Introduction
Childhood traumatic experiences, such as abuse and neglect, have a tremendous impact on individuals’ general development and psychopathology outcomes (Clinchard, Casas, & Kim-Spoon, Reference Clinchard, Casas and Kim-Spoon2026; Green et al., Reference Green, McLaughlin, Berglund, Gruber, Sampson, Zaslavsky and Kessler2010; Lippard & Nemeroff, Reference Lippard and Nemeroff2023). A substantial number of systematic reviews and meta-analyses have suggested that childhood trauma increases the risk for psychosis (Giannopoulou et al., Reference Giannopoulou, Georgiades, Stefanou, Spandidos and Rizos2023; Inyang et al., Reference Inyang, Gondal, Abah, Minnal Dhandapani, Manne, Khanna, Challa, Kabeil and Mohammed2022; Matheson et al., Reference Matheson, Shepherd, Pinchbeck, Laurens and Carr2013; Varese et al., Reference Varese, Smeets, Drukker, Lieverse, Lataster, Viechtbauer, Read, van Os and Bentall2012). More frequent childhood traumatic experiences bear increased risk on psychosis, as individuals with multiple exposures to childhood trauma have an increased risk to develop psychosis compared to individuals with single trauma in nonclinical populations (Shevlin, Houston, Dorahy, & Adamson, Reference Shevlin, Houston, Dorahy and Adamson2008). Additionally, traumatic childhood experiences are associated with higher levels of psychotic symptoms in patients with psychotic disorders (Bailey et al., Reference Bailey, Alvarez-Jimenez, Garcia-Sanchez, Hulbert, Barlow and Bendall2018). Levels of experienced childhood trauma are approximately three times higher among patients with psychosis compared to healthy controls (Matheson et al., Reference Matheson, Shepherd, Pinchbeck, Laurens and Carr2013; Varese et al., Reference Varese, Smeets, Drukker, Lieverse, Lataster, Viechtbauer, Read, van Os and Bentall2012).
The search for factors potentially explaining the link between childhood trauma and psychosis has consistently resulted in negative cognitive schemata about the self as a relevant mediator (Alameda et al., Reference Alameda, Rodriguez, Carr, Aas, Trotta, Marino, Vorontsova, Herane-Vives, Gadelrab, Spinazzola, di Forti, Morgan and Murray2020; Bloomfield et al., Reference Bloomfield, Chang, Woodl, Lyons, Cheng, Bauer-Staeb, Hobbs, Bracke, Kennerley, Isham, Brewin, Billings, Greene and Lewis2021; Giannopoulou et al., Reference Giannopoulou, Georgiades, Stefanou, Spandidos and Rizos2023; Rosenfield, Jiang, & Pauselli, Reference Rosenfield, Jiang and Pauselli2022; Ucok et al., Reference Ucok, Noyan, Guloksuz, Saka, Alptekin, Atbasoglu, Akturan, Karadayı, Tatar, Akdede, Binbay, Altınyazar, Ulaş, Yalınçetin, Gümüş-Akay, Cihan, Soygür, Cankurtaran, Kaymak and van Os2024). Cognitive schemas correspond to core beliefs about oneself or others, developed throughout life and with the potential to change over time (Moscovitch, Moscovitch, & Sheldon, Reference Moscovitch, Moscovitch and Sheldon2023). They influence how individuals perceive themselves and others, interpret events, and anticipate others’ behavior, affecting how they relate to themselves and how they behave in the social world. Childhood trauma has been proposed to lead to the formation of negative self-schemas, which may heighten stress sensitivity, the tendency to interpret neutral stimuli as threatening, and enhance susceptibility to develop paranoia or other psychotic experiences (Garety et al., Reference Garety, Kuipers, Fowler, Freeman and Bebbington2001; Kesting & Lincoln, Reference Kesting and Lincoln2013). Negative self-schemas are apparent across the whole psychosis vulnerability spectrum (Bortolon et al., Reference Bortolon, Capdevielle, Boulenger, Gely-Nargeot and Raffard2013; Hardy et al., Reference Hardy, Emsley, Freeman, Bebbington, Garety, Kuipers, Dunn and Fowler2016; Stowkowy et al., Reference Stowkowy, Liu, Cadenhead, Cannon, Cornblatt, McGlashan, Perkins, Seidman, Tsuang, Walker, Woods, Bearden, Mathalon, Heinssen and Addington2016) and are related with psychotic symptoms (Smith et al., Reference Smith, Fowler, Freeman, Bebbington, Bashforth, Garety, Dunn and Kuipers2006).
In addition to cognitive self-schemas, which operate at a reflective level, disturbances of a prereflective or basic self-experience have also been suggested to mediate the association between childhood trauma and psychotic experiences (Gawęda et al., Reference Gawęda, Prochwicz, Adamczyk, Frydecka, Misiak, Kotowicz, Szczepanowski, Florkowski and Nelson2018). In the last decade, basic self-disturbance is increasingly recognized as a fundamental feature of psychosis (Sass & Parnas, Reference Sass and Parnas2003; Sass, Pienkos, Nelson, & Medford, Reference Sass, Pienkos, Nelson and Medford2013). A substantial body of research suggests that basic self-disturbances are more common in schizophrenia spectrum disorders compared to other mental disorders (Raballo, Poletti, Preti, & Parnas, Reference Raballo, Poletti, Preti and Parnas2021) or healthy controls (Burgin et al., Reference Burgin, Reniers and Humpston2022; Hur, Kwon et al., Reference Hur, Kwon, Lee and Park2014) and indicates associations with higher psychopathology, lower functioning, and lower personal recovery (Henriksen, Raballo, & Nordgaard, Reference Henriksen, Raballo and Nordgaard2021; Jelsma et al., Reference Jelsma, Dijkstra, Wiemer, Vellinga, de Koning, Cahn, Simons, van der Pluijm and de Haan2025; Raballo et al., Reference Raballo, Poletti, Preti and Parnas2021). Two studies confirmed a mediating effect of both cognitive biases and basic self-disturbances in the association between traumatic life events and psychotic proneness in nonclinical populations (Gawęda et al., Reference Gawęda, Prochwicz, Adamczyk, Frydecka, Misiak, Kotowicz, Szczepanowski, Florkowski and Nelson2018; Pionke-Ubych et al., Reference Pionke-Ubych, Frydecka, Cechnicki, Nelson and Gawęda2021). Additionally, one study confirmed mediation of bodily self-disturbance on the association between childhood trauma and psychotic symptoms in patients with schizophrenia (Torregrossa et al., Reference Torregrossa, Liu, Armstrong, Heckers and Sheffield2024).
