Timing Dependent Plasticity Research Articles

Synaptic neural circuit inspired by side-gated graphene synaptic transistors for neuromorphic computing

Abstract With the rapid advancement of artificial intelligence, the energy consumption bottleneck inherent in the von Neumann computing architecture poses a significant obstacle to the future development of edge computing, artificial intelligence, and information technology. Consequently, it is crucial to develop synaptic neural circuits that exhibit memory and learning properties through synaptic plasticity. Drawing inspiration from the side-gated graphene synaptic transistor, we have designed a synaptic neural circuit comprising four key components: pre-voltage input, synaptic weight modulation, electric double-layer effect, and post-membrane current response. Through comprehensive simulations, we have successfully mimicked various synaptic behaviours, including long-term and short-term synaptic plasticity, paired-pulse facilitation, spiking-rate-dependent plasticity, spiking time-dependent plasticity, and Pavlovian associative learning. This approach establishes a robust framework for designing synaptic neural network circuits with advanced learning capabilities, thereby enhancing the practical applications of neural networks and machine learning.

Analytical solutions of two-liner tunnels constructed in time-dependent plastic rocks considering loading-unloading stress paths

Analytical solutions of two-liner tunnels constructed in time-dependent plastic rocks considering loading-unloading stress paths

Nonlinear Ion Dynamics Enable Spike Timing Dependent Plasticity of Electrochemical Ionic Synapses.

Programmable synaptic devices that can achieve timing-dependent weight updates are key components to implementing energy-efficient spiking neural networks (SNNs). Electrochemical ionic synapses (EIS) enable the programming of weight updates with very low energy consumption and low variability. Here, the strongly nonlinear kinetics of EIS, arising from nonlinear dynamics of ions and charge transfer reactions in solids, are leveraged to implement various forms of spike-timing-dependent plasticity (STDP). In particular, protons are used as the working ion. Different forms of the STDP function are deterministically predicted and emulated by a linear superposition of appropriately designed pre- and post-synaptic neuron signals. Heterogeneous STDP is also demonstrated within the array to capture different learning rules in the same system. STDP timescales are controllable, ranging from milliseconds to nanoseconds. The STDP resulting from EIS has lower variability than other hardware STDP implementations, due to the deterministic and uniform insertion of charge in the tunable channel material. The results indicate that the ion and charge transfer dynamics in EIS can enable bio-plausible synapses for SNN hardware with high energy efficiency, reliability, and throughput.

Resistive switching and synaptic characteristics in ZnO@β-SiC composite-based RRAM for neuromorphic computing

The advancement of neuromorphic computing in resistive random-access memory (RRAM) is crucial for the rapid expansion of artificial intelligence. Conventional metal oxide-based RRAM faces challenges in mimicking synaptic activity, leading to the exploration of new resistive switching (RS) materials. This study introduces a ZnO@β-SiC composite-based RRAM device that exhibits biological synapse-like functionality. The device shows self-compliance and forming-free RS at ∼0.8 V, where it also mimics synaptic responses such as potentiation, depression, and paired-pulse facilitation at low voltage stimuli (∼0.6 V, 40 ms) with learning and forgetting behavior. Moreover, the synaptic plasticity is analyzed through spike rate dependent plasticity, spike number dependent plasticity, and spike time dependent plasticity. Further, the transition from short-term plasticity to long-term plasticity is observed under more training pulses and lower interval stimuli. The observed RS mechanism and synaptic functionalities are explained by the electric field-driven formation and dissolution of conducting filaments of oxygen vacancies. The chemical properties and local electronic structure have been examined by x-ray photoelectron spectroscopy and x-ray absorption spectroscopy. To elucidate the atomistic memristive behavior and the contribution of different electrical parameters in RRAM, detailed conductive atomic-force microscopy and impedance analysis have been carried out.

Astrocytes Mediate Psychostimulant-Induced Alterations ofSpike-Timing Dependent Synaptic Plasticity.

