Spiking Neural Networks Based on OxRAM Synapses for Real-Time Unsupervised Spike Sorting

DualSort: online spike sorting with a running neural network

Objective.Spike sorting, i.e. the detection and separation of measured action potentials from different extracellularly recorded neurons, remains one of the bottlenecks in deciphering the brain. In recent years, the application of neural networks (NNs) for spike sorting has garnered significant attention. Most methods focus on specific sub-problems within the conventional spike sorting pipeline, such as spike detection or feature extraction, and attempt to solve them with complex network architectures. This paper presents DualSort, a simple NN that gets combined with downstream post-processing for real-time spike sorting. It shows high efficiency, low complexity, and requires a comparatively small amount of human interaction.Approach.Synthetic and experimentally obtained extracellular single-channel recordings were utilized to train and evaluate the proposed NN. For training, spike waveforms were labeled with respect to their associated neuron and position in the signal, allowing the detection and categorization of spikes in unison. DualSort classifies a single spike multiple times in succession, as it runs over the signal in a step-by-step manner and uses a post-processing algorithm that transmits the network output into spike trains. Main results.With the used datasets, DualSort was able to detect and distinguish different spike waveforms and separate them from background activity. The post-processing algorithm significantly strengthened the overall performance of the model, making the system more robust as a whole. Although DualSort is an end-to-end solution that efficiently transforms filtered signals into spike trains, it competes with contemporary state-of-the-art technologies that exclusively target single sub-problems in the conventional spike sorting pipeline.Significance.This work demonstrates that even under high noise levels, complex NNs are not necessary by any means to achieve high performance in spike detection and sorting. The utilization of data augmentation on a limited quantity of spikes could substantially decrease hand-labeling compared to other studies. Furthermore, the proposed framework can be utilized without human interaction when combined with an unsupervised technique that provides pseudo labels for DualSort. Due to the low complexity of our network, it works efficiently and enables real-time processing on basic hardware. The proposed approach is not limited to spike sorting, as it may also be used to process different signals, such as electroencephalogram (EEG), which needs to be investigated in future research.

NeuSort: an automatic adaptive spike sorting approach with neuromorphic models

Objective.Spike sorting, a critical step in neural data processing, aims to classify spiking events from single electrode recordings based on different waveforms. This study aims to develop a novel online spike sorter, NeuSort, using neuromorphic models, with the ability to adaptively adjust to changes in neural signals, including waveform deformations and the appearance of new neurons.Approach.NeuSort leverages a neuromorphic model to emulate template-matching processes. This model incorporates plasticity learning mechanisms inspired by biological neural systems, facilitating real-time adjustments to online parameters.Results.Experimental findings demonstrate NeuSort's ability to track neuron activities amidst waveform deformations and identify new neurons in real-time. NeuSort excels in handling non-stationary neural signals, significantly enhancing its applicability for long-term spike sorting tasks. Moreover, its implementation on neuromorphic chips guarantees ultra-low energy consumption during computation.Significance.NeuSort caters to the demand for real-time spike sorting in brain-machine interfaces through a neuromorphic approach. Its unsupervised, automated spike sorting process makes it a plug-and-play solution for online spike sorting.

Electroencephalogram based brain-computer interface: Applications, challenges, and opportunities

Brain-Computer Interfaces (BCI) is an exciting and emerging research area for researchers and scientists. It is a suitable combination of software and hardware to operate any device mentally. This review emphasizes the significant stages in the BCI domain, current problems, and state-of-the-art findings. This article also covers how current results can contribute to new knowledge about BCI, an overview of BCI from its early developments to recent advancements, BCI applications, challenges, and future directions. The authors pointed to unresolved issues and expressed how BCI is valuable for analyzing the human brain. Humans' dependence on machines has led humankind into a new future where BCI can play an essential role in improving this modern world.

