Epileptic brain network from scalp EEG: identifying the epileptic driver by connectivity analysis on brain waveforms

Epilepsy is a neurological disorder characterized by seizures, i.e. abnormal synchronous activity of neurons in the brain. During a focal seizure the epileptic activity spreads rapidly from the ictal onset region to neighboring brain areas. ElectroEncephaloGraphy (EEG) is a commonly used technique to diagnose epilepsy. EEG has a high temporal resolution which allows us to investigate the dynamics of the underlying brain activity. Due to the rapid propagation of a seizure, the seizure can originate from a network of brain regions which are simultaneously active before being noticeable on the EEG. In this paper we investigate two state of the art source localization techniques, the Recursive Applied and Projected (RAP) and the pre-correlated and orthogonally projected (POP) multiple signal classification (MUSIC) algorithm, to identify the location of the driver behind the simulated epileptic brain network. Furthermore we investigate the applicability of connectivity analysis to identify the source driving the underlying brain network. We showed that the POP-MUSIC algorithm outperforms the RAP-MUSIC algorithm to identify the locations of the simultaneous brain activity. Furthermore, we showed the feasibility of identifying the driver behind a brain network by POP-MUSIC algorithm followed by connectivity analysis