Association between specific childhood trauma and symptom dimensions in patients with obsessive-compulsive disorder: dimensional and network perspectives

ABSTRACT

Background: Childhood trauma has been increasingly recognised as an important environmental risk factor in the development and maintenance of obsessive-compulsive disorder (OCD). However, the specific ways in which different trauma types influence distinct OCD symptom dimensions remain unclear. This study aimed to investigate the multidimensional impact of childhood trauma on OCD using a network analysis approach.

Methods: A total of 410 patients with OCD participated in this study. Factor analysis and exploratory graph analysis were conducted to identify symptom dimensions. Network analysis was used to explore connections between childhood trauma types and symptom dimensions.

Results: Four distinct symptom dimensions were identified: Compulsive Behavioural Tendencies, Obsessive Cognitive Tendencies, Pathological Obsessions and Compulsions, and Emotional Distress. Emotional Distress exhibited the strongest connectivity in the network and served as a key bridge linking childhood trauma to psychopathology. Among the childhood trauma types, emotional abuse and physical neglect demonstrated the highest bridge centrality, highlighting their substantial impact on OCD development and persistence. Network Comparison Tests indicated no significant sex differences in network structure, global strength, or centrality.

Conclusions: Emotional abuse and physical neglect appear to impact OCD symptomatology primarily through emotional dysfunction. These findings emphasise the need for incorporating emotion-focused strategies and trauma-informed interventions into OCD treatment protocols. Longitudinal studies are warranted to better understand the causal relationships between childhood trauma, emotional dysregulation, and OCD.

HIGHLIGHTS

• Four OCD symptom dimensions were identified: Compulsive Behavioural Tendencies, Obsessive Cognitive Tendencies, Pathological Obsessions and Compulsions, and Emotional Distress.

• In the trauma–psychopathology network, emotional abuse and physical neglect show the strongest links to OCD symptoms.

• In the trauma–psychopathology network, Emotional Distress dimension serves as a key bridge linking childhood trauma to OCD.

1. Introduction

Obsessive-Compulsive Disorder (OCD) is a psychiatric condition characterised by persistent, intrusive, and uncontrollable thoughts, impulses, urges, images, or sounds (obsessions) and/or repetitive behaviours or mental acts (compulsions) that individuals feel driven to perform in an effort to reduce anxiety or distress. According to the most recent World Mental Health (WMH) surveys, the lifetime prevalence of OCD is 4.1%, and the disorder is characterised by a persistent course, early onset, and insufficient treatment (Stein et al., 2025). OCD is a mental disorder with a significant disease burden. It has a profound impact on patients’ quality of life and social functioning, leading to substantial emotional suffering (Collaborators, 2022; Organization, 2017).

Childhood trauma, defined as adverse experiences in childhood or adolescence that exceed an individual’s coping capacity, is a significant transdiagnostic environmental risk factor in the aetiology of obsessive-compulsive disorder (Piras & Spalletta, 2020). These may include emotional neglect, emotional abuse, physical neglect, physical abuse, and sexual abuse (Bernstein et al., 2003). Research has shown that individuals exposed to prolonged traumatic environments may internalise the negative messages associated with the trauma, leading to low self-esteem, emotional dysregulation, and attachment disturbances. These factors can increase susceptibility to developing obsessive-compulsive disorder (Boger et al., 2020). Misiak and colleagues have proposed that childhood trauma may indirectly elevate the risk of developing obsessive-compulsive disorder by disrupting psychological mechanisms and perceptual functions, particularly stress perception, emotional regulation, and attachment patterns (Misiak et al., 2017).

However, several questions regarding the relationship between childhood trauma and OCD psychopathology remain unresolved. First, different trauma types are often combined into a composite score, limiting the ability to examine their distinct effects (Bauer et al., 2022). Prior research has shown that different forms of childhood trauma may differentially impact mental health. For instance, emotional abuse has been consistently associated with internalising symptoms such as depression, anxiety, and obsessive-compulsive tendencies, often due to its detrimental effects on self-esteem and emotional regulation (Kadivari et al., 2023; Li et al., 2016). Physical abuse has been confirmed to be linked with externalising problems (Cui & Liu, 2020). Emotional neglect, characterised by a lack of emotional support or validation, has been shown to contribute to difficulties in forming secure attachments and regulating emotions, thereby increasing vulnerability to disorders like OCD (Kim & Cicchetti, 2010; Spinhoven et al., 2014). Sexual abuse, meanwhile, is often associated with a broad range of psychological sequelae, including posttraumatic stress disorder (PTSD), dissociation, and heightened risk for various anxiety disorders (Molnar et al., 2001; Chung & Chen, 2017). These distinctions illustrate the value of examining trauma types separately rather than aggregating them into a single score.

