Introduction
Post-traumatic stress disorder (PTSD) is a relatively common and predictable psychological syndrome Reference Miller(1). PTSD occurs in a proportion of individuals exposed to severe psychological trauma Reference Kasai, Yamasue, Araki, Sakamoto and Kato(2) and in which the individual responds with fear, helplessness or horror Reference Danckwerts and Leathem(3). Individuals with PTSD suffer from intrusive memories about the traumatic event, persistent avoidance of stimuli associated with the trauma, and persistent symptoms of increased arousal, symptoms which become uncontrollable and disabling Reference Bremner and Charney(4). Because of its debilitating nature, PTSD has emerged as an important public health problem in the general population Reference Sareen, Cox, Stein, Afifi, Fleet and Asmundson(5).
In recent years, a great deal of research has been directed towards understanding the etiology, phenomenology, neurobiology, clinical characteristics and treatment of PTSD Reference Nemeroff, Bremner, Foa, Mayberg, North and Stein(6). However, a number of core psychological processes underlying PTSD have yet to be elucidated Reference Shin, Rauch and Pitman(7,Reference Liberzon and Sripada8). Over the past decade, findings from neuroimaging studies have allowed for tremendous advances in our understanding of the experience of emotions in healthy individuals and the dysregulation of these processes associated with PTSD. These studies have been useful in both generating hypotheses on the neurobiology of normative human responses to trauma and complementing our understanding of the wide-ranging alterations in trauma survivors who develop PTSD.
Structural neuroimaging studies have focused primarily on hippocampal volumetry Reference Geuze, Vermetten and Bremner(9) as well as the prefrontal cortex Reference Geuze, Vermetten, Goebel and Westenberg(10) and other brain structures. Hippocampal morphology has been correlated with severity of PTSD symptomatology Reference Gilbertson, Shenton and Ciszewski(11,Reference Villarreal and King12). However, the results have been inconsistent, with studies reporting significant reductions or increases, as well as unchanged volumes. For example, studies have shown that patients with PTSD are associated with bilateral lower hippocampal volume (Reference Bossini and Castrogiovanni13–Reference Li, Chen, Li, Zhang, He and Lin18), which are considered to be either because of atrophy of the hippocampus as a consequence of suffering from PTSD because of excessive stress Reference Bremner, Randall and Scott(19,Reference Gurvits, Shenton and Hokama20) or that hippocampal volume to be a risk factor for developing PTSD Reference Gilbertson, Shenton and Ciszewski(11). Other studies report unchanged hippocampal volumes in female patients with chronic PTSD traumatised by intimate partner violence Reference Fennema-Notestine, Stein, Kennedy, Archibald and Jernigan(21), those traumatised at the same air show plane crash Reference Jatzko, Rothenhofer and Schmitt(22), elderly PTSD patients Reference Golier, Harvey, Legge and Yehuda(23) and adult burn patients Reference Winter and Irle(24). One study noted opposite trends in abused juveniles Reference Tupler and De Bellis(25). However, a recent meta-analysis confirmed the presence of significantly smaller hippocampal and left amygdala volumes in patients with PTSD compared to controls with and without trauma exposure Reference Karl, Schaefer, Malta, Dorfel, Rohleder and Werner(26). The findings of previous studies suggest that abnormal hippocampal volume was not a necessary and sufficient condition of PTSD.
Several studies have shown that the medial prefrontal cortex, which includes the anterior cingulate cortex and medial frontal cortex, are involved in the process of extinction of fear conditioning and the retention of extinction Reference Milad, Rauch, Pitman and Quirk(27). Research on abnormalities in the prefrontal cortex in PTSD patients suggest decreased volume (Reference Fennema-Notestine, Stein, Kennedy, Archibald and Jernigan21,Reference Carrion, Weems and Eliez28–Reference Hakamata, Matsuoka and Inagaki30), while some findings suggested increased volume of the middle-inferior and ventral regions of the prefrontal cortex Reference Richert, Carrion, Karchemskiy and Reiss(29).
