Association Between Psychotic and Dissociative Symptoms: Further Investigation Using Network Analysis

ABSTRACT

The association and overlap between psychotic and dissociative phenomena have been increasingly recognized. Previous studies found that psychotic symptoms are closely associated with post-traumatic and dissociative symptoms and that these trauma-related phenomena may mediate the relationship between trauma and psychotic symptoms. It remained less explored which specific post-traumatic and dissociative symptom clusters are particularly associated with psychotic symptoms. This cross-sectional study used a data-driven approach (network analysis) to explore the associations among different psychotic and post-traumatic/dissociative symptom clusters in an online convenience predominantly female sample (N = 468)(59.2% had ever seen a psychiatrist). Participants completed well-established multidimensional measures that assessed different symptom clusters of psychosis, dissociation, and PTSD. In addition, multiple mediation analysis was conducted to examine which post-traumatic/dissociative symptoms could mediate the relationship between childhood and adulthood trauma and different psychotic symptoms. Our results confirmed previous findings that PTSD and dissociative symptoms are closely associated with psychotic symptoms. More importantly, both data-driven and multiple mediation analysis results indicated that identity dissociation was particularly associated with perceptual anomalies and bizarre experiences, while emotional constriction was particularly associated with negative symptoms. It is important to screen for trauma and dissociation and provide trauma-and dissociation-informed care when working with people at risk of or experiencing psychosis. Further longitudinal studies using more representative samples are needed.

KEYWORDS:

• Trauma
• dissociation
• psychosis
• Complex post-traumatic stress disorder (PTSD)
• depression

Psychotic experiences such as hallucinations and delusions are traditionally believed to be indicative of severe mental disorders (e.g., schizophrenia) and therefore are regarded as considerable mental health issues that should receive medical attention, although studies showed that these experiences are not uncommon in the general population too (Johns & van Os, 2001; Kelleher et al., 2012). Even though psychotic symptoms are major features of patients with psychotic disorders, patients with trauma-related disorders, including post-traumatic stress disorder (PTSD), borderline personality disorder (BPD) and dissociative disorders, also commonly report psychotic symptoms; and, more importantly, the psychotic symptoms, especially auditory hallucinations, could have similar clinical presentations in these two categories of people (e.g., McCarthy-Jones & Longden, 2015; Merrett et al., 2016; Moskowitz et al., 2017). Ross et al. (1990) and Laddis and Dell (2012) even found that severely traumatized and dissociative patients report more psychotic symptoms than patients with schizophrenia. In a systematic review, Renard et al. (2017) suggested that there are considerable overlapping symptoms in psychotic and dissociative disorders. A recent meta-analysis also reported that dissociation has robust relationships with psychotic symptoms (r = .437) in both clinical and nonclinical samples, including hallucinations (r = .461), delusions (r = .418) and paranoia (r = .447) (Longden et al., 2020). Moreover, both psychosis and dissociation have a close association with psychological trauma (Read et al., 2008). The association between trauma/adversities and some psychotic symptoms was also found to be mediated by post-traumatic and dissociative symptoms (Bloomfield et al., 2021; Williams et al., 2018). This literature indicates that psychotic symptoms are closely associated with both post-traumatic and dissociative symptoms and that these trauma-related phenomena may mediate the relationship between trauma and psychotic symptoms. In line with these empirical findings, some scholars (e.g., Moskowitz et al., 2009) have proposed that some, if not most, psychotic symptoms may be “traumatic in origin and dissociative in kind.” However, there are two major research gaps that require further investigation, and these research gaps drove our present research.

First, the specific associations among different specific post-traumatic and dissociative symptoms and psychotic symptoms still require more research. As reviewed in Longden et al. (2020), most previous studies relied on the Dissociative Experiences Scale (DES), the Tellegen Absorption Scale (TAS) or other dissociation measures that did not comprehensively assess different clusters of dissociative symptoms. For instance, identity dissociation, which may be particularly associated with voice-hearing experiences (Fung et al., 2020) and is a core feature of complex dissociative disorders (Brand & Loewenstein, 2010; Steinberg & Schnall, 2000), is not specifically assessed in these measures (except for one to two items). In addition, the relationships between psychotic symptoms and different complex PTSD (CPTSD) symptom clusters included in the new ICD-11 remains unexplored. Therefore, the primary objective of this study was to use a data-driven approach (network analysis) to investigate the associations among different specific psychotic and post-traumatic/dissociative symptom clusters, using data collected with well-established multidimensional measures that assess different symptom clusters of psychosis, dissociation and PTSD. A better understanding of the associations among these clinical symptoms can improve the classification, prevention and treatment of these mental health conditions.

