Figure 1. (Left) Representative Z-spectrum from AAV2 (5.25 x 10⁸ viral genomes/μL) acquired at 800MHz using 5s continuous wave saturation at saturation power of 5.7μT demonstrates significant variation between 0 – 2ppm compared to background poly-L-Lysine (PLL) CEST contrast. (Right) CEST contrast derived from Z-spectra was quantified via MTRasym and showed substantial contrast between 0.6-0.8ppm in AAV2. The source is exchangeable hydroxyl protons on the capsid surface.

Figure 4. (Left) MTRasym spectra generated from Z-spectra acquired in phantoms containing varying capsid densities of AAV2, expressed as viral genomes per microliter, at pH 6.0 using saturation power of 6.3μT. (Right) CEST contrast demonstrated a correlation (contrast = 7*10^-8* viral concentration + 1.9%) with capsid density.

Figure 1. Corpus callosum demyelination in cuprizone mice. Demyelination follows a rostro-caudal pattern. A) Lateral callosal demyelination rostrally. B) Medial callosal demyelination caudally.

Figure 4. Cuprizone mice experience severe demyelination after 12 weeks on cuprizone diet compared to controls (reduced green fluorescence). Near complete demyelination seen in medial corpus callosum (A, E) and cortex (B, F). Partial demyelination was seen in the striatum (C, G) and the thalamus (D, H). Statistics were performed using Student's t-test (I). p < 0.05*; <0.01**, <0.001***.

Brain-wide BOLD activations upon optogenetic stimulation of MD versus VPM at different frequencies. (A) Illustration of atlas-based ROI definitions in the sensorimotor cortices, higher-order cortical and limbic regions, thalamus and brainstem, and basal ganglia. (B) BOLD activation maps upon optogenetic stimulation of MD and VPM at different frequencies (Bonferroni-corrected p<0.001). MD and VPM stimulation recruited distinct brain-wide targets, but both evoked brain-wide cross-modal BOLD activations at 8Hz. (C) BOLD profiles extracted from atlas-based ROIs.

Brain-wide BOLD activations upon 8Hz optogenetic stimulation of MD versus VPM at different number of pulses. (A) Atlas-based ROI definitions in the sensorimotor cortices, higher-order cortical and limbic regions, thalamus and brainstem, and basal ganglia. (B) BOLD activation maps upon 8Hz stimulation of MD and VPM at different number of pulses (Bonferroni-corrected p<0.001). MD and VPM stimulations showed similar decreases of brain-wide cross-modal BOLD activations when varying the number of pulses from 24 to 16/8/96. (C) BOLD profiles extracted from atlas-based ROIs.

Figure 1. Evoked functional changes to a set of 2 different stimulation frequencies. Main response area is marked with red circle and located in primary somatosensory area, barrel field (SSp-bfd).

A) Areas in the mice brain show positive functional changes to 4 Hz stimulation frequency.

B) Areas in the mice brain show negative functional changes to 20 Hz stimulation frequency.

Figure 3. Measured local calcium response. Trace displayed is the signal average within 6 mice. Subfigures A), B), C), D), refer to measured calcium signal response at respectively 2, 4, 8 and 20 Hz of stimulation frequency. Inset figure represents continuous calcium trace of all 7 different stimulation frequencies, blue solid shadowed area captures relative position of the main figure signal within stimulation trial. Units of axis on inset and main figures are same. Yellow line on the main figure denotes signal from same experimental condition but opposite whisker pad stimulation.

Figure 1. Sample images showing the MRI-derived g-ratio maps for a WT mouse (a) and a HET mouse (c). The genu region is indicated in (a). EM images with 1000x zoom and segmentation of myelin (red) and axons (green) for the same two mice are shown in (b) and (d) for the WT and HET mice respectively.

Figure 3. Error and scatter plot showing the MRI measurements of AVF, MVF and g-ratio of the WT mice (blue) and HET mice (red).

