Network analysis of childhood maltreatment and internet addiction in adolescents with major depressive disorder

Abstract

Background

Childhood maltreatment (CM) is closely linked to internet addiction (IA), especially in adolescents with Major Depressive Disorder (MDD). Previous studies have shown that adolescents who experience CM are more likely to develop IA and other psychological problems. This study aims to explore the complex relationship between CM and IA through network analysis, particularly identifying the core symptoms and bridge symptoms to better understand the development of IA in these adolescents.

Methods

A cross-sectional study was conducted in seven hospitals in Anhui Province, China, involving 332 adolescents diagnosed with MDD using DSM-5 criteria. The Childhood Trauma Questionnaire - Short Form (CTQ-SF) and the Internet Addiction Test (IAT) were used to assess CM and IA symptoms, respectively. Gender-based network analysis was also performed.

Results

Network analysis constructed a CM-IA network and identified core and bridge symptoms. “Depress/moody/nervous being offline”, “Request an extension for longer time”, “Sleep loss due to late-night logins”, and “emotional abuse” emerged as central symptoms within the CM-IA network. Additionally, “emotional abuse”, “sexual abuse”, and “complaints about online time” were identified as key bridge symptoms linking CM and IA. These symptoms demonstrated significant connectivity, underscoring their critical role in linking CM and IA.

Conclusion

The findings highlight the complex relationship between CM and IA in adolescents with MDD. Specific symptoms, such as emotional abuse and online-related symptoms, play important roles in understanding and intervening in adolescent IA. Future interventions should target these core and bridge symptoms for more effective prevention and treatment.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12888-024-06224-x.

Introduction

Adolescence is a critical period for rapid physical and brain maturation, accompanied by significant hormonal changes and neural restructuring [1, 2]. During this phase, adolescents are particularly susceptible to mental disorders, with depression being notably prominent [3] and becoming one of the most common mental disorders globally [4]. The “Global Burden of Disease Study 2020” [5] indicated that during adolescence, major depressive disorder (MDD) significantly increases the Disability-Adjusted Life Years (DALY), becoming a major influencing factor. In China, MDD has become one of the leading causes of disability and life loss [6]. Research indicates that adolescents with MDD are significantly more likely to experience incomplete education, with an odds ratio (OR) of 1.76 [7]. They are also at greater risk for Internet addiction (IA) [8], self-harm, and suicidal behavior [9], which further exacerbate their social and academic difficulties. Additionally, MDD in adolescents is strongly linked to depression, anxiety, and suicide later in life [10], highlighting the need for increased attention to adolescent mental health. As one of the challenges faced by adolescents with MDD, IA further worsens their social and academic outcomes.

In addition to MDD, adverse childhood experiences can also significantly impact adolescent behavior throughout their lifetime [11], particularly concerning internet addiction. Childhood maltreatment (CM) includes emotional, physical, and sexual abuse, as well as neglect, and has lasting effects on mental health throughout an individual’s life [12]. However, there is no standardized definition of CM, and different organizations, such as WHO and UNICEF, interpret it differently, which complicates research and intervention efforts [13]. Adolescents exposed to CM are at increased risk of developing behavioral issues like IA, as they often use internet engagement to cope with negative emotions [14]. CM has been linked to altered brain structures involved in threat detection, emotional regulation, and reward processing, as well as increased use of maladaptive emotion regulation strategies, which collectively heighten vulnerability to IA [15, 16]. Recent studies have found a significant association between childhood maltreatment and internet addiction, with maladaptive emotion regulation and peer support playing key roles in mediating and moderating this relationship [16, 17]. A meta-analysis further emphasized that different types of child abuse and study timing significantly impact this association, highlighting the need for tailored interventions [18].

Although IA lacks a universally accepted definition and is not formally recognized in modern psychiatric classification systems, it is generally characterized by excessive internet use, difficulty controlling online behavior, and significant functional impairment [19]. Young [20] categorized IA into several subtypes, including online gaming, communication, information seeking, and others. IA has been shown to impact adolescent brain development, impair decision-making, and contribute to issues like anxiety and depression [21–27]. Additionally, gender differences may influence the CM-IA relationship, with female adolescents more likely to develop emotional issues and males more prone to behavioral problems following abuse [28]. Investigating these gender differences can help develop more effective, gender-specific interventions.

