Assessing chronic pain using brain entropy mapping

使用脑熵图评估慢性疼痛

• 批准号: 10598873
• 负责人: Ze Wang
• 金额: $ 44.19万
• 依托单位: UNIVERSITY OF MARYLAND BALTIMORE
• 依托单位国家: 美国
• 财政年份: 2022
• 资助国家: 美国
• 起止时间: 2022-09-30 至 2024-08-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10598873
• 关键词: Address; Adult; Affect; Affective; Anxiety; Area; Attention; Big Data; Biological Markers; Brain; Brain imaging; Brain region; Chronic; Cognitive; Complex; Data; Data Analyses; Disease; Emotional; Ensure; Entropy; Evaluation; Functional Magnetic Resonance Imaging; Future; Goals; Health; Human; Image; Individual; Individual Differences; Intervention; Knowledge; Lead; Limbic System; Location; Machine Learning; Magnetic Resonance Imaging; Maps; Measures; Mental Depression; Methods; Modeling; Monitor; Motor; Motor Cortex; Opioid; Pain; Pain Research; Pain intensity; Pain interference; Pain management; Paper; Parietal Lobe; Perception; Personal Satisfaction; Pharmaceutical Preparations; Pilot Projects; Prefrontal Cortex; Problem Solving; Process; Psyche structure; Regulation; Research; Research Personnel; Rest; Role; Sensorimotor functions; Sensory; Severities; Specificity; Statistical Data Interpretation; Stress; Testing; Time; Translational Research; Trauma; Validation; Work; accurate diagnosis; attentional control; base; biobank; chronic pain; chronic pain patient; cingulate cortex; clinical pain; cognitive control; connectome; deep learning; deep learning model; diagnostic tool; driving force; experience; high reward; human data; indexing; innovation; machine learning model; multitask; neuroimaging; neuromechanism; neuroregulation; novel; pain perception; pain processing; predictive modeling; prognostic tool; success; technique development; therapy development; tool; traumatic stress; treatment effect

Chronic pain is one of the most prevalent health problems in the world but remains poorly understood and challenging to treat or manage. Two major barriers to progress are the unclear brain mechanism of chronic pain and the complexity to model the multifaceted individual differences in pain experience, making it difficult to accurately diagnose pain or monitor pain progression or treatment effects. To solve this problem, we need big data and sophistic models. In this novel project, we will use the large (n~100000) UK Biobank (UKB) data and deep machine learning (DL) to address the two critical problems in chronic pain research. We propose to use resting state fMRI (rsfMRI) because pain perception and processing are ongoing in the brain which can be characterized by the spontaneous brain activity measured by rsfMRI. We will focus on temporal coherence (TC) of rsfMRI given its fundamental role in brain functions and our leading expertise in this research topic. We initiated the concept of brain entropy (BEN) mapping as a tool to measure regional brain TC and our systematic work has demonstrated the high test-retest stability, sensitivity to causal effects, specificity to focal stimulation, different diseases, and drug states, as well as the potential as an intervention target through neuromodulations or medication. We also showed that TC/BEN contains unique information that can not be characterized by other neuroimaging measures. Aim 1 of this project will use TC/BEN mapping to find a potential chronic pain brain circuit where the resting TC is positively correlated with chronic pain and individual differences of pain experience. Aim 2 will use resting TC to build a multi-task DL-based pain prediction model. This project represents the first- of-its-kind to study TC in chronic pain. It will bring new knowledge about chronic pain brain mechanisms (resting TC alterations and associations) and a DL-based quantitative pain prediction model. Research rigor and method/finding generalizability will be ensured by the use of by far the largest rsfMRI data. These high innovations may lead to intervention targets for pain treatment or intervention development and a quantitative tool to evaluate individual differences in pain or pain progression. Feasibility of this project is guaranteed by the existing large data from UKB, our years of work experience in related research fields, the strong pilot data, and the strong team expertise. Success of this pilot project will immediately lead to large size future important studies in this new research direction.

慢性疼痛是世界上最普遍的健康问题之一，但仍然知之甚少， 难以治疗或管理。进展的两个主要障碍是不清楚慢性疼痛的大脑机制 以及模拟疼痛体验中多方面个体差异的复杂性，使得难以 准确诊断疼痛或监测疼痛进展或治疗效果。为了解决这个问题，我们需要大 数据和诡辩模型。在这个新的项目中，我们将使用大型（n~100000）英国生物银行（UKB）数据， 深度机器学习（DL）解决了慢性疼痛研究中的两个关键问题。我们建议使用 静息状态功能磁共振成像（rsfMRI），因为疼痛的感知和处理是在大脑中进行的， 其特征在于通过rsfMRI测量的自发脑活动。我们将重点关注时间相干性（TC） rsfMRI在大脑功能中的基本作用和我们在这一研究课题中的领先专业知识。我们发起 脑熵（BEN）映射作为测量局部脑TC的工具的概念和我们的系统工作 已经证明了高的重测稳定性，对因果效应的敏感性，对局灶刺激的特异性， 不同的疾病和药物状态，以及通过神经调节作为干预靶点的潜力 或者药物治疗我们还表明，TC/BEN包含的独特信息，不能用其他方法来表征。 神经影像学测量。该项目的目标1将使用TC/BEN映射来找到潜在的慢性疼痛脑 回路，其中静息TC与慢性疼痛和疼痛体验的个体差异正相关。 Aim 2将使用静息TC构建基于多任务DL的疼痛预测模型。该项目是第一个- 研究TC在慢性疼痛中的作用。它将带来关于慢性疼痛大脑机制（休息）的新知识。 TC改变和关联）和基于DL的定量疼痛预测模型。研究的严谨性和 方法/发现的普遍性将通过使用迄今为止最大的rsfMRI数据来确保。这些高 创新可能会导致疼痛治疗或干预发展的干预目标和定量 评估疼痛或疼痛进展的个体差异的工具。本项目的可行性由 UKB现有的大量数据，我们在相关研究领域的多年工作经验，强大的试点数据， 强大的团队专业知识。这一试点项目的成功将立即导致未来大规模的重要研究 在这个新的研究方向。