Ghosh Nirmalya, Holshouser Barbara, Oyoyo Udo, Barnes Stanley, Tong Karen, Ashwal Stephen
Department of Pediatrics, Loma Linda University School of Medicine, Loma Linda, CA, USA.
Dev Neurosci. 2017;39(5):413-429. doi: 10.1159/000475545. Epub 2017 Jun 27.
During human brain development, anatomic regions mature at different rates. Quantitative anatomy-specific analysis of longitudinal diffusion tensor imaging (DTI) and magnetic resonance spectroscopic imaging (MRSI) data may improve our ability to quantify and categorize these maturational changes. Computational tools designed to quickly fuse and analyze imaging information from multiple, technically different datasets would facilitate research on changes during normal brain maturation and for comparison to disease states. In the current study, we developed a complete battery of computational tools to execute such data analyses that include data preprocessing, tract-based statistical analysis from DTI data, automated brain anatomy parsing from T1-weighted MR images, assignment of metabolite information from MRSI data, and co-alignment of these multimodality data streams for reporting of region-specific indices. We present statistical analyses of regional DTI and MRSI data in a cohort of normal pediatric subjects (n = 72; age range: 5-18 years; mean 12.7 ± 3.3 years) to establish normative data and evaluate maturational trends. Several regions showed significant maturational changes for several DTI parameters and MRSI ratios, but the percent change over the age range tended to be small. In the subcortical region (combined basal ganglia [BG], thalami [TH], and corpus callosum [CC]), the largest combined percent change was a 10% increase in fractional anisotropy (FA) primarily due to increases in the BG (12.7%) and TH (9%). The largest significant percent increase in N-acetylaspartate (NAA)/creatine (Cr) ratio was seen in the brain stem (BS) (18.8%) followed by the subcortical regions in the BG (11.9%), CC (8.9%), and TH (6.0%). We found consistent, significant (p < 0.01), but weakly positive correlations (r = 0.228-0.329) between NAA/Cr ratios and mean FA in the BS, BG, and CC regions. Age- and region-specific normative MR diffusion and spectroscopic metabolite ranges show brain maturation changes and are requisite for detecting abnormalities in an injured or diseased population.
在人类大脑发育过程中，不同的解剖区域以不同的速度成熟。对纵向扩散张量成像（DTI）和磁共振波谱成像（MRSI）数据进行特定解剖区域的定量分析，可能会提高我们量化和分类这些成熟变化的能力。设计用于快速融合和分析来自多个技术上不同数据集的成像信息的计算工具，将有助于研究正常大脑成熟过程中的变化，并与疾病状态进行比较。在本研究中，我们开发了一套完整的计算工具来执行此类数据分析，包括数据预处理、基于DTI数据的基于纤维束的统计分析、从T1加权磁共振图像自动解析脑解剖结构、从MRSI数据分配代谢物信息，以及对这些多模态数据流进行共配准以报告区域特异性指标。我们对一组正常儿科受试者（n = 72；年龄范围：5 - 18岁；平均12.7 ± 3.3岁）的区域DTI和MRSI数据进行了统计分析，以建立标准数据并评估成熟趋势。几个区域的几个DTI参数和MRSI比率显示出显著的成熟变化，但在整个年龄范围内的百分比变化往往较小。在皮质下区域（基底神经节[BG]、丘脑[TH]和胼胝体[CC]的组合），最大的组合百分比变化是各向异性分数（FA）增加10%，主要是由于BG（12.7%）和TH（9%）的增加。N - 乙酰天门冬氨酸（NAA）/肌酸（Cr）比率最大的显著百分比增加出现在脑干（BS）（18.8%），其次是BG（11.9%）、CC（8.9%）和TH（6.0%）的皮质下区域。我们发现脑干、BG和CC区域的NAA/Cr比率与平均FA之间存在一致、显著（p < 0.01）但较弱的正相关（r = 0.228 - 0.329）。年龄和区域特异性的标准磁共振扩散和波谱代谢物范围显示了大脑成熟变化，并且是检测受伤或患病群体异常情况所必需的。
