Network Analysis of Symptom-Level Interrelationships Between Adverse Childhood Experiences and Insomnia Symptoms in Adolescents: A Large-Sample Study

Abstract

Background

Sleep is essential for the physical and mental health of adolescents. Previous international studies have indicated a strong association between adverse childhood experiences (ACEs) and insomnia, although relevant reports focusing on Chinese adolescents remain limited.

Methods

This online cross-sectional study ultimately included 6670 adolescents aged 12–18 years. Childhood adversity and insomnia symptoms were assessed using the Adverse Childhood Experiences Questionnaire and the Insomnia Severity Index (ISI), respectively. Statistical analyses included chi-square tests to compare ACE prevalence between groups, binary logistic regression to identify risk factors for insomnia symptoms, and Network analysis explored interrelationships among specific ACEs and insomnia symptoms, using centrality indices (strength, closeness, betweenness) to evaluate symptom influence.

Results

The prevalence of ACEs was 47.6% and insomnia symptoms was 20.6%. ACEs showed a cumulative dose-effect relationship with insomnia. Adolescents with ACEs had a significantly higher incidence of insomnia than those without. Physical abuse (OR = 1.41), emotional neglect (OR = 1.32), social isolation or exclusion (OR = 2.52), and witnessing community violence (OR = 1.57) were identified as significant risk factors for insomnia symptoms. Network analysis showed that ACE10 (family crime, strength z = 2.084) and ACE12 (Isolated or excluded, betweenness z = 3.594) were the most central ACE nodes. The strongest cross-domain connections were ISI4 (sleep dissatisfaction) and ACE12 (edge weights = 0.081), and between ISI1 (difficulty falling asleep) and ACE12 (edge weights = 0.077).

Conclusion

ACEs and insomnia symptoms are prevalent and closely interrelated among Chinese adolescents. These findings highlight the need to strengthen the assessment of ACEs in this population and implement timely interventions to reduce the incidence of insomnia. However, as this is a cross-sectional study from a single province, further research is needed in terms of its generalizability.

Introduction

Adverse childhood experiences (ACEs) are defined as negative events that occur before the age of 18. These negative events often exceed their own coping ability, causing the individual to remain in a state of helplessness and anxiety for a long time and having an extremely adverse impact on the individual. These negative events mainly include: emotional abuse, being neglected, violence, experiences in traffic accidents, emergencies, encountering diseases, being bullied, separation, sexual abuse, etc.1,2 Research shows that approximately one-third of the global population had ACEs.3 The problem of adverse childhood adversity among adolescents is more prominent. Systematic reviews and meta-analyses have documented the occurrence rate of ACEs among the global children and adolescent population is 57.7%, with the highest being 78.9% in South America and Latin America, and relatively lower at 48% in Asia.4

ACEs can have lasting effects on adolescents’ physical and mental health, thereby increasing their risk of chronic diseases such as diabetes,5 cardiovascular diseases,6 and frequent headache.7 In terms of mental illness, meta-analyses have shown that ACEs significantly increase the risk of depressive disorder (OR = 2.50), non-suicidal self-harm (with emotional abuse OR = 2.91; sexual abuse OR = 2.72).8,9 Furthermore, ACEs are associated with the occurrence of anxiety disorders, Internet addiction, schizophrenia and bipolar disorder in adolescents.10–12

ACEs are closely related to sleep disorders in adolescents. Systematic reviews and meta-analyses indicate that children and adolescents with ACEs have a 3.91 times higher risk of insomnia symptoms, a 3.15 times higher risk of nightmares, and an average reduction of 12.1 minutes in sleep duration.13 ACEs can also affect the circadian rhythm disorder of adolescents, including increased social jet lag and difficulty falling asleep.14 The impact of ACEs on sleep health persists throughout the lifespan. A 22-year longitudinal study shows that specific ACEs such as physical abuse, parental imprisonment, emotional abuse and neglect, foster home placement and parental alcohol abuse, and community violence are significantly associated with persistent insomnia symptoms from adolescence to middle age.15

The above research indicates that ACEs were widespread among adolescents, with a relatively high incidence rate, and it increases the risk of physical and mental diseases. Our previous research has shown that ACEs may lead to the occurrence of bruxism by increasing perceived stress.16 However, the relationship between ACEs and adolescent insomnia symptoms remains underexplored, particularly in non-Western samples. The Chinese adolescent group may have cultural specificity that could uniquely shape the ACE-insomnia relationship. Research shows that Chinese parents are more likely to scold and physically punish their children for poor academic performance. In Chinese culture, corporal punishment is regarded as “normal” and even a form of “loving parenting”, which leads to more emotional insomnia problems among teenagers.17 Furthermore, due to factors such as economic income, many Chinese parents go to work elsewhere, resulting in a lack of family support and companionship. Especially in rural areas, there are often left-behind children, which can further aggravate the insomnia symptoms of adolescent.18