In the current study, we aimed to test an integrative model linking childhood trauma to psychotic symptoms through negative self-schemas and basic self-disturbance in different subgroups across the psychosis liability spectrum, namely in patients with psychotic disorders, unaffected siblings, and healthy controls. Earlier studies have mainly focused on an overall impact of trauma, while more recent studies suggest that experiences of abuse and neglect may have a distinct role in relation to psychotic symptoms (Alameda et al., Reference Alameda, Christy, Rodriguez, Salazar de Pablo, Thrush, Shen, Alameda, Spinazzola, Iacoponi, Trotta, Carr, Ruiz Veguilla, Aas, Morgan and Murray2021; Alkema et al., Reference Alkema, Marchi, van der Zaag, van der Sluis, Warrier, Ophoff, Kahn, Cahn, Hovens, Riese, Scheepers, Penninx, Cecil, Oldehinkel, Vinkers and Boks2024). In addition, an extensive body of literature suggests that abuse and neglect should be treated as separate constructs, as they differ conceptually (abuse as external threat, neglect as a deprivation) (McLaughlin, Sheridan, & Lambert, Reference McLaughlin, Sheridan and Lambert2014), include distinct environmental and developmental etiologies and exert different impacts on psychopathology (VanMeter et al., Reference VanMeter, Nivison, Englund, Carlson and Roisman2021). Therefore, we aimed to test mediation pathways for abuse and neglect separately. Previous research indicating mediational pathways for negative self-schemas (Alameda et al., Reference Alameda, Rodriguez, Carr, Aas, Trotta, Marino, Vorontsova, Herane-Vives, Gadelrab, Spinazzola, di Forti, Morgan and Murray2020; Bloomfield et al., Reference Bloomfield, Chang, Woodl, Lyons, Cheng, Bauer-Staeb, Hobbs, Bracke, Kennerley, Isham, Brewin, Billings, Greene and Lewis2021; Ucok et al., Reference Ucok, Noyan, Guloksuz, Saka, Alptekin, Atbasoglu, Akturan, Karadayı, Tatar, Akdede, Binbay, Altınyazar, Ulaş, Yalınçetin, Gümüş-Akay, Cihan, Soygür, Cankurtaran, Kaymak and van Os2024) and self-disturbance (Gawęda et al., Reference Gawęda, Pionke, Krężołek, Prochwicz, Kłosowska, Frydecka, Misiak, Kotowicz, Samochowiec, Mak, Błądziński, Cechnicki and Nelson2018; Pionke-Ubych et al., Reference Pionke-Ubych, Frydecka, Cechnicki, Nelson and Gawęda2021) on the association between trauma and psychosis in clinical and nonclinical samples warrants investigating these pathways in different samples across the psychosis liability spectrum. Including siblings may provide insights into familial risk and vulnerability given that this group shares genetic and environmental factors with patients, while assessing pathways across the whole vulnerability spectrum might provide insights into general mechanisms independent of illness-related factors, but associated with vulnerability for psychosis. Moreover, distinguishing different traumatic experiences while assessing these pathways may uncover trauma-specific vulnerabilities and guide more specific targets of awareness or prevention across the psychosis vulnerability spectrum while highlighting the importance of primary prevention of childhood trauma. We hypothesized to find an indirect effect of both abuse and neglect on psychotic experiences through both negative-self schemas and basic self-disturbance across all groups of illness variability. We hypothesized that indirect effects would be stronger for patients than for nonaffected siblings, based on prior research highlighting the negative impact of both negative self-schemas and self-disturbance in patients (Henriksen et al., Reference Henriksen, Raballo and Nordgaard2021; Sundag et al., Reference Sundag, Ascone, de Matos Marques, Moritz and Lincoln2016; Ucok et al., Reference Ucok, Noyan, Guloksuz, Saka, Alptekin, Atbasoglu, Akturan, Karadayı, Tatar, Akdede, Binbay, Altınyazar, Ulaş, Yalınçetin, Gümüş-Akay, Cihan, Soygür, Cankurtaran, Kaymak and van Os2024).
Method
Study design and participants
Participants included in the present cross-sectional study were part of the Genetic Risk and Outcome of Psychosis (GROUP) study (Korver et al., Reference Korver, Quee, Boos, Simons, de Haan and investigators2012). GROUP is a multisite naturalistic follow-up study (with assessments three and six years after baseline) designed to investigate risk and protective factors influencing the onset and course of psychotic disorders in patients, their unaffected siblings, and healthy controls. Patients were recruited in selected representative geographical areas of the Netherlands and Dutch-speaking Belgium through their clinicians. Inclusion criteria were an age range from 16 to 50 years, mastery of the Dutch language and meeting DSM-IV-TR criteria for a nonaffective psychotic disorder, assessed with the Comprehensive Assessment of Symptoms and History (CASH) (Andreasen, Flaum, & Arndt, Reference Andreasen, Flaum and Arndt1992) or the Schedules for Clinical Assessment in Neuropsychiatry (SCAN) (Schutzwohl, Kallert, & Jurjanz, Reference Schutzwohl, Kallert and Jurjanz2007). First-degree family members were contacted through the selected patients. Controls were recruited through random mailings to addresses in the areas where the patients resided. The full procedure of the GROUP study has been described elsewhere (Korver et al., Reference Korver, Quee, Boos, Simons, de Haan and investigators2012). GROUP release 8.0 was used for the current analyses.
