Cerebral vasculature plays a critical role in brain function. Accurate characterization of its normative organization and distribution is essential for identifying vascular abnormalities. However, existing cerebrovascular templates do not adequately account for age-related anatomical variability and typically rely on structural-image-driven registration without explicitly leveraging vascular information for alignment. To address these limitations, we present a set of open, age-specific cerebrovascular templates constructed from time-of-flight magnetic resonance angiography (TOF-MRA) data of 1,288 healthy adults spanning the adult lifespan (19-92 years). Fine-scale vascular structures were automatically segmented using CereVessPro and registered through a novel two-stage framework that integrates age-specific structural alignment with vessel-guided refinement. The resulting resources comprise a series of openly accessible maps, including age-specific MRA templates, vessel probability maps, vessel density maps, and vessel radius maps from young to older adulthood. This framework is designed to improve vascular correspondence across subjects by explicitly addressing age-dependent anatomical differences and incorporating vascular features during alignment. Validation analyses revealed biologically plausible age-related changes in vascular morphology and vessel calibre, particularly in major cerebral arteries. Together, these data provide an age-resolved reference for cerebrovascular mapping, improving vascular alignment and supporting more reliable group-level analyses. This resource may facilitate more sensitive detection of vascular abnormalities in both research and clinical contexts.

Background: Over 50% of sickle cell anemia (SCA) patients will have silent or overt cerebral strokes, which have been linked to cognitive impairment. L-glutamine supplementation has been FDA approved to reduce vaso-occlusive crises frequency. However, glutamine’s effect on stroke and cognitive impairment in SCA is unknown. Measurement of neurometabolites using conventional MR spectroscopy has been limited to a single voxel due to technical challenges. A novel, high-resolution MR spectroscopic imaging (MRSI) technique, termed SPICEx, permits measurement of critical neural substrates and metabolites. In this pilot study, we assessed the neurometabolomic profile of patients with SCA compared to healthy controls using MRSI. Hypothesis: As a marker of stroke risk and brain health, we hypothesized that SCA patients have lower cerebral glutamine compared to controls. Methods: Adults with SCA and healthy controls without history of stroke underwent 3T brain MRI. SPICEx [RG1] produced 3D whole brain (WB) neurometabolite maps (2x3x3 mm 3 , acquisition time 12 minutes) including voxel-wise glutamine. MRI metrics of white matter (WM) microstructural disruption (increased mean diffusivity, MD) were also obtained. Gray matter (GM), WM and WB glutamine measures were compared between SCA vs. controls using Mann-Whitney U test. Spearman’s correlation assessed the relationship between cerebral glutamine, MD, and volume. Results: Of 8 SCA patients and 9 controls, we found lower WB, WM, and WM to WB glutamine ratio in SCA vs. controls (P=0.08, P=0.02, and P=0.01, respectively). Across the cohort, low GM glutamine was associated with reduced GM volume (ρ=0.5, p=0.04), and low WM glutamine was associated with elevated MD, (ρ=–0.5, p=0.04). Two SCA patients were prescribed 8-12 weeks of L-glutamine supplementation, and their WB glutamine increased compared to their first scan. Conclusion: Patients with SCA have reduced cerebral glutamine, likely due to increased energy expenditure from inflammation and erythropoiesis, leading to insufficient amino acid substrates for neural processing. Glutamine is a precursor for glutamate, GABA, and glutathione, the major antioxidant counteracting oxidative stress. The relative reduction of glutamine in the white matter may align with the propensity for silent infarcts occurring in the deep white matter. We are enrolling a larger cohort to investigate the role of L-glutamine supplementation to prevent stroke and cognitive impairment in SCA.

PURPOSE:To develop an effective deep learning (DL)-based method to denoise arterial spin labeling (ASL) data. METHODS:Conventional DL-based ASL denoising methods often suffer from overfitting and poor generalization when training data are limited. The proposed method overcame this problem using two strategies: (i) perform data augmentation to create large training data and (ii) denoise in-distribution and out-of-distribution components of the target perfusion-weighted image separately. Specifically, Image-to-Image Schrödinger Bridge (I2SB)-based distribution remapping transforms were applied to the large public ASL datasets so that their intensity distribution matched that of the data to be denoised. U-Net-based DL denoisers were trained on the remapped data to capture in-distribution features. High-SNR outputs from the DL-denoiser were incorporated into a Bayesian model to reconstruct the out-of-distribution features with sparsity constraints, generating denoised images for cerebral blood flow (CBF) quantification. RESULTS:Simulation studies highlighted the importance of distribution remapping for effective data augmentation in limited-data scenarios. Both simulation and in vivo experiments showed that the proposed method outperformed state-of-the-art approaches, achieving an average SNR improvement of approximately 7 dB. Evaluations on multiple datasets confirmed robust and generalizable performance across different ASL sequences and imaging protocols. To demonstrate clinical potential, our method was applied to denoising stroke patient data (using only one-sixth of total averages with ˜83% reduction in scan time) and produced comparable CBF maps to the conventional ASL method. CONCLUSION:The proposed method enables effective ASL denoising with limited training data. It has the potential to accelerate ASL acquisition, enhance image quality, and improve clinical utility.