At cellular and circuit levels, drug addiction is considered a dysregulation of synaptic plasticity. In addition, dysfunction of the glutamate transporter 1 (GLT-1) in the nucleus accumbens (NAc) has also been proposed as a mechanism underlying drug addiction. However, the cellular and synaptic impact of GLT-1 alterations in the NAc remain unclear. Here we show in the NAc that 10 days withdraw after 5 days treatment with cocaine or amphetamine decreases GLT-1 expression in astrocytes, which results in the prolongation of the excitatory postsynaptic potential (EPSP) decay kinetics in D1 receptor-containing medium spiny neurons (D1R-MSNs). Using the spike timing dependent plasticity (STDP) paradigm, we found that enlargement of EPSP duration results in switching the LTP elicited in control animals to LTD in psychostimulant-treated mice. In contrast to D1-MSNs, D2-MSNs did not display changes in EPSP kinetics and synaptic plasticity. Notably, the psychostimulant-induced synaptic transmission and synaptic plasticity effects were absent in IP3R2-/- mice, which lack astrocyte calcium signal, but were mimicked by the selective astrocytes stimulation with DREADDs. Finally, ceftriaxone, which upregulates GLT-1, restored normal GLT-1 function, EPSP kinetics, and synaptic plasticity in psychostimulant-treated mice. Therefore, we propose that cocaine and amphetamine increase dopaminergic levels in the NAc, which stimulates astrocytes and downregulates the GLT-1. The decreased GLT-1 function prolonged the EPSP kinetics, leading to the modulation of the STDP, transforming the LTP observed in control animals into LTD in psychostimulant-treated mice. Present work reveals a novel mechanism underlying the synaptic plasticity changes induced by these drugs of abuse.

Consciousness driven Spike Timing Dependent Plasticity

Consciousness driven Spike Timing Dependent Plasticity

An ultra low frequency spike timing dependent plasticity based approach for reducing alcohol drinking

Alcohol use disorder (AUD) is a chronic relapsing brain disorder characterized by an impaired ability to stop or control alcohol consumption despite adverse social, occupational, or health consequences. AUD affects nearly one-third of adults at some point during their lives, with an associated cost of approximately $249 billion annually in the U.S. alone. The effects of alcohol consumption are expected to increase significantly during the COVID-19 pandemic, with alcohol sales increasing by approximately 54%, potentially exacerbating health concerns and risk-taking behaviors. Unfortunately, existing pharmacological and behavioral therapies for AUD are associated with poor success rates, with approximately 40% of individuals relapsing within three years of treatment.Pre-clinical studies have shown that chronic alcohol consumption leads to significant changes in synaptic function within the dorsal medial striatum (DMS), one of the brain regions associated with AUD and responsible for mediating goal-directed behavior. Specifically, chronic alcohol consumption has been associated with hyperactivity of dopamine receptor 1 (D1) medium spiny neurons (MSN) and hypoactivity of dopamine receptor 2 (D1) MSNs within the DMS. Optogenetic, chemogenetic, and transgenic approaches have demonstrated that reducing the D1/D2 MSN signaling imbalance decreases alcohol self-administration in rodent models of AUD.Here, we present an electrical stimulation approach that uses ultra-low (≤ 1 Hz) frequency (ULF) spike-timing-dependent plasticity (STDP) in mouse models of AUD to reduce DMS D1/D2 MSN signaling imbalances by stimulating D1-MSN afferents into the GPi and ACC glutamatergic projections to the DMS in a time-locked stimulation sequence. Our data suggest that GPi/ACC ULF-STDP selectively decreases DMS D1-MSN hyperactivity leading to reduced alcohol consumption without evoking undesired affective behaviors using electrical stimulation rather than approaches requiring genetic modification. This work represents a step towards fulfilling the unmet need for a reliable method of treating severe AUD through cell-type-specific control with clinically available neuromodulation tools.

