儿童青少年期是大脑快速发育，认知能力与行为发展的黄金期。其中，工作记忆能力作为执行功能的核心成分在学龄儿童青少年期经历着重要发展。个体在儿童青少年期的工作记忆能力与其他认知能力如推理、思维等密切相关，甚至能够预测个体的学业成绩以及个人成就。同时，学龄儿童青少年也是焦虑等情绪性因素的核心敏感期，长期焦虑不仅导致儿童认知能力发展受阻，而且还会影响大脑功能的发育。因此，探索学龄儿童工作记忆发展的大脑机制有利于揭示儿童脑智发育的一般性规律，对促进儿童认知发展有着重要意义。

基于脑智发育与认知发展的交互式特异化理论，儿童大脑在发育过程中往往需要协同多个脑区共同支持某一特定的认知任务。默认网络(default mode network，DMN)被认为是大脑的全局工作空间(Global workspace)，用于汇聚来自内部和外部的信息从而帮助个体预测和适应不断变化的环境需求；额顶网络(fronto-parietal network，FPN)则涉及广泛的外源性认知加工，对于执行功能(如工作记忆)和认知灵活性等至关重要。近年来，随着无创脑功能成像技术迅猛发展，越来越多的研究发现——大脑默认网络与额顶网络的功能交互，在高级认知功能中发挥着关键作用。然而，绝大多数研究仅针对已成熟的大脑展开，这两个网络究竟如何相互作用进而支持学龄儿童的认知发展，仍然知之甚少。同时，学龄儿童青少年时期也是焦虑等情绪性因素的核心敏感期，长期焦虑导致儿童认知能力发展受阻，同时也影响大脑功能的发育。然而特质焦虑在儿童发育过程中如何影响大脑默认与额顶网络的功能交互并进一步影响工作记忆发展的机制还尚不明确。基于这一前提，我们提出了本论文的关键科学问题：学龄儿童青少年期大脑默认网络与额顶网络的功能交互究竟如何支撑工作记忆发展，以及特质焦虑如何影响这两个网络间的功能交互以及工作记忆的发展？

为此，本研究以人脑默认网络与额顶网络的功能交互作为出发点，利用无创功能性磁共振成像技术fMRI和经典工作记忆范式，结合横断和纵向追踪设计，通过三个研究，系统考察大脑默认与额顶网络的功能交互发育对学龄儿童青少年工作记忆能力的促进作用及特质焦虑对其影响特征。主要包括以下三个方面：第一：学龄儿童在工作记忆任务中其大脑默认网络与额顶网络的功能交互与成人是否存在差异以及如何随年龄发展变化？第二：学龄儿童青少年阶段大脑默认网络与额顶网络功能交互的纵向发育如何在工作记忆发展过程中发挥作用？第三：特质焦虑这一关键情绪性因素如何影响学龄儿童大脑默认与额顶网络功能交互发育及工作记忆发展?

研究一使用数字n-back工作记忆范式通过比较329名学龄儿童(6-12岁)和103名成人(18-28岁)在工作记忆表现以及大脑功能之间的差异，揭示了学龄儿童的工作记忆与成人相比仍然不够成熟同时还伴随着弱于成人的与任务相关的默认网络负激活与额顶网络激活。通过多体素神经成熟度和大脑功能耦合的分析，我们发现大脑默认与额顶网络呈现出发育的异质性，通过支持向量回归分析发现大脑默认与额顶网络脑区二者的激活一起能够更好地预测儿童工作记忆的行为表现。除此以外，儿童默认-额顶网络间功能耦合强于成人，且静息态下的内在(intrinsic)网络间耦合介导了年龄与任务态下默认-额顶网络间的功能耦合之间的关系。后续分析发现后扣带回-背外侧前额叶皮层(PCC-DLPFC)的局部耦合成为网络间功能耦合中最突出的连接通路，该通路进一步与后扣带回在任务中的激活共同介导了儿童年龄与在工作记忆表现之间的关系。该研究探讨了学龄儿童工作记忆能力发展过程中大脑激活，多体素神经成熟度以及功能耦合的发育与成人之间的差异，且发现大脑默认网络通过与额顶网络功能的交互从而促进儿童工作记忆表现，揭示了两个大脑网络在儿童高级认知功能发展过程中的促进作用。

在研究一的基础上，研究二从纵向发育的角度，继续对528名正常发育的儿童青少年(6-15岁)的工作记忆能力发展过程中的大脑功能纵向发育规律进行探讨。研究二揭示了儿童工作记忆能力在6-15岁阶段的纵向发育轨迹，表明个体工作记忆能力在儿童青少年阶段经历持续发展。同时大脑默认与额顶网络的激活及多体素活动模式也体现出纵向发育的异质性。随后，对大脑功能耦合的分析结果表明，1-back条件下儿童与成人在网络间功能耦合的空间相似性有着线性的年龄发育效应。与此同时，通过比较第三年和第一年之间儿童工作记忆的发展变化，发现以默认网络后扣带回(PCC)为核心区域的局部网络间功能耦合的发育与儿童工作记忆表现的改善有着显著的相关性，使用偏最小二乘回归模型发现网络间局部功能耦合的发育可以预测儿童工作记忆表现的发展。本研究在研究一的基础上揭示了儿童青少年工作记忆能力的纵向发育轨迹，并且进一步发现以默认网络中后扣带回为核心的局部网络间功能耦合的发育对儿童工作记忆的发展有促进作用。

