Kombine Derin Öğrenme Tabanlı Epileptik Nöbet Teşhisi

Öz

Epilepsi hastalığı yaygın nörolojik hastalıklardan bir tanesi olarak öne çıkmaktadır. Epilepsi hastalığının teşhisinde elektroensefalografi (EEG) kullanılarak beynin sinirsel aktivitesi gözlemlenir ve bu da epilepsi hastalığının teşhisine olanak sağlar. Günümüzde genel olarak biyolojik sinyallerden hastalık teşhisinde klasik makine öğrenmesi yöntemleri sıklıkla kullanılmakla birlikte son yıllarda derin öğrenme yapıları ön plana çıkmaktadır. Derin öğrenme ağları sinyallerden özellik çıkarımına gerek duymaması, özellikler için ek bir çaba gerektirmemesi, insan kaynaklı hesaplama hatalarının önüne geçmesi ve zaman kaybının önüne geçmesi açısından klasik makine öğrenmesi yöntemlerine göre daha avantajlı bir konuma gelmektedir. Bu çalışmada, zaman serisi EEG sinyalini, zaman-frekans bileşenlerini temsil edecek görüntüleri ve ham EEG sinyallerinin sayısal değerlerini kullanarak epilepsi nöbet aktivitesini otomatik bir şekilde tespit eden kombine bir derin öğrenme modeli üzerine çalışılmıştır. Çalışmada Bonn Üniversitesinin halka açık epilepsi veri seti kullanılmıştır. Veri seti sağlıklı ve epilepsi hastası insanlardan kaydedilen A,B,C,D,E şeklinde etiketlenmiş EEG kayıtlarını içermektedir. Bu çalışmada EGG sinyallerinin zaman dizisini ve zamana bağlı EEG sinyallerinin zaman-frekans-görüntü dönüşümlerini kullanarak kombine bir model ortaya koyulmuştur. Sinyalleri görüntülere dönüştürmede CWT ve STFT yöntemleri kullanılmıştır. Oluşturulan modelin CNN girdilerinde STFT görüntüleri kullanıldığında ikili sınıflandırma için %99.47 doğruluk oranı elde edilmiştir. CWT görüntüleri ile ise %99.27 doğruluk oranına ulaşılmıştır. Elde edilen model, EEG verilerinde epilepsi nöbet aktivitesinin olup olmadığını yüksek başarı ile tespit edebilmektedir.

Anahtar Kelimeler

• Epilepsi
• Evrişimli Sinir Ağı
• Yinelenen Sinir Ağı
• EEG
• Epileptik nöbet teşhisi

Combined Deep Learning Based Epileptic Seizure Diagnosis

Öz

Epilepsy stands out as one of the common neurological diseases. In the diagnosis of epilepsy, the neural activity of the brain is observed using electroencephalography (EEG), which allows the diagnosis of epilepsy disease. Although classical machine learning methods are frequently used in the diagnosis of diseases from biological signals, deep learning structures have come to the fore in recent years. Deep learning networks are in a more advantageous position than classical machine learning methods in terms of not requiring feature extraction from signals, requiring no additional effort for features, preventing human-induced computational errors, and preventing time loss. In this study, a combined deep learning model that automatically detects epileptic seizure activity using images representing the time-frequency components of the time domain EEG signal and numerical values of the raw EEG signals was studied. The public epilepsy dataset of the University of Bonn was used in the study. The dataset includes EEG recordings labelled as A, B, C, D, E recorded from healthy and epileptic people. In this study, we made a combined model using the time sequence of EGG signals and time-frequency-image transformations of time-dependent EEG signals. We used CWT and STFT to convert signals to images. We achieved 99.47% accuracy for binary classification when we used STFT images in the CNN inputs of our model. With CWT images, we performed an accuracy rate of 99.27%. The model obtained can detect with high success whether there is epileptic seizure activity in EEG data.

Anahtar Kelimeler

• Epilepsy
• Convolutional Neural Network
• Recurrent Neural Network
• EEG
• Epileptic Seizure Diagnos

Kaynakça

1. Alzubaidi, L., Zhang, J., Humaidi, A. J., Al-Dujaili, A., Duan, Y., Al-Shamma, O., . . . Farhan, L. (2021). Review of deep learning: concepts, CNN architectures, challenges, applications, future directions. Journal of Big Data, 8(1), 53. doi:10.1186/s40537-021-00444-8
2. Andrzejak, R. G., Lehnertz, K., Mormann, F., Rieke, C., David, P., & Elger, C. E. (2001). Indications of nonlinear deterministic and finite-dimensional structures in time series of brain electrical activity: dependence on recording region and brain state. Phys Rev E Stat Nonlin Soft Matter Phys, 64(6 Pt 1), 061907. doi:10.1103/PhysRevE.64.061907
3. Bajaj, N. (2020). Wavelets for EEG Analysis.
4. Beghi, E. (2020). The Epidemiology of Epilepsy. Neuroepidemiology, 54(2), 185-191. doi:10.1159/000503831
5. Brian, P., Avirath, S., Sean, C., Victoria, G., Antoni, V.-C., & Adrien, M. (2021). Brain Informatics. doi:10.21203/rs.3.rs-112880/v1
6. Chen, G. (2016). A Gentle Tutorial of Recurrent Neural Network with Error Backpropagation.
7. Dwi Saputro, I. R., Maryati, N. D., Solihati, S. R., Wijayanto, I., Hadiyoso, S., & Patmasari, R. (2019). Seizure Type Classification on EEG Signal using Support Vector Machine. Journal of Physics: Conference Series, 1201, 012065. doi:10.1088/1742-6596/1201/1/012065
8. Gu, J., Wang, Z., Kuen, J., Ma, L., Shahroudy, A., Shuai, B., . . . Chen, T. (2018). Recent advances in convolutional neural networks. Pattern Recognition, 77, 354-377. doi:https://doi.org/10.1016/j.patcog.2017.10.013

