Figure 1: (a) Example of used vertebral levels from C2 to C5 on structural T1-w image and (b) example of FA map from C3 level with illustration of individual regions-of-interest defined based on probabilistic PAM50 atlas. Light blue - spinal cord (SC), red/light blue - gray matter (GM) and white matter (WM), yellow - ventral columns, light blue - lateral columns, green - dorsal columns.

Figure 2: Variations of individual qMRI metrics across vertebral levels from C2 to C5 for 6 different regions-of-interest (spinal cord, white matter, gray matter, dorsal columns, lateral columns, ventral columns). Mean and standard deviation values are shown.

Figure 1. Data and regions of interest (ROIs). Data consisted of 2 subjects scanned with a 26-shell protocol (35 DWIs + 1 b0 per shell), resulting in 936 image volumes per subject. This covers a range of q-space, including multiple diffusion times, b-values, and angular directions. 16 ROIs were automatically extracted based on registration to the spinal cord atlas, and are shown overlaid on the structural image, the b0 image, and the mean diffusion weighted image.

Figure 2. Multi-compartment models. Models consist of some combination of restricted cylinder compartments (C1-C4), Gaussian compartment (G1-G2), Restricted sphere compartments (S1-S3), and CSF compartments, with 8 different types of constraints used in the modeling process. This results in 480 different possible multi-compartment biophysical models of spinal cord microstructure.

Fig.5 – 7T MRI offers new perspectives for SC characterization. Some atypical presentations, with details not seen on 3T scanners are presented here. (a, b) reference HC, (c) MS lesion in posterior funiculi with atypical ventral gray horn shape, (d) strong vascular hyposignal, (e) central canal dilatation, (f) MS patient presenting with several lesions (blue *not seen at 3T), (g) ALS patient presenting with strong degeneration of lateral funiculi, (h) MND patient presenting with strong hypersignal in GM (snake eyes), (i) MND patient with severe atrophy of the central GM commissure.

Fig. 1 – Anatomical and structural native image quality assessment of 4 MR sequences (TSE, MP2RAGE, MGE, DTI) along the whole volume or for each slice from C1 to C7, using scoring between 0 and 4. Average scores and success rate were calculated for each modality and cohort. Images with a score ≥ 2 were qualified as "GOOD” and considered for the success rate calculation. B1+ and B0 inhomogeneities were evaluated as well using relative variations in %. BMI: body mass index (kg/m2).

Figure4: (A) Connectivity matrix (absolute values) averaged over 15 runs of 5 monkeys at different time points (Pre-SCI, Post-SCI Stage1 (2 weeks), Post-SCI Stage2 (7-8 weeks) and Post-SCI Stage3 (16-22 weeks). (B) Box-plot of the connectivity values at different time points . The dotted line joins the median of the box-plots. Significantly different box-plots using Wilcoxon non-parametric test are denoted by * (p<0.05) and ** (p<0.01) .

Figure 5: (A) Mean +/- SD of power spectra (0.01-0.1Hz) of all ICs (averaged over 15 runs of 5 monkeys ) at pre-and post-SCI stage 1. The shaded bar denotes the region where the drop in connectivity is conspicuously observed.(B) Mean of all IC’s power spectra at different time points i.e. pre- and post SCI stage1-3. (C) Box plot of averaged (0.01-0.1 Hz) spectral power for all ICs at different time points. The dotted line joins the median of the box-plots. Statistically significant difference using Wilcoxon non-parametric test are denoted by * (p <.05) and ** (p<.01).

Figure 1: MRI and histology data for patients from the International Spinal Cord Injury Biobank included in this study. Advanced MRI was collected 9mm above and 9mm below injury epicentre. IDI: injury-death interval; ihMT: inhomogeneous magnetization transfer; MWF: myelin water fraction; DTI: diffusion tensor imaging; RD: radial diffusivity; LFB: Luxol fast blue; FA: fractional anisotropy; AD: axial diffusivity; DBSI: diffusion basis spectrum imaging; FF: fibre fraction; pNF: phosphorylated neurofilament.

