A Novel Bayesian Approach for EEG Source Localization.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• EEG source localization is crucial for understanding brain activity.
• Efficient localization requires appropriate regularization techniques.
• Existing methods may struggle with focal or sparse brain activations.

Purpose of the Study:

• To propose a novel sparse prior for improved EEG source localization.
• To develop an efficient and tractable algorithm using variational Bayesian methods.
• To extend the method for group structures and spatially extended sources.

Main Methods:

• Bayesian framework utilizing a novel sparse prior distribution.
• Variational Bayesian (VB) inference for an iterative, closed-form algorithm.
• Validation using synthetic and real EEG data from auditory and visual stimuli.

Main Results:

• The proposed sparse prior favors focal EEG sources, enhancing localization accuracy.
• Experimental results demonstrate state-of-the-art performance compared to existing methods.
• The method shows superior accuracy, especially when few brain regions are active.

Conclusions:

• The novel sparse prior offers more accurate EEG source localization than current approaches.
• The VB framework provides an efficient and effective computational solution.
• The method is robust and applicable to various EEG source configurations.