Predictive Markers and Mechanisms of Persistent Psychotic-like Experiences in Children: An Adolescent Brain and Cognitive Development Study Analysis

儿童持续性精神病样经历的预测标记和机制：青少年大脑和认知发展研究分析

• 批准号: 10471418
• 负责人: Jason Smucny
• 金额: $ 17.47万
• 依托单位: UNIVERSITY OF CALIFORNIA AT DAVIS
• 依托单位国家: 美国
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-09-01 至 2026-08-31
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10471418
• 关键词: Adolescence; Adolescent; Affect; Age; Algorithms; Artificial Intelligence; Behavior; Brain; California; Child; Childhood; Classification; Clinical; Cognition; Computing Methodologies; DSM-III; Data; Data Set; Decision Making; Development; Device or Instrument Development; Diagnostic; Diffusion Magnetic Resonance Imaging; Distress; Early Intervention; Educational workshop; Environment; Equation; Event; Exposure to; Faculty; Family; Family history of; Functional Magnetic Resonance Imaging; Future; Goals; Grant; Influentials; Interview; Laboratories; Lead; Learning; Machine Learning; Manic; Measurement; Measures; Mental Health; Mental disorders; Mentors; Methods; Modeling; National Institute of Mental Health; Neurocognitive; Nutritional; Output; Patient Self-Report; Perceptual disturbance; Phenotype; Positioning Attribute; Postdoctoral Fellow; Predictive Factor; Prevention strategy; Psychiatrist; Psychiatry; Psychopathology; Psychoses; Questionnaires; Reading; Recommendation; Recording of previous events; Research; Research Assistant; Research Personnel; Research Project Grants; Rest; Risk; Risk Factors; Schedule; Secure; Severities; Societies; Structure; Substance abuse problem; Suicide attempt; Sum; Testing; Time; Training; United States National Institutes of Health; Universities; Vocational Guidance; adverse outcome; base; career; career development; cognitive development; cohort; deep learning; deep learning algorithm; demographics; diagnostic tool; environmental stressor; experience; follow-up; gray matter; ideation; improved; individualized medicine; individualized prevention; innovation; insight; meetings; multimodal data; multimodality; neurocognitive test; neuroimaging; novel; novel diagnostics; personalized intervention; personalized medicine; predictive marker; prevent; programs; psychotic-like experiences; skills; socioeconomics; statistics; tool; trustworthiness; tv watching; white matter

PROJECT TITLE Predictive Markers and Mechanisms of Persistent Psychotic-like Experiences in Children: An Adolescent Brain and Cognitive Development Study Analysis PROJECT SUMMARY/ABSTRACT My long-term career goal is to establish a highly influential and independent research program and become a broad spectrum leader and innovator in computational psychiatry. I therefore propose the following training goals: 1) To gain formalized training in computational psychiatry methods and statistics, including deep learning and structural equation modeling, 2) to obtain a deeper understanding of the field of developmental psychopathology, and 3) to acquire training in developing and sustaining an independent investigator position with his own laboratory. To meet the first goal, I will meet regularly with University of California Davis (UCD) faculty co-mentors Ian Davidson and Emilio Ferrer, attend relevant UCD classes, and complete a Research Plan with the following Specific Aims: 1) Using deep learning from Adolescent Brain and Cognitive Development (ABCD) Study data (demographic/clinical information, neurocognitive testing, neuroimaging data, and environmental metrics), predict child psychotic-like experience (PLE) distress scores the following year (i.e., baseline data predicting year one PQ-BC distress, year one data predicting year two distress, etc.), and 2) Using rules-based guidance from deep learning using explainable AI (XAI) algorithms, use sequential, structural equation modeling (SEM) of latent profiles (i.e., longitudinal trajectories, e.g., emerging, absent, remitting, persistent) to test hypotheses regarding the longitudinal mechanism(s) which may lead to persistent PLEs in the ABCD cohort. In Aim 1, XAI methods will be used identify the rules used by the deep learner to make decisions; in Aim 2, these rules will then be used to create latent constructs for modeling profiles of longitudinal trajectories. If r2>0.80 is consistently achieved for each iterative analysis in Aim 1, it suggests that ABCD instruments may predict PLE severity one year after measurement at a potentially clinically implementable level and help generate future hypotheses for personalized interventions aimed at reducing risk for persistent PLE distress. For Aim 2, insight will be gained with regard to how brain structure, brain function, neurocognitive ability, and environmental influences interact to influence the time course of PLE expression. To meet the second goal, I will receive mentoring including guided readings via monthly meetings with co-mentor Dr. Ellen Leibenluft, a world-renowned psychiatrist specializing in developmental research and Section Chief at the NIMH. I will also attend developmental classes, seminars, and workshops offered by UCD. To meet the third goal, I will attend weekly meetings with mentor Dr. Cameron Carter, from whom I will gain overarching career guidance. This will include learning how to create and oversee a laboratory, obtaining recommendations for society memberships, acquiring networking opportunities, securing grant sponsorships, and receiving RCR-related guidance. I will also attend mentoring and grantsmanship workshops offered by UCD, attend an off-campus workshop providing hands-on activities to provide exposure to fundamental management skills, complete a formalized UCD RCR plan, and gain mentoring experience by advising a research assistant and postdoc.

