Semisupervised Deep Stacking Network with Adaptive Learning Rate Strategy for Motor Imagery EEG Recognition

Practical motor imagery electroencephalogram (EEG) data-based applications are limited by the waste of unlabeled samples in supervised learning and excessive time consumption in the pretraining period. A semisupervised deep stacking network with an adaptive learning rate strategy (SADSN) is proposed to solve the sample loss caused by supervised learning of EEG data and the extraction of manual features. The SADSN adopts the idea of an adaptive learning rate into a contrastive divergence (CD) algorithm to accelerate its convergence. Prior knowledge is introduced into the intermediary layer of the deep stacking network, and a restricted Boltzmann machine is trained by a semisupervised method in which the adjusting scope of the coefficient in learning rate is determined by performance analysis. Several EEG data sets are carried out to evaluate the performance of the proposed method. The results show that the recognition accuracy of SADSN is advanced with a more significant convergence rate and successfully classifies motor imagery.

[1]  Geoffrey E. Hinton,et al.  Reducing the Dimensionality of Data with Neural Networks , 2006, Science.

[2]  Tapani Raiko,et al.  Enhanced Gradient and Adaptive Learning Rate for Training Restricted Boltzmann Machines , 2011, ICML.

[3]  Shen Zhang,et al.  Application of a common spatial pattern-based algorithm for an fNIRS-based motor imagery brain‐computer interface , 2017, Neuroscience Letters.

[4]  Wolfram Burgard,et al.  Deep learning with convolutional neural networks for brain mapping and decoding of movement-related information from the human EEG , 2017, ArXiv.

[5]  Geoffrey E. Hinton.  Training Products of Experts by Minimizing Contrastive Divergence , 2002, Neural Computation.

[6]  G. Pfurtscheller,et al.  Brain–Computer Communication: Motivation, Aim, and Impact of Exploring a Virtual Apartment , 2007, IEEE Transactions on Neural Systems and Rehabilitation Engineering.

[7]  Matthew D. Zeiler.  ADADELTA: An Adaptive Learning Rate Method , 2012, ArXiv.

[8]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[9]  LiKe,et al.  A multiwavelet-based time-varying model identification approach for time-frequency analysis of EEG signals , 2016 .

[10]  Lianghua He,et al.  A Deep Learning Method for Classification of EEG Data Based on Motor Imagery , 2014, ICIC.

[11]  Jian Yang,et al.  Sparse Deep Stacking Network for Image Classification , 2015, AAAI.

[12]  Bin He,et al.  EEG Source Imaging Enhances the Decoding of Complex Right-Hand Motor Imagery Tasks , 2016, IEEE Transactions on Biomedical Engineering.

[13]  Hui Zhang,et al.  A pairwise algorithm for pitch estimation and speech separation using deep stacking network , 2015, 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[14]  Stefan Haufe,et al.  Single-trial analysis and classification of ERP components — A tutorial , 2011, NeuroImage.

[15]  Noel Lopes,et al.  Improving Convergence of Restricted Boltzmann Machines via a Learning Adaptive Step Size , 2012, CIARP.

[16]  Shiliang Sun,et al.  Semi-supervised feature extraction for EEG classification , 2012, Pattern Analysis and Applications.

[17]  J. P. Kulasingham,et al.  Deep belief networks and stacked autoencoders for the P300 Guilty Knowledge Test , 2016, 2016 IEEE EMBS Conference on Biomedical Engineering and Sciences (IECBES).

[18]  Jing Zhang,et al.  Hyperspectral image classification based on deep stacking network , 2016, 2016 IEEE International Geoscience and Remote Sensing Symposium (IGARSS).

[19]  Licheng Jiao,et al.  Wishart Deep Stacking Network for Fast POLSAR Image Classification. , 2016, IEEE transactions on image processing : a publication of the IEEE Signal Processing Society.

[20]  Ugur Halici,et al.  A novel deep learning approach for classification of EEG motor imagery signals , 2017, Journal of neural engineering.

[21]  Heung-Il Suk,et al.  A Novel Bayesian Framework for Discriminative Feature Extraction in Brain-Computer Interfaces , 2013, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[22]  Liang Jin,et al.  Stable dynamic backpropagation learning in recurrent neural networks , 1999, IEEE Trans. Neural Networks.

[23]  Toshihisa Tanaka,et al.  Simultaneous Design of FIR Filter Banks and Spatial Patterns for EEG Signal Classification , 2013, IEEE Transactions on Biomedical Engineering.

[24]  Shiliang Sun,et al.  A review of adaptive feature extraction and classification methods for EEG-based brain-computer interfaces , 2014, 2014 International Joint Conference on Neural Networks (IJCNN).

[25]  Laurenz Wiskott,et al.  Gaussian-binary restricted Boltzmann machines for modeling natural image statistics , 2014, PloS one.

[26]  Shang-Lin Wu,et al.  Common spatial pattern and linear discriminant analysis for motor imagery classification , 2013, 2013 IEEE Symposium on Computational Intelligence, Cognitive Algorithms, Mind, and Brain (CCMB).

[27]  Ke Li,et al.  A multiwavelet-based time-varying model identification approach for time-frequency analysis of EEG signals , 2016, Neurocomputing.

[28]  Shiliang Sun,et al.  Extreme energy difference for feature extraction of EEG signals , 2010, Expert Syst. Appl..

[29]  Shiliang Sun,et al.  The Extreme Energy Ratio Criterion for EEG Feature Extraction , 2008, ICANN.

[30]  Shu Wang,et al.  Oscillation of a class of partial functional population model , 2010 .

[31]  Na Lu,et al.  A Deep Learning Scheme for Motor Imagery Classification based on Restricted Boltzmann Machines , 2017, IEEE Transactions on Neural Systems and Rehabilitation Engineering.