Abstracts

Rationale: Responsive neuromodulation targeting thalamic nuclei has emerged as a promising treatment option for patients with drug resistant epilepsy who are not candidates for traditional surgical approaches. The principal of responsive neuromodulation relies on accurate and early detection of seizures and subsequent electrical disruption of the ictal network. However, deciding which thalamic nuclei to target in an individual patient remains an area of active exploration. Although thalamocortical anatomical networks are well-characterized, whether ictal propagation patterns reliably follow these structural substrates is unknown. To address this question, we characterize thalamocortical ictal networks by cortical lobe in 13 patients with focal seizures captured on simultaneous thalamic and cortical intracranial recordings.

Methods: All patients who underwent stereotactic intracranial EEG (sEEG) recordings between July 2020 and May 1, 2022, with sampling of the thalamus as part of their clinical care at our institution (n=13) were included. For each patient, consensus of the seizure onset zone was obtained during multidisciplinary conference discussions. For each patient found to have focal or multifocal seizures, the lobe(s) of onset for each stereotyped seizure type was categorized. For each seizure type, the propagation delay from the seizure onset zone to the ipsilateral thalamic electrode contacts was quantified by visual analysis. Seizures that involved multiple ipsilateral lobes and thalamic nuclei at onset were excluded from the present analysis. Seizures involving more than 1 cortical lobe at onset without propagation to the thalamic nuclei sampled were included. For each seizure type, thalamocortical propagation patterns were categorized as follows: instantaneous (thalamus recruited within 1s of cortical onset), delayed (thalamus recruited after 1s of cortical onset), and no propagation (thalamus never recruited through the duration of the seizure). The proportion of seizure types that propagated to each thalamic nucleus were calculated for each cortical lobe.

Results: A total of 13 subjects with thalamic sampling during intracranial EEG recording were reviewed. 20 focal electrographic seizure types were captured from 7 subjects (4F, ages 9-53). In our cohort, frontal lobe seizures had inconsistent propagation to the centromedian nucleus region of the thalamus (CM), parietal lobe seizures had consistent instantaneous recruitment of the CM, temporal lobe seizures had consistent instantaneous recruitment of the anterior nucleus region of the thalamus (AN) and consistently did not recruit the CM, and occipital lobe seizures consistently did not recruit the CM (Table 1, Figure 1).

Conclusions: With the growing indications for thalamic neuromodulation as a treatment option, identification of the appropriate thalamic nuclei to target to engage the ictal network is increasingly critical. These data suggest that some common patterns may exist in thalamocortical ictal networks across patients to empirically guide optimal thalamic targets for neuromodulation.

Funding: NIH NINDS R01NS115868