Implantable Ionic Memristors Based on Natural Polymer Heterojunctions

ABSTRACT Memristors emulate the synaptic plasticity of the human brain by exhibiting conductance modulation dependent on prior stimuli, thereby enabling simultaneous information processing and storage. Ionic memristors, which utilize the migration of biological ions (e.g., K + , Cl − ), exhibit history‐dependent synaptic plasticity and offer a more biocompatible approach than conventional solid‐state electronic memristors. However, existing ionic memristors typically employ rigid substrates and biologically incompatible materials, which limit their potential for use in implantable devices. Here, we report a fully biocompatible natural polymer‐based ionic memristor, comprising a hyaluronic acid hydrogel, a chitosan hydrogel, and a polydimethylsiloxane (PDMS) substrate. Hyaluronic acid and chitosan are biodegradable natural polymers widely used in biomedical applications, while PDMS is a biocompatible elastomer demonstrating safety. Furthermore, our memristor operates as a dynamic memristor capable of performing reservoir computing. The reservoir computing implemented via our memristor achieved a prediction accuracy of 98.94%, while reducing the training time by 36.8% compared to systems that do not incorporate the memristors. Notably, the device maintained its computational performance post‐implantation and demonstrated excellent biocompatibility in both in vitro and in vivo evaluations. Collectively, these results highlight the potential of our approach as an implantable platform for bio‐integrated computing and next‐generation biomedical engineering.