心理病理学网络理论、方法与挑战

对于精神障碍这一概念的理解, 传统DSM-ICD分类诊断系统和研究领域标准RDoC均基于潜变量视角, 认为精神障碍的症状由其潜在共同原因所致。这2种观点都忽略了症状间的相互作用。不同于分类和维度视角, Borsboom在2008年对精神障碍的概念化提出了的全新视角——心理病理学网络理论。此理论的核心观点是症状之间的动态因果关系构成了精神障碍。基于心理病理学网络理论的网络分析方法, 主要以结合EBIC的glasso算法估计症状间的偏相关网络, 并通过网络中节点中心性与网络连接性等指标, 来考查精神障碍症状的不同特性。近几年来, 研究者发现心理病理学网络分析方法在对症状间因果关系的推断、核心症状的识别和网络结构的可靠性与可重复性方面仍面临一些挑战。这些挑战为心理病理学网络理论与方法指明了未来可能的发展方向。

Psychopathological network theory,methods and challenges

As for the conceptualization of mental disorders, the traditional DSM-ICD classification diagnostic system, i.e., Diagnostic and Statistical Manual of Mental Disorders(DSM) and International Classification of Diseases(ICD), as well as the Research Domain Criteria (RDoC) proposed by National Institute of Mental Health(NIMH) are both based on the latent variable perspective, assuming that the symptoms of mental disorders have an underlying common cause(the disorder entity or dysfunction in different potential dimensions). However, such latent variable perspective requires local independency between variables, thus both views ignore the interaction between symptoms. In 2008, Borsboom put forward the psychopathological network theory, a new perspective different from the categorical and dimensional views of the conceptualization of mental disorders. This theory focuses on the interaction between symptoms, assuming that mental disorders are directly composed of symptoms and dynamic causal relationship between them. Based on this theory, network methods mainly estimate the partial correlation network of symptoms using the glasso algorithm with EBIC, and examine the different characteristics of mental disorder symptoms using indicators such as node centrality and network connectivity. In recent years, many new network models have emerged, such as Bayesian networks and relative importance networks that can perform causality inferences. With the increasing number of studies that applied psychopathological network theory and methods, this theory and method has clearly become one of the mainstream research theories and methods in the field of current mental health and psychopathological and psychometrics related research. But at the same time, researchers also found some remaining challenges for psychopathological network methods with respect to causality inference of symptoms, identification of central symptoms, and also reliability and replicability of network structures. Accordingly, this review briefly introduced the core idea and basic principles of psychopathology network theory, as well as the most commonly used psychopathology network analysis methods so far, and summarized important applications and values of psychopathology network theory and methods, then synthesized the main challenges that psychopathological network analysis method were currently facing. Finally, corresponding possible solutions were proposed. After reviewing a wide range of related publishments in theories, methods, and empirical since psychopathology network theory was put forward, we provided unique insights into the possible agendas for future research on psychopathological network methods, hoping the challenges and progress in the methodology could also bring new opportunities for the further improvement of psychopathological network theory.