Moreover, OCD symptomatology is highly complex and multidimensional, and it remains unclear whether different trauma types exert distinct effects on specific symptom domains. To capture this heterogeneity, the present study considered three domains of OCD-related psychopathology. The first domain, pathological obsessions and compulsions, reflects maladaptive clinical symptoms that cause marked distress and functional impairment, assessed with the clinician-rated Yale-Brown obsessive compulsive scale (Wu et al., 2016), widely regarded as the gold standard for OCD assessment. The second domain, compulsive tendencies in everyday behaviour, represents trait-like features of compulsivity observable in the general population rather than restricted to clinical cases, assessed with the self-report Obsessive Compulsive Inventory-Revised (OCI-R). As Robbins (2024) observed, ‘Although pathological OCD symptoms are clinically important, there are also compulsive elements of everyday behaviour’ (Robbins, 2024). The third domain, obsessive-compulsive-related cognitions and beliefs, is conceptually distinct from both clinical symptoms and behavioural tendencies. It encompasses dysfunctional attitudes and assumptions, such as threat overestimation, perfectionism, and thought control, which can amplify everyday intrusions into clinically significant obsessions and are measured with the Obsessive Beliefs Questionnaire (OBQ) (Frost & Steketee, 2002). In addition to OCD symptoms, we included dimensional measures of depression, anxiety, and perceived stress, as these symptoms are highly prevalent in OCD even without formal comorbidity (Ruscio et al., 2010; Torres et al., 2016). Childhood trauma is a robust transdiagnostic risk factor for these conditions (Huhne et al., 2024; Wang et al., 2023), and incorporating them allows a more comprehensive understanding of how trauma is associated with mental health through both OCD-specific and transdiagnostic emotional pathways.

To capture the multidimensional nature of childhood trauma and OCD psychopathology, and to elucidate the nuanced effects of trauma on OCD symptoms, network analysis has emerged as a promising methodological approach (Wu et al., 2025). Previous studies have typically used linear regression or structural equation modelling to examine the link between childhood trauma and OCD symptoms (Baldini et al., 2025). However, these methods are limited in handling complex, multidimensional data and cannot accurately capture the distinct effects of specific trauma types on different symptom dimensions. The network approach enables the visualisation and quantification of interactions between various trauma types and OCD symptom dimensions, allowing for the identification of central symptoms and bridge symptoms that play key roles within the trauma – symptom network.

Although our study focused primarily on childhood trauma, it is equally important to acknowledge sex differences in OCD. Epidemiological studies have consistently shown sex-related variations in the onset and course of OCD. Women often present with earlier onset and a predominance of contamination/cleaning symptoms, whereas men are more likely to exhibit symmetry/ordering and hoarding symptoms (Brakoulias et al., 2017; Torresan et al., 2013). Neurobiological and cognitive studies further suggest that males and females may differ in stress perception, emotional regulation, and attachment style, which are critical mechanisms linking trauma exposure to OCD symptomatology (Ferreira et al., 2020). For example, hormonal and neuroendocrine factors may modulate the stress response differently across sexes, potentially influencing vulnerability to trauma-related psychopathology. Importantly, childhood trauma itself may exert sex-specific effects. Prior research has indicated that females are more vulnerable to the long-term emotional consequences of emotional abuse and neglect, while males may show heightened sensitivity to physical abuse or externalising sequelae (Norman et al., 2012; Springer et al., 2007). These distinctions highlight the necessity of disentangling sex-specific pathways through which trauma contributes to OCD heterogeneity. Taken together, considering sex differences is not only crucial for clarifying the complex aetiology of OCD but also carries important clinical implications. Integrating sex perspectives into trauma – OCD research can help identify subgroups at higher risk and inform the development of more precise, individualised interventions.

In summary, this study adopt dimensional and network perspectives to categorise OCD symptoms into distinct dimensions and systematically examine the effects of different types of childhood trauma on each dimension. The aims of this study were to (1) identify symptom dimensions within the multidimensional structure of OCD psychopathology; (2) examine which types of childhood trauma contribute to psychopathology; (3) determine which symptom dimensions are most affected and which function as bridges linking childhood trauma to OCD symptoms; and (4) explore potential sex differences in network structure.

2. Methods

2.1. Sample

Participants were 410 outpatients diagnosed with OCD at Shanghai Mental Health Center from January 2019 to April 2025. Diagnoses were confirmed by senior psychiatrists using the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5). Inclusion criteria: (1) age 16–65 years; (2) $\ge$junior high school education; (3) DSM-5 OCD diagnosis; (4) Yale-Brown Obsessive-Compulsive Scale (Y-BOCS) total score $\ge$16. Exclusion criteria: (1) organic brain disorders or severe physical illnesses; (2) comorbid psychiatric disorders assessed by Mini-International Neuropsychiatric Interview (MINI); (3) pregnancy/breastfeeding; (4) substance abuse or severe suicide attempts.

