Spike-timing-dependent Plasticity Research Articles

An unsupervised STDP-based spiking neural network inspired by biologically plausible learning rules and connections

The backpropagation algorithm has promoted the rapid development of deep learning, but it relies on a large amount of labeled data and still has a large gap with how humans learn. The human brain can quickly learn various conceptual knowledge in a self-organized and unsupervised manner, accomplished through coordinating various learning rules and structures in the human brain. Spike-timing-dependent plasticity (STDP) is a general learning rule in the brain, but spiking neural networks (SNNs) trained with STDP alone is inefficient and perform poorly. In this paper, taking inspiration from short-term synaptic plasticity, we design an adaptive synaptic filter and introduce the adaptive spiking threshold as the neuron plasticity to enrich the representation ability of SNNs. We also introduce an adaptive lateral inhibitory connection to adjust the spikes balance dynamically to help the network learn richer features. To speed up and stabilize the training of unsupervised spiking neural networks, we design a samples temporal batch STDP (STB-STDP), which updates weights based on multiple samples and moments. By integrating the above three adaptive mechanisms and STB-STDP, our model greatly accelerates the training of unsupervised spiking neural networks and improves the performance of unsupervised SNNs on complex tasks. Our model achieves the current state-of-the-art performance of unsupervised STDP-based SNNs in the MNIST and FashionMNIST datasets. Further, we tested on the more complex CIFAR10 dataset, and the results fully illustrate the superiority of our algorithm. Our model is also the first work to apply unsupervised STDP-based SNNs to CIFAR10. At the same time, in the small-sample learning scenario, it will far exceed the supervised ANN using the same structure.

Wi-Fi frame detection via spiking neural networks with memristive synapses

With increasing performance of deep learning, researchers have employed Deep Neural Networks (DNNs) for wireless communications. In particular, mechanisms for detecting Wi-Fi frames that use DNNs demonstrate their excellent performances in terms of detection accuracy. However, DNNs require significant amount of computation resources. Thus, if the DNN-based mechanisms are used in mobile devices or low-end devices, their battery would be quickly depleted. Spiking Neural Networks (SNNs), which are regarded as the next generation of neural network, have advantages over DNNs: low energy consumption and limited computational complexity. Motivated by these advantages, in this paper, we propose a mechanism to detect a Wi-Fi frame using SNNs and show the feasibility of SNNs for Wi-Fi detection. The mechanism is composed of a preprocessing module for converting collected signals and an SNN module for detecting Wi-Fi frames. The SNN module employs Leaky Integrate-and-Fire (LIF) neurons and Spike-Timing-Dependent Plasticity (STDP) learning rule. To reflect the features of an actual neuromorphic hardware system, our SNN module considers memristive synaptic features such as non-linear weight updates. Our experimental study demonstrates the detection capabilities of the proposed mechanism are comparable to those of previous mechanisms that use DNNs, CNNs and RNNs while consuming definitely less energy than the previous mechanism.

A superconducting synapse exhibiting spike-timing dependent plasticity

The recent success of artificial intelligence (AI) systems has been accompanied by a rapid increase in the computational resources needed to successfully train them. This rate of increase threatens the future development of AI systems as they are presently configured. Unsupervised learning, where systems are trained online instead of through offline computation, offers a possible way forward. Here, we present the design of a synaptic circuit made from superconducting electronics capable of spike-timing dependent plasticity (STDP), a form of unsupervised learning. The synapse is constructed from three sub-circuits, each responsible for a part of the synaptic action. We demonstrate the operation of the synapse through numerical simulation and show that it reproduces the hallmark behaviors of STDP. Combined with existing superconducting neuromorphic components like neurons and axons, this synaptic structure could help form a fast, powerful, and energy-efficient Spiking Neural Network.

Highly uniform resistive switching characteristics of Ti/TaOx/ITO memristor devices for neuromorphic system

In this study, we focused on the uniformity of resistance states of Ti/TaOx/ITO devices and the possibility of using them in neuromorphic applications under DC and pulse measurement conditions. The thickness and chemical composition of the devices was verified by transmission electron microscopy (TEM). First, the I-V curves of the devices were controlled with the compliance current and reset voltage. In addition, the multi-level characteristics were demonstrated through DC sweeps and pulses for high-density memory and neuromorphic systems. The linearity of potentiation and depression was then improved to seek high accuracy of pattern recognition in a neural network. Finally, spike-timing-dependent plasticity (STDP) was performed (including potentiation and depression) to mimic Hebbian learning of the nerve system.

