An EEG data processing pipeline for seizure detection
Abstract: Abstract network representations of EEG data are fundamental to understand the neural-circuit mechanisms underlying neurological disorders. However, EEG abstract network generation tools and network data interpretation methods are not standardized and are still to found their place in the clinical settings. Here we present a data processing pipeline to automate the process of generating functional networks and graph representations from EEG data. And, we demonstrate a use case for these abstract network representations to detect seizure activity with graph neural networks, a novel artificial neural network paradigm optimized to learn from graph-structured data.
Note: research and methods dissemination manuscripts are under preparation.
Short-bio: Dr. Alan Diaz holds a Postdoctoral Fellowship at the Neural Systems and Brain Signal Processing Lab (NSBSPL) which is part of the Krembil Research Institute - University Health Network (UHN) in Toronto, ON, Canada. He obtained a PhD in Computer Science from the University of Dublin, Trinity College (Ireland) in 2021. His research focuses on the analysis of electrophysiological signals, mainly intracranial EEG, aiming to identify cross-regional functional subnetworks in the human brain that can serve as biomarkers for neurological disorders such as epilepsy, Parkinson's disease and major depression disorder.