Second, most studies found that post-traumatic and dissociative symptoms could mediate the relationship between trauma and psychotic symptoms (Bloomfield et al., 2021; Williams et al., 2018), but these studies have the same limitations. The potential mediating role of each specific post-traumatic/dissociative symptom requires more research in order to improve preventive interventions that target psychotic symptoms in traumatized populations. Therefore, the second objective of this study was to examine which post-traumatic/dissociative symptom clusters would mediate the relationship between childhood and adulthood trauma and various psychotic symptoms.

Methods

Participants and procedures

We analyzed data collected from an online convenience sample that completed self-report measures of trauma, depression, PTSD symptoms, dissociation and psychotic symptoms. From November 2021 to February 2022, potential participants were recruited through mental health-related social media platforms to complete a Google Form.

Participants were those who met the following inclusion criteria: 1) aged 18 or above; 2) provided informed consent to participate; 3) were able to access the Internet and read English; and 4) who self-reported to have experienced any depressive emotions. Participants who self-reported to have: 1) an immediate need for professional help, 2) a currently unstable mental health status, 3) a reading disorder, dementia or intellectual disabilities, and/or 4) recurrent suicidal ideations, attempts or plans during the previous two weeks were excluded from the online survey. These individuals were advised to seek professional help and not to complete the online survey. The original goal of this project was to investigate the experience of people with depressed emotions, therefore we asked if the participants had experienced any depressive emotions in the past. Ethical approval was obtained from the institutional review board at the Chinese University of Hong Kong. The online survey was anonymous and voluntary, and all participants provided online informed consent before they started the survey. Part of the data from the same project has been reported elsewhere (Fung et al., 2022a, 2022b).

Measures

The online survey included questions about health and sociodemographic characteristics (e.g., age, gender, clinical diagnosis) in addition to the following self-report measures:

The Patient Health Questionnaire-9 (PHQ-9)

The PHQ-9 is a 9-item self-report measure of DSM depression, with good internal consistency (α = .86), test-retest reliability (r = .84), and concurrent validity with another depression measure (r = .77) (Kroenke et al., 2001; Kung et al., 2013). A score of 10 or above on the PHQ-9 indicates the presence of depressive symptoms. The PHQ-9 had good internal consistency in our sample (α = .879).

The Positive and Negative Symptoms Frequency subscales of the Community Assessment of Psychic Experiences (CAPE-P and CAPE-N)

The CAPE is a 42-item self-report measure of psychotic symptoms (Konings et al., 2006). The CAPE has three subscales that can be used to assess psychotic (20 items) and negative (14 items) symptoms of psychosis and depressive symptoms. According to a meta-analysis, both the CAPE-P (meta-analytic mean α = .84) and the CAPE-N (meta-analytic mean α = .81) have good internal consistency, and that the three-factor structure of the CAPE was supported; moreover, the CAPE-P can be further divided into three factors (i.e., bizarre experiences [e.g., thoughts are not your own], delusional ideations [e.g., being persecuted in some way] and perceptual anomalies [e.g., hearing voices]) (see Mark & Toulopoulou, 2016). According to Bukenaite et al. (2017), a cutoff score of 1.47 on the CAPE-P could identify individuals at ultra-high risk for psychosis with a sensitivity of 77% and a specificity of 58%. The CAPE-P (α = .896) and the CAPE-N (α = .8881) had good internal consistency in our sample.

The Multiscale Dissociation Inventory (MDI)

The MDI is a 30-item self-report measure that can comprehensively assess six different psychoform dissociative symptom clusters (i.e., disengagement, identity dissociation, emotional constriction, memory disturbance, depersonalization, derealization); these subscales are moderately intercorrelated (mean r = .39) and have satisfactory internal consistency (α = .74 to 96) (Briere, 2002). Most factors of the MDI had good convergent validity with other dissociation measures (β = .22 to .34), accounting for 64–79% of the variance (Briere et al., 2005). The MDI is also strongly correlated with the DES (r = .73) (Mitchell, 2006), which is the most widely-used dissociation measure. The advantage of using the MDI is that it can assess each specific dissociative symptom cluster that is of interest, including identity dissociation, while it is much shorter than other comprehensive dissociation measures such as the Multidimensional Inventory of Dissociation (MID) (a 218-item self-report measure) or the Structured Clinical Interview for Dissociative Disorders (SCID-D) (an in-depth semi-structured diagnostic interview). The six MDI subscales yield raw scores that can be converted to T-scores according to the MDI Manual (Briere, 2002), so that we could determine how many participants had clinically significant levels of each dissociative symptom cluster. The MDI had excellent internal consistency in our sample (α = .957).

