Abstracts

This is a Late-Breaking abstract.

Rationale: Approximately 3 million adults in the U.S. have active epilepsy, with a third being insufficiently controlled by medication alone. These patients require alternative treatment options. One such alternative treatment option is the continuous electrical deep-brain-stimulation (DBS) of the centromedian nucleus of the thalamus (CMN). CMN-DBS has recently emerged as a promising therapeutic intervention for patients with drug-resistant epilepsy causing tonic-clonic seizures, secondarily generalized seizures, or Lennox-Gastaut syndrome. However, parameter selection for seizure control remains empiric, poorly understood, and time-consuming, with no established consensus on the optimal stimulation location. For example, early investigations suggested that stimulating the ventrolaterally located parvocellular part of CMN improves clinical outcomes. However, a more recent study favors the anteromedial CMN and neighboring mediodorsal and parafascicular nuclei. These discrepancies in the literature highlight a clear need for neurophysiological biomarkers of therapy as changes in seizure control through postoperative programming take weeks to months to observe.

Methods: To overcome this clinical barrier, we propose to determine the patient-specific stimulation location by observing the brain’s response to single pulses of electrical stimulation. Specifically, we quantify the potential extent of network interaction accessible through the implanted DBS leads across the CMN via evoked electroencephalographic (EEG) responses intraoperatively. For this purpose, we stimulate through each pairwise combination of the implanted electrode with short electrical pulses (6 mA intensity, 200 us pulse duration, 1 Hz pulse frequency) while simultaneously recording the evoked responses from the scalp-recorded EEG. Next, we quantify the magnitude of the induced responses for each stimulated location to infer which stimulation location causes maximum network engagement, measured through the root-mean-square value of the scalp response within a 500 ms window after stimulation. Finally, we test whether this biomarker translates into improved clinical outcomes.

Results: To this end, we have tested our approach in two patients that underwent bilateral DBS of the CMN. We found that the magnitude and morphology of evoked EEG responses across lobes and hemispheres depend on the stimulation location within the CMN, highlighting the differential effect on large-scale network engagement. These results provided the basis for DBS programming to maximize the impact upon desired networks. The outcome of this initial study showed an encouraging early (3-month) nine-fold reduction in seizure occurrence for the first patient.

Conclusions: With further verification, our systematic characterization of CMN-DBS evoked EEG biomarkers will enable more optimal DBS programming and improved patient outcomes. The proposed approach will provide a quantitative framework upon which electrode location and parameter selection may be based on.

Funding: This work was supported by the NIH (R01-MH120194, R01-MH122258, R01-EB026439, P41-EB018783, U01-NS108916, U24-NS109103) and an 2022 American Epilepsy Society Seed Grant to JTW.