Figure 4: Spearman correlations between Luxol fast blue (LFB) staining for myelin phospholipids and three myelin-sensitive MRI metrics (radial diffusivity (RD), myelin water fraction (MWF), inhomogeneous magnetization transfer (ihMT)). All three metrics had significant, moderately strong correlations with LFB. MWF had the strongest correlation, RD the weakest. T_1D filtering slightly increased the strength of the correlation with ihMT.

Figure-4. Percentage of HRF variability in the gray matter. In both the runs, RH consistently showed about 6% variability across quadrants, spinal levels (=slices) and subjects. TTP showed higher variability than RH (7–9%) but seemed less consistent across runs. This lesser consistency could perhaps be due to the poorer temporal resolution of our data (VAT=2.08s) (this is also a limitation of the study), as TTP estimates are quantified in units of VAT.

Figure-3. Simple statistics of response height (RH), time-to-peak (TTP) and full-width at half max (FWHM) in the cervical spinal cord. Mean ± standard deviation of HRF parameters aggregated across all quadrants, levels and subjects is shown for both runs. The last column comparing runs 1 and 2 shows that aggregated HRF parameters are consistent across runs. The last row comparing gray (GM) and white matter (WM) shows that RH and TTP are significantly higher in the GM.

Figure 1

a) 2D-T2w gradient echo image with the corresponding ²³Na MRS voxel (SV) positioning in the cervical spinal cord in yellow, b) representative ²³Na spectra from in vivo and phantom scans, used to quantify total sodium concentration. Example of multi-shell diffusion NODDI maps for c) fiso (fraction of free protons) and d) ficvf (fraction of restricted protons) for extracellular and intraneurite space evaluation, d) macromolecular tissue volume (MTV) map to quantify MTV in the spinal cord.

Figure 2.

Results for total sodium concentration adjusted for gender in the cervical spinal cord (aTSC), in both healthy controls (HC) and cervical myelopathy patients (CM). a) boxplots for HC and CM aTSC (* p<0.05).

Regression variable plots for aTSC in HC (blue) and CM (red) groups. aTSC correlations to b) ficvf (fraction of restricted protons, an indication of intracellular space) c) fiso (fraction of free protons, a measure of extracellular space) and d) macromolecular tissue volume (MTV).

Figure 1. Overview of template registration pipeline. Initially, T2-weighted scans are registered to the template (top row). DTI data acquired during the same scan session are then registered to the anatomical data and PAM 50 objects are warped to diffusion data (bottom row) to generate the pediatric WM spine atlas.

Figure 3. Spinal cord WM atlas: A. An atlas of spinal white matter tracts derived from the Gray’s Anatomy atlas B. Generated white matter atlas of pediatric spinal cord overlaid on a b0 image. The selected tracts are labeled with multiple colors. C. The C3 level is marked in green color on the sagittal T2-weighted scan.

Figure 4. (A) Connectivity matrix at different time-points arranged such that the connectivities from same community are next to each other. Red boxes on the matrix highlight the communities formed using graph theory principles. (B) The resulting communities are shown in different colors overlaid on a grid matrix. The columns of grid matrix represent the seven ROI and the rows represent the five slices/segments. No IR component was observed from slice 2 and kept white in color. Note community structures change and hence their number and colors don’t match across time points.

Figure 3: (A) Connectivity matrix averaged over 14 runs of 5 monkeys at different time points with intra-slice and neighboring inter-slice connectivities highlighted in black triangles and black box respectively. Box-plot of (B) intra-slice connectivities and (C) neighboring inter-slice connectivities at different time points. Box-plots for intra-slice and neighboring inter-slice are computed from their corresponding significant connectivity measures. Significantly different box-plots (Wilcoxon non-parametric test) are denoted by * (p<0.05) and * * (p<0.01) . .

