Introduction
There is consistent evidence that developmental trauma (including maltreatment in childhood and adolescence) increases risk of psychosis (McGrath et al., Reference McGrath, McLaughlin, Saha, Aguilar-Gaxiola, Al-Hamzawi, Alonso and Kessler2017). People experiencing psychosis are more than twice as likely to have experienced developmental trauma (Varese et al., Reference Varese, Smeets, Drukker, Lieverse, Lataster, Viechtbauer and Bentall2012) and 15 times more likely to have experienced childhood sexual abuse (CSA) than people without psychosis (Bebbington et al., Reference Bebbington, Bhugra, Brugha, Singleton, Farrell, Jenkins and Meltzer2004). Evidence that developmental trauma causes psychosis fulfils Bradford Hill criteria (Hill, Reference Hill1965), including strong and consistent associations between trauma and psychosis (Varese et al., Reference Varese, Smeets, Drukker, Lieverse, Lataster, Viechtbauer and Bentall2012); temporal relationships (Kelleher et al., Reference Kelleher, Keeley, Corcoran, Ramsay, Wasserman, Carli and Cannon2013); dose effects (Duhig et al., Reference Duhig, Patterson, Connell, Foley, Capra, Dark and Scott2015; Longden, Sampson, & Read, Reference Longden, Sampson and Read2016; Schäfer & Fisher, Reference Schäfer and Fisher2011); and increased risk of conversion from at-risk states to first-episode psychosis (Brew, Doris, Shannon, & Mulholland, Reference Brew, Doris, Shannon and Mulholland2018). Developmental trauma may account for up to a third of psychosis cases (McGrath et al., Reference McGrath, McLaughlin, Saha, Aguilar-Gaxiola, Al-Hamzawi, Alonso and Kessler2017), and is associated with poor prognosis and treatment response (Cakir, Tasdelen Durak, Ozyildirim, Ince, & Sar, Reference Cakir, Tasdelen Durak, Ozyildirim, Ince and Sar2016; Misiak & Frydecka, Reference Misiak and Frydecka2016); the latter may be suggestive of distinct and/or additional neurobiological mechanisms underlying psychotic phenomena. Despite this, we lack a precise mechanistic understanding of how developmental trauma alters brain structure and function to give rise to psychosis. This may represent a barrier to developing more effective treatments for this patient group (Bloomfield et al., Reference Bloomfield, Yusuf, Srinivasan, Kelleher, Bell and Pitman2020).
Various brain alterations are associated with psychosis (Bloomfield, Buck, & Howes, Reference Bloomfield, Buck and Howes2016). Childhood and adolescence are sensitive periods for brain development (Goddings & Giedd, Reference Goddings and Giedd2014), including synaptic pruning, synaptogenesis, and myelination (Miller et al., Reference Miller, Duka, Stimpson, Schapiro, Baze, McArthur and Sherwood2012). Developmental trauma can disrupt brain development to cause lasting changes in structure and function (McCrory, Gerin, & Viding, Reference McCrory, Gerin and Viding2017; Teicher, Samson, Anderson, & Ohashi, Reference Teicher, Samson, Anderson and Ohashi2016). These include reduced volume of the hippocampus and anterior cingulate cortex, altered fiber tract density in the corpus callosum, and altered sensory systems (Teicher et al., Reference Teicher, Samson, Anderson and Ohashi2016). Animal research using stress paradigms indicates potential processes underlying these alterations include aberrant dendritic arborization and inhibition of neurogenesis (Czéh et al., Reference Czéh, Michaelis, Watanabe, Frahm, De Biurrun, Van Kampen and Fuchs2001; Magariños, McEwen, Flügge, & Fuchs, Reference Magariños, McEwen, Flügge and Fuchs1996). In parallel, several neurocognitive domains of direct relevance to psychotic symptomatology are particularly sensitive to the effects of developmental trauma, including aberrant amygdalar responsivity during threat processing, striatal reward processing dysfunction, impaired frontal emotion regulation, and executive control (McCrory et al., Reference McCrory, Gerin and Viding2017). Therefore, it is likely that developmental trauma results in changes to brain structure and function that can give rise to psychosis (Read, Fosse, Moskowitz, & Perry, Reference Read, Fosse, Moskowitz and Perry2014). It is, therefore, imperative to understand the underlying neurobiological mechanisms accounting for this.
It has been proposed that adult survivors of developmental trauma with psychosis represent a distinct clinical phenotype from those who have not experienced developmental trauma, underlined by differences in brain structure and function (Read et al., Reference Read, Fosse, Moskowitz and Perry2014). Such phenotypes have also been proposed clinically using subgroups such as ‘traumatic psychosis’, ‘neurodevelopmental psychosis’, and ‘psychotic PTSD’ to describe distinct manifestations of psychosis (Bloomfield et al., Reference Bloomfield, Chang, Woodl, Lyons, Cheng, Bauer-Staeb and Lewis2021; Stevens, Spencer, & Turkington, Reference Stevens, Spencer and Turkington2017). Given the implications of understanding underlying mechanisms for developing targeted treatments, we sought to address whether there are differences in brain structure and function within patients with psychosis according to whether they have or have not survived developmental trauma. Our hypothesis was that within people experiencing psychosis, there are structural and functional brain differences between those with or without a developmental trauma history. We tested this by systematically reviewing the neuroimaging literature of people experiencing psychosis with and without a history of developmental trauma.
Methods and materials
Search strategy
We preregistered this systematic review with PROSPERO (CRD42018105021) and followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) (Moher et al., Reference Moher, Liberati, Tetzlaff and Altman2009). Preliminary search was conducted on 3rd January 2019, using PubMed and MedLine (M.A.P.B., M.M), and a definitive search was conducted on 19th May 2024 using PubMed, MedLine, Web of Science, and PsychINFO via Ovid (M.A.P.B., T-Y. C.). For both searches, we used a combination of AND/OR operators using the search string: (child* OR adolesc* OR develop*) AND (schizophrenia OR psychosis OR ‘psychotic’) AND (abus* OR maltreat* OR trauma* OR advers* OR neglect) AND (‘gray matter’ OR ‘magnetic resonance imaging’ OR ‘connectivity’ OR ‘salience network’ OR ‘resting state’ OR ‘default mode’ OR ‘white matter’ OR ‘DTI’ OR ‘PET’ OR ‘SPECT’ OR ‘Computed Tomography’). The numerous possible iterations were also input individually into the search engines, and additional results that did not appear in systematic searches were noted. A librarian was consulted on the search strategy.
