Modified Empirical Mode Decomposition to process EEG signal including artefacts
-
1
The University of West Bohemia, Department of Computer Science and Engineering, Czechia
The continuous EEG signal is considered non-stationary. Therefore it is necessary to use time-frequency domain methods to process the EEG signal. One of such methods is Hilbert-Huang transform (HHT). HHT transform was designed especially for processing non-stationary signals. The first part of HHT is the Empirical Mode Decomposition (EMD), which decomposes the signal into Intrinsic Mode Functions (IMFs) defined by the signal itself.
During the EMD we locate local extrema (maxima and minima) and interpolate a cubic spline running between them, so we can get the upper and lower envelope. But when we need to create the complete envelope to cover the whole signal, we have to add additional extremum points, because detected extrema don’t cover the beginning and the end of the signal. Mirror method and Slope-based method in were designed to deal with this task. But these methods don’t reflect specifics of the EEG signal, such as artefacts (significant change of amplitude and frequency). When the artefact is located near the beginning or the end of the signal, then the estimated additional extrema don’t fit to the trend of the signal.
I have proposed two methods, which estimate additional extrema to create complete envelopes. The first method includes first and last sample of the signal to the sets of minima and maxima. The second one is based on the Mirror method. It calculates new extrema prior the first sample/past the last sample of the signal as the mirror images of the first two extrema symmetric to the first/last sample of the signal. If the first estimated extremum is a minimum, we have to ensure, that its value isn’t greater than the value of first sample. If the value of the minimum is greater, then we replace the minimum value with the first sample.
With these two methods the EEG signal could be processed without previous artefact rejection. But there is significant difference between the modified mirror method and the first-last method. Modified mirror method is approximately six times faster than the first-last method. It means that the number of modified mirror method iterations during the sifting process (part of EMD) is six times smaller.
Keywords:
brain machine interface
Conference:
4th INCF Congress of Neuroinformatics, Boston, United States, 4 Sep - 6 Sep, 2011.
Presentation Type:
Poster Presentation
Topic:
Brain machine interface
Citation:
Ciniburk
J
(2011). Modified Empirical Mode Decomposition to process EEG signal including artefacts.
Front. Neuroinform.
Conference Abstract:
4th INCF Congress of Neuroinformatics.
doi: 10.3389/conf.fninf.2011.08.00076
Copyright:
The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers.
They are made available through the Frontiers publishing platform as a service to conference organizers and presenters.
The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated.
Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed.
For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions.
Received:
17 Oct 2011;
Published Online:
19 Oct 2011.
*
Correspondence:
Dr. Jindrich Ciniburk, The University of West Bohemia, Department of Computer Science and Engineering, Pilsen, Czechia, ciniburk@kiv.zcu.cz