Using Artificial Intelligence to Predict Cognitive Training Response in Amnestic Mild Cognitive Impairment

使用人工智能预测遗忘型轻度认知障碍患者的认知训练反应

• 批准号: 10572105
• 负责人: JOSEPH M GULLETT
• 金额: $ 16.01万
• 依托单位: UNIVERSITY OF FLORIDA
• 依托单位国家: 美国
• 财政年份: 2023
• 资助国家: 美国
• 起止时间: 2023-05-01 至 2028-04-30
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10572105
• 关键词: Activities of Daily Living; Address; Adherence; Alzheimer' s Disease; Alzheimer' s disease risk; Analysis of Variance; Artificial Intelligence; Attention; Behavior Therapy; Biometry; Brain; Brain region; Characteristics; Clinical; Clinical Trials; Cognitive; Cognitive Therapy; Cognitive aging; Complex; Control Groups; Data; Data Set; Decision Making; Diagnosis; Disease; Disease Progression; Educational Intervention; Educational workshop; Effectiveness; Elderly; Florida; Foundations; Functional Magnetic Resonance Imaging; Future; Healthcare; Home; Impairment; Individual; Infrastructure; Intervention; Intervention Trial; Investigation; Investigative Techniques; Life; Logistic Regressions; Machine Learning; Magnetic Resonance Imaging; Memory; Mentored Patient-Oriented Research Career Development Award; Mentors; Mentorship; Methodology; Modeling; Neuroanatomy; Neurobiology; Neurocognitive; Neurodegenerative Disorders; Neurosciences; Outcome; Patients; Pattern; Performance; Persons; Population; Populations at Risk; Predictive Factor; Prefrontal Cortex; Proxy; Public Health; Quality of life; Research; Research Personnel; Resources; Rest; Risk; Sample Size; Sampling; Science; Short-Term Memory; Structure; Superior temporal gyrus; Techniques; Thinking; Training; Training Programs; Training and Education; Universities; Work; amnestic mild cognitive impairment; behavioral clinical trial; career; career development; clinical translation; cognitive performance; cognitive training; cohort; comparison control; design; executive function; experience; gray matter; group intervention; high risk; improved; individual patient; innovation; machine learning model; mild cognitive impairment; multimodal neuroimaging; neural; neuroimaging; normal aging; payment; personalized intervention; personalized medicine; predicting response; predictive modeling; processing speed; professor; random forest; recruit; responders and non-responders; response; skills; support vector machine; tool; treatment effect; vector; white matter

Project Summary/Abstract: This K23 Clinical Trial project will provide Dr. Gullett, an Assistant Professor at the University of Florida, the direct mentored-training needed to address important questions related to intervention response in an amnestic mild cognitive impairment (aMCI) diagnosed population at risk for Alzheimer’s disease. As a neuropsychologist, Dr. Gullett has gained clinical experience in the assessment of neurodegenerative diseases including Alzheimer’s disease and its precursor, MCI, as well as research experience using structural neuroimaging to investigate various clinical disorders. The support provided by the K23 mechanism through the NIA will provide Dr. Gullett with the protected mentored-training needed to build on his current skills and become an expert in clinical neuroscience, machine learning, and behavioral interventions for mild cognitive impairment and Alzheimer’s disease. Career development and training plan: Dr. Gullett’s training plan consists of foundational formal coursework in 1) clinical trials, 2) MCI and Alzheimer’s disease effects, and 3) biostatistics and machine learning investigative techniques. These foundations will be directly applied through mentorship by experts in the fields of behavioral cognitive interventions, neuroimaging, and machine learning, as well as a proposed in-person workshop in functional neuroimaging analysis. This mentored-training plan will provide Dr. Gullett with the expertise to not only carry out the proposed project, but to become a unique and invaluable resource for future collaborative efforts applying neuroscience-based machine learning tools to investigate personalized interventions for Alzheimer’s disease. Research plan: The proposed project will provide the clinical trials training needed for Dr. Gullett to establish the effectiveness of a planned take-home, 12-week cognitive training program in patients with amnestic mild cognitive impairment (N=75; Aim 1). The expert mentorship team proposed has decades of experience in behavioral clinical trials interventions, which will provide the applicant with design and methodology guidance, as well as the recruitment infrastructure and resources needed to successfully carry out the proposed project. Further, this project will provide training in multi-modal neuroimaging-based machine learning to determine the baseline neural, cognitive, and functional factors that distinguish aMCI patients who respond to treatment from those who do not (Aim 2). This innovative approach will ultimately allow the applicant to investigate which of a myriad of features aMCI patients possess at a baseline assessment are the most salient predictors of their ability to improve from a well-validated cognitive training intervention. A project such as this will enable Dr. Gullett to develop a unique skillset to facilitate an R01-level academic career tasked with providing individual aMCI patients personalized interventions based on their own unique neurobiological and cognitive features.

项目摘要/摘要： 这个K23临床试验项目将为佛罗里达大学助理教授Gullett博士提供直接的 需要指导的培训，以解决与遗忘症轻度干预反应相关的重要问题 认知障碍(AMCI)诊断为阿尔茨海默病的高危人群。作为一名神经心理学家，古利特博士 在评估神经退行性疾病包括阿尔茨海默病和 它的前体MCI以及利用结构神经成像研究各种临床的经验 精神错乱。K23机制通过NIA提供的支持将为Gullett博士提供受保护的 辅导式培训需要在他现有技能的基础上，成为临床神经科学、机器方面的专家 学习，以及对轻度认知障碍和阿尔茨海默病的行为干预。 职业发展和培训计划：Gullett博士的培训计划包括1) 临床试验，2)MCI和阿尔茨海默病的影响，以及3)生物统计学和机器学习研究 技巧。这些基金会将由行为学领域的专家通过指导直接应用 认知干预、神经成像和机器学习，以及拟举办的功能性 神经影像分析。这项有指导的培训计划将为Gullett博士提供专业知识，不仅可以执行 提议的项目，但要成为未来合作努力应用的独特和宝贵的资源 以神经科学为基础的机器学习工具，以研究阿尔茨海默病的个性化干预。 研究计划：拟议的项目将为Gullett博士提供所需的临床试验培训，以建立 计划好的为期12周的带回家的认知训练计划对遗忘型轻度认知障碍患者的疗效 损害(N=75；目标1)。建议的专家指导团队在行为临床方面拥有数十年的经验 试验干预，这将为申请人提供设计和方法指导，以及征聘 成功开展拟议项目所需的基础设施和资源。此外，该项目将提供 基于多模式神经成像的机器学习的训练，以确定基线神经、认知和 区分对治疗有反应和无反应的MCI患者的功能因素(目标2)。这 创新的方法将最终允许申请者调查MCI患者的众多特征中的哪一个 在基线评估中拥有的是最显著的预测指标，可以预测他们从经过充分验证的 认知训练干预。这样的项目将使Gullett博士能够发展一套独特的技能，以促进 R01级学术生涯的任务是为aMCI患者提供个性化的干预 具有独特的神经生物学和认知特征。