Microscopic diffusion anisotropy in the human brain: Reproducibility, normal values, and comparison with the fractional anisotropy
Microscopic diffusion anisotropy in the human brain: Reproducibility, normal values, and comparison with the fractional anisotropy
复制标题
DOI：
10.1016/j.neuroimage.2015.01.025
复制
发表时间：
2015-04-01
期刊：
影响因子：
5.7
通讯作者：
Finsterbusch, Juergen
中科院分区：
文献类型：
--
作者：
Lawrenz, Marco;Brassen, Stefanie;Finsterbusch, Juergen
人类对组织微结构的神经成像，如轴突密度和完整性，在临床和神经科学研究中是关键。大多数研究依赖于扩散张量成像(DTI)及其衍生的测量，最突出的是分数各向异性(FA)。然而，FA还取决于纤维取向分布，这是一种更宏观的组织特性。最近引入的所谓微观扩散各向异性的测量方法，即细胞或微观水平上的扩散各向异性，克服了这一限制，因为它们与轴突和纤维的取向分布无关。在这项研究中，我们评估了两个衡量微观扩散各向异性的I-MA和MA指数，用于人类神经科学和临床研究的可行性。这两个指数都反映了细胞的偏心率，但虽然I-MA也取决于细胞大小，但MA与细胞大小无关，与FA一样，在0到1之间变化。在一组19名健康志愿者的全脑测量中，我们测量了平均值和变异性，评估了它们在会话内和会话之间的重复性，并比较了MA和FA在选定感兴趣区(ROI)的值。会话内和会话间的比较没有显示出实质性的差异，但MA的重复性比I-MA要好得多(会话间的变异系数为10.5%对28.9%)。总体而言，MA的重复性低于FA，但在定义的ROI中具有可比性，组间比较所需的平均组大小相似(约60名参与者，相对差异为5%)。MA指数的组平均值一般大于FA(7个ROI的平均+/-标准差为0.83+/-0.10对0.58+/-0.13)，且在白质脑ROI之间的差异较小。即使在一些灰质感兴趣区，也可以观察到与白质感兴趣区相当的MA值。此外，与FA组相比，MA组白质ROI值的组内差异较小(志愿者的平均标准差为0.038比0.049)，这可能是由于影响FA的纤维取向分布存在显著差异。这些结果表明，MA(I)应优先于I-MA，(Ii)具有与FA相当的重复性和群体大小要求；(Iii)对纤维取向分布的敏感性低于FA；以及(Iv)对组织微结构的差异或变化可能比FA更敏感。R1.1(C)2015 Elsevier Inc.保留所有权利。
Human neuroimaging of tissue microstructure, such as axonal density and integrity, is key in clinical and neuroscience research. Most studies rely on diffusion tensor imaging (DTI) and the measures derived from it, most prominently fractional anisotropy (FA). However, FA also depends on fiber orientation distribution, a more macroscopic tissue property. Recently introduced measures of so-called microscopic diffusion anisotropy, diffusion anisotropy on a cellular or microscopic level, overcome this limitation because they are independent of the orientation distributions of axons and fibers. In this study, we evaluate the feasibility of two measures of microscopic diffusion anisotropy I-MA and MA indices, for human neuroscience and clinical research. Both indices reflect the eccentricity of the cells but while I-MA also depends on the cell size, MA is independent of the cell size and, like FA, scaled between 0 and 1. In whole-brain measurements of a group of 19 healthy volunteers, we measured average values and variability, evaluated their reproducibility, both within and between sessions, and compared MA to FA values in selected regions-of-interest (ROIs). The within-and between-session comparison did not show substantial differences but the reproducibility was much better for the MA than I-MA (coefficient of variation between sessions 10.5% vs. 28.9%). The reproducibility was less for MA than FA overall, but comparable in the defined ROIs and the average group sizes required for between-group comparisons was similar (about 60 participants for a relative difference of 5%). Group-averaged values of MA index were generally larger and showed less variation across white-matter brain ROIs than FA (mean +/- standard deviation of seven ROIs 0.83 +/- 0.10 vs. 0.58 +/- 0.13). Even in some gray-matter ROIs, MA values comparable to those of white matter ROIs were observed. Furthermore, the within-group variation of the values in white matter ROIs was lower for the MA compared to the FA (mean standard deviation over volunteers 0.038 vs. 0.049) which could be due to significant variability in the distribution of fiber orientation contributing to FA. These results indicate that MA (i) should be preferred to I-MA, (ii) has a reproducibility and group-size requirements comparable to those of FA; (iii) is less sensitive to the fiber orientation distribution than FA; and (iv) could be more sensitive to differences or changes of the tissue microstructure than FA. R1.1 (C) 2015 Elsevier Inc. All rights reserved.