鉴于儿童青少年期是脑智发育容易受到负性情绪如特质焦虑等影响的敏感期，我们通过研究三考察了特质焦虑对儿童工作记忆发展以及对发育中默认与额顶网络功能交互的影响特征。脑像受试者为来自研究二中的儿童青少年，在脑像追踪的第二个时间点前后对153名儿童青少年(9-15岁)测量了特质焦虑。结果发现，高特质焦虑学龄儿童工作记忆发展更差，即特质焦虑水平越高的学龄儿童青少年在工作记忆任务中的行为表现越差、纵向增长较慢。同时，我们还发现高特质焦虑会导致左侧背外侧前额叶皮层(DLPFC_L)和右侧角回(AG_R)的激活以及多体素成熟度的减少。在脑功能耦合方面：高特质焦虑水平会削弱额顶网络内以及默认-额顶网络间的功能连接强度。本研究结果表明：特质焦虑在儿童工作记忆能力发展以及大脑网络间的功能交互上有着负性影响，通过抑制大脑默认-额顶网络的功能交互进一步抑制儿童青少年工作记忆的发展。

综合以上三个研究，我们发现儿童青少年的工作记忆能力经历持续关键的发展，在此阶段内个体的工作记忆表现显著弱于成人并表现出纵向发育趋势。伴随着工作记忆表现的改善，大脑默认与额顶网络在任务态下的激活以及多体素成熟度指标以及网络内功能耦合三个层面上均体现出发育的异质性，具体来说，大脑默认网络相对于额顶网络更加接近成人的成熟模式，且默认网络通过与额顶网络功能交互促进儿童青少年工作记忆表现的发展，其中以默认网络中后扣带回(PCC)作为核心区域的网络间局部功能耦合最为突出，据此我们推测大脑默认网络或许通过为相对未成熟的额顶网络发挥脚手架功能从而支持儿童青少年认知功能的发展。同时，特质焦虑通过影响大脑默认与额顶网络的功能交互进而影响儿童青少年工作记忆的发展。总体而言，本研究结合交互式特异化理论，融合了认知神经科学和发展心理学的研究方法与技术手段并结合多模态纵向追踪大样本儿童青少年脑像数据，分别从认知行为-多体素神经成熟度-大脑功能网络的发育等多个维度，全面系统地揭示了人脑默认网络通过增强与额顶网络之间的功能交互来支持工作记忆能力的发展规律，同时发现高特质焦虑可能会通过削弱默认网络与额顶网络之间的功能交互，进而影响学龄儿童工作记忆发展的现象。本研究结果为后续进一步解析儿童青少年脑智发育规律与认知能力发展的机理奠定了基础。

The human brain undergoes protracted development from childhood to adulthood, with significant improvement in a wide spectrum of learning, problem solving and executive functions. Working memory (WM), as a core component of executive function, also undergoes a critical and prolonged development during childhood and adolescence. With the rapid development of cognitive ability, childhood and adolescence is also a susceptibility period and a core risk period for emotional factors such as anxiety. Anxiety not only hinders the development of children's cognitive ability, but also affects the development of brain anctivity. Therefore, exploring the development of WM in children and adolescents, as well as the brain functional basis and influencing factors that support the development of this ability will help us to analyze the laws of children's brain development. The interactive specialization theory holds that children's brains often need to cooperate with multiple brain regions to support a specific cognitive task during development. The default mode network (DMN) is considered as the brain's global workspace, which integrates information from both internal and external sources to predict and adapt to changing environmental needs. The fronto-parietal network (FPN) is involved in a wide range of exogenous cognition and is essential for cognitive flexibility and executive function. A growing body of research has found that functional interaction of DMN-FPN supports performance on higher-level cognitive tasks in adults. However, how the two networks interact to support cognitive development in children remains elusive, and how anxiety as a trait influences the functional interaction of DMN-FPN development during childhood and affects the development of WM still remains unknown. Based on the interactive specialization theory, we raised the key scientific questions of this study: how does the functional interaction between DMN and FPN in school-age children and adolescence support WM development, and how trait anxiety affects the development of WM and functional interaction between these two networks?

This study focuses on the functional interaction of DMN-FPN in human brain, using the non-invasive functional magnetic resonance imaging technology fMRI and the classical WM paradigm and combining with cross-sectional and longitudinal designs to systematically examine the development of DMN-FPN functional interaction on the promotion of school-age children and adolescents WM ability and the influence of trait anxiety on WM development. It mainly includes the following three aspects: First: what’s the difference of DMN-FPN functional interaction during WM performance between school-age children wiht adults, and how does it develop with age? Second: how does the longitudinal development of the DMN-FPN functional interaction in school-age children and adolescents play a role in WM development? Third: how does trait anxiety affect the development of WM and DMN-FPN functional interactive in school-age children and adolescent?