In addition to findings related to the hippocampus and medial prefrontal cortex, many current functional neuroimaging studies have identified other brain areas such as the prefrontal cortex, temporal lobe, parietal lobe, limbic lobe and others that may also be implicated in PTSD (Reference Bonne, Gilboa and Louzoun31–Reference Geuze, Westenberg, Heinecke, De Kloet, Goebel and Vermetten36). The findings of functional neuroimaging studies suggest that there are more brain areas that may be affected in PTSD. However, only a few studies have found corresponding structural abnormalities in these brain areas.
In contrast to the considerable research on subcortical structure volumetry, few studies to date have been directed to grey matter reductions in the cortex. It is evident that structural neuroimaging studies will allow for the testing of hypotheses of an association between PTSD and abnormal grey matter. Although volumetry findings reveal changes in the volume of specific brain regions, most of these studies defined particular regions-of-interests (ROIs) and measured their size and hemispheric asymmetry using traditional morphometric techniques with high-resolution magnetic resonance images (MRI). The disadvantage of this method is that some important brain areas may be neglected, and the process of drawing ROIs may introduce additional error. Furthermore, the measurement of volume may not accurately reflect changes in the internal structure of the brain. In recent years, a fully automated voxel-based morphometry (VBM) technique has allowed for the examination of cerebral asymmetries across the entire brain directly (Reference Corbo, Clement, Armony, Pruessner and Brunet37–Reference Luders, Gaser, Jancke and Schlaug39), which can compensate for the subjectivity of ROI approaches. The VBM technique has been used for assessing regional grey-matter density (GMD) in PTSD patients and revealed abnormal GMD in the hippocampus, anterior cingulate cortex and insula (Reference Emdad, Bonekamp and Sondergaard15,Reference Jatzko, Rothenhofer and Schmitt22,Reference Richert, Carrion, Karchemskiy and Reiss29,Reference Corbo, Clement, Armony, Pruessner and Brunet37,Reference Kasai, Yamasue, Gilbertson, Shenton, Rauch and Pitman38).
Most previous PTSD studies in the West focused on the disorder caused by various traumatic events, such as war (Reference Vythilingam, Luckenbaugh and Lam17,Reference Bremner, Randall and Scott19,Reference Pavic, Gregurek and Rados40), disaster (Reference Jatzko, Rothenhofer and Schmitt22,Reference Abe, Yamasue and Kasai41–Reference Yamasue, Kasai and Iwanami43) and sexual abuse Reference Bremner, Vythilingam and Vermetten(14,Reference Tupler and De Bellis25). Although considerable research has focused on rape-related PTSD, limited studies have been carried out in the context of Mainland China. In this study, rape was defined as an event that occurred without the victim's consent, that involved the use or threat of force to penetrate the victim's vagina or anus by penis, tongue, fingers or object, or the victim's mouth by penis Reference Tjaden and Thoennes(44). Interestingly, evidence indicates that the incidence rate of PTSD induced by rape is the highest among all kinds of trauma Reference Rothbaum, Foa, Riggs, Murdock and Walsh(45).
In the current study, we used VBM to explore differences in GMD between VoR with and without PTSD, as well as in healthy comparison (HC) subjects. Based on findings from previous neuroimaging studies (Reference Milad, Rauch, Pitman and Quirk27–Reference Geuze, Westenberg, Heinecke, De Kloet, Goebel and Vermetten36), we hypothesised that VoR with PTSD would show structural changes in extensive brain areas, including the prefrontal, temporal, parietal and limbic regions, compared to VoR without PTSD and to HC.
Material and methods
Subjects
We conducted a cross-sectional study on VoR and HC in Guangdong Province, People's Republic of China. Subjects living and working in Guangdong who met the following criteria were included: females, at least 18 years old, right-handed, with an educational attainment above secondary school level. Exclusion criteria for VoR and HC were a history of neurological or brain trauma and alcohol or drug use/abuse. Additional exclusion criteria for VoR included a previous or current psychiatric diagnosis other than PTSD, and for HC included any previous or current Diagnostic and Statistical Manual of Mental Disorders-Fourth Edition (DSM-IV) psychiatric disorder. The Ethics Committee of the Second Xiangya Hospital, Central South University, the People's Republic of China approved the study protocol. Written informed consent was obtained from all subjects in the study. The sample size for this study was projected based on previous research (Reference Lindauer, Vlieger and Jalink16,Reference Fennema-Notestine, Stein, Kennedy, Archibald and Jernigan21,Reference Jatzko, Rothenhofer and Schmitt22,Reference Abe, Yamasue and Kasai41) where sample sizes ranging from 8 to 20 were reported.