The present study used data on childhood trauma from the first follow-up (3 years) assessment and data on self-disturbances, negative schemas, and severity of (positive and negative) psychotic experiences from the second follow-up (6 years) assessment in participating centers in the Amsterdam region, based on availability. Data from 278 participants (110 patients, 120 siblings, and 48 healthy controls) for whom complete data for outcome variables were available were included in the current study. The GROUP study was approved by the Medical Ethics Committee of the Academic Medical Centre of Utrecht (approval number 04/003-O). Written informed consent was obtained from participants before they were enrolled in the study.
Measurements
Childhood trauma
The Dutch Childhood Trauma Questionnaire (CTQ) was used to assess experienced childhood trauma (Bernstein et al., Reference Bernstein, Fink, Handelsman, Foote, Lovejoy, Wenzel, Sapareto and Ruggiero1994). The CTQ is a 25-item self-report questionnaire and measures physical, emotional, and sexual abuse and physical and emotional neglect before the age of 18. Items are rated on a 5-point Likert rating scale ranging from 1 (never true) to 5 (very often true). For all statistical analyses, two childhood trauma subscales were used: abuse (comprising physical, emotional, and sexual abuse) and neglect (comprising physical and emotional neglect). The abuse and neglect subscale scores were computed as the sum of the individual item responses, with higher scores reflecting higher rates of experienced childhood trauma. The CTQ has been demonstrated to be an adequate measurement with good criterion-related validity and reliability, and exhibits high internal consistency (Bernstein et al., Reference Bernstein, Stein, Newcomb, Walker, Pogge, Ahluvalia, Stokes, Handelsman, Medrano, Desmond and Zule2003; Thombs, Bernstein, Lobbestael, & Arntz, Reference Thombs, Bernstein, Lobbestael and Arntz2009).
Psychotic symptoms
The presence of (sub)clinical psychotic experiences was assessed with the Community Assessment of Psychic Experiences (CAPE) (Konings et al., Reference Konings, Bak, Hanssen, van Os and Krabbendam2006). The CAPE is a 42-item self-report questionnaire which measures the prevalence and distress of positive, negative, and depressive symptoms over the last 3 years. Items are rated on a 4-point Likert rating scale ranging from 0 (absent) to 3 (almost always present). In the current study, only the subscales for positive and negative psychotic symptoms were used. A total sum score was calculated, with a higher score reflecting a higher frequency of experienced positive (20 items) and negative symptoms (14 items). The CAPE has been shown to have strong reliability and validity to assess psychotic-like experiences in both clinical (Hansen et al., Reference Hansen, Torgalsbøen, Røssberg, Romm, Andreassen, Bell and Melle2013) and nonclinical samples (Konings et al., Reference Konings, Bak, Hanssen, van Os and Krabbendam2006).
Disturbed self-awareness and depersonalization
The Self-Experience Lifetime Frequency Scale (SELF) was used to assess self-disturbance phenomena (Heering et al., Reference Heering, Goedhart, Bruggeman, Cahn, de Haan, Kahn, Meijer, Myin-Germeys, van Os and Wiersma2016). The SELF is a 12-item self-report questionnaire which captures two aspects of altered self-experience, namely ‘disturbed self-awareness’ and ‘depersonalization’. The former corresponds to a reduced sense of ownership and agency, and diminished self-affection (Parnas et al., Reference Parnas, Moller, Kircher, Thalbitzer, Jansson, Handest and Zahavi2005; Zandersen & Parnas, Reference Zandersen and Parnas2019), while the latter corresponds to a disturbed sense of self-presence and the first-person perspective (Parnas et al., Reference Parnas, Moller, Kircher, Thalbitzer, Jansson, Handest and Zahavi2005). The first domain (disturbed self-awareness) contains items based on the Examination of Anomalous Self-Experience (EASE) (Henriksen et al., Reference Henriksen, Raballo and Nordgaard2021) and the Comprehensive Assessment of At-Risk Mental States (CAARMS) (Yung et al., Reference Yung, Yuen, McGorry, Phillips, Kelly, Dell’Olio, Francey, Cosgrave, Killackey, Stanford, Godfrey and Buckby2005). The second domain (depersonalization) contains items from the Depersonalization Severity Scale (DSS) (Simeon, Guralnik, & Schmeidler, Reference Simeon, Guralnik and Schmeidler2001). In the EASE, comparable items are captured under items 2.2 ‘distorted first-person perspective’ and 2.3 ‘other states of depersonalization’ (Henriksen et al., Reference Henriksen, Raballo and Nordgaard2021). Participants were asked to report about their lifetime frequency and level of distress of the 12 items on a 5-point Likert rating scale. Frequency and distress scores range from 0 (never) to 4 (all the time/severely distressed). A total SELF score was calculated by summing the scores of all items for frequency and distress, with higher scores reflecting higher levels of self-disturbance. Good internal consistency of the SELF has previously been demonstrated in a combined sample of clinical and nonclinical populations, with Cronbach’s alpha of the two separate SELF domains ranging from 0.79 to 0.88 (Heering et al., Reference Heering, Goedhart, Bruggeman, Cahn, de Haan, Kahn, Meijer, Myin-Germeys, van Os and Wiersma2016). The full questionnaire can be found in Supplementary Table S1.
