Adding Biomarkers for High Frequency Oscillations and Spike-Gamma from Interictal Data to VR to assist in SOZ localization
Abstract number :
1.036
Submission category :
1. Basic Mechanisms / 1C. Electrophysiology/High frequency oscillations
Year :
2025
Submission ID :
1154
Source :
www.aesnet.org
Presentation date :
12/6/2025 12:00:00 AM
Published date :
Authors :
Presenting Author: James Evans, BS – University of Illinois Urbana Champaign
Andres Maldonado, MD – OSF Healthcare
Ansh Desai, MD – OSF Healthcare
Reid Jockisch, BS – Jump Trading Simulation and Education Center
Connor Davey, BA – Jump Trading Simulation and Education Center
Graham Huesmann, MD, PhD – Carle Foundation Hospital
Fadi Mikhail, MD – Carle Foundation Hospital
Aaron Anderson, PhD – University of Illinois Urbana Champaign
Matthew Bramlet, MD – OSF Healthcare
Bradley Sutton, PhD – University of Illinois Urbana Champaign
Rationale: Stereoelectroencephalographic (SEEG) electrodes are used to localize the seizure onset zone (SOZ) in patients with drug resistant epilepsy. Manually analyzing SEEG data to determine prospective SOZs is time consuming. Interictal SEEG biomarkers have been used for research purposes to perform interictal source localization which may improve localization time and accuracy. In this study, we examine two promising biomarkers, high frequency oscillations (HFOs) and spike-gamma (SG), and compare them to each other and to clinically resected tissue. As virtual reality (VR) models have proven useful for navigating patient anatomy, we will add these to a VR presurgical planning tool [J. L. Evans, Front Neuroinform 2024] to assist in SOZ localization.
Methods: We analyzed data from two patients who underwent evaluation and treatment for drug resistant epilepsy at OSF Saint Francis Medical Center in Peoria, Illinois. All patients underwent pre-implantation MRI, post-implantation CT, and post-resection CT. SEEG data were recorded using DIXI SEEG electrodes sampled at 1,024 Hz. All patients were seizure free (Engel I) 6 months post resection. MRI and CT images were co-registered; electrode locations were extracted using the methods described in SEEG4D [J. L. Evans 2024]. SEEG data was processed into HFOs using PyHFO, an open-source tool to compute HFOs [Y. Zhang,J Neural Eng. 2024]. SG signals were determined by first computing and evaluating signal spikes, then checking the presence of gamma band oscillations 3σ above the mean power as described in [J. Thomas,Ann Neurol. 2023]. We evaluated the correlation between contacts resected and those with high rates of SG and HFO signals.
Results: We evaluated a total of 211 contacts. We calculated HFOs with PyHFO and SG following the procedures listed. We found a correlation coefficient of r=0.275 between contacts with high rates of SG and HFOs, showing significant correlation p< 0.05. We show the contacts plotted based on their respective biomarker rates color coded based on their resection status. We also visualize a 3D VR object of the subject where the contacts size is based off their power, and their color based off HFO rates colored low-to-high (blue to red). This model closely aligns with the clinical mapping of electrodes and the resected site.
Basic Mechanisms