APSIPA Transactions on Signal and Information Processing > Vol 13 > Issue 1

Riemannian Manifold-based Epileptic Seizure Detection Using Transfer Learning and Artifact Rejection Techniques

Kazi Mahmudul Hassan, Tokyo University of Agriculture and Technology, Japan, Xuyang Zhao, Tokyo University of Agriculture and Technology, Japan, Hidenori Sugano, Juntendo University School of Medicine, Japan, Toshihisa Tanaka, Tokyo University of Agriculture and Technology, Japan
 
Suggested Citation
Kazi Mahmudul Hassan, Xuyang Zhao, Hidenori Sugano and Toshihisa Tanaka (2024), "Riemannian Manifold-based Epileptic Seizure Detection Using Transfer Learning and Artifact Rejection Techniques", APSIPA Transactions on Signal and Information Processing: Vol. 13: No. 1, e33. http://dx.doi.org/10.1561/116.20240032

Publication Date: 04 Dec 2024
© 2024 K. M. Hassan, X. Zhao, H. Sugano, and T. Tanaka
 
Subjects
Pattern recognition and learning,  Biological and biomedical signal processing,  Multiresolution signal processing,  Nonlinear signal processing,  Statistical signal processing,  Statistical/Machine learning,  Classification and prediction,  Deep learning
 
Keywords
ElectroencephalogramanomalyseizureartifactRiemannian potato
 

Share

Open Access

This is published under the terms of CC BY-NC.

Downloaded: 99 times

In this article:
Introduction 
Previous Works 
Materials and Methods 
Experiment 
Results, Discussion and Future Direction 
Conclusion 
Appendices 
References 

Abstract

Improvement in technology and the availability of electroencephalogram (EEG) data have raised the demand for automated seizure detection in long-term EEG recordings. This study proposes a framework to automate seizure detection from long-term EEG by combining anomaly detection, artifact removal, and seizure detection techniques, along with Riemannian manifold and transfer learning approaches. First, the method identifies potential EEG segments for seizures using Riemannian manifold-based features from covariance matrices. Next, it removes extra-physiological artifacts using power-based features. Finally, it uses Riemannian potato-based features to classify the remaining segments with a LightGBM classifier. The method’s performance was evaluated on two datasets–a private dataset (Juntendo) and a public dataset (Siena)–using leave-one-patient-out cross-validation. For the Juntendo dataset, the method achieved an average performance across all subjects with a sensitivity of 89.9%, specificity of 96.8%, precision of 33.3%, and an F1–score of 44.5%. On the Siena dataset, the method achieved a sensitivity of 63.8%, specificity of 98.7%, precision of 32.4%, and an F1–score of 40.5%. Processing EEG data in multiple stages helps reduce the class imbalance problem. Therefore, automating the seizure detection process will ease the practitioner’s workload.

DOI:10.1561/116.20240032