G Model
ARTICLE IN PRESS
NSM-6861; No. of Pages 8
Journal of Neuroscience Methods xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Journal of Neuroscience Methods journal homepage: www.elsevier.com/locate/jneumeth
Computational Neuroscience
SSVEP recognition using common feature analysis in brain–computer interface Yu Zhang a,∗ , Guoxu Zhou b,1 , Jing Jin a,2 , Xingyu Wang a,2 , Andrzej Cichocki b,c,1 a Key Laboratory for Advanced Control and Optimization for Chemical Processes, Ministry of Education, East China University of Science and Technology, Shanghai 200237, China b Laboratory for Advanced Brain Signal Processing, RIKEN Brain Science Institute, Wako-shi, Saitama 351-0198, Japan c System Research Institute, Polish Academy of Sciences, Warsaw 00-901, Poland
h i g h l i g h t s • • • •
We propose a common feature analysis method to exploit the common features from EEG as references for SSVEP recognition. EEG data recorded from ten healthy subjects are used to validate effectiveness of the CFA method for SSVEP recognition. Experimental results indicate that CFA significantly outperformed CCA and MCCA methods in using a short time window. Our study confirms the proposed CFA method is promising for the development of a high-speed SSVEP-based BCI.
a r t i c l e
i n f o
Article history: Received 23 December 2013 Received in revised form 22 March 2014 Accepted 24 March 2014 Keywords: Brain–computer interface (BCI) Electroencephalogram (EEG) Canonical correlation analysis (CCA) Common feature analysis (CFA) Steady-state visual evoked potential (SSVEP)
a b s t r a c t Background: Canonical correlation analysis (CCA) has been successfully applied to steady-state visual evoked potential (SSVEP) recognition for brain–computer interface (BCI) application. Although the CCA method outperforms the traditional power spectral density analysis through multi-channel detection, it requires additionally pre-constructed reference signals of sine–cosine waves. It is likely to encounter overfitting in using a short time window since the reference signals include no features from training data. New method: We consider that a group of electroencephalogram (EEG) data trials recorded at a certain stimulus frequency on a same subject should share some common features that may bear the real SSVEP characteristics. This study therefore proposes a common feature analysis (CFA)-based method to exploit the latent common features as natural reference signals in using correlation analysis for SSVEP recognition. Results: Good performance of the CFA method for SSVEP recognition is validated with EEG data recorded from ten healthy subjects, in contrast to CCA and a multiway extension of CCA (MCCA). Comparison with existing methods: Experimental results indicate that the CFA method significantly outperformed the CCA and the MCCA methods for SSVEP recognition in using a short time window (i.e., less than 1 s). Conclusions: The superiority of the proposed CFA method suggests it is promising for the development of a real-time SSVEP-based BCI. © 2014 Elsevier B.V. All rights reserved.
1. Introduction
∗ Corresponding author. Tel.: +86 021 64253581. E-mail addresses: [email protected] (Y. Zhang), [email protected] (G. Zhou), [email protected] (J. Jin), [email protected] (X. Wang), [email protected] (A. Cichocki). 1 Tel.: +81 048 4679668. 2 Tel.: +86 021 64253581.
As an advanced communication system, brain–computer interface (BCI) provides a direction connection between human brain and computer, thereby assisting to re-establish communicative and environmental control abilities for people with severe motor disabilities (Wolpaw et al., 2002; Gao et al., 2003; Hoffmann et al., 2008). In the last couple of years, BCIs were mainly designed with steady-state visual evoked potential (SSVEP), event-related potential or sensorimotor rhythm recorded by electroencephalogram
http://dx.doi.org/10.1016/j.jneumeth.2014.03.012 0165-0270/© 2014 Elsevier B.V. All rights reserved.
Please cite this article in press as: Zhang Y, et al. SSVEP recognition using common feature analysis in brain–computer interface. J Neurosci Methods (2014), http://dx.doi.org/10.1016/j.jneumeth.2014.03.012
G Model NSM-6861; No. of Pages 8 2
ARTICLE IN PRESS Y. Zhang et al. / Journal of Neuroscience Methods xxx (2014) xxx–xxx
(EEG) (Wang et al., 2008; Allison et al., 2008; Zhang et al., 2012a, 2013a; Sellers and Donchin, 2006; Li et al., 2010; Pfurtscheller et al., 2006; Li and Zhang, 2010). SSVEP-based BCI has been increasingly studied since it requires less training to the user and usually provides relatively higher information transfer rate (ITR) (Allison et al., 2010; Guger et al., 2012). SSVEP is a periodic brain activity elicited at the same frequency as flicker frequency and also at its harmonics over occipital scalp region (see Fig. 1), when subject focuses attention on a flickering stimulus (Müller-Putz et al., 2005; Cheng et al., 2002). SSVEP-based BCI is developed to translate EEG signals recorded from the subject into computer commands through recognizing the SSVEP by typically using power spectral density analysis (PSDA) with fast Fourier transform (FFT) (Cheng et al., 2002; Wang et al., 2006; Müller-Putz et al., 2008; Hwang et al., 2012). However, the PSDA method is sensitive to noise with a single or bipolar channel, and requires relatively long time window to estimate the spectrum with sufficient frequency resolution. These drawbacks result in relatively low SSVEP recognition accuracy when the time window is not enough long (e.g.,
ARTICLE IN PRESS
NSM-6861; No. of Pages 8
Journal of Neuroscience Methods xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Journal of Neuroscience Methods journal homepage: www.elsevier.com/locate/jneumeth
Computational Neuroscience
SSVEP recognition using common feature analysis in brain–computer interface Yu Zhang a,∗ , Guoxu Zhou b,1 , Jing Jin a,2 , Xingyu Wang a,2 , Andrzej Cichocki b,c,1 a Key Laboratory for Advanced Control and Optimization for Chemical Processes, Ministry of Education, East China University of Science and Technology, Shanghai 200237, China b Laboratory for Advanced Brain Signal Processing, RIKEN Brain Science Institute, Wako-shi, Saitama 351-0198, Japan c System Research Institute, Polish Academy of Sciences, Warsaw 00-901, Poland
h i g h l i g h t s • • • •
We propose a common feature analysis method to exploit the common features from EEG as references for SSVEP recognition. EEG data recorded from ten healthy subjects are used to validate effectiveness of the CFA method for SSVEP recognition. Experimental results indicate that CFA significantly outperformed CCA and MCCA methods in using a short time window. Our study confirms the proposed CFA method is promising for the development of a high-speed SSVEP-based BCI.
