Classification of Four-Class Motor Imaginary EEG Signals with Deep Learning Using Empirical Mode Decomposition and Welch Method

Öz

In electroencephalogram (EEG) based brain-computer interface (BCI) applications, it is very important to extract features from motor imagery (MI) signals obtained by imagining related limb movements and to classify them. Many different feature extraction methods and classification algorithms have been used in studies on MI-EEG signals. However, significant differences have been observed between the classification accuracies obtained as the number of classes increased in these signals. In the proposed method, feature extraction method including power spectral density (PSD) information of signals is presented. By applying empirical mode decomposition (EMD) to the raw EEG data, signals at different frequency levels were obtained. The PSD values of these signals were calculated using the welch method. The PSD values obtained were combined in a feature vector. Using the generated feature vectors, the long-short term memory (LSTM) network, a popular deep learning algorithm, was trained. The comparisons of the test accuracies obtained as a result of the training on the basis of individuals and channels were made in detail. As a result of the comparison, it was observed that the channels at the center of the scalp are more successful than the other channels.

Anahtar Kelimeler

• Empirical Mode Decomposition
• Brain-Computer Interface
• Long-Short Term Memory
• Welch Method.

Ampirik Mod Ayrıştırması ve Welch Yöntemini Kullanarak Dört Sınıflı Motor Hayali EEG Sinyallerinin Derin Öğrenme ile Sınıflandırılması

Öz

Elektroensefalogram (EEG) tabanlı beyin-bilgisayar arayüzü (BBA) uygulamalarında, kişilerin ilgili uzuv hareketlerini hayal etmesiyle elde edilen motor hayali (MI) sinyallerinden özellik çıkarmak ve bunları sınıflandırmak oldukça önemli bir konudur. MI-EEG sinyalleriyle ilgili yapılan çalışmalarda, birçok farklı özellik çıkarma yöntemleri ve sınıflandırma algoritmaları kullanılmıştır. Fakat bu sinyallerde sınıf sayısı arttıkça elde edilen sınıflandırma başarıları arasında belirgin farklar gözlemlenmiştir. Önerilen yöntemde, sinyallerin güç spektral yoğunluğu (PSD) bilgilerini içeren özellik çıkarma yöntemi sunulmuştur. Ham EEG verilerine ampirik mod ayrıştırması (EMD) uygulanarak farklı frekans seviyelerindeki sinyaller elde edilmiştir. Bu sinyallerin PSD değerleri welch yöntemi kullanılarak hesaplanmıştır. Elde edilen PSD değerleri bir öznitelik vektöründe birleştirilmiştir. Oluşturulan öznitelik vektörlerini kullanarak, popüler bir derin öğrenme algoritması olan uzun-kısa dönem hafıza (LSTM) ağı eğitilmiştir. Eğitim sonucunda elde edilen test başarılarının, kişiler ve kanallar bazındaki karşılaştırmaları detaylı olarak yapılmıştır. Karşılaştırma sonucunda kafa derisinin merkez noktasında bulunan kanalların, diğer kanallara göre daha başarılı oldukları görülmüştür.

Anahtar Kelimeler

• Ampirik Mod Ayrıştırması
• Beyin-Bilgisayar Arayüzü
• Uzun-Kısa Dönem Hafıza
• Welch Yöntemi.

Kaynakça

1. Amin, S. U., Alsulaiman, M., Muhammad, G., Mekhtiche, M. A., Hossain, M. S. (2019). Deep Learning for EEG motor imagery classification based on multi-layer CNNs feature fusion. Future Generation computer systems, 101, 542-554.
2. Wang, L., Zhang, X., Zhong, X., Zhang, Y. (2013). Analysis and classification of speech imagery EEG for BCI. Biomedical signal processing and control, 8 (6), 901-908.Congress on Computer Science and Engineering (APWC on CSE) (s. 34-39). IEEE.
3. Aydemir, O., & Kayikcioglu, T. (2014). Decision tree structure based classification of EEG signals recorded during two dimensional cursor movement imagery. Journal of neuroscience methods, 229, 68-75.
4. Li, F., He, F., Wang, F., Zhang, D., Xia, Y., Li, X. (2020). A Novel Simplified Convolutional Neural Network Classification Algorithm of Motor Imagery EEG Signals Based on Deep Learning. Applied Sciences, 10 (5), 1605.
5. Wang, T., Deng, J., He, B. (2004). Classifying EEG-based motor imagery tasks by means of time–frequency synthesized spatial patterns. Clinical Neurophysiology, 115 (12), 2744-2753.
6. Kam, T. E., Suk, H. I., Lee, S. W. (2013). Non-homogeneous spatial filter optimizationfor Electroencephalogram (EEG)-based motor imagery classification. Neurocomputing, 108, 58-68.
7. Tosun, M., & Kasım, Ö. (2020). Novel eye-blink artefact detection algorithm from raw EEG signals using FCN-based semantic segmentation method. IET Signal Processing, 14(8), 489-494.
8. Selim, S., Tantawi, M. M., Shedeed, H. A., Badr, A. (2018). A CSP\AM-BA-SVM Approach for Motor Imagery BCI System. IEEE Access, 6, 49192-49208.

