Explainable Machine Learning Methods for Classification of Brain States during Visual Perception

The aim of this work is to find a good mathematical model for the classification brain states during visual perception with focus on interpretability results. To achieve it, we use deep learning models different activation functions and optimization methods their comparison best considered dataset 31 EEG channels trials. estimate influence features process make method more interpretable, SHAP library technique. We that Adagrad worst one FTRL. In addition, only works well both linear tangent models. results could be useful EEG-based brain–computer interfaces (BCIs) in part choosing appropriate machine correct training BCI intelligent system.