Figure 3. Statistical comparison（Mann-Whitney U tests) of gray matter volume change, NAA/Cr, NAA/(Cho+Cr) and SUVR in the regions that contains EZ/PZ/NIZ. * P < 0.05, ** P < 0.01, *** P < 0.001.

Figure 2. SEEG recordings from patient #4 with TLE arising from left hippocampus. (A) Examples of ictal discharges around the time of seizure onset; (B) Location of the SEEG electrodes shown on the 3D brain image.

Figure 2. Representative images showing a 36-year-old man with Left Mesial Temporal Sclerosis (arrow). The signal and morphological abnormalities compatible with hippocampal sclerosis are equally seen in both 3D Wave-CAIPI SPACE FLAIR (A) and in the 3D standard SPACE FLAIR (B).

Figure 3. Representative images showing a 32-year-old man with Tuberous Sclerosis presenting with cortical tubers (arrows) and radial bands (arrowheads). The standard SPACE FLAIR (A) sequence showed greater image degradation due to motion, contrasting with the better performance of the Wave-CAIPI SPACE FLAIR (B) sequence that was preferred over the standard sequence in the head-to-head comparison.

Fig. 1. (A-C) The three candidate neural models. (D) Regions of interest (ROIs) of the three models were obtained from the Talairach Daemon atlas in the Montreal Neurological Institute space. (E) The brown window indicates the anatomical locations of the left hippocampus (HP, green ROI) and mammillary body (MB, purple ROI). Cyan ROI: anterior cingulate gyrus (ACG), blue ROI: anterior thalamic nuclei (ATN), red ROI: parahippocampal gyrus (PHG), yellow ROI: posterior cingulate gyrus (PCG).

Fig. 3. Post-hoc between-group comparison results of the 10 paths in the simplified HDC model using the Mann–Whitney U test with Benjamini–Hochberg correction. Abbreviation: MTS = patients with Mesial Temporal Sclerosis, N.S. = Non-Significant path.

Figure 2. Thalamocortical functional connectivity changes and their clinical significance in mTBI.

(A) Patients with mTBI exhibited significantly reduced thalamo-DMN anticorrelation (black arrows) during WM 1-back and 2-back task conditions compared with the HCs.

Thalamo-DMN anticorrelation strength exhibited (B) a significant positive correlation with participants’ PSQI and PCSQ scores and (C) a significant negative correlation with participants’ arithmetic ability and WMI during the WM 1-back and 2-back task conditions.

Figure 1. Structural evidence of two TCD origins in mTBI revealed by DTI.

(A) Significantly decreased FA (top row) and increased RD (bottom) at the bilateral thalamic borders (yellow arrows) were observed in patients with mTBI compared with HCs (p<0.01, FDR corrected). The ovals covered by translucent dark blue indicate the location of the bilateral thalamus. These maps were masked by the threshold of group-averaged FA > 0.2.

(B) A significant reduction in thalamocortical tract density was found in the mTBI group compared with the HC group (p<0.01, FDR corrected).

Fig. 3. Changing trajectories of seed-based functional connection measures with interaction effect at p<0.01 with left rostral anterior cingulate cortex (A), right rostral anterior cingulate cortex (B) and right rostral middle frontal regions (C) as seeds. * indicates significance after FDR correction. (D). Functional connections with interaction effect at p<0.01 with the trends of inactive fighters demonstrating recovery whereas active fighter showing decline or no change across two time points.

Fig.1. CNS Vital Signs test scores. (A). Changing trajectories along time for inactive fighters (blue) and active fighters (red). (B). Fixed effects (p-values) in linear mixed effect model. * indicates significance after FDR correction.

Fig.2. Examples of OSS and NOSS maps of brain surface from both vertical (AP) and horizontal (LR) vibrations.

Fig.4. Group comparison of cortical NOSS among different brain lobes.

Representative HP metabolites heatmaps from a Sham and a rTBI mouse showing higher [1-¹³C]lactate levels in rTBI in subcortical areas. Higher HP [1-¹³C]pyruvate and HP 13C urea levels can be observed in the entire brain of rTBI compared to Sham. Heatmaps of HP [1-¹³C]lactate/pyruvate ratio show lower values in cortical areas in rTBI.