9. Kumar, S. U., Inbarani, H. H. (2017). PSO-based feature selection and neighborhood rough set-based classification for BCI multiclass motor imagery task. Neural Computing and Applications, 28 (11), 3239-3258.
10. Rodríguez-Bermúdez, G., García-Laencina, P. J. (2012). Automatic and adaptive classification of electroencephalographic signals for brain computer interfaces. Journal of medical systems, 36 (1), 51-63.
11. Ge, S., Wang, R., Yu, D. (2014). Classification of four-class motor imagery employing single-channel electroencephalography. PloS one, 9 (6), e98019.
12. Yuyi, Z., Surui, L., Lijuan, S., Zhenxin, L., Bingchao, D. (2017). Motor imagery eeg discrimination using hilbert-huang entropy.
13. Kim, C., Sun, J., Liu, D., Wang, Q., Paek, S. (2018). An effective feature extraction method by power spectral density of EEG signal for 2-class motor imagery-based BCI. Medical biological engineering computing, 56 (9), 1645-1658.
14. Xie, X., Yu, Z. L., Lu, H., Gu, Z., & Li, Y. (2016). Motor imagery classification based on bilinear sub-manifold learning of symmetric positivedefinite matrices. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 25(6), 504-516.
15. Kevric, J., Subasi, A. (2017). Comparison of signal decomposition methods in classification of EEG signals for motor-imagery BCI system. Biomedical Signal Processing and Control, 31, 398-406.
16. Amin, S. U., Alsulaiman, M., Muhammad, G., Bencherif, M. A., Hossain, M. S. (2019). Multilevel weighted feature fusion using convolutional neural networks for EEG motor imagery classification. IEEE Access, 7, 18940-18950.
17. Jirayucharoensak, S., S. Pan-Ngum ve P. Israsena, (2014). "EEG-Based Emotion Recognition Using Deep Learning Network with Principal Component Based Covariate Shift Adaptation," The Scientific World Journal, C. s. 10.
18. Zhang, Z., Duan, F., Sole-Casals, J., Dinares-Ferran, J., Cichocki, A., Yang, Z., Sun, Z. (2019). A novel deep learning approach with data augmentation to classify motor imagery signals. IEEE Access, 7, 15945-15954.
19. Sakai, A., Minoda, Y., & Morikawa, K. Data augmentation methods for machine-learning-based classification of bio-signals. In 2017 10th Biomedical Engineering International Conference (BMEiCON) (pp. 1-4). IEEE.
20. Brunner, C., Leeb, R., Müller-Putz, G., Schlögl, A., Pfurtscheller, G. (2008). BCI Competition 2008–Graz data set A. Institute for Knowledge Discovery (Laboratory of Brain-Computer Interfaces), Graz University of Technology, 16.
21. Pigorini, A., Casali, A. G., Casarotto, S., Ferrarelli, F., Baselli, G., Mariotti, M., Rosanova, M. (2011). Time–frequency spectral analysis of TMS-evoked EEG oscillations by means of Hilbert–Huang transform. Journal of neuroscience methods, 198 (2), 236-245.
22. Alkan, A., Kiymik, M. K. (2006). Comparison of AR and Welch methods in epileptic seizure detection. Journal of Medical Systems, 30 (6), 413-419.
23. Alhagry, S., Fahmy, A. A., El-Khoribi, R. A. (2017). Emotion recognition based on EEG using LSTM recurrent neural network. Emotion, 8 (10), 355-358.