Very high density EEG elucidates spatiotemporal aspects of early visual processing

Outcome: TDLC's Marlene Behrmann and Michael Tarr at Carnegie Mellon University (with Pulkit Grover, Shawn Kelly, and Amanda Robinson as lead author) showed that whereas standard human EEG systems based on spatial Nyquist estimates suggest that 20-30 mm electrode spacing suffices to capture neural signals on the scalp, “super-Nyquist” density EEG (“SND”) with Nyquist density (“ND”) arrays can capture neural signals, especially in the high frequency range, with greater accuracy (see Figure 1a for montage of 128 electrodes 14mm apart; 1b classification accuracy from 6-way analysis of the possible stimuli; and 1c for the left/right visual field stimuli of medium).

Impact/benefits: That we obtain better classification accuracy in the temporal and frequency domains provides a proof of concept for the development of super-Nyquist systems and offers the possibility of examining temporal dynamics in cortex in a noninvasive fashion. Such systems might also serve to diagnose various disorders including migraine, sleep and concussion. Also, because EEG less expensive and easier to implement than the similarly time-precise method of magnetoencephalography (MEG) or of functional MRI, there is the possibility that the high-density EEG systems might be exploited in underserviced areas.

Background/Explanation: Relative to functional magnetic resonance imaging (fMRI) and, to a lesser extent, MEG, EEG signals are believed to yield relatively low spatial resolution. This is thought to be due to the disproportionate decay of high-spatial frequencies (which carry high resolution information) during volume conduction from electrical sources in the brain to electrodes on the scalp. At the same time, the concrete limits of EEG’s spatial resolution are not well understood, and a critical question remains unanswered: does increasing the density of EEG electrodes enable the extraction of high-resolution spatial information? This study provided an answer to this question through an experiment designed to test whether, in human early visual cortex, high-density EEG can capture higher-resolution spatial neural information as compared to present-day, standard low-density systems. We demonstrated that this is indeed the case. As a non-invasive neuroimaging method that is almost unrivalled in its temporal precision, high-density electroencephalography (EEG) is extremely useful in studying neural signals, their representation and temporal dynamics.

Behrmann and Tarr