A new multimodal imaging strategy for integrating fMRI with EEG

Area of Science:

• Neuroimaging
• Biophysics
• Computational Neuroscience

Background:

• Multimodal functional neuroimaging combining fMRI and EEG is crucial for understanding brain activity.
• fMRI-EEG mismatches, stemming from differences in electric and hemodynamic signals, introduce distortions.
• Accurate source localization is essential for interpreting combined fMRI and EEG data.

Purpose of the Study:

• To propose and evaluate a novel two-step algorithm, the "Twomey algorithm," for fMRI-constrained cortical source imaging.
• To address and correct distortions caused by fMRI-EEG mismatches in multimodal neuroimaging.
• To assess the robustness of the Twomey algorithm against invalid fMRI priors.

Main Methods:

• A two-step approach: 1) imposing an fMRI-derived "hard" spatial prior to solve the EEG inverse problem, and 2) re-entering the result as an initial guess into an EEG least squares fitting with Twomey regularization.
• Comparison with the 90% fMRI-constrained Wiener filter and EEG-alone weighted minimum norm methods.
• Evaluation using randomized computer simulations, examining effects of fMRI invisible sources, extra regions, and displacement.

Main Results:

• The Twomey algorithm, utilizing fMRI priors, achieved higher spatial resolution compared to EEG-alone methods.
• fMRI invisible sources were identified as the primary cause of errors in integrated fMRI-EEG source imaging.
• The Twomey algorithm demonstrated greater robustness than the Wiener filter approach when subjected to invalid fMRI priors.

Conclusions:

• The proposed Twomey algorithm offers a significant advancement for fMRI-constrained cortical source imaging.
• This method effectively corrects fMRI-EEG mismatches, enhancing spatial resolution in multimodal neuroimaging.
• The Twomey algorithm's robustness makes it a valuable tool for integrating fMRI and EEG data.