A Comparison between Different Machine Learning Approaches Combined with Anodic Stripping Voltammetry for Copper Ions and pH Detection in Cell Culture Media

Recently, the scientific community has shown a great interest about Organ-on-Chip (OoC) devices, special kind of micro-fabricated platforms capable recapitulating human physiology implementing traditional cell culture methods and concept in vivo studies. Copper ions represent cellular micronutrient that must be monitored for its potential hazardous effects. The application electrochemical analysis heavy metal detection quantification commercial media presents several issues due to electrolyte complexity interferents. In fact, best our knowledge, there is lack applications OoC devices implement Anodic Stripping Voltammetry as an ion dosing technique reasons reported above. considering just peak intensity value from measurement, it turns out challenging quantify concentration since other or molecules may interfere with measurement. With aim overcome these issues, present work aims develop automated system based on machine learning algorithms demonstrate possibility build reliable forecasting model copper three different (MEM, DMEM, F12). Effectively, combining measurements multivariate algorithm leads higher classification accuracy. Two pH conditions, i.e., physiological (pH 7.4) acidic 4), were considered establish how influences experimental datasets obtained using square-wave anodic stripping voltammetry (SWASV) used carry trained model. proposed method led significant improvement Cu2+ accuracy (96.6% SVM 93.1% NB MEM) well being able monitor solution.