Institute of Biomedical Engineering, National Taiwan University, Taipei, Taiwan.
Neuroimage. 2010 Jan 1;49(1):32-43. doi: 10.1016/j.neuroimage.2009.08.031. Epub 2009 Aug 21.
Diffusion tensor imaging (DTI) has already proven to be a valuable tool when investigating both global and regional microstructural white matter (WM) brain changes in the human aging process. Although subject to many criticisms, voxel-based analysis is currently one of the most common and preferred approaches in such DTI aging studies. In this context, voxel-based DTI analyses have assumed a 'linear' correlation when finding the significant brain regions that relate age with a particular diffusion measure of interest. Recent literature, however, has clearly demonstrated 'non-linear' relationships between age and diffusion metrics by using region-of-interest and tractography-based approaches. In this work, we incorporated polynomial regression models in the voxel-based DTI analysis framework to assess age-related changes in WM diffusion properties (fractional anisotropy and axial, transverse, and mean diffusivity) in a large cohort of 346 subjects (25 to 81 years old). Our novel approach clearly demonstrates that voxel-based DTI analyses can greatly benefit from incorporating higher-order regression models when investigating potential relationships between aging and diffusion properties.
弥散张量成像（DTI）已经被证明是一种有价值的工具，可用于研究人类衰老过程中整体和局部脑白质（WM）微观结构的变化。尽管受到许多批评，体素基分析仍然是此类 DTI 衰老研究中最常用和首选的方法之一。在这种情况下，体素基 DTI 分析在发现与特定扩散测量值相关的与年龄相关的显著大脑区域时，假设存在“线性”相关性。然而，最近的文献通过使用基于感兴趣区域和轨迹的方法清楚地证明了年龄与扩散指标之间存在“非线性”关系。在这项工作中，我们在体素基 DTI 分析框架中纳入了多项式回归模型，以评估 346 名受试者（25 至 81 岁）的 WM 扩散特性（各向异性分数、轴向、横向和平均扩散系数）与年龄相关的变化。我们的新方法清楚地表明，当研究衰老与扩散特性之间的潜在关系时，体素基 DTI 分析可以从纳入更高阶回归模型中受益匪浅。