The 5-item Somatoform Dissociation Questionnaire (SDQ-5)

The SDQ-5 is a self-report measure of somatoform dissociative symptoms, with satisfactory internal consistency (α = .80) and excellent diagnostic validity (Nijenhuis et al., 1996, 1997). According to Nijenhuis (2010), to identify dissociative disorders, the cutoff score for the SDQ-5 was ≥ 8. The SDQ-5 had poor internal consistency in this sample (α = .539).

The International Trauma Questionnaire (ITQ)

The ITQ is an 18-item self-report measure that assesses six post-traumatic symptom clusters recognized in ICD-11 (i.e., reexperiencing, avoidance, sense of current threat, affective dysregulation, negative self-concept, disturbances in relationships) (Cloitre et al., 2018). The ITQ has very good internal consistency (α = .89 to .94) and concurrent validity with another PTSD measure (r = .89) (Cloitre et al., 2021). The PTSD (α = .844) and DSO (α = .821) symptom subscales of the ITQ had good internal consistency in our sample.

Brief Betrayal Trauma Survey (BBTS)

The BBTS is a 24-item self-report measure that assesses both childhood and adulthood traumatic events, with good test-retest reliability over three years (Goldberg & Freyd, 2006).

Data analysis

The demographic information about the participants, including age, gender, level of dissociation, level of PTSD, as well as the score of psychotic symptoms, was analyzed and revealed with SPSS (Version 27).

The exploration of the relationship network among dissociation, PTSD, psychotic symptoms, and other psychological related variables is conducted via the R Version 4.0. To investigate the relationship, the exploratory network analysis using the Gaussian Graphical Model was performed through the “bootnet” packages (Epskamp et al., 2018). The exploratory network analysis does not require any prior assumptions or hypotheses about the direction of relationships between variables in order for the underlying structure of the data to be revealed on a data-driven basis. It should be noted that the network analysis did not involve hypothesis testing. The EBICglasso was applied to fit a Gaussian graphical model according to the extended Bayesian information criteria in order to optimize the regularization of the parameters. The network’s accuracy and robustness (stability) were assessed with Bootstrapping (1000 bootstraps). The qgraph program is used to depict the networks in order to illustrate the centrality, strength, and stability of the relationships. The weight of the edges between the parameters would also be shown, described as the level of interaction between the variables (please refer to the footnotes regarding the definition of each term). In addition, the ggmFit function is used to reveal the estimated network’s model fitThe directional associations (direct effect, indirect effect and total effect) between network paths were analyzed using the “lavaan” package by transforming the network qgraph into a lavaan model with the “graph2lavaan” function. In network analysis, “betweenness” refers to the amount of influence a node (i.e., variable) has in the network. “Strength” refers to the degree to which each node is connected to other nodes in the network. “Closeness” refers to the average shortest distance between nodes. “Edge-Weight” indicates the type of interaction (positive or negative), and the absolute value of the edge is indicated by the strength centrality.

Furthermore, to achieve the second objective of the study, LISREL 8.8 (Wijanto, 2008) was used to conduct the multiple mediation analysis, to investigate which post-traumatic and dissociative symptom clusters significantly mediate the relationship between trauma and different psychotic symptoms.

Results

Sample characteristics

A total of 468 participants gave informed consent and completed the online survey. Their ages ranged from 18 to 65. This is a regionally diverse sample as participants were from 19 different countries/regions, including: 23.1% from the United Kingdom, 22.0% from Canada, 17.3% from Singapore, 13.7% from the United States, 7.1% from New Zealand, and < 5% from each of the remaining 14 countries/regions. The sample characteristics are reported in Table 1. Most of the measured clinical symptoms were endorsed by the majority of the participants (see Table 1).

Table 1. Sample characteristics (N = 468).

Variables	Mean	SD	Frequency (%)
Age (range = 18 to 65)	25.6	8.56
Gender (Female)			91.0%
Marital status (married)			18.4%
Education level (Bachelor’s degree or above)			42.5%
Being unemployed			41.0%
Financially dependent on partners, friends, family or other people			52.1%
Financially dependent on social welfare			14.5%
Ever seen a psychiatrist			Used to = 36.1%; Current = 23.1%
Ever been hospitalized for mental health issues			21.4%
Types of childhood trauma (BBTS)	3.05	2.28
Types of adulthood trauma (BBTS)	2.73	2.25
Depressive symptoms (PHQ-9)	17.68	6.25	87.6% PHQ-9 ≥ 10
Re-experiencing (ITQ)	4.07	2.45	75.2%^a
Avoidance (ITQ)	5.11	2.40	85.3%^a
Sense of current threat (ITQ)	4.99	2.59	82.3%^a
Affective dysregulation (ITQ)	5.68	1.80	94.7%^a
Negative self-concept (ITQ)	5.64	2.42	87.4%^a
Disturbances in relationships (ITQ)	5.90	2.05	92.7%^a
Disengagement (MDI)	15.55	5.12	64.7%^b
Depersonalization (MDI)	12.27	5.80	66.9%^b
Derealization (MDI)	12.68	5.57	52.4%^b
Emotional Constriction (MDI)	13.98	6.33	54.9%^b
Memory Disturbance (MDI)	11.61	5.24	51.3%^b
Identity Dissociation (MDI)	7.22	4.07	15.0%^b
Somatoform dissociative symptoms (SDQ-5)	6.63	2.32	25.4%^c
Positive Symptoms of Psychotic Disorders (CAPE-P total score)	1.75	0.53	64.1%^d
Delusional ideations (CAPE-P)	2.08	0.64
Bizarre experiences (CAPE-P)	1.60	0.63
Perceptual anomalies (CAPE-P)	1.24	0.50
Negative Symptoms of Psychotic Disorders (CAPE-N)	2.59	0.64

