we have extended an existing event-based algorithm (Lagorce et al., 2017), which introduced novel spatio-temporal features as a Hierarchy Of Time-Surfaces (HOTS). Built from asynchronous events captured by a neuromorphic camera, these time surfaces allow to encode the local dynamics of a visual scene and to create an efficient event-based pattern recognition architecture. Inspired by neuroscience, we have extended this method to improve its performance. First, we add a homeostatic gain control on the activity of neurons to improve the learning of spatio-temporal patterns (Grimaldi et al., 2021). We also provide a new mathematical formalism that allows an analogy to be drawn between the HOTS algorithm and Spiking Neural Networks (SNN). Following this analogy, we transform the offline pattern categorization method into an online and event-driven layer. This classifier uses the spiking output of the network to define new time surfaces and we then perform the online classification with a neuromimetic implementation of a multinomial logistic regression. These improvements not only consistently increase the performance of the network, but also bring this event-driven pattern recognition algorithm fully online. The results have been validated on different datasets: Poker-DVS (Serrano-Gotarredona and Linares-Barranco, 2015), N-MNIST (Orchard, Jayawant et al., 2015) and DVS Gesture (Amir et al., 2017). This demonstrates the efficiency of this bio-realistic SNN for ultra-fast object recognition through an event-by-event categorization process.

Exploring the Architectural Biases of the Canonical Cortical Microcircuit

The cortex plays a crucial role in various perceptual and cognitive functions, driven by its basic unit, the canonical cortical microcircuit. Yet, we remain short of a framework that definitively explains the structure-function relationships of this fundamental neuroanatomical motif. To better understand how physical substrates of cortical circuitry facilitate their neuronal dynamics, we employ a computational approach using recurrent neural networks and representational analyses. We examine the differences manifested by the inclusion and exclusion of biologically-motivated inter-areal laminar connections on the computational roles of different neuronal populations in the microcircuit of two hierarchically-related areas, throughout learning. Our findings show that the presence of feedback connections correlates with the functional modularization of cortical populations in different layers, and provides the microcircuit with a natural inductive bias to differentiate expected and unexpected inputs at initialization. Furthermore, when testing the effects of training the microcircuit and its variants with a predictive-coding inspired strategy, we find that doing so helps better encode noisy stimuli in areas of the cortex that receive feedback, all of which combine to suggest evidence for a predictive-coding mechanism serving as an intrinsic operative logic in the cortex.

Stochastic photonic spiking neuron for Bayesian inference with unsupervised learning

Stochasticity is an inherent feature of biological neural activities. We propose a noise-injection scheme to implement a GHz-rate stochastic photonic spiking neuron (S-PSN). The firing-probability encoding is experimentally demonstrated and exploited for Bayesian inference with unsupervised learning. In a breast diagnosis task, the stochastic photonic spiking neural network (S-PSNN) can not only achieve a classification accuracy of 96.6%, but can also evaluate the diagnosis uncertainty with prediction entropies. As a result, the misdiagnosis rate is reduced by 80% compared to that of a conventional deterministic photonic spiking neural network (D-PSNN) for the same task. The GHz-rate S-PSN endows the neuromorphic photonics with high-speed Bayesian inference for reliable information processing in error-critical scenarios.

Synaptic self-organization of spatio-temporal pattern selectivity

Spiking model neurons can be set up to respond selectively to specific spatio-temporal spike patterns by optimization of their input weights. It is unknown, however, if existing synaptic plasticity mechanisms can achieve this temporal mode of neuronal coding and computation. Here it is shown that changes of synaptic efficacies which tend to balance excitatory and inhibitory synaptic inputs can make neurons sensitive to particular input spike patterns. Simulations demonstrate that a combination of Hebbian mechanisms, hetero-synaptic plasticity and synaptic scaling is sufficient for self-organizing sensitivity for spatio-temporal spike patterns that repeat in the input. In networks inclusion of hetero-synaptic plasticity that depends on the pre-synaptic neurons leads to specialization and faithful representation of pattern sequences by a group of target neurons. Pattern detection is robust against a range of distortions and noise. The proposed combination of Hebbian mechanisms, hetero-synaptic plasticity and synaptic scaling is found to protect the memories for specific patterns from being overwritten by ongoing learning during extended periods when the patterns are not present. This suggests a novel explanation for the long term robustness of memory traces despite ongoing activity with substantial synaptic plasticity. Taken together, our results promote the plausibility of precise temporal coding in the brain.

Spintronic Integrate-Fire-Reset Neuron with Stochasticity for Neuromorphic Computing

Spintronics has been recently extended to neuromorphic computing because of its energy efficiency and scalability. However, a biorealistic spintronic neuron with probabilistic "spiking" and a spontaneous reset functionality has not been demonstrated yet. Here, we propose a biorealistic spintronic neuron device based on the heavy metal (HM)/ferromagnet (FM)/antiferromagnet (AFM) spin-orbit torque (SOT) heterostructure. The spintronic neuron can autoreset itself after firing due to the exchange bias of the AFM. The firing process is inherently stochastic because of the competition between the SOT and AFM pinning effects. We also implement a restricted Boltzmann machine (RBM) and stochastic integration multilayer perceptron (SI-MLP) using our proposed neuron. Despite the bit-width limitation, the proposed spintronic model can achieve an accuracy of 97.38% in pattern recognition, which is even higher than the baseline accuracy (96.47%). Our results offer a spintronic device solution to emulate biologically realistic spiking neurons.

Attractive and repulsive effects of sensory history concurrently shape visual perception

Background

Sequential effects of environmental stimuli are ubiquitous in most behavioral tasks involving magnitude estimation, memory, decision making, and emotion. The human visual system exploits continuity in the visual environment, which induces two contrasting perceptual phenomena shaping visual perception. Previous work reported that perceptual estimation of a stimulus may be influenced either by attractive serial dependencies or repulsive aftereffects, with a number of experimental variables suggested as factors determining the direction and magnitude of sequential effects. Recent studies have theorized that these two effects concurrently arise in perceptual processing, but empirical evidence that directly supports this hypothesis is lacking, and it remains unclear whether and how attractive and repulsive sequential effects interact in a trial. Here we show that the two effects concurrently modulate estimation behavior in a typical sequence of perceptual tasks.

Results

We first demonstrate that observers’ estimation error as a function of both the previous stimulus and response cannot be fully described by either attractive or repulsive bias but is instead well captured by a summation of repulsion from the previous stimulus and attraction toward the previous response. We then reveal that the repulsive bias is centered on the observer’s sensory encoding of the previous stimulus, which is again repelled away from its own preceding trial, whereas the attractive bias is centered precisely on the previous response, which is the observer’s best prediction about the incoming stimuli.

Conclusions

Our findings provide strong evidence that sensory encoding is shaped by dynamic tuning of the system to the past stimuli, inducing repulsive aftereffects, and followed by inference incorporating the prediction from the past estimation, leading to attractive serial dependence.