(A) T2 image and overlaid grid used for HP ¹³C MRSI analyses, red voxels indicate cortical areas and their corresponding HP ¹³C spectra for Sham and rTBI. Quantitative analyses of HP ¹³C (B) lactate, (C) pyruvate, (D) urea and (E) lactate/pyruvate for cortical areas. (F) T2 weighted image and HP ¹³C grid, red voxels indicate subcortical areas and corresponding HP ¹³C spectra. Quantitative analyses of HP ¹³C (G) lactate, (H) pyruvate, (I) urea and (J) lactate/pyruvate for subcortical areas.

Figure1: From left to right, axial sagittal coronal view of tracts reconstruction of cingulum (blue), superior longitudinal fasciculus (purple), inferior longitudinal fasciculus (yellow), uncinate fasciculus (light blue).

Figure 2: From left to right, sagittal (rgb color direction) and coronal view of tracts reconstruction of cerebello-thalamo-corticalmand (left purple and rightt red) cerebro-ponto-cerebellar tract (left light blue and right green).

Figure 2: A) Examples of qualitative and quantitative images. Example T1 and T2 maps, alongside co-registered SWI and FLAIR images are shown from a 34-year-old female mTBI patient at one month after brain injury. Note the stark T1 and T2 difference between grey and white matter. B) Boxplots of T1 and T2 distributions in mTBI patients and controls. Note that T1 and T2 from mTBI did not differ from controls for any regions. See Fig 1 for abbreviations of regions.

Figure 3: Correlations between T1 and T2 with patient assessments. A) T1 and T2 correlation frequencies measured in outlined regions with the neurological assessments (RPQ, BTACT, GOSE) are plotted for those correlations with p<0.05. Note the higher frequency of T1 correlations with clinical outcome at time 2 compared to time point 1. B) The same associations (p<0.05) are plotted against the correlation coefficient, but without separating the outcome measures. T1 showed more positive associations with clinical outcome at time 2, which were also generally stronger than those for T2.

Figure 3. Statistical results from the voxelwise analysis of long-range FCS showing two clusters ((a); CANSOF_FCS = 0.85 ± 0.17, CTL_FCS = 1.22 ± 0.32; P < 0.0001) and ((c); CANSOF_FCS = 2.06 ± 0.45, CTL_FCS = 1.56 ± 0.28; P = 0.0010). The data-driven reconstruction of associated white-matter network for each clusters (a,b) and (c,d) are shown respectively. CTLs = controls, CANSOF= Canadian special operations forces command, FCS = functional connectivity strength, ROI = region of interest, WM = white matter

Figure 4. Statistical results from the voxelwise analysis of the white-matter network associated with the long-range functional connectivity cluster in Fig. 3c,d (CANSOF_MD = 10.00 ± 2.13, CTL_MD = 7.43 ± 0.67; P < 0.0001). CTLs = controls, CANSOF= Canadian special operations forces command

Figure 1: Example T1-weighted image (left) with ROIs superimposed (caudate: R=yellow, L=dark blue; putamen: R=red; L=dark green; globus pallidus: R=pink, L=light blue; thalamus: R=light green, L=brown), first and final echo gradient echo (GE) magnitude image, and susceptibility map calculated using iterative Tikhonov regularisation and weak harmonic (WH) QSM for a representative control subject. WH QSM was used here as its reduced noise and residual background fields yielded significant $$$\chi$$$ differences unlike the iterative Tikhonov technique.

Figure 2: Box plots of deep gray matter regions with significant group differences in age-corrected mean susceptibility. HC=healthy controls, LTLE=left temporal lobe epilepsy, RTLE=right temporal lobe epilepsy. * indicates p<0.05, ** indicates p<0.01, *** indicates p<0.001.

Figure 2. Illustration of a group comparison of whole-brain COMMIT weights between of HCs and OIE. Multiple connections showed significant alterations both ipsilaterally and contralaterally. Group comparisons were performed using one-tailed t-tests. Significance was thresholded at p<0.001 uncorrected. Age and gender were accounted for in HCs whereas age, gender, side of epileptic focus, duration of epilepsy and age of onset of epilepsy were accounted for in OIE patients. Matrices in both HCs and OIE patients were masked based on a similarity threshold calculated in HCs.