Figure 1. Experimental design for awake rsfMRI acquisitions. A) Mouse cradle for rsfMRI acquisitions. B) Body weight measurements in 5 representative mice across different sessions. C) Experimental timeline for habituation protocols [Abv; PS1: post-surgery day-1, PS10: post-surgery day-10, Hab: Start of habituation, HS1: habituation session-1, HS2: habituation session-2, HS3: habituation session-3, Scan: rsfMRI acquisitions]

Figure 3. rsfMRI connectivity networks differ between awake and anesthesia. A) Seed-location for whole-brain connectivity quantifications in (B). B) Whole-brain group-averaged connectivity matrices for awake, anaesthesia and mean correlation differences across states. c) rsfMRI connectivity maps across states, and corresponding difference maps for six representative networks.

FIGURE 1. MRI sequences and protocols. (A) Three different saturation modules (CPMG, onVDMP, and onSL) followed by the same readout module (RARE). DGE MRI protocol that used in DGE experiments with 50% w/w (B) and 25%/12.5% w/w (C) D-glucose injection.

FIGURE 2. Comparison results of three DGE MRI methods with 50% D-glucose injection. Parenchyma (A) and CSF (B) DGE images acquired by CPMG, onVDMP and onSL. Parenchyma (C) and CSF (D) DGE curves acquired by CPMG, onVDMP and onSL. DGE images were averaged over sets of 4 for display (7 out of 28). Fitted ΔSmax of parenchyma (E) and CSF (F). SNR of parenchyma (G) and CSF (H).

(a) QUTE-CE MRI Angiographic Quality at 3T. Imaging parameters were: TR=4ms, TE=0.1, Flip=20°, 0.4mm isotropic, 6:40 scan time, (b) Images from Cones-AFI and Resultant B1+ maps. The top two rows are from sequential ordering, with higher spoiler gradient areas for the second row. Some residual transverse magnetization affects a signal level in the center of FoV, resulting high B1+ values. The bottom row images are from golden-angle ordering, and the resultant B1+ map is really similar to Bloch-Siegert B1+ map.

Maximum intensity projection images (MIPs) of the whole rat head at 8 months are displayed (a) pre-contrast and (b) post-contrast 14mg/kg ferumoxytol. Unique contrast-enhanced vascular MIPs are obtained and (c,d) the brain is segmented . (c) pre- and (d) post-contrast images demonstrates that the time-of-flight signal enhancement is limited to the arteries at the periphery of the FoV. (e-g) MIPs of a rat head after intrathecal injection demonstrate glymphatic (e) pre-contrast (f) after 56 minutes. (g) The dynamic mixing is visualized from 7-56min in subtraction images.

MR images of an ICH mouse at different time points (From left to right: before, day 1, day 3, day 7, day 14) Pixels values of the ipsilateral and the contralateral brain were extracted from the regions within the red and blue circles respectively. (A) T2-weighted reference image, (B) AREX(rNOE), (C) AREX(APT), (D) T1 map, (E) T2 map.

(A) T1 and T2 fitting with iron content (R² for T1 = 0.9316, Y = -734.4*log X + 1769; R² for T2 = 0.9792, Y = -96.96*log X + 153.6). (B) AREX value with iron content. Two-way ANOVA was performed for statistical analysis as compare with the value of 0.15625mM. *P<0.05, **P<0.01, ***P<0.001. (C) A reference T2-weighted image of the phantom. Values were taken by averaging the pixels values of each tube.

Figure 5: Representative T₂-weighted images of a CPZ mouse pre- (W0), 5 weeks (W5) post- cuprizone exposure, and 4 additional weeks (W9) post cuprizone exposure termination. A strong gray-white matter contrast which was observed at baseline decreased at week 5 (red arrows) and increased back (blue arrows) after recovery (W9). Unlike in the corpus callosum, this change was not visually obvious in the cerebral cortex.

Figure 1: A reduction in OEF, and CMRO₂ in the cortex of Cuprizone (CPZ) mice but no change in CBF during demyelination. CTRL (black, n=9) and CPZ (blue, n=11) mice were imaged pre- (W0), 5 weeks (W5) post- cuprizone exposure, and 4 additional weeks (W9) post cuprizone challenge termination. Each symbol represents a different mouse (* - p ≤ 0.05, ** - p ≤ 0.01, *** - p ≤ 0.001).