Remarks.

^aEndorsement of a complex PTSD symptom cluster was defined as a score of 2 (Moderately) or above on the ITQ.

^bEndorsement of a psychoform dissociative symptom cluster was defined according to the MDI Manual (Briere, 2002). The raw cutoff score for each of the six subscales was ≥ 14, ≥9, ≥12, ≥13, ≥11, and ≥ 10, respectively.

^cThe cutoff score for the SDQ-5 was ≥ 8 (Nijenhuis, 2010).

^dThe cutoff score for CAPE-P total score was ≥ 1.47 (Bukenaite et al., 2017).

The estimated network

The explorative network analysis was used to explore the possible connections among different dissociation symptoms, PTSD symptoms and psychotic symptoms with other psychological experience variables. The inputted variables included: 1. Dissociation – disengagement, depersonalization, derealization, emotional constriction, memory disturbance, identity dissociation, and somatoform dissociation; 2. Complex PTSD symptoms – reexperiencing, avoidance, sense of current threat, affective dysregulation, and relationship disturbance; 3. Psychotic Symptoms – delusional ideations, bizarre experiences, perceptual anomalies, and negative symptoms. 4. Other psychological experiences – Depression. The network model fit was considered as somewhat adequate, with X²(1099.84/153) = 7.18, CFI = 0.93, TLI = 0.85, RMSEA = 0.084 (90% CI = 0.075–0.092), SRMR = 0.044. Furthermore, with a centrality stability (CS) coefficient of 0.75 and an edge stability coefficient of 0.75, the network centrality was found to be stable. Among all the variables, the memory disturbance features indicated by the Multiscale Dissociation Inventory own the highest centrality (1.14), with closeness at 0.005 and betweenness at 36. Table 2 displays the summary of strength stability, closeness and betweenness of the variables.

Table 2. Summary of the centrality measures.

	Strength (Centrality)	Closeness	Betweenness
Depression (PHQ)	0.96	0.004	22
Re-experiencing (ITQ)	0.80	0.004	10
Avoidance (ITQ)	0.80	0.004	6
Sense of current threat (ITQ)	0.76	0.004	20
Affective dysregulation (ITQ)	0.90	0.004	16
Negative self-concept (ITQ)	0.77	0.004	0
Disturbances in relationships (ITQ)	0.85	0.004	8
Disengagement (MDI)	1.00	0.005	28
Depersonalization (MDI)	1.23	0.005	40
Derealization (MDI)	1.12	0.005	36
Emotional Constriction (MDI)	0.92	0.005	36
Memory Disturbance (MDI)	1.14	0.005	34
Identity Dissociation (MDI)	0.72	0.004	24
Somatoform dissociative symptoms (SDQ-5)	0.55	0.003	0
Negative Symptoms of Psychotic Disorders(CAPE-N)	0.97	0.005	40
Delusional ideations (CAPE-P)	0.90	0.004	30
Bizarre experiences (CAPE-P)	0.95	0.004	6
Perceptual anomalies (CAPE-P)	0.76	0.004	2

Here we focus on reporting which post-traumatic/dissociative symptom clusters had the strongest connection with each psychotic symptom cluster. For the edge weight matrix of other nodes, please refer to Table 3.

Table 3. Edge weight matrix.