The final sample included 217 males (52.9%) and 193 females (47.1%), with a mean age of 29.57 $\pm$ 8.19 years. Educational attainment was relatively high, with 55.6% (n = 228) of participants having completed between 12 and 16 years of formal education and 18.0% (n = 74) having attained more than 16 years. A majority of the sample was single (63.2%, n = 259), while 33.4% (n = 137) were married. Regarding socioeconomic indicators, the largest proportion of participants (41.0%, n = 168) reported a per capita monthly household income from 5,001 to 10,000 CNY (detailed distributions for all demographic variables are provided in Supplementary Table S1).

2.2. Measures

2.2.1. Childhood Trauma Questionnaire (CTQ)

The CTQ assesses five types of childhood trauma (physical neglect, physical abuse, emotional neglect, emotional abuse, and sexual abuse) before age 16 (Bernstein et al., 2003). It contains 28 items rated from 1 (Never) to 5 (Always). Internal consistency was good (Cronbach’s $\alpha$ = 0.758).

2.2.2. Beck Depression Inventory – II (BDI-II)

The BDI-II includes 21 items assessing depressive severity, rated on a 4-point scale (0–3) (Beck et al., 1961). Higher scores indicate greater severity. Internal consistency was excellent (Cronbach’s $\alpha$ = 0.927).

2.2.3. Beck Anxiety Inventory (BAI)

The BAI has 21 items measuring anxiety severity, rated 0 from (Not at all) to 3 (Severe) (Beck et al., 1988). Higher scores indicate greater anxiety. Internal consistency was excellent (Cronbach’s $\alpha$ = 0.930).

2.2.4. Perceived Stress Scale (PSS)

The PSS consists of 10 items assessing perceived stress, rated from 0 (Never) to 4 (Very Often) (Cohen et al., 1983). Higher scores indicate greater perceived stress. Internal consistency was good (Cronbach’s $\alpha$ = 0.810).

2.2.5. Obsessive Beliefs Questionnaire-44 (OBQ-44)

The OBQ-44 measures obsessive beliefs across Responsibility/Threat Estimation, Perfectionism/Certainty, and Importance/Control of Thoughts (Obsessive Compulsive Cognitions Working, 2005). It includes 44 items rated 1 (Strongly Disagree) to 7 (Strongly Agree). Internal consistency was excellent (Cronbach’s $\alpha$ = 0.950).

2.2.6. Obsessive-Compulsive Inventory-Revised (OCI-R)

The OCI-R (Foa et al., 2002) has 18 items measuring OCD severity across six subscales: Washing, Checking, Ordering, Obsessing, Hoarding, Neutralising (rated 0–4) (Foa et al., 2002). Higher scores indicate greater severity. Internal consistency was good (Cronbach’s $\alpha$ = 0.875).

2.2.7. Yale-Brown Obsessive-Compulsive Scale (Y-BOCS)

The clinician-administered Y-BOCS includes 10 items (obsessions and compulsions) rated from 0 (no symptoms) to 4 (extreme symptoms) (Goodman et al., 1989). Higher total scores indicate greater OCD severity. Internal consistency was acceptable (Cronbach’s $\alpha$ = 0.748).

2.3. Data analysis

2.3.1. Extraction of symptom dimensions

Before conducting exploratory factor analysis (EFA) and exploratory graph analysis (EGA), the 124 items were aggregated into 14 specific symptoms, derived from the empirically validated factor structures of the original scales. This aggregation retains a broad spectrum of psychopathological symptoms while reducing model redundancy and enhancing the robustness of the resulting factor and network structures. A Supplementary Table S2 details 14 specific symptoms, their source scale, and corresponding internal consistency.

An EFA was first conducted to examine the underlying symptom dimensions in patients with OCD. The analysis was performed using the robust maximum likelihood estimation method (MLR) with oblique rotation. The suitability of the data for factor analysis was assessed using the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy and Bartlett’s test of sphericity. A KMO value greater than 0.80, along with a significant result on Bartlett’s test, indicated that the data were appropriate for factor analysis (Shrestha, 2021). The optimal number of factors was determined through parallel analysis, which is widely regarded as one of the most robust and reliable methods for identifying the appropriate number of factors (Lim & Jahng, 2019). In addition, several conventional model fit indices were used to evaluate the adequacy of the factor model, including the comparative fit index (CFI), Tucker-Lewis index (TLI), standardised root mean square residual (SRMR), and root mean square error of approximation (RMSEA). A factor structure was considered to exhibit good model fit if the CFI and TLI values exceeded 0.90, and both SRMR and RMSEA values were below 0.08.

