Table of Contents Author Guidelines Submit a Manuscript
Computational and Mathematical Methods in Medicine
Volume 2017, Article ID 1861645, 17 pages
https://doi.org/10.1155/2017/1861645
Research Article

A Removal of Eye Movement and Blink Artifacts from EEG Data Using Morphological Component Analysis

1Graduate School of Life Science and Systems Engineering, Kyushu Institute of Technology (KYUTECH), Kitakyushu, Japan
2RIKEN Brain Science Institute, Wako, Japan
3Artificial Intelligence Research Center, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Japan

Correspondence should be addressed to Balbir Singh; pj.ca.hcetuyk.niarb.ude@hgnis-riblab-anawa

Received 3 October 2016; Revised 25 November 2016; Accepted 15 December 2016; Published 17 January 2017

Academic Editor: Michele Migliore

Copyright © 2017 Balbir Singh and Hiroaki Wagatsuma. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

EEG signals contain a large amount of ocular artifacts with different time-frequency properties mixing together in EEGs of interest. The artifact removal has been substantially dealt with by existing decomposition methods known as PCA and ICA based on the orthogonality of signal vectors or statistical independence of signal components. We focused on the signal morphology and proposed a systematic decomposition method to identify the type of signal components on the basis of sparsity in the time-frequency domain based on Morphological Component Analysis (MCA), which provides a way of reconstruction that guarantees accuracy in reconstruction by using multiple bases in accordance with the concept of “dictionary.” MCA was applied to decompose the real EEG signal and clarified the best combination of dictionaries for this purpose. In our proposed semirealistic biological signal analysis with iEEGs recorded from the brain intracranially, those signals were successfully decomposed into original types by a linear expansion of waveforms, such as redundant transforms: UDWT, DCT, LDCT, DST, and DIRAC. Our result demonstrated that the most suitable combination for EEG data analysis was UDWT, DST, and DIRAC to represent the baseline envelope, multifrequency wave-forms, and spiking activities individually as representative types of EEG morphologies.