Abstracts

Rationale: Stereo-EEG with multiple intracerebral depth electrodes is being increasingly used for pre-surgical epilepsy monitoring. Standard methods of evaluation with these recordings rely on visually identifying EEG changes in specific contacts to indicate the seizure onset zone. Non-linear information transfer measures from such recordings can potentially provide additional information that may not be apparent with conventional linear measures or on visual evaluation of EEG (Yang, et. al., 2013; Mierlo, et. al., 2014; Yang, et. al., 2012; Zhu, et. al., 2015). Normalized transfer entropy (NTE), a specific non-linear functional measure has been shown to be a robust indicator of aberrant information transfer in parkinsonism that can be corrected with therapy with resultant behavioral improvements (Gilmour, et. al., 2011). Therefore, we tested the applicability of NTE as a measure of peri-ictal and ictal information transfer in epilepsy. Methods: We evaluated NTE between depth electrode contacts located within deep brain structures and the superficial contacts that were located in the cortex in 2 patients with pharmacoresistant epilepsy (one extratemporal, one temporal) undergoing stereo-EEG. Comprehensive assessments of conventional linear measures of information transfer and visual EEG evaluation were also performed and correlated. Results: Patient 1 had inadequate seizure control (Engel Class 2) following right anterior temporal lobectomy, performed after noninvasive video-EEG monitoring and other presurgical testing that had suggested a mesial temporal focus. Stereo-EEG with 15 depth electrodes in the right frontotemporal region indicated ictal EEG onset in the right orbitofrontal region. NTE function was very high, with specific clear peaks between the contacts in deep brain areas and their corresponding contacts in the superficial cortical regions, in 13 out of 15 electrodes including in the posterior and anterior orbitofrontal depth electrodes, suggesting multifocal peri-ictal and ictal information transfer. Patient 2 had nonlesional temporal lobe epilepsy with bitemporal independent interictal spikes. Stereo-EEG with 8 depth electrodes (bilateral temporal and orbitofrontal) electrodes showed ictal EEG onset in the right amygdala and hippocampus in 24 recorded seizures. However, there was no NTE between the amygdala or hippocampus and the neocortical electrode contacts on the right side. In contrast, there was clear NTE in 2 depth electrodes between the left amygdala and deep contacts in the posterior orbitofrontal cortex and the corresponding superficial contacts in the temporal and frontal cortices suggesting multifocal information transfer. Conclusions: These patients illustrate the occurrence of higher order information transfer in the nonlinear domain in the peri-ictal and ictal periods. These preliminary findings from 2 patients suggest that NTE may be an important measure of information transfer from deep brain structures to the neo-cortex in temporal and extratemporal epilepsy. Further studies are necessary to determine the clinical importance of including nonlinear informational transfer measures such as NTE in making surgical decisions, planning the extent of surgical resection and reducing surgical failures.