创伤后应激障碍与重性抑郁障碍特异性脑网络识别与量化研究

The post-traumatic stress disorder (PTSD) and major depressive disorder (MDD) are very common mental disorders after the experiencing the psychological trauma. It is difficult to differentiate these two disorders due to the high overlaps. However, these two disorders after trauma have their own characteristics in clinical feature, treatment and prognosis, which implicate that the different biological mechanism of the overlaps between PTSD and MDD. Unfortunately, the medicine used in these two disorders are same. The poor efficacy on patients may due to the lack of specific therapeutic targets of PTSD and MDD. Because of the lack of neuroimaging studies including both PTSD and MDD, it is difficult to differentiate them from neuroimaging information. This study will explore the abnormalities of brain network of PTSD and MDD together using the functional MRI and diffusion tensor imaging and investigate the distinct neural mechanism related to the overlaps. We try to find the biomarkers of intelligent diagnosis and specific treatment of PTSD and MDD by quantitative brain network indexes.

创伤后应激障碍（post-traumatic stress disorder，PTSD）和重性抑郁障碍（major depressive disorder, MDD）在个体经历重大应激事件后均十分常见，由于两者存在大量临床症状的重叠，因此在诊断上容易混淆。然而创伤后的PTSD和MDD的临床特征、治疗和预后又有各自的特点，这提示了这些重叠症状背后存在着不同的生物学机制。目前临床中用于治疗两种疾病的药物大致相同，缺乏各自特异的治疗靶点，治疗效果欠佳。由于缺乏同时纳入这两类疾病的影像学研究，目前尚不清楚它们重叠症状的生物学基础，也不能利用影像学信息将两者区分开。本研究创新性地运用功能磁共振和弥散张量成像技术同时探索PTSD患者和创伤后的MDD患者的脑网络异常，寻找这两者症状重叠的疾病背后不一样的脑机制，力图利用量化的脑网络指标为创伤后应激障碍和重性抑郁障碍的智能诊断和特异化治疗寻找生物学标记。

创伤后应激障碍（post-traumatic stress disorder，PTSD）和和重性抑郁障碍（major depressive disorder, MDD）是十分常见且严重影响患者社会心理功能和生存质量的精神障碍。PTSD仅在创伤事件后发生，而MDD在创伤事件后发病率也大幅提高，而两种疾病之间的共病特别普遍，但在临床中，PTSD和MDD治疗上有差异，且两者共病的患者症状更重，治疗效果和临床结局更差，存在更高的自杀风险。因此，识别PTSD和共病MDD的特异性脑网络特征，探索两者之间的脑网络差异，寻找这两种临床症状重叠的疾病背后不一样的脑机制，并用脑网络指标为PTSD的智能诊断和特异化治疗寻找生物学标记将有助于重大创伤后幸存者的客观诊断、评估和治疗决策。本项目采集了创伤后应激障碍患者和重性抑郁患者的脑网络，利用组间比较结合机器学习等技术探索了其诊断和治疗的生物学标记，发现PTSD患者存在特征性的脑网络异常，体现在发现默认网络、视觉网络、躯体运动网络、边缘网络和背侧注意网络内的连接和网络间连接异常，以及全脑网络的组织方式异常，其中枕叶视觉网络的“异化”可能是PTSD患者减轻症状的代偿机制，而PTSD共病MDD可以通过杏仁核亚区的功能连接差异来区分。本项目发现利用利用多变量机器学习技术结合脑影像学指标可以对PTSD的客观智能诊断、疗效预测，以及认知功能评估构建模型，为PTSD的风险预测和智能化诊疗决策指明了方向。