Statistical Analysis Core

摘要 产生丰富的多层次行为、基因类型和成像数据将带来的好处 PPG取决于对这些数据的分析质量。统计分析(SA)核心将形成一个中心 为私营部门小组提供统计和数据分析方面的专门知识和支助，其使命是加强科学研究 通过确保数据的最佳使用来取得进展。因为数据分析中的最佳结果是通过 将数据分析方面的专业知识与相关主题的专业知识相结合，SA核心一直是 旨在将耶鲁大学统计系的统计专家与精选数据结合在一起- 项目1-5中对科学非常熟悉的定向人员。此次合作将 开发和应用最强大和最具启发性的分析方法，保持严格的统计有效性 而不会人为地限制方法的复杂性。各个项目将执行分析以 实现他们的特定目标，SA核心将在需要时提供帮助。SA Core还将负责 执行使用来自多个项目的数据的分析，以满足跨 队列、年龄和数据类型。SA核心描述中详细介绍了几个综合目标。例如, 我们将使用在产前和新生儿期进行的纵向fMRI连接性测量 项目1和项目4预测自闭症的严重程度并预测对一种新的社会价值培训的反应 项目5中提出的协议。我们将最大限度地采用项目2中的大脑连接网络 与学龄儿童持续关注研究项目1中的新生儿有关，这些新生儿太小，不能 选择性注意力要被有效地测量。我们将介绍诱导性神经生物学特征 从项目3到功能磁共振网络的多能干细胞连通性以及注意力和自闭症的测量 来自项目2的学龄。在通过仔细应用经典的 例如方差分析、回归和混合效应模型等方法，SA Core将寻求应用更现代的、 在统计理论和方法方面的成熟发展，以开发改进的分析。如中所述 SA核心描述，成员被选为在统计方面具有独特和互补的专业知识 预期对项目和综合目标具有重要意义的领域，如高维度统计 (这适用于在本PPG中为每个参与者测量多个变量)、维度 简化、统计计算、纵向分析和网络模型(将适用于两个大脑 PPG中的连通性网络和遗传网络)。SA核心的一个附带好处是招聘 周博士和内加班博士，从事最高级别统计研究的学者，到 自闭症研究。我们还预计会有更多的耶鲁大学统计专业的研究生参与 目前正在进入充满活力的扩张期的耶鲁儿童自闭症研究部门 学习中心。 1

Summary The benefits to come from producing the wealth of multi-level behavioral, genotypic, and imaging data in this PPG depend on the quality of the analyses of these data. The Statistical Analysis (SA) Core will form a center of expertise and support in Statistics and data analysis for the PPG, with a mission of enhancing scientific progress by ensuring optimal use of data. As optimal results in data analysis are achieved through a combination of expertise in data analysis with expertise in relevant subject matter, the SA Core has been designed to bring statistical experts from the faculty of the Yale Statistics Department together with select data- oriented personnel from Projects 1-5 who are intimately familiar with the science. This collaboration will develop and apply the most powerful and revealing analytic methods, maintaining rigorous statistical validity without artificially limiting the sophistication of the methodology. The individual Projects will perform analyses to achieve their specific aims and the SA Core will assist as needed. The SA Core will also be responsible for performing analyses that use data from more than one project to address integrative aims that span across cohorts, ages, and data types. Several integrative aims are detailed in the SA Core description. For example, we will use longitudinal fMRI connectivity measurements made in the prenatal and neonatal periods from Project 1 and Project 4 to predict autism severity and also to predict response to a novel social value training protocol proposed in Project 5. We will adapt brain connectivity networks from Project 2 derived as maximally associated with sustained attention in school-age children to study neonates in project 1 who are too young for selective attention to be measured effectively. And we will relate neurobiological characteristics of induced pluripotent stem cells from Project 3 to fMRI network connectivity and measures of attention and autism in school age from Project 2. After establishing baseline analyses through careful applications of classical methods such as ANOVA, regression, and mixed-effects models, the SA Core will seek to apply more modern, sophisticated developments in statistical theory and methodology to develop improved analyses. As detailed in the SA Core description, members were chosen to have distinct and complementary expertise in statistical areas expected to be of importance to the projects and integrative aims, such as high dimensional statistics (which applies whenever many variables are measured for each participant as in this PPG), dimension reduction, statistical computing, longitudinal analysis, and network models (which will apply to both brain connectivity networks and genetic networks in this PPG). A collateral benefit of the SA Core is the recruitment of Dr. Zhou and Dr. Negahban, scholars working at the highest levels of statistical research, to the field of autism research. We also anticipate increased participation by graduate students from the Yale Statistics Department, which is currently entering a vibrant period of expansion, in the autism research of the Yale Child Study Center. 1