Jonas Munch Nielsen1, Ivan C Zibrandtsen2, Paolo Masulli3, Torben Lykke Sørensen4, Tobias S Andersen5, Troels Wesenberg Kjær6. 1. Department of Neurology, Zealand University Hospital, 4000 Roskilde, Denmark; Department of Clinical Medicine, University of Copenhagen, 2200 Copenhagen N, Denmark. Electronic address: jomun@regionsjaelland.dk. 2. Department of Neurology, Zealand University Hospital, 4000 Roskilde, Denmark. 3. Department of Applied Mathematics and Computer Science DTU Compute, Section of Cognitive Systems, Technical University of Denmark, 2800 Kgs. Lyngby, Denmark; iMotions A/S, 1621 Copenhagen V, Denmark. 4. Department of Ophthalmology, Zealand University Hospital, 4000 Roskilde, Denmark; Department of Clinical Medicine, University of Copenhagen, 2200 Copenhagen N, Denmark. 5. Department of Applied Mathematics and Computer Science DTU Compute, Section of Cognitive Systems, Technical University of Denmark, 2800 Kgs. Lyngby, Denmark. 6. Department of Neurology, Zealand University Hospital, 4000 Roskilde, Denmark; Department of Clinical Medicine, University of Copenhagen, 2200 Copenhagen N, Denmark.
Abstract
OBJECTIVE: To explore the possibilities of wearable multi-modal monitoring in epilepsy and to identify effective strategies for seizure-detection. METHODS: Thirty patients with suspected epilepsy admitted to video electroencephalography (EEG) monitoring were equipped with a wearable multi-modal setup capable of continuous recording of electrocardiography (ECG), accelerometry (ACM) and behind-the-ear EEG. A support vector machine (SVM) algorithm was trained for cross-modal automated seizure detection. Visualizations of multi-modal time series data were used to generate ideas for seizure detection strategies. RESULTS: Three patients had more than five seizures and were eligible for SVM classification. Classification of 47 focal tonic seizures in one patient found a sensitivity of 84% with a false alarm rate (FAR) of 8/24 h. In two patients each with nine focal nonmotor seizures it yielded a sensitivity of 100% and a FAR of 13/24 h and 5/24. Visual comparisons of features were used to identify strategies for seizure detection in future research. CONCLUSIONS: Multi-modal monitoring in epilepsy using wearables is feasible and automatic seizure detection may benefit from multiple modalities when compared to uni-modal EEG. SIGNIFICANCE: This study is unique in exploring a combination of wearable EEG, ECG and ACM and can help inform future research on monitoring of epilepsy.
OBJECTIVE: To explore the possibilities of wearable multi-modal monitoring in epilepsy and to identify effective strategies for seizure-detection. METHODS: Thirty patients with suspected epilepsy admitted to video electroencephalography (EEG) monitoring were equipped with a wearable multi-modal setup capable of continuous recording of electrocardiography (ECG), accelerometry (ACM) and behind-the-ear EEG. A support vector machine (SVM) algorithm was trained for cross-modal automated seizure detection. Visualizations of multi-modal time series data were used to generate ideas for seizure detection strategies. RESULTS: Three patients had more than five seizures and were eligible for SVM classification. Classification of 47 focal tonic seizures in one patient found a sensitivity of 84% with a false alarm rate (FAR) of 8/24 h. In two patients each with nine focal nonmotor seizures it yielded a sensitivity of 100% and a FAR of 13/24 h and 5/24. Visual comparisons of features were used to identify strategies for seizure detection in future research. CONCLUSIONS: Multi-modal monitoring in epilepsy using wearables is feasible and automatic seizure detection may benefit from multiple modalities when compared to uni-modal EEG. SIGNIFICANCE: This study is unique in exploring a combination of wearable EEG, ECG and ACM and can help inform future research on monitoring of epilepsy.
Authors: Sebastian Böttcher; Elisa Bruno; Nino Epitashvili; Matthias Dümpelmann; Nicolas Zabler; Martin Glasstetter; Valentina Ticcinelli; Sarah Thorpe; Simon Lees; Kristof Van Laerhoven; Mark P Richardson; Andreas Schulze-Bonhage Journal: Sensors (Basel) Date: 2022-04-26 Impact factor: 3.847