a r t i c l e
i n f o
Article history: Received 23 December 2013 Received in revised form 22 March 2014 Accepted 24 March 2014 Keywords: Brain–computer interface (BCI) Electroencephalogram (EEG) Canonical correlation analysis (CCA) Common feature analysis (CFA) Steady-state visual evoked potential (SSVEP)
a b s t r a c t Background: Canonical correlation analysis (CCA) has been successfully applied to steady-state visual evoked potential (SSVEP) recognition for brain–computer interface (BCI) application. Although the CCA method outperforms the traditional power spectral density analysis through multi-channel detection, it requires additionally pre-constructed reference signals of sine–cosine waves. It is likely to encounter overfitting in using a short time window since the reference signals include no features from training data. New method: We consider that a group of electroencephalogram (EEG) data trials recorded at a certain stimulus frequency on a same subject should share some common features that may bear the real SSVEP characteristics. This study therefore proposes a common feature analysis (CFA)-based method to exploit the latent common features as natural reference signals in using correlation analysis for SSVEP recognition. Results: Good performance of the CFA method for SSVEP recognition is validated with EEG data recorded from ten healthy subjects, in contrast to CCA and a multiway extension of CCA (MCCA). Comparison with existing methods: Experimental results indicate that the CFA method significantly outperformed the CCA and the MCCA methods for SSVEP recognition in using a short time window (i.e., less than 1 s). Conclusions: The superiority of the proposed CFA method suggests it is promising for the development of a real-time SSVEP-based BCI. © 2014 Elsevier B.V. All rights reserved.
1. Introduction
∗ Corresponding author. Tel.: +86 021 64253581. E-mail addresses: [email protected] (Y. Zhang), [email protected] (G. Zhou), [email protected] (J. Jin), [email protected] (X. Wang), [email protected] (A. Cichocki). 1 Tel.: +81 048 4679668. 2 Tel.: +86 021 64253581.
As an advanced communication system, brain–computer interface (BCI) provides a direction connection between human brain and computer, thereby assisting to re-establish communicative and environmental control abilities for people with severe motor disabilities (Wolpaw et al., 2002; Gao et al., 2003; Hoffmann et al., 2008). In the last couple of years, BCIs were mainly designed with steady-state visual evoked potential (SSVEP), event-related potential or sensorimotor rhythm recorded by electroencephalogram
http://dx.doi.org/10.1016/j.jneumeth.2014.03.012 0165-0270/© 2014 Elsevier B.V. All rights reserved.
Please cite this article in press as: Zhang Y, et al. SSVEP recognition using common feature analysis in brain–computer interface. J Neurosci Methods (2014), http://dx.doi.org/10.1016/j.jneumeth.2014.03.012
G Model NSM-6861; No. of Pages 8 2
ARTICLE IN PRESS Y. Zhang et al. / Journal of Neuroscience Methods xxx (2014) xxx–xxx
(EEG) (Wang et al., 2008; Allison et al., 2008; Zhang et al., 2012a, 2013a; Sellers and Donchin, 2006; Li et al., 2010; Pfurtscheller et al., 2006; Li and Zhang, 2010). SSVEP-based BCI has been increasingly studied since it requires less training to the user and usually provides relatively higher information transfer rate (ITR) (Allison et al., 2010; Guger et al., 2012). SSVEP is a periodic brain activity elicited at the same frequency as flicker frequency and also at its harmonics over occipital scalp region (see Fig. 1), when subject focuses attention on a flickering stimulus (Müller-Putz et al., 2005; Cheng et al., 2002). SSVEP-based BCI is developed to translate EEG signals recorded from the subject into computer commands through recognizing the SSVEP by typically using power spectral density analysis (PSDA) with fast Fourier transform (FFT) (Cheng et al., 2002; Wang et al., 2006; Müller-Putz et al., 2008; Hwang et al., 2012). However, the PSDA method is sensitive to noise with a single or bipolar channel, and requires relatively long time window to estimate the spectrum with sufficient frequency resolution. These drawbacks result in relatively low SSVEP recognition accuracy when the time window is not enough long (e.g.,
