An Artificial Nerve with Complementary Memristor for Neuromorphic Self‐Protection
Qiaoling Tian, Xiaoning Zhao, Xichen Tang, Ye Tao, Zhongqiang Wang, Xiaoxiao He, Jianbo Cao, Ya Lin, Haiyang Xu, Yichun Liu
Abstract
Abstract Protective inhibition (PI) behavior is critical to human self‐protection by reducing the physiological arousal of overloaded stimulation. Neuromorphic hardware mimicking PI behavior can facilitate advancements in the field of bio‐realistic artificial intelligence and remains to be studied. Herein, a specially engineered complementary memristor with limited Cu embedding in porous amorphous carbon is proposed to emulate PI behavior. The depletion of Cu produces complementary switching behavior, sharing similarities with the PI response of its biological counterpart. Integrating the memristor, a resistor, and various sensors yields artificial sensory afferent nerves that intelligently demonstrate PI behavior in response to stimuli including pressure, light, and temperature. As a proof of concept, an artificial motion sensory nerve (AMSN) is constructed by sending the feedback signals of the memristor to intelligently stimulate the sciatic nerve of an anesthetized mouse to perform neuromodulation. Upon the increase of stimuli, the overstrain of the muscle is successfully inhibited, mimicking the PI response of the biological nerve. Furthermore, an AMSN is also constructed with a 3 × 3 memristor array to shed light on the potential for large area and multimodal sensing. This work provides a fundamental synaptic unit for developing bio‐realistic artificial nerve for neuromorphic self‐protection.