Selection criteria
We sought to address whether there were differences in brain structure and function between patients with psychosis who have survived developmental trauma compared to patients with psychosis who have not. We assessed studies against predetermined criteria for inclusion in the review: original studies in English up until 19th May 2024 of adult participants between the ages of 18–65; studies comparing or measuring brain structure and function in participants with psychosis, either in two groups (with or without developmental trauma) or along a gradient of developmental trauma exposure, via neuroimaging techniques specified in the search (MRI, DTI, PET, SPECT, or CT); studies were included of participants across the natural history of psychosis, including clinical high-risk states, first-episode or chronic stages; both medicated and unmedicated patients; and individuals at familial high risk. High-risk patients who had not experienced psychosis were included as there is evidence that trauma induces vulnerability to psychosis across the spectrum of severity (Bechdolf et al., Reference Bechdolf, Thompson, Nelson, Cotton, Simmons, Amminger and Yung2010). Developmental trauma was frequently defined using the Childhood Trauma Questionnaire Short Form, but there was substantial heterogeneity with other studies using alternative questionnaires, including the Childhood Life Events Questionnaire, the Early Trauma Inventory, the Traumatic Experiences Check-List, and the Childhood Experiences of Care and Abuse Questionnaire. Peer victimization or neighborhood-level exposures, such as crime, were not considered, as these are not reliably measured in many neuroimaging studies. Exclusion criteria were: studies without measures of brain structure and/or function; studies only including healthy participants; and studies involving both underage and adult participants that did not distinguish between age groups. In studies involving patient groups selected for a particular trait (e.g. history of violence), these data were disregarded, as they are not representative of the general patient population.
Screening of abstracts and full text was done by two researchers, APN and CW, using Covidence. The same two researchers performed the data extraction using a pre-made data extraction chart in Covidence. Any disagreements were then discussed with the lead author, MAPB, where a final decision was made.
Quality and risk of bias assessment
The Newcastle–Ottawa quality assessment scale was used to assess methodological quality and risk of bias (Wells et al., Reference Wells, Shea, O’Connell, Peterson, Welch, Losos and Tugwell2011). Rating results are presented below in Table 6.
Results
The selection process is presented in Figure 1. We identified 31 suitable published studies. Thirteen studies used structural magnetic resonance imaging (MRI), five used diffusion tensor imaging (DTI), eleven used functional MRI (fMRI), one used positron emission tomography (PET), and one study used MRI and fMRI. Nineteen of the included studies met criteria for good quality (Aas et al., Reference Aas, Kauppi, Brandt, Tesli, Kaufmann, Steen and Melle2017; Asmal et al., Reference Asmal, Kilian, du Plessis, Scheffler, Chiliza, Fouche and Emsley2019; Domen et al., Reference Domen, Michielse, Peeters, Viechtbauer, van Os and Marcelis2019; Egerton et al., Reference Egerton, Valmaggia, Howes, Day, Chaddock, Allen and McGuire2016; Frissen, van Os, Peeters, Gronenschild, & Marcelis, Reference Frissen, van Os, Peeters, Gronenschild and Marcelis2018; Habets, Marcelis, Gronenschild, Drukker, & Van Os, Reference Habets, Marcelis, Gronenschild, Drukker and Van Os2011; Hoy et al., Reference Hoy, Barrett, Shannon, Campbell, Watson, Rushe and Mulholland2012; Kumari et al., Reference Kumari, Gudjonsson, Raghuvanshi, Barkataki, Taylor, Sumich and Das2013; Kumari et al., Reference Kumari, Uddin, Premkumar, Young, Gudjonsson, Raghuvanshi and Das2014; Neilson et al., Reference Neilson, Bois, Gibson, Duff, Watson, Roberts and Lawrie2017; Peeters, et al., Reference Peeters, Gronenschild, Van De Ven, Habets, Goebel, Van Os and Marcelis2015a,Reference Peeters, van de Ven, Gronenschild, Patel, Habets and Goebelb; Quidé et al., Reference Quidé, O’Reilly, Rowland, Carr, Elzinga and Green2017a,Reference Quidé, Ong, Mohnke, Schnell, Walter, Carr and Greenb; Dauvermann et al., Reference Dauvermann, Mothersill, Rokita, King, Holleran, Kane and Donohoe2021; Quidé, Girshkin, Watkeys, Carr, & Green, Reference Quidé, Girshkin, Watkeys, Carr and Green2021; King et al., Reference King, Mothersill, Holleran, Patlola, McManus, Kenyon and Donohoe2022; Costello et al., Reference Costello, Dauvermann, Tronchin, Holleran, Mothersill, Rokita and Cannon2023; Xie et al., Reference Xie, Cai, Liu, Wei, Zhao, Dai and Li2023). All but three studies assessed substance use in patients (Aas et al., Reference Aas, Navari, Gibbs, Mondelli, Fisher, Morgan and Dazzan2012; Aas et al., Reference Aas, Kauppi, Brandt, Tesli, Kaufmann, Steen and Melle2017; Costello et al., Reference Costello, Dauvermann, Tronchin, Holleran, Mothersill, Rokita and Cannon2023). Out of 31 studies, 20 excluded participants with current and past substance dependence (Allen et al., Reference Allen, Azis, Modinos, Bossong, Bonoldi, Samson and McGuire2018; Asmal et al., Reference Asmal, Kilian, du Plessis, Scheffler, Chiliza, Fouche and Emsley2019; Barker et al., Reference Barker, Bois, Johnstone, Owens, Whalley, McIntosh and Lawrie2016a,Reference