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18
(1) Depression (PHQ)	1.00	0.02	0.01	0.02	0.13	0.40	0.03	0.16	0.00	0.00	0.01	0.04	0.00	0.00	0.15	0.00	0.00	0.00
(2) Re-experiencing (ITQ)	0.02	1.00	0.29	0.27	0.02	0.00	0.00	0.00	0.03	0.00	0.00	0.09	0.00	0.00	0.00	0.02	0.00	0.06
(3) Avoidance (ITQ)	0.01	0.29	1.00	0.19	0.09	0.04	0.07	0.00	0.05	0.00	0.04	0.00	0.00	0.01	0.00	0.00	0.00	0.00
(4) Sense of current threat (ITQ)	0.02	0.27	0.19	1.00	0.08	0.00	0.00	0.00	0.00	0.10	0.00	0.01	0.00	0.00	0.00	0.09	0.00	0.00
(5) Affective dysregulation (ITQ)	0.13	0.02	0.09	0.08	1.00	0.15	0.19	0.08	0.02	0.00	0.12	0.00	0.00	0.00	0.02	0.00	0.00	0.00
(6) Negative self-concept (ITQ)	0.40	0.00	0.04	0.00	0.15	1.00	0.11	0.00	0.02	0.00	0.00	0.00	0.00	0.00	0.06	0.00	0.00	0.00
(7) Disturbances in relationships (ITQ)	0.03	0.00	0.07	0.00	0.19	0.11	1.00	0.05	0.02	0.00	0.08	0.00	0.00	0.01	0.21	0.00	−0.03	−0.03
(8) Disengagement (MDI)	0.16	0.00	0.00	0.00	0.08	0.00	0.05	1.00	0.19	0.14	0.00	0.32	0.00	0.00	0.06	0.00	0.00	0.00
(9) Depersonalization (MDI)	0.00	0.03	0.05	0.00	0.02	0.02	0.02	0.19	1.00	0.52	0.14	0.02	0.16	0.03	0.00	0.00	0.03	0.00
(10) Derealization (MDI)	0.00	0.00	0.00	0.10	0.00	0.00	0.00	0.14	0.52	1.00	0.06	0.12	0.00	0.12	0.01	0.00	0.01	0.03
(11) Emotional Constriction (MDI)	0.01	0.00	0.04	0.00	0.12	0.00	0.08	0.00	0.14	0.06	1.00	0.17	0.00	0.03	0.24	0.00	0.00	0.02
(12) Memory Disturbance (MDI)	0.04	0.09	0.00	0.01	0.00	0.00	0.00	0.32	0.02	0.12	0.17	1.00	0.12	0.05	0.01	0.09	0.05	0.04
(13) Identity Dissociation (MDI)	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.16	0.00	0.00	0.12	1.00	0.06	0.00	0.00	0.17	0.21
(14) Somatoform dissociative symptoms (SDQ-5)	0.00	0.00	0.01	0.00	0.00	0.00	0.01	0.00	0.03	0.12	0.03	0.05	0.06	1.00	0.04	0.06	0.05	0.09
(15) Negative Symptoms of Psychotic disorders(CAPE-N)	0.15	0.00	0.00	0.00	0.02	0.06	0.21	0.06	0.00	0.01	0.24	0.01	0.00	0.04	1.00	0.11	0.06	0.00
(16) Delusional ideations (CAPE-P)	0.00	0.02	0.00	0.09	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.09	0.00	0.06	0.11	1.00	0.40	0.13
(17) Bizarre experiences (CAPE-P)	0.00	0.00	0.00	0.00	0.00	0.00	−0.03	0.00	0.03	0.01	0.00	0.05	0.17	0.05	0.06	0.40	1.00	0.16
(18) Perceptual anomalies (CAPE-P)	0.00	0.06	0.00	0.00	0.00	0.00	−0.03	0.00	0.00	0.03	0.02	0.04	0.21	0.09	0.00	0.13	0.16	1.00

For negative symptoms of psychosis, we found that emotional constriction (edge-weight = 0.24, standardized edge-weight/effect size = 2.64) and disturbances in relationships (edge-weight = 0.21, standardized edge-weight/effect size = 2.75) had the strongest connection.

For delusional ideations, sense of current threat (edge-weight = 0.09, standardized edge-weight/effect size = −0.5) and memory Disturbance (edge-weight = 0.09, standardized edge-weight/effect size = −0.12) had the strongest connection.

For bizarre experiences, identity dissociation (edge-weight = 0.17, standardized edge-weight/effect size = 1.12) had the strongest connection.

For perceptual anomalies, identity dissociation (edge-weight = 0.21, standardized edge-weight/effect size = 0.09) also had the strongest connection.