VoR subjects
VoR subjects were recruited in two stages. In the first stage, 53 potential subjects were recruited from (a) four public psychological consulting clinics, (b) referred to the clinics by a non-government organisation specialising in assisting victims of sexual assault and (c) through advertisements in local newspapers requesting VoR for brain imaging studies. As an incentive for participation, 6 months free psychological counseling and medical therapy were offered. In the second stage, psychiatrists in the four consulting clinics explained the study to the 53 potential subjects and requested consent to participate in this study. As a result, 23 VoR met the inclusion/exclusion criteria and gave written informed consent.
HC subjects
The HC subjects for this study were also recruited through a two-stage process. First, after obtaining the major demographic characteristics (i.e. gender, race, age, height, weight and educational years) of VoR, we published notices across the province to recruit female health workers with similar demographic characteristics. A total of 65 volunteers agreed to participate and were screened according to inclusion/exclusion criteria. The health status of the HC was determined based on health check reports, as well as an illness history interview conducted by a doctor and a psychiatrist. In the second stage, the exact number of HC was recruited to best match the major demographic characteristics of each case of VoR with PTSD.
In this study, all subjects were measured with the Trauma History Questionnaire Reference Green and Stamm(46) and the PTSD Checklist Civilian Version (PCL-C) Reference Ruggiero, Ben, Scotti and Rabalais(47). In addition, two independent, clinically experienced psychiatrists interviewed VoR subjects using the Clinician-Administered PTSD Scale (CAPS) Reference Blake, Weathers and Nagy(48). The PCL-C was used to predict PTSD diagnoses, and the CAPS was used to differentiate PTSD and non-PTSD VoR subgroups. A senior psychiatrist confirmed the final diagnosis of PTSD.
MRI data acquisition
Images were obtained from using a research-dedicated Siemens Avanto 1.5 T MRI scanner. The T1-weighted anatomical images were acquired using a three-dimensional gradient-echo sequence, with time of repletion (TR) = 11 ms, echo time(TE) = 4.94 ms, number of averages = 1, matrix = 256 × 224 pixels, field of view = 256 × 224 mm, with a flip angle of 15°. One-hundred and seventy-six sagittal slices with a 1-mm slice thickness were acquired with no interslice gap. There was a voxel resolution of 1 × 1 × 1 mm3. The total acquisition time was 5 min and 34 s.
MRI data analysis
VBM were implemented by using the Statistical Parametric Mapping software (SPM2) (Wellcome Department of Imaging Neuroscience, London, England; www.fil.ion.ucl.ac.uk) Reference Friston, Holmes, Worsley, Poline, Frith and Frackowiak(49). Images were firstly spatially normalised to the Montreal Neurological Institute (MNI) space with the standard T1-MRI template Reference Mazziotta, Toga, Evans, Fox and Lancaster(50) implemented in the SPM2 program, and re-sliced into a final voxel size of 1 × 1 × 1 mm3 using tri-linear interpolation. The spatially normalised images were then segmented into three compartments: grey matter, white matter and cerebrospinal fluid, respectively. Finally, the segmented grey matter images from VoR with PTSD, VoR without PTSD and HC were smoothed with a 12-mm full-width at half-maximum isotropic Gaussian kernel Reference Ashburner and Friston(51). Between-group comparisons of grey matter images were performed in the general linear model.
Because we are particularly interested in exploring increases/decreases in GMD in VoR with PTSD compared to VoR without PTSD and HC, two-sample t-tests were performed in the VBM analysis in a voxel-by-voxel manner. Consistent with previous studies Reference Liberzon, King, Britton, Phan, Abelson and Taylor(52,Reference Hou, Liu and Wang53), the significance threshold was set to p < 0.005 corrected for multiple comparisons with a minimal cluster size of > 50 voxels. The significant regions were superimposed onto SPM2's standard T1-weighted brain images.