Post-silicon nano-electronic device and its application in brain-inspired chips

As information technology is moving toward the era of big data, the traditional Von-Neumann architecture shows limitations in performance. The field of computing has already struggled with the latency and bandwidth required to access memory (“the memory wall”) and energy dissipation (“the power wall”). These challenging issues, such as “the memory bottleneck,” call for significant research investments to develop a new architecture for the next generation of computing systems. Brain-inspired computing is a new computing architecture providing a method of high energy efficiency and high real-time performance for artificial intelligence computing. Brain-inspired neural network system is based on neuron and synapse. The memristive device has been proposed as an artificial synapse for creating neuromorphic computer applications. In this study, post-silicon nano-electronic device and its application in brain-inspired chips are surveyed. First, we introduce the development of neural networks and review the current typical brain-inspired chips, including brain-inspired chips dominated by analog circuit and brain-inspired chips of the full-digital circuit, leading to the design of brain-inspired chips based on post-silicon nano-electronic device. Then, through the analysis of N kinds of post-silicon nano-electronic devices, the research progress of constructing brain-inspired chips using post-silicon nano-electronic device is expounded. Lastly, the future of building brain-inspired chips based on post-silicon nano-electronic device has been prospected.

Phase-change memtransistive synapses for mixed-plasticity neural computations

In the mammalian nervous system, various synaptic plasticity rules act, either individually or synergistically, over wide-ranging timescales to enable learning and memory formation. Hence, in neuromorphic computing platforms, there is a significant need for artificial synapses that can faithfully express such multi-timescale plasticity mechanisms. Although some plasticity rules have been emulated with elaborate complementary metal oxide semiconductor and memristive circuitry, device-level hardware realizations of long-term and short-term plasticity with tunable dynamics are lacking. Here we introduce a phase-change memtransistive synapse that leverages both the non-volatility of the phase configurations and the volatility of field-effect modulation for implementing tunable plasticities. We show that these mixed-plasticity synapses can enable plasticity rules such as short-term spike-timing-dependent plasticity that helps with the modelling of dynamic environments. Further, we demonstrate the efficacy of the memtransistive synapses in realizing accelerators for Hopfield neural networks for solving combinatorial optimization problems.

Emergence of probabilistic representation in the neural network of primary visual cortex

During the early development of the mammalian visual system, the distribution of neuronal preferred orientations in the primary visual cortex (V1) gradually shifts to match major orientation features of the environment, achieving its optimal representation. By combining computational modeling and electrophysiological recording, we provide a circuit plasticity mechanism that underlies the developmental emergence of such matched representation in the visual cortical network. Specifically, in a canonical circuit of densely-interconnected pyramidal cells and inhibitory parvalbumin-expressing (PV+) fast-spiking interneurons in V1 layer 2/3, our model successfully simulates the experimental observations and further reveals that the nonuniform inhibition plays a key role in shaping the network representation through spike timing-dependent plasticity. The experimental results suggest that PV + interneurons in V1 are capable of providing nonuniform inhibition shortly after vision onset. Our study elucidates a circuit mechanism for acquisition of prior knowledge of environment for optimal inference in sensory neural systems

•

Computational and experimental methods are combined to representation in mice V1

•

Nonuniform inhibition plays a key role in shaping the network representation

•

PV + interneurons provide nonuniform inhibition shortly after vision onset

Toward Efficient Processing and Learning With Spikes: New Approaches for Multispike Learning

Spikes are the currency in central nervous systems for information transmission and processing. They are also believed to play an essential role in low-power consumption of the biological systems, whose efficiency attracts increasing attentions to the field of neuromorphic computing. However, efficient processing and learning of discrete spikes still remain a challenging problem. In this article, we make our contributions toward this direction. A simplified spiking neuron model is first introduced with the effects of both synaptic input and firing output on the membrane potential being modeled with an impulse function. An event-driven scheme is then presented to further improve the processing efficiency. Based on the neuron model, we propose two new multispike learning rules which demonstrate better performance over other baselines on various tasks, including association, classification, and feature detection. In addition to efficiency, our learning rules demonstrate high robustness against the strong noise of different types. They can also be generalized to different spike coding schemes for the classification task, and notably, the single neuron is capable of solving multicategory classifications with our learning rules. In the feature detection task, we re-examine the ability of unsupervised spike-timing-dependent plasticity with its limitations being presented, and find a new phenomenon of losing selectivity. In contrast, our proposed learning rules can reliably solve the task over a wide range of conditions without specific constraints being applied. Moreover, our rules cannot only detect features but also discriminate them. The improved performance of our methods would contribute to neuromorphic computing as a preferable choice.

A Hebbian Approach to Non-Spatial Prelinguistic Reasoning

This research integrates key concepts of Computational Neuroscience, including the Bienestock-CooperMunro (BCM) rule, Spike Timing-Dependent Plasticity Rules (STDP), and the Temporal Difference Learning algorithm, with an important structure of Deep Learning (Convolutional Networks) to create an architecture with the potential of replicating observations of some cognitive experiments (particularly, those that provided some basis for sequential reasoning) while sharing the advantages already achieved by the previous proposals. In particular, we present Ring Model B, which is capable of associating visual with auditory stimulus, performing sequential predictions, and predicting reward from experience. Despite its simplicity, we considered such abilities to be a first step towards the formulation of more general models of prelinguistic reasoning.

Opportunities for neuromorphic computing algorithms and applications

Neuromorphic computing technologies will be important for the future of computing, but much of the work in neuromorphic computing has focused on hardware development. Here, we review recent results in neuromorphic computing algorithms and applications. We highlight characteristics of neuromorphic computing technologies that make them attractive for the future of computing and we discuss opportunities for future development of algorithms and applications on these systems.

Local dendritic balance enables learning of efficient representations in networks of spiking neurons

How can neural networks learn to efficiently represent complex and high-dimensional inputs via local plasticity mechanisms? Classical models of representation learning assume that feedforward weights are learned via pairwise Hebbian-like plasticity. Here, we show that pairwise Hebbian-like plasticity works only under unrealistic requirements on neural dynamics and input statistics. To overcome these limitations, we derive from first principles a learning scheme based on voltage-dependent synaptic plasticity rules. Here, recurrent connections learn to locally balance feedforward input in individual dendritic compartments and thereby can modulate synaptic plasticity to learn efficient representations. We demonstrate in simulations that this learning scheme works robustly even for complex high-dimensional inputs and with inhibitory transmission delays, where Hebbian-like plasticity fails. Our results draw a direct connection between dendritic excitatory-inhibitory balance and voltage-dependent synaptic plasticity as observed in vivo and suggest that both are crucial for representation learning.

A Fully Spiking Hybrid Neural Network for Energy-Efficient Object Detection

This paper proposes a Fully Spiking Hybrid Neural Network (FSHNN) for energy-efficient and robust object detection in resource-constrained platforms. The network architecture is based on a Spiking Convolutional Neural Network using leaky-integrate-fire neuron models. The model combines unsupervised Spike Time-Dependent Plasticity (STDP) learning with back-propagation (STBP) learning methods and also uses Monte Carlo Dropout to get an estimate of the uncertainty error. FSHNN provides better accuracy compared to DNN based object detectors while being more energy-efficient. It also outperforms these object detectors, when subjected to noisy input data and less labeled training data with a lower uncertainty error.