A robust spike sorting method based on the joint optimization of linear discrimination analysis and density peaks

Spike sorting is a fundamental step in extracting single-unit activity from neural ensemble recordings, which play an important role in basic neuroscience and neurotechnologies. A few algorithms have been applied in spike sorting. However, when noise level or waveform similarity becomes relatively high, their robustness still faces a big challenge. In this study, we propose a spike sorting method combining Linear Discriminant Analysis (LDA) and Density Peaks (DP) for feature extraction and clustering. Relying on the joint optimization of LDA and DP: DP provides more accurate classification labels for LDA, LDA extracts more discriminative features to cluster for DP, and the algorithm achieves high performance after iteration. We first compared the proposed LDA-DP algorithm with several algorithms on one publicly available simulated dataset and one real rodent neural dataset with different noise levels. We further demonstrated the performance of the LDA-DP method on a real neural dataset from non-human primates with more complex distribution characteristics. The results show that our LDA-DP algorithm extracts a more discriminative feature subspace and achieves better cluster quality than previously established methods in both simulated and real data. Especially in the neural recordings with high noise levels or waveform similarity, the LDA-DP still yields a robust performance with automatic detection of the number of clusters. The proposed LDA-DP algorithm achieved high sorting accuracy and robustness to noise, which offers a promising tool for spike sorting and facilitates the following analysis of neural population activity.

From End to End: Gaining, Sorting, and Employing High-Density Neural Single Unit Recordings

The meaning behind neural single unit activity has constantly been a challenge, so it will persist in the foreseeable future. As one of the most sourced strategies, detecting neural activity in high-resolution neural sensor recordings and then attributing them to their corresponding source neurons correctly, namely the process of spike sorting, has been prevailing so far. Support from ever-improving recording techniques and sophisticated algorithms for extracting worthwhile information and abundance in clustering procedures turned spike sorting into an indispensable tool in electrophysiological analysis. This review attempts to illustrate that in all stages of spike sorting algorithms, the past 5 years innovations' brought about concepts, results, and questions worth sharing with even the non-expert user community. By thoroughly inspecting latest innovations in the field of neural sensors, recording procedures, and various spike sorting strategies, a skeletonization of relevant knowledge lays here, with an initiative to get one step closer to the original objective: deciphering and building in the sense of neural transcript.

Classification of Whisker Deflections From Evoked Responses in the Somatosensory Barrel Cortex With Spiking Neural Networks

Spike-based neuromorphic hardware has great potential for low-energy brain-machine interfaces, leading to a novel paradigm for neuroprosthetics where spiking neurons in silicon read out and control activity of brain circuits. Neuromorphic processors can receive rich information about brain activity from both spikes and local field potentials (LFPs) recorded by implanted neural probes. However, it was unclear whether spiking neural networks (SNNs) implemented on such devices can effectively process that information. Here, we demonstrate that SNNs can be trained to classify whisker deflections of different amplitudes from evoked responses in a single barrel of the rat somatosensory cortex. We show that the classification performance is comparable or even superior to state-of-the-art machine learning approaches. We find that SNNs are rather insensitive to recorded signal type: both multi-unit spiking activity and LFPs yield similar results, where LFPs from cortical layers III and IV seem better suited than those of deep layers. In addition, no hand-crafted features need to be extracted from the data—multi-unit activity can directly be fed into these networks and a simple event-encoding of LFPs is sufficient for good performance. Furthermore, we find that the performance of SNNs is insensitive to the network state—their performance is similar during UP and DOWN states.

Engineering Spiking Neurons Using Threshold Switching Devices for High-Efficient Neuromorphic Computing

Inspired by the human brain, the spike-based neuromorphic system has attracted strong research enthusiasm because of the high energy efficiency and powerful computational capability, in which the spiking neurons and plastic synapses are two fundamental building blocks. Recently, two-terminal threshold switching (TS) devices have been regarded as promising candidates for building spiking neurons in hardware. However, how circuit parameters affect the spiking behavior of TS-based neurons is still an open question. Here, based on a leaky integrate-and-fire (LIF) neuron circuit, we systematically study the effect of both the extrinsic and intrinsic factors of NbO x -based TS neurons on their spiking behaviors. The extrinsic influence factors contain input intensities, connected synaptic weights, and parallel capacitances. To illustrate the effect of intrinsic factors, including the threshold voltage, holding voltage, and high/low resistance states of NbO x devices, we propose an empirical model of the fabricated NbO x devices, fitting well with the experimental results. The results indicate that with enhancing the input intensity, the spiking frequency increases first then decreases after reaching a peak value. Except for the connected synaptic weights, all other parameters can modulate the spiking peak frequency under high enough input intensity. Also, the relationship between energy consumption per spike and frequency of the neuron cell is further studied, leading guidance to design neuron circuits in a system to obtain the lowest energy consumption. At last, to demonstrate the practical applications of TS-based neurons, we construct a spiking neural network (SNN) to control the cart-pole using reinforcement learning, obtaining a reward score up to 450. This work provides valuable guidance on building compact LIF neurons based on TS devices and further bolsters the construction of high-efficiency neuromorphic systems.