Directed network analysis

The network paths were transformed into a lavaan model in order to investigate the possible directed association among the variables. Aligning with the purpose of exploring which mental health variables (post-traumatic and dissociative symptoms, and depression) would influence psychotic symptoms, the following directed paths were identified in the network. First, derealization showed a direct effect on delusional ideation (β = 0.49, p < .001). Second, somatoform dissociation showed a direct effect on memory disturbance (β = 0.71, p < .001). Third, the sense of current threat also showed a direct effect on the negative symptoms (β = 0.75, p < .001). Furthermore, depression showed an indirect effect toward negative symptoms (β = 0.03, p < .001) and bizarre experience (β = 0.02, p < .001) through disengagement, identity dissociation and negative self-concept. Reexperiencing has revealed an indirect effect toward negative symptoms (β = 0.33, p < .001) and Bizarre Experience (β = 0.15, p < .001) through affective dysregulation. Reexperiencing also had indirect effect toward delusional ideation through emotional constriction (β = 0.19, p < .001). Moreover, beyond the direct effect between the sense of current threat and the negative symptoms, the pathway has further extended toward disturbance in relationships and derealization, then predicted delusional ideations (β = 0.04, p < .001). In the same pathway, derealization has also caused memory disturbance, and predicted depression (β = 0.009, p < .001). Likewise, beyond the direct effect between derealization and delusional ideations, it predicted depression through somatoform dissociation and memory disturbance (β = 0.08, p < .001). Finally, instead of PTSD or dissociation leading to psychotic symptoms, perceptual anomalies (β = 0.68, p < .001) showed a direct effect on reexperiencing. Table 4 displays the direct, indirect effects, and total effects of the paths and variables. In addition, a meaningful pathway was identified. In the lavaan mode, a negative pathway was identified between identity dissociation and avoidance, through negative self-concept, affective dysregulation and depersonalization (β =-0.66, p < .001). However, a positive direct effect from avoidance toward identity dissociation was revealed (β = 0.55, p < .001).

Table 4. Direct, indirect effects, and total effects of the network paths.

	Direct Effect	Indirect Effect	Total Effect
Derealization → Delusional Ideations	0.49**	/	0.49**
Derealization → Delusional Ideations → Somatoform Dissociation → Memory Disturbance → Depression	/	0.08**	0.08**
Somatoform Dissociation → Memory Disturbance	0. 71**	/	0. 71**
Sense Of Current Threat → Negative Symptoms of Psychotic disorders	0. 75**	/	0. 75**
Sense Of Current Threat → Negative Symptoms of Psychotic disorders → Disturbance In Relationships → Derealization → Delusional Ideations	/	0.04**	0.04**
Depression → Disengagement → Identity Dissociation → Negative Self-Concept →Negative Symptoms of Psychotic disorders	/	0.03**	0.03**
Depression → Disengagement → Identity Dissociation → Negative Self-Concept → Negative Symptoms of Psychotic disorders → Bizarre Experience	/	0.02**	0.02**
Re-Experiencing → Affective Dysregulation → Negative Symptoms of Psychotic disorders	/	0.33**	0.33**
Re-Experiencing → Affective Dysregulation →Negative Symptoms of Psychotic disorders → Bizarre Experience	/	0.15**	0.15**
Re-Experiencing → Emotional Constriction → Delusional Ideations	/	0.19**	0.19**
Perceptual Anomalies → Re-Experiencing	0.68**	/	0.68**

**p < .001.

Multiple mediation analysis

Multiple mediation analysis was performed in order to evaluate the degree of mediation between (childhood/adulthood) trauma and different psychotic symptoms, with 13 mediators being added into each analysis.

We first report the possible mediators in the relationship between childhood trauma and psychotic symptoms. We found that sense of current threat (β = 0.05, p < .01) and memory disturbance (β = 0.09, p < .01) were the strongest mediation path toward delusional ideations; negative self-concept was identified as the strongest mediating factor toward negative symptoms (β = 0.05, p < .01); memory disturbance demonstrated a strong mediating effect toward bizarre experience (β = 0.08, p < .01); and, identity dissociation and somatoform dissociation were the strongest mediators in accounting for perceptual anomalies (β = 0.12, p < .01).

We then report the possible mediators in the relationship between adulthood trauma and psychotic symptoms. We found that emotional constriction (β = 0.06, p < .01) demonstrated the strongest mediating effect in predicting negative symptoms; sense of current threat provided the strongest mediating effect in predicting delusional ideations (β = 0.05, p < .01); identity dissociation is shown as the strongest mediating factor toward bizarre experience (β = 0.09, p < .01); and, identity dissociation also displayed the strongest mediating effect in predicting perceptual anomalies (β = 0.11, p < .01).

Table 5 shows the summary of the multiple mediation analysis outcomes (Direct, Indirect and Total Effect) (also see Table S6 in Supplementary files).

Table 5. Multiple Mediation Analysis Path (direct, indirect and total effect) (top 3 mediators).