Figure 1. Flowchart. Raw images were processed using Tractoflow. The output of Tractoflow and CIVET-calculated surfaces were used to build the tractogram using SET. COMMIT filtering was then performed. In parallel, the Freesurfer-calculated surfaces were used to generate Brainnetome parcels. The COMMIT-weighted tractogram and Brainnetome parcellations were used to derive structural connectivity matrices. Matrices of patients with right-sided OIE or TLE were then side-flipped and bundles that were anatomically dissimilar between controls were excluded in all matrices.

Summary

Overview of post-processing of MRI and FDG PET.

Figure 2. Simulated (A) and experimental (B) MTR_asym curves in the normal brain tissues for B₁ values of 1, 2, 3, and 4μT. The experimental curves (B) represented the mean MTR_asym spectra of the normal brain tissues in 9 TSC patients.

Figure 3. Comparison of experimental (A) and simulated (B-F) MTR_asym spectra with the proposed 2-stage method at four B₁ levels. The MTR_asym curves in normal tissues (red lines) were simulated with the parameters listed in Table 1. As for the MTR_asym spectra in cortical tubers (black lines), they were obtained via varying T₁ (part B), T₂ (part C), M_MT (part D), M_amide (part E), and M_amine (part F) according to Table 2B, in order to generate the mRMSE between simulated and experimental CEST contrast for each of the candidate Bloch-McConnell parameters.

Figure 1. Regions of interest traced on quantitative susceptibility maps on three representative slices showing deep gray matter nuclei in A) the basal ganglia; blue: caudate nucleus, orange: putamen, cyan: globus pallidus, green: thalamus, red: pulvinar thalamus. B) midbrain; purple: red nucleus, yellow: substantia nigra. C) cerebellum; orange: dentate nucleus.

Figure 2. Nigrosome-1 sign in the substantia nigra. The top row shows bilateral presence of N1 in A) tSWI, B) QSM, C) SWI and D) T2*W of a healthy control. E, F) tSWI and QSM images of an iRBD patient with unilateral loss of N1 in the left hemisphere. G, H) tSWI and QSM images of an iRBD patient with bilateral loss of N1.

Figure 2. Absolute perfusion image using voxel-wise calibration (left) and after correction volume effects around the edge of the brain (right)

Figure 3. CBF map extracted from the kinetic model (left). The image of estimated PVs of gray matter (center), and the estimated gray matter perfusion (right)

Example MS-qMRI maps of (A) T1, (B) T2, (C) PD, and (D) Spatial Entropy for a 15-year-old female.

Scatterplots of the mean T1 versus spatial entropy of the white matter within the intracranial matter. The solid lines represent linear regression of females (circle) and males (triangle). Distribution plots of the variables are provided to highlight gender related differences.

A comparison of (a) BMI and (b) cognitive function within bariatric surgery candidates before (Bar-pre), 2 weeks after (Bar-post 2wk), and 14 weeks after surgery (Bar-post 14wk). Compared to pre-surgery BMI, Bariatric surgery candidates have a significantly reduced BMI 2 weeks after surgery (p = 0.001) and 14 weeks after surgery (p = 0.0001). Cognitive function increases and is significantly greater 2 weeks post-surgery (p = 0.026) and 14 weeks post-surgery (p = 0.003) when compared to pre-surgery.

A comparison of CMRO₂ among bariatric surgery candidates before, 2 weeks after (Bar-post 2wk), and 14 weeks after surgery (Bar-post 14wk), young healthy controls (YHC) and age matched healthy controls (AM HC). Before bariatric surgery, patients have a significantly higher CMRO₂ than AM HC (p = 0.02), as well as 2 weeks post-surgery (p = 0.003).

S1 overview. A) T2 hyperintense lesion located at the base of the skull in the temporal lobe. Several pathways with anomalies interdigitate in the vicinity of the lesion. The shaded area (bottom right) shows the mean population profile (bold purple line, shaded area = +/-1 z-score). Although the IFO signal did not extend beyond the shaded area (C, D), the proposed anomaly detection framework identified abnormalities in that region (B, pink areas). Bold orange line: original tract-profiles. Dotted purple line: reconstructed representation learned from the network.