Subsequently, the study employed EGA, a recently developed approach for identifying latent dimensions, to further assess the robustness of the extracted symptom dimensions (Golino et al., 2021). This method integrates the Gaussian graphical model (GGM) with the Walktrap algorithm for weighted networks to accurately estimate the underlying dimensional structure. To assess the stability of the dimensions identified through EGA, the study further applied bootstrap exploratory graph analysis (bootstrap EGA) (Christensen & Golino, 2021). Specifically, 5,000 bootstrap samples were generated based on the parameter distribution, and EGA was applied to each sample to construct a sampling distribution of the dimensional structure. Additionally, bootstrap EGA assessed the stability of item assignments within each dimension to further evaluate the robustness of the identified structure.

Finally, the core symptom dimensions in patients with obsessive-compulsive disorder were identified by integrating the results of exploratory factor analysis (EFA) and exploratory graph analysis (EGA). The severity of each dimension was quantified using factor scores. Each symptom was uniquely assigned to a single dimension, with no cross-loadings. Modification indices, together with the HiTOP transdiagnostic dimensional theory (Watson et al., 2022), were examined to determine the final factor structure.

2.3.2. Constructing the ‘Childhood Trauma–Symptom Dimension’ network

In this study, various types of childhood trauma and symptom dimensions were treated as nodes to estimate a partial correlation network representing the ‘Childhood Trauma-Symptom Dimension’ structure. The network was regularised using the graphical least absolute shrinkage and selection operator (GLASSO) and selected based on the extended Bayesian information criterion (EBIC) with a hyperparameter $\gamma$ = 0.5, ensuring a parsimonious and interpretable model (Epskamp & Fried, 2018). To evaluate the stability of the network model, a nonparametric bootstrap procedure with 1,000 resamples was performed, and 95% confidence intervals were calculated for the edge weights (Epskamp et al., 2018). In addition, the case-dropping bootstrap method was used to estimate the correlation stability coefficients. A coefficient greater than 0.5 is considered to indicate good stability (Epskamp et al., 2018).

To investigate the importance of individual nodes within the ‘trauma-symptom’ network, this study estimated two key centrality metrics: strength centrality and expected influence. Strength centrality reflects the overall connectivity of a node and is calculated as the sum of the absolute values of all edge weights connected to that node. In contrast, expected influence is computed using the raw edge weights (without taking the absolute value), capturing the directional impact of the node within the network. The study also estimated bridge centrality metrics, including bridge strength centrality and bridge expected influence, to further identify nodes that serve as ‘bridges’ between the childhood trauma and symptom clusters. These indices quantify the strength of associations between specific types of childhood trauma and symptom dimensions across the two clusters. The centrality and bridge centrality measures employed in this study are well-established in the field of psychopathology. They have been widely used to explain the maintenance of psychiatric conditions and to identify critical targets for clinical intervention (Epskamp et al., 2018).

Additionally, network comparison tests were conducted to assess differences in network structure between sex-stratified groups. Using 5,000 permutations and a seed value of 123, the study tested for network invariance, global strength invariance, and centrality invariance. The Benjamini-Hochberg correction was used to access potential different edges.

3. Results

3.1. Identification of symptom dimensions

3.1.1. Exploratory factor analysis (EFA)

The symptom data in this study were suitable for factor analysis. Specifically, the KMO measure was 0.85 (greater than 0.80), and Bartlett’s test of sphericity was significant (Bartlett’s $K^2$=  5997.8, p < .001). In addition, the results of parallel analysis showed that the eigenvalues of the four factors extracted from the actual data were all higher than the average eigenvalues from the simulated data, indicating the presence of four underlying factors (see Supplementary Figure S1).

The results of the exploratory factor analysis indicated that the four-factor model demonstrated excellent fit to the data (CFI = 0.963, TLI = 0.918, SRMR = 0.022, RMSEA = 0.064). 14 specific symptoms were grouped into four distinct symptom dimensions: Dimension 1 primarily reflected the behavioural tendencies of individuals with high obsessive compulsive traits in daily life, including compulsive washing, checking, ordering, hoarding, and neutralising, and was labelled ‘Compulsive Behavioural Tendencies’. Dimension 2 captured the characteristic cognitive processing styles of such individuals, including heightened threat estimation, perfectionism, and thought control, and was labelled ‘Obsessive Cognitive Tendencies’. Dimension 3 represented pathological dysfunctions commonly observed in individuals with OCD, specifically pathological obsessions and compulsions, and was labelled ‘Pathological Obsessions and Compulsions’. Dimension 4 reflected emotional dysregulation in daily life, encompassing obsessive ideation, depressive symptoms, anxiety symptoms, and perceived stress, and was labelled ‘Emotional Distress’. The factor loadings for these four symptom dimensions are presented in Table 1.

Table 1.

Factor loadings of 14 specific symptoms on four symptom dimensions.

