Maddah Mahnaz, Kubicki Marek, Wells William M, Westin Carl-Fredrik, Shenton Martha E, Grimson W Eric L
Computer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology, Cambridge, MA, USA.
Med Image Comput Comput Assist Interv. 2008;11(Pt 1):917-24. doi: 10.1007/978-3-540-85988-8_109.
This paper presents a tract-oriented analysis of diffusion tensor (DT) images of the human brain. We demonstrate that unlike the commonly used ROI-based methods for population studies, our technique is sensitive to the local variation of diffusivity parameters along the fiber tracts. We show the strength of the proposed approach in identifying the differences in schizophrenic data compared to controls. Statistically significant drops in fractional anisotropy are observed along the genu and bilaterally in the splenium, as well as an increase in principal eigenvalue in uncinate fasciculus. This is the first tract-oriented clinical study in which an anatomical atlas is used to guide the algorithm.
本文提出了一种针对人脑扩散张量(DT)图像的基于纤维束的分析方法。我们证明,与群体研究中常用的基于感兴趣区域(ROI)的方法不同,我们的技术对沿纤维束的扩散率参数的局部变化敏感。我们展示了所提出方法在识别精神分裂症数据与对照组差异方面的优势。观察到在胼胝体膝部和双侧压部的各向异性分数有统计学意义的下降,以及钩束主特征值的增加。这是第一项使用解剖图谱指导算法的基于纤维束的临床研究。