Clustering of Early Cortical Responses to Median Nerve Stimulation from Average and Single Trial MEG and EEG Signals

Abstract

Median nerve electrical stimulation (MNES) produces early and strong averaged magnetoencephalography (MEG) or electroencephalography (EEG) signals, despite considerable single trial (ST) variability, demonstrated in separate MEG and EEG studies. Here, simultaneous MEG/EEG recordings are used to assess whether same or different aspects of ST variability are influencing EEG and MEG. Clustering techniques provided groupings for the ST timeseries for cortical responses to MNES derived from one modality. These groupings were applied to the corresponding ST timeseries derived from the other modality to quantify the similarity in variability captured by MEG and EEG signals. Estimates of early cortical activity elicited by MNES derived from MEG and EEG signals were very similar, provided ongoing mu rhythm was removed. Similarity between EEG and MEG estimates included both results based on average signals and measures of ST variability. Either MEG or EEG can provide a robust measure of the early cortical activity elicited by MNES as well as of its variability. Reliable indices of early cortical responses to MNES can be derived from either MEG or EEG data. These indices can be based on average signals, as is routinely done with clinical EEG, but it could also rely on hitherto little utilized measures of ST variability.