Hierarchical statistical modeling for multivariate signal processing and its applications

2015 Fiscal Year Final Research Report

• Project/Area Number: 25730155
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Soft computing
• Research Institution: Advanced Telecommunications Research Institute International
• Principal Investigator: Hirayama Jun-ichiro 株式会社国際電気通信基礎技術研究所, 脳情報通信総合研究所, 専任研究員 (80512269)
• Research Collaborator: Hyvärinen Aapo ヘルシンキ大学 ; Kiviniemi Vesa オウル大学病院 ; OGAWA Takeshi 株式会社国際電気通信基礎技術研究所(ATR) ; KAWANABE Motoaki 株式会社国際電気通信基礎技術研究所(ATR) ; YAMASHITA Okito 株式会社国際電気通信基礎技術研究所(ATR) ; ISHII Shin 京都大学
• Project Period (FY): 2013-04-01 – 2016-03-31
• Keywords: 機械学習 / 生体信号処理 / 脳機能イメージング / ブレイン・マシンインターフェイス

Outline of Final Research Achievements

We developed novel multivariate data analysis and signal processing methods to find and extract characteristic patterns from data, focusing on applications to the analysis of functional brain activity measurements. Our methods seek to find basis patterns (coactivation patterns, modules) of brain activity, thereby characterizing the global variability of functional brain connectivity, in a unified manner by explicitly modeling hierarchical structures underlying data. We evaluated each method with simulations, electro/magnetoencephalography and functional MRI data, which demonstrated the advantages over existing methods as well as applicability to functional brain imaging data analysis. Future applications to cognitive neuroscience or neural engineering can be expected.

Free Research Field

機械学習