Related Experiment Video
Updated: Jun 7, 2025
09:32
Cortical Source Analysis of High-Density EEG Recordings in Children
Published on: June 30, 2014
21.3K
Investigating the impact of the regularization parameter on EEG resting-state source reconstruction and functional
F Leone1, A Caporali2, A Pascarella3
1Department of Psychology, Sapienza University of Rome, via dei Marsi 78, Rome, 00185, Italy; IRCCS Fondazione Santa Lucia, via Ardeatina, 354, Rome, 00179, Italy.
Neuroimage
|November 9, 2024
Summary
Finding the right regularization parameter is key for accurate electroencephalography (EEG) source localization and functional connectivity analysis. This study identifies optimal parameters for both, improving EEG data interpretation.
Area of Science:
- Neuroscience
- Biomedical Engineering
- Signal Processing
Background:
- Accurate electroencephalography (EEG) source localization is essential for understanding brain activity, particularly resting-state network dynamics and functional connectivity.
- Minimum norm estimation (MNE) in EEG source analysis faces challenges, notably the selection of an appropriate regularization parameter.
- The optimal regularization for source localization may differ from that required for connectivity analysis, presenting a critical trade-off.
Purpose of the Study:
- To determine the optimal regularization coefficient for EEG source estimation that maximizes reconstruction accuracy across varying signal-to-noise ratios.
- To evaluate the impact of different regularization parameters on source localization and functional connectivity estimation.
- To provide guidance on selecting regularization parameters for specific EEG analysis goals.
Main Methods:
- Simulated synthetic EEG signals from three resting-state networks (Motor, Visual, Dorsal Attention) with varying signal-to-noise ratios.
- Applied minimum norm estimation with a range of regularization parameters.
- Assessed performance using Region Localization Error, source extension, and source fragmentation metrics.
- Validated results with real functional connectivity data.
Main Results:
- The optimal regularization coefficient for accurate functional connectivity estimation was found to be 10-2.
- A regularization coefficient of 10-1 was preferred for analyses focused solely on source localization.
- Performance metrics demonstrated a clear trade-off between source localization accuracy and connectivity estimation based on regularization choice.
Conclusions:
- The choice of regularization parameter significantly impacts EEG source analysis outcomes, affecting both localization and connectivity estimation.
- A regularization value of 10-2 is recommended for optimal functional connectivity analysis using EEG.
- For studies prioritizing source localization accuracy, a regularization value of 10-1 is more appropriate.
