An Izhikevich Model Neuron MOS Circuit for Low Voltage Operation

Yuki Tamura, Satoshi Moriya, Tatsuki Kato, Masao Sakuraba, Yoshihiko Horio, Shigeo Sato

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The Izhikevich neuron model has attracted attention because it can reproduce various neural activities although it is described by simple differential equations and is expected to be applied to engineering. Among a few MOS circuits inspired by the Izhikevich model, the circuit proposed by Wijekoon and Dudek in 2008 exhibits the simplest structure, and it is practical. However, the power supply voltage of the circuit is 3.3 V. To implement such a neuron MOS circuit using state-of-the-art semiconductor manufacturing process, we must redesign the circuit to operate it with a lower supply voltage. Thus, we analyzed their circuit operation by SPICE simulation assuming a 1.0 V supply voltage and found that the bias voltage ranges to generate specific spike activities were limited. In addition, we clarified the discrepancies between the Izhikevich neuron model and the original circuit. In this study, we propose a new Izhikevich model neuron circuit based on these findings and investigate the circuit dynamics by null-cline analysis and SPICE simulation. The dynamics of the proposed MOS circuit are close to those of the Izhikevich model and various spikes are generated. Furthermore, we successfully enlarged the bias voltage range for specific spikes.

Original languageEnglish
Title of host publicationArtificial Neural Networks and Machine Learning – ICANN 2019
Subtitle of host publicationTheoretical Neural Computation - 28th International Conference on Artificial Neural Networks, 2019, Proceedings
EditorsIgor V. Tetko, Pavel Karpov, Fabian Theis, Vera Kurková
PublisherSpringer-Verlag
Pages718-723
Number of pages6
ISBN (Print)9783030304867
DOIs
Publication statusPublished - 2019 Jan 1
Event28th International Conference on Artificial Neural Networks, ICANN 2019 - Munich, Germany
Duration: 2019 Sep 172019 Sep 19

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11727 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference28th International Conference on Artificial Neural Networks, ICANN 2019
CountryGermany
CityMunich
Period19/9/1719/9/19

Keywords

  • Analog neuron circuit
  • Izhikevich neuron model
  • Null-cline analysis
  • SPICE simulation

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint Dive into the research topics of 'An Izhikevich Model Neuron MOS Circuit for Low Voltage Operation'. Together they form a unique fingerprint.

Cite this