Kochunov P, Lancaster J, Thompson P, Toga A W, Brewer P, Hardies J, Fox P
Research Laboratory Center, University of Texas Health Science Center at San Antonio, 78284, USA.
Neuroimage. 2002 Oct;17(2):922-7.
The goal of regional spatial normalization is to remove anatomical differences between individual three-dimensional brain images by warping them to match features of a single target brain. Current target brains are either an average, suitable for low-resolution brain mapping studies, or a single brain. While a single high-resolution target brain is desirable to match anatomical detail, it can lead to bias in anatomical studies. An optimization method to reduce the individual anatomical bias of the ICBM high-resolution brain template (HRBT), a high-resolution MRI target brain image used in many laboratories, is presented. The HRBT was warped to all images in a group of 27 normal subjects. Displacement fields were averaged to calculate the "minimal deformation target" (MDT) transformation for optimization. The greatest anatomical changes in the HRBT, following optimization, were observed in the superior precentral and postcentral gyri on the right, the right inferior occipital, the right posterior temporal lobes, and the lateral ventricles. Compared with the original HRBT, the optimized HRBT showed better anatomical matching to the group of 27 brains. This was quantified by the improvements in spatial cross-correlation and between the group of brains and the optimized HRBT (P < 0.05). An intended use of this processing is to create a digital volumetric atlas that represents anatomy of a normal adult brain by optimizing the HRBT to the group consisting of 100+ normal subjects.
区域空间归一化的目标是通过对个体三维脑图像进行变形,使其与单个目标脑的特征相匹配,从而消除它们之间的解剖差异。当前的目标脑要么是适合低分辨率脑图谱研究的平均脑,要么是单个脑。虽然单个高分辨率目标脑有助于匹配解剖细节,但可能会导致解剖学研究中的偏差。本文提出了一种优化方法,以减少许多实验室使用的高分辨率MRI目标脑图像——国际神经成像基准联盟高分辨率脑模板(ICBM HRBT)的个体解剖偏差。将ICBM HRBT变形到一组27名正常受试者的所有图像上。对位移场进行平均,以计算用于优化的“最小变形目标”(MDT)变换。优化后,ICBM HRBT中最大的解剖学变化出现在右侧中央前回和中央后回、右侧枕下回、右侧颞叶后部以及侧脑室。与原始的ICBM HRBT相比,优化后的ICBM HRBT与27个脑组成的组在解剖学上匹配得更好。这通过空间互相关以及脑组与优化后的ICBM HRBT之间的改进进行了量化(P < 0.05)。这种处理的一个预期用途是通过将ICBM HRBT优化到由100多名正常受试者组成的组,创建一个代表正常成人大脑解剖结构的数字体积图谱。