Abstracts

Rationale: Normal brain functioning is emerging from a complex interplay among brain regions within a large-scale neuronal network. Focal epilepsy alters this network causing seizures. Studying the alterations of this network is crucial for identifying epileptogenic nodes, which may help delineate the epileptogenic zone (EZ) in patients undergoing surgery. Here, we aim to characterize alterations in functional connectivity (FC) of epileptogenic brain networks based on interictal and ictal intracranial EEG (iEEG) recordings. We hypothesize that FC increases in epileptogenic nodes of the network compared to non-epileptogenic nodes.

Methods: We analyzed iEEG recordings from 31 children (11.41 years old ± 5.94) with drug resistant epilepsy (DRE) who underwent resective neurosurgery. Patients were dichotomized into having good (Engel 1; 22 patients) and poor (Engel ≥2; 9 patients) surgical outcomes. On the iEEG data, we identified one-minute of data containing: (i) interictal activity without frequent spikes (“No Spikes”); (ii) interictal activity with frequent spikes (“Spikes”); (iii) pre-ictal activity before the onset of a clinical seizure (“Pre-Ictal”); (iv) ictal activity during a clinical seizure (“Ictal”); and (v) post-ictal activity after the end of a clinical seizure (“Post-Ictal”) (Fig. 1a). For each segment, we computed three FC measures [i.e., Amplitude Envelope Correlation (AEC), Correlation (CORR), and Phase Locked value (PLV)] in the low (30-60 Hz) and high gamma (60-90 Hz) bands (Fig. 1a), which characterize synchronization of neural populations before and during seizures and are mostly found in the epileptogenic tissue. Using the Wilcoxon signed-rank test, we tested the difference in mean FC between nodes inside vs. outside resection and the clinically defined seizure onset zone (SOZ) (Fig. 1b).

Results: We observed higher mean FC values (p < 0.05) for the nodes inside compared to outside resection in good outcome patients for the following metrices: (i) AEC (low and high gamma) for all conditions (Fig. 2a); (ii) PLV (low and high gamma) for “Post-Ictal”; and (iii) CORR (low gamma) for “Spikes” and “Ictal.” Similar findings yield the analysis for the clinically defined SOZ (Fig. 2b). For poor outcome patients, FC values between nodes inside vs. outside resection (Fig. 2c) and SOZ (Fig. 2d) did not reveal significant findings. In both good and poor outcome patients, AEC values of “No Spikes” were lower compared to all other conditions in both frequency bands (p < 0.05). Similarly, AEC values were lower in “Pre-Ictal” compared to “Spikes,” “Ictal,” and “Post-Ictal” (p < 0.05).