Multifocal slow potential generation revealed by high-resolution EEG and current density reconstruction

Abstract

In this work we used high-resolution EEG (123 channels) and current density reconstruction (CDR) to analyze the generators of slow potentials (SPs) in 31 healthy individuals. SPs were obtained during a task-performance feedback anticipation paradigm. The task consisted of a visual paired-associate memory test, with correct performance on single trials indicated by pleasant visual stimuli and incorrect performance by an unpleasant sound. We used realistic models of each subject's head based on their magnetic resonance images (MRIs) to estimate the potentials in the intracranial compartments and to define the source space using individual cortical geometry. Source reconstruction was performed by an Lp-norm minimization algorithm. Results showed a multifocal pattern of current density foci in various association cortices, including prefrontal areas 9 and 10 of Brodmann in all subjects. Posterior cortical areas also contributed importantly to the SP, for instance extrastriate area 19 and parietal area 7, in 90% of the subjects. According to our modeling, we conclude that even the pure stimulus-anticipation SP obtained here, as opposed to traditional motor-task contigent negative variation (CNVs), is not exclusively prefrontal in origin, being generated by multiple association areas. We discuss our results with respect to new possibilities in large-scale cortical physiology and with respect to their application in psychiatry.