Purpose: Although fMRI constrained EEG/MEG source imaging can enhance spatiotemporal resolutions of functional neuroimaging, it has been reported that hard fMRI constraint can result in misidentification of neuronal sources if severe mismatches exist between fMRI activations and EEG/MEG sources. In our previous works, we proposed an approach to address this issue, which automatically adjusts the strength of fMRI constraints by considering the mismatch level. The previous studies proved to be useful particularly when one wants to obtain actual EEG/MEG source locations and uses fMRI prior information as an auxiliary tool to enhance focality of the distributed sources. However, some loss of concentration in the reconstructed images was inevitable when distinct mismatches existMethods: In this study, instead of automatically extending the prior activation regions, we classified and labeled distinct prior activation regions, compared each of the fMRI prior activation regions with each of the thresholded EEG/MEG activation regions, excluded the matched EEG/MEG activation regions, and produced modified prior activation regions having smaller areas than the extended prior activation regions obtained using the conventional approach. Results: A series of realistic EEG simulations, assuming fMRI invisible and discrepancy sources, showed that the improved approach not only compensated for the distortions due to the mismatched activations, but also maintained the high spatial resolution of the fMRI constrained EEG/MEG source imaging. Conclusions: Our simulation results demonstrated that the proposed technique can be a promising option to deal with mismatches between fMRI and EEG/MEG sources in the fMRI constrained EEG/MEG source imaging.
- Cortical source imaging
- Electroencephalography (EEG)
- Functional magnetic resonance imaging (fMRI)
- Magnetoencephalography (MEG)
- Multimodal neuroimaging