Self-injurious thoughts and behaviors (SITB) represent a pressing public health concern (World Health Organization [WHO], 2021). During adolescence, two particularly common types of SITB are suicidal ideation (SI), which refers to thoughts about ending one’s life (Hatkevich et al., 2019; Nock, 2010), and non-suicidal self-injury (NSSI), defined as the deliberate and direct destruction of one’s body tissue without suicidal intent (Nock, 2010). Suicide has become the second leading cause of death among youth (Centers for Disease Control and Prevention [CDC], 2021). According to surveys, the prevalence of SI among Chinese adolescents is 15.4 % (Chang et al., 2024). Similarly, the prevalence of NSSI among adolescents is significant, exceeding 22 % (Lang and Yao, 2018). Therefore, identifying risk factors for SITB in adolescents and developing corresponding interventions are crucial. Among these factors, depressive symptoms—common psychological symptoms in adolescents—have been widely recognized as shared risk factors for SITB (Turecki and Brent, 2016). This study thus examines the differential associations between specific depressive symptoms and NSSI or SI to provide more precise potential prevention targets for adolescent SITB.

Self-injurious behavior is non-suicidal, but it is strongly correlated and comorbid with SI. Researchers have found that NSSI positively predicts SI (Ribeiro et al., 2016), and SI, in turn, increases the risk of NSSI (Fox et al., 2015). Many studies have employed person-centered approaches (e.g., latent profile analysis or parallel latent class growth models) to explore the relationship between NSSI and SI (de Neve-Enthoven et al., 2024; Giletta et al., 2015; Shen et al., 2024), revealing significant overlap in their occurrence among adolescents. These findings suggest that NSSI and SI may share common risk factors or be simultaneously associated with a third variable (Cavanagh et al., 2003), with depressive symptoms playing a particularly critical role (Shen et al., 2024; Zhang et al., 2019).

Although NSSI and SI exhibit high comorbidity, they remain distinct in several important ways. By definition, NSSI lacks suicidal intent (Nock, 2010). According to the functional model of self-injury (Nock and Prinstein, 2004), NSSI serves to alleviate negative emotions or gain interpersonal support, with its primary function being the management and reduction of psychache. In contrast, SI is more strongly associated with feelings of hopelessness, social isolation, or a sense of meaninglessness, with its core rooted in the inability to tolerate psychache (Shneidman, 1993). These differences in characteristics may contribute to the varying associations between the two behaviors and specific depressive symptoms. However, prior research has predominantly investigated the relationships between depressive symptoms and NSSI or SI from a latent variable perspective, which may obscure the nuanced relationships between individual symptoms and SITB. There is a pressing need for methods capable of capturing the complex interrelations among symptoms and identifying the symptoms most critical for targeted intervention.

The method of network analysis, rooted in the network theory of psychopathology, utilizes graph theory to construct a relational network among observed variables, emphasizing the relationships between symptoms and overall network characteristics (Fried et al., 2017). In this framework, nodes represent variables (e.g., symptoms), and edges represent the statistical relationships between them. Network analysis provides indicators that describe associations and structures among observed variables, which traditional latent variable models often fail to capture (Bringmann et al., 2019). By identifying high-centrality nodes, researchers can uncover symptoms that activate and sustain the entire network. Furthermore, identifying edges with high weights between SITB and depressive symptoms allows for the determination of strong associations between specific variables.

Previous studies have extensively utilized network analysis to investigate the relationships between depressive symptoms and NSSI or SI. One cross-sectional study identified that, among depressive symptoms, only negative affect and negative self-esteem were associated with NSSI in adolescents (Lei et al., 2024). Another study exploring the comorbidity network of depressive symptoms and NSSI revealed that guilt played a pivotal role in bridging the two symptom clusters (Niu et al., 2024). Longitudinal evidence has also indicated that somatic symptoms and psychomotor retardation, together with depressed affect, are key symptoms contributing to the worsening of NSSI (Zhou et al., 2023). Regarding SI, researchers found that loneliness and sadness were the depressive symptoms most strongly associated with SI (Gijzen et al., 2021). Similarly, another study reported that sad mood and guilt were the depressive symptoms most closely linked to SI in adolescents (Xu et al., 2024).

However, most previous studies have not simultaneously included both NSSI and SI within a network, instead focusing solely on the associations between depressive symptoms and either NSSI or SI. This approach fails to fully reveal the potential differences in how depressive symptoms interact with these two phenomena. Although there is a study that has incorporated different forms of SITB into network analyses, it also primarily relied on cross-sectional designs (Wang et al., 2024; Jia et al., 2025). As a result, they are unable to adequately capture the longitudinal associations between depressive symptoms and SITB, thereby limiting insights into developmental interventions for the future.

In summary, addressing the limitations of previous studies, this study adopts a longitudinal design to construct and compare cross-sectional and longitudinal networks of depressive symptoms with NSSI and SI. The primary focus is to examine differences in specific depressive symptoms closely associated with NSSI and SI across the two types of networks. This approach aims to deepen our understanding of the interaction mechanisms between adolescent depressive symptoms and SITB. Furthermore, the findings can provide symptom-level scientific evidence to inform intervention strategies aimed at reducing the risk of SITB among adolescents. Based on existing theories of SITB and prior empirical findings, we hypothesized that SI and NSSI would share certain depressive symptoms (e.g., sad mood), but NSSI would be more strongly related to somatic symptoms, whereas SI would be more closely related to affective symptoms. These associations may differ between the mean and slope networks.