Mathematical and computational modeling of suicidal thoughts and behaviors

项目总结/摘要 自杀是一个毁灭性的公共卫生问题。美国每年有4万多人死于自杀， 使其成为第十大死亡原因，并造成超过300亿美元的生产力损失和医疗费用。 不幸的是，尽管科学进步已经导致其他主要死亡原因的显着下降， (e.g.,在过去的世纪里，目前的自杀率几乎与早期相同。 1900年代为了改善自杀的预测和预防，更好地了解自杀的风险机制， 自杀想法和行为背后的保护过程是必要的。计算精神病学认为 这对推进自杀研究，特别是通过建立和测试正式的理论，是有希望的。 尽管许多有影响力的自杀理论已经存在了几十年，但这些理论都是口头上的， 这使得它们在性质上未被充分指定（由于语言固有的不精确性）。另一方面，在一项研究中， 使用计算精神病学的工具，形式理论在数学方程中被实例化， 计算机代码。这就需要更具体和精确的确切强度，形式和时间尺度， 理论效果。事实上，形式理论已经导致了其他科学领域的重大进步和突破。 涉及复杂系统的理解和预测的领域（例如，生态系统、气候）。因此，在本发明中， 拟议的项目旨在通过使用数学和 计算建模，以建立，评估和测试自杀的正式理论。候选人和指导 团队，包括自杀和复杂动态系统计算建模方面的领先专家， 发展了一个关于自杀想法和行为的初步理论，包括认知，情感， 行为和社会因素。该项目的目标1是使用微分形式化这些关联 方程（一系列数学模型，指定函数及其 导数，并且对于模拟复杂系统随时间的变化非常有用）。目标二：转型 将这些数学方程转化为计算机代码，以模拟人工数据， 理论所暗示的行为。这一步将允许评估该理论是否能够 产生基本的，已知的关于自杀想法和行为的现象。最后，Aim 3将利用 数据来自一项正在进行的NIMH资助的自杀想法和行为的密集纵向研究（N = 300）， 根据实时收集的经验数据评估基于理论的模拟人工数据。拟议 这项研究最大的潜在影响是发展和评估自杀想法的第一个正式理论， 行为使用数学和计算建模，以及促进研究计划 结合了NIMH的两个主要优先事项：自杀预防和计算精神病学。合在一起，这可以 推进对自杀想法和行为的机械理解，以提高自杀预测， 并提供有关预防自杀的新治疗和干预目标的可操作信息。

Project Summary/Abstract Suicide is a devastating public health concern. More than 40,000 people die by suicide in the US each year, making it the 10th leading cause of death and responsible for >$30 billion in lost productivity and medical costs. Unfortunately, whereas scientific advances have led to significant declines in other leading causes of death (e.g., pneumonia, tuberculosis) over the past century, the current suicide rate is nearly identical to the early 1900s. In order to improve prediction and prevention of suicide, a better mechanistic understanding of risk and protective processes underlying suicidal thoughts and behaviors is needed. Computational psychiatry holds such promise for advancing suicide research, particularly through building and testing formal theories. Although many influential suicide theories have existed for decades, these have all been instantiated verbally, which renders them underspecified by nature (due to the inherent imprecision of language). On the other hand, using tools from computational psychiatry, formal theories are instantiated in mathematical equations and computer code. This requires more specificity and precision of the exact strength, form, and time scale of theorized effects. Indeed, formal theories have led to significant advances and breakthroughs in other scientific fields concerned with the understanding and prediction of complex systems (e.g., ecosystems, climate). Thus, the proposed project aims to address this major gap in suicide research by using mathematical and computational modeling to build, evaluate, and test a formal theory of suicide. The candidate and mentorship team, including leading experts in suicide and computational modeling of complex dynamic systems, have developed a preliminary theory of suicidal thoughts and behaviors encompassing cognitive, affective, behavioral, and social factors. Aim 1 of the project is to formalize each of these associations using differential equations (a family of mathematical models that specify the relationship between functions and their derivatives and are extremely useful for modeling change in complex systems over time). Aim 2 is to transform these mathematical equations into computer code to simulate artificial data, allowing for direct observation of the behavior implied by the theory. This step will allow for an evaluation of whether the theory is able to produce fundamental, known phenomena about suicidal thoughts and behaviors. Finally, Aim 3 will leverage data from an ongoing NIMH-funded intensive longitudinal study of suicidal thoughts and behaviors (N = 300) to evaluate the theory-based simulated artificial data against empirical data collected in real-time. The proposed study’s greatest potential impacts are to develop and evaluate the first formal theory of suicidal thoughts and behaviors using mathematical and computational modeling, as well as to promote a program of research uniting the two major NIMH priorities of suicide prevention and computational psychiatry. Together, this can advance the mechanistic understanding of suicidal thoughts and behaviors to improve suicide prediction, as well as provide actionable information about novel treatment and intervention targets for suicide prevention.