多稳态
爆裂
混乱的
记忆电阻器
生物神经元模型
瞬态(计算机编程)
神经元
计算机科学
生物系统
物理
统计物理学
控制理论(社会学)
人工神经网络
人工智能
神经科学
非线性系统
生物
控制(管理)
操作系统
量子力学
作者
Hairong Lin,Chunhua Wang,Yichuang Sun,Wei Yao
标识
DOI:10.1007/s11071-020-05687-3
摘要
The theoretical, numerical and experimental demonstrations of firing dynamics in isolated neuron are of great significance for the understanding of neural function in human brain. In this paper, a new type of locally active and non-volatile memristor with three stable pinched hysteresis loops is presented. Then, a novel locally active memristive neuron model is established by using the locally active memristor as a connecting autapse, and both firing patterns and multistability in this neuronal system are investigated. We have confirmed that, on the one hand, the constructed neuron can generate multiple firing patterns like periodic bursting, periodic spiking, chaotic bursting, chaotic spiking, stochastic bursting, transient chaotic bursting and transient stochastic bursting. On the other hand, the phenomenon of firing multistability with coexisting four kinds of firing patterns can be observed via changing its initial states. It is worth noting that the proposed neuron exhibits such firing multistability previously unobserved in single neuron model. Finally, an electric neuron is designed and implemented, which is extremely useful for the practical scientific and engineering applications. The results captured from neuron hardware experiments match well with the theoretical and numerical simulation results.
科研通智能强力驱动
Strongly Powered by AbleSci AI