Visual exposure enhances stimulus encoding and persistence in primary cortex

Experience shapes sensory responses, already at the earliest stages of cortical processing. We provide evidence that, in the primary visual cortex of anesthetized cats, brief repetitive exposure to a set of simple, abstract stimuli expands the range and decreases the variability of neuronal responses that persist after stimulus offset. These refinements increase the stimulus-specific clustering of neuronal population responses and result in a more efficient encoding of both stimulus identity and stimulus structure, thus potentially benefiting simple readouts in higher cortical areas. Similar results can be achieved via local plasticity mechanisms in recurrent networks, through self-organized refinements of internal dynamics that do not require changes in firing amplitudes.

The brain adapts to the sensory environment. For example, simple sensory exposure can modify the response properties of early sensory neurons. How these changes affect the overall encoding and maintenance of stimulus information across neuronal populations remains unclear. We perform parallel recordings in the primary visual cortex of anesthetized cats and find that brief, repetitive exposure to structured visual stimuli enhances stimulus encoding by decreasing the selectivity and increasing the range of the neuronal responses that persist after stimulus presentation. Low-dimensional projection methods and simple classifiers demonstrate that visual exposure increases the segregation of persistent neuronal population responses into stimulus-specific clusters. These observed refinements preserve the representational details required for stimulus reconstruction and are detectable in postexposure spontaneous activity. Assuming response facilitation and recurrent network interactions as the core mechanisms underlying stimulus persistence, we show that the exposure-driven segregation of stimulus responses can arise through strictly local plasticity mechanisms, also in the absence of firing rate changes. Our findings provide evidence for the existence of an automatic, unguided optimization process that enhances the encoding power of neuronal populations in early visual cortex, thus potentially benefiting simple readouts at higher stages of visual processing.

Neural implementations of Bayesian inference

Bayesian inference has emerged as a general framework that captures how organisms make decisions under uncertainty. Recent experimental findings reveal disparate mechanisms for how the brain generates behaviors predicted by normative Bayesian theories. Here, we identify two broad classes of neural implementations for Bayesian inference: a modular class, where each probabilistic component of Bayesian computation is independently encoded and a transform class, where uncertain measurements are converted to Bayesian estimates through latent processes. Many recent experimental neuroscience findings studying probabilistic inference broadly fall into these classes. We identify potential avenues for synthesis across these two classes and the disparities that, at present, cannot be reconciled. We conclude that to distinguish among implementation hypotheses for Bayesian inference, we require greater engagement among theoretical and experimental neuroscientists in an effort that spans different scales of analysis, circuits, tasks, and species.

Stochastic binary synapses having sigmoidal cumulative distribution functions for unsupervised learning with spike timing-dependent plasticity

Spike timing-dependent plasticity (STDP), which is widely studied as a fundamental synaptic update rule for neuromorphic hardware, requires precise control of continuous weights. From the viewpoint of hardware implementation, a simplified update rule is desirable. Although simplified STDP with stochastic binary synapses was proposed previously, we find that it leads to degradation of memory maintenance during learning, which is unfavourable for unsupervised online learning. In this work, we propose a stochastic binary synaptic model where the cumulative probability of the weight change evolves in a sigmoidal fashion with potentiation or depression trials, which can be implemented using a pair of switching devices consisting of serially connected multiple binary memristors. As a benchmark test we perform simulations of unsupervised learning of MNIST images with a two-layer network and show that simplified STDP in combination with this model can outperform conventional rules with continuous weights not only in memory maintenance but also in recognition accuracy. Our method achieves 97.3% in recognition accuracy, which is higher than that reported with standard STDP in the same framework. We also show that the high performance of our learning rule is robust against device-to-device variability of the memristor's probabilistic behaviour.

Thalamo-cortical spiking model of incremental learning combining perception, context and NREM-sleep

The brain exhibits capabilities of fast incremental learning from few noisy examples, as well as the ability to associate similar memories in autonomously-created categories and to combine contextual hints with sensory perceptions. Together with sleep, these mechanisms are thought to be key components of many high-level cognitive functions. Yet, little is known about the underlying processes and the specific roles of different brain states. In this work, we exploited the combination of context and perception in a thalamo-cortical model based on a soft winner-take-all circuit of excitatory and inhibitory spiking neurons. After calibrating this model to express awake and deep-sleep states with features comparable with biological measures, we demonstrate the model capability of fast incremental learning from few examples, its resilience when proposed with noisy perceptions and contextual signals, and an improvement in visual classification after sleep due to induced synaptic homeostasis and association of similar memories.

We created a thalamo-cortical spiking model (ThaCo) with the purpose of demonstrating a link among two phenomena that we believe to be essential for the brain capability of efficient incremental learning from few examples in noisy environments. Grounded in two experimental observations—the first about the effects of deep-sleep on pre- and post-sleep firing rate distributions, the second about the combination of perceptual and contextual information in pyramidal neurons—our model joins these two ingredients. ThaCo alternates phases of incremental learning, classification and deep-sleep. Memories of handwritten digit examples are learned through thalamo-cortical and cortico-cortical plastic synapses. In absence of noise, the combination of contextual information with perception enables fast incremental learning. Deep-sleep becomes crucial when noisy inputs are considered. We observed in ThaCo both homeostatic and associative processes: deep-sleep fights noise in perceptual and internal knowledge and it supports the categorical association of examples belonging to the same digit class, through reinforcement of class-specific cortico-cortical synapses. The distributions of pre-sleep and post-sleep firing rates during classification change in a manner similar to those of experimental observation. These changes promote energetic efficiency during recall of memories, better representation of individual memories and categories and higher classification performances.

Robust cortical criticality and diverse dynamics resulting from functional specification

Despite the recognition of the layered structure and evident criticality in the cortex, how the specification of input, output, and computational layers affects the self-organized criticality has not been much explored. By constructing heterogeneous structures with a well-accepted model of leaky neurons, we find that the specification can lead to robust criticality rather insensitive to the strength of external stimuli. This naturally unifies the adaptation to strong inputs without extra synaptic plasticity mechanisms. Low degree of recurrence constitutes an alternative explanation to subcriticality other than the high-frequency inputs. Unlike fully recurrent networks where external stimuli always render subcriticality, the dynamics of networks with sufficient feedforward connections can be driven to criticality and supercriticality. These findings indicate that functional and structural specification and their interplay with external stimuli are of crucial importance for the network dynamics. The robust criticality puts forward networks of the leaky neurons as promising platforms for realizing artificial neural networks that work in the vicinity of critical points.