(A) Heatmap organized by tissue type and (B) by vertebral level (C3-C7) in a BH scan. Two structures of motion are notable in these plots: a more prolonged movement indicated by the bracket that appears predominantly in the R_z and T_y motion traces, and a brief movement indicated by the arrows that is most pronounced in R_x, R_z, and T_y. The prolonged movement is discernible in GM and in WM tissues. The brief movement appears to have a more continuous gradation along the y-axis when organized by vertebral level, and the signal increases in magnitude from C3 to C7.

Three heatmaps of a BH scan organized first by (A) vertebral level (C3-C7). Red arrows indicate the location of BHs which coincide with the P_ETCO₂ and HR traces above the heatmap. (B) Heatmap reorganized by GLM derived t-statistics for HR. (C) Heatmap reorganized by GLM derived t-statistics for P_ETCO₂. Motion regressors were also included in the GLM but are not shown here. For simplicity of the color scale, the absolute values of the t-statistics are shown. Though the organization is changed, the structured variation from the BH paradigm is visible in all three heatmap variations.

FIG. 4. QSM image of SPIO.Mag: Amplitude map, in which nano-iron oxide is black; QSM: The post-processed susceptibility distribution diagram, in which nano-iron oxide is white. We observed characteristic SPIO signals in the T2* hypointense signal region.

FIG. 5. Trend graphs of T2* values and susceptibility values changing over time.The susceptibility values per unit pixel decrease with time, and the total volume of iron increases with time. The T2* values also show a trend of increasing with time.

Figure 3: Cerebral Metabolic rates of glucose oxidation (CMR_Glc(Ox)) in A. Cerebral cortex and B. Spinal cord. The bar graph represents the mean±SD of the group.

Figure 4: Cerebral Metabolic rates of acetate oxidation (CMR_Ace(Ox)) in A. Cerebral cortex and B. Spinal cord. The bar graph represents the mean±SD of the group.

Fig.5: Mean normal appearing T₁-MP2RAGE map calculated over all MS patients (B) (after lesion filling) and age matched HC (A) in the PAM50 space and lesion distribution map (C). T₁ values in NA tissues were statistically higher in MS patients (NAWM= 969±64/NAGM=970±47ms) as compared to HC_MS (902±22 and 939±21). Impairments were more particularly marked in NAaiGM in the upper cervical levels and in the whole WM, especially in CST/LST (p<0.0001).

Fig.2: Representative T₁-MP2RAGE maps obtained on HC (A), MS (B) and ALS (C) and the selected ROIs. For the MS patient, red arrows point on GM/WM lesions

Fig 1. The pipeline of segment, normalization, scaling, feature extraction, feature selection, and model building.

Fig 4. The partial dependence plots, the local dependence plots, and the accumulated dependence plots for each feature.

Figure 2. Sagittal and axial views at various cervical cord levels showing clusters of significant cord atrophy progression at 1-year follow-up (color-coded for t-values, p<0.001 for display purposes) in: A) multiple sclerosis (MS) patients, considered as a whole; B) relapsing-remitting (RR) MS patients; and C) progressive (P) MS patients. Red boxes indicate clusters of cord atrophy progression significant at the time x group interaction vs HC. Abbreviations: L=left; R=right; A=anterior; P=posterior.

Figure 1. Sagittal and axial views at various cervical cord levels showing clusters of significant cord atrophy differences (color-coded for t-values, p<0.001 for display purposes) in: A) multiple sclerosis (MS) patients, considered as a whole, vs healthy controls (HC); B) clinically isolated syndrome (CIS) patients vs HC; C) relapsing-remitting (RR) MS vs CIS patients; and D) progressive (P) MS vs RRMS patients. Abbreviations: L=left; R=right; A=anterior; P=posterior.

Figure1. Representative metric maps of a case with multiple sclerosis (41-year-old woman) and analysis process.

Figure 2. The results of all metrics values of white matter of spinal cords at each spinal level in patients with MS and NMOSD.