Recently, several studies have explored the link between CM and IA. For example, a meta-analysis involving over 21,000 adolescents across 19 studies [29] found a significant positive correlation between CM and IA. Further analysis [30] showed that emotional neglect (OR = 3.062) and physical neglect (OR = 2.328) are independently associated with IA. Despite these findings, most existing studies rely on regression analysis, which captures only linear relationships and fails to illustrate the complex interactions between multiple factors. To address this gap, this study aims to provide a more comprehensive understanding of the complex relationships between different subtypes of childhood maltreatment CM and internet addiction IA in adolescents. Specifically, we use Network Analysis (NA) to explore how core symptoms of adolescent IA and subtypes of CM interrelate, with a particular focus on gender differences. By identifying key core and bridge symptoms, we seek to uncover potential intervention targets and provide insights into how these factors contribute to IA development in adolescents with MDD.

Methods

Study design and participants

A cross-sectional study was conducted in four general hospitals (Chaohu, Bengbu, Suzhou, Bozhou) and three psychiatric hospitals (Hefei, Ma’anshan, Fuyang) located in Anhui Province, China, between January and July 2021. Adolescent participants aged 12 to 18 were consecutively enlisted from psychiatric outpatients and inpatients in these hospitals. MDD was diagnosed by two board-certified psychiatrists using the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5) criteria [31], which included a comprehensive clinical evaluation and a structured interview. To ensure a standardized assessment, we also employed the Centre for Epidemiologic Studies Depression Scale (CES-D) to evaluate the severity of depressive symptoms in all participants.

Exclusion criteria included individuals with other psychiatric or neurological disorders, intellectual disabilities, or those who could not provide informed consent. Specifically, 14 participants were excluded following the second interview due to a diagnosis of bipolar disorder, 7 participants had incomplete questionnaire data, and 3 participants were diagnosed with organic brain disorders upon discharge. Out of 356 invited adolescents, 332 completed the assessment, resulting in a robust participation rate of 93%.

The study protocol (202009-kyxm-04) was approved by the Medical Ethics Committee of Chaohu Hospital of Anhui Medical University prior to commencement. All procedures were conducted in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. Written consent was obtained from all eligible participants and their guardians after briefing them about the study’s objectives and procedures.

Sociodemographic Information

Sociodemographic data, including age, gender, grade level, and only-child status, were gathered using structured questionnaires.

Measurement tools

The Childhood Trauma Questionnaire - Short Form (CTQ-SF)

The CTQ-SF is a widely used questionnaire designed to assess CM in individuals aged 12 and older [32]. It consists of 28 items, with 25 dedicated to evaluating various forms of maltreatment experienced during childhood, such as emotional abuse (EA), physical abuse (PA), sexual abuse (SA), emotional neglect (EN), and physical neglect (PN). The scoring for each item ranges from 1 (“never”) to 5 (“often”), and the total score for each subscale is the sum of scores for its five items, with higher scores indicating more severe abuse [33]. It is important to note that the scores for 7 items (item 2, 5, 7, 13, 19, 26, and 28) are to be reverse coded. As a result, the total score for each subscale ranges from 5 to 25. In previous research, the Chinese version of the CTQ-SF [34, 35] demonstrated commendable reliability, with Cronbach’s alpha values of 0.84 for emotional abuse, 0.79 for physical abuse, 0.86 for sexual abuse, 0.78 for emotional neglect, and 0.83 for physical neglect, and an overall scale reliability of 0.80 in our cohort.

The Internet Addiction Test (IAT)

The evaluation of IA utilized the authenticated Chinese version of the Internet Addiction Test (IAT), as reported by Chin and Leung [36] and Lai et al. [37]. This instrument consists of 20 questions and employs a 5-point Likert scale for respondents to assess the frequency of their internet-related symptoms, ranging from 1 (indicating very rarely) to 5 (indicating very often). The cumulative score from this test serves as an indicator of the severity of an individual’s internet addiction, with higher scores reflecting more severe addiction levels. The reliability of the Chinese version of the IAT in our study was excellent, with a Cronbach’s alpha of 0.93, indicating high internal consistency.