Experience-dependent contextual codes in the hippocampus

The hippocampus contains neural representations capable of supporting declarative memory. Hippocampal place cells are one such representation, firing in one or few locations in a given environment. Between environments, place cell firing fields remap (turning on/off or moving to a new location) to provide a population-wide code for distinct contexts. However, the manner by which contextual features combine to drive hippocampal remapping remains a matter of debate. Using large-scale in vivo two-photon intracellular calcium recordings in mice during virtual navigation, we show that remapping in the hippocampal region CA1 is driven by prior experience regarding the frequency of certain contexts and that remapping approximates an optimal estimate of the identity of the current context. A simple associative-learning mechanism reproduces these results. Together, our findings demonstrate that place cell remapping allows an animal to simultaneously identify its physical location and optimally estimate the identity of the environment.

CRBA: A Competitive Rate-Based Algorithm Based on Competitive Spiking Neural Networks

In this paper we present a Competitive Rate-Based Algorithm (CRBA) that approximates operation of a Competitive Spiking Neural Network (CSNN). CRBA is based on modeling of the competition between neurons during a sample presentation, which can be reduced to ranking of the neurons based on a dot product operation and the use of a discrete Expectation Maximization algorithm; the latter is equivalent to the spike time-dependent plasticity rule. CRBA's performance is compared with that of CSNN on the MNIST and Fashion-MNIST datasets. The results show that CRBA performs on par with CSNN, while using three orders of magnitude less computational time. Importantly, we show that the weights and firing thresholds learned by CRBA can be used to initialize CSNN's parameters that results in its much more efficient operation.

Clone-structured graph representations enable flexible learning and vicarious evaluation of cognitive maps

Cognitive maps are mental representations of spatial and conceptual relationships in an environment, and are critical for flexible behavior. To form these abstract maps, the hippocampus has to learn to separate or merge aliased observations appropriately in different contexts in a manner that enables generalization and efficient planning. Here we propose a specific higher-order graph structure, clone-structured cognitive graph (CSCG), which forms clones of an observation for different contexts as a representation that addresses these problems. CSCGs can be learned efficiently using a probabilistic sequence model that is inherently robust to uncertainty. We show that CSCGs can explain a variety of cognitive map phenomena such as discovering spatial relations from aliased sensations, transitive inference between disjoint episodes, and formation of transferable schemas. Learning different clones for different contexts explains the emergence of splitter cells observed in maze navigation and event-specific responses in lap-running experiments. Moreover, learning and inference dynamics of CSCGs offer a coherent explanation for disparate place cell remapping phenomena. By lifting aliased observations into a hidden space, CSCGs reveal latent modularity useful for hierarchical abstraction and planning. Altogether, CSCG provides a simple unifying framework for understanding hippocampal function, and could be a pathway for forming relational abstractions in artificial intelligence.

In situ Parallel Training of Analog Neural Network Using Electrochemical Random-Access Memory

In-memory computing based on non-volatile resistive memory can significantly improve the energy efficiency of artificial neural networks. However, accurate in situ training has been challenging due to the nonlinear and stochastic switching of the resistive memory elements. One promising analog memory is the electrochemical random-access memory (ECRAM), also known as the redox transistor. Its low write currents and linear switching properties across hundreds of analog states enable accurate and massively parallel updates of a full crossbar array, which yield rapid and energy-efficient training. While simulations predict that ECRAM based neural networks achieve high training accuracy at significantly higher energy efficiency than digital implementations, these predictions have not been experimentally achieved. In this work, we train a 3 × 3 array of ECRAM devices that learns to discriminate several elementary logic gates (AND, OR, NAND). We record the evolution of the network's synaptic weights during parallel in situ (on-line) training, with outer product updates. Due to linear and reproducible device switching characteristics, our crossbar simulations not only accurately simulate the epochs to convergence, but also quantitatively capture the evolution of weights in individual devices. The implementation of the first in situ parallel training together with strong agreement with simulation results provides a significant advance toward developing ECRAM into larger crossbar arrays for artificial neural network accelerators, which could enable orders of magnitude improvements in energy efficiency of deep neural networks.

Computational Efficiency of a Modular Reservoir Network for Image Recognition

Liquid state machine (LSM) is a type of recurrent spiking network with a strong relationship to neurophysiology and has achieved great success in time series processing. However, the computational cost of simulations and complex dynamics with time dependency limit the size and functionality of LSMs. This paper presents a large-scale bioinspired LSM with modular topology. We integrate the findings on the visual cortex that specifically designed input synapses can fit the activation of the real cortex and perform the Hough transform, a feature extraction algorithm used in digital image processing, without additional cost. We experimentally verify that such a combination can significantly improve the network functionality. The network performance is evaluated using the MNIST dataset where the image data are encoded into spiking series by Poisson coding. We show that the proposed structure can not only significantly reduce the computational complexity but also achieve higher performance compared to the structure of previous reported networks of a similar size. We also show that the proposed structure has better robustness against system damage than the small-world and random structures. We believe that the proposed computationally efficient method can greatly contribute to future applications of reservoir computing.

Threshold switching memristor-based stochastic neurons for probabilistic computing

Biological neurons exhibit dynamic excitation behavior in the form of stochastic firing, rather than stiffly giving out spikes upon reaching a fixed threshold voltage, which empowers the brain to perform probabilistic inference in the face of uncertainty. However, owing to the complexity of the stochastic firing process in biological neurons, the challenge of fabricating and applying stochastic neurons with bio-realistic dynamics to probabilistic scenarios remains to be fully addressed. In this work, a novel CuS/GeSe conductive-bridge threshold switching memristor is fabricated and singled out to realize electronic stochastic neurons, which is ascribed to the similarity between the stochastic switching behavior observed in the device and that of biological ion channels. The corresponding electric circuit of a stochastic neuron is then constructed and the probabilistic firing capacity of the neuron is utilized to implement Bayesian inference in a spiking neural network (SNN). The application prospects are demonstrated on the example of a tumor diagnosis task, where common fatal diagnostic errors of a conventional artificial neural network are successfully circumvented. Moreover, in comparison to deterministic neuron-based SNNs, the stochastic neurons enable SNNs to deliver an estimate of the uncertainty in their predictions, and the fidelity of the judgement is drastically improved by 81.2%.

A generative spiking neural-network model of goal-directed behaviour and one-step planning

In mammals, goal-directed and planning processes support flexible behaviour used to face new situations that cannot be tackled through more efficient but rigid habitual behaviours. Within the Bayesian modelling approach of brain and behaviour, models have been proposed to perform planning as probabilistic inference but this approach encounters a crucial problem: explaining how such inference might be implemented in brain spiking networks. Recently, the literature has proposed some models that face this problem through recurrent spiking neural networks able to internally simulate state trajectories, the core function at the basis of planning. However, the proposed models have relevant limitations that make them biologically implausible, namely their world model is trained ‘off-line’ before solving the target tasks, and they are trained with supervised learning procedures that are biologically and ecologically not plausible. Here we propose two novel hypotheses on how brain might overcome these problems, and operationalise them in a novel architecture pivoting on a spiking recurrent neural network. The first hypothesis allows the architecture to learn the world model in parallel with its use for planning: to this purpose, a new arbitration mechanism decides when to explore, for learning the world model, or when to exploit it, for planning, based on the entropy of the world model itself. The second hypothesis allows the architecture to use an unsupervised learning process to learn the world model by observing the effects of actions. The architecture is validated by reproducing and accounting for the learning profiles and reaction times of human participants learning to solve a visuomotor learning task that is new for them. Overall, the architecture represents the first instance of a model bridging probabilistic planning and spiking-processes that has a degree of autonomy analogous to the one of real organisms.

