Adolescence is a crucial time for the development of healthy mental well-being. In China, adolescents face significant pressures from traditional cultural values and intense societal competition. These pressures include academic demands, expectations from parents, and peer pressure, all contributing to psychological stress and a variety of mental health issues. Notably, depression and anxiety are prevalent, with their incidence rising in recent years. (Shorey et al., 2022; Racine et al., 2021). Zhang et al. (2021) highlighted that 25.6 % of participants exhibited depressive symptoms, 26.9 % displayed anxiety symptoms, and 20.6 % showed concurrent depression and anxiety symptoms. The co-occurrence of depression and anxiety is common during these years, leading to more severe symptoms, a higher suicide risk, and significant impairments in functioning and quality of life (Garber and Weersing, 2010; Kessler et al., 2005; Lamers et al., 2011; Van Loo et al., 2016; Hirschfeld, 2001). This underscores the urgent need for further research into the co-occurrence of depression and anxiety among Chinese adolescents.

The American Psychiatric Association (2013) highlighted the heterogeneity that exists within individuals diagnosed with depression and anxiety. Such heterogeneity suggests unique symptom presentations among diagnosed youth. Traditionally, research on depression-anxiety comorbidity has adopted a variable-oriented approach, which may neglect certain symptoms not covered by diagnostic criteria and miss critical clinical insights (Wang, 2018; Hou and Zhang, 2023). Latent Profile Analysis (LPA) offers a solution. Latent profile analysis (LPA) is characterized as a person-centered approach, aiming to uncover hidden population segments from continuous observed variables. This technique probabilistically categorizes participants into profiles based on similar responses (Wang and Hanges, 2011). This model-driven technique ensures objective cluster selection through rigorous statistical evaluation, surpassing traditional clustering methods in effectiveness (Magidson and Vermunt, 2002), leading to its increasing importance in classifying populations into cohesive subgroups. LPA is also adept at identifying common patterns of comorbidity (Collins and Lanza, 2009). For instance, Hou and Zhou (2021) utilized LPA to ascertain comorbidity patterns of Premenstrual Syndrome (PMS) and depression in Chinese female university students, revealing four distinct profiles. Similarly, Wang et al. (2021) investigated patterns of depression and anxiety in youth affected by the Lushan earthquake, revealing three distinct groups. Nonetheless, a significant research gap persists in exploring such subclassifications among school-going adolescents. Identifying distinct symptomatology profiles among Chinese adolescents is crucial for developing personalized therapeutic interventions. Thus, this study aims to examine the heterogeneity of anxiety and depression comorbidity in Chinese adolescents through LPA.

Negative cognition, characterized by a bias in cognitive processing, directs our focus towards negative stimuli, fostering their misinterpretation and enhanced recall (Cacioppo et al., 2014). This cognitive bias can worsen psychological health and exacerbate psychiatric disorders. The strong association between negative cognition and both depression and anxiety has been consistently demonstrated (Beevers et al., 2019; Nieto et al., 2020; Clark and Beck, 2010). Research into various aspects of negative cognitive biases-such as attention bias (quicker responses to negative stimuli), memory bias (better recall of negative information), and interpretation bias (negative readings of ambiguous scenarios)-highlights their significance in the development, persistence, and recurrence of depression and anxiety (Keller et al., 2019; Duyser et al., 2020; Hirsch et al., 2016; Everaert et al., 2017). Nevertheless, the impact of negative cognitive biases as predictors of specific anxiety-depression comorbidity profiles remains underexplored. Our study aims to address this gap by examining how three dimensions of negative cognitive bias can predict distinct subgroups characterized by depression and anxiety symptoms.

Cognitive biases significantly influence an individual's challenges with emotional regulation-the capacity to modulate emotions in reaction to emotional experiences, which in turn predisposes individuals to depression and anxiety (Joormann and D'Avanzato, 2010; Joormann and Quinn, 2014). Emotional regulation (ER) involves an individual's ability to recognize, assess, respond to, and manage their emotional expressions to achieve personal goals (Rosencrans et al., 2017). Effective ER enhances adaptability across various contexts, improving overall function and mental health (Gross, 2015). Numerous studies indicate that difficulties in emotional regulation are associated with negative outcomes, including depression and anxiety disorders (Everaert and Joormann, 2019; Ehring et al., 2010; Salters-Pedneault et al., 2006). Consequently, this study incorporates emotional regulation difficulties as a key predictor variable.

Furthermore, demographic factors, notably gender and age, have been identified as significant predictors of anxiety and depression. Specifically, evidence consistently indicates that women are more prone to experiencing depression and anxiety compared to men (Liu et al., 2019; Breslau et al., 2017; Gater et al., 1998). Therefore, this study incorporates gender and age as a third predictor variable.

To further substantiate the LPA classifications, this study's third objective was to investigate differences in specific Internet use patterns and learning-related challenges, such as Internet addiction, academic “Lying flat,” and involution, across the identified subgroups. Prior research has explored the link between Internet addiction and mental health issues, including depression and anxiety, revealing a significant association (Stanković and Nešić, 2022; Dong et al., 2020; Zhao et al., 2023; Xue et al., 2023). This study aims to discern if variations in Internet addiction scores exist among the defined subgroups, thereby verifying the precision of the classifications. Regarding academic challenges, while the negative impact of academic stress on anxiety and depression is well-documented (Deng et al., 2022; Quach et al., 2015; Kumaraswamy, 2013), the phenomena of “Lying flat” and involution have been less examined. “Lying flat,” or “Tang Ping” in Chinese, denotes a deliberate decision to reject undue exertion as a form of escapism (Zhou, 2022). Conversely, involution, or “neijuan,” describes the intense academic and occupational pressures faced by newer generations in China, showing a marked positive correlation with anxiety (Yi et al., 2022). However, the relationship between “Lying flat” behaviors and anxiety and depression remains unexplored. Consequently, assessing how different anxiety and depression subgroups vary in their experiences of academic “Lying flat” and involution is imperative.

In summary, this study posits three hypotheses: First, we expect to identify distinct profiles based on severity levels of depression and anxiety. Second, we hypothesize that female participants, characterized by significant negative cognitive biases and reduced emotional regulation abilities, will predominantly fall into the profile with higher levels of depression and anxiety. Third, we anticipate a strong positive correlation between the discussed Internet use and academic variables, and the severity of anxiety and depression among our adolescent cohort.