Importantly, traditional regression methods, while useful for identifying risk factors, assume linear relationships between aggregated ACE scores and insomnia outcomes and cannot capture the complex interrelationships among individual ACE types and specific insomnia symptoms. Network analysis offers a methodological advantage by conceptualizing symptoms as interacting nodes, allowing us to identify which specific ACEs are most centrally connected to particular insomnia symptoms and to detect bridge nodes that may serve as intervention targets. Based on this, we propose the following hypotheses: (1) ACEs and insomnia symptoms will show a cumulative dose-effect association in Chinese adolescents; (2) specific ACE types will be more strongly connected to insomnia symptoms than others; and (3) particular insomnia symptoms will serve as bridge nodes linking ACEs to insomnia. Therefore, this study adopted a large-sample cross-sectional survey to investigate the incidence of ACEs and insomnia symptoms among Chinese adolescents, using a combination of regression analysis and network analysis to reveal the risk factors and network associations of ACEs on insomnia symptoms and identify key intervention targets.

Materials and Methods

Study Population and Design

This online cross-sectional survey investigated adverse childhood experiences and insomnia symptoms among adolescents aged 12–18 years. Using convenience sampling, we recruited participants mainly from Shandong Province, China, between May 1 and May 30, 2022. A psychology teacher used the Wenjuanxing platform (https://www.wjx.cn/app/survey.aspx) to distribute the questionnaire via WeChat and QQ. A total of 7,794 adolescents completed the survey. After excluding those who did not complete the survey and invalid answers, 6,670 adolescents were included in the study. The inclusion criteria are included: (1) Junior and senior high school students; (2) access to QQ or WeChat users; (3) Voluntary participation; (4) ability to read and understand Chinese questionnaires.

This study was approved by the Medical Ethics Committee of Beijing Huilongguan Hospital (Approval No. 2021–18- Department), and all participants and guardians signed electronic informed consent forms before participating in the study.

Evaluation Methods and Quality Control

The questionnaire was created using the Wenjuanxing platform, and a QR code was generated. School teachers distributed the QR code to class-specific group chats via social media (WeChat, QQ). Prior to the survey, the responsible teachers received training on how to explain the completion requirements to students and provide on-site clarification. Participants scanned the QR code using their mobile devices, completed the questionnaire independently, and submitted their responses online. After data collection, quality control personnel screened all responses. Any response meeting one or more of the following criteria was defined as invalid and excluded: duplicate submissions, logically inconsistent answers, incomplete responses, completion time less than 5 minutes, or patterned/careless responding (selecting the same option for all items). Missing data were handled using listwise deletion.

Assessment Tools and Procedure

Sociodemographic Factors

Demographic data included gender, age, grade, whether they were only children, whether they were left-behind children, and academic performance.

Adverse Childhood Experiences

We measured ACEs using the 14-item Chinese version of the Revised Adverse Childhood Experiences Questionnaire (ACEQ-R).19 There are 14 items in the scale, and each item is scored as “0” if it does not appear, and “1” if it does. The total score of the 14 items is accumulated and added to obtain the total score of the individual ACEQ-R. The higher the score, the greater the variety of negative experiences. The results show that the Chinese version of ACEQ-R has good reliability and validity.20 The Cronbach’s α of ACEQ-R in this study was 0.789.

Sleep Quality

To assess sleep quality, participants completed the Insomnia Severity Index (ISI), a widely used 7-item self-report measure of insomnia severity and daily impairment.21 Responses on each item range from 0 to 4, yielding a cumulative score of 0 to 28, with elevated scores indicating greater insomnia severity. Based on total scores, insomnia was categorized as: none (0–7), mild (8–14), moderate (15–21), or severe (22–28). The ISI possesses good reliability and validity; In this study, Cronbach’s α coefficient was 0.877.

Statistical Analysis

We used IBM SPSS version 26.0 (Chicago, IL, USA) for data management and R version 4.3.1 for network analysis. Categorical data were summarized using frequencies and percentages, compared via chi-square tests. Continuous variables were described as means with standard deviations. To determine whether ACEs predicted insomnia symptoms, we fitted multiple logistic regression models, and to adjust for demographic factors such as gender, age, grade, whether they were only children, whether they were left-behind children, and academic performance. Multicollinearity was assessed using the variance inflation factor (VIF). A VIF value greater than 10 indicates the presence of problematic multicollinearity.22 In this study, all VIF values were below 5.046, suggesting that multicollinearity was not a serious concern.