Goal-directed behaviour relies on brain processes supporting planning of actions based on their expected consequences before performing them in the environment. An important computational modelling approach proposes that the brain performs goal-directed processes on the basis of probability distributions and computations on them. A key challenge of this approach is to explain how these probabilistic processes can rely on the spiking processes of the brain. The literature has recently proposed some models that do so by ‘thinking ahead’ alternative possible action-outcomes based on low-level neuronal stochastic events. However, these models have a limited autonomy as they require to learn how the environment works (‘world model’) before solving the tasks, and use a biologically implausible learning process requiring an ‘external teacher’ to tell how their internal units should respond. Here we present a novel architecture proposing how organisms might overcome these challenging problems. First, the architecture can decide if exploring, to learn the world model, or planning, using such model, by evaluating how confident it is on the model knowledge. Second, the architecture can autonomously learn the world model based on experience. The architecture represents a first fully autonomous planning model relying on a spiking neural network.

The covariance perceptron: A new paradigm for classification and processing of time series in recurrent neuronal networks

Learning in neuronal networks has developed in many directions, in particular to reproduce cognitive tasks like image recognition and speech processing. Implementations have been inspired by stereotypical neuronal responses like tuning curves in the visual system, where, for example, ON/OFF cells fire or not depending on the contrast in their receptive fields. Classical models of neuronal networks therefore map a set of input signals to a set of activity levels in the output of the network. Each category of inputs is thereby predominantly characterized by its mean. In the case of time series, fluctuations around this mean constitute noise in this view. For this paradigm, the high variability exhibited by the cortical activity may thus imply limitations or constraints, which have been discussed for many years. For example, the need for averaging neuronal activity over long periods or large groups of cells to assess a robust mean and to diminish the effect of noise correlations. To reconcile robust computations with variable neuronal activity, we here propose a conceptual change of perspective by employing variability of activity as the basis for stimulus-related information to be learned by neurons, rather than merely being the noise that corrupts the mean signal. In this new paradigm both afferent and recurrent weights in a network are tuned to shape the input-output mapping for covariances, the second-order statistics of the fluctuating activity. When including time lags, covariance patterns define a natural metric for time series that capture their propagating nature. We develop the theory for classification of time series based on their spatio-temporal covariances, which reflect dynamical properties. We demonstrate that recurrent connectivity is able to transform information contained in the temporal structure of the signal into spatial covariances. Finally, we use the MNIST database to show how the covariance perceptron can capture specific second-order statistical patterns generated by moving digits.

The dynamics in cortex is characterized by highly fluctuating activity: Even under the very same experimental conditions the activity typically does not reproduce on the level of individual spikes. Given this variability, how then does the brain realize its quasi-deterministic function? One obvious solution is to compute averages over many cells, assuming that the mean activity, or rate, is actually the decisive signal. Variability across trials of an experiment is thus considered noise. We here explore the opposite view: Can fluctuations be used to actually represent information? And if yes, is there a benefit over a representation using the mean rate? We find that a fluctuation-based scheme is not only powerful in distinguishing signals into several classes, but also that networks can efficiently be trained in the new paradigm. Moreover, we argue why such a scheme of representation is more consistent with known forms of synaptic plasticity than rate-based network dynamics.

Slowly activating outward membrane currents generate input-output sub-harmonic cross frequency coupling in neurons

A major challenge in understanding spike-time dependent information encoding in the neural system is the non-linear firing response to inputs of the individual neurons. Hence, quantitative exploration of the putative mechanisms of this non-linear behavior is fundamental to formulating the theory of information transfer in the neural system. The objective of this simulation study was to evaluate and quantify the effect of slowly activating outward membrane current, on the non-linearity in the output of a one-compartment Hodgkin-Huxley styled neuron. To evaluate this effect, the peak conductance of the slow potassium channel (g_K-slow) was varied from 0% to 200% of its normal value in steps of 33%. Both cross- and iso-frequency coupling between the input and the output of the simulated neuron was computed using a generalized coherence measure, i.e., n:m coherence. With increasing g_K-slow, the amount of sub-harmonic cross-frequency coupling, where the output frequencies (1-8 Hz) are lower than the input frequencies (15-35 Hz), increased progressively whereas no change in iso-frequency coupling was observed. Power spectral and phase-space analysis of the neuronal membrane voltage vs. slow potassium channel activation variable showed that the interaction of the slow channel dynamics with the fast membrane voltage dynamics generates the observed sub-harmonic coupling. This study provides quantitative insights into the role of an important membrane mechanism i.e. the slowly activating outward current in generating non-linearities in the output of a neuron.

Discrete Langevin machine: Bridging the gap between thermodynamic and neuromorphic systems

A formulation of Langevin dynamics for discrete systems is derived as a class of generic stochastic processes. The dynamics simplify for a two-state system and suggest a network architecture which is implemented by the Langevin machine. The Langevin machine represents a promising approach to compute successfully quantitative exact results of Boltzmann distributed systems by LIF neurons. Besides a detailed introduction of the dynamics, different simplified models of a neuromorphic hardware system are studied with respect to a control of emerging sources of errors.

Improving the Izhikevich Model Based on Rat Basolateral Amygdala and Hippocampus Neurons, and Recognizing Their Possible Firing Patterns

Introduction

Identifying the potential firing patterns following different brain regions under normal and abnormal conditions increases our understanding of events at the level of neural interactions in the brain. Furthermore, it is important to be capable of modeling the potential neural activities to build precise artificial neural networks. The Izhikevich model is one of the simplest biologically-plausible models, i.e. capable of capturing most recognized firing patterns of neurons. This property makes the model efficient in simulating the large-scale networks of neurons. Improving the Izhikevich model for adapting with the neuronal activity of rat brain with great accuracy would make the model effective for future neural network implementations.

Methods

Data sampling from two brain regions, the HIP and BLA, was performed by the extracellular recordings of male Wistar rats, and spike sorting was conducted by Plexon offline sorter. Further analyses were performed through NeuroExplorer and MATLAB. To optimize the Izhikevich model parameters, a genetic algorithm was used. In this algorithm, optimization tools, like crossover and mutation, provide the basis for generating model parameters populations. The process of comparison in each iteration leads to the survival of better populations until achieving the optimum solution.

Results

In the present study, the possible firing patterns of the real single neurons of the HIP and BLA were identified. Additionally, an improved Izhikevich model was achieved. Accordingly, the real neuronal spiking pattern of these regions' neurons and the corresponding cases of the Izhikevich neuron spiking pattern were adjusted with great accuracy.

