No abstract

Objective: The present study examined the benefit of rapid alternation of EEG and fMRI (a common strategy for avoiding artifact caused by rapid switching of MRI gradients) for detecting experimental modulations of ERPs in combined EEG-fMRI. The study also assessed the advantages of aiding the extraction of specific ERP components by means of signal decomposition using Independent Component Analysis (ICA).Methods: 'Go-nogo' task stimuli were presented either during fMRI scanning or in the gaps between fMRI scans, resulting in 'gradient' and 'no gradient' ERPs. 'Go-nogo' differences in the N2 and P3 components were subjected to conventional ERP analysis, as well as single-trial and reliability analyses.Results: Comparable N2 and P3 enhancement on 'nogo' trials was found in the 'gradient' and 'no-gradient' ERPs. ICA-based signal decomposition resulted in better validity (as indicated by topography), greater stability and lower measurement error of the predicted ERP effects. Conclusions:While there was little or no benefit of acquiring ERPs in the gaps between fMRI scans, ICA decomposition did improve the detection of experimental ERP modulations.Significance: Simultaneous and continuous EEG-fMRI acquisition is preferable to interleaved protocols. ICA-based decomposition is useful not only for artifact cancellation, but also for the extraction of specific ERP components. ABSTRACTObjective: The present study examined the benefit of rapid alternation of EEG and fMRI (a common strategy for avoiding artifact caused by rapid switching of MRI gradients) for detecting experimental modulations of ERPs in combined EEG-fMRI. The study also assessed the advantages of aiding the extraction of specific ERP components by means of signal decomposition using Independent Component Analysis (ICA).Methods: 'Go-nogo' task stimuli were presented either during fMRI scanning or in the gaps between fMRI scans, resulting in 'gradient' and 'no gradient' ERPs. 'Go-nogo' differences in the N2 and P3 components were subjected to conventional ERP analysis, as well as single-trial and reliability analyses.Results: Comparable N2 and P3 enhancement on 'nogo' trials was found in the 'gradient' and 'no-gradient' ERPs. ICA-based signal decomposition resulted in better validity (as indicated by topography), greater stability and lower measurement error of the predicted ERP effects. Conclusions:While there was little or no benefit of acquiring ERPs in the gaps between fMRI scans, ICA decomposition did improve the detection of experimental ERP modulations.Significance: Simultaneous and continuous EEG-fMRI acquisition is preferable to interleaved protocols. ICA-based decomposition is useful not only for artifact cancellation, but also for the extraction of specific ERP components.