Influence of Processing Tools on fMRI Metanalyses

描述(由申请人提供)：定量荟萃分析是将局部神经群体与认知操作联系起来的有效策略。通过将已发表的观察结果分组，定量荟萃分析可以收集聚合的证据，以创建基本认知操作和支持复杂任务的认知操作集合的神经解剖学和认知显性模型。在过去的四年里，荟萃分析策略已经由PI和研究成像中心(RIC)的同事们开发和验证。有效开展荟萃分析的电子环境(脑图)已投入使用。然而，目前的元分析策略并不能纠正图像处理和数据分析中使用的信息学工具之间的差异。目前，不同的信息学工具混淆了结果，特别是空间归一化、统计方法和推论。因此，评估和了解不同信息学工具对荟萃分析和神经成像结果的影响是非常重要的。基于我们目前的工作和NIH正在进行的奖励(题为“认知神经成像中的荟萃分析：方法验证”)，本申请寻求进一步改进和扩展建模，以说明信息学工具的影响，并包括统计参数图像的荟萃分析。这项应用的总体目标是通过评估和最小化不同信息学工具的影响，推动量化荟萃分析作为研究人类大脑功能图谱的方法的使用。将评估不同空间标准化程序对报告的功能激活位置的影响(目标1)。将扩大不同统计分析和统计推断战略对检测功能激活的影响，以包括对共享统计参数图像的分析(目标4)。通过这一应用程序开发的战略、准则、统计模型、电子工具和数据集将通过出版物、互联网和社区演示与脑图界共享。一旦被彻底批评和改进，将这些工具/战略整合到电子环境(BrainMap)中以供公共分发是我们的下一个目标。

DESCRIPTION (provided by applicant): Quantitative metanalysis is an effective strategy for associating localized neural populations with cognitive operations. By grouping published observations, quantitative metanalysis can assemble converging evidence to create neuroanatomically and cognitively explicit models of elementary cognitive operations and of ensembles of cognitive operations supporting complex tasks. Metanalysis strategies have been developed and validated by the PI and colleagues at the Research Imaging Center (RIC) for the last four years. An electronic environment (BrainMap) for effectively carrying out metanalyses is operational. At present, the metanalysis strategy, however, does not correct for differences among the informatics tools used in image processing and data analysis. Currently, results are confounded by different informatics tools, in particular by spatial normalization, statistical methods and inferences. Assessing and understanding effects of different informatics tools on metanalysis and neuroimaging results are therefore very important. Building upon our current work and an ongoing NIH reward (entitled "Metanalysis in Cognitive Neuroimaging: Methods Validations"), the present application seeks to further refine and extend modeling to account for the effects of informatics tools and to include metanalyses with statistical parametric images. The overall objective of this application is then to advance the use of quantitative metanalysis as a research method for functional human brain mapping by assessing and minimizing the effects of different informatics tools. The effects of different spatial normalization procedures on reported locations of functional activations will be assessed (Aim 1). The effects of different statistical analyses and statistical inference strategies on detection of functional activations will be extended to include analysis of shared statistical parametric images (Aim 4). The strategies, guidelines, statistical models, electronic tools, and data sets developed through this application will be shared with the brain-mapping community through publications, Internet and community presentations. Once thoroughly critiqued and refined, integration of these tools/strategies into an electronic environment (BrainMap) for public distribution is our next objective.