Volume 96, Issue 4 p. 1015-1028
RESEARCH ARTICLE
Open Access

An exploration of the relationship between voices, dissociation, and post-traumatic stress disorder symptoms

Emily Piesse

Corresponding Author

Emily Piesse

School of Psychology and Exercise Science, Murdoch University, Murdoch, Western Australia, Australia

Correspondence

Emily Piesse, Murdoch University, Building 440, 90 South Street, Murdoch, WA, 6150, Australia.

Email: [email protected]

Contribution: Conceptualization, Data curation, Formal analysis, Methodology, Project administration, Writing - original draft, Writing - review & editing

Search for more papers by this author
Georgie Paulik

Georgie Paulik

School of Psychology and Exercise Science, Murdoch University, Murdoch, Western Australia, Australia

Perth Voices Clinic, Murdoch University, Murdoch, Western Australia, Australia

School of Psychological Science, University of Western Australia, Nedlands, Western Australia, Australia

Contribution: Conceptualization, Data curation, Formal analysis, Methodology, Supervision, Writing - original draft, Writing - review & editing

Search for more papers by this author
Danielle Mathersul

Danielle Mathersul

School of Psychology and Exercise Science, Murdoch University, Murdoch, Western Australia, Australia

Centre for Molecular Medicine and Innovative Therapeutics, Health Futures Institute, Murdoch University, Murdoch, Western Australia, Australia

War Related Illness and Injury Study Center (WRIISC), Veterans Affairs Palo Alto Health Care System, Palo Alto, California, USA

Contribution: Conceptualization, Data curation, Methodology, Project administration, Supervision, Writing - original draft

Search for more papers by this author
Lee Valentine

Lee Valentine

Orygen and Centre for Youth Mental Health, University of Melbourne, Parkville, Victoria, Australia

Contribution: Data curation, Formal analysis, Writing - original draft

Search for more papers by this author
Ilias Kamitsis

Ilias Kamitsis

Orygen and Centre for Youth Mental Health, University of Melbourne, Parkville, Victoria, Australia

Contribution: Formal analysis, Writing - review & editing

Search for more papers by this author
Sarah Bendall

Sarah Bendall

Orygen and Centre for Youth Mental Health, University of Melbourne, Parkville, Victoria, Australia

Contribution: Conceptualization, Data curation, Formal analysis, Methodology, Supervision, Writing - original draft, Writing - review & editing

Search for more papers by this author
First published: 06 October 2023

Abstract

Objectives

Extensive research has shown voice hearing to be associated with symptoms of Post-Traumatic Stress Disorder (PTSD) and dissociation. However, most studies have adopted a quantitative design, using cross-sectional data sampling methods, precluding temporal relationships between variables from being defined.

Design

Using a qualitative design, this study sought to identify potential symptom relationships by addressing the research question: what is the nature of the temporal relationship between voices, dissociation and PTSD symptoms?

Methods

Seven voice hearers (aged 27 to 68 years) participated in a semi-structured interview exploring voice hearing, PTSD symptoms, and dissociation. The interviews were analysed using Interpretative Phenomenological Analysis.

Results

One superordinate theme was identified in the data. Voices were observed to occur in dynamic interrelationship with PTSD symptoms and dissociation, and were frequently experienced before and after PTSD symptoms and dissociative episodes.

Conclusions

Implications for theoretical understandings of voice hearing and future research are discussed.

Practitioner Points

  • Trauma-affected voice hearers may experience distressing voices, PTSD intrusions, and dissociative symptoms concurrently or sequentially, and the nature of these relationships between symptoms may be meaningful to the individual.
  • An in-depth assessment of the content of and temporal relationship between voices, dissociative symptoms and PTSD intrusions can aid formulation for psychological therapy.
  • Potential directional effects between voices and PTSD symptoms, identified in the current study, may be relevant for the development or refinement of psychological therapies for voice hearing.

INTRODUCTION

Voice hearing, or experiencing auditory verbal hallucinations, refers to the perception of voices without a corresponding external stimulus (Shinn et al., 2020). While common among individuals diagnosed with psychosis (Berry et al., 2017), voice hearing also commonly occurs in individuals with other—non-psychotic—conditions, especially trauma-related conditions such as Post-Traumatic Stress Disorder (PTSD), Borderline Personality Disorder, and Dissociative Identity Disorder (DID; Moskowitz et al., 2017; Shinn et al., 2020), and can occur in non-clinical populations (Beavan et al., 2011). Childhood trauma is a risk factor for developing psychosis in adulthood (Varese, Smeets, et al., 2012), and greater exposure to traumatic events has been found to confer a higher risk for psychosis in a dose–response relationship (Varese, Smeets, et al., 2012). Later trauma is also a risk factor; one study found people who developed a traumatic stress disorder were at significantly greater risk of being diagnosed with schizophrenia spectrum disorders up to at least 5 years after the traumatic event, even after controlling for other variables such as genetic risk (Okkels et al., 2017). Associations between trauma and voice hearing specifically have also been identified. Multiple studies have found childhood trauma (Bentall et al., 2012), and particularly childhood sexual abuse (Bailey et al., 2018; Hardy et al., 2016; Shevlin et al., 2011), to be associated with higher rates of voice hearing.