The Centre for Epidemiologic Studies Depression Scale (CES-D)

The Centre for Epidemiologic Studies Depression Scale (CES-D) was used to assess the severity of depressive symptoms in patients, with each item scored from 0 (“rarely or none of the time”) to 3 (“most or all of the time”) [38]. The CES-D consists of 20 items covering mood, somatic complaints, interactions, and positive affect, reflecting the frequency of symptoms over the past week. The total score ranges from 0 to 60, with higher scores indicating more severe symptoms. The CES-D has been widely validated in both clinical and community settings, demonstrating good reliability and validity. In our study, the Cronbach’s alpha for the CES-D was 0.85, indicating high internal consistency.

Network estimation

Data analysis in this study was performed using the R language within the RStudio environment, version 4.3.2 [39]. Network construction was accomplished through the EBICglasso function and Spearman correlation, creating a sparse Gaussian graphical model. Specifically, we employed an integrative network model that combines different types of CM—such as emotional abuse, physical abuse, and neglect—with core symptoms of IA. This integrative approach allowed us to evaluate multiple factors simultaneously, identifying the complex interactions between CM and IA symptoms, and uncovering core and bridge symptoms that are crucial for understanding the mechanisms linking CM to IA. This method, guided by the Extended Bayesian Information Criterion (EBIC), streamlined the network structure and enhanced interpretability [40]. The analysis primarily focused on strength and the expected influence (EI) index, due to their suitability for psychopathological networks [41]. Bridge Strength and the bridge expected influence (BEI) index were computed to identify bridge symptoms, utilizing the bridge function in the networktools R package [42]. Other centrality metrics such as closeness and betweenness were not used due to their limited effectiveness in revealing psychological variables [43].

Graphical representation of the network was generated using the qgraph R package (version 1.9.8) [44], where nodes (circles) were connected by edges (lines), with the thickness of the edges indicating interaction strength [45]. Positive correlations were depicted in blue, while negative correlations were shown in red. The Fruchterman-Reingold force-directed algorithm was used to cluster nodes with strong associations and place those with weaker associations on the periphery [46]. The R² predictability index, calculated using the mgm R package (1.2–14) [47], was visually represented by the size of a semi-circular area around each node, quantifying the variance each node explained in relation to others in the network.

Estimation of network accuracy and stability

The bootnet R package (version 1.5.6) [48] was used to conduct 1000 bootstrap iterations to assess the accuracy and stability of network edges, with edge precision evaluated by examining the 95% confidence interval of bootstrap edge weights; narrower intervals indicated higher accuracy. The robustness of centrality measures was further assessed by comparing centrality indices from the complete sample with those from a 70% reduced sample. The Centrality Stability coefficient (CS coefficient) was calculated to evaluate network robustness, with a value of ≥ 0.5 indicating high reliability, 0.25 to 0.5 denoting moderate reliability, and < 0.25 indicating less robustness.

Comparisons of network characteristics by gender

We performed gender-based network comparisons using the NetworkComparisonTest package (version 2.2.2) [49]. Effect sizes were determined using Bootstrap-based Spearman correlation, and the average correlation coefficients were reported from 1000 resampled associations. By comparing two networks and visualizing the differences in a corPlot diagram, we effectively identified key distinctions.

Results

Sociodemographic Characteristics of Participants

A total of 332 adolescents participated in the study, with an average age of 15.32 years (SD = 1.62). Among them, 242 (73%) were male and 90 (27%) were female. Regarding grade level, 148 participants (45%) were in middle school, while 184 (55%) were in high school. Additionally, 139 participants (42%) were only children, whereas 193 (58%) had siblings.

Network structure and centrality measures analysis

Figure 1 illustrates the network structure of CM-IA. Out of 300 edges in this network, 131 edges (44%) have non-zero weights, indicating a dense interconnectivity between CM and IA. In terms of node predictability, on average, 43% of the variance can be explained by neighboring nodes. The node “Preoccupation with the Internet” (IAT-15) has the highest predictability in the network, reaching 62%. This is closely followed by “Depress/moody/nervous being offline” (IAT-20) and “Request an extension for longer time” (IAT-16), with predictabilities of 61% and 58%, respectively. The predictability of all other nodes in the network is detailed in Table 2.

Fig. 1.

Predictive Network Model Depicting the Association between Childhood Maltreatment and Internet Addiction in Adolescents with Major Depressive Disorder (MDD)

Table 2.