Figure 1. MRI protocol including T2*-weighted images and Diffusion Tensor-derived maps as conducted at the lumbar enlargement level (T11-L1).

Figure 2. Box plots of cross-sectional areas of total spinal cord, grey and white matter areas in the cervical cord (at C2/3 level) and in the lumbar enlargement (at level T11-L2) of both DCM patients and healthy controls. ***p<0.001, *p<0.05.

Figure 2: Comparison of average CSA measurements at the C1/C2 level using SCT and the in-house technique (LT). Although SCT generally obtained lower CSA measurements, the relative reproducibility of CSA was very high, both on an individual level (see left regression of individual data) and on a group level (compare middle and right). Statistical analysis involved a non-parametric Kruskal-Wallis test with post-hoc multiple comparison correction by Tukey-Kramer. P-values are uncorrected for testing both SCT and LT.

Figure 1: Example spinal cord segmentations at the C1/C2 level, SCT (left), LT (right). MPRAGE images without overlaid segmentation are shown for reference in the center. For display purposes, the SCT segmentations were upsampled to the LT standard and slices were visually matched. Arrows point toward partial segmentation of spinal nerve roots. For subject 1, (top row, female, RRMS, 33 years), mean SCT-CSA was 65.04 mm², compared to LT-CSA=74.03 mm². For subject 2, (bottom row, female, HC, 42 years), mean SCT-CSA was 53.49 mm², compared to LT-CSA=61.26 mm².

Figure 1 Multi-contrast MRI in acute cervical cord contusion

The extent of the cord damage is captured in different modalities in both unilateral- and bilateral contusion injury. Edema (hyperintensity, T₂w), hemorrhage (hypointensity, T₂*w), vasogenic edema (blue, black triangle), and cytotoxic edema (red, white triangle) are evident in their respective contrasts. fDWI images clearly delineates focal axonal damage compared to T₂w, T₂*w and DTI.

Figure 2 Inter-subject registration pipeline

A histologic spinal cord atlas was digitized across all spinal cord levels to generate a template with pseudo T2-weighted contrast used for MRI registration. Manually marked spinal cord level aided initial z-level alignment and straightening followed by x-y plane registration to the template. Registered single subject images were averaged and registered iteratively across all subjects with the group mean image showing clear spinal cord anatomy.

Figure 2. Example slices for a control (M, 35 years old) and a participant with MS (F, 42 years old, EDSS 1.5). Left to right: anatomical T2*-weighted mFFE for segmentation; grey matter horns regions of interest overlaid on average functional run; FA and RD diffusion maps. FA = Fractional Anisotropy. GM = Grey Matter. RD = Radial Diffusivity.

Figure 4. Ventral and dorsal network z-scores by subgroups: healthy controls (HC), MS with low FA/RD, MS with high FA/RD. FA = Fractional Anisotropy. RD = Radial Diffusivity. NS = Non Significant. *Significant at p<.05.

Fig 3. t-test results of 5 parameters between poor & good recovery groups. Results show that patients in poor-recovery group have significant lower rEnhance, slope1, but possess increased FWHM

Table 2. Univariate/multivariate Logistic Regression for prediction of poor postoperative recovery (mJOA≤15)

Fig. 1: A: T2-weighted sagittal turbo spin echo image of the lower spine used for slice positioning. The 20 axial-oblique slices are positioned individually in each subject to encompass the lumbar enlargement and the conus medullaris. B: Corresponding axial slices from caudal (slice 1) to rostral (slice 20) direction. C: In each slice (here: slice 14 is shown), spinal cord and gray matter (GM) are segmented manually and a cerebrospinal fluid (CSF) mask is defined anterior to the spinal cord. The white matter (WM) mask is obtained by subtracting SC and GM masks.