Network analysis was conducted using R 4.3.1 (bootnet v1.5, qgraph v1.9, networktools). The network comprised 21 nodes (eg, 14 ACE items and 7 ISI items). ACE items were binary (0/1) and ISI items were ordinal (0–4). A sparse Gaussian Graphical Model (GGM) was estimated using the EBICglasso method with cor_auto for automatic selection of correlations (polychoric for ordinal/binary data, Pearson for continuous) and a tuning parameter γ = 0.5.23,24 The EBICglasso procedure automatically selects the optimal LASSO penalty parameter λ by evaluating a path of 100 λ values logarithmically spaced between λ_max (the smallest value that shrinks all partial correlations to zero) and λ_max/100, and the final network is selected as the one that minimizes the EBIC. Under this regularization approach, small partial correlations are shrunk to exactly zero, retaining only edges with non-zero regularized coefficients. Centrality indices (strength, closeness, betweenness) were standardized to z-scores.25 Nonparametric bootstrap (1,000 resamples) estimated edge weight accuracy with 95% confidence intervals. Centrality stability was evaluated using case-dropping bootstrap (1,000 iterations), yielding CS-coefficients (CS > 0.25 = acceptable; CS > 0.50 = good).26 Edge and node difference tests were conducted using differenceTest() within bootnet.

Results

Common Method Bias Test

Harman’s one-factor test revealed three factors with eigenvalues >1, the first accounting for 34.60% of the variance, below the 40% critical value, indicating that common method bias does not seriously threaten the findings.

Socio-Demographic Characteristics and Association with Insomnia Symptoms

This study included 6,670 adolescents (mean age 15.84 ± 1.39 years; 46.7% male, 53.4% female). The overall prevalence of insomnia symptoms was 20.7% (n = 1,381). Higher rates were observed among females than males (22.6% vs 18.5%), among senior high school students than junior high school students (25.8% vs 11.7%), and among left-behind children than their non-left-behind peers (23.1% vs 20.3%). No significant difference was found between only children and non-only children (20.7% vs 20.6%). Regarding academic performance, a graded association emerged: the poorer the academic performance, the higher the insomnia prevalence (26.5% for low, 18.9% for medium, and 16.6% for high performance)(Table 1).

Table 1.

Socio-Demographic Characteristics and Associations with Insomnia Symptom (N = 6670)

Variables	n	%	Insomnia Symptoms
			n	%	χ2	Phi	P
Gender					16.545	0.050	<0.001
Male	3116	46.7	578	18.5
Female	3554	53.3	803	22.6
Only child					0.004	0.001	0.950
No	5865	87.9	1215	20.7
Yes	805	12.1	166	20.6
Grade					187.706	0.165	<0.001
Junior high school	2409	36.1	281	11.7
Senior high school	4261	63.9	1100	25.8
Left behind children					5.682	0.029	0.017
No	5358	80.3	1078	20.1
Yes	1312	19.7	303	23.1
Exam results					66.860	0.100	<0.001
Bad	2118	31.8	561	26.5
Moderate	2818	42.2	533	18.9
Good	1762	26.0	287	16.6

The Detection Rate of Insomnia Symptoms in Different Adverse Childhood Experience Exposures

The detection rate of ACEs in adolescents was 47.6%, among which the most common was emotional abuse (28.5%), followed by physical abuse, emotional neglect (19.2%), isolated or excluded (18.9%), and witnessed community violence was the lowest (1.8%). 17.1% of adolescents had 1 ACE, 18.3% had 2–3ACEs, and 12.2% had ≥4 ACEs. The detection rate of insomnia symptoms in adolescents with different types of ACEs exposure was higher than that in adolescents without exposure (P < 0.001). With the increase of ACEs, the detection rate of insomnia symptoms also showed an increasing trend, and the detection rate of insomnia symptoms in adolescents with more than 4 ACEs was 46.0% (Figure 1 and Table 2).

Figure 1.

The number of cases of adverse childhood traumas and adversities experienced by adolescents.

Table 2.

