A Multimodal MMR Study In PTSD After Traffic Accident

Background and ObjectivePost-traumatic stress disorder (PTSD) is characterized by unique symptoms resultingfrom exposure to a traumatic event, such as combat, violent crime, motor vehicle accidentor childhood abuse. Post-traumatic stress disorder was formally established as anindependent psychological disease in1980and gained much attention in current psychiatricnosology. The symptoms of PTSD include persistent re-experiencing of the event,avoidance of stimuli associated with the event, numbing of general responsiveness, andincreased arousal. Motor vehicle accidents are the leading cause for PTSD in the generalpopulation, and it has been found that of those who were in an accident where medicalattention was needed, almost50%developed PTSD. Some studies found thepost-traumatic stress disorder after motor vehicle accidents can lead to high levels ofdepressive symptoms and poor mental health.The main clinical symptoms of post-traumatic stress disorder are related to thecognitive impairment and mood disorders. The clinical diagnostic criteria for posttraumaticstress disorder is based on the identification of some Symptom Categories and scores ofthe PTSD Scale. The Clinician-Administered PTSD Scale (CAPS), post-traumatic stressdisorder self-rating scale (PTSD-SS) and Essen Trauma Inventor (ETI) are the structuredinterview for assessing posttraumatic stress disorder (PTSD) diagnostic status and symptomseverity. However, neuroimaging studies have suggested that structural and functionalchanges exist in PTSD patient’s brain, mostly in cerebral areas relative to emotion andmemory including prefrontal lobe, limbic system and paralimbic system. In recent years,findings from functional and structural neuroimaging studies have allowed for tremendousadvances in understanding of the neural mechanisms underlying PTSD.The core neuropsychological processes underlying PTSD have yet to be elucidated,and the association between the anatomical and functional deficits in PTSD after motorvehicle accidents is largely unknown. The aim of the study was to investigate the changes in cortical thickness, ALFF,FA and brain structural connectivity network in PTSD patientsby the multimodal MRI, and to explore the association among the different factors inPTSD resulting from motor vehicle accidents. The relationship of these factors with thePTSD symptom severity is another key point we intend to elucidate.MaterialsTwenty PTSD patients (age,18–40years; mean,32.92years) who had been involvedin motor vehicle accident were recruited. Diagnosis of PTSD was established with theClinician-Administered PTSD Scale for DSM-IV (CAPS-DX). Twenty healthy controls(age,20–38years; mean,31.53years) individually matched for age, gender and years ofeducation were consecutively recruited. Inclusion criteria for all subjects wereright-handedness, and an IQ>80, as assessed with the Wechsler Intelligence Scale. Patientshad no history of other Axis I psychiatric diagnoses other than depression on the StructuredClinical Interview for DSM-IV Axis I Disorders, whereas controls were free from Axis Idiagnoses on the SCID. Exclusion criteria for both groups were contraindications for MRIand other neuropsychiatric disorders, such as schizophrenia, mental retardation, epilepsy,and head injury with loss of consciousness for more than5min.Methods1,The cortical thickness of the brain was analyzed by Freesurfer, a new method withwhich automated surface reconstruction yields measurements of the cortical thickness foreach subject’s entire brain and computed cross-subject statistics based on the corticalanatomy. Brain surfaces were reconstructed and inflated, and cortical thicknesses wereestimated. Finally, all of the cortical thicknesses were generated and compared byperforming t-tests between PTSD patients and NC subjects. A statistical threshold ofP<0.05(FDR corrected) was used for an exploratory whole brain analysis. Individualcortical thickness values from these significantly different areas were calculated andexported to SPSS12.0for regression analysis.2,The resting-state fMRI data were preprocessed by SPM5.0. REST1.5software wasused to calculate the ALFF. For standardization purposes, the ALFF of each voxel wasdivided by the global mean ALFF value. ALFF maps of the PTSD and NC groups werecompared on a voxel-wise basis by a two-sample t-test in SPM5. A threshold of P<0.05(FDR corrected) was considered statistically significant. Individual ALFF values from these significantly different