Fig. 3: Dependency of signal-to-noise ratio (SNR) of gray matter (GM) (blue curve) and white matter (WM) (orange curve) on the number of echoes. The two curves run largely similar, increasing from 1 to 2 echoes and minimally decreasing afterwards. B: Dependency of contrast-to-noise ratio (CNR) between GM/WM (blue curve) and WM/CSF (orange curve) on the number of echoes. While WM/CSF CNR steadily increases with more echoes, the WM/CSF contrast exhibits a very flat peak at 3-4 echoes. In both subplots, error bars represent standard deviation across subjects (n=10).

Figure2. Two 29-year-old men, prolonged-standing affected L1/2 disc, sedentary affected L5/S1 disc, the nucleus pulposus was poorly demarcated from the annulus fibrosus

Figure 1. L4/L5、L5/S1 NP values were negatively correlated with age

Figure 1: T₂-weighted image of the cervical spinal cord of a DCM patient showing the segmented cord in red and the compression site displayed on the inset.

Figure 2: A: % BOLD signal correlation with spinal cord compression volume when tapping with left hand. B: % BOLD signal correlation with spinal cord compression volume when tapping with right hand.

Figure 1: Relationships between UNI signal and estimated T₁ value. (a) The protocol Pr0, provides the highest CNR at the expense of acquisition time. (b) The protocol Pr1 provides a high CNR, while allowing to cover brain and CSC in less than 8 minutes. It should be noted that a ±20% B₁⁺ variation for a T₁ of 1350 (brain GM) for instance, leads to an estimation error of 8.7%, showing the necessity for B₁⁺ correction. Protocols previously optimized for (c) studying brain^2,14, and (d) SC⁷, independently. (e) Derivative of the UNI signal, providing an indication on tissue discrimination.

Figure 3 : A representation of the UNI-denoised image (no unit) for Pr1 and 3 on brain and Pr1 and 2 along the cervical SC. In the bottom-right image we can see a zoomed-out section of spine, which shows the need for sub-millimetric resolution for imaging SC. We can observe a nice delineation between different structures of brain and SC, with a higher signal difference (contrast) between GM and WM for the optimized protocol 1 (arrows).

(a) Localizer of a volunteer with metallic implants and measured slice stack highlighted in orange. (b) 30 1x1x4mm³ transversal ZOOM EPI acquisitions (c) 30 1x1x4mm³ transversal ZOOM MAVRIC FSE acquisitions with a frequency increment of 500 Hz and 9 steps. The spinal cord is highlighted by red dotted ellipsoids.

(a) Localizer of water phantom with a diameter of 2.8 cm, surrounded by 4 aluminum screws, with 2.9 cm and 3.3 cm to each other. (b) 7 out of 20 inner FOV FSE MAVRIC acquisitions with a resolution of 1x1x4mm³ (c) 7 out of 20 inner FOV FSE MAVRIC acquisitions with a resolution of 1x1x4mm³ and 81 MAVRIC frequency steps of Δω 500 Hz

Outline of data processing, starting with temporal mean of fMRI time series, manually drawn spinal cord masks and comparison; then non-linear registration using the Spinal Cord Toolbox; then the post-registration segmentations and comparison.

Illustration of edge masks and box plots of percent of voxels in agreement along the edges in input masks and registered segmentations across all raters and participants. Edge masks were created by a 2-pixel erosion subtracted from 2-pixel dilation of the expert mask. Bar in box plot equates significance (p < 0.01).

Figure1 Upper row:b=0s/mm2,b=800s/mm2,mean diffusivity(MD),and fraction anisotropy(FA)map.Bottom row:Fiber tractography(FT)results of the iumbosacral nerves,a maximal intensity projection(MIP),FT and T2WI FFE combined,and disply of the segments at each level region of interest(ROI) placement.Dates obained by using diffusion tensor imaging based on SENSE1.6.

Figure2 Upper row:b=0s/mm2,b=800s/mm2,mean diffusivity(MD),and fraction anisotropy(FA)map.Bottom row:Fiber tractography(FT)results of the iumbosacral nerves,a maximal intensity projection(MIP),FT and T2WI FFE combined,and disply of the segments at each level region of interest(ROI) placement.Dates obained by using diffusion tensor imaging based on mutiband SENSE 2.

