Explainable AI-Enhanced Ensemble Protocol Using Gradient-Boosted Models for Zero-False-Alarm Seizure Detection from EEG
Sensors · 28.01.2026 · 26 (3) · 863
DOI: 10.3390/s26030863 ↗Kurz gefasst
Studienprotokoll in der Zeitschrift Sensors, online erschienen am 28.01.2026. Verfasst von Abdul Rehman und 1 weiteren Autor:innen. Beteiligte Einrichtung: Research Institute of Engineering & Technology, Jeonju University, Jeonju 55069, Republic of Korea. Hauptthemen laut MEDLINE: Electroencephalography, Seizures, Artificial Intelligence. Die Arbeit wurde bisher 11-mal zitiert (Crossref) und nennt 40 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Korean government, Ministry of Education.
Abstract (Originalsprache)
Epilepsy affects over 50 million people worldwide, yet automated seizure detection systems either achieve moderate sensitivity with excessive false alarms or rely on uninterpretable deep networks. This study presents a patient-independent EEG-based seizure detection framework that achieved zero false alarms in 24 h with 95% sensitivity in a retrospective evaluation on a CHB–MIT pediatric cohort (n = 6 seizure-positive patients). The pipeline extracts 27 time-, frequency-, and nonlinear-domain features from 5 s windows and trains five ensemble classifiers (XGBoost, CatBoost, LightGBM, Extra Trees, Random Forest) using strict leave-one-subject-out cross-validation. All models achieved segment-level AUC ≥ 0.99. Under zero-false-alarm constraints, XGBoost attained perfect specificity with 0.922 sensitivity. SHAP and LIME analyses suggested candidate EEG biomarkers that appear consistent with known ictal signatures, including temporo-parietal theta-band power, amplitude variability (IQR, RMS), and Hjorth activity. External validation on the Siena Scalp EEG Database (12 adult patients, 37 seizures) demonstrated cross-dataset generalization with 95% event-level sensitivity (Extra Trees) and AUC of 0.86 (Random Forest). Temporal lobe channels dominated feature importance in both datasets, confirming consistent biomarker identification across pediatric and adult populations. These findings demonstrate that calibrated gradient-boosted ensembles using interpretable EEG features achieve clinically safe seizure detection with cross-dataset generalizability.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Research Institute of Engineering & Technology, Jeonju University, Jeonju 55069, Republic of Korea.
- Department of Industrial Engineering, Jeonju University, Jeonju 55069, Republic of Korea.
- Department of Data Science, Jeonju University, Jeonju 55069, Republic of Korea.
ORCID-Profile: Abdul Rehman, Sungchul Mun
Förderung
- Korean government
- Ministry of Education
Zitieren
APA
Rehman, A., & Mun, S. (2026). Explainable AI-Enhanced Ensemble Protocol Using Gradient-Boosted Models for Zero-False-Alarm Seizure Detection from EEG. Sensors, 26(3), 863. https://doi.org/10.3390/s26030863
Vancouver
Rehman A, Mun S. Explainable AI-Enhanced Ensemble Protocol Using Gradient-Boosted Models for Zero-False-Alarm Seizure Detection from EEG. Sensors. 2026;26(3):863. doi:10.3390/s26030863
BibTeX
@article{Rehman2026,
author = {Rehman, Abdul and Mun, Sungchul},
title = {Explainable AI-Enhanced Ensemble Protocol Using Gradient-Boosted Models for Zero-False-Alarm Seizure Detection from EEG},
journal = {Sensors},
year = {2026},
volume = {26},
number = {3},
pages = {863},
doi = {10.3390/s26030863}
}
EUVIMED