Enhanced Automated Determination of Optimal Positions for Electroencephalography (EEG) Electrode Placements Using Data-Driven Hybrid Machine Learning (ML) Models

Electroencephalography (EEG) is a noninvasive technique that records and measures the brain’s electrical activity by placing electrodes on the scalp to detect neuronal electrical signals. Accurate electrode placement is crucial for proper diagnosis and subsequent treatment. This study proposes an innovative hybrid algorithm that aids in the rapid automatic prediction of optimal electrode placements using an optical motion tracking system and a customized stylus. The proposed hybrid algorithm combines numerical models devised in our previous research with a machine learning regression model to estimate the best electrode sites. A graphical user interface (GUI) is used to assist the clinicians in placing the EEG electrodes with reference to the obtained position estimates from the algorithms. Experimental studies conducted show that the suggested hybrid approach produces position estimates that are about 20% more accurate than mathematical models and approximately 83% more accurate than a pure machine learning (ML) model. This demonstrates that the proposed hybrid algorithm is capable of predicting optimal EEG electrode locations with greater accuracy and precision compared with models based solely on mathematics or ML.