Descriptive statistics of measurement items

Item abbreviation	Item content	Mean (SD)	Strength a	EI a	Predictability b
CTQ-SF
EA	Emotional abuse	11.20 (4.66)	1.03	1.03	0.43
PA	Physical abuse	7.37 (3.19)	0.56	0.56	0.28
SA	Sexual abuse	5.81 (2.17)	0.36	0.36	0.12
EN	Emotional neglect	15.60 (5.12)	0.75	0.75	0.43
PN	Physical neglect	10.55 (3.34)	0.63	0.63	0.39
IAT
IAT-1	Stay online longer	3.06 (1.36)	0.67	0.67	0.38
IAT-2	Neglect chores to spend more time online	2.74 (1.31)	0.97	0.97	0.49
IAT-3	Prefer the excitement online	2.90 (1.51)	0.86	0.86	0.50
IAT-4	Form new relationship	2.05 (1.22)	0.42	0.42	0.25
IAT-5	Others complain about your time	3.44 (1.37)	0.98	0.98	0.41
IAT-6	School grades suffer	2.46 (1.28)	0.89	0.89	0.46
IAT-7	Academic efficiency declines	2.11 (1.27)	0.65	0.65	0.31
IAT-8	Check email/SNS before doing things	1.94 (1.07)	0.80	0.80	0.38
IAT-9	Become defensive/secretive Internet use	2.65 (1.43)	0.46	0.46	0.17
IAT-10	Sooth disturbing thoughts	3.23 (1.46)	0.91	0.91	0.40
IAT-11	Anticipation for future online activities	2.52 (1.37)	1.01	1.01	0.56
IAT-12	Fear that is boring and empty without the Internet	2.81 (1.37)	0.89	0.89	0.48
IAT-13	Snap or act annoyed if bothered without being online	2.35 (1.20)	0.93	0.93	0.55
IAT-14	Sleep loss due to late-night logines	2.54 (1.39)	1.07	1.07	0.57
IAT-15	Preoccupation with the Internet	2.63 (1.30)	1.03	1.03	0.62
IAT-16	Request an extension for longer time	2.59 (1.41)	1.13	1.13	0.59
IAT-17	Failure to cut down the time spend online	2.21 (1.32)	0.86	0.86	0.48
IAT-18	Conceal the amount of time spend online	2.10 (1.32)	0.70	0.70	0.36
IAT-19	Spend more time online over going out with others	2.51 (1.48)	0.85	0.85	0.52
IAT-20	Depress/moody/nervous being offline.	2.12 (1.33)	1.14	1.14	0.61

Note: a The values of Strength and EI (Expected Influence) were raw data generated from the network; b These relationships align with the predictability measures obtained by R2, which are displayed as a bar on the edge of the node’s circle.

In the CM community, the connection between the nodes EN “Emotional neglect” and PN “Physical neglect” is the most direct and strong, with a weight of 0.40. The node EA “Emotional abuse” also has a significant direct connection with PA “Physical abuse”, with a weight of 0.38. Additionally, there is a notable link between EA and EN, with a weight of 0.25. In terms of IA symptoms, the nodes IAT-3 “Prefer the excitement online” and IAT-19 “Spend more time online over going out with others” are the most closely connected, with a weight of 0.46, representing the tightest bond in the entire network. Following closely, IAT-1 “Stay online longer” and IAT-2 “Neglect chores to spend more time online” have a close relationship with a weight of 0.37; and IAT-16 “Request an extension for longer time” and IAT-17 “Failure to cut down the time spend online” are also closely linked, with a weight of 0.28. Notably, the differences in these edge weights are statistically significant (refer to Figure S1). Detailed information about the weights of other edges in the CM-IA network can be found in Table S1.

As depicted in Fig. 2, within the CM-IA network, IAT-20 “Depress/moody/nervous being offline” emerges as the node with the highest centrality, followed by IAT-16 “Request an extension for longer time”, IAT-14 “Sleep loss due to late-night logines”, and EA “Emotional Abuse”. Additionally, SA “Sexual Abuse” holds the highest bridging value, succeeded by IAT-5 “Others complain about your time” and EA “Emotional Abuse”. Specifically, the strongest link is observed between EA “Emotional Abuse” and IAT-5 “Others complain about your time”. The centrality and bridge symptoms for all nodes are accessible in Table 1. Figure 3A illustrates satisfactory accuracy, evidenced by the convergence of the black and red lines. Furthermore, the narrow gray band indicates minimal variability during the resampling process. Regarding the stability of the network analysis, the centrality and bridging metrics exhibit excellent stability levels with CS = 0.75 and CS = 0.68, respectively, as shown in Fig. 3B.

