An improved artifacts removal method for high dimensional EEG

Hou, Jing; Morgan, Kyle; Tucker, Don M.; Konyn, Amy; Poulsen, Catherine; Tanaka, Yasuhiro; Anderson, Erik; Luu, Phan

doi:10.1016/j.jneumeth.2016.05.003

Cited by 13 publications

(11 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Secondly, we used the default setting of the Net Station (EGI) to analyze the EEG data because it is approved as a reliable method for the artifact detection [Luu et al, 2011[Luu et al, , 2016Liang et al, 2017]. Although independent component analysis (ICA) has become popular in artifact removal for the analysis of EEG data and possibly improves the accuracy of the results [Hou et al,2016], the ICA approach is usually required to identify artifact components, following decomposition based on either spatial topographies or temporal characteristics or both, which is a subjective process. As a result, the biased ICA results could be obtained by subjective errors.…”

Section: Discussionmentioning

confidence: 99%

Event-Related Potential Evidence of Enhanced Visual Processing in Auditory-Associated Cortex in Adults with Hearing Loss

et al. 2020

View full text Add to dashboard Cite

Objective: The present study investigated the characteristics of visual processing in the auditory-associated cortex in adults with hearing loss using event-related potentials. Methods: Ten subjects with bilateral postlingual hearing loss were recruited. Ten age-and sex-matched normal-hearing subjects were included as controls. Visual ("sound" and "non-sound" photos)-evoked potentials were performed. The P170 response in the occipital area as well as N1 and N2 responses in FC3 and FC4 were analyzed. Results: Adults with hearing loss had higher P170 amplitudes, significantly higher N2 amplitudes, and shorter N2 latency in response to "sound" and "non-sound" photo stimuli at both FC3 and FC4, with the exception of the N2 amplitude which responded to "sound" photo stimuli at FC3. Further topographic mapping analysis revealed that patients had a large difference in response to "sound" and "non-sound" photos in the right frontotemporal area, starting from approximately 200 to 400 ms. Localization of source showed the difference to be located in the middle frontal gyrus region (BA10) at around 266 ms. Conclusions: The significantly stronger responses to visual stimuli indicate enhanced visual processing in the auditoryassociated cortex in adults with hearing loss, which may be attributed to cortical visual reorganization involving the right frontotemporal cortex.

show abstract

Section: Discussionmentioning

confidence: 99%

Event-Related Potential Evidence of Enhanced Visual Processing in Auditory-Associated Cortex in Adults with Hearing Loss

et al. 2020

View full text Add to dashboard Cite

show abstract

“…More recently, Chang, Lim & Im (2016) proposed a new method for the real-time detection of eyeblink artefacts that relies solely on an automatic thresholding algorithm, while Kilicarslan, Grossman & Contreras-Vidal, 2016 extended the use of an adaptive noise cancelling scheme for the detection and removal of ocular artefacts and signal drifts, which are the main sources of EEG contamination in brain–machine interface (BMI) applications. Zou, Nathan & Jafari (2016) proposed a method based on the hierarchical clustering of IC features to detect and remove both physiological and non-physiological artefacts from low spatial resolution EEG recordings for BCI applications, and Hou et al (2016) introduced a modified ICA approach for high density EEG recordings to automatically identify eyeblink components using an artefact relevance index calculated by template matching of each IC. Most recently, Radüntz et al (2017) extended their earlier method to classify artefactual and non-artefactual ICs by using pre-selected features of the IC topoplot patterns and power spectra as input to several previously trained machine learning algorithms.…”

Section: Introductionmentioning

confidence: 99%

“…Although very promising, the above listed methods are affected by limitations that prevent their general application to EEG datasets recorded with different types of electrodes (wet or dry) and layouts. Indeed, some methods considered only a reduced number of IC features (e.g., Barbati et al, 2004 ; Halder et al, 2007 ; Viola et al, 2009 ; Mognon et al, 2011 ; Zou, Nathan & Jafari, 2016 ), heavily manipulated the EEG data to extract the input features to the classifier ( Radüntz et al, 2017 ), or focused on the identification of well-defined artefacts only, such as EMG, EOG and ECG artefacts, often using simultaneously recorded artefactual signals (e.g., Halder et al, 2007 ; Viola et al, 2009 ; Nolan, Whelan & Reilly, 2010 ; Mognon et al, 2011 ; Winkler, Haufe & Tangermann, 2011 ; Chang, Lim & Im, 2016 ; Kilicarslan, Grossman & Contreras-Vidal, 2016 ; Hou et al, 2016 ). Other methods were developed for highly specific applications, such as ictal scalp EEG ( LeVan, Urrestarazu & Gotman, 2006 ), visual evoked potentials ( Nolan, Whelan & Reilly, 2010 ), BCI ( Daly et al, 2015 ; Zou, Nathan & Jafari, 2016 ) and BMI applications ( Kilicarslan, Grossman & Contreras-Vidal, 2016 ).…”

