Jidong Li, Wei Zhao, Chenwei Fu, Zhenpeng Zhai, Pengfei Xu, Xinyuan Diao, Wanlin Guo, Jun Yin

Nature Communications   ，2025，16（1），  7911

Abstract: Biological nervous systems rely on distinct spiking frequencies across a wide range for perceiving, transmitting, processing, and executing information. Replicating this frequency range in an artificial neuron would facilitate the emulation of biosignal diversity but it remains challenging. Here, we develop an ion-electronic hybrid artificial neuron by compactly integrating a nonlinear electrochemical element with a solid-state memristor. This hybrid neuron employing a minimalist architecture exhibits a tunable spiking frequency spanning five orders of magnitude, significantly surpassing the capability of artificial neurons based on electronic devices. Notably, stimuli-dependent ion fluxes enable inherent afferent sensing of liquid flow, temperature, and chemical constituents, eliminating the need for separate, bulky sensors. Connection to biomotor nerves facilitates muscle actuation with frequency-regulated modes. The frequency encoding of a hybrid neuron array allows for the recognition of handwritten patterns. This hybrid neuron design, taking advantage of both ionic and electronic features, offers a promising approach for advanced e-skin and neurointerface technologies.

Link: https://www.nature.com/articles/s41467-025-63195-7