Slow EEG pattern predicts reduced intrinsic functional connectivity in the default mode network: An inter-subject analysis

Abstract

The last two decades have witnessed great progress in mapping neural networks associated with task-induced brain activation.More recently, identification of resting state networks (RSN) paved the way to investigate spontaneous task-unrelated brain activity. The cardinal features characterising RSN are low-frequency fluctuations of blood oxygenation level dependent (BOLD) signals synchronised between spatially distinct, but functionally connected brain areas. Simultaneous EEG/fMRI has been previously deployed to study the neurophysiological signature of RSN by comparing EEG power with BOLD amplitudes. We hypothesised that band-limited EEG power may be directly related to network-specific functional connectivity (FC) of BOLD signal time courses. Hence, we studied the association between individual EEG signature and FC in a core RSN, the so-called default mode network (DMN).Combined EEG/fMRI data of 20 healthy volunteers collected during a 15-minute rest period were analysed. Using an inter-subject analysis design, we demonstrated a network and frequency specific relation between RSN FC and EEG. In a multiple regression model, EEG band-powers explained 70% of DMN FC variance, with significant partial correlations of DMN FC to delta (r=−0.73) and beta (r=0.53) power. The identified EEG pattern has been previously associated with increased alertness. Conversely, an established EEG-derived sedation index (spectral edge frequency SEF95) closely correlated with DMN FC. The study presents an approach that opens a new perspective to EEG/fMRI correlation. Direct evidence was provided for a distinct neurophysiological correlate of DMN FC. This finding further validates the biological relevance of network-specific intrinsic FC and provides an initial neurophysiological basis for interpreting studies of DMN FC alterations.