Stimuli-Responsive Memristive Materials for Artificial Synapses and Neuromorphic Computing

Neuromorphic computing holds promise for building next-generation intelligent systems in a more energy-efficient way than the conventional von Neumann computing architecture. Memristive hardware, which mimics biological neurons and synapses, offers high-speed operation and low power consumption, enabling energy- and area-efficient, brain-inspired computing. Here, recent advances in memristive materials and strategies that emulate synaptic functions for neuromorphic computing are highlighted. The working principles and characteristics of biological neurons and synapses, which can be mimicked by memristive devices, are presented. Besides device structures and operation with different external stimuli such as electric, magnetic, and optical fields, how memristive materials with a rich variety of underlying physical mechanisms can allow fast, reliable, and low-power neuromorphic applications is also discussed. Finally, device requirements are examined and a perspective on challenges in developing memristive materials for device engineering and computing science is given.

Neural Network Training Acceleration With RRAM-Based Hybrid Synapses

Hardware neural network (HNN) based on analog synapse array excels in accelerating parallel computations. To implement an energy-efficient HNN with high accuracy, high-precision synaptic devices and fully-parallel array operations are essential. However, existing resistive memory (RRAM) devices can represent only a finite number of conductance states. Recently, there have been attempts to compensate device nonidealities using multiple devices per weight. While there is a benefit, it is difficult to apply the existing parallel updating scheme to the synaptic units, which significantly increases updating process’s cost in terms of computation speed, energy, and complexity. Here, we propose an RRAM-based hybrid synaptic unit consisting of a “big” synapse and a “small” synapse, and a related training method. Unlike previous attempts, array-wise fully-parallel learning is possible with our proposed architecture with a simple array selection logic. To experimentally verify the hybrid synapse, we exploit Mo/TiO_x RRAM, which shows promising synaptic properties and areal dependency of conductance precision. By realizing the intrinsic gain via proportionally scaled device area, we show that the big and small synapse can be implemented at the device-level without modifications to the operational scheme. Through neural network simulations, we confirm that RRAM-based hybrid synapse with the proposed learning method achieves maximum accuracy of 97 %, comparable to floating-point implementation (97.92%) of the software even with only 50 conductance states in each device. Our results promise training efficiency and inference accuracy by using existing RRAM devices.

Hybrid oxide brain-inspired neuromorphic devices for hardware implementation of artificial intelligence

The state-of-the-art artificial intelligence technologies mainly rely on deep learning algorithms based on conventional computers with classical von Neumann computing architectures, where the memory and processing units are separated resulting in an enormous amount of energy and time consumed in the data transfer process. Inspired by the human brain acting like an ultra-highly efficient biological computer, neuromorphic computing is proposed as a technology for hardware implementation of artificial intelligence. Artificial synapses are the main component of a neuromorphic computing architecture. Memristors are considered to be a relatively ideal candidate for artificial synapse applications due to their high scalability and low power consumption. Oxides are most widely used in memristors due to the ease of fabrication and high compatibility with complementary metal-oxide-semiconductor processes. However, oxide memristors suffer from unsatisfactory stability and reliability. Oxide-based hybrid structures can effectively improve the device stability and reliability, therefore providing a promising prospect for the application of oxide memristors to neuromorphic computing. This work reviews the recent advances in the development of hybrid oxide memristive synapses. The discussion is organized according to the blending schemes as well as the working mechanisms of hybrid oxide memristors.

Memristor-Based Edge Detection for Spike Encoded Pixels

Memristors have many uses in machine learning and neuromorphic hardware. From memory elements in dot product engines to replicating both synapse and neuron wall behaviors, the memristor has proved a versatile component. Here we demonstrate an analog mode of operation observed in our silicon oxide memristors and apply this to the problem of edge detection. We demonstrate how a potential divider exploiting this analog behavior can prove a scalable solution to edge detection. We confirm its behavior experimentally and simulate its performance on a standard testbench. We show good performance comparable to existing memristor based work with a benchmark score of 0.465 on the BSDS500 dataset, while simultaneously maintaining a lower component count.

On-Line Recursive Decomposition of Intramuscular EMG Signals Using GPU-Implemented Bayesian Filtering

OBJECTIVE

Real-time intramuscular electromyography (iEMG) decomposition, which is needed in biofeedback studies and interfacing applications, is a complex procedure that involves identifying the motor neuron spike trains from a streaming iEMG recording.