Conclusion

This study was conducted to elevate our knowledge of neural interactions in different structures of the brain and accelerate the quality of future large-scale neural networks simulations, as well as reducing the modeling complexity. This aim was achievable by performing the improved Izhikevich model, and inserting only the plausible firing patterns and eliminating unrealistic ones.

Noise Helps Optimization Escape From Saddle Points in the Synaptic Plasticity

Numerous experimental studies suggest that noise is inherent in the human brain. However, the functional importance of noise remains unknown. n particular, from a computational perspective, such stochasticity is potentially harmful to brain function. In machine learning, a large number of saddle points are surrounded by high error plateaus and give the illusion of the existence of local minimum. As a result, being trapped in the saddle points can dramatically impair learning and adding noise will attack such saddle point problems in high-dimensional optimization, especially under the strict saddle condition. Motivated by these arguments, we propose one biologically plausible noise structure and demonstrate that noise can efficiently improve the optimization performance of spiking neural networks based on stochastic gradient descent. The strict saddle condition for synaptic plasticity is deduced, and under such conditions, noise can help optimization escape from saddle points on high dimensional domains. The theoretical results explain the stochasticity of synapses and guide us on how to make use of noise. In addition, we provide biological interpretations of proposed noise structures from two points: one based on the free energy principle in neuroscience and another based on observations of in vivo experiments. Our simulation results manifest that in the learning and test phase, the accuracy of synaptic sampling with noise is almost 20% higher than that without noise for synthesis dataset, and the gain in accuracy with/without noise is at least 10% for the MNIST and CIFAR-10 dataset. Our study provides a new learning framework for the brain and sheds new light on deep noisy spiking neural networks.

Flexible Neuromorphic Electronics for Computing, Soft Robotics, and Neuroprosthetics

Flexible neuromorphic electronics that emulate biological neuronal systems constitute a promising candidate for next-generation wearable computing, soft robotics, and neuroprosthetics. For realization, with the achievement of simple synaptic behaviors in a single device, the construction of artificial synapses with various functions of sensing and responding and integrated systems to mimic complicated computing, sensing, and responding in biological systems is a prerequisite. Artificial synapses that have learning ability can perceive and react to events in the real world; these abilities expand the neuromorphic applications toward health monitoring and cybernetic devices in the future Internet of Things. To demonstrate the flexible neuromorphic systems successfully, it is essential to develop artificial synapses and nerves replicating the functionalities of the biological counterparts and satisfying the requirements for constructing the elements and the integrated systems such as flexibility, low power consumption, high-density integration, and biocompatibility. Here, the progress of flexible neuromorphic electronics is addressed, from basic backgrounds including synaptic characteristics, device structures, and mechanisms of artificial synapses and nerves, to applications for computing, soft robotics, and neuroprosthetics. Finally, future research directions toward wearable artificial neuromorphic systems are suggested for this emerging area.

Somatodendritic consistency check for temporal feature segmentation

The brain identifies potentially salient features within continuous information streams to process hierarchical temporal events. This requires the compression of information streams, for which effective computational principles are yet to be explored. Backpropagating action potentials can induce synaptic plasticity in the dendrites of cortical pyramidal neurons. By analogy with this effect, we model a self-supervising process that increases the similarity between dendritic and somatic activities where the somatic activity is normalized by a running average. We further show that a family of networks composed of the two-compartment neurons performs a surprisingly wide variety of complex unsupervised learning tasks, including chunking of temporal sequences and the source separation of mixed correlated signals. Common methods applicable to these temporal feature analyses were previously unknown. Our results suggest the powerful ability of neural networks with dendrites to analyze temporal features. This simple neuron model may also be potentially useful in neural engineering applications.

The authors propose a learning rule for a neuron model with dendrite. In their model, somatodendritic interaction implements self-supervised learning applicable to a wide range of sequence learning tasks, including spike pattern detection, chunking temporal input and blind source separation.

Emergent Inference of Hidden Markov Models in Spiking Neural Networks Through Winner-Take-All

Hidden Markov models (HMMs) underpin the solution to many problems in computational neuroscience. However, it is still unclear how to implement inference of HMMs with a network of neurons in the brain. The existing methods suffer from the problem of being nonspiking and inaccurate. Here, we build a precise equivalence between the inference equation of HMMs with time-invariant hidden variables and the dynamics of spiking winner-take-all (WTA) neural networks. We show that the membrane potential of each spiking neuron in the WTA circuit encodes the logarithm of the posterior probability of the hidden variable in each state, and the firing rate of each neuron is proportional to the posterior probability of the HMMs. We prove that the time course of the neural firing rate can implement posterior inference of HMMs. Theoretical analysis and experimental results show that the proposed WTA circuit can get accurate inference results of HMMs.

Spiking time-dependent plasticity leads to efficient coding of predictions

Latency reduction in postsynaptic spikes is a well-known effect of spiking time-dependent plasticity. We expand this notion for long postsynaptic spike trains on single neurons, showing that, for a fixed input spike train, STDP reduces the number of postsynaptic spikes and concentrates the remaining ones. Then, we study the consequences of this phenomena in terms of coding, finding that this mechanism improves the neural code by increasing the signal-to-noise ratio and lowering the metabolic costs of frequent stimuli. Finally, we illustrate that the reduction in postsynaptic latencies can lead to the emergence of predictions.

An unsupervised neuromorphic clustering algorithm

Brains perform complex tasks using a fraction of the power that would be required to do the same on a conventional computer. New neuromorphic hardware systems are now becoming widely available that are intended to emulate the more power efficient, highly parallel operation of brains. However, to use these systems in applications, we need "neuromorphic algorithms" that can run on them. Here we develop a spiking neural network model for neuromorphic hardware that uses spike timing-dependent plasticity and lateral inhibition to perform unsupervised clustering. With this model, time-invariant, rate-coded datasets can be mapped into a feature space with a specified resolution, i.e., number of clusters, using exclusively neuromorphic hardware. We developed and tested implementations on the SpiNNaker neuromorphic system and on GPUs using the GeNN framework. We show that our neuromorphic clustering algorithm achieves results comparable to those of conventional clustering algorithms such as self-organizing maps, neural gas or k-means clustering. We then combine it with a previously reported supervised neuromorphic classifier network to demonstrate its practical use as a neuromorphic preprocessing module.

Synthesizing Images From Spatio-Temporal Representations Using Spike-Based Backpropagation

Spiking neural networks (SNNs) offer a promising alternative to current artificial neural networks to enable low-power event-driven neuromorphic hardware. Spike-based neuromorphic applications require processing and extracting meaningful information from spatio-temporal data, represented as series of spike trains over time. In this paper, we propose a method to synthesize images from multiple modalities in a spike-based environment. We use spiking auto-encoders to convert image and audio inputs into compact spatio-temporal representations that is then decoded for image synthesis. For this, we use a direct training algorithm that computes loss on the membrane potential of the output layer and back-propagates it by using a sigmoid approximation of the neuron's activation function to enable differentiability. The spiking autoencoders are benchmarked on MNIST and Fashion-MNIST and achieve very low reconstruction loss, comparable to ANNs. Then, spiking autoencoders are trained to learn meaningful spatio-temporal representations of the data, across the two modalities-audio and visual. We synthesize images from audio in a spike-based environment by first generating, and then utilizing such shared multi-modal spatio-temporal representations. Our audio to image synthesis model is tested on the task of converting TI-46 digits audio samples to MNIST images. We are able to synthesize images with high fidelity and the model achieves competitive performance against ANNs.

