Biomimetic afferent nervous system (ANS) is urgently needed in the booming development of neuromorphic chips and artificial intelligent robots. It is necessary to make biomimetic ANS stably respond to the real irritation (heat or touch), not the conventionally used electric stimuli, in a manner of low-power-consumption, flexibility and miniaturization. Here, we present a chip-like flexible ANS device by fusing the ion conductive elastomer (ICE) with the MXene artificial synapse. It can not only successfully generate stable response to the joint flexion behaviors and weak pulse at radial artery, but also the plastic neural response can be simulated. Theoretical explanation for the proposed biomimetic ANS device is investigated by integrating the simulated geometry deformation, ion migration theory and trap assisted tunnel effect. Furthermore, the proposed device is applied to mimic an interesting “operant conditioned reflex behavior” by using a typical “trial and error learning” experiment. The transformations from short-term memory to long-term memory during two operant conditioning reflex stages (training and experiencing) are implemented. Conclusively, the sensory behaviors in biological ANS are simulated by the proposed device, which demonstrates its powerful ability of generating plastic neuro-impulse to real stimuli, thereby granting it a new role to realize the anthropomorphic sensation.

仿生传入神经系统（ANS）是神经形态芯片和人工智能机器人蓬勃发展的迫切需要。有必要使仿生ANS以低功耗、柔性和小型化的方式稳定地响应真实的刺激（热或触摸），而不是传统使用的电刺激。在这里，我们通过将离子导电弹性体 (ICE) 与 MXene 人工突触融合，展示了一种类似芯片的柔性 ANS 设备。它不仅可以成功地对桡动脉的关节屈曲行为和微弱脉搏产生稳定的反应，而且可以模拟可塑性神经反应。通过整合模拟的几何变形、离子迁移理论和陷阱辅助隧道效应，研究了所提出的仿生 ANS 装置的理论解释。此外，所提出的设备通过使用典型的“试错学习”实验来模拟有趣的“操作性条件反射行为”。在两个操作条件反射阶段（训练和体验）中实现了从短期记忆到长期记忆的转变。最终，该装置模拟了生物 ANS 中的感觉行为，证明了其对真实刺激产生可塑性神经冲动的强大能力，从而赋予其实现拟人化感觉的新作用。在两个操作条件反射阶段（训练和体验）中实现了从短期记忆到长期记忆的转变。最终，该装置模拟了生物 ANS 中的感觉行为，证明了其对真实刺激产生可塑性神经冲动的强大能力，从而赋予其实现拟人化感觉的新作用。在两个操作条件反射阶段（训练和体验）中实现了从短期记忆到长期记忆的转变。最终，该装置模拟了生物 ANS 中的感觉行为，证明了其对真实刺激产生可塑性神经冲动的强大能力，从而赋予其实现拟人化感觉的新作用。