There is growing evidence to suggest PTSD symptoms—such as post-traumatic intrusions (e.g., flashbacks, nightmares, or distressing intrusive memories), avoidance, hyperarousal and trauma-related beliefs—are specifically associated with voice hearing (Alsawy et al., 2015; Brand et al., 2020). One large population study found voice hearing was more common among those who had experienced traumatic events and higher reported PTSD symptoms was associated with a greater likelihood of voice hearing (Alsawy et al., 2015). In addition, across non-clinical (i.e., student) and clinical (i.e., psychosis) samples, trauma memory intrusions have been specifically linked with voice hearing (Gracie et al., 2007; Peach et al., 2018). Among individuals with early psychosis, traumatic memory intrusions have been found to mediate the relationship between childhood trauma and voices, with intrusions independently associated with voice-hearing severity (Peach et al., 2018). Hardy (2017) has theorised the link between trauma and psychosis, and suggests that two types of intrusions can lead to psychotic experiences. The first, anomalous intrusions, may explain experiences (psychotic symptoms) that seem unrelated or only thematically linked to past trauma. The second, “memory intrusions may be retrieved at any point along a continuum of contextualisation, with recollections ranging from coherent memories to very fragmented intrusions” (Hardy, 2017, p. 16). These fragmented intrusions are more likely to reflect the auditory, visual and sensory experiences that are characteristic of psychosis, which include voices (Hardy, 2017).

Dissociation is a phenomenon which is common in those who have experienced trauma and is frequently observed in voice hearers across clinical and non-clinical samples (Longden et al., 2019). Some researchers conceptualise dissociative states on a continuum from non-clinical phenomena (i.e., ‘daydreaming’) to clinical symptoms (i.e., amnesia, fugue states) such as those observed in DID (Berry et al., 2017). Others have proposed dissociative symptoms may be grouped into two qualitatively distinct categories: detachment, comprising symptoms of depersonalisation (i.e., feeling alienated or detached from one's body) and derealisation (i.e., feeling the external world to be unreal or false); and compartmentalisation, comprising dissociative amnesia and fugue states (Brown, 2006; Holmes et al., 2005). Studies of individuals with PTSD have shown dissociation to be significantly higher among voice hearers, compared to non-voice hearers (Anketell et al., 2010); and higher dissociation levels predict more extensive voice hearing, whereas PTSD symptoms do not (Wearne et al., 2018). Furthermore, there is evidence that dissociation mediates the relationship between trauma and voice hearing, whereby trauma does not directly precipitate voice hearing, but dissociation in response to trauma increases vulnerability to hearing voices (Allen et al., 1997; Sun, Alvarez-Jimenez, Simpson, et al., 2018; Varese, Barkus, & Bentall, 2012). In a recent meta-analysis from Bloomfield et al. (2021), dissociation was found to predict hallucinations, including voices, among people who had experienced developmental trauma, with a small-to-medium effect size (i.e., pooled Cohen's d = 0.35), consistent with other studies (Varese, Barkus, & Bentall, 2012; Williams et al., 2018). It has also been suggested that while hearing voices can occur within the context of psychotic disorders, it should be considered a dissociative experience (Moskowitz & Corstens, 2008) that results from past trauma or interpersonal stress (Longden et al., 2012).

Despite increasing evidence for dissociation as a co-occurring and potentially mediating variable in the relationship between trauma and voice hearing, the mechanism underlying this observed relationship remains unknown (Pilton et al., 2015; Varese et al., 2011). Moreover, it is unclear how voices, PTSD symptoms and dissociation may be temporally related (Bloomfield et al., 2021; Brand et al., 2020), and what the nature of any interrelationships may be, including directional effects (Pilton et al., 2015). Most studies have employed cross-sectional designs, using questionnaires to measure constructs and analysing correlations between items to estimate relationships between variables, meaning inferences about temporality or causality are not able to be substantiated (Bloomfield et al., 2021; Jongeneel et al., 2020; Pilton et al., 2015; Williams et al., 2018).