Fig. 2.

Analysis of Centrality and Bridge Indices within the Network Structure Connecting Childhood Maltreatment and Internet Addiction. Note. EA: Emotional abuse, PA: Physical abuse, SA: Sexual abuse, EN: Emotional neglect, PN: Physical neglect, IAT-1: Stay online longer, IAT-2: Neglect chores to spend more time online, IAT-3: Prefer the excitement online, IAT-4: Form new relationship, IAT-5: Others complain about your time, IAT-6: School grades suffer, IAT-7: Academic efficiency declines, IAT-8: Check email/SNS before doing things, IAT-9: Become defensive/secretive Internet use, IAT-10: Sooth disturbing thoughts, IAT-11: Anticipation for future online activities, IAT-12: Fear that is boring and empty without the Internet, IAT-13: Snap or act annoyed if bothered without being online, IAT-14: Sleep loss due to late-night logines, IAT-15: Preoccupation with the Internet, IAT-16: Request an extension for longer time, IAT-17: Failure to cut down the time spend online, IAT-18: Conceal the amount of time spend online, IAT-19: Spend more time online over going out with others, IAT-20: Depress/moody/nervous being offline

Table 1.

Demographic characteristics (N = 332)

Variables	Mean (SD) or N (%)
Age (years)	15.32 (1.62)
Gender
Female	242 (73)
Male	90 (27)
Grade Level
Middle school	148 (45)
High school	184 (55)
Only child
Yes	139 (42)
No	193 (58)
Outcome measures
CESD total score a	36.86 (12.54)
IAT total score b	50.73 (16.68)
CTQ-SF total score c	50.50 (12.87)

>Note: ^a CESD: the Centre for Epidemiologic Studies of Depression Symptom Scale; ^b IAT: Internet Addiction Test; ^c CTQ-SF: The Childhood Trauma Questionnaire - Short Form.

Fig. 3.

Network robustness and accuracy

Gender differences

Figure S4 presents the network structures generated for both genders. When comparing the networks generated for females and males, no significant differences were found in terms of global network strength (females: 10.05 vs. males: 9.47; S = 0.57, p = .71) or the distribution of edge weights in the network structure (M = 0.28, p = .12).

Discussion

This study employs network analysis to explore the relationship between CM and IA among 332 adolescents aged 12 to 18 diagnosed with MDD. Participants were recruited from outpatient and inpatient units across multiple hospitals. Additionally, this study aimed to examine the influence of gender differences on the relationship between CM and IA, providing insights into gender-specific factors. The study reveals a dense network of connections between CM and IA, with 44% of the edges showing non-zero weights, indicating a close interaction between the two. This dense interconnectivity highlights the complex and multifaceted interactions between various aspects of CM and the symptoms of IA.

The CM-IA network identifies core symptoms including “Depress/moody/nervous being offline” (IAT-20), “Request an extension for longer time” (IAT-16), “Sleep loss due to late-night logines” (IAT-14), and “Preoccupation with the Internet” (IAT-15). These findings elaborate on the definition of IA as a behavioral disorder where individuals develop an excessive reliance on and lack of control over internet use, leading to significant impairment in psychological, social, academic, or occupational functioning [50]. Additionally, our findings align with previous network analysis studies on IA among Chinese adolescents, particularly in highlighting the importance of symptoms like “Preoccupation with the Internet” and “Request an extension for longer time spent online” as central components of the IA-depression model [51]. Previous research has indicated that adolescents with heightened IA tendencies often show signs of emotional abuse (EA) and a lack of emotional regulation, which can exacerbate depressive symptoms and internet dependency [51].

