Abstract: Psychological resilience, as a crucial capacity for maintaining mental health, has long faced challenges in its conceptual definition and operationalization. The recent process-oriented perspective views psychological resilience not as a fixed trait, but as a dynamic process involving continuous interactions between individuals and multi-systemic environmental factors. Complex dynamic systems theory provides a theoretical foundation for operationalizing this procedural nature of psychological resilience. Drawing on this framework, the present paper delineates two subsystems: the mental health state system, which is composed of various indicators of mental health state, and the influencing factors system, which consists of factors that affect mental health state. The influencing factors system can drive the mental health state system to undergo mere state fluctuations, or it can reshape the system's characteristics while simultaneously inducing state changes. Within this framework, the present paper defines psychological resilience as a system characteristic exhibited by the mental health state system under the drive of an influencing factors system, specifically encompassing two dimensions: stability and tendency toward health. High-level psychological resilience is characterized by a mental health state system that is both stable and inclined toward health. Network analysis serves as a powerful tool for characterizing this framework. In a network, elements are represented as nodes, which refer to the indicators within a system, and the connections between nodes are referred to as edges. By constructing a mental health state network based on indicators of the mental health state system, the parameters of this network can be used to quantify psychological resilience. In this paper, we propose this approach as the "network resilience paradigm." Current research often focuses the mental health state system on psychopathological symptoms, mapping the spontaneous activation tendency (threshold) of symptom networks onto health tendency and edge weights onto stability. A symptom network characterized by low spontaneous activation tendency (a tendency toward health) and weak edge weights is considered to reflect a higher level of symptom network resilience. This line of work has thereby formed a "symptom network resilience" paradigm as a specific instantiation of the broader network resilience approach. Based on the centrality hypothesis in network analysis, key mental health state indicators that determine symptom network resilience can be identified, either by their spontaneous activation tendency (reflecting health tendency) or by their edge weights (reflecting stability). Influencing factors of psychological resilience can also be incorporated into the psychopathological symptom network to form a hybrid symptom-and-resilience-factor network. This allows researchers to identify risk and protective factors for symptom network resilience. Furthermore, it enables researchers to examine the circular interactions between symptom network resilience and its influencing factors, for example, through the personality-resilience-psychopathology network, which considers the fast-changing process of psychopathological symptoms, the slow-changing process of personality, and the reciprocal influences between them. However, this field is still in its initial stage, with evidence largely confined to theoretical reasoning and simulation studies. The present paper discusses several key limitations and future directions in order of urgency. First, with respect to the quantitative accuracy of theoretical constructs, future research needs to develop statistical methods for quantifying individualized network resilience from multimodal data (physiological, environmental, and self-report) through intensive longitudinal designs, rather than relying on cross-sectional self-report data for group-level network estimations. Second, given the novelty of this paradigm, further work is needed to examine how network resilience compares with other resilience paradigms such as self-report-based psychological resilience. Third, the ultimate goal of this paradigm should be clinical practice and application. Future efforts should aim to achieve monitoring, early warning, and intervention for individualized network resilience with the aid of experience sampling methods and smart mobile devices. Finally, future prospects may also extend to resilience at the group level beyond the individual, as well as to alternative approaches to defining psychological resilience within the complex dynamic systems framework beyond network analysis (such as differential equation models).

Key words: psychological resilience, complex dynamic systems, psychopathological symptoms, network theory