	CBT	OCT	PC	ED
Compulsive Washing Tendencies	0.380*
Compulsive Checking Tendencies	0.626*
Compulsive Ordering Tendencies	0.769*
Hoarding Tendencies	0.401*
Compulsive Neutralising Tendencies	0.729*
Threat Estimation		0.806*
Perfectionism		0.862*
Thought Control		0.625*
Pathological Obsessions			0.784*
Pathological Compulsions			0.767*
Obsessive Ideation				0.535*
Depressive Symptoms				0.819*
Anxiety Symptoms				0.822*
Perceived Stress				0.671*

Note. * p < .05. Bolded values indicate the highest factor loading of each symptom on a specific dimension. Only factor loadings greater than 0.3 are presented in the table. CBT = Compulsive Behavioural Tendencies; OCT = Obsessive Cognitive Tendencies; PC = Pathological Obsessions and Compulsions; ED = Emotional Distress.

3.1.2. Exploratory graph analysis (EGA)

This study further validated the core symptom dimensions of obsessive-compulsive disorder using EGA. EGA identified four symptom dimensions, consistent with the results of the exploratory factor analysis. Bootstrap EGA demonstrated good structural stability, with a median number of dimensions equal to 4 and a 95% confidence interval of [3.77, 4.23]. The four-dimensional solution showed the highest replication frequency (0.986), while the three-dimensional and five-dimensional solutions had much lower replication frequencies (0.008 and 0.006, respectively). Additionally, the original network identified by EGA closely aligns with the median network generated through Bootstrap EGA (see Figure 1), further supporting the stability of the identified dimensional structure. Figure 2 presents the reproducibility of symptom assignments to specific dimensions across bootstrap samples. All dimensions showed structural stability coefficients exceeding 0.80, indicating a high degree of robustness in the identified symptom structure.

Figure 1.

Network clusters identified by EGA (left) and median network clusters from Bootstrap EGA (right).

Note. Cluster 1 includes compulsive washing tendencies (CLE), compulsive checking tendencies (CHE), compulsive ordering tendencies (ORD), hoarding tendencies (HOA), and compulsive neutralising tendencies (NEU). Cluster 2 includes obsessive ideation (OBS), depressive symptoms (BDI), anxiety symptoms (BAI), and perceived stress (PSS). Cluster 3 includes pathological obsessions (Yobs) and pathological compulsions (Ycom). Cluster 4 includes threat estimation (RT), perfectionism (PC), and thought control (ICT).

Figure 2.

Proportional likelihood of node assignments to specific clusters based on bootstrap EGA.

Note. Cluster 1 includes compulsive washing tendencies (CLE), compulsive checking tendencies (CHE), compulsive ordering tendencies (ORD), hoarding tendencies (HOA), and compulsive neutralising tendencies (NEU). Cluster 2 includes obsessive ideation (OBS), depressive symptoms (BDI), anxiety symptoms (BAI), and perceived stress (PSS). Cluster 3 includes pathological obsessions (Yobs) and pathological compulsions (Ycom). Cluster 4 includes threat estimation (RT), perfectionism (PC), and thought control (ICT).

In summary, this study identified four core symptom dimensions in patients with obsessive-compulsive disorder: Compulsive Behavioural Tendencies, Obsessive Cognitive Tendencies, Pathological Obsessions and Compulsions, and Emotional Distress. Factor scores were computed for each dimension, and the item-level factor loadings for all four dimensions are provided in Supplementary Table S3.

3.2. The relationship between childhood trauma and symptom dimensions

Figure 3 illustrates the ‘Trauma-Symptom’ network, encompassing five types of childhood trauma within the trauma cluster and four symptom dimensions within the symptom cluster. The overall stability of the network model was satisfactory: edge weight estimates based on the original sample closely aligned with the average estimates derived from bootstrap resampling (see Supplementary Figure S2). The correlation stability coefficient for centrality measures was 0.751, and for bridge centrality measures, 0.517, both indicating moderate to high levels of stability.

Figure 3.

‘Childhood Trauma–Symptom Dimension’ network.

Note: Orange circles represent nodes within the childhood trauma cluster, and blue circles represent nodes within the symptom dimension cluster. The thickness of the edges indicates the strength of partial correlations, with blue edges representing positive associations.

Network centrality analysis revealed that Emotional Distress (ED) was the most central node within the symptom cluster (strength and expected influence both reached 1.865), while emotional abuse (EA) held the highest centrality within the childhood trauma cluster (strength and expected influence both reached 0.601; see Figure 4). These two nodes demonstrated the strongest overall connectivity in the ‘Trauma-Symptom’ network, indicating the closest associations with other nodes across the network.

Figure 4.

Standardised strength centrality and expected influence.

Note. Because all estimated edge weights in the network were positive, strength centrality and expected influence yielded identical standardised values. EA = emotional abuse; PA = physical abuse; SA = sexual abuse; EN = emotional neglect; PN = physical neglect; CBT = Compulsive Behavioural Tendencies; PC = Pathological Obsessions and Compulsions; OCT = Obsessive Cognitive Tendencies; ED = Emotional Distress.

