DEVELOPMENTAL MULTIMODAL IMAGING OF NEUROCOGNITIVE DYNAMICS (DEV-MIND)

项目概要/摘要 研究领域标准 (RDoC) 项目启动了一个开发研究分类的框架 基于基因、行为、电路等转化研究中出现的功能维度 认知生物学参数。近十年后，RDoC 已发展成为一个由功能性组成的矩阵 领域（例如认知系统）、特定领域的结构（例如注意力、感知）和分析单位 用于测量每个结构（例如，生理学、行为、基因等）。许多研究做出了贡献 根据认知生物学参数定义每个构造，虽然这些努力已被广泛 尽管维度 RDoC 结构取得了成功，但其本身在很大程度上仍未得到验证。 缺乏充分的验证是 RFA-MH-19-242 的核心，它要求提案“执行公正的 对现有结构进行数据驱动的验证，可能涉及合并、细分或分层组织 通过在构造之间和内部集成数据来实现它们。”具体来说，自由亚洲电台呼吁进行研究，使用“多重 每个构造的行为任务和分析级别”，“多模式数据融合……公正地分类和 比较结构”和“结构的数据驱动定义，涉及结构和功能数据 大脑状态、网络、电路动力学和信号层次结构如何与基于任务的输出相关 化验。” RFA 还鼓励使用“加速纵向设计，特别强调 发展……以及用于分类、预测和解释发展的尖端计算方法 轨迹。”神经认知和（表观）基因组动力学的发育多模态成像（Dev- MIND）联盟以创新的、大规模的发育多模式神经影像来响应这一号召 研究将利用先前开发的纵向儿科队列和数据融合算法 团队是通过 NSF 支持的 Dev-Cog 项目建立的。具体来说，Dev-MIND 将评估 认知系统领域内三种构造的单一性和潜在的层次结构（即， 注意力、认知控制和工作记忆）使用一系列自定义认知任务、多模式 成像、（表观）基因组分析、加速纵向设计和数据驱动的相似性度量 构建验证测试。我们的神经影像方法将包括基于 脑磁图 (MEG)、基于多模态分割的高分辨率体积 MRI 分析、 以及用于全脑动态功能连接的功能磁共振成像 (fMRI)。这些神经影像学和行为学 性能指标还将与（表观）遗传数据相结合，以确定基因组之间的协方差， 认知和神经活动模式。这种数据驱动的方法将使分类和预测成为可能 每个结构的发展轨迹，是计算精神病学目标的核心。总而言之，这 该项目汇集了领先的研究团队、一系列最先进的神经影像技术，以及 尖端的分析方法，对现有 RDoC 结构进行公正、数据驱动的验证。

Project Summary/Abstract The Research Domain Criteria (RDoC) project initiated a framework for developing research classifications based on functional dimensions emerging from translational research on genes, behaviors, circuits, and other cognitive-biological parameters. Almost a decade later, RDoC has grown into a matrix consisting of functional domains (e.g., cognitive systems), domain-specific constructs (e.g., attention, perception), and units of analysis for measuring each construct (e.g., physiology, behavior, genes, etc.). Numerous studies have contributed to defining each construct in terms of cognitive-biological parameters, and while these efforts have been broadly successful, the dimensional RDoC constructs themselves remain largely unvalidated. This lack of adequate validation is central to RFA-MH-19-242, which requests proposals that “perform unbiased data-driven validation of existing constructs that may involve merging, subdividing, or hierarchically organizing them by integrating data between and within constructs.” Specifically, the RFA calls for studies that use “multiple behavioral tasks and levels of analysis per construct,” “multimodal data fusion … to unbiasedly classify and compare constructs,” and “data-driven definitions of constructs that involve structural and functional data on how brain states, networks, circuit dynamics, and hierarchies in the signals relate to outputs from task-based assays.” The RFA also encourages the use of “accelerated longitudinal designs, with a particular emphasis on development … and cutting-edge computational approaches to classify, predict, and explain developmental trajectories.” The Developmental Multimodal Imaging of Neurocognitive and (Epi)genomic Dynamics (Dev- MIND) Consortium responds to this call with an innovative, large-scale developmental multimodal neuroimaging study that will leverage previously-developed longitudinal pediatric cohorts and data fusion algorithms that this team established through the NSF-supported Dev-Cog project. Specifically, Dev-MIND will evaluate the unitarity and potential hierarchical structure of three constructs within the cognitive systems domain (i.e., attention, cognitive control, and working memory) using a battery of custom cognitive tasks, multimodal imaging, (epi)genomic analysis, an accelerated longitudinal design, and data-driven similarity metrics for construct validation testing. Our neuroimaging approach will include dynamic functional mapping based on magnetoencephalography (MEG), high-resolution volumetric MRI analyses based on multimodal parcellation, and functional MRI (fMRI) for whole-brain dynamic functional connectivity. These neuroimaging and behavioral performance metrics will also be combined with (epi)genetic data to identify covariance between genomic, cognitive, and neural activity patterns. Such data-driven approaches will enable classification and prediction of developmental trajectories per construct, and are central to the goals of computational psychiatry. In sum, this project brings together leading investigative teams, an array of state-of-the-art neuroimaging technology, and cutting-edge analytical methods to perform unbiased, data-driven validation of existing RDoC constructs.