S2 overview. A) The lesion is located anterior to the right primary motor cortex in the supplementary motor area (hyperintense signal on the FLAIR image, hypointense on the RISH map). Tractography show tracts traversing the area (CC4). B) Anomalies were identified in the right CC4, CST, and SLF-I bundles (top right, pink areas). The bold orange line represents the original tract-profiles whereas the dotted purple line represents the reconstructed representation learned from the network. The z-score approach shows less focused anomaly patterns along the tracts (shaded area: +/- 1 Z).

Figure 3. Examples of activation maps highlighting the important Ω_i for prediction of expressive and receptive language score. Each 3D visualization shows the nodes of Ω_i and their contributions (i.e., Z scores of attenuation weights) to increase the prediction accuracy, quantified by the thickness of an individual tube. Thicker tubes indicate that a node is more predictive of observed language scores. Relatively small attention weights (i.e., Z-scores less than 2.5 standard deviations of each map) were omitted for clarity.

Figure 1. The architecture diagram of the proposed CNN+GCN model, which takes the connectome matrix S as input, to predict an output score t. Here, “**” indicates 2D convolution, and “*” indicates 1D convolution. In the final layer, a linear unit is used for language score regression.

Cluster dendrogram and heat map using cortical thickness measures in pediatric patients with epilepsy.

Regional cortical thickness differences between patient subgroups and healthy controls.

Figure 2. Cortical thickness in patients vs HC shown on the inflated surface of the brain. Red and blue show regions with significantly thicker cortex in the patients and HCs respectively.

Figure 1. Cortical areas with statistically significant increased thickness in patients compared with HCs

: Classification result using decision tree method. L : left TLE, R: right TLE.

Analysis of attribute importance using the random forest method.

Figure 1. The statistical comparison of hippocampal volumes, mean T2 and PET SUVR in ipsilateral hippocampus and hippocampal subregions vs contralateral regions. To the left of the line is the ipsilateral ROIs, and to the right of the line is the contralateral ROIs. MR-negative LTLE is shown in blue, and RTLE is shown in orange. The significance level in all patients is indicated by a black asterisk. * P < 0.05, ** P < 0.01, *** P < 0.001.

Figure 3. Receiver operating characteristic (ROC) curve of the classification-models.

Fig.2 Metabolic difference among each group in the left (a) and right (b) DLPFC. *p < 0.05. Abbreviations: DLPFC, dorsolateral prefrontal cortex; HC, healthy control; Ins, Myoinositol; NAA, N-acetyl aspartate; Cho, choline; Cr, creatine; Glx = Glu + Gln (Glu, glutamate, Gln, glutamine).

Table 2. Metabolite concentrations (mmol/kg wet weight) of DLPFC in pre-/post-treatment GTCS patients and HC. Abbreviations: Ins, Myoinositol; NAA, N-acetyl aspartate; Cho, choline; Cr, creatine; Glx = Glu + Gln (Glu, glutamate, Gln, glutamine). ^a p < 0.05, compared with metabolite concentrations on the right DLPFC; ^b p < 0.05, compared with metabolite concentrations on the same side DLPFC of HC; ^c p < 0.05, compared with metabolite concentrations on the same side DLPFC of HC; ^d p < 0.05, compared with metabolite concentrations on the same side DLPFC of pre-treatment patients.

Sagittal, axial and coronal views (line A, B and C) of the MP2RAGE sequence with the postprocessing morphometric analysis results of a 6 years girl having left focal epilepsia. UNI column is a highly contrasted T1 image. T1 relaxometry map is the second column. Morphobox column represents the brain segmentation obtained. A Volumetry and T1 relaxometry Z-score maps are also provided for a visual analysis. We noticed a decreased T1 relaxometry values in the left cortical occipital grey matter.

Figure 1. Neurochemical changes between lesion and contralateral side in epilepsy patients with focal cortical dysplasia and non MRI detectable lesion (* p< 0.05, + p<0.06).

Table 1. Demographic data for patients (n=13) including clinical presentation, seizure frequency, MR diagnosis and post-operative clinical outcome.

