A Wireless, Bias-Free Near-Infrared Photoelectrochemical Platform for Intelligent Neurochemical Monitoring in Freely Behaving Animals

Monitoring of neurochemical dynamics in deep brain regions of freely behaving subjects remains challenging due to the reliance of existing techniques on wired connections, external electrical bias, or shallow optical penetration. Here, we report a wireless, bias-free photoelectrochemical (PEC) sensing platform that enables longitudinal neurochemical monitoring under near-infrared (NIR) illumination. The platform integrates an NIR-responsive triazine graphdiyne photoelectrode, modular conformation-switching molecular probes that transduce target binding into intrinsic photocurrent signals at 0 V, and a miniaturized wireless module for autonomous operation. As a proof of concept, glutathione dynamics are monitored in the substantia nigra of freely moving Parkinson's disease model rats with high sensitivity and selectivity, while an antifouling hydrogel coating ensures stability in complex biological environments. Furthermore, on-device machine learning based on knowledge distillation enables disease-state inference from single-channel PEC signals. This work establishes a generalizable, intelligent PEC interface for minimally invasive neurochemical sensing, providing a versatile foundation for advanced implantable bioelectronic systems.