However, a small subset of researchers has employed ecological momentary assessment (EMA) to investigate potential temporal relationships. EMA involves the repeated administration of questionnaires over a series of days, at undisclosed intervals, to collect time-series data which can quantify symptom frequency and illustrate temporal relationships (Jongeneel et al., 2020). One of the few PTSD studies to explore temporal relationships between voices and other symptoms was undertaken by Brand et al. (2020), who used an EMA design to measure day-to-day symptom patterns among a sample of chronic voice hearers with trauma histories. The results showed voices were more likely to occur when trauma memory intrusions (i.e., flashbacks) were more frequent in the preceding hour. There was no such association between voices and hyperarousal or avoidance.

Two other dissociation-focused research groups have employed EMA designs to explore temporal relationships with voices (Jongeneel et al., 2020; Varese et al., 2011). In one study of patients with schizophrenia spectrum disorders, state dissociation predicted voice hearing in real time (Varese et al., 2011). In a similar study by Jongeneel et al. (2020), results showed that while voices and dissociation symptoms co-occurred contemporaneously (i.e., in the moment of sampling), dissociation did not predict voice hearing, in contrast to the findings of Varese et al. (2011).

A further limitation in the research literature pertains to the relatively few studies employing qualitative methods (McCarthy-Jones & Longden, 2015). While some researchers have included a qualitative component in their studies (Anketell et al., 2011), phenomenological research has not been the core focus. Others have argued that in addition to psychopathological, neurobiological, and psychological explanations of voice hearing, a better phenomenological understanding—a deep knowledge of the lived experience of voice hearers—is required to address evidence gaps, corroborate models, and ultimately improve clinical interventions (Upthegrove et al., 2016).

The present study seeks to address both methodological limitations and theoretical gaps in the literature by employing qualitative methods to address the research question: what is the nature of the relationship between the voices, dissociation, and PTSD symptoms? While dissociation has been suggested to mediate the relationship between trauma and voice hearing (Longden et al., 2012), most research has been cross-sectional (Williams et al., 2018), directional effects have not been substantiated, and not all constructs have been examined in the same study (Brand et al., 2020; Pilton et al., 2015). It is hoped that the semi-structured interview format used in the present study may elicit important information about temporal patterns of symptom onset and offset, which may help explain interrelationships between variables.

METHOD

Design

A mixed-methods research design comprising a phenomenological qualitative interview and quantitative symptom measures was used to explore participants' experiences of voice hearing, dissociation, and PTSD symptoms. Data was collected via a semi-structured interview with each participant and analysed using Interpretative Phenomenological Analysis (IPA; Smith et al., 2009). IPA is a qualitative research framework which prioritises first-person perspectives on phenomena of interest, placing individual experience at the centre of the research endeavour (Smith et al., 2009). IPA was well suited to this study as it aimed to capture the rich lived experience of the participants. IPA design typically involves sample sizes that are small even by qualitative research standards—a sample size of between three and 10 is considered appropriate (Smith et al., 2009). This is due to the in-depth nature of the individual interview and the focus on the depth of understanding of the phenomenology of experiences over the breadth of such experiences (Smith et al., 2009). We aimed to recruit up to 10 participants in the current study. Quantitative measures of voices, PTSD symptoms and dissociation were also employed to describe the cohort.

Participants

Participants were recruited from the research registry of Perth Voices Clinic (PVC), a specialist clinical psychology service for voice hearing. Membership of the registry, and participation in any related projects, is voluntary. Ethics approval for the project (HREC 2021/134) and use of the PVC research registry (HREC 2016/089) was provided by Murdoch University's Human Research Ethics Committee. Purposive sampling was utilised to recruit from the registry, in accordance with IPA principles (Smith et al., 2009), meaning participants were selected based on their assumed ability to convey rich personal experiences of voice hearing as assessed by author GPW, who had detailed knowledge of the PVC research registry as an administrator. Consideration was given to obtaining a diverse sample according to variables such as age, gender, and trauma complexity. Inclusion criteria required participants to be aged 18 years and over, and to have heard voices (i.e., not originating from another person) in the past week.

Eligible registry members were contacted by telephone to ascertain interest in participation. Interested registrants were emailed an information letter and a follow-up phone call was made 1 week later to request participation. Recruitment continued until seven participants had been confirmed. This sample size was deemed appropriate given IPA emphasises the importance of individual perspectives and case-by-case exploration of experiences (Smith et al., 2009).

