2.1 Participants
A total of 68 female subjects participant in this study, including 34 patients with BN and 34 age-matched healthy controls (HC) in the normal weight range with no history of eating disorders or other psychiatric disorders. The patients group recruited from Mental Health Center of West China hospital of Sichuan University. The healthy volunteers were recruited through public advertisement. The structured clinical interview for DSM-V disorders (SCID) was administered by two trained psychiatrists. All subjects were female,non-medicated and right-handed with no lifetime history of major psychiatric disorders, head trauma, substance abuse or dependence and neurologic disease, without metal implants or heart pacemakers. We did not exclude the patients who have mild anxiety and depressive symptoms, as these symptoms were commonly seen in BN individuals[9, 10].
2.2 Clinical psychological assessments
All patients completed the revised version of Core Self-Evaluation Scale(CSES) and the Rosenberg Self-Esteem Scale(RSES) in Chinese. The CSES was revised by Du Jianzheng et.al, with acceptable reliability and validity, which included 10 items and used a 5-point score method. The total score ranged from 10-50 points, which represent the core self-evaluation level of each subject. This scale had good reliability and validity in Chinese population, and the norm of CSES was 36.05±5.21 for female college students. In this study, we compared the CSES score of BN with the national norm.
The RSES included 10 items, each item was rated on 4-point scale, and item3,5,8,9,10 are negtively worded. The total score was the sum of the scores of each item, indicating the levels of individual self-esteem[12]. The institute of education of Tsinghua University investigated the self-esteem of college students in Beijing by using the RSES scale in Chinese.The reasearch found the average RSES score of college students is 28.73±4.48, no significant gender difference was found[12]. It has good representation in Chinese population. In our study, we use this value to compare with the average RSES score of the patients group.
This study was approved by Ethics Committee of the West China hospital of Sichuan University, and all subjects provided written informed consent.
2.3 MRI data acquisition
MR images were obtained using a 3.0T philips MR system. All subjects underwent a 3D T1-weighted volumetric scan with the following parameters: sigittal; matrix size,256×256; field of view(FOV)=256mm×256mm;slice thickness=1mm;no gap; flip angel=7°; repetition time(TR)=8.2ms; echo time(TE)=3.8ms; voxel size=1×1×1mm. All subjects were asked to be relax, keep their eyes closed without fall sleep, not think of anything and keep head motionless as possible.
2.4 Voxel-based morphometry processing
Data analysis was carried out using the voxel-based morphometry(VBM) toolbox by Christian Gaser (VBM8; http://www.neuro.uni-jena.de/vbm/), as part of SPM8 (Wellcome Trust Centre for Neuroimaging, Institute of Neurology, London, UK), run under Matlab R2013b. Structural images were normalized to Montreal Neurological Institute(MNI) standered space and segmented into grey matter(GM), white matter(WM) and cerebrospinal fluid(CSF) using the unified segmentation approach. Finally, the grey matter images were smoothed using a 8mm full-width half-maximum Gaussian kernel for analysis.
2.5 Statistical analysis
A two-sample t-test were conducted using SPM8 software to compare regional GMV differences between patients and HC,with age and total intracranial volume (TIV, the sum of GM, WM and CSF volumes) as covariates. Multiple comparisons were corrected using family-wise error(FWE) approach at a cluster level, with a corrected threshold of P<0.001. An absolute threshold was set at 0.2 in all computations.
The Statistical Package for Social Science (SPSS) version20.0 (SPSSInc, Chicago, https://www.ibm.com/products/spss-statistics) was used to conduct statistical analysis of clinical variables and demographic data, as well as compare total GM, WM, CSF and intracranial volume between patients and control group. Comparison between two groups was conducted by two sample t-test.
Moreover, correlation analyses were conducted using SPSS 20.0. We extracted the brain regions that showed GMV alterations as region of interest(ROI), using the Region of Interest extraction tool in xjView via the SPM toolbox (http://www.alivelearn.net/xjview8/). Pearson correlations were conducted between ROIs and clinical variables after Bonferroni correction.