Table 1: Regional differences in mean R1 values from time 1 to time 3 in the NFS group.*corrected for multiple comparisons

Multi-Frequency Magnetization Transfer (MFMT) Sequence design. MFMT is a saturation prep followed by 3d-FLASH acquisition. MFMT saturation prep allows multiplexing up to four different/independent offset frequencies each with independent bandwidths – thereby simultaneously saturating up to four different frequencies. Sat RF pulses are Gaussian and run with RF phase stepping (RF spoiling).

Representative result Hippocampal transverse plane images qualitatively show a clear increase in contrast between MFMT ‘On’ versus ‘Off’. The yellow arrows specifically point to hippocampal regions and details with obvious contrast increase when MFMT is ‘On’.

Figure (3): In a) The final product of the hippocampus segmentation pipeline, each color represents a specific subfield. In b) Example of manual correction of a lesion that is labeled as DG.

Figure (2) First two stages of Image preparations for registration and segmentation. Top row shows the MPRAGE T1 weighted images, bottom row shows the TSE T2 weighted images

Figure 1. A) High-resolution T1-weighted image with circular ROIs in the globus pallidus (red) and frontal white matter (green) and visibly hyperintensities of the globus pallidus caused by Mn deposition. Similarly, b) T1-map, from which R1=1/T1 was calculated for several brain regions.

Figure 2. Receiver operating curves (ROCs) for PI, R1 in GP, FWM, and SN to distinguish between occupationally exposed subjects from unexposed controls. R1 in the GP yields the highest AUC (0.77) but only 86% sensitivity and 59% specificity. R1 in the FWM allows for similar discrimination between welders and controls with 91% sensitivity and 63% specificity (AUC = 0.74).

Figure 4. 3D IR-UTE imaging of myelin in a control mouse (A) and a mouse four days after open-field LIB injury (B). LFB for the control and mTBI mice shows a significant reduction in staining intensity within the corpus callosum (CC) (red arrows) for the mTBI mouse (C), consistent with demyelination induced by LIB. UTE-measured myelin density for the CC was reduced by ~25% from 10.6±0.8% for the control mouse to 7.9±0.9% for the mTBI mouse (F), largely consistent with LFB staining (C).

Figure 3. Representative axial images of an adult control C57BL/6 mouse: (a) T2-FSE, (b) 3D IR-UTE imaging at TE = 0.020 ms (c), and TE = 2 ms (windowed 10X), where myelin (thin arrows) and bone (thick arrows) signals dropped to near zero, consistent with short T2* relaxation times.

Figure 1: (A): A brief flowchart describing the generation of structural connectivity matrix for every participant. Average symmetric connectivity matrix for the impaired and nonimpaired group is shown. The color map indicates the average strength of connections between any two nodes in the connectivity matrix. (B): Paths showing weaker structural connectivity in boxers. Nodes and edges of the paths are represented in blue and red colors respectively. L and R represent left and right hemisphere.

Figure 4: Rich-club regime for both HC and boxers is shown in the shaded box. Nodes exhibiting rich-club properties are shown as yellow circles for both boxers and HC, and non-rich-club nodes are shown as green circles. Rich-club edge strength, feeder edge strength, and local edge strength are plotted as bar plots for both groups. K:degree; φ-norm: normalized rich-club coefficient

Figure 3: Top: Regions with significantly lower FWFA in boxers as compared to HC are overlaid on MNI template. Middle: Regions with significantly lower FWRD in HC as compared to boxers are overlaid on MNI template. Bottom: Regions with significantly lower FWMD in HC as compared to boxers are overlaid on MNI template. Average values in the cluster are extracted for each participant and shown as a boxplot. Average values in each significant cluster for boxers and HC are shown as blue and red dots, respectively.

Figure 1: Top: Regions with significantly lower ST dMRI-derived FA in boxers as compared to HC are overlaid on MNI template. Middle: Regions with significantly lower ST dMRI-derived RD in HC as compared to boxers are overlaid on MNI template. Bottom: Regions with significantly lower ST dMRI-derived MD in HC as compared to boxers are overlaid on MNI template. Average values in each cluster are extracted for each participant and shown as a boxplot. Average values in the significant cluster for boxers and HC are shown as blue and red dots, respectively.