The Incidence of Insomnia Symptoms Among Adolescents with and without Adverse Childhood Experiences (N = 6670)

Variables	n	%	Insomnia Symptoms
			n	%	χ2	Phi	P
ACE					532.505	0.272	<0.001
No	3496	52.4	424	12.1
1	1140	17.1	222	19.5
2–3	1223	18.3	362	29.6
≥4	811	12.2	373	46.0
Emotional abuse					257.368	0.193	<0.001
No	4770	71.5	748	15.7
Yes	1990	28.5	633	33.3
Physical abuse					222.048	0.179	<0.001
No	5387	80.8	921	17.1
Yes	1283	19.2	460	35.9
Sexual abuse					65.574	0.099	<0.001
No	6409	96.1	1275	19.9
Yes	261	3.9	106	40.6
Emotional neglect					277.238	0.200	<0.001
No	5390	80.8	899	16.7
Yes	1280	19.2	482	37.7
Physical neglect					142.005	0.144	<0.001
No	6246	93.6	1197	19.2
Yes	424	6.4	184	43.4
Parental separation and divorce					49.056	0.085	<0.001
No	6367	95.4	1270	19.9
Yes	303	4.5	111	36.6
Domestic violence					93.911	0.118	<0.001
No	6310	94.6	1234	19.6
Yes	360	5.4	147	40.8
Family substance abuse					13.137	0.044	<0.001
No	6485	97.2	1323	20.4
Yes	185	2.8	58	31.4
Family history of mental illness					38.271	0.076	<0.001
No	6222	93.3	1237	19.9
Yes	448	6.7	144	32.1
Family crime					13.719	0.045	<0.001
No	6436	96.5	1310	20.4
Yes	234	3.5	71	30.3
Peer bullying or victimization					73.982	0.105	<0.001
No	6357	95.3	1256	19.8
Yes	313	4.7	125	39.9
Isolated or excluded					478.727	0.259	<0.001
No	5411	81.1	837	15.5
Yes	1259	18.9	544	43.2
Witness community violence					60.164	0.095	<0.001
No	6547	98.2	1321	20.2
Yes	123	1.8	60	48.8
Low socioeconomic status					41.324	0.078	<0.001
No	6327	94.9	1263	20.0
Yes	343	5.1	118	34.4

Effects of Types of Adverse Childhood Experiences on Insomnia Symptoms in Adolescents

Adjusting for demographics, we conducted binary logistic regressions. Among ACEs, isolation or exclusion showed the strongest association with adolescent insomnia (OR = 2.52, 95% CI = 2.13–2.99), followed by witnessing community violence (OR = 1.57, 95% CI = 1.03–2.39), emotional abuse (OR = 1.51, 95% CI = 1.29–1.77), physical abuse (OR = 1.41, 95% CI = 1.19–1.68), and emotional neglect (OR = 1.32, 95% CI = 1.10–1.59). Demographic risks included senior high school status (OR = 1.90, 95% CI = 1.52–2.36) and older age (OR = 1.20, 95% CI = 1.11–1.29). Better academic performance was protective (moderate: OR = 0.74; good: OR = 0.62) relative to poor performance (Table 3).

Table 3.

The Association Between Adverse Childhood Experiences and Insomnia Symptoms in Adolescents (N = 6670)

	β	SE	OR	95% CI	P	β	SE	OR	95% CI	P
Emotional abuse	0.39	0.08	1.47	1.26–1.72	<0.001	0.41	0.08	1.51	1.29–1.77	<0.001
Physical abuse	0.28	0.09	1.32	1.11–1.56	0.001	0.34	0.09	1.41	1.19–1.68	<0.001
Sexual abuse	0.26	0.15	1.30	0.97–1.75	0.081	0.20	0.16	1.23	0.91–1.67	0.188
Emotional neglect	0.25	0.09	1.28	1.07–1.53	0.008	0.28	0.09	1.32	1.10–1.59	0.003
Physical neglect	0.14	0.13	1.15	0.89–1.48	0.285	0.16	0.13	1.17	0.90–1.52	0.246
Parental separation and divorce	0.03	0.15	1.03	0.77–1.39	0.840	0.04	0.16	1.04	0.77–1.41	0.806
Domestic violence	0.03	0.14	0.84	0.79–1.35	0.839	0.01	0.14	1.01	0.77–1.33	0.945
Family substance abuse	−0.16	0.14	0.85	0.57–1.27	0.423	−0.23	0.21	0.80	0.53–1.19	0.796
Family history of mental illness	0.23	0.13	1.25	0.96–1.63	0.093	0.15	0.14	1.16	0.89–1.52	0.269
Family crime	−0.29	0.19	0.75	0.52–1.09	0.135	−0.34	0.20	0.72	0.49–1.04	0.082
Peer bullying or victimization	−0.11	0.15	0.89	0.67–1.19	0.449	−0.01	0.99	0.99	0.73–1.33	0.937
Isolated or excluded	0.98	0.09	2.66	2.25–3.14	<0.001	0.93	0.09	2.52	2.13–2.99	<0.001
Witness community violence	0.48	0.21	1.62	1.07–2.44	0.022	0.45	0.22	1.57	1.03–2.39	0.038
Low socioeconomic status	0.12	0.14	1.13	0.86–1.48	0.373	0.05	0.14	1.05	0.80–1.38	0.742
Gender						0.09	0.07	1.08	0.95–1.24	0.230
Age						0.18	0.04	1.20	1.11–1.29	<0.001
Only child						0.10	0.10	1.11	0.91–1.35	0.313
Left behind children						−0.12	0.08	0.89	0.76–1.04	0.129
Senior high school						0.64	0.11	1.90	1.52–2.36	<0.001
Academic performance
Moderate						−0.30	0.07	0.74	0.64–0.86	<0.001
Good						−0.38	0.09	0.62	0.58–0.81	<0.001

