A deep neural network classifier for P300 BCI speller based on Cohen’s class time-frequency distribution

This paper presents a new method of predicting the P300 component of an electroencephalography (EEG) signal to recognize the characters in a P300 brain-computer interface (BCI) speller accurately. This method consists of a deep learning model and the nonlinear time-frequency features. It is believed that the combination of the deep model network and extracting the nonlinear features of the EEG led this research to a better prediction of the P300 and, therefore, character recognition. Cohen’s class distribution is used in order to extract the nonlinear features of the EEG. Evaluating all of the kernels, Butterworth found to be more informative and it produced better results. Based on the differences observed between time-frequency responses of target and nontarget signals, specific subbands are selected to extract seven features. A deep-structured neural network, namely stacked sparse autoencoders, is applied for BCI character recognition. This deep network reduces the dimension of feature space by extracting unsupervised features. Then, the features are fed to a Softmax classifier. Afterward, the whole network passes a fine-tuning phase by a supervised backpropagation algorithm. For evaluating the work, Dataset II of BCI Competition III is utilized. Based on the results, this approach would improve the accuracy in both P300 detection and character recognition. This research results in 82.7% and 93.5% accuracy for P300 classification and character recognition, respectively

PDF

___

[1] Amiri S, Fazel-Rezai R, Asadpour V. A review of hybrid brain-computer interface systems. Advances in Human- Computer Interaction 2013. doi: 10.1155/2013/187024
[2] Cecotti H, Graser A. Convolutional Neural Networks for P300 Detection with Application to Brain-Computer Interfaces. IEEE Transactions on Pattern Analysis and Machine Intelligence 2011; 33 (3): 433-445. doi: 10.1109/TPAMI.2010.125
[3] Elshout J, Garcia Molina G. Review of brain-computer interfaces based on the P300 evoked potential. Philips Research Europe 2009.
[4] Lotte F, Bougrain L, Cichocki A, Clerc M, Congedo M et al. A review of classification algorithms for EEG- based brain–computer interfaces: a 10 year update. Journal of Neural Engineering 2018; 15 (3). doi:10.1088/1741- 2552/aab2f2
[5] Vaid S, Singh P, Kaur C. EEG signal analysis for BCI interface: A Review. In: 2015 Fifth International Conference on Advanced Computing & Communication Technologies; Haryana; 2015. pp. 143-147. doi: 10.1109/ACCT.2015.72
[6] Wolpaw JR, Birbaumer N, Heetderks WJ, McFarland DJ, Peckham PH et al. Brain-computer interface technology: a review of the first international meeting. IEEE Transactions on Rehabilitation Engineering 2000; 8(2): 164-173. doi: 10.1109/tre.2000.847807
[7] Olivas-Padilla BE, Chacon-Murguia MI. Classification of multiple motor imagery using deep convolutional neural networks and spatial filters. Applied Soft Computing 2019; 75: 461-472. doi: j.asoc.2018.11.031 1238
GHAZIKHANI and ROUHANI/Turk J Elec Eng & Comp Sci [8] Vallabhaneni A, Wang T, He B. Brain—Computer Interface. In: He B. (eds) Neural Engineering. Bioelectric Engineering. Boston, MA, USA: Springer2005. doi: 10.1007/0-306-48610-5_3
[9] Al-Sharhan S, Bimba A. Adaptive multi-parent crossover GA for feature optimization in epileptic seizure identifi- cation. Applied Soft Computing 2019; 75: 575-587. doi: j.asoc.2018.11.012
[10] Bressler SL, Ding M. Event‐Related Potentials. Wiley Encyclopedia of Biomedical Engineering 2006. doi: 10.1002/9780471740360.ebs0455
[11] Koukkou M, Koenig T, Bänninger A, Rieger K, Diaz Hernandez L et al. Neurobiology of schizophrenia: electrophys- iological indices. In: A. Javed & K. N. Fountoulakis (eds.), Advances in Psychiatry. Boston, MA, USA: Springer International Publishing, 2019, pp. 433–459. doi: 10.1007/978-3-319-70554-5_27
[12] Farwell LA, Donchin E. Talking off the top of your head: toward a mental prosthesis utilizing event-related brain potentials. Electroencephalography and Clinical Neurophysiology 1988; 70 (6), 510-523. doi: 10.1016/0013- 4694(88)90149-6
[13] Fazel-Rezai R, Allison BZ, Guger C, Sellers EW, Kleih SC et al. P300 brain computer interface: current challenges and emerging trends. Frontiers in Neuroengineering 2012; 5, 14. doi: 10.3389/fneng.2012.00014
