Paper Details: Downloads: 1224
Serial Number: P1161146012
Title: Denoising EEG: Which Class of ICA Algorithms Produces the Cleaner Signals?
Authors: Janett Walters-Williams and Yan Li
Abstract: Independent component analysis (ICA) is a popular blind source separation (BSS) technique that has proven to be promising for the analysis of EEG data. A number of ICA approaches have been used for EEG data analysis, and even more ICA algorithms exist, however the impact of using different algorithms on the results is largely unexplored. In this paper, we study the performance of four major classes of algorithms for ICA, namely information maximization, maximization of non-gaussianity, joint diagonalization of cross-cumulant matrices, and second-order correlation based methods when they are applied to EEG data. We propose several analysis techniques to evaluate their performance. The results demonstrate that the classes of ICA algorithms based on using mutual information estimated contrast function proved to provide better separation performance. Our results confirm that mutual information estimation creates better performing ICA algorithms resulting in cleaner EEG signals.
Keywords: EEG Signal, Independent Component Analysis, Denoising
Journal/Conference: International Journal of Bioinformatics and Medical Engineering
Volume: 12
Issue: 1
Submission Date: 11/1/2011 12:00:00 AM
Review Date: 9/11/2012 12:00:00 AM
Publishing Date: 9/16/2012 12:00:00 AM
Article Downloads: 1224
Download:

Facebook