Paths	Direct Effect	Indirect Effect	Total Effect
Childhood Trauma → Negative Symptoms of Psychotic Disorders	−0.01	0.06*	0.3
Through Emotional Constriction		0.104*
Through Negative Self-Concept		0.055*
Through Disturbance In Relationships		0.044*
Childhood Trauma → Delusional Ideations	0.16*	0.06*	0.21*
Through Memory Disturbance		0.086*
Through Sense Of Current Threat		0.052*
Through Identity Dissociation		0.049*
Childhood Trauma → Bizarre Experience	0.06	0.24	0.3
Through Identity Dissociation		0.101*
Through Memory Disturbance		0.076*
Through Somatoform Dissociation		0.035*
Childhood Trauma → Perceptual Anomalies	0.11*	0.21*	0.32*
Through Identity Dissociation		0.122*
Through Somatoform Dissociation		0.122*
Through Derealization		0.051*
Adulthood Trauma → negative symptoms of psychotic disorders	0.21	0.16*	0.21
Through Emotional Constriction		0.065*
Through Disturbance In Relationships		0.032*
Through Re-Experiencing		−0.03*
Adulthood Trauma → Delusional Ideations	0.12*	0.19*	0.31*
Through Memory Disturbance		0.07*
Through Sense Of Current Threat		0.051*
Through Identity Dissociation		0.04*
Adulthood Trauma → Perceptual Anomalies	0.06	0.19*	0.25
Through Identity Dissociation		0.113*
Through Somatoform Dissociation		0.036*
Through Derealization		0.034*
Adulthood Trauma → Bizarre Experience	0.02	0.21*	0.23
Through Identity Dissociation		0.09*
Through Memory Disturbance		0.062*
Through Derealization		0.016*

*p < .05.

Discussion

In a convenience sample of mostly females with some depressive symptoms, this study confirmed previous findings that PTSD and dissociative symptoms are closely associated with psychotic symptoms. We also made the first attempt to further explore the potential role of different clusters of post-traumatic/dissociative symptoms in the relationship between trauma and psychotic symptoms using both exploratory network analysis and multiple mediation analysis. The relationships among different symptom clusters are complicated, and thus require further discussion.

Before we start discussing our findings, we would like to note that our screening results showed that many participants in our sample had clinically significant levels of the measured mental health symptoms. For example, as reported in Table 1, all PTSD symptom clusters were endorsed by the majority of the participants; additionally, 64.1% of participants exhibited high risk for psychosis according to the suggested cutoff of Bukenaite et al. (2017). More importantly, we understand that many psychopathological issues can be observed along a continuum that spans from the realm of normal functioning to the pathological. Therefore, despite the use of online methods to recruit our participants, our findings possess generalizability that extends beyond nonclinical samples.

First of all, our data-driven results have shown some very interesting yet reasonable results regarding the associations among specific symptom clusters. In the network analysis results subsection, we report which post-traumatic/dissociative symptoms had the strongest connection with each psychotic symptom clusters. We found that identity dissociation was particularly associated with perceptual anomalies (edge-weight = 0.21, standardized edge-weight/effect size = 2.75) and bizarre experiences (edge-weight = 0.17, standardized edge-weight/effect size = 1.12), and these results support the idea that identity dissociation (e.g., having dissociated self-states) may be one important dissociative feature that could “generate” different psychotic symptoms, especially hearing voices (Fung et al., 2020). Clinically, dissociated self-states with unresolved needs may cause unfavorable intrusions (e.g., harsh voices, weird visions, thoughts are not one’s own), and these dissociative phenomena may be labeled as psychotic symptoms in the field of psychosis. The data also showed that emotional constriction (e.g., not being able to feel emotions) was associated with negative symptoms (edge-weight = 0.24, standardized edge-weight/effect size = 2.64), highlighting the possibility that some negative symptoms may result from difficulties in the process of integrating certain emotions.

The directed network analysis also revealed that various PTSD/dissociative symptoms had direct/indirect effect toward psychotic symptoms. It is worth noting that derealization showed significant predictions toward different symptoms, including memory disturbance, delusional ideations and depression. In the literature, it has been found that depersonalization/derealization symptoms were significantly associated with psychological distress, including depression (Schlax et al., 2020). Moreover, Martinez et al. (2020) also observed that that derealization was associated with delusional beliefs among dissociative patients but not in schizophrenia patients. Together with our findings, it implies that at least some delusional symptoms may involve dissociative processes, and may be particularly explained by derealization. Another clinically meaningful finding is that reexperiencing had an indirect effect toward negative symptoms, and bizarre experience, and delusional ideations, highlighting the importance of addressing PTSD symptoms even in the context of psychosis (Brand et al., 2018). Nevertheless, there is one unexpected finding, which showed that perceptual anomalies had a direct effect on reexperiencing (β = 0.68, p < .001).