Barker, Bois, Neilson, Johnstone, Owens, Whalley and Lawrieb; Cancel et al., Reference Cancel, Comte, Boutet, Schneider, Rousseau, Boukezzi and Fakra2017; Dauvermann et al., Reference Dauvermann, Mothersill, Rokita, King, Holleran, Kane and Donohoe2021; King et al., Reference King, Mothersill, Holleran, Patlola, McManus, Kenyon and Donohoe2022; Kumari et al., Reference Kumari, Gudjonsson, Raghuvanshi, Barkataki, Taylor, Sumich and Das2013; Kumari et al., Reference Kumari, Uddin, Premkumar, Young, Gudjonsson, Raghuvanshi and Das2014; Poletti et al., Reference Poletti, Mazza, Bollettini, Locatelli, Cavallaro, Smeraldi and Benedetti2015; Quidé et al., Reference Quidé, Girshkin, Watkeys, Carr and Green2021; Quidé et al., Reference Quidé, O’Reilly, Rowland, Carr, Elzinga and Green2017a,Reference Quidé, Ong, Mohnke, Schnell, Walter, Carr and Greenb; Quidé, O’Reilly, Watkeys, Carr, & Green, Reference Quidé, O’Reilly, Watkeys, Carr and Green2018; Ruby, Rothman, Corcoran, Goetz, & Malaspina, Reference Ruby, Rothman, Corcoran, Goetz and Malaspina2017; Xie et al., Reference Xie, Cai, Liu, Wei, Zhao, Dai and Li2023) and seven studies covaried for drug use (Domen et al., Reference Domen, Michielse, Peeters, Viechtbauer, van Os and Marcelis2019; Egerton et al., Reference Egerton, Valmaggia, Howes, Day, Chaddock, Allen and McGuire2016; Frissen et al., Reference Frissen, van Os, Peeters, Gronenschild and Marcelis2018; Habets et al., Reference Habets, Marcelis, Gronenschild, Drukker and Van Os2011; Neilson et al., Reference Neilson, Bois, Gibson, Duff, Watson, Roberts and Lawrie2017; Peeters et al., Reference Peeters, Gronenschild, Van De Ven, Habets, Goebel, Van Os and Marcelis2015a,Reference Peeters, van de Ven, Gronenschild, Patel, Habets and Goebelb).
PRISMA 2020 flow diagram for new systematic reviews, which included searches of databases and registers only.
*Consider, if feasible to do so, reporting the number of records identified from each database or register searched (rather than the total number across all databases/registers).
**If automation tools were used, indicate how many records were excluded by a human and how many were excluded by automation tools.
Structural imaging
Cortical structure
All of three studies investigating cortical structure found that compared to people with psychosis who had not been exposed to developmental trauma (Table 1), developmental trauma is associated with reduced thickness or surface area of the cortex (Barker et al., Reference Barker, Bois, Johnstone, Owens, Whalley, McIntosh and Lawrie2016a; Habets et al., Reference Habets, Marcelis, Gronenschild, Drukker and Van Os2011 ; Neilson et al., Reference Neilson, Bois, Gibson, Duff, Watson, Roberts and Lawrie2017). In one high-quality study (n = 88), people with psychosis exhibited a negative relationship between cortical thickness and developmental trauma, whereas healthy siblings exhibited a positive relationship (Habets et al., Reference Habets, Marcelis, Gronenschild, Drukker and Van Os2011). Similarly, another high-quality study (n = 99) found that in patients with psychosis, developmental trauma reduces the cortical thickness of the right temporal lobe, whereas in healthy controls, an opposite pattern was observed (Neilson et al., Reference Neilson, Bois, Gibson, Duff, Watson, Roberts and Lawrie2017). In both studies (Habets et al., Reference Habets, Marcelis, Gronenschild, Drukker and Van Os2011; Neilson et al., Reference Neilson, Bois, Gibson, Duff, Watson, Roberts and Lawrie2017), there was no relationship between antipsychotic drug exposure and cortical thickness, suggesting that pharmacotherapy was unlikely to be driving the effects. The third study (n = 145) of participants at high familial risk of psychosis used social services involvement as a proxy marker of developmental trauma, finding reduced bilateral hemispheric surface area in the developmental trauma survivor group relative to individuals not exposed to trauma (Barker et al., Reference Barker, Bois, Johnstone, Owens, Whalley, McIntosh and Lawrie2016a). This study found no difference in cortical thickness between the two groups.
Cortical and regional structure findings using magnetic resonance imaging
Abbreviations: BPAD, Bipolar affective disorder; CSA, Childhood sexual abuse; DT, Developmental trauma; FEP, First-episode psychosis; fHR, Familial high-risk; GMV, Gray matter volume; HC, Healthy controls; MRI, Magnetic resonance imaging; P + DT, Psychosis and developmental trauma; P − DT, Psychosis without developmental trauma; SCZ, Schizophrenia; SZA, Schizoaffective disorder; SZP, Schizophreniform disorder.
Global gray matter volume
There is high-quality evidence for the association between developmental trauma and reduced gray matter volume (GMV) across any brain region (Table 1) (Benedetti et al., Reference Benedetti, Radaelli, Poletti, Falini, Cavallaro, Dallaspezia and Smeraldi2011; Cancel et al., Reference Cancel, Comte, Truillet, Boukezzi, Rousseau, Zendjidjian and Fakra2015; Frissen et al., Reference Frissen, van Os, Peeters, Gronenschild and Marcelis2018; Kumari et al., Reference Kumari, Gudjonsson, Raghuvanshi, Barkataki, Taylor, Sumich and Das2013; Kumari et al., Reference Kumari, Uddin, Premkumar, Young, Gudjonsson, Raghuvanshi and Das2014; Poletti et al., Reference Poletti, Vai, Smeraldi, Cavallaro, Colombo and Benedetti2016; Ruby et al., Reference Ruby, Rothman, Corcoran, Goetz and Malaspina2017; Sheffield, Williams, Woodward, & Heckers, Reference Sheffield, Williams, Woodward and Heckers2013).