A Soft-Pruning Method Applied During Training of Spiking Neural Networks for In-memory Computing Applications

Inspired from the computational efficiency of the biological brain, spiking neural networks (SNNs) emulate biological neural networks, neural codes, dynamics, and circuitry. SNNs show great potential for the implementation of unsupervised learning using in-memory computing. Here, we report an algorithmic optimization that improves energy efficiency of online learning with SNNs on emerging non-volatile memory (eNVM) devices. We develop a pruning method for SNNs by exploiting the output firing characteristics of neurons. Our pruning method can be applied during network training, which is different from previous approaches in the literature that employ pruning on already-trained networks. This approach prevents unnecessary updates of network parameters during training. This algorithmic optimization can complement the energy efficiency of eNVM technology, which offers a unique in-memory computing platform for the parallelization of neural network operations. Our SNN maintains ~90% classification accuracy on the MNIST dataset with up to ~75% pruning, significantly reducing the number of weight updates. The SNN and pruning scheme developed in this work can pave the way toward applications of eNVM based neuro-inspired systems for energy efficient online learning in low power applications.

A structure-time parallel implementation of spike-based deep learning

Motivated by the recent progress of deep spiking neural networks (SNNs), we propose a structure-time parallel strategy based on layered structure and one-time computation over a time window to speed up the prominent spike-based deep learning algorithm named broadcast alignment. Furthermore, a well-designed deep hierarchical model based on the parallel broadcast alignment is proposed for object recognition. The parallel broadcast alignment achieves a significant 137× speedup compared to its original implementation on MNIST dataset. The object recognition model achieves higher accuracy than that of the latest spiking deep convolutional neural networks on the ETH-80 dataset. The proposed parallel strategy and the object recognition model will facilitate both the simulation of deep SNNs for studying spiking neural dynamics and also the applications of spike-based deep learning in real-world problems.

Hierarchical Bayesian Inference and Learning in Spiking Neural Networks

Numerous experimental data from neuroscience and psychological science suggest that human brain utilizes Bayesian principles to deal the complex environment. Furthermore, hierarchical Bayesian inference has been proposed as an appropriate theoretical framework for modeling cortical processing. However, it remains unknown how such a computation is organized in the network of biologically plausible spiking neurons. In this paper, we propose a hierarchical network of winner-take-all circuits which can carry out hierarchical Bayesian inference and learning through a spike-based variational expectation maximization (EM) algorithm. Particularly, we show how the firing activities of spiking neurons in response to the input stimuli and the spike-timing-dependent plasticity rule can be understood, respectively, as variational E-step and M-step of variational EM. Finally, we demonstrate the utility of this spiking neural network on the MNIST benchmark for unsupervised classification of handwritten digits.

Deep learning in spiking neural networks

In recent years, deep learning has revolutionized the field of machine learning, for computer vision in particular. In this approach, a deep (multilayer) artificial neural network (ANN) is trained, most often in a supervised manner using backpropagation. Vast amounts of labeled training examples are required, but the resulting classification accuracy is truly impressive, sometimes outperforming humans. Neurons in an ANN are characterized by a single, static, continuous-valued activation. Yet biological neurons use discrete spikes to compute and transmit information, and the spike times, in addition to the spike rates, matter. Spiking neural networks (SNNs) are thus more biologically realistic than ANNs, and are arguably the only viable option if one wants to understand how the brain computes at the neuronal description level. The spikes of biological neurons are sparse in time and space, and event-driven. Combined with bio-plausible local learning rules, this makes it easier to build low-power, neuromorphic hardware for SNNs. However, training deep SNNs remains a challenge. Spiking neurons' transfer function is usually non-differentiable, which prevents using backpropagation. Here we review recent supervised and unsupervised methods to train deep SNNs, and compare them in terms of accuracy and computational cost. The emerging picture is that SNNs still lag behind ANNs in terms of accuracy, but the gap is decreasing, and can even vanish on some tasks, while SNNs typically require many fewer operations and are the better candidates to process spatio-temporal data.

Neuroinspired unsupervised learning and pruning with subquantum CBRAM arrays

Resistive RAM crossbar arrays offer an attractive solution to minimize off-chip data transfer and parallelize on-chip computations for neural networks. Here, we report a hardware/software co-design approach based on low energy subquantum conductive bridging RAM (CBRAM®) devices and a network pruning technique to reduce network level energy consumption. First, we demonstrate low energy subquantum CBRAM devices exhibiting gradual switching characteristics important for implementing weight updates in hardware during unsupervised learning. Then we develop a network pruning algorithm that can be employed during training, different from previous network pruning approaches applied for inference only. Using a 512 kbit subquantum CBRAM array, we experimentally demonstrate high recognition accuracy on the MNIST dataset for digital implementation of unsupervised learning. Our hardware/software co-design approach can pave the way towards resistive memory based neuro-inspired systems that can autonomously learn and process information in power-limited settings.

Constraints of Metabolic Energy on the Number of Synaptic Connections of Neurons and the Density of Neuronal Networks

Neuronal networks in the brain are the structural basis of human cognitive function, and the plasticity of neuronal networks is thought to be the principal neural mechanism underlying learning and memory. Dominated by the Hebbian theory, researchers have devoted extensive effort to studying the changes in synaptic connections between neurons. However, understanding the network topology of all synaptic connections has been neglected over the past decades. Furthermore, increasing studies indicate that synaptic activities are tightly coupled with metabolic energy, and metabolic energy is a unifying principle governing neuronal activities. Therefore, the network topology of all synaptic connections may also be governed by metabolic energy. Here, by implementing a computational model, we investigate the general synaptic organization rules for neurons and neuronal networks from the perspective of energy metabolism. We find that to maintain the energy balance of individual neurons in the proposed model, the number of synaptic connections is inversely proportional to the average of the synaptic weights. This strategy may be adopted by neurons to ensure that the ability of neurons to transmit signals matches their own energy metabolism. In addition, we find that the density of neuronal networks is also an important factor in the energy balance of neuronal networks. An abnormal increase or decrease in the network density could lead to failure of energy metabolism in the neuronal network. These rules may change our view of neuronal networks in the brain and have guiding significance for the design of neuronal network models.