Effect of Cumulative Number of Adverse Childhood Experiences on Insomnia Symptoms in Adolescents

After adjusting for demographic factors, multiple binary logistic regression analyses showed the cumulative number of adverse childhood experiences was a risk factor for insomnia symptoms, and the risk of developing insomnia symptoms increased with the increase in the cumulative number (ACEs = 1, OR = 1.51, 95% CI = 1.37–1.97; ACEs = 2-3, OR = 2.99, 95% CI = 2.53–3.52; ACEs = 4, OR = 6.22, 95% CI = 5.24–7.46) (Table 4).

Table 4.

The Association Between the Number of Adverse Childhood Experiences and Insomnia Symptoms in Adolescents (N = 6670)

	β	SE	OR	95% CI	P	β	SE	OR	95% CI	P
1 ACE	0.56	0.09	1.75	1.47–2.10	<0.001	0.50	0.09	1.51	1.37–1.97	<0.001
2–3 ACEs	1.11	0.08	3.05	2.60–3.57	<0.001	1.09	0.08	2.99	2.53–3.52	<0.001
≥4 ACEs	1.82	0.09	6.17	5.20–7.32	<0.001	1.83	0.09	6.24	5.22–7.46	<0.001
Gender						0.18	0.07	1.20	1.05–1.36	0.007
Age						0.19	0.04	1.20	1.12–1.30	<0.001
Only child						0.10	0.10	1.11	0.91–1.35	0.313
Left behind children						−0.08	0.08	0.92	0.79–1.08	0.303
Senior high school						0.60	0.11	1.83	1.48–2.27	<0.001
Exam results
Moderate						−0.30	0.07	0.74	0.64–0.85	<0.001
Good						−0.38	0.09	0.68	0.58–0.81	<0.001

Network Analysis Between Adverse Childhood Experiences and Insomnia Symptoms

The centrality index of network analysis showed that among adverse childhood experiences, ACEI10 (Family crime) had the highest strength (z = 2.084,95% CI = 0.489–3.679), ACEI12 (Isolated or excluded) had the highest betweenness (z = 3.594, 95% CI = 1.928–5.260) and closeness centrality (z = 2.026,95% CI = 0.584–3.468). Nodes ACEI9 (Family history of mental illness) and ACEI10 were most direct connection (Edge Weight = 0.477), followed by ACEI4 (Emotional neglect) and ACEI5 (Physical neglect) (Edge Weight = 0.450). Meanwhile, in the comparison of node strength differences, ACE10, ACE12, and ACE4 were significantly higher than the other nodes (Supplementary Figure 1). Similarly, in the group with insomnia symptoms, ISI7 (Distress caused by sleep difficulties) exhibited the highest strength (z = 0.030, 95% CI = −0.866–0.926), ISI4 exhibited highest closeness centrality (z = −0.592, 95% CI = −2.038–0.853) and ISI1 (Difficulty in falling asleep) had the highest betweenness (z = 0.426, 95% CI = −0.704–1.556). However, the 95% bootstrap confidence intervals for all three centrality indices included zero, suggesting that these estimates should be interpreted with caution and do not provide strong statistical evidence for these nodes being unequivocally more central than others. Nodes ISI1 and ISI2 (Sleep maintenance) were most directly connected (Edge Weight = 0.488), followed by nodes ISI5 (Interference with daily life) and ISI6 (Noticeability of sleep problems by others) (Edge Weight = 0.386). Among different symptoms, nodes ISI4 and ACEI12 had the most direct connection (Edge Weight = 0.081), followed by nodes ISI1 and ACEI12 (Edge Weight = 0.077). (Figures 2 and 3).

Figure 2.

Network analysis model of adverse childhood experiences and insomnia symptoms in adolescents.

