新生児頭部MR画像を用いた脳形状正規化法  [in Japanese] Brain Shape Normalization in Neonatal Brain MR Images  [in Japanese]

Search this Article

Author(s)

Abstract

脳疾患を超早期に発見・治療を行うため,新生児に対する脳疾患診断法が求められている.従来,MRI(magnetic resonance imaging)を用いた脳疾患診断法の一つとして,VBM(voxel-based morphometry)が用いられる.VBMでは頭部解剖構造を標準脳に正規化することで,統計処理により異常部位を検出する.しかし,標準脳が成人脳由来であるため,正規化法の新生児脳への適用が困難となり,VBMを新生児脳に適用できない.本研究では,新生児脳の正規化法を3種提案する.第1の提案法では新生児の個人脳を標準脳とする.第2の提案法では第1の手法に頭蓋除去を前処理として適用する.第3の提案法では複数の新生児脳を基に標準脳を構築する.実験では,14例の新生児脳(修正齢4週から4週)に対して相互情報量と,前交連(AC)と後交連(PC)の位置と,ACとPCを結ぶ線(AC-PC line)の角度を用いて精度評価を行った.実験結果より,第3の提案法を用いることで,位置誤差が約1.7mm,角度誤差が約1.6°と,従来手法と比較し有意に高精度な結果が得られた.

In order to diagnose the cerebral diseases in the early stage, diagnosing method for neonatal brain must be studied. Voxel-based morphometry (VBM) is an approach to diagnose the cerebral diseases using MR (magnetic resonance) images. VBM normalizes the evaluating brain MR images into a template, and evaluates the distribution of gray matter density statistically. Because the template is produced from adult brain MR images, it is difficult to apply the conventional normalization to the neonatal brain. In this article, three new normalization approaches for the neonatal brain are proposed. The 1st approach defines an individual neonatal brain as a template. The 2nd approach introduces skull stripping into the 1st approach. The 3rd approach produces a neonatal template brain from some neonatal brains. Normalization accuracy was evaluated by using mutual information, position of anterior and posterior commissures (AC and PC) and angle of AC-PC line. The results in 14 neonates (revised age between -4 and 4 weeks) showed that the normalization accuracy of the 3rd approach was significantly higher than the conventional normalization and the other two approaches.

Journal

  • IEICE technical report

    IEICE technical report 110(364), 115-120, 2011-01-12

    The Institute of Electronics, Information and Communication Engineers

References:  12

Codes

  • NII Article ID (NAID)
    110008675413
  • NII NACSIS-CAT ID (NCID)
    AA11370335
  • Text Lang
    JPN
  • Article Type
    ART
  • ISSN
    09135685
  • NDL Article ID
    10970012
  • NDL Source Classification
    ZN33(科学技術--電気工学・電気機械工業--電子工学・電気通信)
  • NDL Call No.
    Z16-940
  • Data Source
    CJP  NDL  NII-ELS 
Page Top