Interictal epileptiform discharges (IEDs) are critical markers in epilepsy, playing a key role in diagnosis and treatment planning; however, their detection remains labor-intensive, subjective, and lacking standardization—particularly in the context of intracranial EEG (iEEG) during sleep. Most existing automated tools were developed for scalp EEG, leaving a significant gap in validated, scalable solutions for deep brain (iEEG) annotation, which limits reproducibility and widespread adoption in both clinical and research settings. This shortfall is clinically important, as accurate and efficient IED detection is directly linked to improved epilepsy management and may help prevent cognitive decline associated with undetected discharges.

The Technology
A gradient boosting (LGBM) algorithm trained on expert-annotated IEDs to enable fully automated, highly accurate IED detection. Feature selection is optimized for interpretability and computational efficiency, with key features inclduing peak-to-peak amplitude, entropy, and Teager-Kaiser energy.
Algorithms was trained using the first open-access, expert-annotated dataset of multichannel iEEG sleep recordings with sub-second resolution, obtained from 25 drug-resistant epilepsy patients at UCLA and Tel Aviv Sourasky Medical Center.
The algorithm achieves 94.3% precision and 94.3% sensitivity in cross-validation, outperforming both a leading commercial software and traditional thresholding techniques. Generalizes well across different patients, clinical sites, and annotators.

Potential Applications
• Clinical Epilepsy Monitoring: Objective, non-invasive rapid annotation of IEDs in presurgical or chronic epilepsy patient management using iEEG.
• Research: High-quality resource for developing, benchmarking, and validating new automated IED detection algorithms and understanding epileptiform brain activity during sleep
• Device Development: Provides a robust annotated dataset for training and validating next-generation closed-loop neuromodulation devices and decision support systems in epilepsy care
• Cognitive Neuroscience: Enables large-scale study of the relationship between sleep physiology, interictal activity, and neurocognitive outcomes in epilepsy

Value Proposition
This technology delivers standardized, objective, and highly generalizable automated IED annotation, reducing subjectivity and ensuring consistent results across centers, patient populations, and recording setups. By accelerating clinical and research workflows and providing a freely available, validated resource, it enables faster, more accurate epilepsy diagnosis, surgical planning, and collaborative innovation in neurotechnology.

Reference:
1. Patent filed
2. Falach et al 2024, Scientific Data

Dataset generation process. (A) Epilepsy patients implanted with depth electrodes were connected to an iEEG acquisition system for clinical monitoring during sleep. (B) The recorded signals underwent offline examination by a neurologist using clinical software. Bottom: annotation examples from two patients and annotators. PHG = parahippocampal gyrus, Bi-A = bi-synchronous amygdala, M-REC = mesial right entorhinal cortex. (C) The final dataset contains raw data and IEDs annotations exported into BIDS format