Measures

The Hamilton Program for Schizophrenia Voices Questionnaire (HPSVQ; Van Lieshout & Goldberg, 2007) is a 13-item, self-report measure of key characteristics of auditory verbal hallucinations (voices). The HPSVQ assesses the content (i.e., quantity of negative content), volume, frequency and degree of distress elicited by an individual's voices, among other aspects of the hallucination. Nine items are scored on a five-point, Likert-type scale ranging from 0 (least severe) to 4 (most severe), with scores summed to give an overall severity score. The remaining four items measure qualitative aspects of voices and are not included in the total score. The HPSVQ has good convergent validity with the Psychotic Symptom Rating Scales—Auditory Hallucinations (PSYRATS-AH), a validated clinical measure of auditory hallucinations which has been used extensively in research settings (Berry et al., 2021).

The Dissociative Experiences Scale-Taxon (DES-T; Waller et al., 1996) is an eight-item, self-report measure derived from the larger DES-II (Carlson & Putnam, 1993). The DES-T provides an indication of whether an individual's dissociation symptoms are pathological (i.e., related to a dissociative disorder) or not (i.e., reflecting ‘normal’ dissociative experiences such as daydreaming). Responses range on a scale from 0% (never) to 100% (always), with the scale increasing in increments of 10%. The DES-II has demonstrated good psychometric properties for research purposes (van Ijzendoorn & Schuengel, 1996) while the DES-T has been shown to have excellent sensitivity as a screening instrument for DID (Ross, 2021).

The PTSD Checklist for DSM-5 (PCL-5; Weathers et al., 2013) is a 20-item, self-report measure which maps onto the DSM-V (American Psychiatric Association, & American Psychiatric Association DSM-5 Task Force, 2013) symptom checklist for PTSD and indicates distress related to symptoms over the past month. Items are scored on a Likert-type scale from 0 (not at all) to 4 (extremely). The PCL-5 has demonstrated good reliability and validity among trauma-exposed college students in the United States (Blevins et al., 2015).

Qualitative interview

A semi-structured interview schedule was developed to address the research question: “what is the nature of the relationship between voices, dissociation, and PTSD symptoms”? The interview schedule comprised five primary questions and follow-up questions or prompts (see Figure S1). The questions pertained to general voice hearing experiences (i.e., “Please describe for me, in as much detail as possible, what happens when you hear your voice/s, from before they occur until after they finish”), as well as PTSD symptoms, dissociative symptoms, and ‘other’ experiences perceived to be related to voices and PTSD or dissociation symptoms. A Lived Experience Consultant associated with the PVC research registry was sent a summary of the aims of the study and a copy of the draft structured interview, and via telephone provided feedback on the structure and content of the interview schedule. The interview was amended according to this feedback. In the early phase of planning this study, the PVC research registry committee was consulted in a face-to-face meeting, which includes three people with lived experience, two of whom represent the Hearing Voices Network.

Procedure

Data was obtained via semi-structured interview between participants and the primary researcher. Interviews were conducted at Murdoch University's Psychology Clinic. Six interviews were conducted in-person during a single session; one interview was conducted across two sessions, via teleconference and telephone, as the participant lived interstate. All participants were offered the opportunity to be accompanied by a support person, but all declined. The semi-structured interview was governed by IPA principles, with rapport-building and follow-up questions (i.e., to encourage elaboration in areas of interest) as appropriate (Smith et al., 2009). Measures to assess PTSD and dissociation symptoms were administered following the interview. A verbal debriefing was provided to participants.

Data analysis

Data were analysed using the IPA guidelines developed by Smith et al. (2009). Initially, data familiarisation was undertaken by reading all participant transcripts several times. Each individual transcript was coded, and emerging ideas, keywords, and phrases were annotated (Smith et al., 2009). When the initial data analysis was completed for each participant, emerging themes were then conceptualised for the whole sample, while also retaining the individuality of lived experience. The contradictions and the elaborations in the text were considered in order to understand the phenomenon at a deep level (Smith et al., 2009). Similar themes were coded and grouped into clusters. Superordinate themes were established which connected a number of related themes together by looking at patterns across the data set. This analysis was primarily conducted by the author (EP) with review by and exploration with authors (SB, LV and IK).

Reflexivity

All interviews were undertaken by author (EP), as part of her Master of Applied Psychology (Clinical) thesis. Senior members of the team were clinical psychologists with several years' experience in clinical practice and conducting research with people with lived experience of voices, trauma, and PTSD and psychosis diagnoses. The authors understood that the profession of clinical psychology and prior work experience may lead to bias when collecting and interpreting the data in this study. In order to minimise any pre-existing ideas from influencing data analysis, the authors aimed to ensure that the analysis remained as close as possible to the participants' descriptions of their lived experience and a co-author (LV), with expertise in IPA and not in trauma and voices, was a key member of the analysis team.

