Joint source separation of simultaneous EEG-fMRI recording in two experimental conditions using common spatial...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Simultaneous electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) offer high spatial and temporal resolution for neuroscientific studies.
• Existing multimodal data-mining methods, like joint Independent Component Analysis (ICA), primarily analyze single experimental conditions.
• There is a need for methods that can extract distinct neural activities across multiple conditions from combined EEG-fMRI data.

Purpose of the Study:

• To introduce a novel data decomposition method, joint common spatial pattern (jCSP), for analyzing simultaneous EEG-fMRI data.
• To develop a method capable of identifying brain source activities that differ significantly between experimental conditions.
• To reveal distinctive group-level jCSP patterns through clustering analysis.

Main Methods:

• Proposed the joint common spatial pattern (jCSP) method, which leverages inter-conditional differences in brain source activity.
• Applied jCSP to a simultaneous EEG-fMRI dataset from 21 subjects.
• Utilized a group analysis with clustering to identify common jCSP patterns across participants.

Main Results:

• The jCSP method successfully separated source activities with strong discriminative power between conditions.
• A distinct dynamic pattern was identified, linking EEG alpha power and fMRI signals during eyes-open resting-state.
• Group analysis revealed reproducible jCSP patterns.

Conclusions:

• The jCSP method is effective for source separation in multimodal EEG-fMRI data, particularly for identifying condition-specific activities.
• This approach enhances the analysis of brain dynamics by exploiting differences between experimental states.
• The findings highlight a specific neural signature associated with the eyes-open resting state.