Development of Biodegradable Substrates and Synaptic Transistors for Next‐Generation Transient Electronics

AbstractUbiquitous electronic gadgets in lives have led to an increase in electronic waste (e‐waste), posing a threat to the environment and ecology that must be addressed. This work demonstrates the use of gelatin, a natural protein, for development of flexible biodegradable substrates and synaptic transistors using the same material as gate dielectric. The fabricated p‐channel transistors exhibit high electrical stability and exceptional synaptic characteristics through spike timing dependent plasticity (STDP), spike voltage dependent plasticity (SVDP), and spike number dependent plasticity (SNDP), respectively upon variation of post‐synaptic current (PSC) with time, amplitude, and number of stimuli. These devices exhibit pulse paired facilitation (PPF) with relaxation time constants in the range of ≈10 ms and regulating modulation amplitude of 1 greatly resembling a biological synapse. Study on the variability among distinct devices and over multiple cycles demonstrate outstanding repeatability of synaptic plasticity. The devices showcase significant PSC values with almost linear SNDP, while consuming an ultralow power of ≈11.7 fJ. Excellent stability is observed when subjected to multiple bending sequences. Complete dissolution of these devices in aqueous environments in an hour without any alteration to temperature or pH confirms excellent biodegradability of these devices leading toward transient neuromorphic circuits and systems that adhere to the concepts of circular economy.

High-Performance Memristive Synapse Based on Space-Charge-Limited Conduction in LiNbO3.

Advancing neuromorphic computing technology requires the development of versatile synaptic devices. In this study, we fabricated a high-performance Al/LiNbO3/Pt memristive synapse and emulated various synaptic functions using its primary key operating mechanism, known as oxygen vacancy-mediated valence charge migration (VO-VCM). The voltage-controlled VO-VCM induced space-charge-limited conduction and self-rectifying asymmetric hysteresis behaviors. Moreover, the device exhibited voltage pulse-tunable multi-state memory characteristics because the degree of VO-VCM was dependent on the applied pulse parameters (e.g., polarity, amplitude, width, and interval). As a result, synaptic functions such as short-term memory, dynamic range-tunable long-term memory, and spike time-dependent synaptic plasticity were successfully demonstrated by modulating those pulse parameters. Additionally, simulation studies on hand-written image pattern recognition confirmed that the present device performed with high accuracy, reaching up to 95.2%. The findings suggest that the VO-VCM-based Al/LiNbO3/Pt memristive synapse holds significant promise as a brain-inspired neuromorphic device.

Inhibition SNN: unveiling the efficacy of various lateral inhibition learning in image pattern recognition

This study presents an enhanced spiking neural network (SNN) with inhibition, referred to as Inhibition SNN, which advances the field of pattern recognition through efficient event-driven computation. We incorporate a winner-take-all (WTA) mechanism into leaky integrate-and-fire neurons. This, combined with spike timing dependent plasticity, simulates the biological learning process. We propose two network prototypes, namely Inhibition V1 and Inhibition V2. Inhibition V1 differentiates between an excitatory layer and an inhibitory layer, while Inhibition V2 uses self-connections in the excitatory layer to implement WTA dynamics, enhancing signal contrast without a separate inhibitory layer. Our experiments on the MNIST dataset show that both Inhibition V1 and Inhibition V2 can match the accuracy of current advanced networks, with Inhibition V2 performing better. An unsupervised learning network of two layers (784–100) achieves 86% accuracy on MNIST and 61% on Fashion-MNIST. The scalability and robustness of these networks in neural computation underscore their potential for practical applications. This research offers a fresh perspective on SNN design, highlighting the impact of inhibition on learning efficiency. With plans to adapt these networks to neuromorphic hardware, Inhibition SNN could pave the way for energy-efficient intelligent systems.