METHODS

We have previously proposed a sequential decomposition algorithm based on a Hidden Markov Model of EMG, which used Bayesian filter to estimate unknown parameters of motor unit (MU) spike trains, as well as their action potentials (MUAPs). Here, we present a modification of this original model in order to achieve a real-time performance of the algorithm as well as a parallel computation implementation of the algorithm on Graphics Processing Unit (GPU). Specifically, the Kalman filter previously used to estimate the MUAPs, is replaced by a least-mean-square filter. Additionally, we introduce a number of heuristics that help to omit the most improbable decomposition scenarios while searching for the best solution. Then, a GPU-implementation of the proposed algorithm is presented.

RESULTS

Simulated iEMG signals containing up to 10 active MUs, as well as five experimental fine-wire iEMG signals acquired from the tibialis anterior muscle, were decomposed in real time. The accuracy of decompositions depended on the level of muscle activation, but in all cases exceeded 85 %.

CONCLUSION

The proposed method and implementation provide an accurate, real-time interface with spinal motor neurons.

SIGNIFICANCE

The presented real time implementation of the decomposition algorithm substantially broadens the domain of its application.

Bridging Biological and Artificial Neural Networks with Emerging Neuromorphic Devices: Fundamentals, Progress, and Challenges

As the research on artificial intelligence booms, there is broad interest in brain-inspired computing using novel neuromorphic devices. The potential of various emerging materials and devices for neuromorphic computing has attracted extensive research efforts, leading to a large number of publications. Going forward, in order to better emulate the brain's functions, its relevant fundamentals, working mechanisms, and resultant behaviors need to be re-visited, better understood, and connected to electronics. A systematic overview of biological and artificial neural systems is given, along with their related critical mechanisms. Recent progress in neuromorphic devices is reviewed and, more importantly, the existing challenges are highlighted to hopefully shed light on future research directions.

An Attention-Based Spiking Neural Network for Unsupervised Spike-Sorting

Bio-inspired computing using artificial spiking neural networks promises performances outperforming currently available computational approaches. Yet, the number of applications of such networks remains limited due to the absence of generic training procedures for complex pattern recognition, which require the design of dedicated architectures for each situation. We developed a spike-timing-dependent plasticity (STDP) spiking neural network (SSN) to address spike-sorting, a central pattern recognition problem in neuroscience. This network is designed to process an extracellular neural signal in an online and unsupervised fashion. The signal stream is continuously fed to the network and processed through several layers to output spike trains matching the truth after a short learning period requiring only few data. The network features an attention mechanism to handle the scarcity of action potential occurrences in the signal, and a threshold adaptation mechanism to handle patterns with different sizes. This method outperforms two existing spike-sorting algorithms at low signal-to-noise ratio (SNR) and can be adapted to process several channels simultaneously in the case of tetrode recordings. Such attention-based STDP network applied to spike-sorting opens perspectives to embed neuromorphic processing of neural data in future brain implants.

Artificial Neuron and Synapse Realized in an Antiferromagnet/Ferromagnet Heterostructure Using Dynamics of Spin-Orbit Torque Switching

Efficient information processing in the human brain is achieved by dynamics of neurons and synapses, motivating effective implementation of artificial spiking neural networks. Here, the dynamics of spin-orbit torque switching in antiferromagnet/ferromagnet heterostructures is studied to show the capability of the material system to form artificial neurons and synapses for asynchronous spiking neural networks. The magnetization switching, driven by a single current pulse or trains of pulses, is examined as a function of the pulse width (1 s to 1 ns), amplitude, number, and pulse-to-pulse interval. Based on this dynamics and the unique ability of the system to exhibit binary or analog behavior depending on the device size, key functionalities of a synapse (spike-timing-dependent plasticity) and a neuron (leaky integrate-and-fire) are reproduced in the same material and on the basis of the same working principle. These results open a way toward spintronics-based neuromorphic hardware that executes cognitive tasks with the efficiency of the human brain.