Based on previous research (Reference Milad, Rauch, Pitman and Quirk27–Reference Geuze, Westenberg, Heinecke, De Kloet, Goebel and Vermetten36), we hypothesised that compared with HC, VoR with PTSD would show grey matter abnormalities in the prefrontal, temporal, parietal and limbic regions. We used the small volume correction tool in the SPM2 package with the specific purpose of restricting comparisons to specific voxels located in these regions. This approach permits the implementation of hypothesis-driven analyses with corrections for the pre-specified ROIs rather than corrections for the whole brain.
Results
Following the initial interview, among the 23 VoR subjects, 13 met the DSM-IV diagnostic criteria for current PTSD and 10 VoR did not meet the criteria for PTSD. Based on the study protocol, 13 HC were recruited to match VoR with PTSD. All subjects were scanned with MRI. However, because of too many head movements during MRI scanning, a total of five subjects (2 PTSD, 2 non-PTSD and 1 HC) were removed. As a result, the final sample consisted of 11 VoR with PTSD (18–31 years), 8 VoR without PTSD (23–33 years) and 12 HC (22–33 years).
The three groups did not differ significantly on major demographics (i.e. age, height, weight and educational years). In addition, the average interval between rape trauma and data acquisition did not differ significantly between VoR with and without PTSD. However, VoR with PTSD scored significantly higher on PTSD symptomatology (p < 0.001) compared to VoR without PTSD and HC. None of the participants in this study received medication prior to neuroimaging acquisition. The results are summarised in Table 1.
Demographic and clinical characteristics of VoR with and without PTSD and healthy comparison
CM, centimeter; KG, kilogram; Interval, time between raped trauma occurrence and scan; SD, standard deviation.
Differences in GMD between VoR with PTSD and HC
The areas found to have abnormal GMD in VoR with PTSD compared to HC are shown in Fig. 1. The cortical areas with decreased GMD in VoR with PTSD compared to HC are listed in Part a of Fig. 1. These areas include the left medial frontal cortex (Fig. 1a-1), right medial frontal cortex (Fig. 1a-2 ), the left middle frontal cortex (Fig. 1a-3), the left middle temporal gyrus (Fig. 1a-4) and the left fusiform cortex (Fig. 1a-5). The areas with increased GMD are listed in Part b of Fig. 1, and include the right posterior cingulate cortex (Fig. 1b-1), the left precentral cortex (Fig. 1b-2), right precentral cortex (Fig. 1b-3), the left inferior parietal lobule (Fig. 1b-4), right inferior parietal lobule (Fig. 1b-5) and the right postcentral cortex (Fig. 1b-6). The MNI coordinates, voxel t values, k values (cluster size > 50), and corresponding Brodmann Areas (BA) are detailed in Table 2. Regions displayed are for p < 0.005.
Part (a) shows the cortical areas with decreased GMD in VoR with post-traumatic stress disorder compared to HC. Part (b) shows the areas with increased GMD. Part (c) shows the cortical areas with decreased GMD in VoR with PTSD compared to VoR without PTSD. Part (d) shows the areas with increased GMD. IPL, inferior parietal lobule; L, left; MFC, middle frontal cortex; mFC, medial frontal cortex; MTG, middle temporal gyrus; PCC, posterior cingulate cortex; R, right.
GMD in victims of rape with PTSD vs. healthy comparison
k, cluster size.
Regions displayed are for p < 0.005, k > 50.
Differences in GMD between VoR with and without PTSD
The areas found to have abnormal GMD in VoR with PTSD compared to the VoR without PTSD are shown in Fig. 1. The cortical areas with decreased GMD are listed in Part c of Fig. 1. These areas include the right uncus (Fig. 1c-1), the left middle temporal gyrus (Fig. 1c-2), and the left fusiform cortex (Fig. 1c-3). The areas with increased GMD are listed in Part d of Fig. 1, and include the left precentral cortex (Fig. 1d-1), the left inferior parietal lobule (Fig. 1d-2), and the right postcentral cortex (Fig. 1d-3). The MNI coordinates, voxel t values, k values (cluster size > 50), and corresponding BA are detailed in Table 3. Regions displayed are for p < 0.005.
GMD in VoR with PTSD vs. VoR without PTSD
k, cluster size.
