P300 Classification for EEG-based BCI system with Bayes LDA, SVM, LassoGLM and a Deep CNN methods

A research project that explores the potential of DL models in detecting ERP (p300) in EEG signals. From EDA to modeling, we developed an end-to-end training and evaluation workflow that achieved an accuracy of 98% on the EPFL dataset. We utilized the achieved results in building a Neurophone and using the famous EMOTIV headset for EEG measurement.