2021
DOI: 10.1101/2021.01.22.427809
|View full text |Cite
Preprint
|
Sign up to set email alerts
|

Automatic classification of ICA components from infant EEG using MARA

Abstract: Automated systems for identifying and removing non-neural ICA components are growing in popularity among adult EEG researchers. Infant EEG data differs in many ways from adult EEG data, but there exists almost no specific system for automated classification of source components from paediatric populations. Here, we adapt one of the most popular systems for adult ICA component classification for use with infant EEG data. Our adapted classifier significantly outperformed the original adult classifier on samples … Show more

Help me understand this report
View published versions

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1

Citation Types

0
3
0

Year Published

2021
2021
2024
2024

Publication Types

Select...
4
3

Relationship

0
7

Authors

Journals

citations
Cited by 9 publications
(3 citation statements)
references
References 41 publications
0
3
0
Order By: Relevance
“…To increase consistency and efficiency for classifying brain and non-brain independent components, automatic classification data processing toolboxes available in EEGLab have been made available. Some of these toolboxes include ICLabel (Pion-Tonachini et al, 2019 ), MARA (Multiple Artifact Rejection Algorithm; Haresign et al, 2021 ), FASTER (Fully Automated Statistical Thresholding for EEG Artifact Rejection; Nolan et al, 2010 ), SASICA (Semi-Automated Selection of Independent Components of the electroencephalogram for Artifact correction; Chaumon et al, 2015 ), ADJUST (Automatic EEG artifact Detection based on the Joint Use of Spatial and Temporal features; Mognon et al, 2011 ), and IC_MARC (Frølich et al, 2015 ). Although these approaches can help to distinguish the brain and non-brain source components from ICA decomposition, visual inspection is still typically advisable.…”
Section: Mobile Eeg Data Processingmentioning
confidence: 99%
“…To increase consistency and efficiency for classifying brain and non-brain independent components, automatic classification data processing toolboxes available in EEGLab have been made available. Some of these toolboxes include ICLabel (Pion-Tonachini et al, 2019 ), MARA (Multiple Artifact Rejection Algorithm; Haresign et al, 2021 ), FASTER (Fully Automated Statistical Thresholding for EEG Artifact Rejection; Nolan et al, 2010 ), SASICA (Semi-Automated Selection of Independent Components of the electroencephalogram for Artifact correction; Chaumon et al, 2015 ), ADJUST (Automatic EEG artifact Detection based on the Joint Use of Spatial and Temporal features; Mognon et al, 2011 ), and IC_MARC (Frølich et al, 2015 ). Although these approaches can help to distinguish the brain and non-brain source components from ICA decomposition, visual inspection is still typically advisable.…”
Section: Mobile Eeg Data Processingmentioning
confidence: 99%
“…First, our events of interest are intrinsically linked with one of the biggest EEG artefacts (i.e., eye movements), and so it is possible that residual artifact in the EEG signal may have contaminated our data. However, our data were processed using algorithms specially designed to clean naturalistic EEG data (66,67), and previous analyses suggest that the electrode locations and frequency bands that we examined should be least affected by artifact, compared with more anterior locations and higher and lower frequencies (68). Additionally, our analyses were carefully designed to preclude this potential confound.…”
Section: Limitations and Strengthsmentioning
confidence: 99%
“…EEG data was pre-processed and cleaned from oculomotor and other contaminatory artefacts using a fully automatic artefact rejection procedure specially designed for naturalistic infant EEG data by Mariott Haresign (66), building on previous related work (59,60). Briefly, this involved the following steps: 1) data were high-pass filtered at 1Hz, 2) line noise at 50Hz was eliminated using the EEGLAB function clean_line.m, 3) data were low-pass filtered at 20Hz, 4) the data were referenced to a robust average reference 5) noisy channels were rejected using the EEGLAB function pop_rejchan.m, 6) the channels identified in the previous stage were then interpolated back, using the EEGLAB function eeg_interp.m, 7) continuous data were automatically rejected (NaN-ed) in a sliding 1s epoch based on a percentage of bad channels (set here at 70% of channels) that exceed 5 standard deviations of the mean channel EEG power, and 8) Independent Component Analyses (ICA) were computed on the continuous data using the EEGLAB function runica.m.…”
Section: Eeg Artefact Rejection and Pre-processingmentioning
confidence: 99%