Frontal gray matter volume
Findings of reduced regional GMV in people with psychosis who experienced developmental trauma were reported in four out of five studies (Table 1) (Benedetti et al., Reference Benedetti, Radaelli, Poletti, Falini, Cavallaro, Dallaspezia and Smeraldi2011; Cancel et al., Reference Cancel, Comte, Truillet, Boukezzi, Rousseau, Zendjidjian and Fakra2015; Poletti et al., Reference Poletti, Vai, Smeraldi, Cavallaro, Colombo and Benedetti2016; Sheffield et al., Reference Sheffield, Williams, Woodward and Heckers2013). One study had a large (n = 302) sample of medicated participants with chronic (> 10 years) bipolar disorder or schizophrenia (Poletti et al., Reference Poletti, Vai, Smeraldi, Cavallaro, Colombo and Benedetti2016). Compared to patients with psychosis with low levels of developmental trauma (P − DT), patients with psychosis who have survived developmental trauma (P + DT) showed reduced GMV in the orbitofrontal cortex (OFC), when compared to healthy controls (Poletti et al., Reference Poletti, Vai, Smeraldi, Cavallaro, Colombo and Benedetti2016). There was some evidence that specific types of trauma were associated with region-specific alterations in brain structure. For example, exposure to CSA was associated with reduced prefrontal GMV (Sheffield et al., Reference Sheffield, Williams, Woodward and Heckers2013), while exposure to emotional neglect was associated with reduced GMV in the dorsolateral prefrontal cortex (DLPFC) which in turn mediated the severity of disorganization symptoms (Cancel et al., Reference Cancel, Comte, Truillet, Boukezzi, Rousseau, Zendjidjian and Fakra2015). Another high-quality study reported a negative association between ratings of psychosocial deprivation and GMV in the left inferior frontal region and left middle frontal precentral gyri (Kumari et al., Reference Kumari, Gudjonsson, Raghuvanshi, Barkataki, Taylor, Sumich and Das2013).
Medial temporal and subcortical gray matter volume
Four out of five studies found that P + DT were associated with smaller medial temporal volumes, specifically reduced amygdala and/or hippocampus volumes (Table 1) (Aas et al., Reference Aas, Navari, Gibbs, Mondelli, Fisher, Morgan and Dazzan2012; Barker et al., Reference Barker, Bois, Johnstone, Owens, Whalley, McIntosh and Lawrie2016a; Hoy et al., Reference Hoy, Barrett, Shannon, Campbell, Watson, Rushe and Mulholland2012; Kumari et al., Reference Kumari, Gudjonsson, Raghuvanshi, Barkataki, Taylor, Sumich and Das2013). There was high-quality evidence of negative relationships between trauma exposure and the volumes of these structures from two studies (Aas et al., Reference Aas, Navari, Gibbs, Mondelli, Fisher, Morgan and Dazzan2012; Kumari et al., Reference Kumari, Gudjonsson, Raghuvanshi, Barkataki, Taylor, Sumich and Das2013). Importantly, one high-quality study (Hoy et al., Reference Hoy, Barrett, Shannon, Campbell, Watson, Rushe and Mulholland2012) (n = 21) of medicated FEP patients reported that 24% of patients met PTSD criteria (using the Posttraumatic Diagnostic Scale) (Foa, Riggs, Dancu, & Rothbaum, Reference Foa, Riggs, Dancu and Rothbaum1993) in relation to their developmental trauma experiences.
For subcortical structures, in one large study (n = 302) (Poletti et al., Reference Poletti, Vai, Smeraldi, Cavallaro, Colombo and Benedetti2016), reduced thalamic GMV was found in P + DT relative to healthy trauma survivors (Table 1). No studies were found measuring or reporting alterations in striatal structures.
White matter
In three out of five studies, there is high-quality evidence that within patients with psychosis, developmental trauma is associated with reduced white matter microstructure measured as reduced functional anisotropy (FA) and increased mean diffusivity (MD) (Table 2) (Asmal et al., Reference Asmal, Kilian, du Plessis, Scheffler, Chiliza, Fouche and Emsley2019; Domen et al., Reference Domen, Michielse, Peeters, Viechtbauer, van Os and Marcelis2019; Poletti et al., Reference Poletti, Mazza, Bollettini, Locatelli, Cavallaro, Smeraldi and Benedetti2015). One study (n = 83) found that connectivity was inversely related to the degree of developmental trauma in white matter tracts linking gray matter structures that also exhibit volumetric deficits described above (Poletti et al., Reference Poletti, Mazza, Bollettini, Locatelli, Cavallaro, Smeraldi and Benedetti2015), including the corpus callosum, cingulum, corona radiata, inferior longitudinal fasciculus, and thalamic radiation. Importantly, there was also prospective evidence of inverse relationship between level of trauma and mean FA observed over time in the patient group (Domen et al., Reference Domen, Michielse, Peeters, Viechtbauer, van Os and Marcelis2019), which was not observed in other groups and remained significant when controlling for medication. Together, both of these studies suggest a dose–response effect of trauma exposure on the extent of white matter alterations. There was also some evidence for particular types of trauma being associated with patterns of structural connectivity from one study (n = 54) of minimally medicated FEP patients (< 4 weeks cumulative lifetime exposure to dopamine antagonists) (Asmal et al., Reference Asmal, Kilian, du Plessis, Scheffler, Chiliza, Fouche and Emsley2019). In that high-quality study, sexual abuse was associated with reduced FA in the inferior fronto-occipital fasciculus, inferior longitudinal fasciculus, and the superior longitudinal fasciculus, whilst emotional neglect was associated with increased FA in the right superior longitudinal fasciculus, relative to patients without experiences of developmental trauma. Two studies found that although both patients with psychosis and individuals with a history of DT showed reduced FA in similar areas, there were no differences between patients with psychosis, as well as DT, and those without (Costello et al., Reference Costello, Dauvermann, Tronchin, Holleran, Mothersill, Rokita and Cannon2023; Xie et al., Reference Xie, Cai, Liu, Wei, Zhao, Dai and Li2023).