Modeling and Analysis of Spike Signal Sequence for Memristor Crossbar Array in Neuromorphic Chips

This paper presents the efficient systematic methods for modeling and analysis of spike signal sequence in crossbar arrays for neuromorphic computing chips. A novel spike signal sequence is proposed, where the ideal spike sequence with only spike time information in the original spiking neural network (SNN) algorithm is mapped onto actual spike waveform by stitching neighboring sequential spikes together with certain overlaps. We thoroughly investigate and analyze the performance of the input encoding as well as the implementation of spike timing dependent plasticity (STDP)-based SNN on memristor crossbar arrays with the proposed spike signal sequence. A detailed circuit model of a crossbar array, consisting of resistance, capacitance and inductance derived by the partial equivalent element circuit (PEEC) method, is created to simulate the training process of SNN. The proposed spike signal sequence is demonstrated that is able to achieve accurate input encoding as well as high recognition accuracy when it is used to perform the classification task on MNIST handwritten digits. The spike signal sequence is further analyzed and assessed in terms of the main factors affecting its encoding accuracy and the parasitic effects of crossbar arrays on its robustness.

A Bio-Plausible Approach to Realizing Heat-Evoked Nociceptive Withdrawal Reflex on the Upper Limb of a Humanoid Robot

In this letter, we present a method for realizing the heat-evoked nociceptive withdrawal reflex (NWR) in the upper limb of a humanoid robot so that it can avoid the potential damage caused by noxious heat. We use a spiking neural network whose structure, encoding scheme, and form of information transmission mimic the reflex arc in humans to improve bio-plausibility. The proper synaptic strengths between the sensory neurons and the interneurons in the first two layers are learned using the bio-plausible reward-modulated spike timing-dependent plasticity learning algorithm. By monitoring the spikes from the motor neuron in the third layer, a reflex matching the intensity of the stimulation can be evoked. Experimental evaluations show that noxious heat stimulation can be detected online and evoke the NWR. The experiments on a full-size humanoid robot show that the method enables robots to avoid potential damage robustly with proper NWR, depending on the site and intensity of the stimulation. We also verify that the method takes advantage of the intrinsic characteristics of its neuromorphic encoding scheme to reproduce essential features of the NWR, e.g., spatial summation effect and temporal summation effect in humans. The improved bio-plausibility and the capability to reproduce the human-like feature.

Multi-neuron connection using multi-terminal floating–gate memristor for unsupervised learning

Multi-terminal memristor and memtransistor (MT-MEMs) has successfully performed complex functions of heterosynaptic plasticity in synapse. However, theses MT-MEMs lack the ability to emulate membrane potential of neuron in multiple neuronal connections. Here, we demonstrate multi-neuron connection using a multi-terminal floating-gate memristor (MT-FGMEM). The variable Fermi level (EF) in graphene allows charging and discharging of MT-FGMEM using horizontally distant multiple electrodes. Our MT-FGMEM demonstrates high on/off ratio over 105 at 1000 s retention about ~10,000 times higher than other MT-MEMs. The linear behavior between current (ID) and floating gate potential (VFG) in triode region of MT-FGMEM allows for accurate spike integration at the neuron membrane. The MT-FGMEM fully mimics the temporal and spatial summation of multi-neuron connections based on leaky-integrate-and-fire (LIF) functionality. Our artificial neuron (150 pJ) significantly reduces the energy consumption by 100,000 times compared to conventional neurons based on silicon integrated circuits (11.7 μJ). By integrating neurons and synapses using MT-FGMEMs, a spiking neurosynaptic training and classification of directional lines functioned in visual area one (V1) is successfully emulated based on neuron’s LIF and synapse’s spike-timing-dependent plasticity (STDP) functions. Simulation of unsupervised learning based on our artificial neuron and synapse achieves a learning accuracy of 83.08% on the unlabeled MNIST handwritten dataset.

