Diagnostic and prognostic biomarkers for subtypes of addiction-related circuit dysfunction

项目摘要 物质使用障碍(SOD)的患病率正在上升，已经成为导致残疾的主要原因， 部分原因是我们对潜在的病理生理学的理解是不完整的。像大多数人一样 神经精神综合征，sod是高度异质性的，不同的机制可能在 一些个体，但不是在其他人，甚至在一个单一的诊断类别。此外，肥皂水经常与 与抑郁、焦虑和其他精神症状一起发生，使识别分子和 电路水平的机制，并将它们与那些涉及情绪和焦虑症的人分开。诊断性 因此，异质性是开发更好的治疗方法、识别生物标记物和 量化不同形式成瘾的风险，并预测治疗反应和复发。最近，我们 开发并验证了一种使用功能磁共振成像发现和诊断抑郁症亚型的方法 功能连通性的测量，这反过来又预测特定亚型的临床症状和 治疗结果。在这里，为了响应PAR-18-062，我们提出了一个二次数据分析，它将 将这种方法扩展到SODS，利用多种深度特征化和大规模的神经成像 数据集。我们的中心假设是，在损伤发生机制上的个体差异 反应抑制和显著归因(IRSA)由不同形式的功能障碍连接所介导 在与成瘾相关的回路中，这反过来又相互作用并引起不同的神经生理学成瘾 子类型。在目标1中，我们将使用统计聚类和机器学习方法来描述这些子类型 并优化分类器(fMRI生物标记物)用于个别患者的诊断，最初重点放在 可卡因成瘾。在目标2中，我们将通过首先将这些亚型特定的生物标记复制到 新的数据集，然后评估其纵向稳定性和预测效用。在目标3中，我们将测试 亚型特有的电路机制一般用于在其他形式的成瘾中调节IRSA功能，以及 明确它们与介导快感缺乏和焦虑唤醒的不同机制之间的相互作用 抑郁和焦虑并存。

Project Summary Substance use disorders (SUDs) are increasing in prevalence and are already a leading cause of disability, due in part to the fact that our understanding of the underlying pathophysiology is incomplete. Like most neuropsychiatric syndromes, SUDs are highly heterogeneous, and distinct mechanisms may be operative in some individuals but not in others, even within a single diagnostic category. Furthermore, SUDs frequently co- occur with depression, anxiety, and other psychiatric syndromes, complicating efforts to identify molecular and circuit-level mechanisms, and disentangle them from those involved in mood and anxiety disorders. Diagnostic heterogeneity is thus a fundamental obstacle to developing better treatments, identifying biomarkers for quantifying risk for different forms of addiction, and predicting treatment response and relapse. Recently, we developed and validated an approach to discovering and diagnosing subtypes of depression using fMRI measures of functional connectivity, which in turn predicted subtype-specific clinical symptom profiles and treatment outcomes. Here, in response to PAR-18-062, we propose a secondary data analysis that would extend this approach to SUDs, leveraging multiple deeply characterized and large-scale neuroimaging datasets. Our central hypothesis is that individual differences in mechanisms underlying impairments in response inhibition and salience attribution (iRISA) are mediated by distinct forms of dysfunctional connectivity in addiction-related circuits, which in turn interact and give rise to distinct neurophysiological addiction subtypes. In Aim 1, we will use statistical clustering and machine learning methods to delineate these subtypes and optimize classifiers (fMRI biomarkers) for diagnosing them in individual patients, focusing initially on cocaine addiction. In Aim 2, we will validate these subtype-specific biomarkers by first replicating them in a new dataset and then evaluating their longitudinal stability and predictive utility. In Aim 3, we will test whether subtype-specific circuit mechanisms generalize to mediate iRISA functions in other forms of addiction, and define their interactions with distinct mechanisms mediating anhedonia and anxious arousal in patients with comorbid depression and anxiety.