Bridge centrality analysis revealed that ED exhibited the highest bridge centrality (bridge strength and bridge expected influence both reached 2.305; see Figure 5). Of all symptom dimensions, only ED was directly associated with childhood trauma, implying that the co-occurrence of childhood trauma and broader psychopathology in OCD patients may be maintained via their shared association with ED. Furthermore, EA (bridge strength and bridge expected influence both reached 0.658) and Physical Neglect (bridge strength and bridge expected influence both reached 0.446) showed elevated levels of bridge centrality, highlighting these two forms of childhood trauma as significant early adverse experiences leading to the development of psychopathology.

Figure 5.

Standardised bridging strength centrality and bridging expected influence.

Note. Because all estimated edge weights in the network were positive, bridging strength and bridging expected influence are mathematically equivalent. EA = emotional abuse; PA = physical abuse; SA = sexual abuse; EN = emotional neglect; PN = physical neglect; CBT = Compulsive Behavioural Tendencies; PC = Pathological Obsessions and Compulsions; OCT = Obsessive Cognitive Tendencies; ED = Emotional Distress.

Finally, network comparison tests revealed no significant differences between male and female OCD patients in the trauma-psychopathology network with respect to overall network structure (maximum edge-weight difference = 0.238, p = .135), global strength (strength difference = 0.250, p = .403), individual edge weights (all p > .05, Benjamini-Hochberg corrected, see Supplementary Table S4 for details), or node centrality (all p > .05, Benjamini-Hochberg corrected, see Supplementary Table S5 for details), suggesting comparable mechanisms linking childhood trauma to psychopathology across males and females.

4. Discussion

The present study aimed to elucidate the relationship between different types of childhood trauma and OCD psychopathology using dimensional and network perspectives. Several key findings emerged, each contributing unique insights into the heterogeneity and underlying mechanisms of OCD.

4.1. Distinct symptom clusters reflecting multidimensional OCD psychopathology

By integrating latent variable and network models, this study revealed four major symptom clusters corresponding to distinct dimensions of OCD-related psychopathology. Compulsive Behavioural Tendencies represented canonical behavioural manifestations of OCD commonly identified in symptom dimension studies (Abramowitz et al., 2010; McKay et al., 2004; Bloch et al., 2008). Previous research has consistently demonstrated that these behavioural dimensions reflect distinct neural and psychological mechanisms (van den Heuvel et al., 2009). Emotional Distress highlighted a significant emotional comorbidity dimension consistent with numerous studies showing high prevalence of emotional disorders in OCD patients (See et al., 2022; Yap et al., 2018). This emotional dimension further suggests shared underlying mechanisms across internalising psychopathologies. Pathological Obsessions and Compulsions aligned closely with severe, chronic OCD presentations frequently discussed in clinical subtyping literature (Gillan & Robbins, 2014; Stein et al., 2019). The pathological severity dimension identified here underscores the clinical importance of early identification and targeted interventions for severe OCD presentations. Lastly, Obsessive Cognitive Tendencies reflected cognitive vulnerabilities linked to OCD as described in cognitive approaches (Knowles & Olatunji, 2023; Wheaton et al., 2019). Research consistently links these cognitive traits have consistently been linked to poorer treatment outcomes and symptom persistence (Keeley et al., 2008). Overall, these findings strongly support the multidimensional nature of OCD, underscoring the importance of diverse symptom profiles in assessment and personalised treatment approaches.

4.2. Emotional abuse and physical neglect as key childhood trauma subtypes

Our network analysis identified emotional abuse and physical neglect as the most central bridge nodes connecting childhood trauma to obsessive-compulsive psychopathology, highlighting their critical role in impacting OCD psychopathology. These findings are in line with recent systematic evidence (Baldini et al., 2025), which demonstrated that emotional abuse was the trauma subtype most consistently associated with increased OCD severity and the presence of specific symptom dimensions such as aggressive and religious obsessions. Consistent with prior research, EA has been widely linked to vulnerabilities characteristic of OCD, including perfectionism, heightened self-criticism, and emotion dysregulation (Limburg et al., 2017; Lochner et al., 2002). These vulnerabilities may function as maladaptive coping strategies, through which individuals attempt to regain control and predictability in response to early emotional instability. As Baldini et al. (2025) proposed, psychological mechanisms and neurobiological changes, such as HPA axis dysregulation, may lead from trauma to persistent OCD symptoms (Baldini et al., 2025). Meanwhile, PN may disrupt foundational emotional and attachment security, rendering individuals more susceptible to stress and fostering maladaptive self-regulatory styles, a process that is less directly explored in existing literature but supported by studies highlighting the indirect effects of neglect on psychiatric vulnerability (Grisham et al., 2011). The current network framework extends these findings by quantifying the integrative role of PN within a transdiagnostic symptom-trauma structure, underscoring its potential relevance even when direct statistical associations are inconsistent across studies.

