Abstracts

Rationale: Cortico-cortical evoked potentials (CCEPs) are an electrophysiological measure of connectivity performed in human patients undergoing intracranial EEG. CCEPs have been shown to reflect both ictal propagation and various functional networks; however, here we seek to understand how the network defined by CCEPs might relate to inter-ictal epileptiform activity.

Methods: We retrospectively analyzed data in 5 human subjects who underwent surgical stereo EEG implantation for long-term seizure monitoring. The contact which was key source of interictal spiking activity was clinically identified. CCEPs were generated by stimulating at this contact with parameters: alternating monophasic square wave, pulse width 100 mcs, amp. 2-4 mA, inter-pulse interval 0.5 Hz. Data from all other intracranial contacts was sampled at 1 kHz, band passed 2-300 Hz; stimulation-locked epochs were excerpted and averaged to generate an evoked potential. During spontaneous activity inter-ictall spikes were automatically identified on the clinically implicated contact using a continuous wavelet transformation (cwt). Data were epoched time-locked to each spike, and an event related potential (ERP) and event related spectral activity was generated at all intracranial contact sites. All data were analyzed using custom written code in Matlab. We analyzed the Pearson correlation coefficient for all variables.

Results: Stimulation at the contact clinically determined to be most relevant to inter-ictal spike activity generated robust evoked potentials in all subjects. The waveform and spatial distribution of the CCEPs and spike time-locked activity was varied across subjects. However, the anatomical distribution of the root mean square (RMS) of the power of the CCEP and spike-ERP was highly correlated (r = .86). This correlation held true even after both the CCEP and ERP were normalized to baseline activity to control for global differences in power at different cortical sites (r = 0.62). For the absolute and normalized data, CCEPs were more correlated with spike ERPs than with contact distance from stimulating channel, suggesting that the correlation reflects a neuro-anatomical network with specific connectivity. There was a weaker but significant correlation of CCEP power with spike-related oscillatory activity, separate from the event related potential. This was most prominent in the alpha and beta range (r = 0.56-.57), but less so in the theta, gamma, and HFO ranges.

Conclusions: The data support the use of CCEPs in elucidating the network underlying epileptiform activity. A larger cohort is needed to validate these findings and identify differences, if any, among epilepsy sustained by different anatomico-functional networks.

Funding: Internal institutional funding