Biocompatible AlOOH-based memristor with biomimicking synaptic functions for artificial nociceptor applications

Electronic devices with biomimetic synaptic characteristics are one of the most critical components in health monitoring and artificial intelligence. The design of neuromorphic responsive devices, however, yet remains a ground challenge. Herein, a biocompatible and electrically responsive boehmite material (γ-AlOOH) is utilized to fabricate artificial memristors to realize mimicking biological synapsefunctions. The obtained AlOOH memristor is flexible, highly transparent, and biocompatible, exhibiting nonvolatile resistive switching behaviours with low threshold voltages, high ON-OFF ratios (>103), good switching durability (>104 cycles), and long retention time (>104 s). Specifically, the device can mimic the biosynaptic functions, such as paired-pulse facilitation (PPF), long-term potentiation/depression (LTP/LTD), amplitude-dependent plasticity (SADP), and spike time dependent plasticity (STDP), and the typical sensitization, desensitization, hyperalgesia, and allodynia behaviours of the biological nociceptors. This study reveals that the AlOOH memristor with excellent artificial synapses is resulted by the high proton mobility along with the OH bond length variation of AlOOH subjected to applied stimuli, such as voltages, pulse width, and pulse intervals. Therefore, the biocompatible AlOOH-based memristors with artificial synapse behaviour have promising potential for the applications in high-performance neuromorphic devices and future artificial intelligence.

Biologically plausible information propagation in a complementary metal-oxide semiconductor integrate-and-fire artificial neuron circuit with memristive synapses

Neuromorphic circuits based on spikes are currently envisioned as a viable option to achieve brain-like computation capabilities in specific electronic implementations while limiting power dissipation given their ability to mimic energy-efficient bioinspired mechanisms. While several network architectures have been developed to embed in hardware the bioinspired learning rules found in the biological brain, such as spike timing-dependent plasticity, it is still unclear if hardware spiking neural network architectures can handle and transfer information akin to biological networks. In this work, we investigate the analogies between an artificial neuron combining memristor synapses and rate-based learning rule with biological neuron response in terms of information propagation from a theoretical perspective. Bioinspired experiments have been reproduced by linking the biological probability of release with the artificial synapse conductance. Mutual information and surprise have been chosen as metrics to evidence how, for different values of synaptic weights, an artificial neuron allows to develop a reliable and biological resembling neural network in terms of information propagation and analysis.

Implementation of threshold- and memory-switching memristors based on electrochemical metallization in an identical ferroelectric electrolyte

The use of an identical electrolyte in electrochemical metallization (ECM)-based neuron and synaptic devices has not yet been achieved due to their different resistive-switching characteristics. Herein, we describe ECM devices comprising the same ferroelectric PbZr0.52Ti0.48O3 (PZT) electrolyte, which can sustain both neuron and synaptic behavior depending on the identity of the active electrode. The Ag/PZT/La0.8Sr0.2MnO3 (LSMO) threshold switching memristor shows abrupt and volatile resistive switching characteristics, which lead to neuron devices with stochastic integration-and-fire behavior, auto-recovery, and rapid operation. In contrast, the Ni/PZT/LSMO memory switching memristor exhibits gradual, non-volatile resistive switching behavior, which leads to synaptic devices with a high on/off ratio, low on-state current, low variability, and spike-timing-dependent plasticity (STDP). The divergent behavior of the ECM devices is attributed to greater control of cation migration through the ultrathin ferroelectric PZT. Thus, ECM devices with an identical ferroelectric electrolyte offer promise as essential building blocks in the construction of high-performance neuromorphic computing systems.

Learning cortical hierarchies with temporal Hebbian updates.

A key driver of mammalian intelligence is the ability to represent incoming sensory information across multiple abstraction levels. For example, in the visual ventral stream, incoming signals are first represented as low-level edge filters and then transformed into high-level object representations. Similar hierarchical structures routinely emerge in artificial neural networks (ANNs) trained for object recognition tasks, suggesting that similar structures may underlie biological neural networks. However, the classical ANN training algorithm, backpropagation, is considered biologically implausible, and thus alternative biologically plausible training methods have been developed such as Equilibrium Propagation, Deep Feedback Control, Supervised Predictive Coding, and Dendritic Error Backpropagation. Several of those models propose that local errors are calculated for each neuron by comparing apical and somatic activities. Notwithstanding, from a neuroscience perspective, it is not clear how a neuron could compare compartmental signals. Here, we propose a solution to this problem in that we let the apical feedback signal change the postsynaptic firing rate and combine this with a differential Hebbian update, a rate-based version of classical spiking time-dependent plasticity (STDP). We prove that weight updates of this form minimize two alternative loss functions that we prove to be equivalent to the error-based losses used in machine learning: the inference latency and the amount of top-down feedback necessary. Moreover, we show that the use of differential Hebbian updates works similarly well in other feedback-based deep learning frameworks such as Predictive Coding or Equilibrium Propagation. Finally, our work removes a key requirement of biologically plausible models for deep learning and proposes a learning mechanism that would explain how temporal Hebbian learning rules can implement supervised hierarchical learning.

