Zhou, O. | bioRxiv Science

bioRxiv2025

Building an Ecosystem of Seizure Localization Methods: Neural Fragility as the First Step

The current treatment for drug-resistant epilepsy (DRE) is surgical intervention, which relies on accurate identification of the seizure onset zone (SOZ) using intracranial EEG (iEEG) data. iEEG analysis with computational epileptogenic zone identification algorithms (CEZIAs) is a promising step towards better SOZ localization and surgical outcomes. A key step in validation and adoption of CEZIAs is to allow for widespread shared development and validation of code and data. We describe a set of three R packages to achieve this goal. Our ecosystem of seizure localization methods involves a straightforward analysis pipeline, standardized data formatting and storage, and completely documented and open-source code. The TableContainer package allows for easy storage and manipulation of table data, serving as groundwork for the Epoch package, which is specifically geared towards iEEG data. The Epoch package allows for cropping, resampling, and visualization of iEEG data and provides publicly downloadable iEEG data for reproducibility. Finally, the EZFragility package uses these two foundational packages to analyze iEEGs for SOZ localization using the Neural Fragility method described by Li et al. EZFragility was built using the same core principles as the original method but included several enhancements in computational efficiency and user experience. It accurately reproduces neural fragility results for both sample patients used in the original paper. This project serves as the first step towards building an open-source, reproducible ecosystem of seizure localization methods in R. Future steps include the addition of other CEZIAs using the framework and sample data already made available by these packages. Significance StatementLocalization of the seizure onset zone (SOZ) is a critical step in surgical treatment of drug-resistant epilepsy. Computational epileptogenic zone identification algorithms (CEZIAs) are promising potential tools to aid in clinical decision-making. However, shared development and verification of CEZIAs is difficult due to obscured source code, variable data structures, and limited data availability. The EZFragility software package1 is the first step in building a collaborative ecosystem of CEZIAs that can be downloaded, tested, and used without these roadblocks. EZFragility1 and its dependent packages TableContainer2 and Epoch3 are freely available on the Comprehensive R Archive Network (CRAN), with source code viewable on GitHub. They provide an open-source framework for CEZIA code, data formatting, and data access, all with extensive documentation.

neuroscience↗

bioRxiv2024

Integrating Data Across Oscillatory Power Bands Predicts the Epileptogenic Zone: the Frequency Range Explorer Epileptogenic Zone (FREEZ) Identification Algorithm

Accurate identification of the seizure onset zone (SOZ) using intracranial electroencephalography (iEEG) remains challenging. Although diverse methods have leveraged spectral features to classify patient outcomes, few approaches focus on identifying individual electrodes within the SOZ or integrate a broad spectrum of frequency ranges within a single model. We developed an interpretable machine learning model that integrates power across delta, theta, alpha, beta, gamma, and high-gamma frequencies over time to identify the SOZ. For 1,511 electrodes implanted across 21 patients, we computed the mean spectral power in each frequency band for the first 20 seconds after seizure onset and analyzed the differences in power between SOZ and non-SOZ electrodes. In patients who were seizure-free after surgery (n = 14), electrodes within the SOZ showed significantly higher area under the curve (AUC) for mean power over time in the first 20 seconds after seizure onset compared to electrodes outside the SOZ in the alpha (p = 0.0272), beta (p = 0.0263), gamma (p = 0.0013), and high gamma (p = 0.0086) ranges. Additionally, electrodes within the SOZ in patients that became seizure-free after surgery had significantly higher AUC compared to equivalent electrodes in patients who did not become seizure-free after surgery (n = 7) in the gamma (p = 0.0145) and high gamma (p = 0.0024) power ranges. We trained a stacked random forest ensemble model using these features over time to label electrodes within the SOZ. Leave-one-out patient cross validation of the machine learning model yielded a 96.6% positive predictive value and 99.9% specificity for identifying electrodes within the SOZ. Our dataset included a diverse array of seizure onset patterns, which were all classified accurately by the model. A second model was trained to predict post-operative seizure freedom, yielding 95.2% accuracy for predicting seizure outcome based on a planned resection. This two-model design mirrors clinical workflow, first localizing SOZ electrodes to support surgical planning, then predicting outcome based on a surgical plan. An advantage of our interpretable machine learning approach is the ability to interrogate our models to understand how predictions are made. For electrode classification, the model weighed beta (0.66 {+/-} 0.07), high gamma (0.54 {+/-} 0.06), and delta (0.51 {+/-} 0.06) power bands most heavily. Viewing the models frequency band weights over time reveals that the model identified a pattern resembling the "fingerprint of the epileptogenic zone", reinforcing the importance of this dominant fundamental neurophysiologic signature of seizure onset.

neuroscience↗