Multimodal integration of high resolution EEG and functional magnetic resonance: a simulation study

In this simulation study, we would like to address some questions related to the use of fMRI a priori constraints in the estimation of the cortical source current density. Namely, we would like to assess the utility to include information as estimated from event-related and block-design fMRI, by using as the dependent variable the correlation between the imposed and the estimated waveforms at the level of cortical region of interests (ROI). A realistic head and cortical surface model was used. Factors used were i) the signal to noise ratio of the scalp simulated data (SNR); ii) the particular inverse operator used to estimate the cortical source activity from the simulated scalp data (INVERSE); iii) the strength of the fMRI priors in the estimation of the current activity (K). Analysis of Variance (ANOVA) results revealed that all the considered factors (SNR, INVERSE, K) significantly afflicts the correlation between the estimated and the simulated cortical activity. For the ROIs analyzed in which a presence of fMRI hotspots were simulated, it was observed that the best estimation of cortical source currents were performed with the inverse operator that use the event-related fMRI information. When the ROI analyzed do not present fMRI hotspots, both minimum norm and fMRI-based inverse operators return statistically equivalent correlation values. Such results open the avenue for the use of fMRI-based inverse operator in the estimation of cortical current strengths from motor and cognitive task in the human brain.