METANALYSIS COGNITIVE NEUROIMAGING: METHODS VALIDATIONS

本研究计划的目的是促进定量元分析作为人类功能脑制图的研究方法的使用。这将通过测试、完善和分发由PI和圣安东尼奥德克萨斯大学健康科学中心（UTHSCSA）的同事开发的元分析策略和软件工具来完成。Fox和他的同事们开发了一种新的策略，用于对人脑绘图文献的功能组织进行元分析建模，称为功能体积模型（FVM）。FVM计算心理操作的大脑位置的空间概率分布。Fox及其同事还开发了一个电子环境（BrainMap），用于编码、检索和可视化人类大脑图谱文献，作为元分析的辅助工具。BrainMap包括一个编码方案（一种层次语义），用于所使用的行为和心理操作的实验映射，称为行为编码方案（BCS）。在本提案中，FVM元分析策略（目标1）和BCS（目标2）将在Fox和UTHSCSA的同事以及华盛顿大学Raichle和Petersen小组的合作下进行验证和完善。FVM策略将通过使用两种替代策略对与语言相关的大脑区域进行建模来验证：原始数据元分析和基于文献的元分析（FVM）。通过访问参与小组的大量图像数据档案，原始数据元分析将成为可能。建模假设和结果将被独立检验（目标1）。行为编码方案将通过编码来自脑映射文献的3,30篇论文样本并分析其全面性、可重复性和准确性来验证（目标2）。BCS的验证是研究者将数据输入BrainMap的必要前提。将在两年一次的研讨会上讨论元分析的一般策略和规则，包括FVM策略的改进、BCS和BrainMap环境（目标3）。BrainMap将不断进行修改，以适应已确定的策略、程序、编码方案和本项目产生的数据（目标3）。战略和指南也将通过出版物和脑制图和信息学社区会议上的讲习班向社区介绍。由于这种元分析方法是新的，与先前的元分析形式有本质上的不同，信息学领域将从新的分析范式的验证中受益。

The objective of this research proposal is to advance the use of quantitative metanalysis as a research method for human functional brain mapping. This will done by testing, refining and distributing strategies and software tools for metanalysis developed by the PI and colleagues at the University of Texas Health Science Center at San Antonio (UTHSCSA). Fox and colleagues have developed a new strategy for metanalytic modeling of the functional organization of the human brain-mapping literature, termed functional volumes models (FVM). FVM computes spatial probability profiles of the brain locations of mental operations. Fox and colleagues have also developed an electronic environment (BrainMap) for encoding, retrieval and visualization of the human brain mapping literature, as an aid to metanalysis. BrainMap includes a coding scheme (a hierarchical semantics) for the behaviors used and mental operations experimentally mapped, termed the Behavioral Coding Scheme (BCS). In the present proposal, the FVM metanalysis strategy (Aim 1) and the BCS (Aim 2) will be validated and refined in a collaboration between Fox and colleagues at UTHSCSA and the Washington University group of Raichle and Petersen. The FVM strategy will be validated by modeling 12, language-related brain areas using two, alternative strategies: original-data metanalysis and literature-based metanalysis (FVM). Original-data metanalysis will be made possible by accessing the extensive image-data archives of the participating groups. Modeling assumptions and results will be independently tested (Aim 1). The behavioral coding scheme will be validated by coding three, 30-paper samples from the brain-mapping literature and analyzing for comprehensivity, reproducibility, and accuracy (Aim 2). Validation of the BCS is a necessary precursor to investigator entry of data into BrainMap. General strategies and rules for metanalysis will be addressed at biannual workshops amongst grant personnel and invited consultants, including refinements of the FVM strategy, the BCS, and the BrainMap environment (Aim 3). BrainMap will be modified in an ongoing manner, to accommodate identified strategies, procedures, coding schemes and data produced by this project (Aim 3). Strategies and guidelines will also be presented to the community through publications and workshops at meetings of the brain mapping and informatics communities. As this approach to metanalysis is new and substantially different from prior forms of metanalysis, the field of informatics stands to benefit from the validation of a new analysis paradigm.