Regions displayed are for p < 0.005, k > 50.
Discussion
This cross-sectional study used VBM to examine GMD abnormalities among VoR with PTSD compared to VoR without PTSD and HC in mainland China. The findings of this study support the hypothesis that changes in GMD are associated with the pathophysiology of rape-induced PTSD.
The frontal cortex
The structural abnormalities in the medial and middle frontal cortex found in VoR with PTSD are supported by previous studies (Reference Shin, Rauch and Pitman7,Reference Fennema-Notestine, Stein, Kennedy, Archibald and Jernigan21,Reference Richert, Carrion, Karchemskiy and Reiss29,Reference Hakamata, Matsuoka and Inagaki30). In addition to such structural findings, functional neuroimaging studies have also shown dysfunction in these cortical areas (Reference Bremner, Vythilingam and Vermetten14,Reference Bremner, Staib, Kaloupek, Southwick, Soufer and Charney54–Reference Morey, Petty, Cooper, Labar and McCarthy59). It is therefore possible that altered GMD in the medial and middle frontal cortex may contribute to their hypofunction. However, in the current study, VoR with PTSD had GMD reductions in the left middle frontal cortex and the bilateral medial frontal cortex relative to the HC, but no significant differences relative to VoR without PTSD. This result suggests that severe psychological trauma may change brain grey matter of the medial and middle frontal cortex, but that such plastic changes to these cortical brain structures may not underlie the pathophysiology of PTSD.
Compared to VoR without PTSD and HC, VoR with PTSD had significant bilateral increases in GMD of the precentral gyrus, which mainly consists of the premotor cortex and the supplementary motor area. Studies indicate the importance of the supplementary motor area in motor tasks that demand retrieval of motor memory Reference Tanji(60). The premotor cortex, located in the precentral cortex (BA6), seems to play a major role in language Reference Duffau, Capelle and Denvil(61). Functional neuroimaging studies on PTSD suggest abnormal functional activities in this cortical area. Empirical studies have shown that PTSD groups are characterised by relatively more activation in the precentral cortex than non-PTSD and HCs (Reference Bonne, Gilboa and Louzoun31,Reference Bremner, Staib, Kaloupek, Southwick, Soufer and Charney54,Reference Lanius, Bluhm, Lanius and Pain55,Reference Shaw, Strother and McFarlane62,Reference Jatzko, Schmitt, Demirakca, Weimer and Braus63). These findings suggest that increases in GMD of the precentral cortex may be involved in the neural basis of motor and linguistic PTSD symptomatology.
The parietal lobule
The parietal lobule, including the superior and inferior parietal lobule (BA7, BA40) is implicated in memory, recognition, and deductive reasoning Reference Xie, Xiao, Bai and Jiang(64,Reference Knauff, Mulack, Kassubek, Salih and Greenlee65). In this study, compared to VoR without PTSD and HC, VoR with PTSD had significant increases in GMD of the inferior parietal lobule. This indicates that the inferior parietal lobule might play an important role in the pathophysiology of PTSD. In addition, one positron emission tomography (PET) study showed that chronic PTSD patients presented relatively diminished activity in the postcentral regions Reference Molina, Isoardi, Prado and Bentolila(33), which mainly consist of the primary somatic sensory cortex and secondary somatic sensory cortex Reference Geng, Eger, Ruff, Kristjansson, Rotshtein and Driver(66). Increases in GMD in the postcentral cortex may be associated with the dysfunction associated with PTSD.
The temporal gyrus
Our study showed significant GMD decreases in the left middle temporal gyrus in VoR with PTSD compared to HC. Studies comparing veterans with and without PTSD revealed that those with PTSD had overactivation of the temporal gyrus during the resting state Reference Molina, Isoardi, Prado and Bentolila(33) and the encoding phase Reference Geuze, Vermetten, Ruf, De Kloet and Westenberg(67) and underactivation of the bilateral middle temporal gyrus in the retrieval phase Reference Geuze, Vermetten, Ruf, De Kloet and Westenberg(67). Functional neuroimaging studies revealed significant activation in left middle temporal gyrus in response to empathy judgments in post-therapy PTSD Reference Farrow, Hunter and Wilkinson(68) and higher levels of activation in the middle temporal gyrus in dissociative PTSD Reference Lanius, Bluhm, Lanius and Pain(55). Empirical evidence suggests that the fusiform cortex is specialised for face processing Reference Rossion, Dricot and Devolder(69,Reference Rhodes, Byatt, Michie and Puce70). Research also indicates relatively augmented activity in the fusiform cortex in patients with PTSD Reference Bonne, Gilboa and Louzoun(31,Reference Molina, Isoardi, Prado and Bentolila33). The reduced GMD of the middle temporal gyrus and fusiform cortex found in this study implicates these regions in the dysfunction of memory and dissociative symptoms in PTSD.
