Data-driven analysis of simultaneous EEG/fMRI using an ICA approach.

Area of Science:

• Neuroscience
• Cognitive Neuroscience
• Neuroimaging

Background:

• Electroencephalography (EEG) offers millisecond temporal resolution for neural event analysis.
• Functional magnetic resonance imaging (fMRI) has second-scale temporal resolution, limiting precise event-related analysis.
• Combining EEG and fMRI is sought to leverage their complementary temporal and spatial resolutions.

Purpose of the Study:

• To develop and validate a data-driven method for analyzing simultaneous EEG/fMRI data.
• To automatically select task-specific electrophysiological independent components (ICs).
• To investigate the relationship between inhibition-related EEG components and fMRI BOLD signals during a visual Go/Nogo task.

Main Methods:

• Single-trial simultaneous EEG/fMRI analysis of a visual Go/Nogo task.
• Development of a data-driven procedure to automatically select task-specific electrophysiological independent components (ICs).
• Correlation analysis between single-trial EEG IC amplitude variability and fMRI BOLD signals.

Main Results:

• The novel method successfully identified task-specific EEG components.
• Positive correlations between fMRI BOLD signal and EEG-derived regressors were found in fronto-striatal regions.
• These correlations varied across different phases of task execution, with earlier phases showing stronger effects.

Conclusions:

• Automated selection of Nogo-related ICs in single-subject EEG/fMRI analysis reveals distinct BOLD responses related to different task phases.
• The developed method is generalizable to other events, such as visual responses.
• This approach enhances the ability to study neurophysiological processes by combining EEG and fMRI.