Section: Introductionmentioning

confidence: 99%

“…Other methods were developed for highly specific applications, such as ictal scalp EEG ( LeVan, Urrestarazu & Gotman, 2006 ), visual evoked potentials ( Nolan, Whelan & Reilly, 2010 ), BCI ( Daly et al, 2015 ; Zou, Nathan & Jafari, 2016 ) and BMI applications ( Kilicarslan, Grossman & Contreras-Vidal, 2016 ). Another common limitation is the prevalent use of simulated artefacts for the validation of the classification system; however, simulated artefacts cannot represent the entire variety of real artefacts possibly affecting brain activity recordings (e.g., Barbati et al, 2004 ; Halder et al, 2007 ; Nolan, Whelan & Reilly, 2010 ; Hou et al, 2016 ). Finally, some methods adopted a simple linear discriminant approach for the classification of the ICs (e.g., Halder et al, 2007 ; Radüntz et al, 2015 ) while others were developed for low-density EEG datasets only (e.g., Zou, Nathan & Jafari, 2016 ) or were tailored to brain monitoring techniques with low temporal resolution, such as fMRI (e.g., De Martino et al, 2007 ; Mantini et al, 2008 ).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A new ICA-based fingerprint method for the automatic removal of physiological artifacts from EEG recordings

Tamburro

Fiedler

Stone

et al. 2018

PeerJ

View full text Add to dashboard Cite

BackgroundEEG may be affected by artefacts hindering the analysis of brain signals. Data-driven methods like independent component analysis (ICA) are successful approaches to remove artefacts from the EEG. However, the ICA-based methods developed so far are often affected by limitations, such as: the need for visual inspection of the separated independent components (subjectivity problem) and, in some cases, for the independent and simultaneous recording of the inspected artefacts to identify the artefactual independent components; a potentially heavy manipulation of the EEG signals; the use of linear classification methods; the use of simulated artefacts to validate the methods; no testing in dry electrode or high-density EEG datasets; applications limited to specific conditions and electrode layouts.MethodsOur fingerprint method automatically identifies EEG ICs containing eyeblinks, eye movements, myogenic artefacts and cardiac interference by evaluating 14 temporal, spatial, spectral, and statistical features composing the IC fingerprint. Sixty-two real EEG datasets containing cued artefacts are recorded with wet and dry electrodes (128 wet and 97 dry channels). For each artefact, 10 nonlinear SVM classifiers are trained on fingerprints of expert-classified ICs. Training groups include randomly chosen wet and dry datasets decomposed in 80 ICs. The classifiers are tested on the IC-fingerprints of different datasets decomposed into 20, 50, or 80 ICs. The SVM performance is assessed in terms of accuracy, False Omission Rate (FOR), Hit Rate (HR), False Alarm Rate (FAR), and sensitivity (p). For each artefact, the quality of the artefact-free EEG reconstructed using the classification of the best SVM is assessed by visual inspection and SNR.ResultsThe best SVM classifier for each artefact type achieved average accuracy of 1 (eyeblink), 0.98 (cardiac interference), and 0.97 (eye movement and myogenic artefact). Average classification sensitivity (p) was 1 (eyeblink), 0.997 (myogenic artefact), 0.98 (eye movement), and 0.48 (cardiac interference). Average artefact reduction ranged from a maximum of 82% for eyeblinks to a minimum of 33% for cardiac interference, depending on the effectiveness of the proposed method and the amplitude of the removed artefact. The performance of the SVM classifiers did not depend on the electrode type, whereas it was better for lower decomposition levels (50 and 20 ICs).DiscussionApart from cardiac interference, SVM performance and average artefact reduction indicate that the fingerprint method has an excellent overall performance in the automatic detection of eyeblinks, eye movements and myogenic artefacts, which is comparable to that of existing methods. Being also independent from simultaneous artefact recording, electrode number, type and layout, and decomposition level, the proposed fingerprint method can have useful applications in clinical and experimental EEG settings.

show abstract

“…The blind source separation [ 22 ] has been discussed widely on the issue of a linear mixture signal, and Independent Component Analysis (ICA) and Principal Component Analysis (PCA) are representative methods. In the case of the EEG signal decomposition, those methods were frequently applied [ 23 , 24 ], especially in the offline analysis. In PCA, the EEG components are decomposed on space/time basis, while, as disadvantage, it is difficult to reconstruct overall signals by the linear combination of principal components (PCs) because of the ignorance of signals with small amplitudes and irregular changes.…”

Section: Introductionmentioning

confidence: 99%

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

Singh

Wagatsuma

2017

Computational and Mathematical Methods in Medicine

View full text Add to dashboard Cite

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.

show abstract

An improved artifacts removal method for high dimensional EEG

Cited by 13 publications

References 27 publications

Event-Related Potential Evidence of Enhanced Visual Processing in Auditory-Associated Cortex in Adults with Hearing Loss

Event-Related Potential Evidence of Enhanced Visual Processing in Auditory-Associated Cortex in Adults with Hearing Loss

A new ICA-based fingerprint method for the automatic removal of physiological artifacts from EEG recordings

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

Contact Info

Product

Resources

About