Infrastructure for hyperaligning fMRI data and estimating functional topographies

用于超对齐功能磁共振成像数据和估计功能拓扑的基础设施

• 批准号: 10689268
• 负责人: Maria I Gobbini
• 金额: $ 64.84万
• 依托单位: DARTMOUTH COLLEGE
• 依托单位国家: 美国
• 财政年份: 2022
• 资助国家: 美国
• 起止时间: 2022-08-23 至 2026-05-31
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10689268
• 关键词: Aging; Algorithms; Alzheimer' s Disease; Anatomy; Architecture; Atlases; Brain; Brain imaging; Clinical; Communities; Computer software; Consumption; Data; Data Aggregation; Data Set; Databases; Functional Magnetic Resonance Imaging; Generations; Grain; Hour; Human; Individual; Individual Differences; Infrastructure; Inosine Dialdehyde; Maps; Modeling; Molecular Conformation; Neurosciences; Participant; Personality Tests; Population; Population Analysis; Principal Investigator; Research; Research Infrastructure; Research Personnel; Research Project Grants; Research Support; Resources; Rest; Sampling; Scanning; Software Tools; Space Models; Standardization; Structure; Surface; System; Time; Work; affective neuroscience; aging brain; behavior prediction; cognitive neuroscience; cognitive testing; connectome; cortex mapping; cost effective; data sharing; experimental analysis; flexibility; high dimensionality; information model; movie; neuroimaging; open source; preservation; response; sharing platform; tool

PROJECT SUMMARY Shared information in cortical functional architecture is embedded in topographies that are idiosyncratic, posing a major impediment for functional brain imaging research. Hyperalignment resolves this problem by projecting information from individual brains into a common model information space. The proposed research project will create HyperBase – research infrastructure that will enable the brain imaging research community to leverage hyperalignment to greatly enrich their data, enable analyses of shared information and individual differences embedded in idiosyncratic fine-scale cortical topographies, and create a data sharing platform for data in the hyperaligned common model information space. The infrastructure will be an optimized, standardized template common model space based on a normative database, turnkey software tools for hyperaligning new brains and estimating individual functional topographies, and a framework for sharing hyperaligned data. These data and tools will provide community infrastructural support for research on a broad range of topics in clinical neuroscience, brain aging, and basic cognitive neuroscience. The proposed database will consist of fMRI data in 60 participants collected during movie viewing, story listening, at rest, and during a large set of functional localizers, augmented with demographic information and cognitive and personality test scores. Specific aims 1. Produce an optimized, standardized template for hyperalignment based on a normative database with open-source software that will allow mapping numerous functional topographies, based on standard localizer data in the normative sample, into new participant brains using only fMRI data collected while the new participants watch a movie, listen to a story, or are at rest. 2. Adapt hyperalignment algorithms to work with a standard template and estimate functional topographies via the template and normative localizer data. Develop new hyperalignment algorithms that increase power, precision, and flexibility. 3. Create a system for sharing functional brain imaging data that are projected into the common information space model, allowing accumulation of data in a framework that affords at a fine-grained level of detail. Hyperalign existing public datasets.

项目摘要 皮质功能体系结构中的共享信息嵌入了地形中 特质，为功能性脑成像研究带来了重大障碍。过度对准 通过将各个大脑的信息投射到通用模型中来解决此问题 信息空间。 拟议的研究项目将创建夸张的研究基础架构 使大脑成像研究界能够利用超对准能够极大地丰富他们的 数据，可以分析共享信息以及嵌入的个体差异 特质细尺度皮质地形，并为数据共享数据共享平台 超平整的通用模型信息空间。基础架构将是优化的， 基于普通数据库的标准化模板公共模型空间，交钥匙软件 用于超一对齐新大脑和估计单个功能地形的工具， 共享超级数据的框架。这些数据和工具将为社区提供 基础设施支持临床神经科学中广泛主题的研究 衰老和基本的认知神经科学。所提出的数据库将由60的fMRI数据组成 参与者在电影观看，故事聆听，休息和大量期间收集的参与者 功能性本地化，增强了人口统计信息以及认知和个性 测试分数。 具体目标 1。基于正常 带有开源软件的数据库，该软件将允许映射众多功能 基于正常样本中的标准本地化数据的地形为新的 参与者的大脑仅使用新参与者看电影时收集的fMRI数据， 听一个故事，或者在休息。 2。适应超数算法与标准模板一起使用并估算功能 通过模板和普通本地化数据的地形。开发新的超容器 增加功率，精度和灵活性的算法。 3。创建一个用于共享功能性脑成像数据的系统 通用信息空间模型，允许在一个框架中加速数据 提供细节的细节。现有公共数据集超级安装。