4.3. Emotional distress as a central node in the psychopathological network

ED shows the highest centrality in the psychopathological network. It occupies a pivotal hub that connects and coordinates different OCD symptom dimensions. Distress sustains compulsive behaviour through negative reinforcement. Compulsions briefly lower aversive arousal. This short-term relief increases the chance that compulsions will recur. Exposure and response prevention (ERP) based on inhibitory learning interrupts this loop. It disconfirms threat expectancies and helps patients form new associations that tolerate distress (Thampy et al., 2025). These observations suggest that distress is not only an accompanying factor. It may also be a proximal mechanism that maintains the compulsive cycle. Two caveats are important. Centrality is not causality. Highly central nodes are not automatically the best causal targets. Longitudinal and experimental studies are needed to establish causal pathways (Castro et al., 2024). Clinical outcomes vary across individuals. Differences are especially marked in those with high disgust sensitivity or difficulties in emotion regulation (Pelzer et al., 2025). We therefore recommend augmenting standard ERP with affect-focused enhancement, which is not ERP itself but a set of emotion-targeted modules delivered within or alongside ERP (e.g. emotion-regulation training, mindfulness practices, disgust-focused strategies) and aimed at targeting proximal maintaining processes (Thampy et al., 2025). By building distress tolerance and cognitive reappraisal while reducing avoidance-driven disgust arousal, these modules weaken the negative-reinforcement sequence characterised by distress, transient relief, and subsequent recurrence (Kuhne et al., 2024). Magnetic Resonance Imaging (MRI) studies indicate that mindfulness engages and reorganises affect-related networks, including the amygdala, insula, and prefrontal circuitry, thereby providing a mechanistic basis for stratified benefit (Xu et al., 2025).

4.4. Emotional distress as a bridge between childhood trauma and psychopathology

Our network and bridge-centrality analyses indicate that ED may serve as a critical bridge between childhood trauma and OCD psychopathology. This bridging role may be explained through two partially overlapping pathways. The first pathway is the distress pathway. It reflects the association between childhood trauma, particularly emotional abuse or neglect, and higher ED. At the neural level, this corresponds to overactivation of circuits preferentially engaged in emotion generation, including the amygdala, ventral striatum, periaqueductal grey, and interoceptive and threat-monitoring regions such as the insula and anterior/mid-cingulate (Moreira et al., 2017; Zhang et al., 2025). Heightened sensitivity in these systems amplifies responses to negative cues and, through shared frontoparietal control pathways, intensifies symptom activation, which is linked to more severe OCD symptoms and poorer treatment response (Boger et al., 2020). Longitudinal and ecological momentary studies likewise demonstrate that elevated state distress is associated with exacerbation of compulsive behaviours (Bischof et al., 2024). This pattern is consistent with recently proposed frameworks that conceptualise both emotion generation and regulation as iterative ‘world-perception-valuation-action’ (WPVA) loops (Zhang et al., 2025). Within this framework, generation operates as a first-level evaluative system relying on rapid responses to external cues, making trauma-exposed individuals especially vulnerable to distress amplification (Zhang et al., 2025). The second pathway is the emotion-dysregulation pathway. In this pathway, early trauma may impair the second-level evaluative mechanisms required for regulation, which in WPVA terms rely on metacognitive appraisal of one’s own emotions and higher-order goals (Dvir et al., 2014; Zhang et al., 2025). This results in reduced use of adaptive strategies such as reappraisal and restricted metacognitive processing. At the neural level, this is reflected in diminished top-down control from regulation-related regions, including the frontal pole, dorsomedial prefrontal cortex, and anterior cingulate, over emotion-reactive systems such as the amygdala. Individuals with OCD and a history of trauma frequently exhibit abnormal connectivity within fronto-limbic circuits (amygdala, anterior cingulate, medial prefrontal cortex) (Boger et al., 2020; Zhang et al., 2025). These alterations may reduce tolerance for intrusive thoughts and accelerate the transition to, and maintenance of, compulsive behaviours. In sum, childhood trauma is unlikely to represent a single or direct cause of OCD. Instead, it may contribute through two interrelated mechanisms: distress amplification linked to generation circuits and regulatory inefficiency due to impaired control circuits. Together, these mechanisms increase symptom severity and elevate the risk of onset or relapse.

4.5. Sex invariance in trauma–psychopathology network structure

Finally, these results suggest that the overall architecture of associations between childhood trauma and psychopathology is comparable across sexes, as both bridge strength and bridge expected influence followed similar patterns for males and females, with ED showing the highest bridging values. This consistency implies that the mechanism linking specific types of trauma (e.g. EA, PN) to psychopathological outcomes through ED is not sex-specific. These findings are consistent with prior research suggesting that the psychological impact of childhood trauma, particularly emotional dysregulation and internalising symptoms, often transcends sex differences (Elmer et al., 2025; Pechtel & Pizzagalli, 2011). They also support the implementation of transdiagnostic and trauma-informed interventions across both sexes (Han et al., 2021). Nonetheless, future research may consider exploring potential sex-specific pathways using larger and more diverse samples.

