高通量腦電波之訊號處理

On high-throughput EEG signal analysis

狄斐立(TILLET, Philippe)

指導教授 ： 林進燈

國立交通大學/電機學院/電機資訊國際學位學程/碩士(2014年)

https://doi.org/10.6842/NCTU.2014.00026

摘要

腦機接口的研究始於1970年代作為美國國家研究項目的一部分。此後，在信號處理的進展結合現代硬體，導致了這樣的系統，其應用範圍從視頻遊戲到生物醫療的應用更為豐富有趣。

這項工作的目的，是提出技術，增強腦電波之訊號處理的狀態。具體來說，我們建議設置一個現代化的硬體和軟體技術。我們透過一個新的實驗驗證了獨立成分分析（ICA），這比參考值更快、更準確的驗證了我們的想法

關鍵字

高性能計算 ； 最优化 ； 獨立成分分析(ICA) ； 腦電波之訊號處理

並列摘要

Research on Brain Computer Interfaces (BCIs) began in the 1970s as part of an American national project.
Since then, the advances in signal processing combined with the emergence of modern hardware led to a surge of interest in such systems, whose applications range from video games to biomedicals. The processing of such increasingly accurate signals has become a strenuous task, which requires to leverage the most recent algorithms on the most recent hardware.

This work seeks to propose techniques to update the current Brain-Computer Interfacing framework so as to take these new problems into account. Specifically, we suggest a set modern hardware and software techniques which could be used to enhance the state of EEG signal analysis. We validate our considerations through a novel implementation of the Independent Component Analysis problem, which largely outperforms the reference in terms of speed and accuracy, when separating dense EEG sources.

並列關鍵字

High-Performance Computing ； Unconstrained Optimization ； Independent Component Analysis (ICA) ； EEG Signal Analysis

參考文獻

[3] E. Agullo, C. Augonnet, J. Dongarra, H. Ltaief, R. Namyst, S. Thibault, and S. Tomov. A Hybridization Methodology for High-Performance Linear Algebra Software for GPUs. In GPU Computing Gems, volume 2, pages 473–484. Morgan Kaufmann, 2010.

[4] L. B. Almeida. Misep – linear and nonlinear ica based on mutual information. Journal of Machine Learning Research, 4:1297–1318, 2003.

[5] S.-I. Amari. Natural gradient works efficiently in learning. Neural Computation, 10(2):251– 276, Feb. 1998.

[6] A. J. Bell and T. J. Sejnowski. An information-maximization approach to blind separation and blind deconvolution. Neural Computation, 7:1129–1159, 1995.

[7] R. H. Byrd, G. M. Chin, W. Neveitt, and J. Nocedal. On the use of stochastic hessian information in optimization methods for machine learning. SIAM Journal on Optimization, 21(3):977–995, 2011.