Finally, we extend previous research by examining the potential mediating effects of different specific post-traumatic/dissociative symptoms in the relationship between trauma and psychotic symptoms. Multiple mediation analyses revealed that the relationship between childhood/adulthood trauma and different psychotic symptoms were mediated by different post-traumatic/dissociative symptoms. We found that identity dissociation was the strongest mediator in the relationship between both childhood and adulthood trauma and bizarre experiences and perceptual anomalies, which are consistent with the above data-driven results (i.e., identity dissociation was particularly associated with perceptual anomalies and bizarre experiences). In addition, emotional constriction was the strongest mediator in the relationship between both childhood and adulthood trauma and negative symptoms. However, for delusional ideations, memory disturbance and sense of current threat were the strongest mediators for childhood and adulthood trauma, respectively. Taken together, these findings imply that identity dissociation should receive more attention when preventing and treating bizarre experiences and perceptual anomalies in the context of trauma, while emotional constriction should be taken into account when preventing and treating negative symptoms. Even some apparently “incomprehensible” symptoms, such as delusions, may also be explained by dissociative processes (Moskowitz et al., 2022). Our findings provide support for the recent development of trauma-informed psychological interventions to prevent and treat psychotic symptoms (Bloomfield et al., 2020; Brand et al., 2018; Ross, 2004), and we recommend that a dissociation-informed approach (e.g., addressing the issues behind emotional constriction [e.g., dissociative phobia], resolving conflicts among dissociated self-states) is needed too. Our findings also support the idea that at least some forms of psychotic symptoms are dissociative in nature (Fung et al., 2020; Moskowitz et al., 2022).

The current study incorporates two sophisticated data analysis tools to elucidate the potential mechanisms and pathways between post-traumatic/dissociative symptoms, trauma, and psychotic symptoms. The network analysis revealed the flow information among variables, which facilitate the identification on how changes in one variable propagate through the network to affect other nodes within the network. As a result, the study comprehensively maps out the potential flow connections between variables, thereby facilitating the development of future hypotheses and investigations into causal effects. Furthermore, the multiple mediation analysis helps researchers uncover the potential mediation mechanisms related to the research questions posited regarding the association between childhood/adulthood trauma and psychotic symptoms. By employing this approach, this study extended previous studies by further investigating which specific post-traumatic/dissociative symptom clusters might particularly mediate the relationship between trauma and psychotic symptoms. As a result, the combination of these two approaches makes a significant contribution to the literature by not only supporting the evaluation and validation of existing hypotheses but also formulating future hypotheses for further investigation (e.g., future studies can test the long-term effects of identity dissociation on psychotic symptoms). For example, our results indicate that the causal effects of identity dissociation on various psychotic symptoms should be further investigated in future longitudinal studies.

Despite the clinical and theoretical implications of our findings, this study has several limitations. First, we used a convenience sample – although it was a regionally diverse sample, most participants were female and reported depressive symptoms. Additionally, we did not measure some important demographic factors (e.g., ethnicity, socioeconomic status) that could affect our findings. Therefore, the generalizability of the findings might be limited. While psychotic experiences are typically associated with mental health problems, not all people having psychotic experiences suffer from diagnosed disorders (Fung et al., 2019; Linscott & van Os, 2013). In addition, people with severe mental health problems (e.g., psychiatric inpatients) may have been underrepresented in this online survey, which also excluded participants with immediate need and current suicidal risks. Therefore, given the limited representativeness of our sample, future research should use more representative clinical and nonclinical samples to replicate our findings. Second, this study relied on self-report data, and the diagnostic status of the participants could not be confirmed. Third, we used a cross-sectional study design, and the predictive and causal relationships among the variables cannot be revealed, but we highlight potential mediating effects that can facilitate follow-up studies in the future. Therefore, further longitudinal studies are needed to elaborate and confirm the predictive relationships among the study variables.

Concluding remarks

This study contributes to the increasing body of knowledge regarding the complex relationship between psychosis and dissociation. Both data-driven and multiple mediation analysis results revealed that identity dissociation was particularly associated with perceptual anomalies and bizarre experiences and that emotional constriction was particularly associated with negative symptoms. Our findings support the idea that at least some psychotic symptoms are dissociative in nature. It is important to regularly screen for trauma, PTSD and dissociative symptoms when working with individuals with psychotic symptoms. More trauma- and dissociation-informed interventions for people at risk of or experiencing psychosis are necessary.

Acknowledgments

The first author received The RGC Postdoctoral Fellowship Scheme 2022/2023 from the Research Grants Council (RGC), Hong Kong.

Data availability statement

Data that support the findings of this study are available from the corresponding authors upon reasonable request.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15299732.2023.2293776

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.