Connectome spectrum electromagnetic tomography: A method to reconstruct electrical brain source networks at...

Area of Science:

• Neuroimaging
• Computational Neuroscience
• Graph Signal Processing

Background:

• The M/EEG inverse problem limits accurate brain activity localization.
• Functional MRI (fMRI) has low temporal resolution, while M/EEG has poor spatial resolution.
• Integrating multimodal data can overcome individual modality limitations.

Purpose of the Study:

• To introduce Connectome Spectrum Electromagnetic Tomography (CSET) for solving the M/EEG inverse problem.
• To evaluate CSET's accuracy and robustness against existing methods.
• To enable high spatio-temporal resolution neuroimaging.

Main Methods:

• CSET combines diffusion MRI-derived structural connectivity with graph signal processing.
• Simulated EEG signals from fMRI responses were used for validation.
• Two EEG datasets for visual-evoked potentials were analyzed.

Main Results:

• CSET accurately captures realistic neurophysiological patterns, outperforming state-of-the-art methods.
• CSET reconstructs brain responses with enhanced accuracy and robustness to noise.
• The method demonstrates high spatio-temporal resolution.

Conclusions:

• CSET offers a significant advancement in neuroimaging by providing accurate spatio-temporal localization.
• This technique overcomes limitations of fMRI's low sampling frequency and M/EEG's poor spatial resolution.
• CSET empowers neuroscientists to conduct more precise brain activity research.