Abbreviations: ISI1, Difficulty in falling asleep; ISI2, Sleep maintenance; ISI3, Waking up early; ISI4, Sleep dissatisfaction; ISI5, Interference with daily life; ISI6, Noticeability of sleep problems by others; ISI7, Distress caused by sleep difficulties; ACE1, Emotional abuse; ACE2, Physical abuse; ACE3, Sexual abuse; ACE4, Emotional neglect; ACE5, Physical neglect; ACE6, Parental separation and divorce; ACE7, Domestic violence; ACE8, Family substance abuse; ACE9, Family history of mental illness; ACE10, Family crime; ACE11, Peer bullying or victimization; ACE12, Isolated or excluded; ACE13, Witness community violence; ACE14, Low socioeconomic status.

Figure 3.

Central indicators of each symptom in adolescents.

Abbreviations: ISI1, Difficulty in falling asleep; ISI2, Sleep maintenance; ISI3, Waking up early; ISI4, Sleep dissatisfaction; ISI5, Interference with daily life; ISI6, Noticeability of sleep problems by others; ISI7, Distress caused by sleep difficulties; ACE1, Emotional abuse; ACE2, Physical abuse; ACE3, Sexual abuse; ACE4, Emotional neglect; ACE5, Physical neglect; ACE6: Parental.

For the accuracy test of edges, each line on the Y-axis represents an edge between nodes, and the X-axis represents the weight of the edge. The red line represented the estimated true edge weights of the network, the black line represented the edge weights of 1000 bootstrap operations, and the gray area represents the confidence interval. In this study, the black line and the red line highly overlap, and the confidence interval was relatively narrow, indicating that the accuracy of the network estimation was high. For the stability test of node centrality, the correlation stability coefficients of closeness and strength were greater than 0.50, indicating that the estimation of node centrality is stable (Supplementary Figure 2 and Supplementary Figure 3).

Discussion

Our research found that the incidences of ACEs and insomnia symptoms among adolescents were both relatively high, at 47.6% and 20.6% respectively. ACEs were closely related to insomnia symptoms. Physical abuse, emotional abuse and emotional neglect, and being isolated and excluded were risk factors for insomnia symptoms. The relationship between ACEs and insomnia symptoms followed a dose-response pattern with a clear cumulative effect. Network analysis indicated that the connection between ISI4 and ACEI12 in insomnia symptoms was the most direct connection.

Our study findings indicated a high prevalence of ACEs among adolescents, with approximately half of the participants reporting at least one such experience. The most frequently reported types included emotional abuse, emotional neglect, physical abuse, and social isolation or exclusion. These patterns appear closely linked to China’s distinctive cultural and educational context. In traditional Chinese culture, the emotions expression is often overlooked, and children’s inner emotional experiences and communication are not paid attention to. In terms of education, the Confucian educational philosophy that “a strict teacher produces outstanding students” and “no beating, no achievement” is deeply rooted and passed down from generation to generation.27,28 Many parents view emotional support as “spoiling”, often motivating their children through criticism and belittling (such as “other people’s children”), and even educating them by beating them.29 Influenced by the mentality of pushing children too hard, parents exercise emotional control over their children. In addition, structural contradictions and parenting disorders during the modernization transformation, such as the emotional neglect of left-behind children in rural areas, are more prominent.30,31 The problem of isolation and exclusion among teenagers is also very serious, with an incidence rate of 18.9%. In China, when collective honor is intense, individuals often deviate from the group due to factors such as appearance, personality, and academic performance (whether extremely poor or excellent), and are thus excluded or isolated. In recent years, with the increase of school bullying, the incidence of being isolated and excluded has also risen.32

Our research found that the incidence of insomnia symptoms among girls was higher than that among boys, which was basically consistent with previous studies.33 This might be related to the fact that during adolescence, girls experience greater fluctuations in hormone levels and are more prone to biological clock disorders. Additionally, adolescent girls are more sensitive to pressures such as interpersonal relationships and academic performance, tend to overthink, internalize their emotions more easily, and are more likely to experience anxiety and depression, which can affect sleep quality.34 Our research found that the better the exam results, the lower the incidence of insomnia symptoms, and good academic performance was a protective factor against insomnia. In the context of Chinese education, both parents and teenagers attach great importance to academic performance, and poor academic performance is closely related to anxiety, depression and insomnia symptoms.35,36 Senior high school students experienced more insomnia symptoms than junior high school students, which is related to the heavy academic pressure, sleep deprivation and severe emotional problems of senior high school students.37,38

