Fully Integrated Memristive Hodgkin-Huxley Neurons With Homeostatic Plasticity
Artificial neurons based on the Hodgkin-Huxley (H-H) models could mimic the richest firing patterns, showing great potential in building high-intelligent systems. Emerging devices, such as NbO2-based threshold-switching devices, exhibit more advantages in constructing H-H neuron circuits compared to conventional transistors. However, the on-chip integration of the memristive H-H neuron circuit remains unexplored, limiting its practical applications in hardware. Here, we design and fabricate a fully integrated memristive H-H neuron circuit and achieve all-or-nothing, refractory period, integrator, class 1 excitation, tonic spiking, subthreshold oscillation, tonic bursting, and mixed-mode firing behaviors. We also demonstrate the homeostatic plasticity based on integrated H-H neuron, specifically, the neuron increases threshold spontaneously when receiving an excessively strong input to avoid the superexcitation in the neuron. This work verifies the feasibility of building an integrated memristive H-H neuron and lays the foundation for building high-bionic neuromorphic systems.
Authors
- Yanting Ding
- Chao Li (ORCID: https://orcid.org/0009-0001-7082-1942)
- Pei Chen
- Lingli Cheng
- Jie Yu (ORCID: https://orcid.org/0000-0002-6768-7236)
- Qi Liu (ORCID: https://orcid.org/0000-0001-7062-831X)
- Yue Yang (ORCID: https://orcid.org/0000-0002-5755-5395)
- Xumeng Zhang (ORCID: https://orcid.org/0000-0002-3828-151X)
Institutions
- Chinese Academy of Sciences (CN)
- Fudan University (CN)
- Institute of Microelectronics (CN)
Publication Details
- Journal
- IEEE Electron Device Letters
- Published
- 2024-09-10
- DOI
- https://doi.org/10.1109/led.2024.3456816
- Citations
- 9
- Primary Topic
- Advanced Memory and Neural Computing
- Type
- article
- Field-Weighted Citation Impact
- 1.88
Funders
- National Natural Science Foundation of China