Objective: To develop an effective method for phase correction of magnetic resonance spectroscopic imaging (MRSI) data. Methods: In many MRSI applications, it is desirable to generate absorption-mode spectra, which requires correction of phase errors in the measured MRSI data. Conventional phase correction methods are sensitive to measurement noise and baseline distortion, often resulting in distorted absorption-mode spectra from MRSI data with low-SNR and long acquisition dead time. This paper proposed a novel model-based method for improved phase correction of MRSI data. The proposed method determined the zeroth-order phase and acquisition dead time using a Lorentzian-based spectral model and performed signal extrapolation using a generalized series model. Absorption-mode spectra were then generated from the phase-corrected and extrapolated MRSI data. Results: The proposed method was evaluated using both simulated data and experimental data acquired from human subjects in multi-nuclei (31P, 2H, and 1H) MRSI experiments. Simulation results demonstrated improved parameter estimation accuracy by the proposed method under various noise levels and dead times. The proposed method also consistently generated high-quality absorption-mode spectra with minimal spectral distortions from experimental data. The proposed method was compared with state-of-the-art methods (including the entropy method and LCModel method) and showed more robust phase correction performance with less spectral distortions. Conclusion: This paper introduced a novel method for phase correction of MRSI data. Results from simulated and in vivo data demonstrated that high-quality absorption-mode spectra could be obtained using the proposed method. Significance: This method will provide a useful tool for processing MRSI data.

Unsupervised brain lesion segmentation, focusing on learning normative distributions from images of healthy subjects, are less dependent on lesion-labeled data, thus exhibiting better generalization capabilities. A fundamental challenge in learning normative distributions of images lies in the high dimensionality if image pixels are treated as correlated random variables to capture spatial dependence. In this study, we proposed a subspace-based deep generative model to learn the posterior normal distributions. Specifically, we used probabilistic subspace models to capture spatial-intensity distributions and spatial-structure distributions of brain images from healthy subjects. These models captured prior spatial-intensity and spatial-structure variations effectively by treating the subspace coefficients as random variables with basis functions being the eigen-images and eigen-density functions learned from the training data. These prior distributions were then converted to posterior distributions, including both the posterior normal and posterior lesion distributions for a given image using the subspace-based generative model and subspace-assisted Bayesian analysis, respectively. Finally, an unsupervised fusion classifier was used to combine the posterior and likelihood features for lesion segmentation. The proposed method has been evaluated on simulated and real lesion data, including tumor, multiple sclerosis, and stroke, demonstrating superior segmentation accuracy and robustness over the state-of-the-art methods. Our proposed method holds promise for enhancing unsupervised brain lesion delineation in clinical applications.

Magnetic resonance imaging (MRI) has revolutionized diagnostic radiology and medicine over the past five decades1,2. However, clinical applications of MRI are still mainly limited to visual examination of macroscopic tissue pathology3,4. Because diseases, such as tumours, multiple sclerosis (MS) and neurodegenerative disorders, are highly heterogeneous, there is a critical need for a non-invasive imaging technology that can provide quantitative biomarkers for tissue characterization for personalized and precision medicine5. Here we introduce a new approach to MRI data acquisition and processing, called 'multiplexed MRI' (MRx), to achieve high-resolution simultaneous multiparametric mapping of several molecules. We demonstrate that MRx can obtain a large set of quantitative structural, physiological and molecular biomarkers of the whole brain in standard clinical settings. We further demonstrate that these biomarkers could define an effective tissue state index for disease subtyping and lesion characterization in tumours and MS. We anticipate that the new quantitative multiplexed imaging capabilities of MRx would substantially enhance the capability of MRI for diagnosis, monitoring and assessment of therapeutic efficacy of many neurological diseases and potentially transform brain imaging for both research and clinical applications.

