Due to the exponential development of mobile phone technology in recent decades, smartphones can now perform tasks that were traditionally exclusive to desktop computers and laptops. They enable us to engage in various online activities such as communication, video calls, online gaming, and shopping, almost anytime and anywhere (Liu et al., 2022). Smartphone usage has become a necessity in our daily lives. As of 2023, over 6.7 billion people worldwide used smartphones to access the internet, which accounts for more than 80 % of the global population (Taylor, 2023). University students quickly embraced smartphone technology, and almost all students have a smartphone. According to a 2021 survey, smartphone ownership in individuals with a college degree or higher significantly surpassed that in other demographic groups (Laricchia, 2021). Despite the benefits and convenience offered by smartphones, excessive smartphone usage characterized by compulsive behaviors can interfere with work, school, and interpersonal relationships, leading to problematic smartphone use (PSU) (Billieux et al., 2015).

Abundant evidence indicates that PSU is closely correlated with various mental health problems, such as depression, stress, and anxiety, and even physical health (Li et al., 2020, Wang et al., 2023). Of the mental health problems, depression has been the most widely studied, as revealed in a meta-analysis by our team (Li et al., 2020). Nevertheless, most previous studies have had a cross-sectional design, making it difficult to determine the causal direction of the relationship (Elhai et al., 2018; Li et al., 2020, Pourafshari et al., 2022, Vally et al., 2021, Wang et al., 2020). Therefore, it is necessary to explore the mechanisms underlying the relationship between depressive symptoms and PSU using a longitudinal design.

Rumination is defined as persistent and repetitive dwelling on one's negative emotions and, distress, and their origins and outcomes, often without taking active steps in coping strategies or problem-solving to alleviate the prevailing dysphoric mood (Nolen-Hoeksema, 1987, Treynor et al., 2003). The causal direction of the relationship between rumination and depressive symptoms remains controversial. Some scholars suggested that rumination might contribute to the development or even the aggravation of depression (Nolen-Hoeksema, 1987, Treynor et al., 2003). While others revealed that rumination might result from depression and that rumination serves as a maladaptive coping strategy in response to the experience of negative depressive affect (Watkins & Roberts, 2020).

In terms of the relationship between rumination and PSU, prior research indicated that rumination is not only a contributing factor but also an outcome of PSU ((Billieux et al., 2015; Davis, 2001, Lian et al., 2021). Rumination plays a role in driving PSU, as individuals who ruminate tend to dwell on their problems, regrets, or negative emotions, and participating in smartphone activities can temporarily alleviate this preoccupation. Furthermore, rumination often prompts individuals to seek validation and reassurance in the digital era, such as using smartphones to share their concerns on social media or perpetually checking for supportive messages (Billieux et al., 2015). Additionally, university students suffering from PSU are more likely to encounter academic and interpersonal difficulties (Celikkalp et al., 2020, Petrucco and Agostini, 2023), which further fuels the cycle of rumination.

Although the association between depressive symptoms and PSU has been extensively documented, the precise mechanisms underpinning this connection remain an ongoing focus of research. Given that depression may cause rumination, which leads to PSU and vice versa, it is plausible that depression is indirectly associated with PSU through rumination. Some previous studies supported a mediating role of rumination in the relationship between negative emotion (e.g., depression, anxiety, and stress) and PSU (Arrivillaga et al., 2022; Elhai et al., 2018; Lian et al., 2021, Liu et al., 2022, Peng et al., 2022, Vally et al., 2021, Wang et al., 2020). However, all of these studies were retrospective, limiting their ability to draw a causal inference. Additionally, the mediating effect of rumination in these studies was unidirectional (Arrivillaga et al., 2022; Elhai et al., 2018; Lian et al., 2021, Liu et al., 2022, Peng et al., 2022, Vally et al., 2021, Wang et al., 2020), and its bidirectional mediating role has never been examined. Therefore, we aimed to explore the potential bidirectional mediating effect of rumination on the relationship between depressive symptoms and PSU severity among medical students, using three-wave cross-lagged panel models (CLPMs) with longitudinal data.

There are two main theories closely related to our investigation: compensatory internet use theory (CIUT) (Kardefelt-Winther, 2014) and the interaction of person-affect-cognition-execution (I-PACE) model (Brand et al., 2019, Brand et al., 2016). CIUT theory posits that negative life events and stressors motivate excessive technology use, driven by the need to alleviate resulting negative emotions, such as depression (Kardefelt-Winther, 2014). CIUT emphasized negative life circumstances as the cause, and problematic internet use (e.g., PSU) as the consequence (Kardefelt-Winther, 2014). Thus, individuals who experience affective psychopathological disorders, such as depression, are more likely to be addicted to technology use to relieve their stress and negative emotions (Yuan et al., 2021). The theory has been previously supported by many empirical studies (Arrivillaga et al., 2022, Dempsey et al., 2019, Vally et al., 2021, Wang et al., 2020, Yuan et al., 2021).

I-PACE model involves a more comprehensive and specific model that incorporates numerous pathways relevant to problematic internet use (Brand et al., 2019, Brand et al., 2016). The first component, personal core characteristics, refers to predisposing factors such as genetic and biological influences, psychopathology, personality, and so on (Brand et al., 2019, Brand et al., 2016). The second component, affective and cognitive responses, encompasses internet-related cognitive and attention bias, expectancies about internet usage, coping style, and the urge to regulate mood (Brand et al., 2019, Brand et al., 2016). The third component includes executive functions, inhibitory control, and the decision to use certain applications (Brand et al., 2019, Brand et al., 2016). In the framework of the I-PACE model, depressive symptoms establish a vulnerable state, while rumination amplifies this vulnerability. As a maladaptive cognitive process, rumination can perpetuate negative emotions and stressors, thus increasing the likelihood of seeking escape or relief through problematic smartphone use. Consequently, the I-PACE model incorporates rumination as a mechanism through which depressive symptoms contribute to PSU. Many research on PSU supported the I-PACE model (Arrivillaga et al., 2022, Dempsey et al., 2019, Vally et al., 2021, Wang et al., 2020, Yuan et al., 2021).

The theoretical models mentioned above conceptualize psychopathology (e.g., depression) as a predictor of PSU, rather than the other way around, and have been typically tested and supported in the literature (Elhai et al., 2019). Nonetheless, there is an alternative possibility: PSU could exacerbate existing depression. This is especially true in cases where excessive smartphone use leads to factors such as social isolation, sleep disruption, increased stress, or decreased physical activity, all of which can negatively impact mental health. For instance, a previous study conducted by Zhao and Lapierre (2020) suggested that PSU might worsen depression symptoms through perceived stress (Zhao & Lapierre, 2020). The current study aimed to examine the underlying longitudinal and reciprocal mediating role of rumination in the relationship between depression and PSU among medical students. Therefore, we proposed four hypotheses based on the CLPMs, as outlined in Fig. 1.

Hypothesis 1

The autoregressive model fits well, and at T + 1 (a specific time point following the initial time 'T'), the values of depressive symptoms, rumination, and PSU significantly regress on their previous values at T. (Fig. 1A).

Hypothesis 2

Rumination longitudinally mediates the association between depressive symptoms and PSU (Fig. 1B).

Hypothesis 3

The mediating effect of rumination on the depression-PSU relationship maintains in the reverse direction (Fig. 1C).

Hypothesis 4

The rumination reciprocally mediates the depression-PSU relationship (Fig. 1D).