Structural connectivity findings using diffusion tensor imaging
Abbreviations: CSA, Childhood sexual abuse; DT, Developmental trauma; DTI, Diffusion tensor imaging; EN, Emotional neglect; FA, Fractional anisotropy; FEP, First-episode psychosis; HC, Healthy controls; MD, Mean diffusivity; MDD, Major depressive disorder; P + DT, Psychosis and developmental trauma; P − DT, Psychosis without developmental trauma; SCZ, Schizophrenia; SZA, Schizoaffective disorder.
Functional imaging
Cerebral perfusion
One study in UHR (n = 77) measured resting state cerebral perfusion of the hippocampus, basal ganglia, and midbrain using arterial spin labelling (Allen et al., Reference Allen, Azis, Modinos, Bossong, Bonoldi, Samson and McGuire2018), relative to healthy volunteers (Table 3). Participants on antipsychotic medication were excluded from the final analyses. There was a positive relationship between level of developmental trauma (CTQ score) and resting state cerebral blood flow in the right hippocampus/subiculum and left parahippocampal gyrus in the UHR group. The whole brain analysis found a negative association between developmental trauma and perfusion in a cluster encompassing the left IFG and superior/medial PFC in the UHR group.
Arterial spin labelling imaging findings
Abbreviations: BPD, Borderline personality disorder, BPAD, Bipolar affective disorder, DT, Developmental trauma, HC, Healthy controls, P + DT, Psychosis and developmental trauma, P − DT, Psychosis without developmental trauma, SCZ, Schizophrenia, SZA, Schizoaffective disorder, UHR, Ultra-high risk.
Resting state
Two high-quality fMRI studies investigated the effects of developmental trauma on functional connectivity in the same sample (n = 228) of patients with schizophrenia (Peeters, et al., Reference Peeters, Gronenschild, Van De Ven, Habets, Goebel, Van Os and Marcelis2015a,Reference Peeters, van de Ven, Gronenschild, Patel, Habets and Goebelb). While there is no significant association between developmental trauma and functional connectivity between regions of the default mode network (Peeters et al., Reference Peeters, van de Ven, Gronenschild, Patel, Habets and Goebel2015b), there is a positive association between trauma exposure and nucleus accumbens–lentiform nucleus connectivity (Peeters et al., Reference Peeters, Van De Ven, Habets, Goebel, Van Os and Marcelis2013). A further study showed increased connectivity between the medial prefrontal cortex and the cerebellum in patients with schizophrenia with high levels of trauma in comparison to those with low levels of trauma (Dauvermann et al., Reference Dauvermann, Mothersill, Rokita, King, Holleran, Kane and Donohoe2021).
Emotional processing
Four of the fMRI studies investigated emotional processing using face matching (Table 4) (Aas et al., Reference Aas, Kauppi, Brandt, Tesli, Kaufmann, Steen and Melle2017; Benedetti et al., Reference Benedetti, Radaelli, Poletti, Falini, Cavallaro, Dallaspezia and Smeraldi2011; Cancel et al., Reference Cancel, Comte, Boutet, Schneider, Rousseau, Boukezzi and Fakra2017; Quidé et al., Reference Quidé, Girshkin, Watkeys, Carr and Green2021). A large, high-quality study (n = 101) of mostly medicated patients with schizophrenia or bipolar spectrum diagnoses found task-induced hyperactivation when differentiating between responses to negative and positive emotional valence in middle temporal and lateral occipital cortex, which was associated with trauma exposure (Aas et al., Reference Aas, Kauppi, Brandt, Tesli, Kaufmann, Steen and Melle2017). One functional connectivity analysis in a smaller sample (n = 21) of medicated schizophrenia patients found that CSA was dose-dependently associated with hypoconnectivity between the amygdala–left posterior cingulate cortex/precuneus and amygdala–right calcarine sulcus (Cancel et al., Reference Cancel, Comte, Boutet, Schneider, Rousseau, Boukezzi and Fakra2017). One study showed that in response to a stressor (an MRI session), patients with high levels of DT showed decreased activation in bilateral temporo-parietal-insular junctions, right middle cingulum, right pre–postcentral gyrus, and left cerebella lobules IV–VI, while there was increased activation in patients with low levels of DT (Quidé et al., Reference Quidé, Girshkin, Watkeys, Carr and Green2021). Finally, Benedetti and colleagues (Benedetti et al., Reference Benedetti, Radaelli, Poletti, Falini, Cavallaro, Dallaspezia and Smeraldi2011) investigated the amygdala, hippocampus, ACC, and PFC as ROIs in a small sample (n = 20) of medicated patients with schizophrenia using fearful and angry faces. Comparing high and low trauma groups, trauma-specific ACC and PFC hyperactivation was found relative to both patient and control groups without trauma, which remained significant when controlling for medication. Collectively, within patients with psychosis, developmental trauma is associated with alterations in emotional processing. MRI findings have recently extended to investigate default mode network hubs in an affective theory of mind (ToM) task in medicated patients with schizophrenia or schizoaffective disorder (n = 47) (Quidé et al., Reference Quidé, Ong, Mohnke, Schnell, Walter, Carr and Green2017b). This high-quality study found a relationship between trauma exposure and posterior cingulate hyperactivation in patients, suggesting that developmental trauma may result in functional brain changes contributing to abnormal self-oriented mental imagery. In the whole brain analysis, trauma exposure was associated with superior frontal hyperactivation and temporo-parietal hypoactivation.
Functional magnetic resonance imaging findings
Executive processing
Response inhibition and working memory have been investigated for which both studies were significant. An ROI analysis of a large sample (n = 112) during a Go/No-Go Flanker task showed that developmental trauma was associated with hyperactivation of the left inferior frontal gyrus (IFG) (Table 4) (Quidé et al., Reference Quidé, O’Reilly, Watkeys, Carr and Green2018). Task-induced IFG hyperactivation was associated with general symptom severity within P + DT, but also in P − DT. A separate study (n = 92) investigated default mode network hubs during visuo-spatial working memory processing (Quidé et al., Reference Quidé, O’Reilly, Rowland, Carr, Elzinga and Green2017a). Trauma exposure was associated with increases in activation of the left inferior parietal lobule, without a behavioral difference in working memory performance between groups, possibly reflecting reduced cortical efficiency and/or compensatory mechanisms.