Negative self-schemas
The Brief Core Schemas Scale (BCSS) was used to assess negative self-schemas. It is a 24-item self-report questionnaire which measures negative self-schemas, negative other-schemas, positive self-schemas, and positive other-schemas. In the current study, the negative self-schema subscale was used based on prior research indicating links with psychotic experiences across the vulnerability spectrum (Smith et al., Reference Smith, Fowler, Freeman, Bebbington, Bashforth, Garety, Dunn and Kuipers2006; Stowkowy et al., Reference Stowkowy, Liu, Cadenhead, Cannon, Cornblatt, McGlashan, Perkins, Seidman, Tsuang, Walker, Woods, Bearden, Mathalon, Heinssen and Addington2016) and a mediating effect for negative self-schemas on the association between trauma and psychotic symptoms (Ucok et al., Reference Ucok, Noyan, Guloksuz, Saka, Alptekin, Atbasoglu, Akturan, Karadayı, Tatar, Akdede, Binbay, Altınyazar, Ulaş, Yalınçetin, Gümüş-Akay, Cihan, Soygür, Cankurtaran, Kaymak and van Os2024). Patients were asked to report whether they held a believe (yes/no) and, if present, indicated the degree of conviction on a 5-point Likert rating scale ranging from 1 (slightly believe) to 4 (totally believe). A total sum score was calculated with a higher score indicating higher negative self-appraisal, in line with previous studies (Ashford et al., 2012). Good internal consistency of the BCSS has previously been demonstrated in both clinical and non-clinical samples (Fowler et al., Reference Fowler, Freeman, Smith, Kuipers, Bebbington, Bashforth, Coker, Hodgekins, Gracie, Dunn and Garety2006).
Covariates
The covariates age and sex were selected a priori based on their putative association with psychotic experiences. Data were collected using a self-report questionnaire, specifically developed for the GROUP study.
Statistical analyses
The analyses were performed using Statistical Package for the Social Sciences (IBM SPSS Statistics) version 28.0. Normality was checked visually for all numerical variables with probability plots and scatterplots of standardized residuals. Pearson’s correlation coefficients were calculated to ensure correlation between the childhood trauma subscales and psychotic symptoms. Next, multiple mediation analyses were conducted by testing significance of the indirect effect (effect of X on Y through a and b, as shown in Figure 1) of childhood trauma on psychotic symptoms through negative self-schema and self-disturbance. The PROCESS macro model 4 was used as a regression path analysis tool for parallel mediation analysis (Hayes, Reference Hayes2017). Separate analyses were used for both childhood trauma subscales.
Illustration of the two models used in the mediation analysis. Covariates age and gender were included for all groups.
Figure 1. Long description
The diagram is divided into two vertically stacked panels under three column headers: Predictors, Mediators, and Outcome.
Top Panel: Childhood Abuse
* Predictor: A box on the left labeled X: childhood abuse.
* Mediators: Two boxes in the center. The top box is M 1: negative self-schema. The bottom box is M 2: self-disturbance.
* Outcome: A box on the right labeled Y: psychotic experiences.
* Paths: Dashed arrows labeled a sub 1 and a sub 2 point from X to M 1 and M 2 respectively. Dashed arrows labeled b sub 1 and b sub 2 point from M 1 and M 2 to Y. A solid horizontal arrow labeled c prime points directly from X to Y.
Bottom Panel: Childhood Neglect
* Predictor: A box on the left labeled X: childhood neglect.
* Mediators: Two boxes in the center. The top box is M 1: negative self-schema. The bottom box is M 2: self-disturbance.
* Outcome: A box on the right labeled Y: psychotic experiences.
* Paths: Dashed arrows labeled a sub 1 and a sub 2 point from X to M 1 and M 2 respectively. Dashed arrows labeled b sub 1 and b sub 2 point from M 1 and M 2 to Y. A solid horizontal arrow labeled c prime points directly from X to Y.
A bootstrapping procedure with n = 5000 bootstrap re-samples was used to assess indirect effects. An indirect effect was considered significant if the 95% bias corrected and accelerated bootstrapped confidence interval (BCa CI) excluded zero. The presented path coefficients are unstandardized. The covariates age and gender were added en bloc to account for possible confounding effects. All statistical tests were two sided, and statistical significance was set at p < .05. We did not apply correction for multiple testing, as the PROCESS macro model 4 does not provide exact p-values for indirect effects, but rather confidence intervals to test the null hypothesis with bootstrapping. The ratio of the joint indirect effect to the total effect was used to estimate the proportion of the total effect that was mediated.
Additionally, based on a previous meta-analysis suggesting particularly strong relationships between abuse and positive symptoms, and between neglect and negative symptoms (Alameda et al., Reference Alameda, Christy, Rodriguez, Salazar de Pablo, Thrush, Shen, Alameda, Spinazzola, Iacoponi, Trotta, Carr, Ruiz Veguilla, Aas, Morgan and Murray2021), we investigated differential associations between abuse and neglect and psychotic symptom dimensions in exploratory analyses. Separate mediation analyses were run by testing significance of the indirect effect of both abuse and neglect on both positive and negative symptoms through negative self-schema and self-disturbance in patients, while adjusting for age and gender.
To investigate whether the mediation effects differed per group, we also conducted an exploratory moderated parallel mediation model in the total sample with group as a moderator, using Hayes model 7 (testing moderation of the a-path) and model 14 (testing moderation of the b-path). Models were run separately for abuse and neglect, adjusting for age and gender.
Results
Descriptive statistics
Sociodemographic and clinical characteristics of included patients, unaffected siblings, and controls are presented in Table 1. Patients had a mean age of 35 years and were predominantly male and of white ethnicity. Siblings and healthy controls were similar in age and ethnic composition but had a higher proportion of women. Correlations between predictor and outcome measures showed significant associations for both trauma subtypes across all groups, indicating that higher rates of experienced childhood trauma were related to more frequent psychotic experiences (see Supplementary Table S2).
Study sample characteristics of included patients, siblings, and healthy controls
Table 1. Long description
The table presents data for Patients (n = 110), Siblings (n = 120), and Controls (n = 48). Values are presented as mean (standard deviation) or number (percentage).
* Age in years: Patients 35.03 (0.74), Siblings 35.10 (0.76), Controls 35.56 (1.63).
* Gender (Male): Patients 92 (83.6%), Siblings 53 (44.2%), Controls 30 (62.5%).
* Gender (Female): Patients 18 (16.4%), Siblings 67 (55.8%), Controls 18 (37.5%).