Connectivity differences between GAERS and NEC at 4 months (3A) and 8 months (3B) mapped on the same-age community structure of GAERS. 56 statistically stronger connections (thick lines) were found in GAERS, while only 15 statistically stronger connections were found in NEC. NBS based on comparison of the 360 strongest connections at α=0.05. Only brain areas with significantly different connections are displayed in color.

Age-dependent connectivity differences in GAERS and NEC, mapped onto respective 4 months community structure. In GAERS (4A) 97 stronger connections (thick lines) were found at 4 months and 22 stronger connections were found at 8 months. In NEC (4B) 9 stronger connections (thick lines) and 45 stronger connections were found at 8 months. NBS based on comparison of the 360 strongest connections at α=0.05. Only brain areas with significantly different connections are displayed in color.

Fig 1. Group level pixel-wise analyses of differences between controls and hTau.P301S mice at 5 mths of age as exhibited by (A) Jacobian determinants showing local volumetric differences around inter-cerebellar fibres (white arrows), (B) FA maps depicting microstructural differences in inter-cerebellar fibres (yellow arrows) and brain stem (red arrow), (C) local volumetric changes as depicted by Jacobian determinants from 2.5 – 5 mths, (D) location of Region of Interest (ROI) analysed; in (A), (B) and (C) all coloured pixels statistically significant p<0.05, TFCE-corrected

Fig 2. ROI analyses of DTI metrices in inter-cerebellar fibres in controls and hTau.P301S mice. ROI is defined in Fig 1D. (A) Fractional Anisotropy (FA) values at 2.5 mths, (B) FA values at 5 mths of age, (C) Mean Diffusivity (MD) values at 2.5 mths, (D) MD values at 5 mths of age, (E) Radial Diffusivity (RD) values at 2.5 mths, (F) RD values at 5 mths of age, (G) Average Diffusivity (AD) values at 2.5 mths, and (H) AD values at 5 mths of age, data statistically significant at p<0.05, FDR-corrected

Figure 1. glycoNOE MRI on brain of the Lafora disease mouse model. (A) Axial and (B) sagittal slices. Data were acquired with an RF saturation power of B₁ =0.7 μT. ROIs are indicated in T_2w images by a red line. The Z-spectrum was fitted with Lorentzian and Gaussian hybrid line-shape, by assuming multi-Lorentzian functions for background (blue mark) and Voigt function for glycoNOE (shown in the bottom row). The glycoNOE map was estimated from the amplitude of the corresponding Voigt line-shape.

Figure 2. Histological images of brain for wild type (WT) and knock-out (KO, Epm2a-/-) mice which are stained with an antibody that detects Lafora bodies. Lafora bodies are abnormal accumulated in KO mouse rather than WT.

Figure 1. Averaged T1 maps of rhesus macaque brain in pre- and post-exposure days on DPE 3 (n=3) and DPE 5-7 (n=8), respectively. There were subtle T1 changes between pre-exposure (baseline, A) and early infection (DPE 3, C). The widespread significant T1 changes were observed in the late infection stage (DPE 5-7, D) compared to baseline (B) in the multiple regions, including the cerebellum, prefrontal, and orbital gyrus (white arrow and dotted circle).

Figure 2. A voxel-based statistical comparison of T1 of rhesus monkey brains between pre- and post-EBOV exposure. A. There were no significant changes in T1 values between baseline (pre-infection) and early infection (DPE 3). B. A significant increase in T1 values was found however in the late EBOV infection stage (DPE 5-7). Statistical significance was considered at corrected p < 0.005 (one-tailed paired t-test, FDR correction). The colorbar indicates T scores.

Figure 1. QSM group average of young (7-8 years) and old (15-20 years) macaque brains. Arrows point out substructures of the basal ganglia such as putamen and globus pallidus. An increase in the magnetic susceptibility of deep gray matter nuclei can be observed with age.

Figure 2. Group average of QSM and R₂^* map for the 4 marmoset age groups. A prominent increase in QSM and R₂^* contrast can be seen with increasing age particularly for deep grey matter nuclei. Arrows indicate substructures of the globus pallidus.

Figure 2: Dictionary Learning analysis of rs-fMRI data. a-h) Regions extracted from 10 components DL. i) top-right: Functional connectivity of extracted regions’ difference between WT and zQ175. Bottom-left: representation of significant (red) and non-significant (blue) differences. Red and green frames highlight somato-motor and Default Mode networks respectively.

