Introduction: Anterior nucleus of the thalamus deep brain stimulation (ANT-DBS) effectively suppresses seizures in well-chosen medically refractory epilepsy patients, but very few patients achieve seizure freedom. To improve the therapy, effective neural biomarkers are necessary to inform clinicians when tuning stimulation parameters. The latest implantable pulse generators allow for recording local field potentials (LFP), which may inform clinicians how neural activity at the implant location responds to therapy.
Methods: LFPs were recorded using Medtronic Percept PC in stim on and off conditions under multiple stimulation parameters as part of an Institutional Review Board (IRB) approved clinical study (#NCT05493722) at the University of Minnesota. Raw time domain data was pre-processed or rejected for artifacts and noise. LFP power spectral densities were estimated using slepian multitapers and baseline 1/F was detrended using an adjusted FOOOF method. Participants were categorized as responders if the average of the three latest seizure frequency reports demonstrated > 50% reduction compared to pre-DBS seizure frequency. Statistical comparisons were made with non-parametric Wilcoxon Rank Sum tests.
Results: In participants where low gamma oscillations (30-45Hz) were observed, multiple stimulation parameters statistically significantly reduced gamma activity (p < 0.05). Responders exhibited statistically significant low gamma oscillation suppression with stimulation as compared to non-responders who did not exhibit gamma oscillations (p < 0.05). Baseline low gamma oscillations changed in power across visits and decreased in responders chronically over time.
Conclusion : Low gamma oscillations were observed in multiple ANT-DBS implanted epilepsy participants and stimulation could reduce these oscillations both acutely within the same visit and chronically over time in responders. Non-responders showed heterogeneous oscillatory activity with minimal modulation of other frequency bands. ANT low gamma band activity may be a useful biomarker for identifying responders and/or optimizing stimulation parameters in medically refractory epilepsy patients.