毕业生知识库

  青少年时期是人类身体和心理快速发育的关键阶段，也是多种精神障碍的高发时期。青少年精神心理发育异常会严重破坏个体未来的发展潜力，并为家庭和社会带来额外的沉重负担。然而，现有针对青少年精神障碍的医疗手段尚无法满足患者多样化的临床需求，其疗效也受到多方面因素的影响和限制。因此，更深入地理解青少年精神障碍的发病机制，已成为临床神经科学中亟待解决的关键问题。安全无创的多模态磁共振成像技术（Magnetic Resonance Imaging，MRI）为从解剖结构、功能活动、连接组等不同视角剖析青少年精神障碍背后的脑生理机制提供了关键途径。尽管越来越多结合MRI的研究表明，青少年精神障碍是高度动态、灵活的脑发育过程和基因、环境等多种因素相互作用的结果，但现有研究对跨模态影像特征互补信息的利用并不充分，在更全面地发掘青少年精神障碍及其相关风险/致病因子作用机制的神经影像学表征上，还面临诸多挑战。为此，本文基于转基因动物模型和青少年随访队列的多模态影像数据、遗传与行为学指标等多维度信息，在多模态融合计算框架下，系统地探究了与罕见单基因拷贝数变异、逆境暴露以及大量常见遗传变异等风险因素相关的多模态脑发育模式，及其与多种青少年精神障碍的关系。本文的主要创新性工作如下：

  （1）利用转基因大鼠模型的纵向多模态磁共振影像数据，本文探究了介导甲基化CpG结合蛋白-2基因（MEthyl-CpG-binding Protein 2，MECP2）功能获得性突变和孤独症样社交功能障碍、运动功能障碍的多模态脑异常发育模式。在有监督的多模态融合计算框架下，通过整合大鼠模型跨越整个青春期的3个发育节点的纵向静息态功能影像特征和结构影像特征，发现MECP2的功能获得性突变对一个主要由背内侧前额叶和压后皮层构成的大鼠默认模式样网络的功能发育具有显著的调控效应，同时还引起了海马灰质的过度生长和丘脑的萎缩。进一步通过与行为指标的相关性分析，发现该默认模式样网络在青春期的功能异常发育与大鼠成年后的孤独症样社交缺陷和运动缺陷显著相关。以上结果揭示了孤独症样行为缺陷背后的一条“基因→脑发育”调控通路，对于促进孤独症神经生物学机制的理解具有十分重要的意义。

  （2）本文提出了一种更精确描述大脑网络水平跨模态交互作用的结构−功能耦合指标（StruCture−FunCtion，SC-FC），并系统性地探讨了青少年重度抑郁症及其伴不同环境风险暴露和具有不同临床特征的亚组中SC−FC耦合的重构方式。基于168例单中心临床青少年抑郁症和101例健康对照的多模态磁共振影像数据，本文发现青少年抑郁症患者在18个大脑亚区表现出显著的SC−FC耦合增强。这些亚区主要涉及视觉网络、默认模式网络和突显网络的枢纽脑区，如双侧腹内侧顶枕沟、右腹侧后扣带回、脑岛等区域。在更精细临床亚组上的差异分析进一步揭示了SC−FC耦合重构模式的异质性，尤其是在伴自杀尝试的抑郁亚组中，发现了海马旁回SC−FC耦合的特异性降低，以及在伴重大生活事件暴露的抑郁亚组中，发现了额叶−边缘回路SC−FC耦合特异且广泛地升高。以上结果为剖析青少年抑郁症的神经病理学机制和临床异质性提供了一个全新的视角。

  （3）基于大规模青少年随访队列，本文发掘了注意缺陷/多动障碍（Attention Deficit/Hyperactivity Disorder, ADHD，简称多动症）遗传风险的多模态影像表征，并考察了其与常见青少年精神障碍和多方面行为特征的关联。结果表明，在脑区水平，ADHD相关的遗传变异对青少年背外侧前额叶、顶下小叶和脑干神经核团等区域的局部结构和功能发育具有显著的调控作用；在大规模脑功能网络水平，ADHD相关的遗传变异显著影响了视觉网络、感觉运动网络和认知控制网络之间的功能交互协调性。同时，研究还发现了ADHD遗传风险与大规模脑功能网络活动的关联性具有一定程度的跨人种泛化性。在多疾病差异性分析中，研究证实了多种精神障碍在与ADHD遗传风险关联的多层次功能活动模式上表现出显著异常，提示了精神障碍的基因重叠与共享的神经病理学机制之间的潜在联系。最后，通过个体化预测分析，研究进一步揭示了该多模态影像表征与显著受ADHD影响的行为特征之间的关联。以上结果共同表明，ADHD的多基因遗传结构对青少年脑发育具有多层次、跨模态的调控效应；同时，从神经生物学层面证实了ADHD相关的常见遗传变异对个体行为特质具有全方面的塑造作用。