Figure 1. Voxel-wise comparison (T-score map) of CBF in the group with TBI to the NC group. The colored voxels indicate regions with statistically significant CBF deficit in the group with TBI compared to the NC group (p=0.005, uncorrected, minimum cluster size (k) = 50 voxels). Cross sectional view through the CBF deficit cluster in: (a) the Thalamus and (b) the left Hippocampus. Colored voxels are overlaid onto a group averaged CBF map from all participants in both TBI and NC groups. CBF: cerebral blood flow.

Figure 3. Negative correlation between hippocampus CBF (a – d) and TBI-related symptoms of (a) Fatigue (p=0.002), (b) Anxiety (p=0.01), (c) Depression (p=0.01), and (d) Sleep Impairment (p=0.008) and negative correlation between rostral anterior cingulate cortex CBF (e – h) and TBI-related symptoms of (e) Fatigue (p=0.0003), (f) Anxiety (p=0.045), (g) Sleep Disturbance (p=0.01), and (h) Sleep Impairment (p=0.01). Significant correlations are denoted as: * p<0.05; ** p<0.005; *** p<0.0005

Video 1: (Link: https://youtu.be/QznHb3jMKeo) aMRI of a sheep pre and post a mild traumatic brain injury, showing increased brain motion post-injury. aMRI data were acquired with a cine bSSFP sequence, with a FOV of 23cm², matrix size = 194 x 240, TR/TE/flip-angle = 35ms/1.7ms/43°, voxel size = 1.2 mm³, 20 cardiac phases, scan time = 2:04 minutes.

Figure 4: The blood velocity vector field in the sheep intracranial carotid arteries pre and post a mild traumatic brain injury. The blood velocity vectors were increased after the impact delivery. The cine 4D flow datasets were acquired with peripheral pulse gating with imaging parameters as follows: FOV = 35 cm², matrix size = 168 x 208, venc = 80 cm/s, 60 slices, pixel size = 1mm x 1mm, slice thickness = 1.5 mm, scan time = 4:58 minutes.

Figure 1: TBSS results showing comparisons between three groups for DTI, DKI, WMTI metrics. Clusters of significantly decreased/increased voxels in (top) CS-SRC and (bottom) CS-nSRC (p < 0.05) are shown in red/blue, respectively, and overlaid on the standard template, together with the mean FA skeleton (green). No significant differences were found between CS-nSRC and CS-SRC groups. Also, AD was not significant in all comparisons.

Figure 1: Comparison of 3D Gradient-Recalled Multi-Echo (SWAN) sequences at 1.5T (left column), 7T (central column), and QSM at 7T (right column). Hypo-intense signals in SWAN sequences (yellow circles) can be differentiated in paramagnetic formations (magnetic susceptibility >0, red circles), associable to vascular malformations, or mineralizing microangiopathy (magnetic susceptibility >0, blue circles).

Figure 2: Examples of lesion representation at 1.5T (first column), 7T (second column), and QSM at 7T (third column). For each lesion surrounded by a colored box, the fourth column reports the corresponding zoomed depictions Whilst at 1.5T images their localization seems to be at the corticomedullary junction, 7T proves, thanks to its spatial resolution, that most lesions have an intracortical distribution.

Figure 2. Each row of panels shows one of two CVS lesions seen in one SLE subject experiencing NP events in axial, sagittal and coronal planes. Inserts in white rectangles are magnified views of the lesion, circled in yellow.

Figure 1. Pipeline to generate FLAIR* images.

Fig 1 Significant differences in resting-state FC between mTBI patients and healthy controls. The seed was located in the left parahippocampal gyrus and the cold/warm color showed the regions that have lower/higher correlation with seed point.

Table 1 GMV changes between mTBI patients and healthy controls. Multiple comparison correction was performed using a threshold (p < 0.001) of individual voxel and a cluster level p < 0.05(FWE corrected)

Figure 5: Sensitivity, specificity, accuracy, and area under the receiver operating curve (AUROC) is shown for each machine learning (ML) algorithm (RBFN: blue, Linear SVM: orange, Nonlinear SVM: gray, and random forest (yellow) for various features. Black dotted line represent the performance of any random classifier and dotted-dash lines represent 95th percentile of the benchmark measure for the respective ML algorithm across a various combination of the feature set, and are shown in the same colors as the bar-plot.

