面向婴幼儿大脑动态发育分析的精细脑图谱研究

脑图谱是神经科学研究必的关键工具之一，但在针对婴幼儿的脑发育研究中，尤其是在时空异质发育的影响下，基于成人的低分辨单一脑图谱难以满足实际分析需求。本项目面向针对婴幼儿的精细、动态脑图谱的重大缺失，从个体配准、边界增强、弹性边界的图谱绘制等角度对婴幼儿脑图谱展开系统的研究。研究个体差异造成脑区边界模糊的问题，提出新的个体功能边界的精细表征，构建高效减小个体差异的多模态纵向配准框架；针对部分细小脑区边界微弱的问题，提出通过滤波的方式增强线状特征并融合多模态梯度以增强脑区边界，有效提高脑图谱精度；研究以全局边界为引导，各年龄边界为约束的基于弹性边界的动态脑图谱构建方法；最终绘制一组高度契合婴幼儿大脑发育时空异质性，符合大脑动态发育轨迹的婴幼儿脑图谱，从而推进发育神经科学研究，为神经发育型疾病早期、自动诊断提供支持。

Brain parcellation maps are essential tools for neuroscientific research. However, due to the heterogeneity and heterochroneity of infantile brain development, the low-resolution and static adult brain parcellations are not applicable for infant developmental research. In consideration of the urgent need for high-resolution and dynamic infant brain parcellations, this project conducts a series of research from the aspects of cortical registration, edge enhancement, and dynamic cortical parcellation methods. We intend to cope with the edge blurriness due to large inter-subject variation by proposing a new functional edge representation with fine details, and accordingly a longitudinal multi-modal registration framework. To further enhance those weak edges between tiny brain areas, we propose to use a filtering method and also take into account multi-modal boundaries, so as to effectively improve the parcellation resolution. To build brain parcellations with adaptive dynamic edges at different ages while having area correspondences among all ages, we design a framework that integrates age-common parcellation as guidance, and age-specific edges as constraint, so that to build dynamic parcellation maps based on deformable edges. Finally, a set of brain parcellations will be developed, which integrate the heterogeneity and heterochroneity of brain development trajectories. This set of brain parcellation will be essential tools for developmental neuroscientific studies and will provide support for the early diagnosis of brain developmental disorders.