A channel rejection method for attenuating motion-related artifacts in EEG recordings during walking

Anderson S. Oliveira*, Bryan R. Schlink, W. David Hairston, Peter König, Daniel P. Ferris

*Kontaktforfatter

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningpeer review

34 Citationer (Scopus)

Abstract

Recording scalp electroencephalography (EEG) during human motion can introduce motion artifacts. Repetitive head movements can generate artifact patterns across scalp EEG sensors. There are many methods for identifying and rejecting bad channels and independent components from EEG datasets, but there is a lack of methods dedicated to evaluate specific intra-channel amplitude patterns for identifying motion-related artifacts. In this study, we proposed a template correlation rejection (TCR) as a novel method for identifying and rejecting EEG channels and independent components carrying motion-related artifacts. We recorded EEG data from 10 subjects during treadmill walking. The template correlation rejection method consists of creating templates of amplitude patterns and determining the fraction of total epochs presenting relevant correlation to the template. For EEG channels, the template correlation rejection removed channels presenting the majority of epochs (>75%) correlated to the template, and presenting pronounced amplitude in comparison to all recorded channels. For independent components, the template correlation rejection removed components presenting the majority of epochs correlated to the template. Evaluation of scalp maps and power spectra confirmed low neural content for the rejected components. We found that channels identified for rejection contained ~60% higher delta power, and had spectral properties locked to the gait phases. After rejecting the identified channels and running independent component analysis on the EEG datasets, the proposed method identified 4.3 ± 1.8 independent components (out of 198 ± 12) with substantive motion-related artifacts. These results indicate that template correlation rejection is an effective method for rejecting EEG channels contaminated with motion-related artifact during human locomotion.

OriginalsprogEngelsk
Artikelnummer225
TidsskriftFrontiers in Neuroscience
Vol/bind11
Udgave nummerAPR
Antal sider17
ISSN1662-4548
DOI
StatusUdgivet - 26 apr. 2017

Fingeraftryk

Dyk ned i forskningsemnerne om 'A channel rejection method for attenuating motion-related artifacts in EEG recordings during walking'. Sammen danner de et unikt fingeraftryk.

Citationsformater