An Izhikevich Model Neuron MOS Circuit for Low Voltage Operation

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

研究成果: Conference contribution


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.

ホスト出版物のタイトルArtificial Neural Networks and Machine Learning – ICANN 2019
ホスト出版物のサブタイトルTheoretical Neural Computation - 28th International Conference on Artificial Neural Networks, 2019, Proceedings
編集者Igor V. Tetko, Pavel Karpov, Fabian Theis, Vera Kurková
出版社Springer Verlag
出版ステータスPublished - 2019
イベント28th International Conference on Artificial Neural Networks, ICANN 2019 - Munich, Germany
継続期間: 2019 9 172019 9 19


名前Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
11727 LNCS


Conference28th International Conference on Artificial Neural Networks, ICANN 2019

ASJC Scopus subject areas

  • 理論的コンピュータサイエンス
  • コンピュータ サイエンス(全般)


「An Izhikevich Model Neuron MOS Circuit for Low Voltage Operation」の研究トピックを掘り下げます。これらがまとまってユニークなフィンガープリントを構成します。