Abstracts

Rationale: Invasive electroencephalography (iEEG), such as electrocorticography (ECoG) and stereoelectroencephalography (SEEG), often involves acquisition over hundreds of channels. Ictal activity observed during continuous monitoring or cortico-coritcal evoked potentials (CCEPs) to single-pulse electrical stimulation can be difficult to quickly interpret over such high-density arrays. Here we describe innovations in the ordering, registration, and coloring of such data, so as to provide an informative summary for subsequent detailed review. Methods: We demonstrate the method for organizing and visualizing CCEPs, averaged over 30 trials using Brainstorm (Tadel et al. 2011). A traditional stacked graph of the average observed response to left superior temporal gyrus stimulation is seen in Fig. 1 (see inset for a schematic of contact locations). Separately, we import the MRI, then segment and label the 3D regions, here using Brainsuite (Shattuck et al. 2002). Each contact is registered to the MRI and assigned one of the six basic hemispheric regions of interest, pre-frontal, frontal, central, parietal, temporal, or occipital (Fig. 2, bottom right). Over a user selected time region of interest (-100ms to 400ms here) the average energy (root mean square, RMS) of each waveform is calculated. The waveforms are separated by hemisphere and ordered in decreasing RMS value, then plotted using the MATLAB (The Mathworks Inc.) “area” function, with left side waveforms appearing above the x-axis, and right side waveforms below it. Finally, the region colors assigned to each contact color the area plot. Results: Fig. 2 shows an example stimulation protocol. Each subplot gives the results from stimulation of a contact pair (top row in the left hemisphere, bottom row in the right, ordered by anatomical location). The height of the area under the curve indicates the summed response across all contacts at each point in time, with the strongest individual RMS responses nearest the x-axis. The color of the areas directly correspond to the colors shown in the inset; because all area plots are drawn to the same scale we immediately observe that, in this instance, the largest response can be seen in the upper right subplot, from stimulation of the left superior temporal gyrus. The prevalent cyan color reveals that this response was dominated by activity in the temporal region (note that this subplot and Fig. 1 reference the same data). Conclusions: This novel plotting approach allows us to combine spatial and temporal information in a single summary plot, over many stimulation points and with high-density arrays. We exploit the advanced and automated 3D segmentation analysis of software such as Brainsuite to identify, order, and color the contacts. The waveforms are ordered within the subplots by user preference, then the area plot is colored by these regions. The resulting single figure provides a powerful summary that draws attention to relevant data that needs further detailed analyses. Funding: Supported in part by the NIH under awards R01NS089212 and R01EB009048.