A Method of Expanding EEG Data Based on Transfer Learning Theory

A brain-computer interface (BCI) is a technology that provides a direct communication channel between the human brain and the external world. Its performance is usually measured based on the classification accuracy of electroencephalography (EEG) signals. Recently, a convolutional neural network (CNN) is used to classify EEG signals that achieve a high classification accuracy, but the lack of tagged data limits the development of this approach. This paper proposes a method based on transfer learning theory to prove the accuracy of the data collected and achieve the purpose of data expansion. To prove that the subject has the right motor imagination, a fast and simple CNN with two-dimensional energy maps for each electrode is proposed, which serves as the input to classify EEG data, and then the gathered data and the contest data are compared using the same training set.

[1]  Klaus-Robert Müller,et al.  The non-invasive Berlin Brain–Computer Interface: Fast acquisition of effective performance in untrained subjects , 2007, NeuroImage.

[2]  M. Çiçek,et al.  Interhemispheric asymmetry of EEG alpha activity at rest and during the Wisconsin Card Sorting Test: relations with performance , 2001, Biological Psychology.

[3]  S. Jarcho.  Equal-Area Projections and the Azimuthal Equidistant Projection in Maps of Disease. , 1945, American journal of public health and the nation's health.

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

[5]  Ivan Laptev,et al.  Learning and Transferring Mid-level Image Representations Using Convolutional Neural Networks , 2014, 2014 IEEE Conference on Computer Vision and Pattern Recognition.

[6]  L. Euginia,et al.  A wavelet based de-noising technique for ocular artifact correction of the electroencephalogram , 2004, Proceedings. The 8th Russian-Korean International Symposium on Science and Technology, 2004. KORUS 2004..

[7]  L. Shoker,et al.  Removal of eye blinking artifacts from EEG incorporating a new constrained BSS algorithm , 2004, Processing Workshop Proceedings, 2004 Sensor Array and Multichannel Signal.

[8]  Tzyy-Ping Jung,et al.  Independent Component Analysis of Electroencephalographic Data , 1995, NIPS.

[9]  Yijun Wang,et al.  Visual and Auditory Brain–Computer Interfaces , 2014, IEEE Transactions on Biomedical Engineering.

[10]  Peter Alfeld,et al.  A trivariate clough-tocher scheme for tetrahedral data , 1984, Comput. Aided Geom. Des..

[11]  Cuntai Guan,et al.  Filter Bank Common Spatial Pattern (FBCSP) in Brain-Computer Interface , 2008, 2008 IEEE International Joint Conference on Neural Networks (IEEE World Congress on Computational Intelligence).