Neural Networks for Directed Connectivity Estimation in Source-Reconstructed EEG Data

Directed connectivity between brain sources identified from scalp electroencephalography (EEG) can shed light on the brain’s information flows and provide a biomarker of neurological disorders. However, as volume conductance results in activity being mix activities originating multiple sources, correct interpretation their is formidable challenge despite source localization applied with some success. Traditional approaches rely statistical assumptions that usually do not hold for EEG, calling model-free approach. We investigated several types Artificial Neural Networks estimating Connectivity Reconstructed EEG Sources assessed accuracy respect to ground truths. show Long Short-Term Memory neural network Non-Uniform Embedding yields most promising due its relative robustness differing dipole locations. conclude certain architectures compete already established methods analysis.