Scholarlink

<p>Understanding the dynamic firing activities of neural networks and their regulatory mechanisms is critical for advancing neuromorphic computing and unraveling brain function. Locally active memristors, a crucial category of memristive devices, have been widely applied to simulate synapse-inspired processes within neural networks, yet the connection between memristor properties and neuronal electrical behaviors remains to be further investigated. In this study, a locally active memristor is employed as a synapse to couple a FitzHugh–Nagumo (FHN) neuron and a Hindmarsh–Rose (HR) neuron. By tuning the neuron model parameters and coupling strength, the dynamic firing behaviors of the system are analyzed, revealing rich periodic states and diverse bifurcation phenomena. The phase synchronization of the coupled neurons is further investigated with varying coupling strengths. Additionally, complexity analysis based on the spectral entropy (SE) metric is performed to verify the high pseudo-randomness of the sequences generated by the coupled system, highlighting its great potential for secure communication applications. Finally, PSIM-based analog circuit simulations are conducted to validate the consistency between numerical simulation results and the physical feasibility of both the memristor and the coupled neuron system. This work not only deepens the understanding of memristor-modulated neural dynamics and synchronization mechanisms but also provides valuable theoretical and experimental support for the design, control, and practical applications of neuromorphic systems and large-scale neural networks.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Locally active memristive synapse modulate firing dynamics and synchronization in heterogeneous FHN-HR neuronal networks

  • 期刊论文

  • 发表日期：

    2026年5月29日

查看全文
  • Xinhao Li,
  • Xiuquan Liu,
  • Ming Liu,
  • Penghe Ge,
  • Dongpo Hu

摘要

Understanding the dynamic firing activities of neural networks and their regulatory mechanisms is critical for advancing neuromorphic computing and unraveling brain function. Locally active memristors, a crucial category of memristive devices, have been widely applied to simulate synapse-inspired processes within neural networks, yet the connection between memristor properties and neuronal electrical behaviors remains to be further investigated. In this study, a locally active memristor is employed as a synapse to couple a FitzHugh–Nagumo (FHN) neuron and a Hindmarsh–Rose (HR) neuron. By tuning the neuron model parameters and coupling strength, the dynamic firing behaviors of the system are analyzed, revealing rich periodic states and diverse bifurcation phenomena. The phase synchronization of the coupled neurons is further investigated with varying coupling strengths. Additionally, complexity analysis based on the spectral entropy (SE) metric is performed to verify the high pseudo-randomness of the sequences generated by the coupled system, highlighting its great potential for secure communication applications. Finally, PSIM-based analog circuit simulations are conducted to validate the consistency between numerical simulation results and the physical feasibility of both the memristor and the coupled neuron system. This work not only deepens the understanding of memristor-modulated neural dynamics and synchronization mechanisms but also provides valuable theoretical and experimental support for the design, control, and practical applications of neuromorphic systems and large-scale neural networks.