Our research revealed that emotional abuse and emotional neglect were risk factors for insomnia symptoms. Adolescents are in a crucial stage of development, and their prefrontal cortex has not yet fully matured.39 Emotional neglect and abuse might affect the prefrontal cortex maturation,40,41 which could in turn affect emotional regulation and may be linked to internalizing symptoms such as anxiety and depression – conditions known to correlate with insomnia.42–45 Additionally, emotional neglect and abuse appear to be associated with dysfunction of the hypothalamic-pituitary-adrenal (HPA) axis,46 which might contribute to abnormal cortisol rhythms and sleep regulation difficulties.47,48 Early-life stressors, such as maternal separation, can disrupt the formation of parent-child attachment, lead to emotional neglect, and trigger problems such as anxiety, depression, and insomnia.49 Although maternal separation was not measured in the present study, it warrants attention in future research.

Experiencing physical abuse, physical violence or witnessing violence may lead to post-traumatic stress disorder, causing nightmares, night terror symptoms and insomnia. Additionally, the trauma experience may cause individuals to remain in a state of excessive alertness for a long time, with an increased perception of stress.50 Even in a safe environment, it is difficult for them to relax, resulting in difficulty falling asleep or waking easily. The stress-diathesis model posits that insomnia is the result of the interaction between an individual’s inherent susceptibility traits (biological or psychological characteristics) and subsequent stress factors (environmental or psychosocial stress).51 Sleep reactivity reflects an individual’s susceptibility to sleep disruption in response to stress, characterized by difficulties in sleep initiation and maintenance.52 Studies have shown that physical and emotional abuse can further lead to the occurrence of insomnia by enhancing sleep reactivity.53

Our research has found that among all the adverse childhood experiences, being isolated or excluded has the most significant impact on the insomnia symptoms. The risk of insomnia symptoms for adolescents who are isolated or excluded is 2.52 times higher than that of those who are not. According to the 3P hypothesis theory of insomnia, social support is a protective factor for insomnia.54 However, long-term isolation or exclusion weakens social support, especially the support from peers. This leads to intense negative emotions such as loneliness, low self-esteem, and helplessness in individuals, which can cause insomnia through various mechanisms, such as generating anxiety and depression, activating rumination thinking, causing excessive awakening at night.55,56 Additionally, isolated and excluded adolescents may choose social media for communication, forming a social media addiction, disrupting sleep patterns, and long-term exposure to blue light itself also affects insomnia.57 ACE12 (being isolated or excluded) also has the highest betweenness and closeness centrality. The high betweenness centrality shows it acts as a key bridge connecting different types of ACEs—many pathways between other ACEs go through this node. The high closeness centrality shows it sits at the center of the ACE network, with the shortest average distance to all other nodes. Together, these findings suggest that isolation or exclusion not only links different ACEs but may also spread or amplify their mutual influences. Thus, interventions targeting isolation or exclusion in adolescents could bring broad structural benefits. However, this interpretation is based on cross-sectional data and requires validation through longitudinal or experimental studies.

The results of this study indicate a significant dose-response relationship between exposure to childhood adversities and the severity of insomnia symptoms in adolescents. This finding suggests that the impact of childhood adversity on sleep function does not follow a simple “all or nothing” pattern, but rather exhibits a clear cumulative effect characteristic.58 As the number of trauma types increases, it may form a “toxin stress load” through damage to multiple systems and multiple levels, affecting the sleep quality of adolescents. This cumulative effect is also present in other mental disorders such as schizophrenia, bipolar disorder, and major depressive disorder.59–61 In our study, the incidence of insomnia symptoms in adolescents with more than 4 types of adversities increased by more than 6 times. In this study, more than 10% of adolescents had 4 or more types of childhood adversity experiences. Therefore, when assessing childhood adversity in adolescents, we should pay attention to the assessment of the multiplicity and severity of adversity experiences and intervene promptly.

Unlike factor analysis, which treats family crime and parental divorce as interchangeable indicators,62 our network analysis revealed that family crime exhibited the ACE10 (family crime) highest strength centrality. Research has shown that family crime is not an isolated event but is closely associated with unique attributes such as social stigma, shame and secrecy, and dramatic family restructuring.63 Furthermore, family crime maintains strong direct associations with multiple other ACEs, including parental divorce, emotional neglect, physical neglect, and family economic hardship.64 This finding extends prior factor-analytic studies and establishes, at the network level, family crime as a central “driver” node within the childhood adversity co-occurrence system.

