Individualized brain systems and depression

项目摘要/摘要 这项建议的目标是推进严重抑郁障碍(MDD)的神经模型。以前的研究 MDD和相关疾病在对个体进行推断时依赖于组级别的信息 大脑，并产生了有限的翻译和临床影响。这种小组级别的方法是有限的，因为 强有力的证据表明，大脑在其组织中表现出很大的个体变异性。这份提案描述了 一种基于新的计算神经成像方法的计算精神病学方法，将提供 改进了对MDD患者的大脑进行映射的细节，包括与诊断状态、症状 和行为特征，以及预测治疗反应。 更具体地说，该团队提出了一种基于功能磁共振成像的高级脑成像方法，该方法将用于 在个人层面上深刻描述脑功能系统丰富的组织结构 (产生“个人化的大脑系统”)。拟议的研究将通过利用700多个 通过数据共享获取的现有数据集。这一建议是可行的，部分原因是数据共享和 PI和团队先前的研究提供了坚实的理论和方法基础。MDD是一种 这项提议的重点特别有希望，因为它(1)高度异质，因此是 绘制个体变异性图；(2)高度流行，是全球疾病负担的主要贡献者；以及 (3)MDD患者治疗后缓解不到1/3。 这项建议的具体目标是：(1)绘制MDD的个性化大脑系统图；(2)表征 个体化的大脑系统与核心的MDD症状和行为缺陷之间的关系；最后， (3)阐明个体化大脑系统与MDD临床试验结果之间的预测关系 机械上截然不同的处理。除了理论驱动的研究外，这项建议还包括发展 补充数据驱动的机器学习方法，只使用个性化的大脑系统 对特定患者做出临床上有意义的预测的功能。这将包括预测诊断 状态、症状和行为特征，以及治疗结果。 精准医学对包括心脏病学和肿瘤学在内的几个医学领域产生了相当大的影响。我们 在精神病学方面还没有看到类似的发展，部分原因是映射关系的挑战 在精神疾病和大脑的临床特征中。计算神经成像技术的发展 办法，包括目前提案中的办法，现在为应对这一挑战提供了新的机会。 和翻译鸿沟。

PROJECT SUMMARY / ABSTRACT The goal of this proposal is to advance neural models of major depressive disorder (MDD). Prior studies of MDD and related conditions have relied on group-level information when making inferences about individual brains, and have yielded limited translation and clinical impact. Such group-level approaches are limited given robust evidence that the brain exhibits substantial individual variability in its organization. This proposal describes a computational psychiatry approach rooted in new computational neuroimaging methods that will provide improved detail in mapping the brains of individuals with MDD, including in relation to diagnostic status, symptom and behavioral profiles, and predicting treatment response. More specifically, the team proposes an advanced fMRI-based brain mapping approach that will be used to deeply characterize the rich organizational structure of functional brain systems at the level of individuals (yielding “individualized brain systems”). The proposed research will be completed by leveraging over 700 existing datasets acquired through data sharing. This proposal is feasible, in part due, to data sharing and the strong theoretical and methodological foundations provided by the PI and the team’s prior research. MDD is a particularly promising focus for this proposal given that it is (1) highly heterogeneous and thus an ideal target for mapping individual variability; (2) highly prevalent and the leading contributor to global disease burden; and that (3) fewer than one in three MDD patients remit after treatment. The Specific Aims of this proposal are to: (1) Map individualized brain systems in MDD; (2) Characterize relations between individualized brain systems and core MDD symptoms and behavioral deficits; and, finally, to (3) Explicate predictive relations between individualized brain systems and MDD clinical trial outcomes to three mechanistically distinct treatments. In addition to theory-driven studies, this proposal includes the development of a complementary data-driven machine learning approach that will use only individualized brain system features to make clinically meaningful predictions about specific patients. This will include predicting diagnostic status, symptom and behavioral profiles, and treatment outcomes. Precision medicine has considerably impacted several medical fields, including cardiology and oncology. We have yet to see similar developments in psychiatry, given, in part, due to the challenge of mapping relations among clinical features of mental illness and the brain. The development of computational neuroimaging approaches, including those in the current proposal, now provide new opportunities to address this challenge and translational gap.