Authors :
Presenting Author: Nina Moergeli, MSc – ETH Zurich
Giovanna Aiello, BSc, MSc. – ETH Zurich; Lennart Stieglitz, Prof. Dr. med. – University Hospital Zurich; Christian Baumann, Prof. Dr. med. – University Hospital Zurich; Lukas Imbach, PD Dr. med. – Swiss Epilepsy Clinic (Klinik Lengg); Rafael Polania, Prof. Dr. – ETH Zurich
Rationale:
The Anterior Nucleus of the Thalamus (ANT) has been suggested as a target for Deep Brain Stimulation (DBS) in drug-resistant epilepsy (DRE) due to its connectivity profile and role in the subcortical propagation of epileptic seizures. However, its electrophysiological properties and reasons for variable treatment responses are unclear. Recent studies highlight the importance of active contact choice for treatment optimization, with more dorsal contacts appearing more beneficial. Moreover, treatment response based on seizure frequency tracking can be incomplete and qualitative.
Methods:
By recording out-patient (194 ± 222.8 days) bipolar ANT Local Field Potentials and contacts’ impedance via DBS leads in an awake state, we have characterized the properties of the ANT in a cohort of 16 DRE patients (8 responders, 8 non-responders). The ANT-LFPs were analyzed in terms of spectral properties and entropic content (sample entropy) in each bipolar configuration (0-1, 0-2 and 0-3), with 0 being the most ventral and 3 being the most dorsal contact. We then observed whether differences were arising when considering separately responders and non-responders.
Results: Contact 2 showed the lowest impedance values (820.5 ± 240 Ω) throughout the recording period (p[0,2] < 0.001, p[1,2] < 0.01, p[2,3]: 0.987), with decreasing trends observed for 2 and 3 (R(2) = -0.37, R(3) = -0.40, both p < 0.001). Choosing the contact with the lowest impedance has been associated with symptom improvement and better response, particularly with upper contacts close to the ventricle, and these findings also indicate the absence of significant foreign body reaction, although we argue that active contacts delivering stimulation could have a lower impedance. Montage 0-2 exhibited the highest entropy (p[0-1 vs 0-2], p[0-2 vs 0-3] < 0.001), with a significant increase over time (Pearson's R(0-2) = 0.316, p < 0.01). Non-responders had higher entropy compared to responders (z(nr-r)=2.044, p< 0.05). In the spectral domain, spectral power in 0-2 was consistently higher than in 0-1 and 0-3 beyond θ frequencies (α: p < 0.01, β
1 < 0.05, β
2,γ: p < 0.001). β
1 oscillations (12-20 Hz) decreased over time for contacts 0-1 and 0-3 (R(0-1) = -0.21, p< 0.05, R(0-3) = -0.198, p = 0.058). Interestingly, non-responders exhibited significantly higher spectral power than responders across all frequency bands except δ (z(r-nr) = 5.27, p< 0.01), with the largest difference in θ and γ bands (z(r-nr)=-5.623, z(r-nr)=-6.639 respectively, p< 0.001).