RESULTS

A total of seven participants aged 27 to 68 years (M = 48.71 years, SD = 15.88) were interviewed for the study. Four participants were female (cisgender), two were male (cisgender), and one identified as demiboy. Interview times ranged from 33 to 90 min (M = 59.18 min; SD = 19.11 min). All participants were assigned a unique numeric identifier for quantitative data collection to ensure anonymity. While it was not an inclusion criterion for the study, all participants had received treatment at PVC. Furthermore, all participants had received some form of trauma-specific therapy, either at PVC or elsewhere. Further details of participant characteristics are contained in Table S1.

Data from these measures is presented in Table S2. Overall, participants scored higher on the HPSVQ (M = 24.14, SD = 6.87) than the sample of schizophrenia and schizoaffective disorder patients (M = 18.8, SD = 6.0) whose data was used to validate the measure (Van Lieshout & Goldberg, 2007), indicating the current sample was likely at the higher end of voice-hearing severity.

Four out of seven participants reported PTSD-related symptoms during interview. Three participants described flashbacks and one reported nightmares, while two identified symptoms which they perceived to be related to trauma but were not flashbacks, nightmares, images, or intrusive memories (i.e., somatic hallucinations). In addition, two participants reported chronic paranoia and related delusions, which were described as occurring in relation to the participants' voices. Scores on the PCL (cut score = 33; six participants scored higher than the cut score, refer Table S2) were comparable with those obtained in a study examining a large cohort of PTSD treatment-seeking military personnel (Wortmann et al., 2016), indicating the current sample was relatively acute regarding PTSD symptoms. This was congruent with participants' self-reported diagnoses of PTSD (i.e., five of the seven participants reported a previous PTSD diagnosis).

Six out of seven participants reported experiencing dissociative symptoms (i.e., losing track of time, detachment and derealisation). Scores on the DES-T indicated three participants experienced severe dissociative symptoms (refer Table S2). Participants also reported PTSD and psychosis symptoms during the interview phase of data collection (see Table S1).

One superordinate theme relevant to the research question was identified following analysis of the data. The superordinate theme, termed “interrelationship”, refers to the observed interplay described by participants between their voices and symptoms of PTSD or dissociation.

Superordinate theme: Voices occur in dynamic interrelationship with symptoms of PTSD and dissociation

This theme describes the nature of temporal relationships which occurred between participants' voices, PTSD symptoms and experiences of dissociation. The relationship patterns identified were portrayed as dynamic and reciprocal, rather than static and linear. There was no one consistent pattern of a relationship between voices and PTSD symptoms or dissociation. Rather, voices were frequently experienced before and after PTSD symptoms (e.g., flashbacks) and dissociative episodes. Moreover, for some participants, voices were perceived to trigger or amplify the PTSD symptom or dissociative experience (i.e., evidence of a dynamic relationship). Participant descriptions further illustrated a functional relationship between hearing voices, PTSD symptoms, and/or dissociation. Participants described voices as providing commentary on symptoms which were imminent—appearing to anticipate their onset—or immediately after their resolution, at times offering reassurance that symptoms had alleviated. For a few participants, dissociative states served as a coping mechanism that provided distance from distressing voices.

Subordinate theme 1: Voices occur in close temporal proximity to PTSD symptoms

For the four participants with PTSD intrusions (three with flashbacks; one with nightmares), voices often occurred adjacent to their intrusions—immediately before onset or after offset of the PTSD intrusion. Among participants who experienced flashbacks, John described a “dark voice” occurring prior, seemingly as a signal: “just before I have the flashback, I'll feel like the dark voice is saying, like, “here we go””. It appeared the voice—if not directly causing the flashback—was anticipating its occurrence and providing commentary accordingly. Conversely, Jane did not experience voices prior to a flashback event, but sometimes heard them immediately afterwards making supportive statements, which served to locate the traumatic experience related to the flashback (i.e., childhood abuse) in the past. Her flashbacks often occurred spontaneously, were quite vivid, and took the form of bodily memories (e.g., of being strangled). She referred to the voices that occurred afterwards as “the good ones”, which would say: “It's okay, that was back then, you are here now, do not go there … you are here and you are safe”. Similarly, Patricia described typically experiencing a trauma-related trigger (e.g., a sexually explicit joke) which precipitated her voices, and discrete flashbacks after hearing the voices: “I'll be driving home and the voices will start … And then I'll get home and I'll try and black it out and go to sleep. But then the flashbacks occur, and then I have the voices again. And then they continue until I end up, as I say, do whatever they say to do [self-harm]”. Each participant described a consistent pattern of voices discrete from but in close temporal proximity to their flashbacks. Laura, who reported trauma-related nightmares, similarly described voices occurring immediately before or after she woke up (i.e., following a nightmare): “[Voices are] usually before, because I'll know if I'm not in a good state. And when I wake up from my nightmare, the voices are there. Yeah. They're more after, because I'm asleep”.

