3D Printed PEDOT:PSS-Based Conducting and Patternable Eutectogel Electrodes for Machine Learning on Textiles

The development of medical wearables necessitates novel electrodes for cutaneous electrophysiology. Traditional hydrogel electrodes, while effective, pose limitations for long-term use due to skin irritation and the need for frequent replacement. Dry electrodes, although easier to integrate into wearable devices, often suffer from poor skin contact and high noise. Poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS) has emerged as a leading material for bioelectronics due to its mixed ionic/electronic conductivity, mechanical flexibility, and biocompatibility. However, its poor mechanical stability limits its practical application in 3D-printed structures.<br/>This study introduces a novel approach to fabricating 3D-printed, PEDOT:PSS-based conducting eutectogel electrodes using a deep eutectic solvent (DES) and polyethylene glycol diacrylate (PEGDA). The DES, composed of choline and lactic acid, enhances the electrical conductivity and mechanical stability of the material. The PEGDA enables the use of standard stereolithographic printing of high-resolution micro-structures. The electrodes were printed in four patterns—flat, pyramidal, striped, and wavy—to assess the impact of geometry on skin conformability and mechanical contact. These electrodes were embedded in textiles and used to generate body surface potential maps (BSPMs) of the forearm during different finger movements. BSPMs for the letters B, I, and O in sign language were recorded and used to train a logistic regression classifier.<br/>The fabricated electrodes demonstrated improved biocompatibility, mechanical stability, and electrical conductivity. When embedded in textiles, these electrodes effectively captured BSPMs, which displayed distinctive patterns corresponding to different finger movements. The logistic regression classifier trained on these BSPMs achieved an accuracy of 97%, significantly surpassing the 86% accuracy of a control array of bare silver electrodes. The performance of the classifiers was assessed by means of confusion matrices and principal component analysis, which showed reliable separation of the classes B, I, and O.<br/>This work presents a significant advancement in the field of wearable electronics for medical applications. The novel PEDOT:PSS-based eutectogel electrodes, with their enhanced performance and stability, offer a promising solution for long-term cutaneous electrophysiology recordings. The ability to accurately classify sign language letters from forearm EMG recordings demonstrates the potential for these electrodes in developing online sign-language translation systems and other brain-machine interfaces. This methodology paves the way for affordable, scalable, and geometry-customised wearables embedded in textiles, significantly advancing machine-interface electronics.