  综上所述，本文在多模态磁共振影像融合计算框架下，通过充分挖掘不同模态的MRI数据所提供的互补信息及其与遗传因素、行为指标的关联，从不同的视角提供了对青少年精神障碍病理学机制的重要见解。

    The adolescence is a critical phase of rapid physical and psychological development in humans, yet it is also highly vulnerable for mental disorders. Adolescent mental disorders impose huge burdens on families and society, however, the efficacy of existing medical interventions for adolescent mental disorders are doubtful and limited by various factors, underscoring the urgent need for deepening the understanding of their pathogenesis. The safe and non-invasive multimodal magnetic resonance imaging (MRI) technology provides a valuable chance to dissect the brain physiological mechanisms underlying adolescent mental disorders from different perspectives. Although increasing MRI studies have indicated that adolescent mental disorders originate from the interactions between the highly flexible brain development and risk factor exposures, most of them did not fully leverage the complementary information across multimodal imaging features. Therefore, there are still huge challenges to comprehensively characterize the neurobiological bases of adolescent mental disorders. To this end, based on transgenic animal models and large-scale adolescent cohort samples，this dissertation have systematically explored multimodal brain development patterns related to various risk factors and their relationships with adolescent mental disorders, within a multimodal fusion computational framework. The main contributions of this dissertation are as follows:

    1. Leveraging longitudinal multimodal brain MRI data from transgenic rat models, this dissertation reveled the brain development patterns mediating the Methyl-CpG-binding Protein 2 (MECP2) duplication and autism-like behavioral deficits. Under a supervised multimodal fusion framework, by integrating resting-state functional imaging features and structural imaging features across three developmental milestones of the rat model, we discovered that MECP2 duplication had a significant effect on the functional development of a default-mode-like network in rats, which primarily composed of the dorsomedial prefrontal cortex and retrosplenial cortex, while also causing overgrowth of hippocampus and atrophy of the thalamus. Further correlation analysis revealed that the functional development of the default-mode-like network during adolescence significantly related to autism-like social and motor deficits in adult phase. These results reveal a “gene→brain” regulatory pathway behind autism-like behavioral deficits, which is of great significance for promoting the understanding of the neurobiological mechanisms of autism.

    2. By defining a more precise structure−function coupling (SC−FC) measure that characterizes interactions between multimodal connectome, this dissertation has systematically explored the SC−FC coupling reconfiguration in adolescents with major depressive disorder (MDD) and its subgroups with different environmental risk exposures and clinical characteristics. Based on multimodal MRI data of 168 adolescents with depression and 101 healthy controls, we found that adolescents with depression exhibited significant SC−FC coupling increase in 18 brain regions, mainly involving hubs of the visual network, default mode network, and salience network. Further subgroup analyses related to more refined clinical subgroups have revealed the heterogeneity of SC−FC coupling reconfiguration pattern, especially in the depressive subgroup with suicide attempts, where a unique reduction of SC−FC coupling in the parahippocampal gyrus was found, and in the depressive subgroup with exposure to major life events, a unique widespread increase of SC−FC coupling in the frontal-limbic circuit was found. These results provide a new perspective for dissecting the neuropathological mechanisms and clinical heterogeneity of adolescent depression.

    3. Based on a large-scale adolescent cohort, this dissertation has identified a multimodal imaging pattern related to the genetic risk of attention-deficit/hyperactivity disorder (ADHD) and has explored the associations between the imaging pattern and other common adolescent psychiatric disorders, as well as various behavioral traits. The results have shown that at the regional level, ADHD-related genetic variations have a significant regulatory effect on the regional structure and function development of dorsolateral prefrontal cortex, inferior parietal lobule, and brainstem nuclei; at the large-scale brain functional network level, ADHD-related genetic variations significantly affect the functional interactions between the visual/sensorimotor network and cognitive control network. In addition, we found that the association between ADHD genetic risk and large-scale brain functional network activity has a remarkable cross-ethnic generalizability. In the cross-diagnostic comparison analysis, the study confirmed that various mental disorders exhibit significant abnormalities in the identified functional patterns associated with ADHD genetic risk, suggesting potential links between the genetic overlap and the shared neuropathological mechanisms of mental disorders. Finally, through individualized predictive analysis, we further revealed the associations between the identified multimodal imaging pattern and behavioral traits which significantly affected by ADHD. These results collectively indicate that the polygenic structure of ADHD has a multilevel, cross-modal regulatory effect on adolescent brain, and confirm that ADHD-related common genetic variations play an important role in shaping the broad behavioral traits.

    In summary, under the MRI fusion computational framework, this dissertation provides important insights into the pathophysiological mechanisms of adolescent mental disorders from different perspectives, by fully exploiting the complementary information across different MRI modalities and exploring the associations between imaging features and genetic factors/behavioral measures.