Neuroimaging-based biomarkers for two components of pain

描述（由申请人提供）：本申请涉及广泛的挑战领域（03）生物标志物的发现和验证，以及特定的挑战主题03- da -101，疼痛生物标志物。疼痛是影响很大一部分人口的生活质量和生产力的核心健康问题。据估计，美国每年因疼痛造成的工作时间损失高达650亿美元。这一申请的挑战在于定义稳健且有意义的生物标志物，作为疼痛相关过程的客观定量测量。目前，疼痛评估几乎完全基于患者的自我报告，这些报告本质上受到生物伤害感受（疼痛相关）过程与患者口头或书面疼痛描述之间复杂关系的限制。当自我报告不可用或不可靠时，客观的生物标志物对理解疼痛和预测疼痛都很有用。它们将加速疼痛研究的步伐，并在几个方面加强对病人的护理。首先，生物标志物可以作为临床试验和治疗的中间结果指标，使临床试验成本更低，治疗更符合患者的个人需求。其次，生物标志物可以帮助开发直接针对特定大脑系统的疼痛新干预措施，即通过非侵入性脑刺激，这将为疼痛管理开辟新的可能性。第三，对生物标志物的研究可以提高法律背景下关于疼痛的决策质量。我们概述了一项使用功能磁共振成像（fMRI）与其他方法相结合的建议，以开发和验证在急性实验诱导疼痛环境中报告的疼痛体验的两个组成部分的生物标志物。这一努力的核心部分是使用机器学习算法来开发基于功能磁共振成像的最佳疼痛预测器。基于机器学习的生物标记物本质上是跨大脑区域的活动模式，具有最大的预测准确性和区分有效性，用于分离物理和非物理（例如，社会或情感）疼痛。这些模式可以作为跟踪疼痛处理的多个组成部分的基础，而不依赖于自我报告。然而，我们不是简单地尝试开发预测自我报告的替代措施，而是特别强调识别对疼痛报告做出不同贡献的可分离的脑过程成分。通过识别潜在疼痛成分过程的生物标志物，我们的目标是开发客观的、基于大脑的“中间表型”测量，这些测量可以独立针对治疗，因此它们本身就很有用，而不仅仅是为疼痛报告提供替代测量的能力。

DESCRIPTION (provided by applicant): This application addresses broad Challenge Area (03) Biomarker Discovery and Validation, and Specific Challenge topic 03-DA-101, Biomarkers for Pain. The challenge Pain is a central health problem that affects quality of life and productivity for a large segment of the population. Lost work time due to pain costs America an estimated $65 billion annually. The challenge in this request for applications is to define robust and meaningful biomarkers that can serve as objective, quantitative measures of pain-related processes. Currently, pain assessment is based almost exclusively on patients' self- reports, which are inherently limited by the complex relationship between biological nociceptive (pain-related) processes and patients' verbal or written descriptions of pain. Objective biomarkers would be useful for both understanding pain and for predicting pain when self-reports are unavailable or unreliable. They would accelerate the pace of research on pain and enhance patient care in several ways. First, biomarkers could serve as intermediate outcome measures in clinical trials and treatments, making clinical trials less costly and treatments better matched to patients' individual needs. Second, biomarkers could help develop new interventions for pain that directly target specific brain systems, i.e., with non-invasive brain stimulation, which would open up new possibilities for pain management. Third, research on biomarkers could improve the quality of decision-making about pain in legal contexts. Our approach We outline a proposal for using functional magnetic resonance imaging (fMRI), in conjunction with other methods, to develop and validate biomarkers for two components of reported pain experience in an acute, experimentally induced pain setting. A central part of this endeavor is the use of machine-learning algorithms to develop optimal fMRI-based predictors of pain. Machine-learning based biomarkers are essentially patterns of activity across brain regions with maximal predictive accuracy and discriminative validity for separating physical and non-physical (e.g., social or emotional) pain. These patterns can serve as the basis for tracking multiple components of pain processing without relying on self-report. Rather than simply trying to develop alternative measures that predict self-reports, however, we place a particular emphasis on identifying separable component brain processes that make different contributions to pain reports. By identifying biomarkers for component processes underlying pain, we aim to develop objective, brain-based measures of "intermediate phenotypes" that could be independently targeted for treatment, and so are useful in their own right beyond their ability to provide surrogate measures for pain reports. PUBLIC HEALTH RELEVANCE: The ability to identify the brain components that contribute to pain in a particular individual would allow care providers to select treatments appropriate for the individual. We have specific reason to believe that at least two separable components can be identified. New results from our laboratory suggest that at least two distinct brain networks make separable contributions to reported pain. One network, the "pain-processing network" (PPN), responds to painful peripheral stimulation (e.g., heat on the skin), and includes classic pain-processing regions. A second network, the "emotional appraisal network" (EAN), does not respond to stimulation of the body, but is involved in the cognitive generation of emotion. In our recent fMRI work on pain, both networks appear to make independent contributions to predicting how much pain a person will report in response to a given stimulus. Our goal is to develop biomarkers based on these two putative components. We plan to achieve these broad goals using converging evidence from three complementary approaches, each of which will be pursued in parallel and addressed in one Specific Aim. In Aim 1, we will develop fMRI- based pain biomarkers using machine learning techniques. We will use a set of fMRI data from several acute thermal pain experiments collected in our laboratory over the past 3 years (combined N = 157). These datasets have highly homogenous scanning and experimental protocols and were designed with the goal of fMRI-based pain prediction in mind. Their existence will allow us to accelerate the pace of computational biomarker development. In Aim 2, we will test whether the fMRI-based biomarkers we develop are causally related to pain, using combined fMRI and transcranial magnetic stimulation (TMS). Biomarkers that are causally related to pain are most likely to be useful as targets for drug or other treatments. We will stimulate regions within the PPN and EAN during fMRI scanning, and test whether effects on activity and connectivity in each network predict changes in pain experience. In Aim 3, we will extend our biomarker validation to chronic pain. We will test whether biomarkers developed in Aims 1 and 2 predict patterns of hypersensitivity to sensory stimuli in patients with focal lesions in the PPN and EAN. Together, these approaches provide three complementary ways of developing and validating fMRI-based biomarkers for pain. ) PUBLIC HEALTH RELEVANCE: This application addresses broad Challenge Area (03) Biomarker Discovery and Validation, and Specific Challenge topic 03-DA-101, Biomarkers for Pain. Pain is a central health problem that affects quality of life and productivity for a large segment of the population, but research and clinical care are hampered by the lack of objective, quantitative measures of biological processes that contribute to pain. Pain-processing biomarkers would be useful for both understanding the generation of pain within the brain (and individual differences therein) and for predicting pain when self-reports are unavailable or unreliable. They would accelerate the pace of research on pain and enhance patient care by a) serving as intermediate outcome measures in clinical trials and treatments, making clinical trials less costly and treatments better matched to patients' individual needs; and b) providing a foundation for new treatments that target the brain systems involved directly; and c) providing guidelines for decision-making about pain in legal contexts.