Figure 3: Diffusion Tensor Imaging results after TBSS analysis. Green regions represent mean white matter skeleton as used in TBSS pipeline. Red-yellow zones show clusters of voxels with significant decrease of FA in zQ175 mice. Blue-light blue zones show clusters of voxels with increased RD in zQ175 mice (p<0.05).

Figure 1: The fiber Tracts pass through Thalamus (Top Row) and Hippocampus (Bottom Row) without parcellation. Fiber Tract Numbers (FTN) and corresponding Fractional Anisotropy (FA) with Control and CTCL groups are also shown. t-test * p0.05.

Figure 2: After Parcellation, Fiber Tract Numbers (FTN) and Fractional Anisotropy (FA) for (A) Hippocampus-Caudate.Putamen, (B) Hippocampus-Thalamus and (C) Thalamus-Caudate-Putamen are shown. t-test * p 0.05 and ** p 0.01.

Figure 1 (A) 6-OHDA is microinjected into the medial forebrain bundle (MFB) to unilaterally lesion dopamine ascending pathways. SNc, substantia nigra pars compacta. (B) Schema of experimental protocol. fMRI data were acquired for 50 min. For functional connectivity analysis, BOLD images were analyzed in two time-windows of 10 min, 10 min before and 30 min after a L-DOPA injection.

Figure 3 ROI-ROI matrices of correlation coefficients in (A) sham-operated mice and (B) 6-OHDA-lesioned mice before injection. Color bar, correlation coefficient. Thirty-six ROIs were classified in the cortex (including somatosensory cortex), subcortex (including striatum, GPi, GPe, SNR and SNC), and thalamus (including STN, MD and CM). (C) Differences between 6-OHDA-lesioned and sham mice before L-DOPA treatment. Left lower triangle panel shows t-values and right upper triangle panel shows significant difference (p < 0.05, network based statistic). Color bar, t-values.

Figure 2: Fixel-based analysis outcome for comparison zQ175 WT>HET. The statistical maps show significant differences (wildtype > heterozygous) for fiber density (FD) (top), fiber cross-section (FC) (middle) and the (fiber density x cross-section) (FDC) combination metric (bottom). Significant fixels (P_FWE<0.05) are mapped to a 200.000 streamline tractogram and coloured based on their percentage effect and superimposed on a fiber-orientation distribution (fod) template (n=9 WT, n=9 HET).

Figure 3: Region based analysis of fixel, tensor and kurtosis metrics. A) Overview of tracts of interest, selected following FBA. B) P-values corresponding to unpaired t-tests (wildtype vs heterozygous) for each metric in each region/tract of interest are displayed in the table. Green boxes indicate the statistical tests that survive FDR correction. Blue and red characters indicate a resp. lower or higher group average in zQ175 heterozygous compared to zQ175 wildtype animals.

(a) Orientational similarity (SIM) maps of different voxel sizes and field strengths. (b) Averaged SIM of single-fiber and crossing-fiber groups.

(a) Deviation angle (Angle_Dev) maps of different voxel sizes and field strengths. (b) Averaged Angle_Dev of single-fiber and crossing-fiber groups.

Figure 4. In vivo results. A&B. T₁ (A) and T₂ (B) maps of a 4- (top) and 6-year-old (bottom) monkey in the coronal, axial, and sagittal views. The red arrow indicates lower T₂ value in globus pallidus in the 6-year-old monkey.

Figure 3. Ex vivo results. A&B. T₁ (A) and T₂ (B) maps by IR-SE (T₁) or SE (T₂), MRF without B₁ correction, and MRF with B₁ correction. The red arrow indicates T₂ overestimation. C. Maps of B₁ factor of the conformal coil. D. Color-coded ROIs. E&F. Boxplots of T₁ (E) and T₂ (F) values in selected ROIs.

Figure 2. Adoptive transfer of CD8 T cells decreased brain bleeding and cleared virus. MR images of the turbinates (A), OB (B), and brain (C) from non-treated and CD8 T cell treated mice on 6 (upper panel) and 11-dpi (lower panel). (D-F) Quantification of hypointensity spots revealed that CD8 T cells transfer protected brain vessel integrity. (G) CD8 T cells reduced viral titers on 6-dpi. (H-I) IHC study revealed that CD8 T cells infiltrated and proliferated in the GL (H) and core (I) of the OB.

