Label-Free Classification of L-Histidine Vs Artificial Human Sweat Using Laser Scribed Electrodes and a Multi-Layer Perceptron Neural Network

A challenge in wearable technology lies in the realtime monitoring of molecular biomarkers associated with human health. Electrochemical sensors are one of the most useful tools for this purpose and are commonly used in health monitoring devices. Electrochemical biosensing is particularly convenient when used in user-friendly, low-cost devices for testing noninvasive body fluids such as sweat and saliva. However, achieving high selectivity and specificity in measurements depends on the complexity of the biomarker and the stability of the biomarker capture molecule. In this study, laser-scribed electrodes (LSEs) were manufactured using a CO2 laser cutter on Polyimide for the label-free classification of sweat components. Cyclic voltammetry experiments were performed on artificial human sweat and the sweat component L-Histidine. The resulting voltammogram data served as input to train a Multi-Layer Perceptron Neural Network (MLP-NN) algorithm capable of classifying L-Histidine and artificial sweat.