BackgroundThe relationship between different dimensions of empathy and individual symptoms of depression during the COVID-19 pandemic remains unclear, despite the established link between empathy and depression. The network analysis offers a novel framework for visualizing the association between empathy and depression as a complex system consisting of interacting nodes. In this study, we investigated the nuanced associations between different dimensions of empathy and individual symptoms of depression using a network model during the pandemic.Methods1,177 students completed the Chinese version of the Interpersonal Reactivity Index (IRI), measuring dimensions of empathy, and the Chinese version of the Patient Health Questionnaire-9 (PHQ-9), measuring symptoms of depression. First, we investigated the nuanced associations between different dimensions of empathy and individual depressive symptoms. Then, we calculated the bridge expected influence to examine how different dimensions of empathy may activate or deactivate the symptoms of depression cluster. Finally, we conducted a network comparison test to explore whether network characteristics such as empathy-depression edges and bridge nodes differed between genders.ResultsFirst, our findings showed that personal distress was positively linked to symptoms of depression. These symptoms involved psychomotor agitation or retardation (edge weight = 0.18), sad mood (edge weight = 0.12), trouble with concentrating (edge weight = 0.11), and guilt (edge weight = 0.10). Perspective-taking was found to be negatively correlated with trouble with concentrating (edge weight = −0.11). Empathic concern was negatively associated with suicidal thoughts (edge weight = −0.10) and psychomotor agitation or retardation (edge weight = −0.08). Fantasy was not connected with any symptoms of depression. Second, personal distress and empathic concern were the most positive and negative influential nodes that bridged empathy and depression (values of bridge expected influence were 0.51 and −0.19 and values of predictability were 0.24 and 0.24, respectively). The estimates of the bridge expected influence on the nodes were adequately stable (correlation stability coefficient = 0.75). Finally, no sex differences in the studied network characteristics were observed.ConclusionsThis study applied network analysis to reveal potential pathways between different dimensions of empathy and individual symptoms of depression. The findings supported the existing theoretical system and contribute to the theoretical mechanism. We have also made efforts to suggest interventions and preventions based on personal distress and empathic concern, the two most important dimensions of empathy for depressive symptoms. These efforts may help Chinese university students to adopt better practical methods to overcome symptoms of depression during the COVID-19 pandemic.