Subordinate theme 2: Voices occur in close temporal proximity to dissociative episodes

Five of the six participants who described dissociative symptoms experienced these as occurring temporally adjacent to their voices. For John, voices were perceived to be causally related to dissociative episodes. He described an indirect relationship between voices and dissociation, whereby his “dark voice” would trigger negative emotional states—agitation and anxiety—which served to increase his vulnerability to dissociation. In these instances, dissociation was a way of managing (or detaching from) the distressing emotions caused by negative voices. According to John's account, voices were present both immediately prior to and after a typical dissociative event, which he described as follows: “It's … like a little safe area for me, that nothing gets you, it just [pause] There's just nothing. There's just these butterflies that fly around there. There's this blue one that lands on my finger and, and all of a sudden, I'll be back and the voices will be back … When I dissociate, like I say, they sort of disappear for a little while. But when I come back … they are just there”. This description also illustrates that while hearing voices and dissociation are temporally connected, for John they are distinct experiences.

Four participants recalled hearing voices while experiencing a dissociative state. Patricia and Jennifer described losing their sense of time due to focusing on their voices and the content of what the voices were saying (i.e., experiencing a sense of lost time or not realising how much time had passed by virtue of having been preoccupied by their voices for that period of time). Jennifer reported: “the longer I go on, listening to them [the voices] and speaking back to them, the more I lose sense of [time], you know, probably the more I'm drifting into a dissociative state”. Patricia specifically mentioned that she would lose track of time trying to “quieten” the voices and resist “doing what they are saying”. She reported: “By the time I do that, it's like six hours later, so I've lost five hours, six hours or whatever it is, I've lost it and I do not know where the hell it's gone because I've been … trying to get rid of the voices myself without ringing people”. Laura, who also experienced voices while dissociating, described a “split” in her sense of self which occurred when she felt unable to “handle” her voices:

It's like I'm still hearing them [voices], but then it's, it's not me hearing them. Yeah. Like they are there, but I cannot acknowledge them as being there because that's not me”. She goes on to say: “It does not happen that frequently. But when it does, it's usually because my voices have been so bad. And, I do not know, I feel free. And my voices are saying the usual things, but it's, it's over there. … It's hard to explain because I do not, like, I'm aware of it, but I'm not. But I feel like Laura's split in two. The voices and all the bad stuff, the paranoia and all that, it's over there. And right now, I'm Laura number two. And I'm free.

As evidenced in her description, the amplification in voice activity appears to trigger the dissociation—the “split-self” feeling—but the voices remain active during the dissociative episode. However, similar to John, for Laura hearing voices and dissociation are distinct experiences. While the voices cause Laura distress, the split in herself (the emergence of “Laura number two”) serves as a coping mechanism whereby she distances herself from the distressing voices.

In contrast to Laura's experience, Lexy described having dissociative experiences whereby a voice/persona would “take control” of their body and cause “heightened distress”. Lexy reported: “He'll [the voice] make it worse by telling me the basic things, “you are horrible”, “everyone hates you” … And it reaches the point that I just cannot deal with it anymore … and I disconnect. … I'll start shaking usually and then like, it'll feel like I'm being forced out of my body”.

In contrast to the other five participants, Jane did not experience voices in close temporal proximity to her dissociative episodes. She was the only participant to experience discrete episodes of dissociation without voices present. She described entering a state where “nothing else is existing around you, you are just in that … feel-good space … Nothing else exists outside that. It's almost like when you are … in an anaesthetic and you are coming out”.