The limbic lobe
In the present study, VoR with PTSD had significant increases in GMD in the right posterior cingulate cortex compared to HC. Meta-analyses have revealed that the prominent themes in the posterior cingulate cortex are episodic memory retrieval and pain Reference Nielsen, Balslev and Hansen(71), visuospatial processing and assessment of threat Reference Nemeroff, Bremner, Foa, Mayberg, North and Stein(6), as well as fear conditioning Reference Doronbekov, Tokunaga and Ikejiri(72). Comparison of connectivity maps by functional connectivity analyses Reference Lanius, Williamson and Densmore(56) revealed that subjects with PTSD showed greater correlations in interregional brain activity than subjects without PTSD in the right posterior cingulate cortex (BA 29). Functional neuroimaging studies have found increased activation in the posterior cingulate cortex in victims with PTSD compared to victims without PTSD and to healthy controls (Reference Bremner, Vythilingam and Vermetten14,Reference Bremner, Staib, Kaloupek, Southwick, Soufer and Charney54,Reference Doronbekov, Tokunaga and Ikejiri72–Reference Bremner, Narayan, Staib, Southwick, McGlashan and Charney74). This indicates that dysfunction in the posterior cingulate cortex may underlie pathological symptoms provoked by traumatic reminders of sexual assault among VoR with PTSD.
It should be noted that whilst the sample size (11 VoR with PTSD) of this study meets the threshold of research reported in the literature Reference Jatzko, Rothenhofer and Schmitt(22,Reference Hou, Liu and Wang53), it was nevertheless limited in terms of its statistical power. A further limitation is the potential for selection bias in both the VoR Group and HC. Given the social stigma attached to VoR in the Chinese cultural context that often results in sexual shame, fear and anxiety over disclosure of the rape, guilt over derogating family honor, self-scrutiny and self-blame after the fact, and even blame of the victim, contribute to the potential selection bias in the VoR Group. Regarding the selection bias in HC, it was not our intention for this group (12 HC) to be representative of the general population, nor were the 65-pooled controls. This group was particularly designed to match the VoR with PTSD regarding the major demographic characteristics. However, given the fact that HC were recruited from healthy workers, selection bias from the healthy-worker effect should be borne in mind.
The use of a cross-sectional design is another significant limitation of this study. The shared variance estimated between variables in a cross-sectional design does not allow for a critical examination of causal relationships among them. Consequently, we cannot state categorically whether changes in GMD were the cause or the effect of trauma exposure/PTSD. Future studies using case-control or longitudinal designs need to be conducted to explore experiences of rape and associated PTSD symptomatology. In addition, the impact of the specific social and cultural meanings of rape and the impact on the individual's post-traumatic response and ability to cope should also be investigated.
Conclusion
The abnormal GMD of cerebral regions in VoR with PTSD supports the hypothesis that PTSD is associated with structural plastic changes to brain grey matter. The results suggest that the medial frontal cortex, precentral cortex, posterior cingulate cortex, postcentral cortex and inferior parietal lobule are likely to contribute to the neural mechanisms underlying PTSD.
Acknowledgments
We acknowledgement support from grant from the National Natural Science Foundation of China (30830046 to Lingjiang Li and Shuangge Sui), the National Science and Technology Program of China (2007 BAI17B02 to Lingjiang Li and Shuangge Sui), the National 973 Program of China (2009CB918303, 2006CB5000800 to Lingjiang Li); Program of Chinese Ministry of Education (20090162110011 to Lingjiang Li).
