Electrical Engineering and Computer Science, University of California at Berkeley, CA, USA.
Department of Radiology, Ruijin Hospital, Shanghai Jiaotong University School of Medicine, Shanghai, China.
Neuroimage. 2018 May 1;171:176-189. doi: 10.1016/j.neuroimage.2018.01.008. Epub 2018 Jan 8.
Longitudinal brain atlases play an important role in the study of human brain development and cognition. Existing atlases are mainly based on anatomical features derived from T1-and T2-weighted MRI. A 4D developmental quantitative susceptibility mapping (QSM) atlas may facilitate the estimation of age-related iron changes in deep gray matter nuclei and myelin changes in white matter. To this end, group-wise co-registered QSM templates were generated over various age intervals from age 1-83 years old. Registration was achieved by combining both T1-weighted and QSM images. Based on the proposed template, we created an accurate deep gray matter nuclei parcellation map (DGM map). Notably, we segmented thalamus into 5 sub-regions, i.e. the anterior nuclei, the median nuclei, the lateral nuclei, the pulvinar and the internal medullary lamina. Furthermore, we built a "whole brain QSM parcellation map" by combining existing cortical parcellation and white-matter atlases with the proposed DGM map. Based on the proposed QSM atlas, the segmentation accuracy of iron-rich nuclei using QSM is significantly improved, especially for children and adolescent subjects. The age-related progression of magnetic susceptibility in each of the deep gray matter nuclei, the hippocampus, and the amygdala was estimated. Our automated atlas-based analysis provided a systematic confirmation of previous findings on susceptibility progression with age resulting from manual ROI drawings in deep gray matter nuclei. The susceptibility development in the hippocampus and the amygdala follow an iron accumulation model; while in the thalamus sub-regions, the susceptibility development exhibits a variety of trends. It is envisioned that the newly developed 4D QSM atlas will serve as a template for studying brain iron deposition and myelination/demyelination in both normal aging and various brain diseases.
纵向脑图谱在研究人类大脑发育和认知方面发挥着重要作用。现有的图谱主要基于 T1 和 T2 加权 MRI 得出的解剖学特征。一个 4D 发育定量磁化率映射 (QSM) 图谱可能有助于估计深部灰质核中铁的年龄相关性变化和白质髓鞘的变化。为此，在 1-83 岁的各个年龄间隔内生成了基于群组的配准 QSM 模板。通过结合 T1 加权和 QSM 图像实现了配准。基于所提出的模板，我们创建了一个准确的深部灰质核分割图 (DGM 图)。值得注意的是，我们将丘脑分为 5 个亚区，即前核、中核、外侧核、丘脑枕和内髓板。此外，我们通过将现有的皮质分割和白质图谱与所提出的 DGM 图谱相结合，构建了一个“全脑 QSM 分割图谱”。基于所提出的 QSM 图谱，使用 QSM 对富含铁的核的分割准确性得到了显著提高，特别是对于儿童和青少年受试者。估计了每个深部灰质核、海马体和杏仁核的磁化率随年龄的变化。我们的自动化图谱分析系统地证实了先前关于深部灰质核中基于手动 ROI 绘图的磁化率随年龄变化的发现。海马体和杏仁核的磁化率发展遵循铁积累模型；而在丘脑亚区中，磁化率发展呈现出多种趋势。预计新开发的 4D QSM 图谱将作为研究正常衰老和各种脑部疾病中脑铁沉积和髓鞘化/脱髓鞘的模板。
