Scholarlink

<p>Neuromorphic computing aims to replicate the brain’s energy-efficient and parallel processing abilities through the utilization of spiking neural networks (SNN). The leaky-integrate-and-fire (LIF) neuron model serves as a fundamental component of SNN, balancing computational simplicity and biological accuracy. This study puts forward a novel LIF neuron that integrates memristor and memcapacitor, thereby improving biological realism. The memristor serves as a variable resistive element, crucial for controlling current flow and enabling efficient spike generation, while the memcapacitor functions as a dynamic charge storage element. This work illustrates the enhancement of neuron spike dynamics and energy efficiency through the use of memcapacitors in contrast to conventional capacitors. To further enhance energy efficiency, the neurons are integrated with adaptive functionality which reduces the spiking frequency gradually when exposed to constant input stimuli. Adaptive LIF neurons use volatile and non-volatile memristors for short- and long-term plasticity. Non-adaptive neurons are better for stable, precise activities because they generate constant spiking responses. An in-depth analysis of adaptive and non-adaptive LIF neuron circuits is presented in terms of energy efficiency and spiking dynamics. The findings reveal the potential of the proposed memcapacitor-based LIF neuron models as foundational elements for scalable, energy-efficient, and biologically inspired neuromorphic computing systems.</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

Implementation and Analysis of Non-adaptive and Adaptive LIF Neuron Circuit Utilizing Memristor and Memcapacitor Elements

  • 期刊论文

  • 发表日期：

    2025年5月19日

查看全文
  • Deepthi M S,
  • Shashidhara H R,
  • Ranjana Bijjur,
  • Sahana H B

摘要

Neuromorphic computing aims to replicate the brain’s energy-efficient and parallel processing abilities through the utilization of spiking neural networks (SNN). The leaky-integrate-and-fire (LIF) neuron model serves as a fundamental component of SNN, balancing computational simplicity and biological accuracy. This study puts forward a novel LIF neuron that integrates memristor and memcapacitor, thereby improving biological realism. The memristor serves as a variable resistive element, crucial for controlling current flow and enabling efficient spike generation, while the memcapacitor functions as a dynamic charge storage element. This work illustrates the enhancement of neuron spike dynamics and energy efficiency through the use of memcapacitors in contrast to conventional capacitors. To further enhance energy efficiency, the neurons are integrated with adaptive functionality which reduces the spiking frequency gradually when exposed to constant input stimuli. Adaptive LIF neurons use volatile and non-volatile memristors for short- and long-term plasticity. Non-adaptive neurons are better for stable, precise activities because they generate constant spiking responses. An in-depth analysis of adaptive and non-adaptive LIF neuron circuits is presented in terms of energy efficiency and spiking dynamics. The findings reveal the potential of the proposed memcapacitor-based LIF neuron models as foundational elements for scalable, energy-efficient, and biologically inspired neuromorphic computing systems.