全球约有8%-12%的人终身受抑郁症困扰(Kessler et al., 2013)，严重影响其生活质量。目前抑郁研究存在一些问题:症状分类无科学依据;有效基因位点无法识别;治疗方案有效性差。忽略抑郁的异质性可能是导致这些问题的原因之一。目前以量表总分作为潜变量抑郁水平的取向需要革新。网络分析作为一种候选方法被提出。网络分析通过构造抑郁多个症状的网络图像，可以明确症状夕间的相互影响樟式，并目找到核心疗状。
    本研究采用网络分析方法构建青少年抑郁症状网络图像，并探究不同基因型，不同环境下网络模式的差异。研究被试为北京双生子研究数据库中780名青少年，年龄在11-17岁之间(M=13.6,SD=1.8)，性别平衡。采用儿童抑郁量表(Children's Depression Inventory,  CDI)来测量抑郁症状;针对BDNF基因rs6265位点多态性进行基因分型，得到AA,  AG和GG三组变异;使用修订版的生活事件量表(Life Events Checklist)来测量环境指标。
    780名青少年的抑郁症状网络图像结果显示抑郁情绪是最核心症状，其次是自我评价和忧虑;3种不同基因型的抑郁症状网络图像结果显示存在差异，AA组网络图像稀疏，AG和GG组的网络图像聚集;AA组自我评价和忧虑中介能力最强，GG组抑郁情绪和疲惫感中介能力最强，AG组两者兼有;不同基因型在不同环境中的抑郁症状网络图像结果显示AG和GG组对环境变化更敏感，恶劣环境网络和普通环境网络存在显著差异，而 AA组对环境变化不敏感。
    研究结论为BDNF基因rs6265位点对青少年抑郁症状网络有显著影响，且AA组表现出的核心症状更偏认知症状;GG组和AG组表现出的核心症状更偏情绪和生理反应，且对环境变化敏感。网络分析作为一种新方法传递了更多信息，需进一步研究与规范。

     About 8% to 12% of people worldwide suffer from depression (Kessler et al.,2013), which seriously affects their quality of life. At present, there are some problems in the study of depression: lacking scientific basis for classification of symptoms; genes with modest effect could not be identified; treatment of depression is no better than the placebo effect. Ignoring the heterogeneity of depression may be a cause of these problems. Now sum-score of measurements is used to represent depression as a latent variable, which needs to be reformed. Network analysis now is a candidate method. By constructing the network graph of depressive symptoms, we can clearly identify the patterns of interaction between symptoms and find the core symptoms.
In this study, network graph of adolescent depressive symptoms was established by network analysis method, and the differences of network patterns in different genotypes and environments were explored. The study subjects were 780 adolescents in the Beijing Twin Study Database, aged between 11 and 17 years (M=13.6, SD=1 .8). The children's Depression Inventory (CDI) was used to measure the depressive symptoms. For the rs6265 polymorphism of the BDNF gene, the AA, AG and GG groups were genotyped. The revised life events scale (Life Events Checklist) was used to measure environmental indicators.
The network graphs of 780 adolescents showed that depressive emotion was the core symptom, followed by self-evaluation and worry. Three different genotypes of depressive symptoms showed differences in network graphs. AA group network graph was sparse, while AG and GG network graphs were dense. Self-evaluation and worry showed higher betweeness in AA group, while depressive emotion and fatigue showed higher betweeness in GG group. The network graphs of depressive symptoms of different genotypes in different environment showed that AG and GG group were more sensitive to environmental changes. There were significant differences between harsh environmental network and common environmental network. While the AA group was not sensitive to environmental changes.
    The results showed that the rs6265 polymorphism of BDNF gene had a significant effect on the network pattern of depressive symptoms of adolescents, and the core symptoms of AA group showed more cognitive symptoms. The core symptoms of GG group and AG group showed more emotional and physiological symptoms, which were more sensitive to environmental changes. Network analysis as a new method to convey more information and there needs further study and specification.