TY - GEN
T1 - An Izhikevich Model Neuron MOS Circuit for Low Voltage Operation
AU - Tamura, Yuki
AU - Moriya, Satoshi
AU - Kato, Tatsuki
AU - Sakuraba, Masao
AU - Horio, Yoshihiko
AU - Sato, Shigeo
N1 - Funding Information:
This study was supported by the Cooperative Research Project Program of the Research Institute of Electrical Communication, Tohoku University; the Program on Open Innovation Platform with Enterprises, Research Institute and Academia (OPERA) from Japan Science and Technology Agency (JST); JSPS KAKENHI (Grant Nos, 17K18864 and 18J12197); and JST CREST Grant Number JPMJCR18K4, Japan.
Publisher Copyright:
© 2019, Springer Nature Switzerland AG.
PY - 2019
Y1 - 2019
N2 - 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.
AB - 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.
KW - Analog neuron circuit
KW - Izhikevich neuron model
KW - Null-cline analysis
KW - SPICE simulation
UR - http://www.scopus.com/inward/record.url?scp=85072865968&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85072865968&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-30487-4_55
DO - 10.1007/978-3-030-30487-4_55
M3 - Conference contribution
AN - SCOPUS:85072865968
SN - 9783030304867
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 718
EP - 723
BT - Artificial Neural Networks and Machine Learning – ICANN 2019
A2 - Tetko, Igor V.
A2 - Karpov, Pavel
A2 - Theis, Fabian
A2 - Kurková, Vera
PB - Springer Verlag
T2 - 28th International Conference on Artificial Neural Networks, ICANN 2019
Y2 - 17 September 2019 through 19 September 2019
ER -