Python-Based Circuit Design for Fundamental Building Blocks of Spiking Neural Network

Spiking neural networks (SNNs) are considered a crucial research direction to address the “storage wall” and “power wall” challenges faced by traditional artificial intelligence computing. However, developing SNN chips based on CMOS (complementary metal oxide semiconductor) circuits remains a challenge. Although memristor process technology is the best alternative to synapses, it is still undergoing refinement. In this study, a novel approach is proposed that employs tools to automatically generate HDL (hardware description language) code for constructing neuron and memristor circuits after using Python to describe the neuron and memristor models. Based on this approach, HR (Hindmash–Rose), LIF (leaky integrate-and-fire), and IZ (Izhikevich) neuron circuits, as well as HP, EG (enhanced generalized), and TB (the behavioral threshold bipolar) memristor circuits are designed to construct the most basic connection of a SNN: the neuron–memristor–neuron circuit that satisfies the STDP (spike-timing-dependent-plasticity) learning rule. Through simulation experiments and FPGA (field programmable gate array) prototype verification, it is confirmed that the IZ and LIF circuits are suitable as neurons in SNNs, while the X variables of the EG memristor model serve as characteristic synaptic weights. The EG memristor circuits best satisfy the STDP learning rule and are suitable as synapses in SNNs. In comparison to previous works on hardware spiking neurons, the proposed method needed fewer area resources for creating spiking neurons models on FPGA. The proposed SNN basic components design method, and the resulting circuits, are beneficial for architectural exploration and hardware–software co-design of SNN chips.

Reconfigurable Low-Power TiO2 Memristor for Integration of Artificial Synapse and Nociceptor

Bio-mimetic advanced electronic systems are emerging rapidly, engrossing their applications in neuromorphic computing, humanoid robotics, tactile sensors, and so forth. The biological synaptic and nociceptive functions are governed by intricate neurotransmitter dynamics that involve both short-term and long-term plasticity. To emulate the neuronal dynamics in an electronic device, an Ag/TiO2/Pt/SiO2/Si memristor is fabricated, exhibiting compliance current controlled reversible transition of volatile switching (VS) and non-volatile switching (NVS). The origin of the VS and NVS depends on the diameter of the conducting filament, which is explained using a field-induced nucleation theory and validated by temporal current response measurements. The switching delay of the device is used to determine the characteristic nociceptive behaviors such as threshold, relaxation, inadaptation, allodynia, and hyperalgesia. The short-term and long-term retention loss attributed to the VS and NVS, respectively, is used to emulate short-term memory and long-term memory of the biological brain in a single device. More importantly, synergistically modulating the VS-NVS transition, the complex spike rate-dependent (SRDP) and spike time-dependent plasticity (STDP) with a weight change of up to 600% is demonstrated in the same device, which is the highest reported so far for TiO2 memristors. Furthermore, the device exhibits very low power consumption, ∼3.76 pJ/spike, and can imitate synaptic and nociceptive functions. The consolidation of complex nociceptive and synaptic behavior in a single memristor facilitates low-power integration of scalable intelligent sensors and neuromorphic devices.

Effect of Al Concentration on Ferroelectric Properties in HfAlOx‐Based Ferroelectric Tunnel Junction Devices for Neuroinspired Applications

Since HfOx‐based ferroelectric tunnel junctions (FTJs) are attractive compared to perovskite‐based FTJs and other emerging memory devices, they are being actively studied recently. They have advantages such as a simple metal–insulator–metal structure, complementary metal oxide semiconductor (CMOS) compatibility, non‐destructive operation, and low power consumption. Moreover, doped HfOx‐based FTJs are in the spotlight in terms of neuromorphic engineering as a way of advancing from the von Neumann structure. In particular, Al dopant is effective for inducing ferroelectric properties due to its smaller radius than that of Hf. The optimal concentration of Al varies depending on the device materials and the annealing conditions during deposition. Therefore, in‐depth research is required for neuromorphic applications. Herein, the properties of FTJ devices according to Al doping concentrations are analyzed. Subsequently, using the device with the highest remanent polarization, neuromorphic applications are implemented, including spike‐timing‐dependent plasticity (STDP), paired‐pulse facilitation (PPF), long‐term potentiation, and depression. The characteristics in different frequency regions are also studied to satisfy the demand for fast switching. Finally, the FTJ device is used as a physical reservoir in reservoir computing for efficient processing of time‐dependent inputs.

