An innovation approach for epileptic fMRI data analysis based on spatio-temporal filtering method

2011 Fiscal Year Final Research Report

• Project/Area Number: 21500295
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Single-year Grants
• Section: 一般
• Research Field: Bioinformatics/Life informatics
• Research Institution: The Institute of Statistical Mathematics
• Principal Investigator: MIWAKEICHI Fumikazu 統計数理研究所, モデリング研究系, 准教授 (30360647)
• Co-Investigator(Kenkyū-buntansha): WATANABE Job 平成帝京大学, 薬学部, 准教授 (90409756)
• Project Period (FY): 2009 – 2011
• Keywords: 時空間データ解析 / 異常値検出 / てんかん

Research Abstract

For the diagnosis of epilepsy, concurrent EEG and fMRI recording has been studied to detect epileptic spike train related brain blood flow. The standard method of data analysis is based on regression analysis and it employs convoluted epileptic spike trains on EEG data with canonical HRF as a reference function. It enables to graphically show epileptic blood flows in three dimensional space and detect the location of epileptic foci. Moreover it is possible to estimate some interaction or network system between foci. This achievement is very helpful for medication or surgical treatment objectives.
However, there are some points in the data analysis to be improved. In the standard analysis, a canonical HRF in common shape is used for all regions in the brain and for all subjects. Therefore only the activation, whose temporal changing is morphologically similar to the reference function, can be detected. In this study, we introduce an innovation approach based on auto regressive(AR) model to detect dynamical difference during seizure comparing to seizure free epoch without employing reference function. Moreover we developed an algorithm to ecstatically evaluate the significance of abnormal blood flow and map on the anatomical brain image.

Research Products

• [Journal Article] 単一試行生体イメージングデータ解析のためのAR型モデルに基づく時空間フィルタリング法 (2012)
  • Author(s): 三分一史和,越久仁敬,岡田泰昌,川合成治,田村義保,石黒真木夫
  • Journal Title: 統計数理 Volume: (in press)
• [Journal Article] Detection and Visualization Method of Dynamic State Transition for Biological Spatio-Temporal Imaging Data (2011)
  • Author(s): F. Miwakeichi, Y. Oku, Y. Okada, S. Kawai, Y. Tamura, M. Ishiguro
  • Volume: 30(3)
  • Pages: 859-66
• [Journal Article] Spatio-temporal correlations from fMRI time series based on the NN-ARx model (2010)
  • Author(s): J. Bosch-Bayard, J. Riera-Diaz, R. Biscay-Lirio, K. Wong, A. Galka, O. Yamashita, N. Sadato, R. Kawashima, E. Aubert-Vazquez, R. Rodriguez-Rojas, P. Valdes-Sosa, F. Miwakeichi and T. Ozaki
  • Volume: 9(4)
  • Pages: 381-406
• [Presentation] 時空間統計学の理論と経済・脳信号データ分析への応用 (2011)
  • Author(s): 三分一史和
  • Organizer: 統計関連学会連合大会チュートリアルセッション
  • Place of Presentation: アクロス福岡
  • Year and Date: 2011-09-04
• [Presentation] A Spatio-Temporal Analysis for Detecting Neural Activation in the Imaging Data (2011)
  • Author(s): F. Miwakeichi, Y. Oku, Y. Okada, S. Kawai, Y. Tamura, M. Ishiguro
  • Organizer: The 58th World Statistics Congress
  • Place of Presentation: The Convention Centre
  • Year and Date: 2011-08-26
• [Presentation] An innovation approach for epileptic fMRI data analysis based on statistical time series modeling (2010)
  • Author(s): F. Miwakeichi, A. Galka, F. Moeller, M. Siniatchkin
  • Organizer: 6th Annual Meeting of the Organization for Human Brain Mapping, Pala u de Congressos de Catalunya
  • Place of Presentation: Barcelona, Spain
  • Year and Date: 20100607-08
• [Presentation] 脳・神経データの時空間解析 (2010)
  • Author(s): 三分一史和
  • Organizer: 神経科学と統計科学の対話
  • Place of Presentation: 統計数理研究所
  • Year and Date: 2010-12-19
• [Remarks]
  • URL: http://www.ism.ac.jp/souran/kenkyusya/miwakeichi_fumikazu.html