Figure 1. MRI detected microbleeds in turbinates, OB, and frontal brain since 4-dpi and monitored the vessel breakdown during VSV brain infection. MR images of the turbinates (A), OB (B, C), and brain (D, E) from normal mouse and VSV-infected mice on 4, 6, and 11-dpi. (D, E) The inserted figures were the enlarged views of the framed bleeding sites in the full views. Red arrow, bleeds. (F) Quantification of volume of hypointensity at the turbinates and numbers of hypointensity spots at the OB and brain. (G) Viral titers at the OB on 6 and 11-dpi. (H) IHC staining of the OB and brain.

Group comparison between CCI rat and healthy controls. The insula and nucleus accumbens showed significantly increased Alff in the CCI group (red area). The Alff was found significantly reduced in the motor cortex (blue area).

Immunohistochemical map of CCI rat. The dotted-line box and white arrows indicate the expression level of the p-Erk increase significantly in insula, S1 and NAc.

Figure 2. Left hemisphere principal long fascicles atlas of the chimpanzee brain. Views : Left- coronal, middle - sagital, right - frontal. Abbreviations : Sup. : Superior, Post : Posterior, Ant : Anterior

Figure 1. Steps allowing the obtaining of the long white matter bundles intra-subject clusters including a registration of the different anatomical and diffusion images to the template, a whole brain deterministic tractography, leading to an intra-subject fiber clustering.

Figure 2 - Single-subject axial slices of A) raw diffusion data B) Neurite Density Index (NDI), C) Orientation Dispersion Index (ODI), D) Isotropic Volume Fraction (IsoVF), E) Fractional Anisotropy (FA), F) Mean Diffusivity (MD), G) Axial Diffusivity H), Radial Diffusivity.

Figure 3 – Mean ± SEM metrics within the corpus callosum for injured and control subjects. Note, injury 1 took place at timepoint 1 and injury 2 took place at timepoint 4. For NDI and ODI, a repeated measures ANOVA was used to determine individual differences amongst the injured and control groups. Post-hoc pairwise comparisons (Bonferroni corrected) were then determined between individual timepoints within each group. Statistically significant (p < .05) pairwise differences are indicated by *.

Increased functional connectivity areas (red) in rTg4510 mouse. EC: Entorhinal cortex.

AT8 (1000x) immunohistochemical stain of 2.5 m/o rTg4510 (A: entorhinal cortex; B: frontal cortex) and littermate (C : Hippocampus; D: frontal cortex)

Fig. 5. Longitudinal evolution of CAPs for each group. A & B: 5 main CAPs based on the 2-week data for CTL & STZ group, respectively. C & D: longitudinal changes in 4 metrics of the 5 CAPs for CTL and STZ group, respectively. In CTL rats, no significant difference was detected with time. However, major longitudinal differences were found in the STZ group. CAP 1 & 2 occurred increasingly less and tended to become transit states with short duration, resilience and higher betweenness. Those CAPs cover mostly visual and somatosensory cortex, PPC, RSC and striatum.

Fig. 2. Intergroup comparison of CAPs at 2 weeks. A & B: 5 main CAPs based on CTL & STZ, respectively. C & D: intergroup differences in 4 metrics for CAPs in A & B, respectively. No significant differences were found when CTL was the reference (C) while the main CAPs 2, 3 & 4 identified from the STZ group (D) had significantly more occurrences and higher resilience than their counterpart in CTL. These CAPs mainly cover RSC, PPC, ACC, visual, motor and somatosensory cortex, thalamus, hypothalamus and striatum.

Figure 1. Three representative hold-out test scans. A) The transverse slice approximately through center of the lesion is shown representative scans, along with B) Ground truth and C) CNN-predicted ROIs, using the trained model indicated in Table 1, row 8. D) An overlay of the 2 ROIs is shown, with overlapping area indicated in yellow.