Deep Learning With Spiking Neurons: Opportunities and Challenges

Spiking neural networks (SNNs) are inspired by information processing in biology, where sparse and asynchronous binary signals are communicated and processed in a massively parallel fashion. SNNs on neuromorphic hardware exhibit favorable properties such as low power consumption, fast inference, and event-driven information processing. This makes them interesting candidates for the efficient implementation of deep neural networks, the method of choice for many machine learning tasks. In this review, we address the opportunities that deep spiking networks offer and investigate in detail the challenges associated with training SNNs in a way that makes them competitive with conventional deep learning, but simultaneously allows for efficient mapping to hardware. A wide range of training methods for SNNs is presented, ranging from the conversion of conventional deep networks into SNNs, constrained training before conversion, spiking variants of backpropagation, and biologically motivated variants of STDP. The goal of our review is to define a categorization of SNN training methods, and summarize their advantages and drawbacks. We further discuss relationships between SNNs and binary networks, which are becoming popular for efficient digital hardware implementation. Neuromorphic hardware platforms have great potential to enable deep spiking networks in real-world applications. We compare the suitability of various neuromorphic systems that have been developed over the past years, and investigate potential use cases. Neuromorphic approaches and conventional machine learning should not be considered simply two solutions to the same classes of problems, instead it is possible to identify and exploit their task-specific advantages. Deep SNNs offer great opportunities to work with new types of event-based sensors, exploit temporal codes and local on-chip learning, and we have so far just scratched the surface of realizing these advantages in practical applications.

STDP Allows Close-to-Optimal Spatiotemporal Spike Pattern Detection by Single Coincidence Detector Neurons

•

Repeating spike patterns exist and are informative. Can a single cell do the readout?

•

We show how a leaky integrate-and-fire (LIF) can do this readout optimally.

•

The optimal membrane time constant is short, possibly much shorter than the pattern.

•

Spike-timing-dependent plasticity (STDP) can turn a neuron into an optimal detector.

•

These results may explain how humans can learn repeating visual or auditory sequences.

Repeating spatiotemporal spike patterns exist and carry information. How this information is extracted by downstream neurons is unclear. Here we theoretically investigate to what extent a single cell could detect a given spike pattern and what the optimal parameters to do so are, in particular the membrane time constant τ. Using a leaky integrate-and-fire (LIF) neuron with homogeneous Poisson input, we computed this optimum analytically. We found that a relatively small τ (at most a few tens of ms) is usually optimal, even when the pattern is much longer. This is somewhat counter-intuitive as the resulting detector ignores most of the pattern, due to its fast memory decay. Next, we wondered if spike-timing-dependent plasticity (STDP) could enable a neuron to reach the theoretical optimum. We simulated a LIF equipped with additive STDP, and repeatedly exposed it to a given input spike pattern. As in previous studies, the LIF progressively became selective to the repeating pattern with no supervision, even when the pattern was embedded in Poisson activity. Here we show that, using certain STDP parameters, the resulting pattern detector is optimal. These mechanisms may explain how humans learn repeating sensory sequences. Long sequences could be recognized thanks to coincidence detectors working at a much shorter timescale. This is consistent with the fact that recognition is still possible if a sound sequence is compressed, played backward, or scrambled using 10-ms bins. Coincidence detection is a simple yet powerful mechanism, which could be the main function of neurons in the brain.

Optimal Localist and Distributed Coding of Spatiotemporal Spike Patterns Through STDP and Coincidence Detection

Repeating spatiotemporal spike patterns exist and carry information. Here we investigated how a single spiking neuron can optimally respond to one given pattern (localist coding), or to either one of several patterns (distributed coding, i.e., the neuron's response is ambiguous but the identity of the pattern could be inferred from the response of multiple neurons), but not to random inputs. To do so, we extended a theory developed in a previous paper (Masquelier, 2017), which was limited to localist coding. More specifically, we computed analytically the signal-to-noise ratio (SNR) of a multi-pattern-detector neuron, using a threshold-free leaky integrate-and-fire (LIF) neuron model with non-plastic unitary synapses and homogeneous Poisson inputs. Surprisingly, when increasing the number of patterns, the SNR decreases slowly, and remains acceptable for several tens of independent patterns. In addition, we investigated whether spike-timing-dependent plasticity (STDP) could enable a neuron to reach the theoretical optimal SNR. To this aim, we simulated a LIF equipped with STDP, and repeatedly exposed it to multiple input spike patterns, embedded in equally dense Poisson spike trains. The LIF progressively became selective to every repeating pattern with no supervision, and stopped discharging during the Poisson spike trains. Furthermore, tuning certain STDP parameters, the resulting pattern detectors were optimal. Tens of independent patterns could be learned by a single neuron using a low adaptive threshold, in contrast with previous studies, in which higher thresholds led to localist coding only. Taken together these results suggest that coincidence detection and STDP are powerful mechanisms, fully compatible with distributed coding. Yet we acknowledge that our theory is limited to single neurons, and thus also applies to feed-forward networks, but not to recurrent ones.

Neural Simpletrons: Learning in the Limit of Few Labels with Directed Generative Networks

We explore classifier training for data sets with very few labels. We investigate this task using a neural network for nonnegative data. The network is derived from a hierarchical normalized Poisson mixture model with one observed and two hidden layers. With the single objective of likelihood optimization, both labeled and unlabeled data are naturally incorporated into learning. The neural activation and learning equations resulting from our derivation are concise and local. As a consequence, the network can be scaled using standard deep learning tools for parallelized GPU implementation. Using standard benchmarks for nonnegative data, such as text document representations, MNIST, and NIST SD19, we study the classification performance when very few labels are used for training. In different settings, the network's performance is compared to standard and recently suggested semisupervised classifiers. While other recent approaches are more competitive for many labels or fully labeled data sets, we find that the network studied here can be applied to numbers of few labels where no other system has been reported to operate so far.

General differential Hebbian learning: Capturing temporal relations between events in neural networks and the brain

Learning in biologically relevant neural-network models usually relies on Hebb learning rules. The typical implementations of these rules change the synaptic strength on the basis of the co-occurrence of the neural events taking place at a certain time in the pre- and post-synaptic neurons. Differential Hebbian learning (DHL) rules, instead, are able to update the synapse by taking into account the temporal relation, captured with derivatives, between the neural events happening in the recent past. The few DHL rules proposed so far can update the synaptic weights only in few ways: this is a limitation for the study of dynamical neurons and neural-network models. Moreover, empirical evidence on brain spike-timing-dependent plasticity (STDP) shows that different neurons express a surprisingly rich repertoire of different learning processes going far beyond existing DHL rules. This opens up a second problem of how capturing such processes with DHL rules. Here we propose a general DHL (G-DHL) rule generating the existing rules and many others. The rule has a high expressiveness as it combines in different ways the pre- and post-synaptic neuron signals and derivatives. The rule flexibility is shown by applying it to various signals of artificial neurons and by fitting several different STDP experimental data sets. To these purposes, we propose techniques to pre-process the neural signals and capture the temporal relations between the neural events of interest. We also propose a procedure to automatically identify the rule components and parameters that best fit different STDP data sets, and show how the identified components might be used to heuristically guide the search of the biophysical mechanisms underlying STDP. Overall, the results show that the G-DHL rule represents a useful means to study time-sensitive learning processes in both artificial neural networks and brain.