Discovery Science of Human Brain Function Across the Lifespan

描述(申请人提供)：面对绘制大脑功能和整个生命周期中精神疾病的起源的挑战，功能神经成像社区正在遵循分子遗传学开创的先例，转向发现科学。1000功能连接计划(FCP)曾经是一个遥远的目标，它证明了对人脑功能进行发现科学的可行性。利用与静息状态功能磁共振成像(R-fMRI)轻松共享数据的优势，FCP汇集了来自全球各地独立收集的1200多个个体的数据集。FCP数据集立即展示了大规模R-fMRI研究的力量，揭示了与性别和年龄相关的大脑内在功能结构的广泛差异，这些差异在典型样本量(例如，n=20-100)下很难检测到。FCP的努力代表了有效、开放的数据共享的开端，曾经难以获得20-30个用于分析的数据集的研究人员突然获得了访问1200多个数据集的权限，用于方法创新和数据挖掘。在FCP初步成功的基础上，我们最近呼吁建立预期的数据共享，使用NKI-Rockland试点样本作为我们的原型，这是一个表型丰富的生命周期样本。通过26种精神病学、行为和认知评估工具获得的300多个表型变量，NKI-Rockland样本展示了一种数据收集模型，该模型非常适合于允许神经成像方法使用在分子遗传学中成功应用的数据挖掘和发现方法。神经成像仪现在可以识别大脑与行为的关系，并描绘出它们在整个生命周期中的动态轨迹。在确定了生成表型丰富的数据集并在预期的基础上与科学界公开共享这些数据集的可行性后，拟议的工作旨在生成一个更精心构建的寿命样本，规模更大，最大限度地代表社区，并使用适当的策略进行采样，以描绘大脑内在功能结构的度量标准轨迹。具体地说，我们建议在四年内招募1000名参与者(6-85岁)，对发育早期和老年晚期进行密集抽样，在这些时期，与年龄相关的变化梯度最大，模型拟合技术最容易产生人为结果。招聘和招生策略将受到仔细控制，以最大限度地提高样本的社区代表性，并最大限度地减少机会主义招聘中常见的偏见。在收集的数据集中，70%将被随机选择用于发现，其余30%将用于严格测试发现阶段产生的假设。与为试点NKI-Rockland样本建立的模型一致，作为这项提议的一部分产生的数据将通过FCP和相关的国际神经成像数据共享倡议(INDI)每周预期地共享。

DESCRIPTION (provided by applicant): Faced with the challenge of charting brain function and the origins of mental illness across the lifespan, the functional neuroimaging community is following the precedent set by molecular genetics in turning to discovery science. Once a distant goal, the 1000 Functional Connectomes Project (FCP) demonstrated the feasibility of conducting discovery science for human brain function. Capitalizing on the ease of data-sharing with resting state fMRI (R-fMRI), the FCP pooled datasets from over 1200 individuals independently collected at sites throughout the world. The FCP dataset immediately demonstrated the power of large-scale R-fMRI investigations, revealing widespread differences in the brain's intrinsic functional architecture related to sex and age that are not easily detected with typical sample sizes (e.g., n = 20-100). The FCP effort represented the inauguration of effective, open data-sharing, with researchers who once struggled to obtain 20-30 datasets for analyses suddenly granted access to over 1200 datasets for methods innovation and data mining. Building on the initial success of the FCP, we recently called for the establishment of prospective data-sharing, using the pilot NKI-Rockland Sample, a phenotypically rich, life-span sample, as our prototype. With more than 300 phenotypic variables obtained across 26 psychiatric, behavioral, and cognitive assessment tools, the NKI- Rockland Sample exemplifies they type of data collection model ideally suited to allow neuroimaging methods to employ the data mining and discovery approach applied successfully in molecular genetics. Neuroimagers can now identify brain-behavior relationships and delineate their dynamic trajectories across the lifespan. Having established the feasibility of generating phenotypically rich datasets and openly sharing them with the scientific community on a prospective basis, the proposed work aims to generate a more carefully constructed lifespan sample, larger in scale, maximally representative of the community and sampled using appropriate strategies for the delineation of normative trajectories for metrics of the brain's intrinsic functional architecture. Specifically, we propose to recruit 1000 participants (ages 6-85) over a four-year period, densely sampling early developmental and advanced aging periods, where age-related gradients of changes are maximal and model-fitting techniques are most prone to artifactual results. Recruitment and enrollment strategies will be carefully controlled to maximize the community representativeness of the sample and minimize biases commonly encountered with opportunistic recruiting. Seventy percent of the collected datasets will be randomly selected for discovery, while the remaining 30% will serve to rigorously test hypotheses generated during the discovery phase. Consistent with the model established for the pilot NKI-Rockland Sample, data generated as part of this proposal will be shared prospectively, on a weekly basis, via the FCP and the associated International Neuroimaging Data-sharing Initiative (INDI).