9. Hussain, L. (2018). Detecting epileptic seizure with different feature extracting strategies using robust machine learning classification techniques by applying advance parameter optimization approach. Cognitive neurodynamics, 12(3), 271-294. doi:10.1007/s11571-018-9477-1
10. Indolia, S., Goswami, A. K., Mishra, S. P., & Asopa, P. (2018). Conceptual Understanding of Convolutional Neural Network- A Deep Learning Approach. Procedia Computer Science, 132, 679-688. doi:https://doi.org/10.1016/j.procs.2018.05.069
11. Jaiswal, A. K., & Banka, H. (2017). Local pattern transformation based feature extraction techniques for classification of epileptic EEG signals. Biomedical Signal Processing and Control, 34, 81-92. doi:https://doi.org/10.1016/j.bspc.2017.01.005
12. Kıymık, M. K., Güler, İ., Dizibüyük, A., & Akın, M. (2005). Comparison of STFT and wavelet transform methods in determining epileptic seizure activity in EEG signals for real-time application. Computers in Biology and Medicine, 35(7), 603-616. doi:https://doi.org/10.1016/j.compbiomed.2004.05.001
13. Mursalin, M., Zhang, Y., Chen, Y., & Chawla, N. (2017). Automated epileptic seizure detection using improved correlation-based feature selection with random forest classifier. Neurocomputing, 241, 204-214.
14. Rajoub, B. (2020). Chapter 2 - Characterization of biomedical signals: Feature engineering and extraction. In W. Zgallai (Ed.), Biomedical Signal Processing and Artificial Intelligence in Healthcare (pp. 29-50): Academic Press.
15. Ravi Kumar, M., & Srinivasa Rao, Y. (2019). Epileptic seizures classification in EEG signal based on semantic features and variational mode decomposition. Cluster Computing, 22(6), 13521-13531. doi:10.1007/s10586-018-1995-4
16. Sharmila, A., & Geethanjali, P. (2016). DWT Based Detection of Epileptic Seizure From EEG Signals Using Naive Bayes and k-NN Classifiers. IEEE Access, 4, 7716-7727. doi:10.1109/ACCESS.2016.2585661
17. Sheoran, P., Rathee, N., & Saini, J. S. (2020, 27-28 Feb. 2020). Epileptic Seizure Detection using Bidimensional Empirical Mode Decomposition and Distance Metric Learning on Scalogram. Paper presented at the 2020 7th International Conference on Signal Processing and Integrated Networks (SPIN).
18. Shi, X., Wang, T., Wang, L., Liu, H., & Yan, N. (2019, 18-21 Nov. 2019). Hybrid Convolutional Recurrent Neural Networks Outperform CNN and RNN in Task-state EEG Detection for Parkinson's Disease. Paper presented at the 2019 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC).
19. Shoka, A., Dessouky, M., El-Sherbeny, A., & El-Sayed, A. (2019). Literature Review on EEG Preprocessing, Feature Extraction, and Classifications Techniques. Menoufia Journal of Electronic Engineering Research, 28(ICEEM2019-Special Issue), 292-299. doi:10.21608/mjeer.2019.64927
20. Siddiqui, M. K., Morales-Menendez, R., Huang, X., & Hussain, N. (2020). A review of epileptic seizure detection using machine learning classifiers. Brain Informatics, 7(1), 5-5. doi:10.1186/s40708-020-00105-1
21. Singh, A., & Trevick, S. (2016). The Epidemiology of Global Epilepsy. Neurologic clinics, 34(4), 837-847. doi:10.1016/j.ncl.2016.06.015
22. van Mierlo, P., Vorderwülbecke, B. J., Staljanssens, W., Seeck, M., & Vulliémoz, S. (2020). Ictal EEG source localization in focal epilepsy: Review and future perspectives. Clinical Neurophysiology, 131(11), 2600-2616. doi:https://doi.org/10.1016/j.clinph.2020.08.001
23. Wang, Y., Dai, Y., Liu, Z., Guo, J., Cao, G., Ouyang, M., . . . Zhao, G. (2021). Computer-Aided Intracranial EEG Signal Identification Method Based on a Multi-Branch Deep Learning Fusion Model and Clinical Validation. Brain sciences, 11(5). doi:10.3390/brainsci11050615
24. Xu, S., Wang, Z., Sun, J., Zhang, Z., Wu, Z., Yang, T., . . . Cheng, C. (2020). Using a deep recurrent neural network with EEG signal to detect Parkinson's disease. Annals of translational medicine, 8(14), 874-874. doi:10.21037/atm-20-5100
25. Zhao, W., Zhao, W., Wang, W., Jiang, X., Zhang, X., Peng, Y., . . . Zhang, G. (2020). A Novel Deep Neural Network for Robust Detection of Seizures Using EEG Signals. Computational and Mathematical Methods in Medicine, 2020, 9689821. doi:10.1155/2020/9689821