Figure 4. Per-animal predicted volumes and dice similarity coefficients. A) Distribution of DSC achieved during training (network in Figure 2, row 8) vs. volume (units are number of voxels) of ground truth ROIs. B) Predicted ROI volume vs. ground truth volume (units are number of voxels). Prediction is performed using trained network shown in Figure 2, row 8.

Figure 2: Structure of the network. Nonlocal position-aware block (NPAB) is deployed in the bottleneck of U-Net. The number of channels indicates below convolution blocks. The parameters are shared and reused in two training stages.

Figure 4: Comparison of different methods performance in same-domain(T2) and cross-domain(SWI and ASL).

Figure 1: Schematic depiction of two CARS (magenta) and axonal (cyan) signals (a) and representative axial and coronal Maximum Intensity Projection views of the healthy (b) and demyelinated (c) spinal cords with CARS (myelin, magenta) and Thy1–CFP fluorescence (axons, cyan) contrasts.

Figure 2: Rostro-caudal profiles of ihMTR, CARS signal and axonal density normalized to the mean of their values at +3 mm and -3 mm from the lesion, measured in (a) PBS-incubated spinal cord and (b) LPC-incubated spinal cord. (c) Dorso-ventral profile of absolute ihMTR in the triangle of WM dorsal tracts.

Figure 2: Demonstration of the respiratory correction in a full CEST spectrum collected in the spinal cord with no correction (blue) and implemented correction (orange) demonstrated.

Figure 1:

A: Raw sagittal spinal cord CEST image.

B: CEST spectrum of spinal cord with the interleaved non-saturated scans demonstrating the global respiratory effect.

Figure 2. Right frontal lobe anaplastic oligodendroglioma in a 51-year-old male (Grade III). The lesion in the right frontal lobe was iso-intense with perilesional edema on T2WI (A) and showed no enhancement on contrast-enhanced T1WI (B). The lesion showed an iso-intensity on diffusion-weighted image (C) and the focus of elevated arteriolar perfusion (arrowhead) on CBVa map (D). On review 1, the lesion was diagnosed as a low-grade glioma. On review 2, the lesion was diagnosed as a high-grade glioma.

Figure 1. Right frontal lobe anaplastic astrocytoma (Grade III) in a 39-year-old male. The lesion in the right frontal lobe was hyper-intense on T2WI (A) and showed no enhancement on contrast-enhanced T1WI (B). The lesion showed a slight hyper-intensity on diffusion-weighted image (C) and the focus of elevated arteriolar perfusion (arrowhead) on CBVa map (D). On review 1, the lesion was diagnosed as a low-grade glioma. On review 2, the lesion was diagnosed as a high-grade glioma.

Figure 2: ROI to ROI connectivity between the 3 groups : A- Less than 5 years B- 5 to 8 years C- More than 8 years.

Figure 1: Seed to seed connectivity between the 3 groups : A- Less than 5 years B- 5 to 8 years C- More than 8 years.

Figure 1. Representative DTI and NODDI parameter maps. The each side optic radiation was drawn on the FA map.

Figure 4. DTI and NODDI Parameters and Postoperative Visual Field Improvement. Ordinal logistic regression was performed using 3 ordinal outcomes (no improvement or worse after surgery, Δ VFIS = 0 ~ 4; mild improvement after surgery, Δ VFIS = -4 ~ -1; and marked improvement after surgery, Δ VFIS = -8 ~ -5).

Figure 1. Axial T2WI and T2 MWF maps of a 7-year old female HC (A) and two patients with MLD, 4 and 9-years old at (B, C respectively) with varying severity of leukodystrophy. Both patients have reduced MWF compared with HC in A. Although slight hyperintensity is noticed on T2WI; the patients have markedly low myelin on MWF.

Figure 2. Shows TBSS maps derived from DTI studies from the patients and HCs. Compared with HCs, FA map shows widespread significant reduction (A), MD and RD maps show increase (B and C ) and AD map shows a limited reduction in WM of the patients.

Two examples (a-d, meningioma, WHO grade II; e-h, malignant B-cell lymphoma) of visual comparison between DTI and NODDI-based CST reconstruction. NODDI-based tractography with ODI has better performance in the region of peritumoral edema and fewer false-positive tracts (c & g, white arrow). After the vasogenic edema subsided, the DTI-based tractography with FA ≤ 0.2 shows CST similar to the pre-surgery NODDI-based CST (d & h).

