Adaptive Biomimetic Neuronal Circuit System Based on Myelin Sheath Function

Abstract

Brain-inspired neuromorphic computing architectures are receiving significant attention in the consumer electronics field owing to their low power consumption, high computational capacity, and strong adaptability, where highly biomimetic circuit design is at the core of neuromorphic network research. Myelin sheaths are crucial cellular components in building stable circuits in biological neurons, capable of adaptively adjusting the conduction speed of neural signals. However, current research on neuronal circuits relies on simplified mathematical models and overlooks the adaptive functionality of myelin sheaths. This paper is based on the dynamic mechanism of myelination, utilizing physical devices such as memristors and voltage-controlled variable capacitors to simulate the physiological functions of myelinsheaths, and other organelles. Furthermore, adaptive biomimetic neuronal circuit system (ABNCS) is constructed by connecting various devices according to the physiological structure of neurons. PSpice simulations show that the ABNCS can adjust its parameters autonomously as the number of action potentials(APs) increase, which modifies the neuron’s activation criteria and firing rate. Through circuit experiments, PSpice simulations were further validated. Implementing myelin sheath functions int he neuronal circuit improves adaptability and reduces power consumption, and when combined with artificial synapses to construct neural networks, can form more stable neural circuits.