Sustainable vertically-oriented graphene-electrode memristors for neuromorphic applications

Neuromorphic computing, an innovative field in electronic and computing engineering, aims to enhance computing paradigms by simulating brain processes. Memristors, a two-terminal device, hold promise in revolutionising neuromorphic architectures by circumventing the Von-Neumann bottleneck. The performance and applicability of memristors heavily rely on the materials and fabrication processes employed. Graphene exhibits unique properties that can be leveraged in memristor design. Moreover, graphene stands out as a material with the potential for large-scale, sustainable production through Plasma Enhanced Chemical Vapour Deposition (PECVD). Notably, the properties of graphene-electrode memristors vary with minor structural differences induced by different PECVD temperatures. This paper reports the synthesis of graphene electrodes by time- and cost-effective PECVD from a sustainable plant extract for memristors. In addition, this paper delves into investigating how these structural variations impact the properties of graphene memristors and explores their potential exploitation in neuromorphic applications for implementing the well-known Spike Timing Dependent Plasticity (STDP) learning mechanism. The paper also utilises the developed STDP learning to perform an unsupervised spike-based pattern classification task.

A Self-Driven Ga2O3 Memristor Synapse for Humanoid Robot Learning.

In recent years, the rapid development of brain-inspired neuromorphic systems has created an imperative demand for artificial photonic synapses that operate with low power consumption. In this study, a self-driven memristor synapse based on gallium oxide (Ga2O3) nanowires is proposed and demonstrated successfully. This memristor synapse is capable of emulating a range of functionalities of biological synapses when exposed to 255nm light stimulation. These functionalities encompass peak time-dependent plasticity, pulse facilitation, and memory learning capabilities. It exhibits an ultrahigh paired-pulse facilitation index of 158, indicating exceptional learning performance. The transition from short-term memory to long-term memory can be attributed to the remarkable relearning capabilities. Furthermore, the potential applications of the memristor synapse is showcased through the successful manipulation of a humanoid intelligent robot. Upon establishing artificial intelligence (AI) systems, the control commands originating from the synaptic device can drive the humanoid robot to perform various actions. Based on the memristor synapses, the autonomous feedback system of the humanoid robot facilitates a good collaboration between robotic actions and bio-inspired light perception. Therefore, this research opens up an effective way to advance the development of neuromorphic computing technologies, AI systems, and intelligent robots that demand ultra-low energy consumption.

Supervised learning of spatial features with STDP and homeostasis using Spiking Neural Networks on SpiNNaker

Artificial Neural Networks (ANN) have gained significant popularity thanks to their ability to learn using the well-known backpropagation algorithm. Conversely, Spiking Neural Networks (SNNs), despite having broader capabilities than ANNs, have always posed challenges in the training phase. This paper shows a new method to perform supervised learning on SNNs, using Spike Timing Dependent Plasticity (STDP) and homeostasis, aiming at training the network to identify spatial patterns. Spatial patterns refer to spike patterns without a time component, where all spike events occur simultaneously. The method is tested using the SpiNNaker digital architecture. A SNN is trained to recognise one or multiple patterns and performance metrics are extracted to measure the performance of the network. Some considerations are drawn from the results showing that, in the case of a single trained pattern, the network behaves as the ideal detector, with 100% accuracy in detecting the trained pattern. However, as the number of trained patterns on a single network increases, the accuracy of identification is linked to the similarities between these patterns. This method of training an SNN to detect spatial patterns may be applied to pattern recognition in static images or traffic analysis in computer networks, where each network packet represents a spatial pattern. It will be stipulated that the homeostatic factor may enable the network to detect patterns with some degree of similarity, rather than only perfectly matching patterns. The principles outlined in this article serve as the fundamental building blocks for more complex systems that utilise both spatial and temporal patterns by converting specific features of input signals into spikes. One example of such a system is a computer network packet classifier, tasked with real-time identification of packet streams based on features within the packet content.

Flexible Artificial Ag NPs:a-SiC0.11:H Synapse on Al Foil with High Uniformity and On/Off Ratio for Neuromorphic Computing.

