Participants
This study was approved by the ethic committee of the Second Hospital of Hebei Medical University. A complete description of the study was provided to every subject, after which written informed consent was obtained from each participant.
Inclusion criteria included: Age 18–30 years, regardless of gender, and all women don’t take oral contraceptives; Junior high school education or above; Han Chinese, Right-handed; Did not take any medication within the imaging hours; The subjects fully understand the content of the study, wishing to participate and complete the study, and sign the informed consent.
Exclusion criteria: People with a history of mental illness; There are mental disorders in first-degree relatives; With a history of alcohol and drug dependence; With cerebral organic and serious physical diseases; Previous seizures or a family history of epilepsy; Patients who are receiving hormone therapy; Patients with current infection, trauma, and associated immune or other medical diseases; All contraindications exist, or magnetic resonance examination reveals abnormalities in brain structure.
Measures Of Depression, Anxiety And Personality Traits
Depressive symptoms of the participants were measured by the Beck Depression Inventory (BDI) and anxiety level was determined by the Spielberger State-Trait Anxiety Inventory (STAI): STAI-Y1 for state anxiety and STAI-Y2 for trait anxiety, respectively. Personality traits was tested by Three-Dimensional Personality Questionnaire (TPQ). Participants were administered the TPQ, BDI and STAI before the MRI scanning to assess personality traits and the presence of depressive or anxiety symptoms, respectively.
Measures Of Abuse And Neglect
Childhood Trauma Questionnaire (CTQ) were adopted to assess all participants' childhood traumatic experiences and this scale has been shown to have acceptable psychometric properties (Cronbach α = 0.90) (29). The questionnaire consists of five following subscales: childhood physical abuse, emotional abuse, sexual abuse, physical neglect, and emotional neglect. In this study, the CTQ-Short Form (CTQ-SF) that consist of 28 items were used. Cutoff scores are 13 or higher for emotional abuse, physical abuse ≥ 10, sexual abuse ≥ 8, emotional neglect ≥ 15, physical neglect ≥ 10. Participants with a score above this threshold of the subscales are considered the presence of childhood trauma. According to the total CTQ, the participants were divided into childhood trauma group and healthy control group. To determine the different effects of abuse and neglect on functional connectivity of the brain, scores for abuse were calculated using physical, emotional, and sexual abuse, and scores for neglect were calculated using emotional and physical neglect.
Mri Acquisition
3D-T1 structural image: Sagittal high resolution structural image scan was performed in T1-FFE sequence. The scan baseline is parallel to the anterior and posterior joint lines.TR = 7.8 ms, TE = 3.8 ms, turn Angle 8°, matrix = 252×227, voxel size 1×1×1 mm, FOV = 250×250×180 mm, layer thickness 1mm, layer spacing 0mm, scanning layer number 180, NEX = 1.
rs-fMRI: Before scanning, subjects were asked to remain awake, close their eyes, and be relaxed, avoid any body activity, and try not to do any systematic thinking. EPI-GRE sequence was adopted: TR = 2000 ms, TE = 30 ms, FA = 90°, layer thickness 4mm, interval 0mm, 33 scanning layers, FOV = 24 cm×24 cm, matrix = 64×64, 180 time points were collected, scanning time 360 s, and 5940 images were obtained.
Data Pre-processing
Conventional MRI images were diagnosed by two experienced imaging diagnosticians, excluding subjects with obvious lesions and multiple foci of abnormal signals. The dicom format images were converted to NIFTI format, CONN18.b software was used to preprocess the resting state functional images. A reference volume and its skull-stripped version were generated using a custom methodology of CONN-default pipeline for analyses in MNI-space. The fMRI reference was then co-registered to the T1WI reference, and then slice-time correction, head movement correction with removal of non-steady state volumes and spatial smoothing with an isotropic. The functional images were registered and resampled to the MNI standard space by gaussian kernel with 8 mm full-width half-maximum. Finally, the BOLD time series were band-pass filtered (0.008–0.09 Hz).
Independent Component Analysis (Ica)
CONN18.b software based on the stable group ICA algorithm (G1 Fast ICA + GICA 3 Back-projection) were used for ICA. Group Independent Component Analyses (gICA) were performed on two groups of preprocessed resting state data. First of all, dimensionality reduction is performed on fMRI data of subjects through principal component analysis. Two-to-three-step principal component analysis is used to achieve multiple nesting, and 99% data difference is generally guaranteed to determine independent components. Then according to the maximum algorithm (infomax algorithm), the data after dimensionality reduction was decomposed by ICA. Finally, the spatial brain map of each subject's brain network component and its corresponding time courses were reconstructed by group-ICA back-construction method to obtain individual components, which were then converted into 2 values to obtain the time courses of relevant network components and the synchronization degree index of voxel blood-oxygen level dependent signals. It was used as the final statistical analysis of group level.
Graph Theoretical Analysis
Graph theory analysis was performed to compare the network topological properties. In total, DMN, SMN, SAN, DAN, FPN networks were included for analysis. The global efficiency, local efficiency, average path length, clustering coefficient and degree were evaluated using two sample t-tests a threshold of P < 0.05 false discovery rate (FDR) corrected with age, gender, educational level, BDI score, STAI score, TPQ score as covariates.
Statistical Analyses
Unpaired t test was used to compare the unpaired data that conformed to normal distribution and homogeneity of variance. The Wilcoxon rank-sum test was applied for others continues variables. Chi-square (χ2) test was used for categorical variables. Continues variables were expressed as mean and standard deviations. The median with interquartile range was employed for nonparametric continuous variables.
The correlations between the topological properties of RSN with abuse and neglect were analyzed by ordinary least squares regression. Age, sex, years of education, anxiety, depression scale and TPQ were included as covariates for multiple comparisons to avoid any confounding effects. In order to determine whether findings were specific to abuse/neglect, abuse was covaried in neglect models, and vice versa. Collinearity checks can be found in Supplement data.
Mediation models were tested with abuse/neglect scores as the predictor, the topological properties of RSN alterations as the mediator, and psychological trait (anxiety and TPQ) as the outcome variable. Sex, Age, Years of education and BDI were included as covariates. Moderated mediation models were then run to test the role of sex as a moderator. Age, years of education, anxiety, depression scale and TPQ were included as covariates. Mediation analyses were conducted using the PROCESS based on SPSS with model 4, utilizing 5,000 bootstrap samples to estimate the 95% confidence.