Online Program Home
My Program

Abstract Details

Activity Number: 73
Type: Contributed
Date/Time: Sunday, July 31, 2016 : 4:00 PM to 5:50 PM
Sponsor: Section on Statistical Learning and Data Science
Abstract #320460
Title: Robust Clustering Methods for Time-Evolving Brain Signals
Author(s): Tianbo Chen* and Ying Sun and Carolina Euan and Hernando Ombao
Companies: KAUST and King Abdullah University of Science and Technology and CIMAT and University of California at Irvine
Keywords: EEG data ; Spectral analysis ; Functional median ; Time series clustering ; Robustness ; Time-evolving

Brain activity following stimulus presentation and during resting state are often the result of highly coordinated responses of large numbers of neurons both locally and globally. Coordinated activity of neurons can give rise to oscillations which are captured by electroencephalograms (EEG). In this work, we propose robust clustering methods for identifying synchronized brain regions where the EEGs show similar oscillations or waveforms. We identify the functional median of the smoothed periodograms using band depth for each EEG channel. Then, the channels are clustered by two different measures, the distance for the median spectra, and the coherence that measures the correlation in channels. Our simulation studies show that the proposed median-based clustering algorithm is more robust in identifying the true clusters compared to functional means when outliers are present. When applied to resting state EEG data, the method partly confirms the segmentation based on the anatomy of the cortical surface. In addition, we illustrate the dynamics of spectrally synchronized brain regions during resting state by visualizing the time-evolving clusters of the EEG channels in 3D environment.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2016 program

Copyright © American Statistical Association