DISCUSSION

This study sought to identify and describe how voices, PTSD symptoms, and dissociation may be temporally related. The results suggest both PTSD symptoms and dissociation are closely associated with voices in real time. As outlined above, the patterns of symptom onset and offset appeared dynamic and reciprocal, rather than linear. Among those participants who experienced flashbacks, voice episodes occurred immediately prior to or after the flashback. Moreover, these voice episodes appeared related to the symptom itself, with voices anticipating, precipitating, or commenting on the flashback. No participant reported voices occurring during a flashback, suggesting a clear demarcation between the two symptoms. However, there was some variability among participants who reported dissociation. One participant reported distinct episodes of dissociation absent of voices, while others reported experiencing voices during episodes of dissociation. This could be interpreted as the voice-hearing episodes can precipitate/trigger and maintain the dissociation. Another interpretation of the latter part of this finding is that voices were expressed through a dissociative part of the personality, in that voices were an integral part of the dissociation itself (Moskowitz et al., 2017; Moskowitz & Corstens, 2008).

These results are broadly consistent with the theoretical and experimental research literature. The finding that PTSD symptoms occur in close temporal relationship to voices is consistent with cross-sectional research which has found an association between voices and greater frequency of trauma memory intrusions (Peach et al., 2018). Furthermore, it supports the findings of an EMA study, which found trauma memory intrusions were associated with voice hearing in real time—specifically, voices were more likely to occur when trauma memory intrusions were more frequent across a specified timeframe of 1 h (Brand et al., 2020). In that study, directional effects were unable to be detected due to the study's design—as each data point represented a cross-sectional sampling of trauma memory intrusions and voices—meaning the temporal order of symptoms could not be identified. In the current study, among the four participants who experienced flashbacks or nightmares (i.e., PTSD intrusion symptoms), three experienced voices immediately after symptom offset, while one participant described entering a dissociative state immediately after the flashback, followed by voices. Three of the participants also experienced voices preceding the intrusion symptom. These patterns are congruent with the evidence that voices and PTSD symptoms are associated and, moreover, are suggestive of temporal relationships in both directions. Specifically, in this small sample, voices were reported to occur both directly before and after PTSD intrusion symptoms.

The findings are also consistent with research which suggests voices and dissociative symptoms are temporally associated (Jongeneel et al., 2020). It should be noted that there was greater variability in participants' accounts of their dissociation symptoms—as they pertained to voices—compared to PTSD intrusion symptoms. One participant described episodes of reduced awareness akin to a fugue state without voices (e.g., “it's like nothing else exists”, “like when … you are in an anaesthetic”), while others described experiences of either detachment (e.g., “I could hear the voices, but they were talking to someone else”) or absorption (e.g., “I hear them and I'm sort of there, imprisoned in myself”) in which voices were present. This variability poses the question as to whether voices are differentially related to dissociative states, depending on the type of dissociation experienced. Some researchers have proposed dissociation does not exist on a unitary spectrum but may comprise discrete subtypes of detachment and compartmentalisation (Brown, 2006; Holmes et al., 2005). Detachment (i.e., depersonalisation, derealisation) may be more common in PTSD than other dissociation symptoms (Anketell et al., 2011; Kilcommons & Morrison, 2005; Sun, Alvarez-Jimenez, Lawrence, et al., 2018), although this has been disputed (Pilton et al., 2015). It is possible that the observed differences in participants' descriptions of dissociation with or without voices in the current study reflected a qualitative difference in the type of dissociation being experienced. Nevertheless, voices appeared temporally related to dissociation regardless of the type of symptom being experienced. Moreover, for a few participants dissociation appeared to be a way of distancing themselves from distressing voices. Dissociation is considered a coping mechanism that allows people to detach from trauma-related distress (Schimmenti, 2018). It is thus possible that for these participants, the function of dissociation was to help them cope with the emotional distress caused by negative voices.

Two of the three participants (John and Patricia) who reported experiencing voices prior to flashbacks scored highest on the PTSD measure (PCL-5) and described their voices as “horrible” in the past week on the HPSVQ—the most negative rating of all of the participants. It is possible that voices for these participants represent post-traumatic intrusions. This aligns with theories that voices may be fragmented post-traumatic intrusions (Hardy, 2017), or decontextualised memories that are dissociated from conscious awareness (Longden et al., 2012). It is also consistent with evidence that voices may sometimes represent trauma-related flashbacks which have been disconnected from memory (Peach et al., 2021). Furthermore, all participants had received some form of therapy for voices and trauma; most had received imagery rescripting for traumatic memories. No association was found between therapy type and participants' descriptions of the connections between their voices, PTSD symptoms, and dissociative experiences. Given the small size of the current study and other potentially mediating variables not measured—such as psychological mindedness, emotion regulation, and therapeutic alliance—it is not possible to speculate further. However, this therapy may have contributed to how they conceptualised any connections between voice hearing, trauma, and dissociation.