* Ethnicity (White): Patients 90 (84.1%), Siblings 103 (85.8%), Controls 41 (91.1%).
* Education in years: Patients 13.14 (0.40), Siblings 14.54 (0.42), Controls 14.48 (0.47).
* D S M I V classification for Patients: Schizophrenia 65 (59.1%), Schizoaffective disorder 19 (17.3%), Psychotic disorder 12 (10.9%), Other psychotic disorder 12 (10.9%), and Schizophreniform 2 (1.8%).
* Illness duration in years (Patients only): 11.21 (0.49).
* Antipsychotic medication (Patients only): Yes 57 (51.8%), No 27 (24.5%), Unknown 23 (23.6%).
* Psychotic experiences (C A P E): Patients 19.89 (11.80), Siblings 8.53 (7.06), Controls 7.60 (6.39).
* Childhood trauma (C T Q) Abuse: Patients 20.86 (6.87), Siblings 19.16 (16.02), Controls 18.77 (5.39).
* Childhood trauma (C T Q) Neglect: Patients 18.31 (5.51), Siblings 16.02 (5.19), Controls 15.26 (4.77).
* Negative self-schema (B C S S): Patients 2.39 (0.31), Siblings 1.12 (0.20), Controls 0.94 (0.20).
* Self-disturbance (S E L F): Patients 21.58 (1.55), Siblings 4.02 (0.74), Controls 4.40 (1.01).
Note: Data are in number (%) or mean (standard deviation). n, number; DSM, Diagnostic and Statistical Manual of Mental Disorders; CAPE, Community Assessment of Psychotic Experiences; CTQ, Childhood Trauma Questionnaire; SELF, Self-Experience Lifetime Frequency Scale; BCSS, Brief Core Schemas Scale; NOS, Not Otherwise Specified.
Mediation analyses
Table 2 shows the results of the mediation analyses performed in patients, siblings, and healthy controls. Mediation effects exist if, for instance, the a1 and the b1 paths are significant, along with the indirect ab effect. In patients, the association between abuse and psychotic experiences was mediated through negative self-schemas (Figure 2a). Self-disturbance phenomena did not function as a mediator. The association between neglect and psychotic experiences was partly mediated through self-disturbance phenomena, while negative self-schemas did not function as mediator.
Mediating effects of negative self-schema and self-disturbance phenomena on the association between childhood trauma and psychotic experiences
Table 2. Long description
The figure contains three tables. Each table lists paths from X to M 1 (a 1-path), X to M 2 (a 2-path), M 1 to Y (b 1-path), M 2 to Y (b 2 path), Direct effect c prime, Indirect effect a b, and Total effect c.
1. Patients Table: For Abuse, the indirect effect is 0.208 (p < 0.05) and the total effect is 0.380 (p < 0.05), with 54.7 percent mediated. For Neglect, the indirect effect is 0.232 (p < 0.05) and the total effect is 0.659 (p < 0.05), with 35.2 percent mediated.
2. Siblings Table: For Abuse, the indirect effect is 0.294 (p < 0.05) and the total effect is 0.337 (p < 0.05), with 87.2 percent mediated. For Neglect, the indirect effect is 0.237 (p < 0.05) and the total effect is 0.319 (p < 0.05), with 74.3 percent mediated.
3. Controls Table: For Abuse, the indirect effect is 0.364 (p < 0.05) and the total effect is 0.539 (p < 0.05), with 67.5 percent mediated. For Neglect, the indirect effect is 0.357 (p < 0.05) and the total effect is 0.582 (p < 0.05), with 61.3 percent mediated.
Variables are defined as X (trauma abuse or neglect), M 1 (negative self-schema), M 2 (self-disturbance), and Y (psychotic experiences). Beta values and 95 percent C I are provided for all paths.
Note: All analyses were adjusted for age and gender. X, trauma abuse or neglect (predictor); M1, negative self-schema (mediator); M2, self-disturbance (mediator); Y, psychotic experiences (outcome).
a Effects were tested using bootstrapping procedures with confidence intervals (95% CI). When 0 is not included in the 95% confidence interval, significance of the effect is indicated.
b Percentage of the total effect that is mediated.
* Significant at the p < .05 level, highlighted in bold.
Mediation model for patients (a), siblings (b), and healthy controls (c) depicting the direct (c’) and indirect (ab = effect of X on Y through a and b) effects of childhood trauma (abuse and neglect) on psychotic experiences. *p < .05, highlighted in bold. Significant mediators are highlighted in gray.
Figure 2. Long description
A three-panel flowchart labeled a, b, and c. Each panel contains two mediation models. The horizontal flow moves from Predictors (X) on the left, through Mediators (M 1 and M 2) in the center, to Outcome (Y) on the right.
Panel a: Patients.
* Childhood abuse model: Path a sub 1 to M 1 (negative self-schema, gray) is 0.10 asterisk. Path b sub 1 to Y is 1.17 asterisk. Path a sub 2 to M 2 (self-disturbance) is 0.42. Path b sub 2 to Y is 0.22 asterisk. Direct effect c prime is 0.17. Indirect effect a b is 0.21 asterisk.
* Childhood neglect model: Path a sub 1 to M 1 is 0.07. Path b sub 1 to Y is 1.16 asterisk. Path a sub 2 to M 2 (gray) is 0.76 asterisk. Path b sub 2 to Y is 0.19 asterisk. Direct effect c prime is 0.43 asterisk. Indirect effect a b is 0.23 asterisk.
Panel b: Siblings.
* Childhood abuse model: Path a sub 1 to M 1 (gray) is 0.13 asterisk. Path b sub 1 to Y is 1.68 asterisk. Path a sub 2 to M 2 (gray) is 0.29 asterisk. Path b sub 2 to Y is 0.29 asterisk. Direct effect c prime is 0.04. Indirect effect a b is 0.29 asterisk.