4.6. Clinical implications

Our findings support the inclusion of trauma-informed care in the treatment of OCD. We identify feasible augmentation targets. Clinicians should routinely assess trauma histories. Priority should be given to emotional abuse and physical neglect. Eye movement desensitisation and reprocessing or trauma-informed cognitive behavioural therapy may be added to standard ERP according to patient needs. This is especially important for treatment-resistant cases, where trauma-related symptoms are often overlooked (Baldini et al., 2025; Boger et al., 2020; Stein et al., 2019). At the mechanistic level, inhibitory-learning ERP disconfirms threat expectancies and builds new associations that tolerate distress. These processes reduce short-term relief from distress and weaken the likelihood of compulsive recurrence. Translational work provides growing support for this pathway (Thampy et al., 2025). Beyond core ERP, training in distress tolerance and acceptance, mindfulness or acceptance skills, and affect labelling may be useful. These modules can limit cross-node activation in symptom networks. They may also improve adherence. Meta-analytic and systematic evidence suggests that mindfulness-based and acceptance-based programmes reduce OCD severity. Higher-quality randomised trials with longer follow-up are still needed (Burkle et al., 2025; Riquelme-Marín et al., 2022). Interventions that target emotional dysfunction directly also show efficacy. Examples include emotion-focused cognitive behavioural therapy, mindfulness-based strategies, and dialectical behaviour therapy. Dialectical behaviour therapy (DBT) may be particularly suitable for patients with a history of emotional abuse and marked emotion dysregulation (Ahovan et al., 2016; Didonna, 2019; Yap et al., 2018).

4.7. Limitation and future direction

First, this study employed a cross-sectional design, which precludes establishing temporal order and therefore does not permit causal inference regarding the relationship between childhood trauma and OCD symptoms. Future research should advance causal inference and support precision augmentation. Prospective longitudinal designs with multimodal measurement are recommended. Studies should test a sequential mediation from childhood trauma to heightened distress reactivity and impaired regulation, and then to OCD onset and severity. Second, the neurobiological mechanisms linking childhood trauma and OCD have not been systematically examined; future multimodal work combining neuroimaging and neurophysiological indices (e.g. EEG, MRI, TMS-EEG) is warranted to clarify these pathways. Third, all participants were recruited from a single mental health centre in Shanghai and showed relatively homogeneous educational attainment, which may further limit generalisability to more diverse populations. Fourth, patients with comorbid PTSD were not included in this study, and PTSD symptoms were not assessed. This decision was made to avoid the considerable diagnostic heterogeneity of PTSD, which could obscure specific associations between childhood trauma and OCD. Notably, PTSD as defined by DSM-5 can manifest in up to 636,120 distinct symptom profiles (Galatzer-Levy & Bryant, 2013), underscoring its complexity and variability. Nonetheless, given the strong link between PTSD and childhood trauma (Lokhammer et al., 2022), excluding comorbid conditions inevitably limits the real-world relevance of our findings, as they cannot be generalised to patients with comorbid OCD and PTSD. Future research should therefore extend beyond OCD-only cohorts to include comorbid OCD – PTSD populations and employ larger, prospective longitudinal designs to clarify how trauma shapes the onset, severity, and joint development of these disorders. Finally, the instruments used did not differentially assess ‘fear-based’ versus ‘disgust-based’ symptomatology. Consequently, we could not directly test whether childhood trauma is differentially associated with these two dimensions. Prior research suggests that disgust-evoking obsessions may be more treatment-resistant than fear-evoking ones and that trauma may contribute differentially to these patterns (Bragdon et al., 2021; Salmani et al., 2022; Schwartz et al., 2025). Accordingly, future work should incorporate measures that specifically capture fear and disgust responses to further elucidate their roles in the trauma – OCD pathway.

5. Conclusion

Collectively, this study provides a comprehensive perspective on OCD’s multidimensional psychopathology, highlighting ED’s central role and its potential function as a bridge linking childhood trauma with broader symptom dimensions within this network structure. The results emphasise the indirect yet profound impact of EA and PN through emotional dysfunction. Future research should further explore causal relationships among these variables and validate targeted therapeutic strategies. Integrating emotion regulation training and trauma-informed care into OCD interventions appears particularly promising and could substantially enhance clinical outcomes and patient well-being.

Funding Statement

This work was sponsored by the National Natural Science Foundation of China [grant number 82230045] and the Integrated Innovation Team of Shanghai Mental Health Center.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

Due to the sensitive nature of the data and ethical considerations, the datasets generated and/or analysed during the current study are not publicly available. However, they are available from the corresponding author upon reasonable request.

Supplemental Material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/20008066.2025.2580890