Authors :
Presenting Author: Pariya Salami, Ph.D. – Massachusetts General Hospital
Pierre Bourdillon, MD, PhD – Hospital Foundation Adolphe de Rothschild
Angelique Paulk, PhD – Massachusetts General Hospital
Peter Hadar, MD, MS – Massachusetts General Hospital; Harvard Medical School
Omar Alamoudi, PhD – The University of Texas Health Science Center at Houston
Mark Richardson, MD, PhD – Massachusetts General Hospital
Sandipan Pati, MD – University of Minnesota
Sydney Cash, MD, PhD – Massachusetts General Hospital
Rationale:
The thalamus is believed to play a critical role in seizure spread and generalization, yet the contributions of individual thalamic nuclei remain poorly understood. Emerging evidence suggests that distinct nuclei participate differently in epileptic networks depending on seizure type. If the thalamus is indeed not a monolithic structure when it comes to seizure dynamics, therapeutic options that target the thalamus for seizure control may benefit from a nucleus- and seizure-specific approach. In particular, knowledge of nucleus-specific connectivity patterns will allow clinicians to devise interventions that are tailored to the functional role of each nucleus in seizure initiation, propagation, or termination.Methods:
We analyzed 308 seizures recorded from 44 patients undergoing intracranial EEG monitoring, each with at least one thalamic electrode. Seizures were classified by onset region and spread pattern. For each seizure, we identified the nucleus showing early recruitment (within 5 s of onset). Functional connectivity between seizure-onset zones (SOZ) and individual thalamic nuclei was computed using a nonlinear measure of connectivity (h²; measured in AnyWave). Connectivity was assessed from 15 seconds before seizure onset to 15 seconds after seizure termination. To account for differences in seizure duration, each seizure was divided into 100 equal time intervals. Connectivity was compared across five phases: pre-seizure, onset, middle (progression), end, and post-seizure. Differences between different thalamic nuclei were evaluated during these periods. The Kruskal-Wallis test was used to assess group differences in each phase.
Results:
Connectivity between thalamic nuclei and the SOZ varied significantly across seizure types. Considering all seizures together, the pulvinar consistently exhibited the highest connectivity across alpha and beta bands during seizure onset and progression (p < 0.05), while the centromedian (CM) showed the lowest. When stratifying by seizure onset region, in seizures originating from mesial temporal regions, both the pulvinar and anterior nucleus of the thalamus (ANT) showed similarly high connectivity. In seizures with broad onset patterns the pulvinar maintained the strongest connectivity, compared to CM and ANT, whereas ANT displayed the weakest connectivity to the SOZ, suggesting a broader role for the pulvinar across seizure types. Notably, connectivity differences across thalamic nuclei diminished by seizure termination (end phase), with all nuclei exhibiting stronger connectivity during this phase, suggesting a shared role in facilitating seizure termination.
Conclusions:
Our findings demonstrate that thalamocortical connectivity is both nucleus-specific and seizure-type dependent. The pulvinar emerges as a central hub in early seizure propagation across diverse seizure types. These results support a precision-medicine approach to thalamic stimulation, tailored to seizure phenotype and thalamic involvement. Future studies should investigate whether targeting high-connectivity nuclei like the pulvinar can improve outcomes of neuromodulation therapies in epilepsy.
Funding:
W81XWH-22-1-0315