Figure 2: Top: Cluster showing a significant correlation between GM and years of fighting in impaired boxers. Mean GM density was extracted from each impaired boxer and plotted as a scatterplot. Middle: Cluster showing a significant correlation between WM and psychomotor speed in impaired boxers. Mean WM density was extracted from each impaired boxer and plotted as a scatterplot. Bottom: Cluster showing a significant correlation between WM density and psychomotor speed in nonimpaired boxers. Mean WM density was extracted from each nonimpaired boxer and plotted as a scatterplot.

(A)Pre-acupunctured differences of function connectivity between CPSD and HC groups. (B)The Pre- and post-acupunctured difference of functional connectivity in HC group. (C)The Pre- and post-acupunctured difference of functional connectivity in CPSD group. The red globule is the seed point and the blue globule is the region in functional connection to the seed point. The solid red line represents the enhanced connection, and the dotted red line represents the reduced connection strengt

Correlation between scores of the ANT task and ALFF and ReHo of CPSD group before acupuncture. (A) The ALFF value of the right parahippoclial gyrus was positively correlated with the accuracy of ANT task (r=0.637, P= 0.001) and negatively correlated with the omission rate (r=-0.427, P=0.047). (B) The ReHo value of the right superior frontal gyrus was negatively correlated with the reaction time in ANT task (r= -0.514P =0.014).

Figure 2 The differences of FA values between left and right hemispheres

Figure 3 The comparison of white matter asymmetry index among control and PWMLs groups.

The processing of the MEGA-PRESS spectrum averaged over normal group of subjects. GABA signal fit error = 8%.

Voxel location in posterior cingulate cortex, size 40x20x20 mm

Figure 1. Significant clusters of gray matter alteration in BINT participants vs. first responders. Whole-brain voxel based morphometry was performed to determine volume differences between BINT and occupational stress groups. Color map indicates scale for t-statistic.

Figure 2. Testosterone, cortisol and T/C ratio in BINT vs. chronic stress groups. Testosterone and cortisol are presented as Z scores of hormone concentrations, while the T/C ratio is calculated from raw values in pg/mL.

Fig. 1 demonstrates the results of the voxel-based analysis of λ1, λ2, MD, and RD. In the ME/CFS patients, there were significant decreases in λ1, λ2, MD, and RD in the pons of the brain stem.

Fig. 2 demonstrates the results of the voxel-based analysis of λ3 and RD. In the ME/CFS patients, there were significant increases in λ3 and RD in the medulla (white arrow) and cuneus (black arrow) region.

Figure 2: The difference of CBF in each step. The CBF values were represented by f. Compared with resting state, the regional and global mean CBF values of control state both decreased. Compared with before vibration (f_A), the regional and global CBF values both decreased after two vibration (f_B and f_C). And compared with the first vibration, the CBF decreased greater after the second vibration.

Figure 1: An illustration of the experimental workflow. In the first step, 3D-ASL was used to measure the CBF of subjects in the resting state. Scan time was 4.5 minutes, and the scan interval was 3 minutes. Then the subjects wear an actuator (off) to be scanned with 3D-ASL again (Control). In the second step, the subjects first wear an actuator (off) to be scanned with 3D-ASL (A). And then the actuator was turned on for 5minutes. After the vibration, 3D-ASL was used again to obtain the CBF of the subjects (B). To ensure the stability of the results, the vibration and measurement was repeated (C).

Fig. 4. The NAAG (left) and NAA (right) MEGA-PRESS spectra summed over the controls (blue) and patient (red) groups. The difference between the spectra are shown in black. Summed NAAG signal (δ = 2.6 ppm) significantly reduced in patient group as compared to controls.

Fig. 1. Typical VOI localization: dorsolateral pre-frontal area (WM-dominant brain region) VOI, 50×19×27 mm3

Figure 2. The area under the ROC curve of train and validation set based on radiomics model

Table 1. Clinical statistics in the diagnosis.