* Childhood neglect model: Path a sub 1 to M 1 (gray) is 0.09 asterisk. Path b sub 1 to Y is 1.68 asterisk. Path a sub 2 to M 2 (gray) is 0.32 asterisk. Path b sub 2 to Y is 0.28 asterisk. Direct effect c prime is 0.08. Indirect effect a b is 0.24 asterisk.
Panel c: Healthy controls.
* Childhood abuse model: Path a sub 1 to M 1 (gray) is 0.10 asterisk. Path b sub 1 to Y is 2.64 asterisk. Path a sub 2 to M 2 is 0.35. Path b sub 2 to Y is 0.30 asterisk. Direct effect c prime is 0.18. Indirect effect a b is 0.36 asterisk.
* Childhood neglect model: Path a sub 1 to M 1 (gray) is 0.12 asterisk. Path b sub 1 to Y is 2.55 asterisk. Path a sub 2 to M 2 is 0.20. Path b sub 2 to Y is 0.32 asterisk. Direct effect c prime is 0.25. Indirect effect a b is 0.36 asterisk.
In siblings, the associations between both trauma subtypes (i.e., abuse and neglect) and psychotic experiences were mediated through negative self-schemas and self-disturbance phenomena (Figure 2b).
In healthy controls, the associations between both trauma subtypes (i.e., abuse and neglect) and psychotic experiences were mediated through negative self-schemas, while self-disturbance phenomena did not function as mediator (Figure 2c).
Findings of exploratory analyses on associations between abuse and positive, and neglect and negative symptoms, can be found in Supplementary Tables S3–S6. The total effect of abuse on positive symptoms was not significant, and negative self-schemas and self-disturbance did not have a significant individual mediation effect. The association between abuse and negative symptoms was mediated through negative self-schemas, while self-disturbance did not function as a significant mediator. The association between neglect and negative symptoms was mediated through self-disturbance, while negative self-schemas did not function as significant mediator. Although the total effect of neglect on positive symptoms was not significant, self-disturbance had a significant individual mediation effect.
The exploratory moderated parallel mediation model in the total sample showed significant moderation between neglect and group-status in model 7 (Supplementary Tables S7–S8), while no moderation was observed between abuse and group-status. In model 14 there was no significant moderation by group-status (Supplementary Tables S9–S10).
Discussion
Summary of main findings
The current study aimed to examine the mediating role of self-disturbance phenomena and negative-self schemas in the relationship between two childhood trauma subtypes (i.e., abuse and neglect) and psychotic experiences in patients with psychotic disorders, unaffected siblings and healthy controls. The study is novel in examining mediation patterns across the whole vulnerability spectrum and for trauma subtypes. In line with our hypothesis, negative self-schemas mediated the associations across all groups, except for neglect in patients. Self-disturbance phenomena partially mediated the association between neglect and psychotic experiences in patients but, contrary to expectations, not between abuse and psychotic experiences. In siblings, self-disturbance phenomena mediated the associations between both abuse and neglect and psychotic experiences, while no mediating effect was found in healthy controls. Additionally, an exploratory model testing moderation of group-status in the total sample showing significant moderation of group-status for neglect in one (a-path), but not the other, model (b-path) and no significant moderation of group-status for abuse suggested no consistent differences in indirect effects across the groups. This finding was not in line with our hypothesis of stronger indirect effects in patients.
Comparison with previous literature and proposed mechanisms: Negative self-schemas
Current findings add to prior literature that the association between negative self-schemas and psychotic symptom severity (Humphrey et al., Reference Humphrey, Bucci, Varese, Degnan and Berry2021) also exists at subclinical levels. The mediating role of negative self-schemas in the association between overall trauma and psychosis was previously reported in two meta-analyses (Alameda et al., Reference Alameda, Rodriguez, Carr, Aas, Trotta, Marino, Vorontsova, Herane-Vives, Gadelrab, Spinazzola, di Forti, Morgan and Murray2020; Bloomfield et al., Reference Bloomfield, Chang, Woodl, Lyons, Cheng, Bauer-Staeb, Hobbs, Bracke, Kennerley, Isham, Brewin, Billings, Greene and Lewis2021) and further described in two reviews (Giannopoulou et al., Reference Giannopoulou, Georgiades, Stefanou, Spandidos and Rizos2023; Rosenfield et al., Reference Rosenfield, Jiang and Pauselli2022). Additionally, a recent study including a large sample of patients with schizophrenia reported mediation for negative self-schema in the relationship between both sexual and emotional abuse, and various positive symptoms (Ucok et al., Reference Ucok, Noyan, Guloksuz, Saka, Alptekin, Atbasoglu, Akturan, Karadayı, Tatar, Akdede, Binbay, Altınyazar, Ulaş, Yalınçetin, Gümüş-Akay, Cihan, Soygür, Cankurtaran, Kaymak and van Os2024). Current findings add that these mediation patterns also exist at subclinical levels, as mediation effects of negative self-schemas were observed across all groups, suggesting that the mediation effect is not illness dependent. Given that self-schemas are universally developed, negative self-schemas may be related to psychotic experiences across the whole vulnerability spectrum (Tariq, Reid, & Chan, Reference Tariq, Reid and Chan2021).
