Ensemble methods for representation of fMRI, EEG/MEG data in graph form for classification of brain states

The goal of this work is to propose a method for representing functional magnetic resonance imaging data and electroencephalography/magnetoencephalography data in the form of graphs reflecting the interactions of brain regions for brain state classification tasks. This method is based on ensemble learning to improve generalizability, accuracy, robustness and parallelism of the classification model, which is important when analyzing such large, complexly structured and noisy data as neuroimaging data.