Image quality comparisons of cervical MRI between Con-MRI and CS-MRI.A-C: Con-MRI. D-F: CS-MRI.

Image quality comparisons of thoracic MRI between Con-MRI and CS-MRI.A-C: Con-MRI. D-F: CS-MRI

Fig. 2 Calculated DTI and MTR metrics for the non-compressed slice (No. 1) and the compressed slice (No. 2).

Fig. 3 Spearman correlation between mJOA and FA (left), mJOA, and MTR (middle). Both 2 metrics are extracted from DC at C2 level. The direct correlation between FA and MTR is also significant (right).

Figure 3: Anatomical comparison of a healthy subject scanned (a) at 3T with a 4-channel coil that wrapped around the thigh and (b) at 7T with a 28-channel QED coil. The individual fascicles of the sciatic nerves are clearly visible in the 7T image, but are more difficult to make out in the 3T image.

Figure 4: (a) A 7T image of a subject with peripheral neuropathy and (b) a DESS ADC map computed in the framed section of panel a. (c) A conventional EPI DWI diffusion map for comparison. The EPI map not only has worse resolution but also has severe distortion.

Fig. 2: Mean IVIM parameters maps across HCs (N=11) at mid C1, C2, C3 levels in the template space (0.5mm isotropic, average of 15 slices) for each diffusion-encoding direction (A-P: anteroposterior, R-L: right-left, I-S: inferosuperior) and averaged across directions. The mean MEDIC image across subjects helps visualize GM and evaluate the quality of the template registration. Below, the mean signal profile is shown for each direction. The linear fit on high b-values helps visualize the deviation from the diffusion-only signal decay, illustrating the challenge along the I-S axis.

Fig. 3: Slice-wise f_IVIM and D^* maps in a representative healthy volunteer (30 years old man) for diffusion encoding in the transverse plane (⊥, mean across maps with A-P and R-L diffusion encoding) and along the I-S axis. The mean maps across slices (after registration based on the MEDIC middle slice image) are shown at the bottom. The slices position is displayed on the sagittal anatomic image (T₂-weighted turbo spin-echo, 0.3×0.3mm² in-plane resolution, 2.75mm slice thickness). The MEDIC image resliced to the IVIM data resolution helps visualize the GM shape along with the IVIM maps.

Fig.1 Different Views of Electric Stimulation(Case1 Red, Case2 Green) and Magnetic Stimulation(Blue). The electrodes in Case2 is 1.3cm more distal than Case 1 and the distal side electrode is closer to the nerve below it than Case 1)

PNS response by (a)ES and (b)MS. In(b), refence magnetic field=0.0825mT/A is at 1.2cm above the center of the coil

Figure 5: Two representative MS patients: Patient A (25yr female, EDSS=0) and Patient B (38yr female, EDSS=1) without (top) and with (bottom) Patch2Self denoising applied to the dMRI data. Lesions in the lateral columns (white outlines) of both patients are more conspicuous in the denoised diffusion volumes, including one lesioned area (dashed outline) that is not apparent in the corresponding mFFE anatomical image for Patient A.

Figure 1: Flow diagram of pre-processing steps for Patch2Self denoising of spinal cord diffusion MRI data.

Bar graph with included standard deviation bars comparing volumetric blood flow of the internal carotid and vertebral arteries with and without LBNP. Note the drop in volumetric blood flow after the application of LBNP (p=.007).

Bar graph with included standard deviation bars comparing internal jugular vein cross-sectional area with and without LBNP. Note the drop in cross-sectional area after the application of LBNP (p=.03).

Figure 3: Exemplary MAVRIC DW ADC maps of the cord in a CSM subject at the indicated levels (A), which are (B) unfused (above the hardware) and (C) fused by the hardware construct. The clear reduction in ADC in this subject at the fused levels is readily visible.

