Connectomes-related to Active Methamphetamine-dependence Project (CAMP)

尽管有报道称，2000年代初甲基苯丙胺的使用量有所下降，但在美国，与精神兴奋剂相关的过量死亡人数（主要涉及甲基苯丙胺）从1999年至2017年增加了约1800%。目前，只有在发生灾难性事件（例如，过量用药、逮捕、失业）后，才能确定最需要干预的IDM。因此，在这些后果发生之前，迫切需要有客观的手段来识别有这种破坏性后果风险的IDM。以前的研究表明，依赖甲基苯丙胺（IDM）的个体的生物、心理和社会因素的独立测量与关键结果（如物质使用模式）之间存在相关性。然而，没有研究利用生物、心理和社会措施的综合力量来预测IDM的结果。最近成功地将神经成像（生物学）和社会心理测量与先进的机器学习技术相结合，以预测药物使用和其他疾病的治疗或诊断结果，为在IDM中取得类似成功奠定了先例。本项目旨在收集来自IDM （Aim 1）的第一个神经成像（磁共振成像[MRI]）、心理（如焦虑、抑郁和精神病症状）和社会（如人际暴力史、同伴网络药物使用）数据集。MRI数据将使用人类连接组项目中采用的最先进的序列来收集，所有数据将公开提供，使全球科学界有机会从这些独特的数据中获得无限的见解。在这里，数据将用于机器学习模型，以预测IDM的两个关键结果：6个月期间的物质使用模式和职业功能（目标2）。作为该项目的一部分开发的机器学习模型将导致未来的资助提案，重点是扩展纵向分析（例如，3-5年）并获得额外的参与者进行模型评估。这项研究的长期目标是发展预测模型，量化IDM改善或恶化成瘾和成瘾相关后果的可能性。通过这种方式，该项目将为客观的生物心理社会模型提供基础，这些模型旨在预测并最终预防IDM的更大危害。

Despite reports of declining methamphetamine use in the early 2000’s, psychostimulant-related overdose deaths in the US, of which methamphetamine is primarily involved, increased ~1,800% from 1999-2017. Currently, identifying IDM who are in the greatest need for intervention is only discovered after catastrophic events have occurred (e.g., overdose, arrest, job loss). Thus, there is an urgent necessity for objective means of identifying IDM at-risk for such devastating consequences, before these consequences can occur. Previous studies have shown correlations between independent measures of biological, psychological, and social factors and critical outcomes (e.g., substance use patterns) in individuals dependent upon methamphetamine (IDM). However, no research has leveraged the combined power of biological, psychological, and social measures to the predict outcomes in IDM. Recent success combining neuroimaging (biological) and psychosocial measures with advanced machine-learning techniques to predict treatment or diagnostic outcomes in substance use and other disorders establish a precedent for achieving similar success in IDM. This project seeks to collect the first ever neuroimaging (magnetic resonance imaging [MRI], psychological (e.g., anxiety, depression, and psychosis symptoms), and social (e.g., interpersonal violence histories, peer network drug use) dataset from IDM (Aim 1). MRI data will be collected using state-of-the art sequences adopted from the Human Connectome Project and all data will made openly available allowing for the global scientific community opportunities to gain limitless insights from these unique data. Here, data will be used in machine-learning models for the prediction of two critical outcomes in IDM: substance use patterns and occupational functioning over a 6-month time period (Aim 2). Machine-learning models developed as part of this project will result in future funding proposals focusing on extending longitudinal analyses (e.g., 3-5 years) and acquiring additional participants for model evaluation. The long-term goal of this research is the development of predictive models quantifying an IDM’s probability of improving or worsening addiction and addiction-related consequences. In this way, this project will provide the foundation for objective, biopsychosocial models tailored toward the prediction and eventual prevention of greater harms to IDM.