areas were calculated and exported to SPSS12.0for regressionanalysis.3, DTI data were pre-processed and analyzed by Functional Software Library (FSL,Version) and SPM8. The brain areas with abnormal FA and MD in the PTSD patients wereanalyzed by matrix. The average FA and MD were extracted, and the relationships of FAand MD with PCL_C score were analyzed. To determine the nodes of SCN (structureconnective network) in each subject, regions of interest were defined in native diffusionspace. The derived transformation parameters were inverted and used to warp theautomated anatomical labeling (AAL) regions of interest from MNI space to the nativediffusion space with nearest-neighbor interpolation. The small-world properties of weightedstructure brain network were investigated.Results1, The visualization of difference maps of the cortical thickness clearly showed thefocal cortical thinning areas. In the left hemisphere, a significant decrease in corticalthickness was observed in the medial prefrontal cortex (BA10), the riangular part of theinferior frontal gyrus (BA45) and anterior cingulate cortex (BA32) of PTSD patients(P<0.05, FDR-corrected). In the right hemisphere, a significant decrease in corticalthickness was found in the superior temporal gyrus of PTSD patients compared with thecontrol group (P<0.05, FDR-corrected).2, By comparing the ALFF values of the PTSD with NC subjects, we found that thePTSD patients exhibited increased ALFF values mainly in the left meidal prefrontal cortex(BA10), the anterior cingulated (BA32) and the right cerebellum (P<0.05, FDR-corrected)Linear regression revealed a strong linear negative correlation between CAPS scoresand cortical thickness of BA10, and a weaker linear correlation between CAPS scores andcortical thickness of BA32, indicating that thinner cortical thickness is associated withgreater PTSD scores. There was no significant association between CAPS scores and thecortical thickness of BA45and superior temporal gyrus. Linear regression of CAPS scoreswith ALFF showed that there was only a linear positive correlation between CAPS scoresand ALFF of BA10, indicating that higher ALFF of BA10is associated with greater PTSDscores. The overlapped areas were BA10and BA32in cortical thickness and ALFF. Linearregression revealed that a linear negative correlation existed between cortical thickness and the ALFF of BA10, but no significant correlation was found between ALFF and thecortical thickness of BA32.3, In the PTSD patients, the FA values of bilateral middle frontal cortex(MFC), rightsuperior prefrontal cortex (SPFC), left putamen was significantly lower (p<0.005,voxel>15). The MD values of bilateral MFC, left amygdala, anterior cingulate cortex(ACC),left insula and left globus pallidus(GP) were significantly higher (p<0.005, voxel>15).Pearson correlation analysis suggested that FA and MD values of abnormal cerebral areahad no significant correlation with PTSD, but FA value of right MFC had significantlynegative correlation trend and MD value of right MFC and left amygdala had significantlypositive correlation trend. Both PTSD group and control group showed higher localefficiency but almost identical global efficiency relative to random networks, which can bediagnosed as the feature of small world topology. Nonparametric Wilcoxon rank sum testrevealed that the differences of specific network metrics between PTSD group and controlgroup were significant. The PTSD group showed significant lower values in characteristicpath lengthLw(p=0.007) and normalized characteristic path lengthλ(p=0.038)compared with normal control group. No significant differences (p>0.05) were found inlocal clustering ofCw,γandσ. As to network efficiency, the comparisons revealedslightly increase in global efficiency (p=0.104) but unchanged local efficiency (p>>0.05) ascompared with normal control subjects.ConclusionsIn this study in patients with PTSD after motor vehicle accidents by multimodal MRIneuroimaging, there are abnormal structure and function of medial prefrontal cortex(mPFC), anterior cingulate cortex (ACC), temporal lobe, parietal lobe parahippocampalgyrus and amygdala. It can be concluded that the post-traumatic stress disorder after motorvehicle accidents is a mental illness caused by change of structure and function in manybrain regions. In the multimodal study, the left meidal prefrontal cortex and anteriorcingulated were the overlapped areas in cortical thickness,ALFF and FA. They are seen askey brain regions to the post-traumatic stress disorder.