Machine Learning-Empowered\nMulticolor Detection and\nDiscrimination of Dopaminergic Agents: Utilizing Gold Nanorods for\nGenerating Rainbow Signals

The demand for rapid and sensitive multianalyte detection\nof dopaminergic\nagents in assessing the progression of Parkinson’s disease\n(PD) has attracted recurring interest in recent years. In this study,\na multicolor sensor was devised to multiplex detection and discrimination\nof levodopa (l-DOPA), carbidopa (Carbi), and benserazide\n(Benz)key dopaminergic compounds. Notably, the presence of\ndopaminergic agents inhibits the etching of gold nanorods (AuNRs)\nby <i>N</i>-bromosuccinimide (NBS), with varying degrees\nof inhibition depending on the concentration and properties of the\nanalytes. This interaction induces alterations in the size and aspect\nratio of the AuNRs, resulting in a red shift within the plasmonic\nband. Remarkably, this shift in the plasmonic band of AuNRs corresponds\nto color variations akin to a rainbow in the solution, providing a\nvaluable opportunity for precise visual detection. Statistical analysis\nof the collected data set was performed using two machine learning\ntechniques: linear discriminant analysis (LDA) for pattern recognition\nand partial least-squares regression (PLSR) for regression analysis.\nLDA effectively discriminated dopaminergic agents and their binary\nand ternary mixtures with 100% accuracy, while PLSR enabled quantitative\nanalysis over a wide concentration range: 1.17–25.0, 2.63–25.0,\nand 1.18–25.0 μmol L<sup>–1</sup>, with detection\nlimits of 0.93, 0.88, and 0.39 μmol L<sup>–1</sup>, for l-DOPA, Carbi, and Benz, respectively. Furthermore, we validated\nthe practical applicability of the sensor assay by identifying and\nquantifying dopaminergic agent in medicines tablets containing mixtures\nof these compounds, affirming its utility and reliability for on-site\nmedicines quality control.