Multi-day recalibration-free electrophysiology interface based on carbon nanotube thin film dry electrodes

Abstract Electrophysiological signals provide crucial and reliable information for decoding behavioral and cognitive human states. However, existing electrode interfaces are susceptible to artifacts or drift over time, thereby requiring frequent machine-learning model retraining. Here we present an end-to-end design of dry electrodes based on carbon nanotube (CNT) thin films, synthesized by a roll-to-roll floating catalyst chemical vapor deposition process. By balancing structural and electrical properties, we achieved low electrode-skin impedance, high signal fidelity, minimal motion artifacts, high breathability, and biocompatibility, outperforming the gold standard wet electrodes. In a multichannel myoelectric interface demonstration, these CNT electrodes maintained stable signals and consistent task performance across 14 days using only minutes of training data and no further model retraining. To our knowledge, this is the first demonstration of a multi-day recalibration-free dry electrode system under naturalistic conditions, achieved by stable electrode–skin interface, rather than ongoing algorithmic adaptation.