INTRODUCTION:Microglial activation can either support neuronal function or exacerbate damage, contributing to Alzheimer's disease (AD) progression. We investigated spatial relationships among microglial activation, neuronal health, and amyloid beta (Aβ) in the AD spectrum. METHODS:Forty healthy controls, 37 patients with mild cognitive impairment (MCI), and 62 patients with AD underwent whole-brain high-resolution 1H-magnetic resonance spectroscopic imaging (MRSI), [1 8F]DPA-714, and [1 8F]AV-45 positron emission tomography (PET). Regional and voxel-wise analyses assessed changes and associations of microglial activation with N-acetylaspartate (NAA) and Aβ. RESULTS:MCI and AD patients showed higher microglial activation and lower NAA, correlating with cognitive decline. In controls and MCI, microglial activation correlated positively with NAA and Aβ in early amyloid-accumulating regions. Conversely, negative correlations with NAA emerged in the hippocampus in MCI and extended to temporal and occipital regions in AD. DISCUSSION:For the first time, we identified two distinct spatial association patterns between [1 8F]DPA-714 PET and NAA, shedding light on the complex interplay between neuroinflammation and neuronal health in AD.

Pseudo-healthy image inpainting is an essential preprocessing step for analyzing pathological brain MRI scans. Most current inpainting methods favor slice-wise 2D models for their high in-plane fidelity, but their independence across slices produces discontinuities in the volume. Fully 3D models alleviate this issue, but their high model capacity demands extensive training data for reliable, high-fidelity synthesis-often impractical in medical settings. We address these limitations with a hierarchical diffusion framework by replacing direct 3D modeling with two perpendicular coarse-to-fine 2D stages. An axial diffusion model first yields a coarse, globally consistent inpainting; a coronal diffusion model then refines anatomical details. By combining perpendicular spatial views with adaptive resampling, our method balances data efficiency and volumetric consistency. Our experiments show our approach outperforms state-of-the-art baselines in both realism and volumetric consistency, making it a promising solution for pseudo-healthy image inpainting. Code is available at https://github.com/dou0000/3dMRIConsistent-Inpaint.

In volume-to-volume translations in medical images, existing models often struggle to capture the inherent volumetric distribution using 3D voxelspace representations, due to high computational dataset demands. We present Score-Fusion, a novel volumetric translation model that effectively learns 3D representations by ensembling perpendicularly trained 2D diffusion models in score function space. By carefully initializing our model to start with an average of 2D models as in TPDM, we reduce 3D training to a fine-tuning process and thereby mitigate both computational and data demands. Furthermore, we explicitly design the 3D model's hierarchical layers to learn ensembles of 2D features, further enhancing efficiency and performance. Moreover, Score-Fusion naturally extends to multi-modality settings, by fusing diffusion models conditioned on different inputs for flexible, accurate integration. We demonstrate that 3D representation is essential for better performance in downstream recognition tasks, such as tumor segmentation, where most segmentation models are based on 3D representation. Extensive experiments demonstrate that Score-Fusion achieves superior accuracy and volumetric fidelity in 3D medical image super-resolution and modality translation. Beyond these improvements, our work also provides broader insight into learning-based approaches for score function fusion.

INTRODUCTION:Understanding the neurometabolic changes associated with amyloid-β (Aβ) deposition is important for early Alzheimer's disease (AD) diagnosis, but their spatial relationships remained unexplored due to technical limitations. METHODS:We investigated the relationship between Aβ deposition and neuronal and glial metabolites using high-resolution 3D magnetic resonance spectroscopic imaging (MRSI) (8-min scan, 2 × 3 × 3 mm3 resolution) and Aβ-positron emission tomography (Aβ-PET) imaging. N-acetylaspartate, myo-inositol, and creatine maps were obtained from 174 participants: 39 controls, 65 mild cognitive impairment (MCI), and 70 AD patients. RESULTS:N-Acetylaspartate levels were negatively correlated with Aβ, while myo-inositol levels were positively correlated globally. Regional associations with Aβ include N-acetylaspartate reductions in frontal cortex, anterior cingulate cortex, and precuneus, and myo-inositol increases in precuneus, lateral temporal, and lateral parietal cortices. Combined MRSI and PET biomarkers achieved the highest diagnostic accuracy for MCI and AD . DISCUSSION:Hybrid high-resolution 3D MRSI and Aβ-PET imaging provides valuable insights into Aβ's impact on neurometabolic changes, improving early AD diagnosis. HIGHLIGHTS:Hybrid 3D magnetic resonance spectroscopic imaging-positron emission tomography (MRSI-PET) imaging reveals Aβ deposition impact on neurometabolism in Alzheimer's disease (AD). N-acetylaspartate (NAA) as a neuronal metabolic marker is negatively associated with Aβ globally and locally. Myo-inositol (mI) as a glial metabolic marker is positively associated with Aβ globally and locally. Combining 3D magnetic resonance spectroscopic imaging (MRSI) and PET biomarkers improves diagnostic accuracy for mild cognitive impairment (MCI) and AD.