Figure 1: A) Sagittal plane of isotropic T2w MAVRIC SL acquisition indicating fusion hardware and the location of displayed axial sections. B) Axial reformat of MAVRIC acquisition at C5 level indicating proximity of hardware to spinal cord. C) Conventional FOCUS (reduced field-of-view EPI) DWI b=0 image is non-diagnostic and completely disrupted due to hardware presence. D) MAVRIC DWI b=0 image allows for diffusion-weighted analysis of cord, even in near vicinity of hardware.

FIG.1.Methods for measurement of mMI and mean T2 value in a 53-year-old patient with bilateral DON (top row) and a 50-year-old GO patient without DON(bottom row). Axial fat-suppressed T2-weighted images in orbit with DON(A) and without DON(B). Oblique coronal T2WI of Dixon show mMI at 21mm with DON(C) and without DON(D) [mMI was the maximum value between (a+b)/c and (d+e)/f)]. E.F Coronal T1WI was presented as anatomical reference. T2 value of optic nerve of the eye with DON was 207.3ms(G) and the eye without DON was 105.5ms(H) respectively.

FIG.2. MRI indicators for prediction of DON.Receiver operating characteristic curve analysis(A) and decision curve analysis (B) showed that combination of mMI at 21mm and optic nerve mean T2 value demonstrated optimal diagnostic efficiency for DON.

Figure 1: (a) Localizers showing the typical geometric setup for cortico-spinal fMRI: the brain (orange) and spinal cord (blue) volumes and the isocenter (red cross). (b, c) in vivo EPI example images of the (b) brain (16 of 60 slices), and (c) spinal cord (6 of 12 slices) using SMS acceleration with a factor of 3 for the brain volume only.

Figure 3: Results of SNR measurements on an in vivo volunteer covering a brain volume only and combined brain and spinal cord volumes in the same acquisition: (a) spinal cord without simultaneous multi-slice imaging (SMS) and (b-d) brain volume (b) without SMS acceleration and (c,d) with SMS acceleration factors of 2 and 3, respectively.

Figure 1: (A) schematic of framework methods. Axial image of the forearm from (B) T1-weighted anatomical image, (C) FA map, (D) DEC-TDI. In B-D, the axis is centred at the same location. (E) shows a three-dimensional reconstruction of the ulnar nerve generated by volume rendering the nerve DEC-TDI image and overlaying on semi-opaque T1 volume render. In D and E, the colours indicate directionality (red for left-right, blue for proximal-distal, and green for anterior-posterior).

Figure 2: Measurements of (A) track-weighted fibre orientation distribution, (B) track-weighted apparent diffusion coefficient, (C) track-weighed fractional anisotropy, (D) track-weighted axial diffusivity and (E) track-weighted radial diffusivity. Profiles have been realigned such that a displacement of 0 mm roughly corresponds with the head of the radial bone, shown as vertical black line in (F). Individual subject test and re-test measurements are colour matched, and are shown with solid and dashed lines, respectively. Average test and average re-test are shown in black.

Figure 1. R2* map of sample 4. The left figure is a sagittal plane of the sample. Slides within the red lines were selected for superior colliculus depth analysis. The right figure is an axial plane of the sample. Red dash line is the second order polynomial fit to the boundary between periaqueductal gray and superior colliculus. The solid lines in different colors are the orthogonal vectors that were used for the depth analysis.

