Graph theoretical analysis of resting-state MEG data: Identifying interhemispheric connectivity and the default mode

Area of Science:

• Neuroimaging
• Brain Connectivity
• Magnetoencephalography (MEG)

Background:

• Resting-state magnetoencephalography (MEG) studies face challenges in demonstrating interhemispheric connectivity and the default mode network (DMN).
• Previous MEG analyses used seed-based correlations, but graph theoretic approaches offer a more comprehensive view of network topology.

Purpose of the Study:

• To compare graph theoretic brain connectivity maps derived from MEG data with and without signal leakage correction.
• To evaluate the impact of signal leakage correction on identifying interhemispheric connectivity and the default mode network (DMN) in resting-state MEG.

Main Methods:

• Acquired 8-minute resting-state eyes-open MEG data from 22 adolescent males.
• Processed MEG data using an automated pipeline, projected to source space, and applied time-domain signal leakage correction.
• Performed voxel-wise correlation analysis and generated graph theoretic degree maps, identifying hubs and assessing interhemispheric connectivity using laterality indices.

Main Results:

• Uncorrected MEG data produced symmetric, midline networks resembling fMRI, but lacked significant interhemispheric connectivity.
• Signal leakage correction revealed the default mode network (DMN) with hubs in the posterior cingulate and biparietal areas.
• Corrected MEG data demonstrated robust interhemispheric connectivity in graph theoretic analyses.

Conclusions:

• Graph theoretic analysis of uncorrected MEG data yields artifactual symmetric networks due to signal leakage.
• Signal leakage correction is essential for accurately mapping resting-state brain networks, including the DMN and interhemispheric connections, using MEG.
• Corrected MEG analysis provides reliable insights into functional brain connectivity.