Abstracts

Rationale: About 30% of children with drug-resistant epilepsy (DRE) continue to have seizures post-surgery. Defining a re-operation plan for a second surgery is particularly challenging in these patients; thus, any novel noninvasive technique which can add to the test battery is highly valued. Since epilepsy is increasingly conceptualized as a network disorder, understanding how brain regions are connected may be crucial for planning treatment. Our purpose is to estimate brain source connectivity using scalp EEG in children that had repeated epilepsy surgery and assess how brain connectivity changes over time.

Methods: We studied nine children with DRE who had repeated surgery (RS-group) and 12 patients who had one surgery followed by seizure-freedom (SF-group). We analyzed scalp-EEG epochs, without spikes, from three time-points (Figure1a-b): before first surgery, in-between two surgeries, and after reoperation. We estimated Functional Connectivity between cortical regions in different frequency bands (alpha, theta, delta, beta, gamma): graph-theory was applied to estimate centrality of each region within the network (Figure1c-f). We compared pre- and post-surgery connectivity of all brain regions (far or adjacent to resection), for RS-group and SF-group (Wilcoxon-signed-rank test). Figure 1 presents an overview of the data analysis pipeline and methodology.

Results: In the SF group, global-centrality increased after surgery in all frequency bands, both in proximity and far from resection (p< 0.001). On the other hand, when looking at the RS group, after failed surgery, we found that theta, alpha and beta networks did not present changes in centrality close to the resection (adjacent regions) or presented a decrease far from resection (see Figure 2). In these same frequency bands, when the 2nd (repeated) surgery was successful, we observed again an increase in the global-centrality (as in the SF group) far from the resection (theta: p< 0.001; alpha: p=0.01; beta: p< 0.001) as well as in its proximity (theta: p=0.02). Figure 2 illustrates the results in the three frequency bands that showed different behaviors in the two groups of patients.

Conclusions: We present first evidence that EEG-based connectivity analysis at source level can reveal epilepsy networks changes following one or two epilepsy surgery in children with DRE. Post-surgical seizure recurrence is associated with a decrease or no changes in the centrality of the untouched brain regions within the network. Conversely, successful surgeries are associated with an increase in the overall centrality of the untouched regions. We can speculate that unaltered or diminished centrality, following epilepsy surgery, is a consequence of an overall compensation effect, where the untouched cortical regions aim to isolate the persistent epileptogenic focus. Differently, an overall connectivity increase after first surgery possibly indicates re-organization of the brain network, where the non-epileptogenic regions gain centrality, since the epileptogenic focus has been successfully removed.

Funding: Please list any funding that was received in support of this abstract.: None.