[14] Amiri S, Fazel-Rezai R, Asadpour V. A review of hybrid brain-computer interface systems. Advances in Human- Computer Interaction 2013; 2013: 1. doi: 10.1155/2013/187024
[15] Song C, Jeon S, Lee S, Ha HG, Kim J et al. Augmented reality-based electrode guidance system for reliable electroencephalography. BioMedical Engineering OnLine 2018; 17(1): 64. doi: 10.1186/s12938-018-0500-x
[16] Vařeka L, Prokop T, Mouček R, Mautner P, Štěbeták J. Application of stacked autoencoders to P300 experi- mental data. In: International Conference on Artificial Intelligence and Soft Computing; 2017. pp. 187–198. doi: 10.1007/978-3-319-59063-9_17
[17] Cömert Z, Boopathi AM, Velappan S, Yang Z, Kocamaz AF. The influences of different window functions and lengths on image-based time-frequency features of fetal heart rate signals. In: 2018 26th Signal Processing and Communications Applications Conference (SIU); 2018. pp. 1-4. doi: 10.1109/SIU.2018.8404247
[18] Iscan Z, Dokur Z, Demiralp T. Classification of electroencephalogram signals with combined time and frequency features. Expert Systems with Applications 2011; 38 (8), 10499-10505. doi: 10.1016/j.eswa.2011.02.110
[19] Murugappan M, Nagarajan R, Yaacob S. Comparison of different wavelet features from EEG signals for classifying human emotions. In: 2009 IEEE Symposium on Industrial Electronics & Applications; Kuala Lumpur, Malaysia; 2009. pp. 836-841. doi: 10.1109/ISIEA.2009.5356339
[20] Ebrahimzadeh E, Alavi SM, Bijar A, Pakkhesal A. A novel approach for detection of deception using smoothed pseudo Wigner–Ville distribution (SPWVD). Journal of Biomedical Science and Engineering 2013; 6 (01): 8. doi: 10.4236/jbise.2013.61002
[21] Cheng W, Ren Z, Qiao X. Improved Wigner–Ville distribution for extracting the visual evoked potential P300. Journal of Shanxi University (Natural Science Edition) 2013; 2, 18.
[22] Liu M, Wu W, Gu Z, Yu Z, Qi F et al. Deep learning based on Batch Normalization for P300 signal detection. Neurocomputing 2018; 275: 288-297. doi: 10.1016/j.neucom.2017.08.039
[23] Sobhani A. P300 Classification using Deep Belief Nets. MSc, Colorado State University, Fort Collins, Colorado, USA, 2014.
[24] Vařeka L, Mautner P. Stacked autoencoders for the P300 component detection. Frontiers in Neuroscience 2017; 11: 302. doi: 10.3389/fnins.2017.00302
[25] Ghazikhani H, Rouhani M. A stacked autoencoders approach for a P300 Speller BCI. In: 8th International Con- ference on Computer and Knowledge Engineering (ICCKE2018); Mashhad, Iran; 2018. pp. 25-26. doi: 10.1109/IC- CKE.2018.856653
26] Yoon K, Kim K. Multiple kernel learning based on three discriminant features for a P300 speller BCI. Neurocom- puting 2017; 237: 133–144. doi: 10.1016/j.neucom.2016.09.053
[27] Rakotomamonjy A, Guigue V. BCI competition III: dataset II-ensemble of SVMs for BCI P300 speller. IEEE Transactions on Biomedical Engineering 2008; 55 (3): 1147–1154. doi: 10.1109/TBME.2008.915728
[28] Konopko K. An implementation of the Cohen’s class time-frequency distributions on a massively parallel processor. Prz. Elektrot 2012; 88 (9B): 289-291.
[29] Ochoa JB. EEG signal classification for brain computer interface applications. Ecole Polytechnique Federale De Lausanne 2002; 7: 1-72.
[30] Kavallieratou E, Fakotakis N, Kokkinakis G. Skew angle estimation in document processing using Cohen’s class distributions. Pattern Recognition Letters 1999; 20 (11): 1305-1311. doi: 10.1016/S0167-8655(99)00097-5
[31] Cohen L. Time-frequency distributions-a review. Proceedings of the IEEE 1989; 77(7): 941-981. doi: 10.1109/5.30749.
[32] Poyil AT, Nasimudeen KM, Aljahdali S. Significance of Cohen’s class for time frequency analysis of signals. International Journal of Computer Applications 2013; 72 (12): 1-8. doi: 10.5120/12543-8960
[33] Bengio Y, Lamblin P, Popovici D, Larochelle H. Greedy layer-wise training of deep networks. In: NIPS’06: Proceedings of the 19th International Conference on Neural Information Processing; Cambridge, MA, USA; 2006. pp. 153-160.
[34] Zhao C, Zhang L. Spectral-spatial stacked autoencoders based on low-rank and sparse matrix decomposition for hy- perspectral anomaly detection. Infrared Physics & Technology 2018; 92: 166-176. doi: 10.1016/j.infrared.2018.06.001
[35] Hoffmann U, Garcia G, Vesin JM, Ebrahimi T. Application of the evidence framework to brain-computer interfaces. In: The 26th Annual International Conference of the IEEE Engineering in Medicine and Biology Society; San Francisco, CA, USA; 2004. pp. 446-449