Group Independent Component Analysis for Event-Related EEG Data
Author Information
Author(s): Eichele Tom, Rachakonda Srinivas, Brakedal Brage, Eikeland Rune, Calhoun Vince D.
Primary Institution: University of Bergen
Hypothesis
Can group independent component analysis (ICA) effectively decompose event-related EEG data?
Conclusion
Group ICA is adequate for decomposing single trials with physiological jitter and reconstructs event-related sources with high accuracy.
Supporting Evidence
- Group ICA models yield overall lower reconstruction accuracy than individual ICA.
- Reconstruction accuracy for the entire timecourse decreases with increasing latency jitter.
- Group ICA allows for straightforward population tests for timecourses and topographies.
Takeaway
This study shows how scientists can use a special method to look at brain waves from many people at once, helping them understand how our brains respond to sounds.
Methodology
The study used a group-level temporal ICA model for event-related EEG analysis, employing PCA for data reduction and back-reconstruction of individual data.
Limitations
Responses with poor time/phase-locking are not satisfactorily reconstructed, limiting the method's visibility to evoked activity.
Participant Demographics
32 healthy participants took part in the auditory oddball experiment.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website