A neuromorphic computing network based on SiCx memristor paves the way for a next-generation brain-like chip in the AI era. Up to date, the SiCx-based memristor devices are faced with the challenge of obtaining flexibility and uniformity, which can push forward the application of memristors in flexible electronics. For the first time, we report that a flexible artificial synaptic device based on a Ag NPs:a-SiC0.11:H memristor can be constructed by utilizing aluminum foil as the substrate. The device exhibits stable bipolar resistive switching characteristic even after bending 1000 times, displaying excellent flexibility and uniformity. Furthermore, an on/off ratio of approximately 107 can be obtained. It is found that the incorporation of silver nanoparticles significantly enhances the device's set and reset voltage uniformity by 76.2% and 69.7%, respectively, which is attributed to the contribution of the Ag nanoparticles. The local electric field of Ag nanoparticles can direct the formation and rupture of conductive filaments. The fitting results of I-V curves show that the carrier transport mechanism agrees with Poole-Frenkel (P-F) model in the high-resistance state, while the carrier transport follows Ohm's law in the low-resistance state. Based on the multilevel storage characteristics of the Al/Ag NPs:a-SiC0.11:H/Al foil resistive switching device, we successfully observed the biological synaptic characteristics, including the long-term potentiation (LTP), long-term depression (LTD), and spike-timing-dependent plasticity (STDP). The flexible artificial Ag NPs:a-SiC0.11:H/Al foil synapse possesses excellent conductance modulation capabilities and visual learning function, demonstrating the promise of application in flexible electronics technology for high-efficiency neuromorphic computing in the AI period.

Parallel hyperparameter optimization of spiking neural networks

Hyperparameter optimization of spiking neural networks (SNNs) is a difficult task which has not yet been deeply investigated in the literature. In this work, we designed a scalable constrained Bayesian based optimization algorithm that prevents sampling in non-spiking areas of an efficient high dimensional search space. These search spaces contain infeasible solutions that output no or only a few spikes during the training or testing phases, we call such a mode a “silent network”. Finding them is difficult, as many hyperparameters are highly correlated to the architecture and to the dataset. We leverage silent networks by designing a spike-based early stopping criterion to accelerate the optimization process of SNNs trained by spike timing dependent plasticity and surrogate gradient. We parallelized the optimization algorithm asynchronously, and ran large-scale experiments on heterogeneous multi-GPU Petascale architecture. Results show that by considering silent networks, we can design more flexible high-dimensional search spaces while maintaining a good efficacy. The optimization algorithm was able to focus on networks with high performances by preventing costly and worthless computation of silent networks.

Neuromorphic learning and recognition in WO3−x thin film-based forming-free flexible electronic synapses

In pursuing advanced neuromorphic applications, this study introduces the successful engineering of a flexible electronic synapse based on WO3-x, structured as W/WO3-x/Pt/Muscovite-Mica. This artificial synapse is designed to emulate crucial learning behaviors fundamental to in-memory computing. We systematically explore synaptic plasticity dynamics by implementing pulse measurements capturing potentiation and depression traits akin to biological synapses under flat and different bending conditions, thereby highlighting its potential suitability for flexible electronic applications. The findings demonstrate that the memristor accurately replicates essential properties of biological synapses, including short-term plasticity (STP), long-term plasticity (LTP), and the intriguing transition from STP to LTP. Furthermore, other variables are investigated, such as paired-pulse facilitation, spike rate-dependent plasticity, spike time-dependent plasticity, pulse duration-dependent plasticity, and pulse amplitude-dependent plasticity. Utilizing data from flat and differently bent synapses, neural network simulations for pattern recognition tasks using the Modified National Institute of Standards and Technology dataset reveal a high recognition accuracy of ∼95% with a fast learning speed that requires only 15 epochs to reach saturation.

Tunable memory behavior in light stimulated artificial synapse based on ZnO thin film transistors

Abstract Optoelectronic synapses are inevitable for realizing neuromorphic vision systems, which require the integration of image recognition, memory and image processing into a single platform. In this work, we present a three terminal optoelectronic synapse created using zinc oxide (ZnO) thin film transistor. The persistent photoconductivity (PPC) of ZnO thin film is utilized to demonstrate the synaptic behaviour. The change in conductance of the device under UV illumination has been interpreted as the weight change in the synapse. The basic synaptic functions such as sensory memory, short term memory, long term memory, duration time dependent plasticity and paired pulse facilitation have been successfully demonstrated. The device shows a PPF index of 160\%, comparable to other optoelectronic synapses reported in literature. Further, to corroborate the existing theory that PPC is caused by oxygen vacancies, additional characterizations are carried out and the presence of oxygen vacancies is detected in the fabricated ZnO device. Subsequently, pattern recognition of MNIST handwritten dataset has been performed using the conductance tuning curves of the proposed ZnO TFT based synapses in a neural network architecture, thereby demonstrating their feasibility to be used in neuromorphic applications.

