Authors :
Presenting Author: Natalia Sucher Munizaga, BA – UCSF
Ehsan Tadayon, MD – Neurology – Tufts Medical Center; Robert Knowlton, MD – Neurology – University of California San Francisco; Benjamin Speidel, MS – Neurology – University of California San Francisco; Jonathan Kleen, MD PhD – Neurology – University of California San Francisco
Rationale: Seizure semiology provides a real-time window into the anatomical regions dynamically engulfed by pathological activity. Many clinical semiology signs and their localizations have been described well before intracranial EEG (ICEEG) recordings, and they are actively used in clinical practice, including epilepsy surgery decision-making. However, to this day very few studies have used ICEEG in a quantitative manner to validate the regional substrates (functional anatomy) underlying semiology features. We used a combination of automated signal processing, signal-to-noise principles, and data visualization approaches to create a functional atlas of the brain regions activated when specific motor semiology features appear clinically in real-time.
Methods: We evaluated ICEEG recordings of patients (n=7) with medically-refractory seizures who underwent intracranial implantation of grid, strip and depth electrodes (ranging from 96-184 electrodes per patient). We selected patients who had similar clinical semiology features (behavioral change, head turn, progression to bilateral tonic-clonic behavior), and electrographic seizure spread patterns (mesial temporal onset followed by spread to lateral temporal lobe and then parietal and frontal cortices). Blinded to the ICEEG data, we reviewed video recordings and comprehensively annotated head, face, and limb behaviors throughout the peri-ictal period at 5 Hz time resolution. We used a line length transform of the ICEEG data as a surrogate for seizure activity intensity at each electrode, mapping the change in activity to personalized vertex and anatomical parcellation of cortical surfaces.
Results: Compared to other regions, the precentral gyrus showed the most consistent significantly positive increases (p < 0.05, one-sided t-tests) in seizure activity during upper proximal limb and mouth tonic activity (100% of patients), head version (86%), eye deviation (83%), and upper distal limb tonic activity (71%). We binned vertex-level activity into subregional coordinates across patients and found that contralateral head version and eye deviation tended to involve superior precentral and caudal middle frontal gyri regions, whereas tonic face, arm, and leg activity tended to map the length of the precentral gyrus compared to other regions. Peripheral coverage regions showed coinciding seizure activity suggestive of diffuse seizure spread.