CMStatistics 2022: Start Registration
View Submission - CMStatistics
B1730
Title: Club Exco: Clustering brain extreme communities from multi-channel EEG data Authors:  Matheus Guerrero - King Abdullah University of Science and Technology (Saudi Arabia) [presenting]
Hernando Ombao - King Abdullah University of Science and Technology (KAUST) (Saudi Arabia)
Raphael Huser - King Abdullah University of Science and Technology (Saudi Arabia)
Abstract: Current methods for clustering nodes in a brain network are determined by cross-dependence measures, which are computed from the entire range of values of the EEG signals. One limitation of these measures is that they do not distinguish whether the signals are dependent (or synchronous) only at large amplitudes or over the entire range of values. We develop the Club Exco method for clustering brain-extreme communities to overcome these shortcomings. Club Exco uses a spherical k-means procedure applied to the ``pseudo-angles'', derived from extreme amplitudes of EEGs, to cluster multi-channel EEG data. With this approach, a cluster center is considered an extremal prototype, revealing nodes sharing the same extreme behavior (i.e., large amplitudes of the signal from one node are in tandem with large amplitudes of the others). Non-extreme-value techniques cannot identify this important feature. Club Exco serves as a tool to classify EEG channels into mutually asymptotically (in)dependent groups. It provides insights into how the brain network organizes itself during an extreme event (e.g., epileptic seizure) in contrast to normal states. We apply Club Exco to investigate the differences in EEG brain connectivity networks of a patient diagnosed with epilepsy, a chronic neurological disorder affecting more than 50 million people globally. Our method reveals a substantial difference in the organization of the alpha band in the brain network compared to coherence-based methods.