Heterogeneous neural network based on locally active memristor with multiple firing patterns

IF 2.5 3区工程技术 Q3 COMPUTER SCIENCE, HARDWARE & ARCHITECTURE

Integration-The Vlsi Journal Pub Date : 2025-08-04 DOI:10.1016/j.vlsi.2025.102490

Ke Meng, Yinghong Cao, Xianying Xu, Jun Mou

{"title":"Heterogeneous neural network based on locally active memristor with multiple firing patterns","authors":"Ke Meng, Yinghong Cao, Xianying Xu, Jun Mou","doi":"10.1016/j.vlsi.2025.102490","DOIUrl":null,"url":null,"abstract":"<div><div>With the development of information technology and neurobiology, there is an urgent need for a memory element with bionic properties to simulate the interactions among neurons. On this basis, a novel locally active memristor (LAM) model is designed, of which the memory properties are utilized to construct a coupled system of three-dimensional Hopfield neural network (HNN) and Hindmarsh–Rose (HR) neurons to simulate neuronal activities. First, the nonvolatility and local activity of the memristor is verified by the power-off plot (POP) and its direct current (DC) <span><math><mi>V</mi></math></span> - <span><math><mi>I</mi></math></span> plot. The device exhibits typical bistable resistance-switching behavior, maintaining two stable resistance states upon power-off, which confirms its non-volatile memory characteristics. A significant negative differential resistance (NDR) region observed in DC <span><math><mi>V</mi></math></span> - <span><math><mi>I</mi></math></span> curves directly verifies its local activity, indicating potential for active signal processing. Second, the complex dynamical behavior of HNN-HR is probed by numerical simulation, adjusting the coupling strength, synaptic weights and HR neuron parameters to demonstrate the bionic properties. The research results show that not only are multiple hidden attractor structures exhibited by the model, but also typical nonlinear phenomena such as transient chaos and intermittent chaos can be reproduced by it, and the dynamic transition between different chaotic firing modes can be realized. In addition, the phenomena of multi- state coexisting attractors and the expansion and migration of attractor topological structures are observed in the model. Finally, by means of the TMS320F28335 digital signal processing (DSP) platform, through the system architecture featuring the MAX3232 communication interface for interaction with the computer and the DAC8552 D/A converter for output to the oscilloscope, the generation of attractors of the HNN-HR model is achieved, and the feasibility of its application in digital circuits is verified. The construction of neural networks by simulating biological synapses through memristors offers a promising avenue for exploring brain function and its bionics.</div></div>","PeriodicalId":54973,"journal":{"name":"Integration-The Vlsi Journal","volume":"105 ","pages":"Article 102490"},"PeriodicalIF":2.5000,"publicationDate":"2025-08-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Integration-The Vlsi Journal","FirstCategoryId":"5","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0167926025001476","RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"COMPUTER SCIENCE, HARDWARE & ARCHITECTURE","Score":null,"Total":0}

引用次数: 0

Abstract

With the development of information technology and neurobiology, there is an urgent need for a memory element with bionic properties to simulate the interactions among neurons. On this basis, a novel locally active memristor (LAM) model is designed, of which the memory properties are utilized to construct a coupled system of three-dimensional Hopfield neural network (HNN) and Hindmarsh–Rose (HR) neurons to simulate neuronal activities. First, the nonvolatility and local activity of the memristor is verified by the power-off plot (POP) and its direct current (DC)

V

I

plot. The device exhibits typical bistable resistance-switching behavior, maintaining two stable resistance states upon power-off, which confirms its non-volatile memory characteristics. A significant negative differential resistance (NDR) region observed in DC

V

I

curves directly verifies its local activity, indicating potential for active signal processing. Second, the complex dynamical behavior of HNN-HR is probed by numerical simulation, adjusting the coupling strength, synaptic weights and HR neuron parameters to demonstrate the bionic properties. The research results show that not only are multiple hidden attractor structures exhibited by the model, but also typical nonlinear phenomena such as transient chaos and intermittent chaos can be reproduced by it, and the dynamic transition between different chaotic firing modes can be realized. In addition, the phenomena of multi- state coexisting attractors and the expansion and migration of attractor topological structures are observed in the model. Finally, by means of the TMS320F28335 digital signal processing (DSP) platform, through the system architecture featuring the MAX3232 communication interface for interaction with the computer and the DAC8552 D/A converter for output to the oscilloscope, the generation of attractors of the HNN-HR model is achieved, and the feasibility of its application in digital circuits is verified. The construction of neural networks by simulating biological synapses through memristors offers a promising avenue for exploring brain function and its bionics.

查看原文本刊更多论文

基于多放电模式局部有源忆阻器的异质神经网络

随着信息技术和神经生物学的发展，迫切需要一种具有仿生特性的记忆元件来模拟神经元之间的相互作用。在此基础上，设计了一种新的局部有源忆阻器（LAM）模型，利用其记忆特性构建三维Hopfield神经网络（HNN）和Hindmarsh-Rose （HR）神经元的耦合系统来模拟神经元的活动。首先，通过断电图（POP）及其直流（DC） V - I图验证了忆阻器的非易失性和局部活度。该器件表现出典型的双稳态电阻开关行为，在断电时保持两个稳定的电阻状态，这证实了其非易失性存储特性。在直流V - I曲线中观察到显著的负差分电阻（NDR）区域，直接验证了其局部活动，表明主动信号处理的潜力。其次，通过数值模拟研究HNN-HR的复杂动力学行为，调整耦合强度、突触权重和HR神经元参数来证明其仿生特性。研究结果表明，该模型不仅能表现出多种隐吸引子结构，还能再现瞬态混沌和间歇混沌等典型非线性现象，并能实现不同混沌发射模式之间的动态过渡。此外，该模型还观察到多态吸引子共存现象以及吸引子拓扑结构的扩展和迁移。最后，利用TMS320F28335数字信号处理（DSP）平台，通过MAX3232通信接口与计算机交互，DAC8552 D/A转换器向示波器输出的系统架构，实现了HNN-HR模型吸引子的生成，并验证了其在数字电路中应用的可行性。通过忆阻器模拟生物突触构建神经网络，为探索脑功能及其仿生学提供了一条有前景的途径。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Integration-The Vlsi Journal 工程技术-工程：电子与电气

CiteScore

3.80

自引率

5.30%

发文量

107

审稿时长

6 months

期刊介绍： Integration''s aim is to cover every aspect of the VLSI area, with an emphasis on cross-fertilization between various fields of science, and the design, verification, test and applications of integrated circuits and systems, as well as closely related topics in process and device technologies. Individual issues will feature peer-reviewed tutorials and articles as well as reviews of recent publications. The intended coverage of the journal can be assessed by examining the following (non-exclusive) list of topics: Specification methods and languages; Analog/Digital Integrated Circuits and Systems; VLSI architectures; Algorithms, methods and tools for modeling, simulation, synthesis and verification of integrated circuits and systems of any complexity; Embedded systems; High-level synthesis for VLSI systems; Logic synthesis and finite automata; Testing, design-for-test and test generation algorithms; Physical design; Formal verification; Algorithms implemented in VLSI systems; Systems engineering; Heterogeneous systems.