Rationale:
Seizure-related loss of consciousness (LOC) in pediatric epilepsy significantly impacts safety, cognitive development, and quality of life. The thalamus is believed to play a central role in sustaining consciousness via its integrative cortical connectivity. Disruption of thalamocortical networks may underlie impaired awareness during seizures, yet mechanistic insights remain limited. Stereo-electroencephalography (sEEG) offers high-resolution intracranial recordings that enable in vivo exploration of these neural dynamics. Here, we present a protocol designed to systematically evaluate thalamocortical connectivity during seizures with and without LOC in pediatric patients with drug-resistant epilepsy. This protocol serves as a foundation for future hypothesis-driven investigations and network-targeted therapeutic strategies.
Methods:
Postoperative CT scans are co-registered with preoperative MRI to localize electrode contacts. Preprocessed sEEG data are exported from the Natus system into MATLAB and Brainstorm for quantitative analysis. Recordings include at least 20 minutes of pre-ictal, ictal, and post-ictal data. Seizures with and without LOC are identified and time-marked using synchronized video-EEG. LOC severity is scored independently by two experienced reviewers. Functional connectivity is assessed via Coherence and Directed Phase Lag Index (dPLI), while effective connectivity is measured using Directed Transfer Function (DTF), a frequency-domain Granger causality method. Graph theory metrics—such as integration, segregation, and global efficiency—are applied to both FC and EC matrices to characterize network topology. Connectivity maps are generated to visualize thalamocortical network dynamics.
Results:
Seven pediatric patients (mean age 14.8 ± 3.4 years; 4 female) underwent thalamic sEEG implantation. Etiologies included focal cortical dysplasia (n=1), periventricular nodular heterotopia (n=1), traumatic brain injury (n=1), prematurity with intraventricular hemorrhage (n=1), neonatal HIE (n=1), and non-lesional epilepsy (n=2). Pre-implant hypotheses included temporal (n=4), frontal (n=1), insular (n=1), and multifocal (n=1) seizure onset. Thalamic sampling included anterior nuclei (n=5), pulvinar (n=4), and centromedian nucleus (n=1). Interictal thalamic abnormalities were present in 4 patients. Thalamic recruitment during seizures occurred within 0.5–13 seconds of cortical onset in 6 of 7 patients.
Conclusions:
This protocol-based approach enables rigorous investigation of thalamocortical connectivity and its role in LOC during epileptic seizures. Ongoing analyses will evaluate frequency-specific connectivity and graph-based network features associated with altered awareness. Findings from this study may enhance patient-specific neuromodulation strategies and advance our understanding of consciousness disruption in epilepsy. Full results will be presented at the AES Annual Meeting.
Funding: The authors received no financial support for this research study