Spiking neural predictive coding for continually learning from data streams

For energy-efficient computation in specialized neuromorphic hardware, we present spiking neural coding, an instantiation of a family of artificial neural models grounded in the theory of predictive coding. This model, the first of its kind, works by operating in a never-ending process of “guess-and-check”, where neurons predict the activity values of one another and then adjust their own activities to make better future predictions. The interactive, iterative nature of our system fits well into the continuous time formulation of sensory stream prediction and, as we show, the model’s structure yields a local synaptic update rule, which can be used to complement or as an alternative to online spike-timing dependent plasticity. In this article, we experiment with an instantiation of our model consisting of leaky integrate-and-fire units. However, the framework within which our system is situated can naturally incorporate more complex neurons such as the Hodgkin-Huxley model. Our experimental results in pattern recognition demonstrate the potential of the model when binary spike trains are the primary paradigm for inter-neuron communication. Notably, spiking neural coding is competitive in terms of classification performance and experiences less forgetting when learning from a task sequence, offering a more computationally economical, biologically-motivated alternative to popular artificial neural networks.

Biological UV Photoreceptors‐Inspired Sn‐Doped Polycrystalline β‐Ga2O3 Optoelectronic Synaptic Phototransistor for Neuromorphic Computing

AbstractIn this study, the authors fabricate Sn‐doped 100‐nm thick polycrystalline β‐Ga2O3 synaptic field‐effect transistors (FETs) emulating optical and electrical spike stimulation. When stimulated by deep ultraviolet (UV) optical spikes or electric voltage spikes at the gate, the devices exhibit several essential synaptic functions of excitatory‐postsynaptic currents (EPSCs), inhibitory‐postsynaptic currents (IPSCs), paired‐pulse facilitation (PPF), spike‐number‐dependent plasticity (SNDP), and spike‐timing‐dependent plasticity (STDP). Following UV optical stimulation, the devices mimic synaptic plasticity with a photogate effect, and the gate voltage stimulation emulates the synaptic weights according to the state of the gate dielectric interface. The β‐Ga2O3 synaptic FET demonstrates synergistic functions in various optoelectronic stimulation modes and successfully mimics the visual memory formation in bees with UV photoreceptors. Moreover, to verify the translation of optoelectrical‐derived synaptic behaviors of β‐Ga2O3 synaptic FETs into artificial neuromorphic computing, handwritten digit image recognition of the Modified National Institute of Standards and Technology dataset is performed using a convolutional neural network, and a learning accuracy of 96.92% is achieved. The realization of these fundamental functions of biological synapses suggests the utility of Ga2O3‐based optoelectronic devices for next‐generation neuromorphic computing.

Influence of Lithium Doping on Volcanic-like Perovskite Memristors and Artificial Synaptic Simulation for Neurocomputing

Perovskite-based memristors have attracted much attention in synaptic simulation due to their outstanding electrical properties and promising potential in neuromorphic computing (NC). In this work, inorganic lead-free perovskite-based memristors composed of Ag/Cs3Bi2–xLixI9–2x (CBLxI)/ITO (x = 0, 0.2, 0.4, 0.6) are fabricated, and the electrical properties, such as endurance, on/off ratio, and retention time, are determined. It is found that the device with x = 0.4 shows good characteristics, such as a set voltage of −0.1 V and a retention time of 104 s. The multilevel storage performance is investigated, and multiple synaptic characteristics, such as paired-pulse facilitation (PPF), spike-voltage-dependent plasticity (SVDP), spike-width-dependent plasticity (SWDP), spike-timing-dependent plasticity (STDP), and learning–forgetting, are simulated. The conductive mechanism of the device is analyzed and discussed with an analogy to natural volcanic rocks, which also have a large surface area, high adsorption, and high chemical inertness. An artificial neural network (ANN) based on the potentiation/depression characteristics is designed and analyzed theoretically, and a pattern recognition rate of 94.25% is accomplished. The strategy and results described in this paper provide insights into the development of nonvolatile memory devices boding well for the adoption of neuromorphic computing for image recognition.