NODDI's volume fraction distributions in the DN group and N group are plotted with a 2D barycentric coordinate system. As compared with the DN group (a), the N group (b) distribution shows a left upper shift in the coordinate system. There is no significant difference in VF_ec (P = 0.733, d), but significant differences in VF_ic (P < 0.001, e) and VF_iso (P = 0.025, c).

Figure 3 Box and whisker plots of DWI, IVIM and DKI metrics in lower-grade gliomas stratified according to IDH1 genotype (IDH1 mutant and IDH1 wild type). Boxes represent the median ± quartiles, with whiskers extending to the maximum and minimum values.

Figure 4 Box and whisker plots of DWI, IVIM and DKI metrics in lower-grade gliomas stratified according to MGMT promoter status (methylated and unmethylated). Boxes represent the median ± quartiles, with whiskers extending to the maximum and minimum values.

Figure 1. NODDI maps of a patient with RI, together with DTI maps, T1WI, T2WI and enhanced images. RI, radiation-induced brain injury; T1WI, T1-weighted imaging; T2WI, T2-weighted imaging; FA, fractional anisotropy; MD, mean diffusivity; AD, axial diffusivity; RD, radial diffusivity; ICVF, intra-neurite volume fraction; ODI, orientation dispersion index; ISOVF, volume fraction of the isotropic compartment.

Figure 2. Differences of DTI parameters including FA (A), MD (B), RD (C) and MD (D) as well as NODDI parameters including ICVF (E), ISOVF (F) and ODI (G) of edema lesions between patients with or without cognitive decline in RI. RI, radiation- induced brain injury; NCF, normal cognitive function; CD, cognitive decline; FA, fractional anisotropy; MD, mean diffusivity. AD, axial diffusivity; RD, radial diffusivity; ICVF, intra-neurite volume fraction; ODI, orientation dispersion index; ISOVF, volume fraction of the isotropic compartment.

Nodes showing a significant negative effect (β<0, p<0.05 FDR-corrected) of Time on CI (A) and BC (B) are shown in blue (non-hubs) and orange (hubs). Nodes are weighted for the statistical significance (p-value). For image display purposes, concerning CI, labels are plotted for most significant nodes (threshold 0.03).

Nodes showing a significant negative effect (β<0, p<0.05 FDR-corrected) of Time × Tumor grade on Cl in LGG patients are represented in blue (non-hubs) and orange (hubs). Nodes are weighted for the statistical significance (p-value). For image display purposes, concerning Cl, labels are plotted for most significant nodes (threshold 0.03).

Figure 1. Axial Wave-SPACE-FLAIR (left), Wave-SWI (middle), and phase images (right) in a 28-year-old patient with confirmed MS demonstrating the presence of paramagnetic rims corresponding to lesions visible on FLAIR (highlighted). A lesion with a central vein sign can be seen in the same Wave-SWI sequence (arrowhead).

Figure 3. Axial Wave-SPACE-FLAIR (left), Wave-SWI (middle), and phase images (right) in a 20-year-old patient with confirmed anti-MOG disease with no visible paramagnetic rim around the lesion visible on FLAIR (highlighted). No lesions with a central vein sign were identified on the Wave-FLAIR and Wave-SWI sequences.

Figure 1. The PET/MR images of glioblastoma recurrence.

Figure 2. Correlation matrix of each parameter pair with significance levels for glioblastoma recurrence (right upper part) and radiotherapy injury (left lower part) groups.

Figure1.The susceptibility of primary motor cortex of ALS patients were significantly higher than those of controls

Figure 2.The susceptibility of SMA of ALS patients were significantly higher than those of controls

Figure 2. Tractography results in representative subjects and group differences. The fibers connecting bilateral occipital lobes in T2DM representative (A) were slender and more diffusely orientated compared with the healthy controls (B). ** indicates P <0.01, *** indicates P <0.001.

Figure 3. Correlations between insulin resistance and the diffusion measures of the occipital fibers. Note that no such correlations were observed in the control group. The first row displays the occipital fibers derived from the HCP1021-1mm template (https://pitt.app.box.com/v/HCP1021-1mm).