Molecular imaging
We identified one high-quality PET study (Egerton et al., Reference Egerton, Valmaggia, Howes, Day, Chaddock, Allen and McGuire2016) of striatal dopamine synthesis capacity in UHR (n = 47) reporting a measure of developmental trauma (Table 5) (Bifulco, Brown, & Harris, Reference Bifulco, Brown and Harris1994). Developmental trauma was associated with elevated striatal dopamine synthesis capacity, particularly in the associative functional striatal subdivision (that is dorsal caudate and putamen), compared to low exposure. However, there was no significant difference in dopamine function between the ultra-high-risk participants who survived developmental trauma and controls with traumatic experiences.
Molecular imaging using 18-F DOPA PET
Abbreviations: CAARMS, Comprehensive assessment of at-risk mental states; DT, Developmental trauma; PET, Positron emission tomography; UHR, Ultra-high risk.
Quality and risk of bias assessment results using the Newcastle–Ottawa assessment scale
Discussion
Our neuroimaging review investigated brain structure and function in survivors of developmental trauma with psychosis across the whole psychotic spectrum including ultra-high-risk individuals. We have found evidence in support of our hypothesis that there are differences in brain structure and function in adults with psychosis who have or have not survived developmental trauma. These included small global, frontal, and subcortical volumes, low corticolimbic connectivity, and alterations in brain function during cognitive processing. Whilst the majority of studies were cross-sectional, there was high-quality prospective evidence of putative trauma-related effects alongside dose effects of trauma exposure on changes in brain structure (Domen et al., Reference Domen, Michielse, Peeters, Viechtbauer, van Os and Marcelis2019) which may suggest causation.
There are several possible interpretations for our findings that there appear to be neuroimaging differences between people with psychosis who report having or having not survived developmental trauma, which are not mutually exclusive. First, trauma-induced changes in brain structure and function may induce vulnerability to psychosis. This interpretation would be consistent with findings from other studies that developmental trauma exposure is associated with changes in structure and function in circuits that are implicated in psychosis (Bloomfield et al., Reference Bloomfield, Buck and Howes2016; Teicher & Samson, Reference Teicher and Samson2016; Xie et al., Reference Xie, Cai, Liu, Wei, Zhao, Dai and Li2023). Moreover, additive interactions with genetic and other environmental factors could possibly also lead to increased illness severity. This is because global brain volume reductions are observed in schizophrenia (Giedd et al., Reference Giedd, Jeffries, Blumenthal, Castellanos, Vaituzis, Fernandez and Rapoport1999) and in those at genetic risk (Cooper, Barker, Radua, Fusar-Poli, & Lawrie, Reference Cooper, Barker, Radua, Fusar-Poli and Lawrie2014). Accumulated trauma-induced changes (Liberzon & Sripada, Reference Liberzon and Sripada2007) may have an additive effect on such volume reductions (Ruby et al., Reference Ruby, Rothman, Corcoran, Goetz and Malaspina2017). This could represent neurobiological pathways to poorer prognosis (Cakir et al., Reference Cakir, Tasdelen Durak, Ozyildirim, Ince and Sar2016; Misiak & Frydecka, Reference Misiak and Frydecka2016). Thus, trauma-induced alterations in brain structure and function may underlie worsened psychosis symptomatology following trauma (Duhig et al., Reference Duhig, Patterson, Connell, Foley, Capra, Dark and Scott2015). Findings of reduced cortical thickness and hippocampal may also provide underlying neurobiological changes to match to a distinct traumatogenic phenotype. However, further research identifying the clinical factors and response to treatment is required to explore this possibility further.
However, it remains unknown which trauma-induced changes in brain structure and function may be associated with resilience. Therefore, an alternative explanation is that brain changes associated with trauma may be adaptations and not necessarily pathological in otherwise healthy individuals. It is also possible that the findings reviewed here are not related to trauma per se but may be due to pre-existing (intrinsic) differences in brain structure and function, and/or variance in etiology of psychotic disorders. Intrinsic brain variations preceding trauma exposure may serve as risk factors underlying the development of psychotic symptoms following the experience of a traumatic event. Prospective, longitudinal studies are needed to elucidate possible phenotypes associated the development of psychosis following trauma exposure. Furthermore, due to the clinical overlap between PTSD and psychosis, these findings may be driven by PTSD. Importantly, PTSD symptoms are often overlooked in patients with psychosis, and studies of patients with psychosis often do not report PTSD symptoms (Zammit et al., Reference Zammit, Lewis, Dawson, Colley, McCann, Piekarski and Bisson2018). Future studies are needed to measure the relationships between developmental trauma, psychopathology, resilience, and alterations in neurobiology to investigate this further.
Findings of divergent brain alterations in patients reporting developmental trauma, according to the presence or absence of psychosis, are striking (Domen et al., Reference Domen, Michielse, Peeters, Viechtbauer, van Os and Marcelis2019; Habets et al., Reference Habets, Marcelis, Gronenschild, Drukker and Van Os2011). One possible interpretation of these findings is that they may reflect resilience and/or compensatory mechanisms, and further work is needed to understand underlying processes. Whilst speculative, possible resilience and vulnerability factors may include susceptibility to stress-induced changes in dendritic arborization and neuronal migration (Lyall et al., Reference Lyall, Shi, Geng, Woolson, Li, Wang and Gilmore2015). Taken together, it remains unknown if potential differences in brain structure and function are due to an additive effect of developmental trauma on psychosis symptomatology or whether psychosis following developmental trauma represents a distinct clinical phenotype and further investigation, including genetic and longitudinal research, is needed to address this.