Current findings also add to the existing literature by showing that the mediation through negative self-schemas exists for both abuse and neglect (with an exception for neglect in patients). One meta-analysis (Alameda et al., Reference Alameda, Rodriguez, Carr, Aas, Trotta, Marino, Vorontsova, Herane-Vives, Gadelrab, Spinazzola, di Forti, Morgan and Murray2020) described a proportion of 47% of total effect mediated by negative self, other and world schemas on the relationship between trauma and psychosis. These findings appear difficult to compare with current mediated percentages (for two mediators simultaneously) but may roughly align with findings in our patient population (54.7% and 35.2% for abuse and neglect respectively). In siblings and controls, this proportion was higher (see Table 2). The observed full mediation of negative self-schemas in the association between abuse and psychotic experiences across all groups is in line with previous hypotheses that traumatic experiences induce negative self-schemas and thus increase the liability for psychotic experiences (Garety et al., Reference Garety, Kuipers, Fowler, Freeman and Bebbington2001; Kesting & Lincoln, Reference Kesting and Lincoln2013). Abuse may contribute to schemas of feeling unloved, worthless, weak or vulnerable, predisposing individuals to interpret neutral stimuli as negative or threatening, thus potentially inducing psychotic symptoms like paranoia or ideas of reference. However, the exploratory analysis on abuse and positive symptoms showed no significant total effect and no significant mediation of negative self-schemas. This finding was unexpected given prior research indicating a particular strong relationship between abuse and positive symptoms (Alameda et al., Reference Alameda, Christy, Rodriguez, Salazar de Pablo, Thrush, Shen, Alameda, Spinazzola, Iacoponi, Trotta, Carr, Ruiz Veguilla, Aas, Morgan and Murray2021).
Interestingly, mediation was not found for negative self-schemas on the association between neglect and psychotic experiences in patients, which contrasts with previous findings in clinical samples (Appiah-Kusi et al., Reference Appiah-Kusi, Fisher, Petros, Wilson, Mondelli, Garety, Mcguire and Bhattacharyya2017; Hardy et al., Reference Hardy, Emsley, Freeman, Bebbington, Garety, Kuipers, Dunn and Fowler2016). Additionally, the exploratory analyses focusing specifically on the association between neglect and both positive and negative symptoms in patients did not show a mediating effect of negative self-schemas. A potential explanation for detecting mediation in siblings and controls but not in patients could be that the effect of neglect on psychotic experiences is explained by negative self-schemas in non-clinical samples but not in clinical samples. It is possible that other mediators (e.g., self-disturbance phenomena) play a more relevant role in patients with psychosis, or that neglect exerts a more direct effect on psychotic experiences in this group.
Comparison with previous literature and proposed mechanisms: Self-disturbance phenomena
Current findings are consistent with a review underscoring the connection between self-disturbance and psychopathology (Henriksen et al., Reference Henriksen, Raballo and Nordgaard2021) and add that this also exists at subclinical levels. The mediating role of self-disturbance phenomena on the relationship between trauma and psychotic experiences has been examined in two non-clinical studies (Gawęda et al., Reference Gawęda, Prochwicz, Adamczyk, Frydecka, Misiak, Kotowicz, Szczepanowski, Florkowski and Nelson2018; Pionke-Ubych et al., Reference Pionke-Ubych, Frydecka, Cechnicki, Nelson and Gawęda2021) and one study focusing on bodily self-disturbance in schizophrenia (Torregrossa et al., Reference Torregrossa, Liu, Armstrong, Heckers and Sheffield2024), all studies confirming mediation. The findings of partial mediation for neglect in patients and full mediation for both abuse and neglect in siblings contribute to the hypothesis that adverse life experiences may lead to psychotic experiences through self-disturbance phenomena. In other words, experiencing childhood trauma might make individuals more vulnerable to a disturbed experience of the sense of self, which could be so disruptive that psychotic experiences emerge, also at subclinical levels. This hypothesis may be supported by suggestions by Gawęda et al. (Reference Gawęda, Pionke, Krężołek, Prochwicz, Kłosowska, Frydecka, Misiak, Kotowicz, Samochowiec, Mak, Błądziński, Cechnicki and Nelson2018) and Pionke-Ubych et al. (Reference Pionke-Ubych, Frydecka, Cechnicki, Nelson and Gawęda2021), who argued that trauma may induce dissociation, which can increase self-disturbance severity and psychotic proneness. These suggestions may be strengthened by evidence connecting dissociation to trauma (Schroeder et al., Reference Schroeder, Langeland, Fisher, Huber and Schäfer2016) and recognizing it as an aspect of self-disturbance (Sass et al., Reference Sass, Borda, Madeira, Pienkos and Nelson2018) in schizophrenia.
Given that the mediational effect of self-disturbance on the association between neglect and psychotic experiences was partial in patients, self-disturbance does not fully account for the relationship between neglect and psychotic experiences. Notably, observing mediation of self-disturbance phenomena for neglect but not for abuse in the patient group contrasts with findings by Torregrossa et al. (Reference Torregrossa, Liu, Armstrong, Heckers and Sheffield2024), who reported strong mediation for bodily self-disturbance on the abuse-positive symptoms relationship. The mediation for neglect and the absence of mediation for abuse was consistent across both main analyses – where the outcome comprised combined positive and negative symptoms – and the exploratory analyses. Another discrepancy with previous studies is the lack of mediation observed in the control group, contrasting with suggestions of self-disturbance serving as risk factor for psychotic experiences in the general population (Torbet, Schulze, Fiedler, & Reuter, Reference Torbet, Schulze, Fiedler and Reuter2015). A possible explanation for this could be a power issue.
Implications
While a substantial body of literature has confirmed the negative impact of childhood trauma on symptoms and functioning in patients with psychosis (Giannopoulou et al., Reference Giannopoulou, Georgiades, Stefanou, Spandidos and Rizos2023), current findings suggest that the relationship between trauma subtypes and psychotic experiences is at least partially explained by negative self-schemas and self-disturbance phenomena. This highlights the potential value of these experiences for risk assessment or early diagnosis for health-practitioners working with individuals across the psychosis vulnerability spectrum. Interventions targeting negative self-schemas or self-disturbance phenomena may prevent the development of psychotic experiences in individuals at risk for psychosis. Literature on self-disturbance as mediator on the association between childhood trauma and psychotic experiences remains scarce, while current findings show discrepancies regarding trauma type. Currently, specific evidence-based interventions targeting negative self-schemas and self-disturbance in psychosis are lacking. However, Acceptance and Commitment Therapy (ACT) may be promising for improving symptom severity, functioning and well-being (Burhan & Karadere, Reference Burhan and Karadere2021; Morris, Johns, & Gaudiano, Reference Morris, Johns and Gaudiano2024; Shi et al., Reference Shi, Cao, Luo, Liu, Yang and Wang2023). Regarding self-disturbance phenomena, body-oriented psychotherapy has been suggested as potential intervention for patients with psychosis (Skodlar & Henriksen, Reference Skodlar and Henriksen2019). Future studies should explore potential interventions targeting negative self-schemas and self-disturbance in these patients.