In Study 1, by comparing differences in WM performance and brain function between 329 school-age children (6-12 years) and 103 adults (18-28 years) using the digital n-back WM paradigm, we found that the WM in school-age children is still immature compared with adults and is accompanied by weaker task-related deactivation and activation in DMN and FPN. Through the analysis of multi-voxel neural maturity and brain functional coupling, we observed the heterogenuouse developmental trajectories between DMN and FPN. Support vector regression analysis revealed that the accumulation of activity in DMN and FPN can better predict the WM behavioral performance in children. In addition, the functional coupling between DMN and FPN in children is stronger than that in adults. Subsequent analysis found that the coupling of the posterior cingulate-dorsolateral prefrontal cortex (PCC-DLPFC) became the most prominent connection pathway in DMN-FPN inter-network functional coupling, which further colobrated with the WM-related activity in PCC to mediate the relationship between children's age with WM performance. Study 1 explored the developmental differences between school-age children with adults in brain activation, multi-voxel neural maturity, and functional coupling during WM development, and found that relatively more mature DMN promotes children’s WM performance by interacting with relatively immaure FPN, revealing the positive role of DMN in the development of higher-order cognitive functions.

Based on Study 1, Study 2 continued to explore the longitudinal development of brain function during the development of WM in 528 normally developing children and adolescents (6-15 years old). Study 2 revealed the longitudinal developmental trajectories of children's WM abilities between the ages of 6 and 15, indicating that individual WM abilities experience prolonged development during childhood and adolescence. At the same time, the activation and multi-voxel activity patterns in DMN and FPN also reflected longitudinal developmental heterogeneity. Subsequently, the results of the analysis of brain functional coupling showed that the spatial similarity of functional coupling between children and adults in the 1-back condition had a linear age developmental effect. At the same time, by comparing the WM developmental changes between the third year and the first year, we found that the development of functional coupling between PCC with FPN was related to the improvement of children's WM performance. By using partial least squares regression model, we observed that the improvements in WM can be predicted by the development of functional coupling between PCC with FPN. Based on Study 1, Study 2 revealed the longitudinal developmental trajectory of WM ability in children and adolescents, and further found that the development of functional coupling between PCC with FPN can promote the WM improvement in children.

Since childhood and adolescence is a sensitive period for brain and intellectual development and can be easily affected by negative emotions such as trait anxiety, we investigated the effects of trait anxiety on the development of children's WM and the DMN-FPN functional interaction in development in Study 3. Children and adolescents’ brain imaging were from Study 2. Trait anxiety was measured in 153 children and adolescents (9-15 years old) before and after the second time point of brain imaging tracking. The results showed that school-age children and andoloscent with high trait anxiety showed slower WM development, in which school-age children and adolescents with higher trait anxiety had worse behavioral performance and slower longitudinal growth in WM. At the same time, we also found that high trait anxiety leads to decrease in activation and multi-voxel maturation of the left dorsolateral prefrontal cortex (DLPFC_L) and right angular gyrus (AG_R). In terms of network functional coupling: high levels of trait anxiety impaired the strength of functional connectivity within FPN and DMN-FPN. The results of this study showed that trait anxiety has a negative impact on the development of children's WM ability and the functional interaction between brain networks. By inhibiting DMN-FPN functional interaction, trait anxiety affected the development of WM in children and adolescents.

Combining the above three studies, we found that WM in children and adolescent undergoes a prolonged and critical development. During this stage, the individual's WM performance is significantly weaker than that of adults and shows a longitudinal development trajectory. Accompanied by the improvement of WM performance, DMN and FPN shows developmental heterogeneity in activation, multi-voxel maturity index, and functional coupling. Relatively DMN is more similar to the mature patterm in adults, while FPN is still immature, and DMN promotes the WM performance development in children and adolescents by interacting with FPN in whicn PCC work as the prominent core region in DMN. We speculate that DMN may support the development of cognitive function in children and adolescents by exerting a scaffolding function for the relatively immature FPN. At the same time, trait anxiety affects WM development in children and adolescents by affecting DMN-FPN functional interaction. In general, this study combines the theory of interactive specialization, and combines technical means of cognitive neuroscience and developmental psychology, and multimodal longitudinal tracking of large sample brain image data of children and adolescents. By interating multivoxel neural maturity-and brain functional network development, we systematically reveal that DMN in human brain supports the WM development by enhancing the functional interaction with FPN, and at the same time reveal that high-trait anxiety may affect the WM development in school-age children and andolescent by impairing DMN-FPN functional interaction. The results of this study laid the foundation for further analysis of the mechanism of brain development and cognitive development in children and adolescents.