One of the limitations of the current study is that it relied upon participant self-report to draw conclusions about temporal relationships between symptoms. There is a risk that participants may not have accurately recalled whether they heard voices prior to or after experiencing a flashback or dissociative episode—phenomena which are inherently subjective (Wearne et al., 2018). It is also possible that participants may have described what they believed to be a representative symptom episode—representative of the entirety of their experience—but in doing so may have omitted less frequent, inconsistent experiences; for instance, omitting episodes in which flashbacks were not accompanied by voices. Future research employing an EMA design examining the relationship between all three constructs—voice hearing, PTSD symptoms, and dissociation—may improve data reliability and corroborate symptom patterns. As others have noted, future research employing longitudinal designs to examine directional effects between co-occurring symptoms is warranted (Bloomfield et al., 2021; Pilton et al., 2015).

Certain characteristics of the current sample should be noted. This was an exploratory study examining voices, PTSD symptoms, and dissociation. Inclusion criteria pertained to voice hearing only; participants were not asked about PTSD or dissociation symptoms prior to interview. However, nearly all participants reported PTSD-type symptoms and six of the seven met criteria for PTSD based on the PCL-5 administered post-interview. Five of the participants reported having received a formal diagnosis of either PTSD or complex PTSD. In addition, all participants reported having received a form of trauma-focused therapy—either imagery rescripting or eye movement desensitisation and reprocessing (EMDR). All imagery rescripting therapy was administered by clinical psychologist and study author GPW. It is possible that the relative homogeneity of the sample (i.e., regarding diagnoses and PTSD symptom acuity) may have reduced the generalisability of the findings; different symptom patterns may have been observed in participants with sub-threshold PTSD symptoms. Likewise, having participants with a diagnosis of DID in the sample may have increased the richness and diversity of experiences, especially regarding dissociation. Having said this, we note that three participants reported high DES-T scores which may be consistent with a dissociative disorder diagnosis, and thus there is the possibility that they may have had an undiagnosed dissociative disorder. Moreover, it is possible that participants' descriptions of symptoms may, in part, have reflected not only therapeutic gains, but also the very fact of being socialised to psychological treatment—specifically, to talking about their trauma and voices. Future research examining the differential effects of therapy modalities on participants' descriptions of their voices, PTSD symptoms, and dissociative experiences may be of interest. For instance, in the current study, one participant referred to engaging in voice “dialoguing” as part of her therapy, consistent with new treatments which emphasise relational aspects of voice hearing and promote engagement with voices to reduce conflict (Longden et al., 2021). However, another participant described an experience of a therapist talking “directly” to their voices as unhelpful. These examples illustrate the variability among the therapeutic interventions administered to participants and their respective responses, which may have influenced how they conceptualised connections between their voices, PTSD symptoms, and dissociative experiences.

Overall, the findings of this qualitative study are consistent with quantitative research which has identified associations between voices, PTSD symptoms, and dissociation. Moreover, the current findings suggest some directional effects between symptoms—particularly in relation to PTSD symptoms—which have not been identified in quantitative research to date (Brand et al., 2020). Relationships identified between symptoms appeared dynamic and reciprocal, rather than static and linear. The relative consistency between participants regarding patterns of PTSD intrusion symptoms and voices should be noted, compared to the greater variability of dissociative symptoms and voices. Further research using real-time data sampling may be warranted to corroborate the temporal patterns identified in this exploratory study.

AUTHOR CONTRIBUTIONS

Emily Piesse: Conceptualization; data curation; formal analysis; methodology; project administration; writing – original draft; writing – review and editing. Georgie Paulik: Conceptualization; data curation; formal analysis; methodology; supervision; writing – original draft; writing – review and editing. Danielle Mathersul: Conceptualization; data curation; methodology; project administration; supervision; writing – original draft. Lee Valentine: Data curation; formal analysis; writing – original draft. Ilias Kamitsis: Formal analysis; writing – review and editing. Sarah Bendall: Conceptualization; data curation; formal analysis; methodology; supervision; writing – original draft; writing – review and editing.

ACKNOWLEDGEMENTS

The authors wish to thank all participants in this study. Without their generosity, this research would not have been possible. The contribution of two Lived Experience Consultants, who reviewed the study protocol and provided valuable feedback, is also gratefully acknowledged. Open access publishing facilitated by Murdoch University, as part of the Wiley - Murdoch University agreement via the Council of Australian University Librarians.

    CONFLICT OF INTEREST STATEMENT

    There are no conflicts of interest in the conducting or reporting of this research.

    DATA AVAILABILITY STATEMENT

    The data will be available throughout the editorial review process and at least 5 years after the date of publication.