Previous studies have reported that when unable to access the internet, adolescents may experience negative emotional reactions such as depression, anxiety, or irritability [52, 53]. However, these emotional reactions were not directly assessed in the current study. These findings suggest that the internet can serve as both an escape from reality and a tool for managing emotions in certain contexts [54]. While such emotional regulation can be beneficial to some extent, excessive and compulsive internet use can lead to delayed sleep and disrupted sleep cycles, worsening the individual’s real-life situation. Our study also highlights “Emotional Abuse” (EA) as another central symptom in the CM-IA model. Emotional abuse can have long-lasting effects on an individual’s ability to regulate emotions and interpersonal relationships [55], undermine their basic sense of safety and trust [56, 57], and increase their tendency to escape reality and seek emotional fulfillment later in life [58]. This drives individuals to use the internet as a coping mechanism to avoid the hardships and emotional pain of real life [59].

Previous studies on preventing problematic internet use have indicated that emotional abuse (EA) can lead to severe psychological issues, including depression and anxiety ​ [60]. This finding supports the inclusion of EA as a central symptom in the CM-IA model. Given these effects, cognitive-behavioral therapy (CBT) has been shown to be effective in addressing emotional regulation challenges caused by EA, helping individuals develop healthier coping strategies and reducing their reliance on the internet.

In our integrative network model, we identified three key bridge symptoms that link IA with CM: Sexual Abuse (SA), Emotional Abuse (EA), and “Others complain about your time” (IAT-5). Notably, there is a significant connection between “Emotional Abuse” (EA) and “Others complain about your time” (IAT-5). Consistent with the findings of Taş [61], we observed a correlation between emotional abuse and IA. Furthermore, a study in South Korea by Kim et al. [62] indicated a positive correlation between sexual abuse and IA (β = 0.20). As covert forms of CM, sexual and emotional abuse can often be difficult to detect and may persist over time, leading to a relatively high incidence rate [63]. The challenges in identifying, defining, and legally substantiating emotional abuse increase the risk of children remaining in harmful environments [64]. These forms of abuse have a lasting negative impact on children’s psychological development, manifesting in heightened levels of depression, anxiety, stress, and neuroticism [65]. Particularly, sexual abuse profoundly affects children’s psychological state and can severely damage their self-esteem and identity [66]. Due to these psychological impacts, such as self-loathing, intense dislike for one’s body and emotions, and resultant social isolation, victims may seek solace and escape in the virtual world. The relative anonymity and control afforded by the internet provide these children with a means to distance themselves from the painful experiences of real life. In this space, children and adolescents may find the support, understanding, and acceptance they lack, temporarily alleviating their inner turmoil. However, prolonged internet use can evolve into a mechanism for escaping reality, ultimately leading to excessive dependence on the internet and the development of IA.

To address the link between CM and IA, comprehensive intervention measures should include cognitive-behavioral therapy to improve emotional regulation and coping strategies, particularly for individuals experiencing “Depress/moody/nervous being offline” (IAT-20). Prior research has demonstrated cognitive-behavioral therapy (CBT) has been shown to be effective in addressing addictive behaviors such as gaming disorder and unspecified internet use disorder, particularly by improving emotional regulation and coping strategies ​ [67]. Given that adolescents experiencing withdrawal from internet use often exhibit symptoms such as depression, anxiety, or irritability, CBT can serve as an appropriate intervention to mitigate these emotional challenges and prevent maladaptive coping behaviors.

Furthermore, for those affected by “Emotional Abuse” (EA) and “Sexual Abuse” (SA), trauma-focused therapy and trauma-informed care (TIC) should be provided to help them process past traumatic experiences. This aligns with our findings, which show a significant association between EA and IA symptoms. TIC aims to create a supportive environment that acknowledges the impact of trauma, reduces re-traumatization, and promotes emotional healing [68, 69]. Since EA can lead to emotional dysregulation and increase the risk of maladaptive coping mechanisms like excessive internet use, TIC can foster resilience and promote healthier coping strategies. Moreover, TIC can address social dysfunction resulting from emotional abuse, reducing the long-term psychological impact of CM [70].

The identification of “Emotional Abuse” and “Sexual abuse” as key bridge symptoms highlights the need for targeted interventions. Emotional dysregulation, closely associated with IA, suggests that cognitive-behavioral therapy (CBT) could help adolescents develop better emotional regulation skills, while trauma-focused therapy may reduce the influence of SA on IA development. Additionally, establishing supportive social networks, enhancing family communication, and educating individuals on recognizing and expressing emotions are crucial strategies for reducing internet dependency.

