Integration and Validation of Lesion Methods for Cognitive Neuroscience

描述（由申请人提供）：病变方法，识别脑损伤部位与认知和行为获得性损伤之间的一致关系，仍然是神经科学中识别高级功能的神经基础不可或缺的方法。本项目的总体目标是将病变方法建立在严谨、定量的基础上，并确定和扩展其局限性，特别是在定位的特异性和可解释性方面。我们将介绍正式整合连接信息的方法，从扩散张量成像，到群体级基于体素的病变缺陷分析。我们将研究现有方法和新方法的有效性、可靠性和解剖学准确性，特别关注受试者群体的影响、不均匀的病变覆盖范围以及灰质和白质损伤的不同作用。我们将通过两种方式验证基于体素的病变研究统计方法：1)基于大量真实脑病变的模拟，捕捉自然病变样本复杂结构的影响；2)在视网膜定位组织的视觉系统中建立一个特征良好的病变数据库，其中将有多光谱MRI数据和广泛的功能评估（视网膜定位fMRI，多焦视觉诱发电位，定量视野和代谢PET数据）。另一个总体目标是确定病变方法的最佳实践，包括确定最优（高效、敏感、有效）的分析形式，并为评估方法的未来改进提供资源。最后，我们将把这个项目中产生的方法、软件、验证数据集和性能基准传播给大脑映射社区。这里提出的工作将使损伤方法能够以前所未有的信心用于识别正常认知和行为的神经系统的基本组成部分。对局灶性脑损伤的行为和认知后果的研究提供了关于认知能力的大脑基础的信息，这是无法从其他来源获得的。拟议的工作将有助于建立病变方法的最佳实践，使其更容易研究和理解由中风和其他神经系统疾病引起的损伤。

DESCRIPTION (provided by applicant): The lesion method, which identifies consistent relationships between sites of brain damage and acquired impairments of cognition and behavior, continues to be an indispensable approach in neuroscience for identifying the neural basis of higher function. The overall goal of this project is to put the lesion method on a rigorous, quantitative footing, and to determine and extend its limits, especially with respect to the specificity and interpretability of localization. We will introduce methods for formally integrating connectivity information, from diffusion tensor imaging, into group-level voxel-based lesion-deficit analyses. We will investigate the validity, reliability, and anatomic accuracy of existing and new methods, with particular attention to the impact of the subject group, non-uniform lesion coverage, and the different roles of gray and white matter damage. We will validate voxel-based statistical methods for lesion studies in two ways: 1) with simulations based on a large set of real brain lesions, that capture the effects of the complex structure of the natural lesion sample; and 2) in a well characterized database of lesions in the retinotopically organized visual system, for which there will be multispectral MRI data and extensive functional assessment (retinotopic fMRI, multifocal visual evoked potentials, quantitative visual fields, and metabolic PET data). Another overarching goal is to determine best practices for the lesion method, including identification of the optimal (efficient, sensitive, valid) forms of analysis and generation of resources for evaluating future improvements in the methods. Finally, we will disseminate methods, software, validation data sets, and performance benchmarks generated in this Project to the brain mapping community. The work proposed here will enable the lesion method to be used with unprecedented confidence to identify the essential components of neural systems for normal cognition and behavior. PUBLIC HEALTH RELEVANCE Studying the behavioral and cognitive consequences of focal brain damage provides information about the brain basis of cognitive abilities that can be gained from no other source. The proposed work will help establish best practices for the lesion method, making it easier to study and understand the impairments that result from stroke and other neurologic diseases.