In the insomnia and ACE network model, ACE12 (Isolated or excluded), IIS4 (sleep dissatisfaction), and ISI1 (difficulty in falling asleep) constituted the most important core of symptoms connecting insomnia and adverse childhood experiences. Adolescence is highly sensitive to the need for social belonging. Unlike active bullying, being isolated or excluded is a passive form of social rejection, manifested as being ignored, left out, or excluded from peer groups.65,66 The harm of this persistent experience of “being unwanted” lies not in a single event but in the ongoing lack of belonging.67 This undermines an individual’s basic sense of security and creates a state of unease, leading to light sleep, frequent awakenings, and sleep dissatisfaction.68 It also easily triggers rumination and hypervigilance, directly interfering with sleep onset. Adolescents who experience isolation or exclusion tend to be quieter, and their suffering is easily overlooked by adults, which may lead to anxiety and depression, and also induce insomnia.69 Of course, this interpretation is based on cross-sectional data, and causal relationships still require validation through longitudinal studies.

This study has several limitations. Firstly, the cross-sectional design with a one-month data collection period precludes causal inferences and does not allow us to observe dynamic changes in the ACE-insomnia network over time. The network edges are undirected, and interpretations should be viewed as hypothesis-generating rather than confirmatory. Secondly, the ACEs rely on retrospective self-reports from adolescents, which may introduce recall bias and shared method variance. Thirdly, the cumulative ACE score aggregates diverse traumatic events (eg, family crime vs emotional abuse) without considering their severity or developmental timing, which may obscure distinct mechanisms. Fourth, residual confounding cannot be ruled out (eg, genetic factors, subsequent stressors). Fifth, the sample drawn from Shandong Province may introduce selection bias and limit generalizability. Finally, not all ACE types (eg, maternal separation) or objective sleep measures were included. Future research should adopt longitudinal designs to examine the temporal stability of the observed network structure and potential causal pathways, while also incorporating more fine-grained ACE assessments (distinguishing types, severity, and developmental timing) and objective sleep measures.

This study has several limitations. Firstly, the cross-sectional design with a one-month data collection period precludes causal inferences and does not allow us to observe dynamic changes in the ACE-insomnia network over time. The network edges are undirected, and interpretations should be viewed as hypothesis-generating rather than confirmatory. Secondly, the ACEs rely on retrospective self-reports from adolescents, which may introduce recall bias and shared method variance. Thirdly, the cumulative ACE score aggregates diverse traumatic events (eg, family crime vs emotional abuse) without considering their severity or developmental timing, which may obscure distinct mechanisms. Fourth, residual confounding cannot be ruled out (genetic factors, subsequent stressors). Fifth, the sample drawn from Shandong Province may introduce selection bias and limit generalizability. Finally, not all ACE types (eg, maternal separation) or objective sleep measures were included. Future research should adopt longitudinal designs to examine the temporal stability of the observed network structure and potential causal pathways, while also incorporating more fine-grained ACE assessments (distinguishing types, severity, and developmental timing) and objective sleep measures.

Conclusion

This study revealed that adverse childhood experiences were common among adolescents (close to 50%, with 12.2% experiencing ≥4 ACEs) and are closely associated with insomnia symptoms in a dose-response manner. Emotional abuse, physical abuse, witnessing community violence, being isolated or excluded were identified as risk factors. Network analysis suggested that Isolated or excluded may be a potential bridge symptom connected to sleep dissatisfaction and difficulty in falling asleep. Given the cross-sectional design, these findings are hypothesis-generating and require longitudinal confirmation. Assessment of ACEs may be relevant for adolescents with insomnia symptoms, but causal intervention targets remain to be validated in future research.

Funding Statement

National Key Research and Development Program of China (code: 2021YFC2501504); Beijing Municipal Administration of Hospitals Incubating Program (code: PX2025066); Beijing Huilongguan Hospital Longyue Plan Fund Project (LY202402); Capital’s Funds for Health Improvement and Research (CFH)(2026-2-2135).

Declaration of Generative AI Use

We did not use generative AI and AI-assisted technologies in the writing process.

Data Sharing Statement

The datasets generated and/or analyzed during the current study are not publicly available but are available from the corresponding author upon reasonable request.

Ethics Approval and Informed Consent

The study received formal approval from the Ethics Committees of Beijing HuiLongGuan Hospital. Written informed consent was obtained from all participants and their guardians, confirming their voluntary participation. The study followed the 1975 Helsinki Declaration (revised in 2008).

Author Contributions

LLW: formal analysis, methodology, project administration, writing original draft, funding acquisition; SJZ, HJL: conceptualization, data curation, investigation; JJL, JZ: data curation, investigation, project administration; XMJ, MQ: conceptualization, investigation, project administration, resources; JXC, YLT: funding acquisition, methodology, investigation, writing – review & editing. All authors took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors declare no competing interests.