Polymer hydrogel electronics with adhesive, self-healing, and anti-ultraviolet capabilities for machine learning-promoted electrophysiological detection

• The mussel-featured hydrogel was designed by multiple dynamic interactions. • The hydrogel showed perfect ductility, self-healing, adhesion, and anti-ultraviolet. • The hydrogel i-skin exhibited fast response/recovery time and wide detection range. • A sensing system realizes machine learning-promoted electrophysiological detection. While adjustable mechanical properties, diverse biochemical features, and good ionic conductivity enabled by molecular engineering, polymer hydrogels as ionic skins (i-skins) are still limited by multiple challenges in the practical human health monitoring and health diagnosis, such as undesirable ductility and contact adhesion, poor functionality, and insufficient sensing ability. In view of this, the strategy of multiple dynamic interactions was proposed, and a mussel characteristic double-network polymer composite hydrogel (PG-BT2) with tannic acid and borax as dynamic media was reasonably designed. Due to the dynamic reversible interactions from the crosslinked network combined with special functional groups, the obtained PG-BT2 hydrogel attains remarkable stretchability with >1200 % elongation and good toughness of 268.6 kJ/m −3 , and also possesses favorable versatility (e.g., self-healing time <2 s, adhesion strength of 4.62 kPa to pigskin, and UV transmittance of 15.3 % at 365 nm). Additionally, the i-skin derived from the PG-BT2 hydrogel exhibits excellent sensing properties, including a low interface contact resistance of 8.3 kΩ at 1 kHz, fast response/recovery time of 0.657 s/0.348 s, and a wide sensing range of 0–1100 % strain. By combining the hydrogel-based i-skin patch with the wireless circuit connected to a Bluetooth module, the constructed sensor system is capable of accurately monitoring micro-expressions and multi-joint movements. The portable system is also demonstrated to capture high-fidelity signals of electrocardiograms available for training artificial intelligence, achieving differentiation and recognition of blood pressure status, which will greatly benefit health management and intelligent medicine.