A Methodical Healthcare Model to Eliminate Motion Artifacts from Big EEG Data

A Methodical Healthcare Model to Eliminate Motion Artifacts from Big EEG Data

Vandana Roy (Research Scholar, Jabalpur Engineering College, Jabalpur, India) and Shailja Shukla (Department of Computer Science, Jabalpur Engineering College, Jabalpur, India)
Copyright: © 2017 |Pages: 19
DOI: 10.4018/JOEUC.2017100105
OnDemand PDF Download:
No Current Special Offers


The Big data as Electroencephalography (EEG) can induce by artifacts during acquisition process which will obstruct the features and quality of interest in the signal. The healthcare diagnostic procedures need strong and viable biomedical signals and elimination of artifacts from EEG is important. In this research paper, an improved ensemble approach is proposed for single channel EEG signal motion artifacts removal. Ensemble Empirical Mode Decomposition and Canonical Correlation Analysis (EEMD-CCA) filter combination are applied to remove artifact effectively and further Stationary Wavelet Transform (SWT) is applied to remove the randomness and unpredictability due to motion artifacts from EEG signals. This new filter combination technique was tested against currently available artifact removal techniques and results indicate that the proposed algorithm is suitable for use as a supplement to algorithms currently in use.
Article Preview

1. Introduction

The design of a big physiological signal is an extensively followed approach that illustrates the present condition of an individual health. To store and handle the Big Data, different technologies are frequently applied in the biomedical and health-care field (Luo & Zhao, 2016) to support healthcare performances. Moreover, various big data approaches and technologies have helped to improve the process performance (Baquero, 2015). Big Data such EEG signals are a good tool for exploring brain activity and are preferred over other large physiological signals because they can be used to detect changes over millisecond time-spans whereas methods such as functional Magnetic Resonance Imaging (fMRI) and Positron Emission Topography (PET) have time resolutions of the order of seconds or minutes. An action potential takes approximately 0.5-130 milliseconds to propagate across a single neuron. Other methods detect by indirect markers of electrical activity in the brain; for example, Single-Photon Emission Computerized Tomography (SPECT) and fMRI record changes in blood flow and PET records changes in metabolic activity, whereas EEG measures the electrical activity of the brain directly.

It is evident that measurement of big physiological signals, even in a surgical environment, is subject to some noise, which in medical terms may be referred to as artifacts. These artifacts are unwanted signals generated by unregulated sources besides the source of interest. There are two main sources of artifacts in neural signals other than the machine and environment. These are the muscular and ocular activities of the individual, which generate low-amplitude, low-frequency electrical pulses which appear in filter range of sensors and recording equipment. Artifact rejection from Big Data is, therefore, a fundamental research topic and is well researched in (Sweeney & McLoone, 2012). This paper considers the cases of artifacts caused by motion in EEG sample taken from Big Data. Numerous applications of Independent Component Analysis (ICA), Wavelet and adaptive filters are proposed in the field of EEG artifact removal (Robertson, 2010) till today.

A common approach for the EEG signal is to reject all EEG epochs containing signal amplitude larger than a specified value. These schemes are firm and do not allow for any adaptation, which results in loss of meaningful data. A component-based, automated separator of artifacts is required to overcome this problem using the linear decomposition of signals into source components. The components give the nature of source information. Therefore, the combinations of the artifact components as a separate source and reconstruction of signals without this artifact source are claimed as artifact-free information. The Fast ICA algorithm using a 30s window size has applied by (Arora & Sachdeva, 2012) using the EEGLAB platform. They developed an automated system for artifact removal based on ICA and Bayesian Classification. The ensemble empirical mode decomposition with canonical correlation analysis (EEMD-CCA) technique for single channel artifact removal has introduced by (Sweeney & McLoone, 2013). The single channel signal is converted into a multi-dimensional signal using the EEMD technique. Then the CCA technique, a second-order statistics technique, is applied to segregate the artifact components from the input signal. Moreover, performance has been shown to be superior to the existing wavelet de-noising and EEMD-ICA technique. A statistical method based on wavelet transform to mimic ocular artifacts in EEG signal have presented by (Kumar & Vimal, 2008). The two techniques for artifact removal, single-channel ICA (SCICA) and wavelet-ICA (WICA) have studied by authors (Mijovic & Van Huffel, 2010) and also compared them with the application of the EEMD-ICA algorithm to a single channel EEG signal. The EEMD-ICA algorithm was tested on simulated data and then applied to real EEG and EMG data for comparison. It was concluded that the SCICA algorithm performed worst when performance was measured in terms of root mean square error (RMSE). The WICA algorithm performed worse in the simulations than SCICA algorithm although its performance was comparable to the EEMD-ICA technique. But still, the SWT has not been widely explored in this context till now.

Complete Article List

Search this Journal:
Volume 34: 3 Issues (2022): 2 Released, 1 Forthcoming
Volume 33: 6 Issues (2021)
Volume 32: 4 Issues (2020)
Volume 31: 4 Issues (2019)
Volume 30: 4 Issues (2018)
Volume 29: 4 Issues (2017)
Volume 28: 4 Issues (2016)
Volume 27: 4 Issues (2015)
Volume 26: 4 Issues (2014)
Volume 25: 4 Issues (2013)
Volume 24: 4 Issues (2012)
Volume 23: 4 Issues (2011)
Volume 22: 4 Issues (2010)
Volume 21: 4 Issues (2009)
Volume 20: 4 Issues (2008)
Volume 19: 4 Issues (2007)
Volume 18: 4 Issues (2006)
Volume 17: 4 Issues (2005)
Volume 16: 4 Issues (2004)
Volume 15: 4 Issues (2003)
Volume 14: 4 Issues (2002)
Volume 13: 4 Issues (2001)
Volume 12: 4 Issues (2000)
Volume 11: 4 Issues (1999)
Volume 10: 4 Issues (1998)
Volume 9: 4 Issues (1997)
Volume 8: 4 Issues (1996)
Volume 7: 4 Issues (1995)
Volume 6: 4 Issues (1994)
Volume 5: 4 Issues (1993)
Volume 4: 4 Issues (1992)
Volume 3: 4 Issues (1991)
Volume 2: 4 Issues (1990)
Volume 1: 3 Issues (1989)
View Complete Journal Contents Listing