Gustation-Inspired Dual-Responsive Hydrogels for Taste Sensing Enabled by Machine Learning.

Human gustatory system recognizes salty/sour or sweet tastants based on their different ionic or nonionic natures using two different signaling pathways. This suggests that evolution has selected this detection dualism favorably. Analogically, this work constructs herein bioinspired stimulus-responsive hydrogels to recognize model salty/sour or sweet tastes based on two different responses, that is, electrical and volumetric responsivities. Different compositions of zwitter-ionic sulfobetainic N-(3-sulfopropyl)-N-(methacryloxyethyl)-N,N-dimethylammonium betaine (DMAPS) and nonionic 2-hydroxyethyl methacrylate (HEMA) are co-polymerized to explore conditions for gelation. The hydrogel responses upon adding model tastant molecules are explored using electrical and visual de-swelling observations. Beyond challenging electrochemical impedance spectroscopy measurements, naive multimeter electrical characterizations are performed, toward facile applicability. Ionic model molecules, for example, sodium chloride and acetic acid, interact electrostatically with DMAPS groups, whereas nonionic molecules, for example, D(-)fructose, interact by hydrogen bonding with HEMA. The model tastants induce complex combinations of electrical and volumetric responses, which are then introduced as inputs for machine learning algorithms. The fidelity of such a trained dual response approach is tested for a more general taste identification. This work envisages that the facile dual electric/volumetric hydrogel responses combined with machine learning proposes a generic bioinspired avenue for future bionic designs of artificial taste recognition, amply needed in applications.