Assessment of mineral compositions on geo-mechanical time dependent plastic creep deformation

Understanding rock mechanical time-dependent plastic deformation enables the researchers to accurately predict rock responses during loading and unloading. This research focuses on the geomechanical aspects of hydrogen storage, particularly rock resistance to creep deformation under varying stresses and pressures. Modelling creep deformation in underground storage mediums has been overlooked and requires further consideration, as it is a critical, cyclical, time-dependent phenomenon essential for hydrogen containment. The study's novelty is demonstrated in its focus on geomechanical properties of rock minerals, providing insights that aid in site-selection criteria for hydrogen storage A creep deformation model with varying mineral compositions is used to assess creep deformation strain. In homogenous cases, calcite showed the highest resistance to creep deformation after three loading cycles, primarily due to its high Young's Modulus. As for the heterogenous cases, the results indicated that the mixture of calcite and quartz was the most creep-resistant rock composition. In terms of rock types, basalt samples demonstrated the highest resistance to creep deformation, while siltstone exhibited the highest susceptibility to deformation. It was observed that as the number of cycles increase, the rock mechanical stability decreases, hence optimizing the number of cycles is a must to ensure hydrogen geo-containment.

Memristor based on carbon nanotube gelatin composite film as artificial optoelectronic synapse for image processing

Photoelectric artificial synapses based on memristors is an effective method to realize neuromorphic computation. This study presents an optoelectronic responsive artificial synapse made of a composite material consisting of gelatin and carbon nanotubes. The memristor demonstrates characteristics of analog resistive switching, the ability to store multiple memory states, and impressive retention properties. It has the capability to induce an excitatory post-synaptic current by means of electrical pulses or pulsed light exposure. The excitatory post-synaptic current can be modulated by the number, amplitude and interval of electrical pulses, as well as the action time, interval and light intensity of optical pulses. The artificial synapse showcases the emulation of fundamental Hebbian learning protocols, including spike timing dependent plasticity and spike amplitude dependent plasticity. In addition, the charge transfer in the carbon nanotube gelatin composite optoelectronic memristor is investigated through first-principles calculations, shedding light on its operational mechanism. Experimental results show that these devices have the potential to be utilized for processing image information, resulting in a significant reduction of input data and training expenses when recognizing handwritten numbers. Overall, the optoelectronic synapse exhibits promising image processing prospects in the field of neuromorphic computing.

Spike-enhanced synapse functions of SnOx-based resistive memory

This paper explores the application of resistive random-access memory (RRAM) devices in emulating various synapse functions. The electrical properties of an indium tin oxide (ITO)/SnOx/TaN device are examined, including the activation process, switching process, endurance (>102), and retention characteristics (104 s). A conduction mechanism for the resistive-switching process based on the migration of oxygen ions in the insulating SnOx film is proposed. Furthermore, the multilevel cell characteristics of the device, demonstrating the control of resistance states by variations in reset voltage and compliance current, were found to be suitable, increasing its data storage capacity. Synapse functions, including potentiation and depression, are tested, and their repeatability is evaluated. A neural network-based Modified National Institute Standards and Technology pattern recognition system is applied to the ITO/SnOx/TaN device, showcasing its capabilities in pattern recognition tasks. This study further explores various synapse functions, such as paired-pulse facilitation, paired-pulse depression, excitatory postsynaptic current, and activity-dependent synaptic plasticity. Hebbian learning rules, including spike rate-dependent plasticity, are demonstrated, along with spike time-dependent plasticity under various spike types, making it suitable for high-functionality neuromorphic applications. Finally, we assembled 3 × 3 synapse arrays and utilized them to form recognizable patterns, demonstrating the potential of the device in implementing neural network functions.