Multimodal Data Analysis and Integration

多模态数据分析与集成

• 批准号: 10639548
• 负责人: Ting Xu
• 金额: $ 32.15万
• 依托单位: NATHAN S. KLINE INSTITUTE FOR PSYCH RES
• 依托单位国家: 美国
• 财政年份: 2017
• 资助国家: 美国
• 起止时间: 2017-04-15 至 2028-03-31
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10639548
• 关键词: Address; Anatomy; Biological Models; Brain; Cognitive; Computer Models; Data; Data Analyses; Diffusion; Dimensions; Electrodes; Electroencephalography; Ensure; Event; Frequencies; Functional Magnetic Resonance Imaging; Goals; Human; Image; Individual; Joints; Link; Macaca; Magnetic Resonance Imaging; Maps; Measurement; Methods; Modality; Modeling; Monkeys; Neurobiology; Output; Physiological; Property; Pupil; Reproducibility; Rest; Role; Signal Transduction; Surface; T2 weighted imaging; Testing; Work; data integration; design; distributed data; dynamic system; functional magnetic resonance imaging/electroencephalography; human data; improved; inter-individual variation; movie; multimodal data; multimodality; neural circuit; neural model; nonhuman primate; spatiotemporal

ABSTRACT: This Conte Center proposal seeks to elucidate the role of slow brain network fluctuations (SBNFs) across species and scales, and to establish computational models of neural circuit dynamics that characterize SBNFs. Data collected in Projects 1-3 will span multiple modalities (EEG, fMRI, ECoG/LFP, and single-unit recordings) and species (humans and nonhuman primates). The overarching goal of the Multimodal Data Analysis & Integration Core (Core-B) is to address the challenge of integrating data across these multiple modalities and analyses. To achieve this goal, Core-B will apply and develop multimodal alignment frameworks that establish a common space for analysis across modalities, individuals, and species (to be used in Projects 1-3), as well as integrate these data to facilitate computational modeling of neural dynamics (to be used in Project 4). Aim 1 will leverage the functional alignment framework to construct a common space for different modalities within- and between-subjects. Aim 2 will refine the cross-species alignment using the functional data from the common tasks to provide the spatial transformation to link human data in Projects 1-2 and macaque data in Project 3. In addition to optimizing cross-modal and cross-species alignment, Aim 3 will construct a framework for decomposing brain dynamics into spatiotemporal states and state-transitions derived jointly from EEG, fMRI, and physiological signals. We will also leverage the cross-species transformation between humans and macaques in Aim 2 to delineate the matched states across species and characterize the species-specific temporal configurations. This framework will provide a state-space in which brain dynamics and their causal properties can be interrogated across individuals and species. The alignment approach and aligned data generated by this Core will be applied to the curated, quality-checked outputs received from Core C after appropriate preprocessing (determined by domain experts from the individual Projects) and will be distributed to Projects 1-4.

摘要：康特中心的这一建议旨在阐明大脑网络缓慢波动的作用 （SBNFs）跨物种和规模，并建立神经回路动力学的计算模型， 表征SBNF。项目1-3中收集的数据将跨越多种模式（EEG、fMRI、ECoG/LFP和 单单位记录）和物种（人类和非人类灵长类动物）。多式联运的总体目标 数据分析和集成核心（Core-B）是为了解决跨这些多个 模式和分析。为了实现这一目标，Core-B将应用和开发多模式对齐 为跨模式、个体和物种的分析建立共同空间的框架（将使用 在项目1-3），以及整合这些数据，以促进神经动力学的计算建模（将 在项目4中使用）。目标1将利用功能调整框架， 不同的模式内-和之间-科目。目标2将使用 公共任务中的功能数据，以提供空间转换，从而链接项目1-2中的人类数据 和猕猴的数据除了优化跨模态和跨物种比对，目标3还将 构建一个框架，用于将大脑动力学分解为时空状态和状态转换 脑电功能磁共振成像和生理信号我们还将利用跨物种转化 目标2中人类和猕猴之间的关系，以描绘物种之间的匹配状态，并描述 物种特有的时间结构。这个框架将提供一个状态空间，在这个状态空间中， 它们的因果性质可以在个体和物种之间进行研究。调整方法和 由该核心生成的对齐数据将应用于从核心接收的策划的、质量检查的输出 C在适当的预处理（由各个项目的领域专家确定）后， 分配给项目1-4。