Figure 1. Glioblastoma with MGMT promoter methylation in a 39-year-old male. (a) Axial T2-FLAIR image. (b) Axial postcontrast T1-weighted image. (c) Axial CBVa map (ml/100 ml). (d) Histogram of CBVa from whole tumor ROI (dashed line). Images show a left thalamic mass with cystic change, which involved the lateral ventricular walls but spared the cortices. The CBVa map shows focal hyperperfusion (arrowhead) in the mass. The corresponding histogram shows a narrow distribution of perfusion in the whole tumor which is mainly concentrated in hypoperfused areas.

Figure 2. Glioblastoma without MGMT promoter methylation in a 44-year-old male. (a) Axial T2-FLAIR image. (b) Axial postcontrast T1-weighted image. (c) Axial CBVa map (ml/100 ml). (d) Histogram of CBVa from whole tumor ROI (dashed line). The images show an inhomogeneous mass in the right temporal lobe, which invaded the right basal ganglia, the cortices, and the subventricular zone. The corresponding histogram shows a wide distribution of perfusion in the whole tumor, with a peak located at around 1.0 ml/100 ml.

Figure 1. Maps of vein density and OEF for representative low grade (LGG), high grade glioma (HGG) cases and control subject.

Figure 2. Differential venous and OEF distribution based on AAL atlas. ^{# *}indicate significant interhemispheric difference of vein density(VD) and OEF in LGGs; ^{^ +} indicate significant interhemispheric difference of VD and OEF in HGGs, resepectively (paired t test, p<0.05, FDR corrected).

Figure 1. Sagittal (A; top row) and axial (A; bottom row) of population maps showing the CTT ROI used in this analysis. The lines in sagittal view at X=5 mm identify the location of the axial slices shown in the bottom row in MNI space. A three dimensional rendering of the CTT is shown in B.

CTT - central tegmental tract; ROI - region of interest; MNI - Montreal Neurological Institute

Figure 3. Correlations between CTT microstructure metrics and thalamus tau-PET SUVR in APOE-ε4 positive subjects (top row; A-D) and APOE-ε4 negative subjects (bottom row; E-H). Significant correlations between CTT microstructure and thalamus tau-PET SUVR are seen in the APOE-ε4 positive group but not in the APOE-ε4 negative group.

Schematic of the CJIVE process. Singular value decomposition (SVD) is applied to the input data matrices (far left) to obtain dimension-reduced PC scores. Next, a permutation test is used to determine the number of components that will lie in the joint subspace. CCA is applied to the PC scores and a weighted average of canonical variables determines the joint subspace. Finally, individual subspaces can be determined by the orthogonal complement of the joint subspace within the respective PC subspaces.

Joint Loadings for component 1. Temporal lobe ROI measures were most prominent among morphometry loadings, including both thickness and volume. ADAS and MMSE dominated cognition loadings (bottom) and both have been shown as excellent diagnostic tools. The proportions of total variation explained by the first joint component were 0.147 for morphometry and 0.467 for cognition. Proportions explained by the second component (not shown) were 0.025 and 0.044 for morphometry and cognition, respectively.

Figure 1: Schematic diagram of all the classification methods used. An artificial neural network (ANN) based classifier was implemented along with a Random forest (RF) of classification trees. Architecture for classification involving denoising autoencoders based on Heinsfeld et al has been shown.

Figure 2: Brain maps showing ROIs associated with each of the 12 networks used in ablation analysis

Figure 4. Correlation matrices between TSC versus local Tau load (A) and versus local Amyloid load (B). Notable correlations are reported at an uncorrected p < 0.05 level of significance and are framed in black.

Figure 3. (A) T-values resulting from voxel-based permutation analyses of our TSC maps (AD > Controls) and (B) atrophy clusters in green (Controls > AD), overlaid to our group template. (C) From the overlay of both TSC and atrophy clusters, one can appreciate the differences between both pathological patterns.

Figure 1: Perfusion averaged over all healthy controls and an exemplary dementia patient participant with semantic variant primary progressive aphasia (svPPA). White arrows indicate regional perfusion deficits.

Figure 3: Within-session and between-session coefficient of variation maps in healthy controls and FTD patients using absolute and relative perfusion.