Common and Individual Feature Analysis and Its Applications in BCI

• Project/Area Number: 26730125
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Intelligent informatics
• Research Institution: Institute of Physical and Chemical Research
• Principal Investigator: ZHOU GUOXU 国立研究開発法人理化学研究所, 脳科学総合研究センター, 研究員 (80608034)
• Project Period (FY): 2014-04-01 – 2016-03-31
• Project Status: Completed (Fiscal Year 2015)
• Budget Amount: ¥3,640,000 (Direct Cost: ¥2,800,000、Indirect Cost: ¥840,000); Fiscal Year 2015: ¥1,560,000 (Direct Cost: ¥1,200,000、Indirect Cost: ¥360,000); Fiscal Year 2014: ¥2,080,000 (Direct Cost: ¥1,600,000、Indirect Cost: ¥480,000)
• Keywords: データ統合 / 共通・個別特徴解析 / 結合成分解析 / テンソル分解

Outline of Final Research Achievements

(1) We proposed a new framework for the common and individual features analysis (CIFA) of multi-block linked data. (2) We developed several efficient algorithms for tensor decompositions, which can be used to extract common features from multi-block high-dimensional data where these common features are dominant. (3) We built a steady-state visual evoked potential (SSVEP) based BCI system and recorded EEG signals of multiple subjects and trials which were under same stimuli. These data were analyzed by CIFA, and we achieved so-far the highest accuracy in target frequency recognition.

Research Products

• [Journal Article] Linked Component Analysis from Matrices to High Order Tensors: Applications to Biomedical Data (2016)
  • Author(s): G Zhou, Q Zhao, Y Zhang, T Adali;, S Xie, A Cichocki
  • Journal Title: Proceedings of the IEEE Volume: 104 Issue: 2 Pages: 310-331
  • DOI: 10.1109/jproc.2015.2474704
  • Peer Reviewed / Open Access / Int'l Joint Research / Acknowledgement Compliant
• [Journal Article] Group Component Analysis for Multiblock Data: Common and Individual Feature Extraction (2016)
  • Author(s): G Zhou, A Cichocki, Y Zhang, D Mandic
  • Journal Title: IEEE Transactions on Neural Networks and Learning Systems Volume: 印刷中 Issue: 11 Pages: 2426-2439
  • DOI: 10.1109/tnnls.2015.2487364
  • Peer Reviewed / Int'l Joint Research / Acknowledgement Compliant
• [Journal Article] Sparse Bayesian Classification of EEG for Brain-Computer Interface (2016)
  • Author(s): Y Zhang, G Zhou, J Jin, Q Zhao, X Wang, A Cichocki
  • Journal Title: IEEE Transactions on Neural Networks and Learning Systems Volume: 印刷中 Issue: 11 Pages: 2256-2267
  • DOI: 10.1109/tnnls.2015.2476656
  • Peer Reviewed / Int'l Joint Research / Acknowledgement Compliant
• [Journal Article] Bayesian robust tensor factorization for incomplete multiway data. (2016)
  • Author(s): Qibin Zhao, Guoxu Zhou, Liqing Zhang, and Andrzej Cichocki
  • Journal Title: IEEE Transactions on Neural Networks and Learning Systems Volume: 27 Issue: 4 Pages: 736-748
  • DOI: 10.1109/tnnls.2015.2423694
  • Peer Reviewed / Open Access / Int'l Joint Research / Acknowledgement Compliant
• [Journal Article] Fast nonnegative tensor factorization based on accelerated proximal gradient and low-rank approximation (2016)
  • Author(s): Y Zhang, G Zhou, Q Zhao, A Cichocki, X Wang
  • Journal Title: Neurocomputing Volume: 印刷中 Pages: 148-154
  • DOI: 10.1016/j.neucom.2015.08.122
  • Peer Reviewed / Int'l Joint Research
• [Journal Article] Optimizing spatial patterns with sparse filter bands for motor-imagery based brain–computer interface (2015)
  • Author(s): Y Zhang, G Zhou, J Jin, X Wang, A Cichocki
  • Journal Title: Journal of neuroscience methods Volume: 255 Pages: 85-91
  • DOI: 10.1016/j.jneumeth.2015.08.004
  • Peer Reviewed / Int'l Joint Research / Acknowledgement Compliant
• [Journal Article] Efficient Nonnegative Tucker Decompositions: Algorithms and Uniqueness (2015)
  • Author(s): Guoxu Zhou, Andrzej Cichocki, Qibin Zhao and Shengli Xie
  • Journal Title: IEEE Transactions on Image Processing Volume: 24 Issue: 12 Pages: 4990-5003
  • DOI: 10.1109/tip.2015.2478396
  • Peer Reviewed / Int'l Joint Research / Acknowledgement Compliant
• [Journal Article] Two Efficient Algorithms for Approximately Orthogonal Nonnegative Matrix Factorization (2015)
  • Author(s): Bo Li, Guoxu Zhou, and Andrzej Cichocki
  • Journal Title: IEEE Signal Processing Letters Volume: 22 Issue: 7 Pages: 843-846
  • DOI: 10.1109/lsp.2014.2371895
  • Peer Reviewed / Acknowledgement Compliant
• [Journal Article] SSVEP recognition using common feature analysis in brain--computer interface (2015)
  • Author(s): Yu Zhang, Guoxu Zho, Jing Jin, Xingyu Wang, and Andrzej Cichocki
  • Journal Title: Journal of Neuroscience Methods Volume: 244 Pages: 8-15
  • DOI: 10.1016/j.jneumeth.2014.03.012
  • Peer Reviewed
• [Journal Article] Robust Bayesian Tensor Factorization for Incomplete Multiway Data (2015)
  • Author(s): Qibin Zhao, Guoxu Zhou, Liqing Zhang, Andrzej Cichocki, Shun-ichi Amari
  • Journal Title: IEEE Transactions on Neural Networks and Learning Systems Volume: PP
  • Peer Reviewed / Acknowledgement Compliant
• [Presentation] Common Components Analysis via Linked Blind Source Separation (2015)
  • Author(s): Guoxu Zhou, Andrzej Cichocki, Danilo P. Mandic
  • Organizer: The 40th IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 2015
  • Place of Presentation: Brisbane, Australia
  • Year and Date: 2015-04-19 – 2015-04-25
• [Presentation] Common Components Analysis via Linked Blind Source Separation (2015)
  • Author(s): G. Zhou, A. Cichocki, and D. Mandic
  • Organizer: The 40th IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 2016
  • Place of Presentation: Brisbane, Australia
  • Year and Date: 2015-04-19
  • Int'l Joint Research