Deep learning–based detection of generalized convulsive seizures using a wrist‐worn accelerometer
Epilepsia · 2025-09 · 66 (S3) · 53-63
DOI: 10.1111/epi.18406 ↗Kurz gefasst
Multizentrische Studie in der Zeitschrift Epilepsia, online erschienen am 23.04.2025. Verfasst von Antoine Spahr und 16 weiteren Autor:innen. Beteiligte Einrichtung: NeuroDigital@NeuroTech, Department of Clinical Neurosciences, Lausanne University Hospital (CHUV), University of Lausanne, Lausanne, Switzerland. Hauptthemen laut MEDLINE: Deep Learning, Accelerometry, Wearable Electronic Devices, Seizures. Die Arbeit wurde bisher 14-mal zitiert (Crossref) und nennt 40 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung.
Abstract (Originalsprache)
Objective
To develop and validate a wrist‐worn accelerometer‐based, deep‐learning tunable algorithm for the automated detection of generalized or bilateral convulsive seizures (CSs) to be integrated with off‐the‐shelf smartwatches.
Methods
We conducted a prospective multi‐center study across eight European epilepsy monitoring units, collecting data from 384 patients undergoing video electroencephalography (vEEG) monitoring with a wrist‐worn three dimensional (3D)–accelerometer sensor. We developed an ensemble‐based convolutional neural network architecture with tunable sensitivity through quantile‐based aggregation. The model, referred to as Episave, used accelerometer amplitude as input. It was trained on data from 37 patients who had 54 CSs and evaluated on an independent dataset comprising 347 patients, including 33 who had 49 CSs.
Results
Cross‐validation on the training set showed that optimal performance was obtained with an aggregation quantile of 60, with a 98% sensitivity, and a false alarm rate (FAR) of 1/6 days. Using this quantile on the independent test set, the model achieved a 96% sensitivity (95% confidence interval [CI]: 90%–100%), a FAR of
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- NeuroDigital@NeuroTech, Department of Clinical Neurosciences, Lausanne University Hospital (CHUV), University of Lausanne, Lausanne, Switzerland.
- Department of Neurology, Clinic Hietzing, Vienna, Austria.
- Karl Landsteiner Institute for Clinical Epilepsy Research and Cognitive Neurology, Medical Faculty Sigmund Freud University, Vienna, Austria.
- Swiss Epilepsy Center, Klinik Lengg, Zurich, Switzerland.
- Department of Clinical Neurophysiology, Aarhus University Hospital, Aarhus, Denmark.
- Danish Epilepsy Centre, Dianalund, Denmark.
ORCID-Profile: Adriano Bernini, Johannes P Koren, Lukas Imbach, Sàndor Beniczky, Sylvain Rheims, Martin Fabricius, Berhard J Steinhoff, Isabelle Beuchat, Jonathan Dan, Philippe Ryvlin
Förderung
- Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung
Zitieren
APA
Spahr, A., Bernini, A., Ducouret, P., Baumgartner, C., Koren, J. P., Imbach, L., Beniczky, S., Larsen, S. A., Rheims, S., Fabricius, M., Seeck, M., Steinhoff, B. J., Beuchat, I., Dan, J., Atienza, D. A., Bardyn, C. E., & Ryvlin, P. (2025). Deep learning–based detection of generalized convulsive seizures using a wrist‐worn accelerometer. Epilepsia, 66(S3), 53-63. https://doi.org/10.1111/epi.18406
Vancouver
Spahr A, Bernini A, Ducouret P, Baumgartner C, Koren JP, Imbach L, et al. Deep learning–based detection of generalized convulsive seizures using a wrist‐worn accelerometer. Epilepsia. 2025;66(S3):53-63. doi:10.1111/epi.18406
BibTeX
@article{Spahr2025,
author = {Spahr, Antoine and Bernini, Adriano and Ducouret, Pauline and Baumgartner, Christoph and Koren, Johannes P and Imbach, Lukas and Beniczky, Sàndor and Larsen, Sidsel A and Rheims, Sylvain and Fabricius, Martin and Seeck, Margitta and Steinhoff, Berhard J and Beuchat, Isabelle and Dan, Jonathan and Atienza, David A and Bardyn, Charles-Edouard and Ryvlin, Philippe},
title = {Deep learning–based detection of generalized convulsive seizures using a wrist‐worn accelerometer},
journal = {Epilepsia},
year = {2025},
volume = {66},
number = {S3},
pages = {53--63},
doi = {10.1111/epi.18406}
}
EUVIMED