Expert-Level Detection of Epilepsy Markers in EEG on Short and Long Timescales
NEJM AI · 26.06.2025 · 2 (7)
DOI: 10.1056/aioa2401221 ↗Kurz gefasst
Fachartikel in der Zeitschrift NEJM AI, erschienen am 26.06.2025. Verfasst von J Li und 33 weiteren Autor:innen. Beteiligte Einrichtung: National Institute of Health Data Science, Peking University, Beijing. Die Arbeit wurde bisher 13-mal zitiert (Crossref) und nennt 36 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: HSRD VA, NINDS NIH HHS, NIA NIH HHS.
Abstract (Originalsprache)
BACKGROUND: Epileptiform discharges, or spikes, within electroencephalogram (EEG) recordings are essential for diagnosing epilepsy and localizing seizure origins. Artificial intelligence (AI) offers a promising approach to automating detection, but current models are often hindered by artifact-related false positives and often target either event- or EEG-level classification, thus limiting clinical utility. METHODS: We developed SpikeNet2, a deep-learning model based on a residual network architecture, and enhanced it with hard-negative mining to reduce false positives. Our study analyzed 17,812 EEG recordings from 13,523 patients across multiple institutions, including Massachusetts General Brigham (MGB) hospitals. Data from the Human Epilepsy Project (HEP) and SCORE-AI (SAI) were also included. A total of 32,433 event-level samples, labeled by experts, were used for training and evaluation. Performance was assessed using the area under the receiver operating characteristic curve (AUROC), the area under the precision-recall curve (AUPRC), calibration error, and a modified area under the curve (mAUC) metric. The model's generalizability was evaluated using external datasets. RESULTS: SpikeNet2 demonstrated strong performance in event-level spike detection, achieving an AUROC of 0.973 and an AUPRC of 0.995, with 44% of experts surpassing the model on the MGB dataset. In external validation, the model achieved an AUROC of 0.942 and an AUPRC of 0.948 on the HEP dataset. For EEG-level classification, SpikeNet2 recorded an AUROC of 0.958 and an AUPRC of 0.959 on the MGB dataset, an AUROC of 0.888 and an AUPRC of 0.823 on the HEP dataset, and an AUROC of 0.995 and an AUPRC of 0.991 on the SAI dataset, with 32% of experts outperforming the model. The false-positive rate was reduced to an average of nine spikes per hour. CONCLUSIONS: SpikeNet2 offers expert-level accuracy in both event-level spike detection and EEG-level classification, while significantly reducing false positives. Its dual functionality and robust performance across diverse datasets make it a promising tool for clinical and telemedicine applications, particularly in resource-limited settings. (Funded by the National Institutes of Health and others.).
Autor:innen und Institutionen
- National Institute of Health Data Science, Peking University, Beijing.
- Harvard Medical School, Boston.
- Neurology Department, Beth Israel Deaconess Medical Center, Boston.
- Institute for Theoretical Physics, Technical University Berlin, Berlin.
- Neurology Department, Washington University in St. Louis, St Louis, MO.
- Ralph H. Johnson VA Medical Center, Charleston, SC.
ORCID-Profile: J Li, D M Goldenholz, M Alkofer, C Sun, F A Nascimento, J J Halford, B C Dean, M Galanti, A F Struck, A S Greenblatt, A D Lam, A Herlopian, C Nwankwo, D Weber, D Maus, H A Haider, I Karakis, J Y Yoo, M C Ng, O Selioutski, O Taraschenko, G Osman, R Katyal, S E Schmitt, S Benbadis, S S Cash, W O Tatum, Z Sheikh, W Y Kong, G Bayas, N Turley, S Hong, M B Westover, J Jing
Förderung
- HSRD VA
- NINDS NIH HHS
- NIA NIH HHS
- NIGMS NIH HHS
- NHLBI NIH HHS
- NCATS NIH HHS
Zitieren
APA
Li, J., Goldenholz, D. M., Alkofer, M., Sun, C., Nascimento, F. A., Halford, J. J., Dean, B. C., Galanti, M., Struck, A. F., Greenblatt, A. S., Lam, A. D., Herlopian, A., Nwankwo, C., Weber, D., Maus, D., Haider, H. A., Karakis, I., Yoo, J. Y., Ng, M. C., …, & Jing, J. (2025). Expert-Level Detection of Epilepsy Markers in EEG on Short and Long Timescales. NEJM AI, 2(7). https://doi.org/10.1056/aioa2401221
Vancouver
Li J, Goldenholz DM, Alkofer M, Sun C, Nascimento FA, Halford JJ, et al. Expert-Level Detection of Epilepsy Markers in EEG on Short and Long Timescales. NEJM AI. 2025;2(7). doi:10.1056/aioa2401221
BibTeX
@article{Li2025,
author = {Li, J and Goldenholz, D M and Alkofer, M and Sun, C and Nascimento, F A and Halford, J J and Dean, B C and Galanti, M and Struck, A F and Greenblatt, A S and Lam, A D and Herlopian, A and Nwankwo, C and Weber, D and Maus, D and Haider, H A and Karakis, I and Yoo, J Y and Ng, M C and Selioutski, O and Taraschenko, O and Osman, G and Katyal, R and Schmitt, S E and Benbadis, S and Cash, S S and Tatum, W O and Sheikh, Z and Kong, W Y and Bayas, G and Turley, N and Hong, S and Westover, M B and Jing, J},
title = {Expert-Level Detection of Epilepsy Markers in EEG on Short and Long Timescales},
journal = {NEJM AI},
year = {2025},
volume = {2},
number = {7},
doi = {10.1056/aioa2401221}
}
EUVIMED