Smart Sensing Nanoplatform with ML-Enhanced Prediction for Simultaneous Neurotransmitter Detection

Abstract

This study aims to develop a sustainable electrochemical sensor integrated with machine learning (ML) for the simultaneous prediction of serotonin (5-HT) and dopamine (DA) concentrations, two key biomarkers of neurological disorders. The novelty of this work lies in the combination of green-synthesized nanomaterials and advanced ML algorithms within a miniaturized, eco-friendly, and cost-effective sensing platform. This integration improves diagnostic accuracy, reduces analysis time, and minimizes reliance on manual interpretation, offering significant potential for clinical applications. Monitoring neurotransmitters is crucial for the early diagnosis of neurological diseases, yet conventional methods are often costly, time-consuming, and dependent on specialized expertise. Electrochemical sensors based on nanomaterials provide a promising alternative, particularly when enhanced by ML to enable accurate, real-time prediction of analyte concentrations. In this work, a screen-printed electrode (SPE) functionalized with green-synthesized silver nanoparticles decorated reduced graphene oxide (AgNPs-rGO) was employed for the simultaneous neurotransmitter detection. Experimental electrochemical signals were systematically collected and used as input features for ML models, with 80% of the dataset applied for training and 20% for validation. Among the tested algorithms, Random Forest Regression (RFR) achieved the highest predictive accuracy, with R² values of 0.984 for 5-HT and 0.897 for DA. By integrating eco-friendly nanomaterials with ML-based predictive modeling on a miniaturized platform, this study provides a rapid, efficient, and sustainable approach with strong potential for improving early diagnosis of neurological disorders.