Multimodal brain-connectivity biomarkers for profiling heterogeneity in early psychosis

项目摘要 标题：用于分析早期精神病异质性的多模式脑连接生物标志物 精神障碍涉及复杂的结构和功能脑连接的功能障碍。但目前的 使用精神疾病诊断和统计手册诊断精神障碍的临床方法 疾病诊断与统计手册（DSM）通常未能根据生物异常对疾病进行分类。识别特定 个别患者的脑系统异常，特别是对于之前处于早期精神病（EP）阶段的患者 不可逆转的大脑改变发生的关键是开发更有效的早期干预方法。为此 目的，我们提出了一种创新的数据驱动的方法来表征大脑的异质性， 不同临床诊断类别的早期精神病患者的异常。创新和 将我们的方法应用于“人类早期精神病连接组项目”的两个受试者数据集 其中高质量的磁共振成像（MRI）数据和临床措施收集了320 患者和80名对照以及CIDAR项目的46名患者和37名对照。精神病的特征- 相关的大脑连接，我们提出了一种新的方法来整合我们的扩散MRI措施， 微观结构，如轴突密度，和我们的静息态功能磁共振成像测量的信息 流经轴突束然后，我们将应用一套系统设计的步骤，包括选择 脑连接特征，典型相关分析和交叉验证，以定义几个新的EP- 基于多模式大脑连接标记的网络。我们的方法将提供新的脑网络 以了解患者特异性异常。该项目的成果可以提供重要的大脑 开发更有效的个性化治疗方法的目标。

Project Summary Title: Multimodal brain-connectivity biomarkers for profiling heterogeneity in early psychosis Psychotic disorders involve dysfunction in complex structural and functional brain connectivity. But the current clinical approach for diagnosing psychotic disorders using the Diagnostic and Statistical Manual of Mental Illness (DSM) usually fails to categorize the diseases based on biological abnormalities. Identifying the specific abnormal brain system of the individual patient, especially for patients at the early psychosis (EP) stage before irreversible brain alterations take place, is key to develop more effective early intervention approaches. For this purpose, we propose to develop an innovative data-driven approach to characterize the heterogeneity of brain abnormalities in early psychosis patients across different clinical diagnostic categories. We will develop and apply our approach to two datasets of subjects from the “Human Connectome Project for Early Psychosis” where high-quality magnetic resonance imaging (MRI) data and clinical measures were collected from 320 patients and 80 controls and the CIDAR project with 46 patients and 37 controls. To characterize psychosis- related brain connectivity, we propose a novel approach to integrate our diffusion MRI measures on microscopic structures, such as axon density, and our resting-state functional MRI measure on the information flow through the axonal bundles. Then we will apply a systematically designed set of steps, including selecting brain connectivity features, canonical correlation analysis, and cross-validation, to define several novel EP- networks based on multimodal brain connectivity markers. Our approach will provide novel brain-network profiles to understand patient-specific abnormalities. Results from this project could provide important brain targets for developing more effective personalized treatment approaches.