Related Experiment Video
Updated: Jun 24, 2026
09:32
Cortical Source Analysis of High-Density EEG Recordings in Children
Published on: June 30, 2014
Validation of regression-based myogenic correction techniques for scalp and source-localized EEG
Brenton W McMenamin1, Alexander J Shackman, Jeffrey S Maxwell
1University of MinnesotaTwin Cities, Minneapolis-Saint Paul, Minnesota, USA. mcmen020@umn.edu
Psychophysiology
|March 21, 2009
Summary
Electromyographic artifacts (EMG) in electroencephalography (EEG) can distort brain activity. While some methods reduce EMG on the scalp, none effectively remove it in source-localized EEG data.
Area of Science:
- Neuroscience
- Biomedical Engineering
- Signal Processing
Background:
- Electromyographic artifacts (EMG) significantly contaminate electroencephalography (EEG) recordings.
- These artifacts, originating from cranial muscles, can obscure genuine neural signals or mimic effects, even in the alpha frequency band (8-13 Hz).
- Previous validation of regression-based EMG correction methods in EEG is limited, and their efficacy in source-localized data remains unexplored.
Purpose of the Study:
- To evaluate the effectiveness of four regression-based techniques for correcting electromyographic artifacts in EEG.
- To assess these methods on both scalp-level EEG data and source-localized data using the LORETA algorithm.
- To validate novel within-subject epoch-wise correction methods.
Main Methods:
- EEG data were collected from 17 participants under conditions with varied neurogenic and myogenic activity.
- Four regression-based artifact correction techniques were assessed: between-subjects, between-subjects using difference-scores, within-subjects condition-wise, and within-subject epoch-wise.
- Performance was evaluated using sensitivity and specificity metrics on both scalp EEG and LORETA-derived source-space data.
Main Results:
- The within-subject epoch-wise regression technique demonstrated superior performance in correcting EMG artifacts on the scalp.
- However, none of the evaluated regression-based techniques proved effective in removing EMG artifacts from source-localized EEG data (LORETA).
- The study provides validation for novel epoch-wise correction methods at the scalp level.
Conclusions:
- Current regression-based methods show limited success in mitigating electromyographic artifacts in source-localized EEG.
- Further development of artifact correction techniques is necessary for accurate source analysis in EEG.
- The findings highlight the challenges in cleaning EEG source data from muscle-related noise.
