Interference-resistant Janus hydrogel electrodes for AI-enabled platform for rehabilitation assessment

Physiological signal acquisition is essential for medical diagnostics and human-machine interfaces. Motion artifacts and sweat-induced instability present persistent challenges for conventional electrodes, fundamentally limiting their reliability in dynamic or long-term use. Here, we report a Janus hydrogel electrode that combines high electrical conductivity (3 × 10⁴ S·m^-1), strong interfacial adhesion to copper surfaces (2079 J·m^-2), and low skin-electrode impedance (∼300 Ω). Distinct from conventional Ag/AgCl electrodes, hydrogel electrode uniquely combines hydrophobic protection, high-strain impedance insensitivity (250%) and rapid self-healing, achieving interference-resistant physiological signal monitoring even under vibration and moisture interferences. Notably, coupling with advanced Transformer algorithms optimized for temporal sequence analysis, the electrodes delivered 97.02% gesture recognition accuracy. This system harnesses these gestures to generate robust and intuitive command inputs, paving the way for potential rehabilitation applications, where future clinical validation is required. This work introduces a multifunctional, biocompatible electrode platform that couples reliable physiological monitoring with AI-powered rehabilitation assessment, opening new avenues for next-generation intelligent wearable bioelectronics.