Evrişimli Sinir Ağları Kullanılarak Epileptik Nöbet Tespiti: CHB-MIT Scalp EEG Veri Kümesi Üzerinde Çoklu Sınıflandırma

Abstract

Epileptik nöbet tespiti, epilepsinin yönetimi ve tedavisinde kritik bir görevdir ve doğru ve etkili tanı araçlarının geliştirilmesini gerektirir. Bu çalışma, CHB-MIT Scalp EEG veri kümesinden yararlanarak, kafa derisi EEG kayıtlarından epileptik nöbetleri tespit etmek için Evrişimsel Sinir Ağları (CNN) kullanan modeller sunar. Analizimizin sağlamlığını artırmak için, epilepsi teşhisi konmuş farklı sayıda denekle bu veri kümesinin birden fazla versiyonunu oluşturduk. Ek olarak, model performansı üzerindeki etkilerini değerlendirmek için temel ve gelişmiş olarak adlandırılan iki tür etiketleme tekniği kullandık. Nöbet tespitindeki etkinliklerini belirlemek için özellikle ResNet50, ResNet101, VGG16, InceptionV3, Xception ve DenseNet121 olmak üzere CNN tabanlı modeller kullandık. Bu çalışma, bu alandaki üç-sınıf sınıflandırma problemini ele alan ilk çalışmadır. DenseNet121 modeli, %85,52'lik bir F1 puanı ile en yüksek performansı elde etmiştir.

Keywords

• EEG
• Epilepsi
• Atak Tespiti
• CNN
• Derin Öğrenme

Epileptic Seizure Detection Using Convolutional Neural Networks: Multi-class Classification on the CHB-MIT Scalp EEG Dataset

Abstract

The detection of epileptic seizures is a critical task in the management and treatment of epilepsy, requiring the development of accurate and effective diagnostic tools. This study presents models using Convolutional Neural Networks (CNN) to detect epileptic seizures from scalp EEG recordings, leveraging the CHB-MIT Scalp EEG dataset. To enhance the robustness of the analysis, multiple versions of this dataset were created with a different number of subjects diagnosed with epilepsy. Additionally, two types of labeling techniques, referred to as basic and advanced, were employed to evaluate their impact on model performance. CNN-based models, particularly ResNet50, ResNet101, VGG16, InceptionV3, Xception, and DenseNet121, were used to assess their effectiveness in seizure detection. This study is the first to address the three-class classification problem specifically with the CHB-MIT Scalp EEG dataset. The DenseNet121 model achieved the highest performance with an F1 score of 85.52%.

Keywords

• Epilepsy
• Seizure Detection
• CNN
• Deep learning
• EEG

References

1. WHO. World Health Organization: epilepsy: epidemiology, aetiology and prognosis, WHO factsheet, 2001.
2. Fisher, R. S., Boas, W. V. E., Blume, W., Elger, C., Genton, P., Lee, P., & Engel Jr, J. Epileptic seizures and epilepsy: definitions proposed by the International League Against Epilepsy (ILAE) and the International Bureau for Epilepsy (IBE), Epilepsia, 2015, 46.4, pp. 470-472.
3. Shoeb, A. H. Application of machine learning to epileptic seizure onset detection and treatment (Doctoral dissertation, Massachusetts Institute of Technology), 2009.
4. CHB-MIT-Scalp-EEG-Seizure-Classification, https://github.com/erencalbay/CHB-MIT-Scalp-EEG-Seizure-Classification, Erişim Tarihi: 31.07.2024.
5. Esmaeilpour, A., Tabarestani, S. S., & Niazi, A. Deep learning‐based seizure prediction using EEG signals: A comparative analysis of classification methods on the CHB‐MIT dataset, Engineering Reports, 2024, e12918.
6. Georgis-Yap, Z., Popovic, M. R., & Khan, S. S. Supervised and Unsupervised Deep Learning Approaches for EEG Seizure Prediction. Journal of Healthcare Informatics Research, 2024, pp. 1-27.
7. Xia, L., Wang, R., Ye, H., Jiang, B., Li, G., Ma, C., & Gao, Z. Hybrid LSTM-Transformer model for the prediction of epileptic seizure using scalp EEG. IEEE Sensors Journal. 2024.
8. Singh, K., & Malhotra, J. Two-layer LSTM network-based prediction of epileptic seizures using EEG spectral features. Complex & Intelligent Systems, 2022, pp. 2405-2418

9. Aslam, M. H., Usman, S. M., Khalid, S., Anwar, A., Alroobaea, R., Hussain, S., ... & Yasin, A. Classification of EEG signals for prediction of epileptic seizures. Applied Sciences, 12(14), 2022, pp. 7251.
10. Jemal, Imene, et al. An interpretable deep learning classifier for epileptic seizure prediction using EEG data, IEEE Access 10, 2022, pp. 60141-60150.
11. Ang, Kai Keng, et al. Filter bank common spatial pattern (FBCSP) in brain-computer interface, 2008 IEEE international joint conference on neural networks (IEEE world congress on computational intelligence) IEEE, 2008.
12. Xu, Yankun, Jie Yang, and Mohamad Sawan. Multichannel synthetic preictal EEG signals to enhance the prediction of epileptic seizures. IEEE Transactions on Biomedical Engineering 69.11, 2022, pp. 3516-3525.
13. Jumaah, Mahmood A., Ammar Ibrahim Shihab, and Akeel Abdulkareem Farhan. Epileptic Seizures Detection Using DCT-II and KNN Classifier in Long-Term EEG Signals. Iraqi Journal of Science, 2020, pp. 2687-2694.
14. Butterworth, S. On the theory of filter amplifiers. Wireless Engineer, 7(6), 1930, pp. 536-541.
15. Duhamel, P., & Vetterli, M. Fast Fourier transforms: a tutorial review and a state of the art. Signal processing, 19(4), 1990, 259-299.
16. Barbe, K., Pintelon, R., & Schoukens, J. Welch method revisited: nonparametric power spectrum estimation via circular overlap. IEEE Transactions on signal processing, 58(2), 2009, pp. 553-565.
17. Lin, J. Divergence measures based on the Shannon entropy. IEEE Transactions on Information theory, 37(1), 1991, pp. 145-151.
18. Jiang, X., Liu, X., Liu, Y., Wang, Q., Li, B., & Zhang, L. Epileptic seizures detection and the analysis of optimal seizure prediction horizon based on frequency and phase analysis. Frontiers in Neuroscience, 17, 2023, 1191683.
19. Shokouh Alaei, H., Khalilzadeh, M. A., & Gorji, A. Optimal selection of SOP and SPH using fuzzy inference system for on-line epileptic seizure prediction based on EEG phase synchronization. Australasian physical & engineering sciences in medicine, 42, 2019, pp. 1049-1068.
20. Huang, G., Liu, Z., Pleiss, G., Van Der Maaten, L., & Weinberger, K. Q. Convolutional networks with dense connectivity. IEEE transactions on pattern analysis and machine intelligence, 44(12), 2019, pp. 8704-8716.