This multifaceted intervention approach aims to address IA by targeting underlying factors, such as emotional abuse, which has been linked to deficits in interpersonal competence and an increased risk of maladaptive coping mechanisms, including digital game addiction [61]. By focusing on these root causes, the intervention seeks to improve emotional regulation, enhance interpersonal skills, and reduce problematic internet behaviors.

The lack of significant gender differences in the CM-IA network could be due to the similar prevalence of emotional dysregulation and internet use patterns among male and female adolescents in this sample. Previous research has shown that both genders may develop IA as a coping mechanism, but the specific symptoms may vary in intensity [71]. One study found no significant gender differences in the relationship between childhood maltreatment and internet addiction, suggesting that this mechanism may have a similar impact on both genders [72]. Another study further showed that the total CTQ score and its subscale scores were positively correlated with IAT scores for both males and females, indicating consistency in this relationship [73]. Moreover, social support plays a crucial role in influencing coping mechanisms and reducing the risk of internet addiction. Research has shown that social support can effectively alleviate trauma-related symptoms, with similar effects for both genders [74]. Therefore, including social support as a key factor in prevention and intervention strategies may benefit all adolescents. Interventions aimed at enhancing social support have the potential to be effective across genders. Additionally, studying how social support influences the development of internet addiction could provide insights for creating more effective intervention strategies.

The limitations of this study warrant attention. Firstly, the observed network structure might be influenced by the specific survey tools used, and different assessment methods could yield varying results. Secondly, despite utilizing self-report tools with high reliability and validity, participant responses may have been compromised by recall bias, potentially affecting the objectivity of the findings. Furthermore, while we used the Childhood Trauma Questionnaire (CTQ) to assess childhood maltreatment, it is important to note that the psychometric properties of the CTQ have been validated among Chinese adolescents [75], which supports the reliability and validity of this tool in this population.

Furthermore, further research is needed to explore the specific role of gender in the relationship between CM and IA, particularly given the limitation of a small male sample size. Future studies should increase the sample size and ensure gender balance for a more comprehensive understanding of these dynamics.

Moreover, although we employed bootstrapping and predictability metrics to assess the robustness and stability of our network model, centrality scores may vary across different samples or studies. Recent literature [76] has highlighted that these scores may not always accurately capture the true influence of symptoms within a network, and the potential instability of centrality scores in complex networks may affect our conclusions regarding the influence of certain key symptoms. Therefore, while centrality analysis provides valuable insights into symptom relationships, these results should be interpreted with caution, particularly across different sample populations, cultural contexts, or clinical conditions. To address this issue, future research should explore alternative methods, such as replicating the study in diverse populations and using dynamic models, to validate and strengthen these findings. This will help establish a more comprehensive understanding of IA and inform more effective prevention strategies.

Conclusion

These findings underscore the need for early identification and targeted interventions for adolescents exposed to childhood maltreatment. Schools and healthcare providers should implement screening programs to identify high-risk individuals. Additionally, integrating cognitive-behavioral interventions and trauma-informed care within mental health services can provide adolescents with coping skills to reduce IA risk.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Author contributions

Huanzhong Liu and Lei Xia were responsible for the study’s design. Song Wang and Feng Geng conducted a comprehensive literature review. Mengyue Gu, Yudong Shi, and Jingyang Gu collected the data. Song Wang, Yating Yang, Ling Zhang, and Mengdie Li performed the data analysis and were responsible for data interpretation. Song Wang wrote the first draft of the paper. Huanzhong Liu critically revised the manuscript. All authors have reviewed and approved the final version of the manuscript.

Funding

A multimodal integrated study of the pathogenesis and clinical intervention of adolescent depression (2022zhyx-B01).

Data availability

The datasets generated and analyzed during the current study are not publicly available due to potential privacy concerns or academic implications, but can be obtained from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

The study protocol (202009-kyxm-04) was approved by the Medical Ethics Committee of Chaohu Hospital of Anhui Medical University prior to commencement. All eligible participants and their guardians provided informed consent after being briefed on the study’s objectives and procedures.

Consent for publication

All participants were given participant information prior to starting the survey and gave informed consent for the publication of the study’s results.

Competing interests

The authors declare no competing interests.