Further longitudinal research is needed to explore the potential causal pathways between psychosis, negative self-schemas, self-disturbance phenomena and psychotic experiences. In addition, given the subjective nature of both negative self-schema and self-disturbance phenomena, qualitative research, such as phenomenological interviewing, could provide deeper insights into patients’ prereflective experiences, enhancing healthcare practitioners’ understanding of patients’ perspectives and potentially improving therapeutic alliance (Dijkstra et al., Reference Dijkstra, Rijkeboer, Noordhof, Boyette, Berendsen, de Koning, Bennen, Hofman and de Haan2024). This approach is thought to better capture patients’ perspectives on their illness and treatment than quantitative studies.
Future studies could account for clustering of trauma experiences (and trauma subtypes) within families to enhance our understanding of trauma-psychosis pathways. While current study focused on mediation effects of negative self-schemas, future research might also investigate effects for negative other schemas, given that prior research also suggests mediation of negative other schemas on the association between trauma and psychosis (Alameda et al., Reference Alameda, Rodriguez, Carr, Aas, Trotta, Marino, Vorontsova, Herane-Vives, Gadelrab, Spinazzola, di Forti, Morgan and Murray2020) and a link between negative other schemas and positive symptoms (Jorovat, Twumasi, Mechelli, & Georgiades, Reference Jorovat, Twumasi, Mechelli and Georgiades2025). Additionally, it would be interesting for future studies investigating these pathways to also consider the impact of variables such as treatment status (Jankowski et al., Reference Jankowski, Bennett, Tao, Neria, Dixon, Wainberg, Sweetland and Patel2025; Kamitsis, Harms, & Bendall, Reference Kamitsis, Harms and Bendall2022) and socioeconomic factors (Tonini, Quide, Whitford, & Green, Reference Tonini, Quide, Whitford and Green2022; Varchmin et al., Reference Varchmin, Montag, Treusch, Kaminski and Heinz2021), given that these factors relate to both trauma and psychosis. Future studies could also investigate the impact of both pharmacological and psychological treatments on the investigated pathways between trauma, self-schemas and self-disturbance, and psychotic experiences.
Strengths and limitations
A strength of this study was the use of specific trauma types, providing more detailed insights than overall trauma. Furthermore, including groups across illness variability enabled insights into the illness-relatedness of the associations while enhancing the validity of findings in the patient group. However, several limitations should be noted. First, combining positive and negative symptoms into one outcome variable could have obscured potential differential effects of trauma on these symptom domains. Second, using only two mediators could have oversimplified our model, potentially inflating their contribution. Third, the SELF as screening instrument does not capture the full range of self-disturbance phenomena as described in more comprehensive assessments such as the EASE (Parnas et al., Reference Parnas, Moller, Kircher, Thalbitzer, Jansson, Handest and Zahavi2005), with some items showing overlap with trauma-related symptoms such as depersonalization, derealization or dissociation (Longden, Madill, & Waterman, Reference Longden, Madill and Waterman2012), raising the possibility for multicollinearity and potentially affecting reliability. Fourth, the cross-sectional and retrospective design of the study limits conclusions regarding causation and raises recall bias concerns in reporting childhood trauma, while the CTQ lacks details about trauma duration and other trauma types such as bullying or parental loss. Fifth, the use of self-reported measures may have introduced self-report biases. Sixth, there was a time gap of 3 years between assessments of childhood trauma and the other variables. As it is widely assumed that experienced childhood trauma is relatively stable over time, we presumed that childhood trauma measured with the CTQ would be a good reflection of the originally experienced childhood trauma and would not have been substantially different if it had been assessed three years later. Seventh, with regard to comparisons between patients and siblings, it is possible that familial traumatic experiences are shared, although the groups were not paired one-to-one in current sample. Lastly, the current study’s relatively small sample sizes, particularly for healthy controls, may pose a power issue. It should also be noted that the psychosis group included a notably higher proportion of males compared to siblings and controls, indicating limited generalizability of current findings to females with psychosis.
Conclusion
The current study examined the mediation of both negative self-schemas and self-disturbance phenomena on the relationship between both abuse and neglect, and psychotic experiences in patients with psychosis, siblings, and healthy controls. Current findings confirm the mediating role of negative self-schemas on the association between trauma and psychotic experiences for clinical and non-clinical populations, and suggest mediation of self-disturbance phenomena for patients and siblings. Negative self-schemas and self-disturbance phenomena may be relevant for risk assessment or early diagnosis and treatment among individuals across the psychosis vulnerability spectrum. Longitudinal research should further examine potential causal pathways between childhood trauma, negative self-schemas, self-disturbance phenomena and psychotic experiences. Most importantly, future research should explore potential interventions targeting these experiences to improve outcomes and incorporate qualitative methods to better understand their impact.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S0033291726105492.
Acknowledgements
We are grateful for the generosity of time and effort by the patients, their families, and healthy participants. Furthermore, we would like to thank all research personnel involved in the GROUP project.
Funding statement
The infrastructure for the GROUP study is funded through the Geestkracht Programme of the Dutch Health Research Council (Zon-Mw, grant number 10-000-1001) with matching funds from participating pharmaceutical companies (Lundbeck, AstraZeneca, Eli Lilly, and Janssen Cilag), universities, and mental healthcare organizations (Amsterdam: Academic Psychiatric Center of the Academic Medical Center and the mental health institutions: GGZ Ingeest, Arkin, Dijk en Duin, GGZ Rivierduinen, Erasmus Medical Center, GGZ Noord Holland Noord).
Competing interests
The authors declare that they have no conflict of interest.
Ethical standard
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.
Open access