Non-invasive Decoding of the Motoneurons: A Guided Source Separation Method Based on Convolution Kernel Compensation With Clustered Initial Points

Despite the progress in understanding of neural codes, the studies of the cortico-muscular coupling still largely rely on interferential electromyographic (EMG) signal or its rectification for the assessment of motor neuron pool behavior. This assessment is non-trivial and should be used with precaution. Direct analysis of neural codes by decomposing the EMG, also known as neural decoding, is an alternative to EMG amplitude estimation. In this study, we propose a fully-deterministic hybrid surface EMG (sEMG) decomposition approach that combines the advantages of both template-based and Blind Source Separation (BSS) decomposition approaches, a.k.a. guided source separation (GSS), to identify motor unit (MU) firing patterns. We use the single-pass density-based clustering algorithm to identify possible cluster representatives in different sEMG channels. These cluster representatives are then used as initial points of modified gradient Convolution Kernel Compensation (gCKC) algorithm. Afterwards, we use the Kalman filter to reduce the noise impact and increase convergence rate of MU filter identification by gCKC. Moreover, we designed an adaptive soft-thresholding method to identify MU firing times out of estimated MU spike trains. We tested the proposed algorithm on a set of synthetic sEMG signals with known MU firing patterns. A grid of 9 × 10 monopolar surface electrodes with 5-mm inter-electrode distances in both directions was simulated. Muscle excitation was set to 10, 30, and 50%. Colored Gaussian zero-mean noise with the signal-to-noise ratio (SNR) of 10, 20, and 30 dB, respectively, was added to 16 s long sEMG signals that were sampled at 4,096 Hz. Overall, 45 simulated signals were analyzed. Our decomposition approach was compared with gCKC algorithm. Overall, in our algorithm, the average numbers of identified MUs and Rate-of-Agreement (RoA) were 16.41 ± 4.18 MUs and 84.00 ± 0.06%, respectively, whereas the gCKC identified 12.10 ± 2.32 MUs with the average RoA of 90.78 ± 0.08%. Therefore, the proposed GSS method identified more MUs than the gCKC, with comparable performance. Its performance was dependent on the signal quality but not the signal complexity at different force levels. The proposed algorithm is a promising new offline tool in clinical neurophysiology.

Spike sorting based on shape, phase, and distribution features, and K-TOPS clustering with validity and error indices

Spike sorting is one of the most important data analysis problems in neurophysiology. The precision in all steps of the spike-sorting procedure critically affects the accuracy of all subsequent analyses. After data preprocessing and spike detection have been carried out properly, both feature extraction and spike clustering are the most critical subsequent steps of the spike-sorting procedure. The proposed spike sorting approach comprised a new feature extraction method based on shape, phase, and distribution features of each spike (hereinafter SS-SPDF method), which reveal significant information of the neural events under study. In addition, we applied an efficient clustering algorithm based on K-means and template optimization in phase space (hereinafter K-TOPS) that included two integrative clustering measures (validity and error indices) to verify the cohesion-dispersion among spike events during classification and the misclassification of clustering, respectively. The proposed method/algorithm was tested on both simulated data and real neural recordings. The results obtained for these datasets suggest that our spike sorting approach provides an efficient way for sorting both single-unit spikes and overlapping waveforms. By analyzing raw extracellular recordings collected from the rostral-medial prefrontal cortex (rmPFC) of behaving rabbits during classical eyeblink conditioning, we have demonstrated that the present method/algorithm performs better at classifying spikes and neurons and at assessing their modulating properties than other methods currently used in neurophysiology.

Nonvolatile Memory Materials for Neuromorphic Intelligent Machines

Recent progress in deep learning extends the capability of artificial intelligence to various practical tasks, making the deep neural network (DNN) an extremely versatile hypothesis. While such DNN is virtually built on contemporary data centers of the von Neumann architecture, physical (in part) DNN of non-von Neumann architecture, also known as neuromorphic computing, can remarkably improve learning and inference efficiency. Particularly, resistance-based nonvolatile random access memory (NVRAM) highlights its handy and efficient application to the multiply-accumulate (MAC) operation in an analog manner. Here, an overview is given of the available types of resistance-based NVRAMs and their technological maturity from the material- and device-points of view. Examples within the strategy are subsequently addressed in comparison with their benchmarks (virtual DNN in deep learning). A spiking neural network (SNN) is another type of neural network that is more biologically plausible than the DNN. The successful incorporation of resistance-based NVRAM in SNN-based neuromorphic computing offers an efficient solution to the MAC operation and spike timing-based learning in nature. This strategy is exemplified from a material perspective. Intelligent machines are categorized according to their architecture and learning type. Also, the functionality and usefulness of NVRAM-based neuromorphic computing are addressed.