Development of Dynamic Resting State Functional Connectivity Machine Learning Framework for Dementia

痴呆症动态静息态功能连接机器学习框架的开发

• 批准号: 10371520
• 负责人: fei jiang
• 金额: $ 14.53万
• 依托单位: UNIVERSITY OF CALIFORNIA, SAN FRANCISCO
• 依托单位国家: 美国
• 财政年份: 2022
• 资助国家: 美国
• 起止时间: 2022-08-15 至 2027-05-31
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10371520
• 关键词: Address; Affect; Algorithms; Alzheimer' s Disease; Area; Award; Benchmarking; Biological Markers; Biometry; Brain; Characteristics; Clinical; Cognitive; Data; Data Analyses; Dementia; Development; Doctor of Medicine; Doctor of Philosophy; Electroencephalography; Epidemiology; Foundations; Functional Imaging; Functional Magnetic Resonance Imaging; Goals; Image; Imaging Techniques; Impaired cognition; Individual; Intervention; Knowledge; Lead; Learning; Mentors; Methodology; Methods; Monitor; Multimodal Imaging; Nature; Neurodegenerative Disorders; Neurologic Dysfunctions; Neurology; Neurosciences; Pattern; Persons; Procedures; Research; Research Personnel; Rest; Risk; Sampling; Symptoms; Technical Expertise; Testing; Time; Training; Treatment Efficacy; United States; Validation; Work; biomedical imaging; career; cognitive change; cognitive performance; cognitive testing; early screening; efficacious treatment; feature extraction; feature selection; high dimensionality; imaging biomarker; imaging modality; machine learning framework; machine learning method; machine learning model; method development; multidisciplinary; multimodal neuroimaging; neuroimaging; neuroimaging marker; next generation; predictive modeling; prevent; radiological imaging; statistical and machine learning; success; therapy design; treatment effect

Project Summary/Abstract The objective of this proposal is to provide a robust course of training for Fei Jiang, Ph.D., a candidate with an excellent foundation in statistical and machine learning research, to enable her to become an independent investigator in the field of quantitative data analysis and statistical/machine learning methods development for neuroimaging research. The proposed research aims to extract dynamic resting-state functional connectivity from multimodality imaging and use them for the prediction of cognitive decline. The central hypothesis is that the resting state functional connectivity changes over the imaging acquisition period, and this dynamic pattern is crucial for the optimal prediction of cognitive decline. Towards proving this hypothesis, a unique machine learn- ing framework is proposed to (1) robustly extract dynamic resting-state functional connectivity from multimodality imaging; (2) identify the important features that are associated with individuals' cognitive scores; and (3) predict cognitive decline using the identified important features. Successful completion of the proposed research will provide the next generation machine learning framework for the extraction and analysis of dynamic resting-state functional connectivity and lead to potential endpoints that can be used in the assessment of treatment effects. Recognizing the multidisciplinary nature of the work proposed, the author will be mentored and work closely with an expert committee from multiple scientific areas of relevance to the project (Neuroimaging, Neurodegenerative disease, Biostatistics): Srikantan Nagarajan (primary mentor), Ph.D., Department of Radiology and Biomedical Imaging, Ashish Raj (co-mentor), Ph.D., Department Radiology and Biomedical Imaging, William W. Seeley (ad- visor), M.D., Ph.D., Department of Neurology, John Kornak (advisor), Ph.D., Department of Epidemiology and Biostatistics, Marilu Gorno Tempini (collaborator), M.D., Ph.D., Department of Neurology, Charles McCulloch (collaborator), Ph.D., Department of Epidemiology and Biostatistics. This committee will be coordinated by Dr. Nagarajan. The goal is that by the end of the K25, Dr. Jiang will have the requisite knowledge, technical skills, and expertise to submit a successful R01 proposal that integrates her expertise in statistical and machine learn- ing methods with a knowledge of the questions and approaches pertaining to imaging in neuroscience, acquired through this training period.

项目总结/摘要 本提案的目的是为蒋飞博士提供一个强大的培训课程，候选人， 在统计和机器学习研究方面有着良好的基础，使她能够成为一名独立的 定量数据分析和统计/机器学习方法开发领域的研究员， 神经影像学研究。该研究旨在提取动态静息态功能连接 并将其用于预测认知能力下降。核心假设是， 静息状态功能连接性在成像采集周期内变化，并且该动态模式是 对认知能力下降的最佳预测至关重要。为了证明这个假设，一个独特的机器学习- 提出了一种新的结构框架：（1）从多模态中鲁棒地提取动态静息态功能连接 成像;（2）识别与个体认知评分相关的重要特征;（3）预测 认知能力下降使用已识别的艾德重要特征。成功完成拟议的研究将 为动态静止状态的提取和分析提供了下一代机器学习框架 功能连接性，并导致可用于评估治疗效果的潜在终点。 认识到拟议工作的多学科性质，作者将得到指导，并与 来自与项目相关的多个科学领域的专家委员会（神经影像学、神经退行性疾病 疾病，生物统计学）：Srikantan Nagarajan（主要导师），博士，放射学和生物医学系 成像，Ashish Raj（共同导师），博士，放射学和生物医学成像部，威廉W。塞利（ad- visor），医学博士，哲学博士、神经病学系，John Kornak（顾问），博士，流行病学系 生物统计学，Marilu Gorno Tempini（合作者），医学博士，哲学博士、神经科，Charles McCulloch （合作者），博士，流行病学和生物统计学系。该委员会将由博士协调。 纳加拉詹我们的目标是，到K25结束时，蒋博士将拥有必要的知识，技术技能， 和专业知识来提交一份成功的R 01提案，该提案将她在统计和机器学习方面的专业知识整合在一起- 学习方法，了解与神经科学成像有关的问题和方法，获得 在这段训练期间。