项目名称 儿童持续性精神病样经历的预测标志物和机制：一个青少年 大脑和认知发展研究分析 项目总结/摘要 我的长期职业目标是建立一个具有高度影响力和独立性的研究项目，并成为 计算精神病学的广泛领导者和创新者。因此，我提出以下培训目标： 1)获得计算精神病学方法和统计学的正式培训，包括深入 学习和结构方程建模，2）以获得更深入的了解该领域的 发展性精神病理学，3）获得发展和维持独立的 研究人员在自己的实验室工作。为了实现第一个目标，我将定期与大学举行会议， 加州戴维斯（UCD）教师共同导师伊恩戴维森和埃米利奥费雷尔，参加相关UCD课程， 完成具有以下具体目标的研究计划：1）使用青少年大脑的深度学习， 认知发展（ABCD）研究数据（人口统计学/临床信息，神经认知测试， 神经影像学数据和环境指标），预测儿童精神病样经历（PLE）的痛苦评分 下一年（即，预测第一年PQ-BC灾难的基线数据，预测第二年的第一年数据 遇险等），和2）使用基于规则的指导，从深度学习使用可解释的人工智能（XAI）算法，使用 潜在分布的顺序结构方程建模（SEM）（即，纵向轨迹，例如，出现， 不存在、缓解、持续），以检验关于可能导致 ABCD队列中的持续性PLEs。在目标1中，将使用XAI方法来识别深度 在Aim 2中，这些规则将用于创建用于建模概要文件的潜在构造 纵向轨迹。如果目标1中的每次迭代分析都一致实现r2>0.80，则表明 ABCD仪器可以预测PLE的严重程度一年后测量在一个潜在的临床可实施的 水平，并帮助产生未来的个性化干预的假设，旨在降低持续性 Ple distress.对于目标2，将获得关于大脑结构、大脑功能、神经认知 能力和环境影响相互作用，影响PLE表达的时间进程。满足 第二个目标，我将通过每月与共同导师艾伦博士的会议接受指导，包括指导阅读 Leibenluft是世界著名的精神病学家，专门从事发展研究，也是NIMH的科长。 我还将参加UCD提供的发展课程，研讨会和研讨会。为了实现第三个目标，我 我将参加与导师卡梅隆卡特博士的每周例会，从他那里我将获得全面的职业指导。 这将包括学习如何创建和监督一个实验室，获得社会的建议， 会员资格，获得网络机会，获得赠款，并接受RCR相关 指导我还将参加由UCD提供的指导和granitarian研讨会，参加校外 提供实践活动的讲习班，以提供基本管理技能，完成a 正式UCD RCR计划，并通过建议研究助理和博士后获得指导经验。