Effects of cortico-cortical paired associative stimulation based on multisensory integration to brain network connectivity in stroke patients: study protocol for a randomized doubled blind clinical trial

Introduction: Brain has a spontaneous recovery after stroke, reflecting the plasticity of the brain. Currently, TMS is used for studies of single-target brain region modulation, which lacks consideration of brain networks and functional connectivity. Cortico-cortical paired associative stimulation (ccPAS) promotes recovery of motor function. Multisensory effects in primary visual cortex(V1) directly influence behavior and perception, which facilitate motor functional recovery in stroke patients. Therefore, in this study, dual-targeted precise stimulation of V1 and primary motor cortex(M1) on the affected hemisphere of stroke patients will be used for cortical visuomotor multisensory integration to improve motor function.MethodThis study is a randomized, double-blind controlled clinical trial over a 14-week period. 69 stroke subjects will be enrolled and divided into sham stimulation group, ccPAS low frequency group, and ccPAS high frequency group. All groups will receive conventional rehabilitation. The intervention lasted for two weeks, five times a week. Assessments will be performed before the intervention, at the end of the intervention, and followed up at 6 and 14 weeks. The primary assessment indicator is the ‘Fugl-Meyer Assessment of the Upper Extremity ’, secondary outcomes were ‘The line bisection test’, ‘Modified Taylor Complex Figure’, ‘NIHSS’ and neuroimaging assessments. All adverse events will be recorded.DiscussionCurrently, ccPAS is used for the modulation of neural circuits. Based on spike-timing dependent plasticity theory, we can precisely intervene in the connections between different cortices to promote the recovery of functional connectivity on damaged brain networks after stroke. We hope to achieve the modulation of cortical visuomotor interaction by combining ccPAS with the concept of multisensory integration. We will further analyze the correlation between analyzing visual and motor circuits and explore the alteration of neuroplasticity by the interactions between different brain networks. This study will provide us with a new clinical treatment strategy to achieve precise rehabilitation for patient with motor dysfunction after stroke.Trial registrationThis trial was registered in the Chinese Clinical Trial Registry with code ChiCTR2300067422 and was approved on January 16, 2023.

Selective control of synaptic plasticity in heterogeneous networks through transcranial alternating current stimulation (tACS)

Transcranial alternating current stimulation (tACS) represents a promising non-invasive treatment for an increasingly wide range of neurological and neuropsychiatric disorders. The ability to use periodically oscillating electric fields to non-invasively engage neural dynamics opens up the possibility of recruiting synaptic plasticity and to modulate brain function. However, despite consistent reports about tACS clinical effectiveness, strong state-dependence combined with the ubiquitous heterogeneity of cortical networks collectively results in high outcome variability. Introducing variations in intrinsic neuronal timescales, we explored how such heterogeneity influences stimulation-induced change in synaptic connectivity. We examined how spike timing dependent plasticity, at the level of cells, intra- and inter-laminar cortical networks, can be selectively and preferentially engaged by periodic stimulation. Using leaky integrate-and-fire neuron models, we analyzed cortical circuits comprised of multiple cell-types, alongside superficial multi-layered networks expressing distinct layer-specific timescales. Our results show that mismatch in neuronal timescales within and/or between cells-and the resulting variability in excitability, temporal integration properties and frequency tuning-enables selective and directional control on synaptic connectivity by tACS. Our work provides new vistas on how to recruit neural heterogeneity to guide brain plasticity using non-invasive stimulation paradigms.

Ultrafast and Low-Power 2D Bi2O2Se Memristors for Neuromorphic Computing Applications

Memristors that emulate synaptic plasticity are building blocks for opening a new era of energy-efficient neuromorphic computing architecture, which will overcome the limitation of the von Neumann bottleneck. Layered two-dimensional (2D) Bi2O2Se, as an emerging material for next-generation electronics, is of great significance in improving the efficiency and performance of memristive devices. Herein, high-quality Bi2O2Se nanosheets are grown by configuring mica substrates face-down on the Bi2O2Se powder. Then, bipolar Bi2O2Se memristors are fabricated with excellent performance including ultrafast switching speed (<5 ns) and low-power consumption (<3.02 pJ). Moreover, synaptic plasticity, such as long-term potentiation/depression (LTP/LTD), paired-pulse facilitation (PPF), and spike-timing-dependent plasticity (STDP), are demonstrated in the Bi2O2Se memristor. Furthermore, MNIST recognition with simulated artificial neural networks (ANN) based on conductance modification could reach a high accuracy of 91%. Notably, the 2D Bi2O2Se enables the memristor to possess ultrafast and low-power attributes, showing great potential in neuromorphic computing applications.