Figure 3. R2* depth profile along the superior colliculus layers. The green region indicates the intermediate layers and the purple region indicates the deepest layer. Solid black lines plot the mean R2* of all vectors in all selected axial planes (95 % confidence interval in shaded grey). Dashed black lines indicates the starting point of the periaqueductal gray. Samples 3 and 4 show one peak in the intermediate layer and another peak in the deep layer. Sample 1and 2 shows similar trend. PAG : periaqueductal gray

Table 3. comparison of GM to WM signal intensity measured from the high resolution PSIR and FFE

Figure1 Axial cervical spinal cord PSIR (a) and multiple FFE (b) images

Figure 4 A and B: Measured SNR on the phantom and a live rat (male Sprague-Dawley rat, 350 g) using Varian single loop coil and optimal quadrature coil specialized to L1-L2 spinal cord imaging. SNR maps were calculated from GRE images with the following parameters: phantom (FOV = 45 × 45 mm2, TR/TE = 1000/4 ms, FA = 20 degree, Matrix = 192 × 192, Bandwidth = 260.4 Hz/pixel), in vivo (FOV = 45 x 45 mm2, TR/TE = 200/2.874 ms, FA = 40 degree, Matrix = 128 x 128, bandwidth 390.6 Hz/pixel). C and D: One-dimensional profiles of the measured SNR along the dotted white line in Figures 4A and B.

Figure 2 CAD drawing and constructed including animal management device (left) and coil management device (right). These devices were designed in SolidWorks (SolidWorks Corp., Santa Monica, CA) and printed (except the long tubes) by a ProJet 3D printer (HD 3500 Plus, 3D Systems, USA).

Table 1 Basic Information Of Volunteers

Fig 1 MRI of a 34-year-old women with C3/4 horizontal cervical spinal cord.A(PSIR imaging),B(DTI and PSIR blending images),C(ROIs)

Figure 4. Example results of the described pipeline. A) and B) provide reformats of the T2w 3D-MSI, while C-D) provide views of the automatically segmented cord performed using the workflow outlined in Figure 3. E-F) provide maps of CSF z-score analysis, whereby the mean T2w signal of the CSF (corrected for signal shading an bias within the cord) is used to identify regions of low CSF (blue areas in the map). The blue arrows highlight a region of missing CSF (cord impingement), which is accurately reflected by the sharp blue region in the z-score map.

Figure 2: Example T1 (A,B) and T2 (C,D) weighted 1.2 mm isotropic acquisitions in orthogonal reformatted (sagittal, A,C – axial, B,D) view planes. The isotropic acquisitions required 5 minutes each and were designed for morphological cord analysis rather than conventional planar diagnostics.

a.Coronal image of brachial plexus.The area of interest(ROI) in the nerve and muscle;b-e.Mip images of brachial plexus of sense2,cs2,cs4,cs6,respectively;the scores of image quality are 4 for b-e.

a-e:Boxplot depiction of CNR,signal of nerves,image noise and image quality,respectively.

Figure 3. Comparison of the conventional FFE with the FFE-CS. LSP magnetic resonance neurography images of a 26-year-old woman without lower back pain. The coronal conventional 3D FFE (a) serves as a reference for the FFE-CS with factor CS2(b) and CS3(c). No difference can be observed among the three types of FFE, all sequences showing compromised image quality. A general noise enhancement could be observed in the FFE sequences with compressed sense. The subjective image quality score for LSP given by the two radiologists was both 3 in the conventional and compressed sensing sequences.

Figure 1.A healthy volunteer, 56-year-old male.

Left: Coronal B-FFE-XD image of the lumbosacral plexus. Right: Coronal B-FFE image of the lumbosacral plexus

Figure 1. The roots of the brachial plexus emerging from the intervertebral foramina (upper left image) and their relationship to the scalene muscles and vasculature of the upper limb (upper right image). The lower image is a simplified schematic of the brachial plexus highlighting (in purple) the spinal roots. Reproduced with permission from Mr Donald Sammut.

Figure 3. A scatterplot showing the negative association between the mean diffusivity of the roots of the brachial plexus and the mean age of adults in the included studies.

Figure 1. female, 48-year-old, trigeminal neuralgia. X-ray image and ROI measurement of left Meckel cavity in transverse, sagittal and coronal position in MRI images.

Table 1. Comparison between the average values of ROI measurement and balloon filling volume of Meckel cavity