Characterizing Schizophrenia Progression via Multi-modal Neuroimaging and Computation

 描述(申请人提供)：精神分裂症(SCZ)是一种导致严重认知障碍的致残性神经发育障碍。SCZ被认为是由影响大规模神经连接的突触障碍引起的。这一观点得到了神经成像研究的支持，这些研究反复显示前额叶皮质(PFC)功能和连接的变化，以及丘脑皮质和联合皮质回路的中断。然而，早期SCZ的复杂神经生物学仍然没有特征性，当干预至关重要时，限制了对早期疾病阶段的治疗。这是改进靶向治疗、预测预后和促进早期发现的主要目标。我们的首要目标是通过最先进的神经成像，纵向描述与认知缺陷相关的早期SCZ并发的功能和结构连接障碍。反过来，我们的目标是通过扩展到神经网络水平的基于生物物理学的计算模型来告知临床神经成像效应背后的突触假说。为了解决这些知识差距，我们将检查早期SCZ患者首次进入耶鲁大学专门治疗早期精神病(STEP)诊所后神经连接障碍的纵向进展。反过来，我们将纵向跟踪患者 在6个月、12个月和24个月后与50名匹配的健康对照组进行比较。为了量化连接障碍，该项目将使用由人类连接组项目(HCP)优化的领先功能和结构方法，与NIMH连接与人类疾病相关倡议保持一致。首先，我们的目标是测试最近发现的PFC和丘脑皮质标志物是否表现出同时(或可分离的)结构和功能改变。这一平衡的纵向设计可以区分早期病程中与临床相关变量相关的“状态”和“特征”神经影像标记物。具体地说，检查药物治疗的效果、治疗依从性、未经治疗的精神病持续时间和症状严重程度，为这些有希望的神经成像标记物的临床应用提供信息。其次，该项目将测试这些神经成像标记物是否与认知障碍的严重程度有关-这是SCZ的一个重要临床特征。我们的目标是通过我们验证的神经成像范式同时检查工作记忆(WM)，以测试结构和功能连接障碍的特定方面是否可以预测WM缺陷。这在SCZ的连接障碍和认知障碍之间提供了一个亟需的联系。最后，为了告知神经连接障碍背后的突触假设，如N-甲基-D-天冬氨酸受体(NMDAR)功能低下导致的皮质兴奋-抑制(E/I)失衡，我们旨在使用基于生物物理学的计算模型，纳入相关细胞细节。我们的目标是通过将电子效应与体内临床神经成像结果相匹配，反复探索控制E/I平衡的突触参数。这种计算精神病学方法可以通过计算拟合帮助解释SCZ的动态神经连接障碍，并为专注于SCZ早期阶段的治疗研究产生新的突触靶点，此时干预是最重要的。

 DESCRIPTION (provided by applicant): Schizophrenia (SCZ) is a disabling neurodevelopmental disorder causing profound cognitive impairment. SCZ is hypothesized to arise from synaptic disturbances affecting large-scale neural connectivity. This view is supported by neuroimaging studies that repeatedly show alterations in prefrontal cortex (PFC) function and connectivity and disruptions across thalamo-cortical and associative cortex circuits. However, the complex neurobiology of early-course SCZ remains uncharacterized, limiting treatments for early illness phases when intervention is crucial. This is a major objective for improving targeted therapies, predicting prognosis, and promoting early detection. Our overarching goal is to longitudinally characterize concurrent functional and structural dysconnectivity in early-course SCZ in relation to cognitive deficits via state-of-the-art neuroimaging. In turn, we aim to inform synaptic hypotheses underlying clinical neuroimaging effects via biophysically-based computational modeling scaled to the level of neural networks. To address these knowledge gaps, we will examine longitudinal progression of neural dysconnectivity in early-course SCZ patients after their initial admission into the Specialized Treatment Early in Psychosis (STEP) Clinic at Yale. In turn, we will follow patients longitudinally at 6, 12, and 24 months later in comparison with 50 matched healthy controls. To quantify dysconnectivity the project will use leading functional and structural methods optimized by the Human Connectome Project (HCP), in line with the NIMH Connectomes Related to Human Disease initiative. First, we aim to test if the recently identified PFC and thalamo-cortical markers exhibit concurrent (or dissociable) structural and functional alterations. This balanced longitudinal design can distinguish `state' versus `trait' neuroimaging markers during early illness course in relation to clinically-relevant variables. Specifically, examining effects of pharmacotherapy, treatment compliance, duration of untreated psychosis, and symptom severity, informs the clinical utility of these promising neuroimaging markers. Second, the project will test if these neuroimaging markers relate to severity of cognitive deficits - a hallmak clinical feature of SCZ. We aim to concurrently examine working memory (WM) via our validated neuroimaging paradigms to test if specific aspects of structural and functional dysconnectivity predict WM deficits. This provides a much-needed link between dysconnectivity and cognitive impairment in SCZ. Finally, to inform synaptic hypotheses behind neural dysconnectivity, such as cortical excitation-inhibition (E/I) imbalance resulting from hypo-function of the N-methyl-D-aspartate glutamate receptor (NMDAR), we aim to use biophysically-based computational models that incorporate relevant cellular detail. We aim to iteratively explore synaptic parameters governing E/I balance by fitting in silico effects with in vivo clinical neuroimaging findings. This computational psychiatry approach can help interpret dynamic neural dysconnectivity in SCZ via computational fits and yield new synaptic targets for treatment studies focused on early SCZ stages, when intervention is most vital.