Related Experiment Video
Updated: Jul 14, 2026
09:36
Extracting Visual Evoked Potentials from EEG Data Recorded During fMRI-guided Transcranial Magnetic Stimulation
Published on: May 12, 2014
Removal of EEG noise and artifact using blind source separation.
S P Fitzgibbon1, D M W Powers, K J Pope
1Cognitive Neuroscience Laboratory, School of Psychology, Flinders University, Adelaide, South Australia.
Summary
Blind signal separation (BSS) effectively removes EEG contamination, but performance varies by contamination type and algorithm. Principal components analysis excels with high-amplitude noise, while other methods suit lower-amplitude artifacts.
Area of Science:
- Neuroscience
- Signal Processing
- Biomedical Engineering
Background:
- Electroencephalography (EEG) is susceptible to various artifacts.
- Blind Signal Separation (BSS) algorithms are utilized for artifact removal in EEG data.
- Objective comparison of BSS algorithm performance on diverse EEG contamination types is needed.
Purpose of the Study:
- To develop and apply a novel framework for evaluating BSS algorithms in EEG.
- To compare the relative performance of different BSS algorithms for common EEG contamination.
- To identify which BSS algorithms are most effective for specific contamination types and levels.
Main Methods:
- A realistic EEG simulation incorporating known signals and contamination was created.
- A novel framework with an objective performance metric was developed for BSS evaluation.
- Multiple BSS algorithms were tested against simulated muscle, blink, saccadic, and tracking artifacts.
Main Results:
- BSS is effective for EEG artifact removal, but performance is highly dependent on contamination type, amplitude, and algorithm choice.
- BSS demonstrated strong performance in separating muscle and blink artifacts.
- Principal Components Analysis (PCA) performed well for high-amplitude contamination, while Second-Order Blind Identification (SOBI) and Infomax were better for lower-amplitude artifacts.
Conclusions:
- BSS is a powerful tool for cleaning EEG data, but algorithm selection is critical.
- The effectiveness of BSS varies significantly across different artifact types and signal-to-noise ratios.
- The developed framework provides a robust method for assessing BSS algorithm performance in EEG analysis.