This study has implications for understanding the neurocognitive processes underlying how developmental trauma may cause psychosis, including through executive and threat processing. In terms of executive function, the PFC is one of the final cortical structures to mature (Huttenlocher, Reference Huttenlocher1990), rendering it especially vulnerable to stressors during development (McCrory et al., Reference McCrory, Gerin and Viding2017; Teicher & Samson, Reference Teicher and Samson2016), and there is evidence that neglect-induced reductions in DLPFC GMV are associated with disorganization symptoms in patients (Cancel et al., Reference Cancel, Comte, Truillet, Boukezzi, Rousseau, Zendjidjian and Fakra2015). Hyperactivation in executive function domains of working memory and response inhibition may reflect attempted compensatory mechanisms necessary to maintain similar levels of behavioral performance on such tasks (Quidé et al., Reference Quidé, O’Reilly, Watkeys, Carr and Green2018; Quidé, et al., Reference Quidé, O’Reilly, Rowland, Carr, Elzinga and Green2017a). Since the PFC is critical for executive function and emotion regulation, dysfunction in regions where there are structural alterations associated with developmental trauma may underlie cognitive impairments (Benedetti et al., Reference Benedetti, Radaelli, Poletti, Falini, Cavallaro, Dallaspezia and Smeraldi2011; Dannlowski et al., Reference Dannlowski, Stuhrmann, Beutelmann, Zwanzger, Lenzen, Grotegerd and Kugel2012; Üçok et al., Reference Üçok, Kaya, Uğurpala, Çıkrıkçılı, Ergül, Yokuşoğlu and Direk2015). Moreover, there is evidence that the PFC is involved in fear extinction (Fullana et al., Reference Fullana, Albajes-Eizagirre, Soriano-Mas, Vervliet, Cardoner, Benet and Harrison2018). Deficits in PFC GMV associated with psychosis patients with developmental trauma may, thus, be an explanation for the maintenance of paranoia.
Findings of smaller hippocampal volumes (Aas et al., Reference Aas, Navari, Gibbs, Mondelli, Fisher, Morgan and Dazzan2012; Barker et al., Reference Barker, Bois, Johnstone, Owens, Whalley, McIntosh and Lawrie2016a; Hoy et al., Reference Hoy, Barrett, Shannon, Campbell, Watson, Rushe and Mulholland2012; Ruby et al., Reference Ruby, Rothman, Corcoran, Goetz and Malaspina2017) are in keeping with studies in adolescent survivors of developmental trauma without psychosis (Opel et al., Reference Opel, Redlich, Zwanzger, Grotegerd, Arolt, Heindel and Dannlowski2014; Teicher et al., Reference Teicher, Anderson, Ohashi, Khan, McGreenery, Bolger and Vitaliano2018) and adults with PTSD or a dissociative disorder (Logue et al., Reference Logue, van Rooij, Dennis, Davis, Hayes, Stevens and Morey2018; Pitman et al., Reference Pitman, Rasmusson, Koenen, Shin, Orr, Gilbertson and Liberzon2012). Glucocorticoid exposure impairs neuronal growth (Czéh et al., Reference Czéh, Michaelis, Watanabe, Frahm, De Biurrun, Van Kampen and Fuchs2001), and the hippocampus is highly sensitive to excessive glucocorticoids (Sapolsky, Reference Sapolsky1996). FEP patients who have survived developmental trauma showed reduced levels of brain-derived neurotrophic factor (BDNF) combined with elevated levels of cortisol, which predicted smaller hippocampal volumes (Mondelli et al., Reference Mondelli, Cattaneo, Murri, Di Forti, Handley, Hepgul and Pariante2011). Hippocampal atrophy is the most consistently reported structural finding in PTSD (Pitman et al., Reference Pitman, Rasmusson, Koenen, Shin, Orr, Gilbertson and Liberzon2012) whereby reduced hippocampal volume may predispose individuals to PTSD and there is also evidence that trauma exposure further reduces hippocampal volumes (Pitman et al., Reference Pitman, Rasmusson, Koenen, Shin, Orr, Gilbertson and Liberzon2012). Furthermore, a recent study found evidence for hippocampal sensitive periods in early life, during which traumatic experiences were associated with reduced hippocampal volume (Humphreys et al., Reference Humphreys, King, Sacchet, Camacho, Colich, Ordaz and Gotlib2019). It is, therefore, possible that hippocampal dysfunction could give rise to psychotic experiences. These findings are also consistent with evidence that PTSD symptoms may be involved in the relationship between developmental trauma and psychosis (Bloomfield et al., Reference Bloomfield, Chang, Woodl, Lyons, Cheng, Bauer-Staeb and Lewis2021). However, given that stress results in changes to hippocampal structure in several psychiatric disorders (Geuze, Vermetten, & Bremner, Reference Geuze, Vermetten and Bremner2005), it is possible that these findings are not specific to psychosis and further research is needed to address the potential role in reduced hippocampal volume in the pathophysiology of psychosis associated with developmental trauma.
