Advancing Electrochemical Screening of Neurotransmitters Using a Customizable Machine Learning-Based Multimodal System

High throughput and automated optical readout systems are already an industry standard in life sciences for screening several reactions at once. However, such high throughput systems are in an inceptive stage for studying electrochemical interactions. This limitation, for example, slows down the process of establishing property-performance relation of novel materials for biochemical sensing. Herein, building on our prior work, we fabricate a low-cost customizable platform to screen response of acetic acid-treated laser induced graphene to identify and quantify four biogenic amine neurotransmitters in artificial saliva, namely dopamine, serotonin, epinephrine, and norepinephrine, which due to similar molecular structures are difficult to differentiate using conventional electrochemical methods. Our analytical platform analyzes multiple sensors at once and processes the data using machine learning to rapidly screen the material-molecule interactions by combining several electrochemical spectral components (fingerprints). Combining multiple spectral features, both within one electrochemical module and across different modules, significantly improves the sensor performance and allows identification of the biomolecules using the same material system. The proposed automated electroanalytical system can be used to screen material-molecule interactions as well as high throughput point-of-care testing for rapid, multiplexed, and low-cost molecular detection.