Integrating Multimodal Brain Imaging Data to Assess Subtle Cognitive Impairment

描述（由申请人提供）：人脑的功能神经影像学研究在理解正常和病理的认知过程中变得越来越重要。复杂的统计分析框架已经发展到定位信号变化和定义参与各种任务的大脑网络。然而，在细微的认知障碍中，例如：例如，暴露相关疾病，退行性疾病的早期阶段，损伤，癌症辅助治疗后的继发性疾病，这些方法对于检测由轻度脑功能障碍引起的脑状态的微小变化往往灵敏度较低。了解疾病机制或细微认知功能障碍的进展需要一种新的统计分析框架，其灵敏度更高，可以测量大脑状态的微小变化。我们已经开发了一种创新的方法，我们成功地应用于测量区域脑血流实验。这些方法使用完善的空间建模程序，在功能脑成像领域是新的，在有效分辨率组或“克里格”中得出统计上最优的空间摘要，初步研究表明，这些方法可以提高信号检测灵敏度，减轻多重测试负担。在这个新的空间建模框架内，我们建议将克里格方法扩展到fMRI和EEG，修改现有的表征大脑网络连接的技术（例如，基于克里格的独立分量分析），并使用基于数据驱动的有效分辨率组派生输入的统计分类器集成成像模式。我们的主要目标是开发这个分析框架，以深入了解轻度认知功能障碍的神经生理机制。实现这一目标可能建议治疗以减轻症状，防止进展，或至少为认知障碍患者提供知情的临床管理。

DESCRIPTION (provided by applicant): Functional neuroimaging studies of the human brain have become increasingly important in the understanding of normal and pathological processes of cognition. Sophisticated statistical analytic frameworks have been developed to locate signal change and define brain networks involved in various tasks. However, in subtle cognitive impairment-e.g., exposure-related illness, early stages of degenerative diseases, injury, secondary illness following adjuvant therapy for cancer-these methods tend to have low sensitivity for detecting small changes in brain states resulting from mild brain dysfunction. An understanding of disease mechanism or progression of subtle cognitive dysfunction requires a novel statistical analytic framework with improved sensitivity to measure small changes in brain states. We have developed an innovative methodology that we successfully applied in measures of regional cerebral blood flow experiments. These methods use well established spatial modeling procedures, new to the functional brain imaging field, to derive statistically optimal spatial summaries within effective resolution groups or "kriging", shown by preliminary studies to improve signal detection sensitivity and mitigate the multiple testing burden. Within this new spatial modeling framework, we propose to extend the kriging methodology to fMRI and EEG, modify existing techniques for characterizing brain networks of connectivity (e.g., kriging-based independent components analysis), and integrate the imaging modalities using a statistical classifier based on derived inputs of data driven effective resolution groups. Our primary goal is to develop this analysis framework to provide insight into the neurophysiological mechanisms of mild cognitive dysfunction. Achieving this goal may suggest treatments to alleviate symptoms, prevent progression, or at minimum, provide an informed clinical management of cognitively impaired patients. PUBLIC HEALTH RELEVANCE: Cognitive impairment is a major health concern, affecting people of all ages. Causes range from traumatic injury to toxin exposures, including chemotherapy for cancer treatment, to degenerative diseases. Mechanisms of damage or disease remain difficult to establish using current methods in functional brain imaging studies due to an inability to measure very small changes in brain states. We propose a new analytic framework using existing technology to improve the ability to measure subtle changes important in the understanding of disease pathology of impaired cognition and to greatly facilitate the integration of information from several imaging modalities with potential implications for clinical management and treatment.