機能的近赤外分光法を用いたブレイン・コンピュータ・インタフェースに対する頭皮血流除去の効果

カテゴリ: 論文誌(論文単位)

グループ名: 【C】電子・情報・システム部門

発行日: 2017/05/01

タイトル(英語): Effectiveness of Scalp-hemodynamics Reduction to Brain-computer Interfaces by Functional Near-infrared Spectroscopy

著者名: 佐藤　貴紀（長岡技術科学大学），南部　功夫（長岡技術科学大学），和田　安弘（長岡技術科学大学）

著者名(英語): Takanori Sato (Nagaoka University of Technology), Isao Nambu (Nagaoka University of Technology), Yasuhiro Wada (Nagaoka University of Technology)

キーワード: ブレイン・コンピュータ・インタフェース，機能的近赤外分光法，頭皮血流，複数距離プローブ配置，一般線形モデル，多値分類　　Brain-computer interfaces，Functional near-infrared spectroscopy，Scalp blood flow，Multidistance probe arrangement，General linear model，Multi-class classification

要約(英語): Brain-computer interfaces (BCIs) are systems that control external devices by decoding information from brain activity signals. Functional near-infrared spectroscopy (fNIRS) has been used in many BCIs because of its simplicity of use and portability. However, hemodynamic changes in the scalp layer (scalp-hemodynamics) often contaminate fNIRS signals, and cause degradation of the detection accuracy of functional brain activities. Although several reduction methods have been proposed, no study has investigated their effects on fNIRS-BCI accuracy. In this study, we investigated the effects of applying scalp-hemodynamics reduction to the classification of for four tasks: ball grasping with left-, right-, or both-hands, or resting without movements. We applied a method that combined short source-detector distance channels with a general linear model. Results showed that the binary-class classification accuracy of left- or right-hand and the multi-class classification accuracy of 3-class grasping were significantly improved, suggesting that the scalp-hemodynamics reduction may provide more accurate fNIRS-BCIs.

本誌: 電気学会論文誌C（電子・情報・システム部門誌） Vol.137 No.5 （2017）

本誌掲載ページ: 717-723 p

原稿種別: 論文／日本語

電子版へのリンク: https://www.jstage.jst.go.jp/article/ieejeiss/137/5/137_717/_article/-char/ja/