Paranoia is a key symptom of psychosis causing high levels of distress (Freeman, Garety, Kuipers, Fowler, & Bebbington, Reference Freeman, Garety, Kuipers, Fowler and Bebbington2002) and is associated with developmental trauma (Read & Argyle, Reference Read and Argyle1999). We found evidence that psychosis in patients reporting developmental trauma is associated small amygdalar volumes (Aas et al., Reference Aas, Navari, Gibbs, Mondelli, Fisher, Morgan and Dazzan2012; Barker et al., Reference Barker, Bois, Johnstone, Owens, Whalley, McIntosh and Lawrie2016a; Hoy et al., Reference Hoy, Barrett, Shannon, Campbell, Watson, Rushe and Mulholland2012; Kumari et al., Reference Kumari, Gudjonsson, Raghuvanshi, Barkataki, Taylor, Sumich and Das2013) and hyperactivation during threat processing (Aas et al., Reference Aas, Kauppi, Brandt, Tesli, Kaufmann, Steen and Melle2017; Benedetti et al., Reference Benedetti, Radaelli, Poletti, Falini, Cavallaro, Dallaspezia and Smeraldi2011). One explanation is that putative trauma-induced structural changes during sensitive periods occur alongside sensitization of threat processing (Humphreys et al., Reference Humphreys, King, Sacchet, Camacho, Colich, Ordaz and Gotlib2019; Teicher et al., Reference Teicher, Samson, Anderson and Ohashi2016). These findings are in keeping with findings that emotional dysregulation is involved in the link between developmental trauma and psychosis (Bloomfield et al., Reference Bloomfield, Chang, Woodl, Lyons, Cheng, Bauer-Staeb and Lewis2021). Potentiated threat detection may develop in adverse environments via attentional biases, as survivors exhibit faster identification of negative valence emotional stimuli than nonmaltreated controls (Masten et al., Reference Masten, Guyer, Hodgdon, McClure, Charney, Ernst and Monk2008). Small ACC volumes may result in impaired top-down amygdalar inhibition which would hyper-potentiate threat detection, removing the ‘brakes’ on an already accelerated system (Humphreys et al., Reference Humphreys, King, Sacchet, Camacho, Colich, Ordaz and Gotlib2019). Reduced activation in the temporo-parieto-insular junctions could also contribute to poorer emotional/threat processing further compounding these issues (Quidé et al., Reference Quidé, Girshkin, Watkeys, Carr and Green2021). Animal models suggest that this may arise from impaired GABA-based inhibition during fear learning (Piantadosi & Floresco, Reference Piantadosi and Floresco2014). This interpretation is consistent with PTSD models that describe a hyper-responsive amygdala alongside a hypo-responsive PFC to threat (Liberzon & Sripada, Reference Liberzon and Sripada2007).
Strengths and limitations
Key strengths of our review include preregistration and the synthesis of multimodal neuroimaging literature across phases of psychosis, adding to previous reviews (Cancel, Dallel, Zine, El-Hage, & Fakra, Reference Cancel, Dallel, Zine, El-Hage and Fakra2019; Thomas et al., Reference Thomas, Rakesh, Whittle, Sheridan, Upthegrove and Cropley2023). However, this review is not without its limitations. These relate to the existing neuroimaging data and to the directionality of causal relationships between traumata, the brain, and psychotic symptoms.
In terms of limitations of the field, there is currently a lack of studies that distinguish between different experiences of developmental trauma in terms of type, severity, and age of exposure. The grouping together of different experiences of trauma into a single entity is a limitation of the current literature study given that different experiences of trauma at different ages are likely to be associated with different effects on the development of brain structure and function. The heterogeneity and small number of published studies for each of the MRI methods is a further limitation. This also limits our ability to detect publication bias across our review. In terms of causality, the cross-sectional nature of the body of research presented here precludes causal inferences of directionality of dynamic changes in brain structure and function assumed to be associated with developmental trauma. We cannot exclude the possibility of reverse causation whereby putative differences in brain structure and function are not caused by trauma exposure, but rather increase risk that a child will be maltreated (Kelleher et al., Reference Kelleher, Keeley, Corcoran, Ramsay, Wasserman, Carli and Cannon2013). There are also a range of possible confounds in this field. For example, given that trauma incidence is higher in more socioeconomically deprived communities (Elliot, Reference Elliot2016), contextual social factors may be responsible for some of the effects. Retrospective assessment of trauma is a recurrent limitation (Okeke, Wilkinson, & Roberts, Reference Okeke, Wilkinson and Roberts2017) and we cannot exclude the possibility that recall bias is influencing our results. However, patients with psychosis under-report and minimize, rather than over-report and exaggerate trauma severity (Church, Andreassen, Lorentzen, Melle, & Aas, Reference Church, Andreassen, Lorentzen, Melle and Aas2017). The majority of studies also employ small samples and were heterogeneous in the types of developmental trauma reported. Whilst most studies accounted for medication dosage by regression analysis or lifetime exposure, it remains possible that our findings could be due to effects of long-term antipsychotics (Fusar-Poli et al., Reference Fusar-Poli, Smieskova, Kempton, Ho, Andreasen and Borgwardt2013). As we did not restrict our research question to survivors of developmental trauma with a diagnosis of schizophrenia, it is possible that our inclusion of studies of patients with other clinical presentations limits the inferences that can be made from our study. On the other hand, our study did not include patients with schizotypy, which is associated with similar brain alterations to schizophrenia (Kirschner et al., Reference Kirschner, Hodzic-Santor, Antoniades, Nenadic, Kircher, Krug and Modinos2022), and there is some evidence for an impact of childhood trauma on gray matter alterations in schizotypy (Quidé et al., Reference Quidé, Watkeys, Tonini, Grotegerd, Dannlowski, Nenadić and Green2024). Finally, as most studies did not report the presence of PTSD symptoms, we cannot exclude the possibility that comorbid PTSD accounts for some of our findings. Future work is, therefore, urgently needed to address these considerations.
Conclusion
Patients with psychosis who have survived developmental trauma may exhibit alterations in brain structure and function compared to those without histories of trauma. There is some overlap with findings in posttraumatic stress disorder which may be pertinent to understanding the neurocognitive basis of psychotic symptoms following developmental trauma. Further research is urgently needed to precisely elucidate neurocognitive mechanisms giving rise to psychosis following developmental trauma. In parallel, we must also elucidate mechanisms of resilience. Understanding these processes may facilitate the development of more effective treatments for trauma survivors to prevent fully established psychosis and aid those experiencing psychosis in achieving remission and recovery.
Acknowledgements
Dr. Bloomfield was funded by a UCL Excellence Fellowship supported by the National Institute for Health Research University College London Hospitals Biomedical Research Centre. This work is part of the Investigating Mechanisms underlying Psychosis Associated with Childhood Trauma (IMPACT) study funded by a UKRI Future Leaders Fellowship to Dr Bloomfield.
Funding statement
This research was funded by a UKRI Future Leaders Fellowship to Dr Michael Bloomfield.
Competing interests
The authors declare none.
Open access