Duchenne muscular dystrophy (DMD) is a genetic disorder characterized by progressive muscle degeneration. Impaired muscle metabolism has been implicated in DMD, and interventions aimed at normalizing and enhancing metabolic function, such as exercise, have been proposed as potential therapeutic strategies. However, the metabolic response to exercise in DMD remains incompletely understood. This study aimed to investigate the acute metabolic response to muscle stimulation mimicking high-intensity exercise. Using phosphorus-31 magnetic resonance spectroscopy (31P-MRS) and fingerprinting (31P-MRSF), changes in phosphocreatine (PCr) recovery rate and creatine kinase (CK) shuttle activity following repeated muscle stimulation were quantified in mdx mice, a well-established mouse model of DMD. The impact of muscle degeneration and aging was assessed by comparing mdx and control mice in two age groups: young (10-12 wk) and adult (20-22 wk). Young control mice exhibited a significant increase in PCr recovery rate and CK rate constant following muscle stimulation, indicating a positive acute metabolic adaptation. In contrast, mdx mice showed no increase in PCr recovery rate and an attenuated increase in CK rate constant, suggesting compromised mitochondrial function and CK shuttle efficiency. These findings indicate that muscle degeneration in DMD impairs acute metabolic response to high-intensity muscle contractions, potentially limiting exercise-induced metabolic benefits. Furthermore, this study demonstrates the utility of 31P-MRS and 31P-MRSF as noninvasive tools to assess muscle metabolism and exercise response in vivo.NEW & NOTEWORTHY The results of the current study demonstrate the impact of muscle degeneration and aging on exercise-induced metabolic response. The absence of an increase in phosphocreatine recovery rate and the attenuated increase in creatine kinase shuttle activity observed in mdx mice suggest impaired acute metabolic adaptation to high-intensity exercise. These findings underscore the potential of phosphorus-31 magnetic resonance spectroscopy and fingerprinting for noninvasive assessment of muscle bioenergetics in vivo.

Multimodal image translation has found useful applications in solving several medical imaging problems. In this paper, we presented a systematic analysis of multimodal images and machine learning-based image translation from an information-theoretic perspective. Specifically, we analyzed the amount of mutual information that exists in some commonly used multimodal images. This analysis revealed varying structural correlation across modalities and tissue-dependence of mutual information. We also analyzed the amount of information transferred and gained in multimodal image translation and provided an upper bound on the information gain. Information-theoretic measures were also proposed to assess the effectiveness of an image translator, and the uncertainty associated with image translation. Numerical results were presented to demonstrate the information gain in practical multimodal image translation, and to validate the proposed upper bound on information gain and the translation error predictor. Finally, several potential applications of our analysis results were discussed, including the image denoising and reconstruction using side information generated by image translation. The findings from this study may prove useful for guiding the further development and application of multimodal image translation.

Deuterium (2H) magnetic resonance spectroscopic imaging (DMRSI) is a newly developed technology for assessing glucose metabolism by simultaneously measuring deuterium-labeled glucose and its downstream metabolites (1) and has a potential to provide a powerful neurometabolic imaging tool for quantitative studies of cerebral glucose metabolism involving multiple metabolic pathways in the human brain. In this work, we developed a dynamic DMRSI method that combines advanced radiofrequency coil and postprocessing techniques to substantially improve the imaging signal-to-noise ratio for detecting deuterated metabolites and enable robust dynamic DMRSI of the human brain at 7 T with very high resolution (HR; 0.7 cc nominal voxel and 2.5 min/image) and whole-brain coverage. Utilizing this capability, we were able to map and differentiate metabolite contents and dynamics throughout the human brain following oral administration of deuterated glucose. Furthermore, by introducing a sophisticated kinetic model, we demonstrated that three key cerebral metabolic rates of glucose consumption (CMRGlc), lactate production (CMRLac), and tricarboxylic acid (TCA) cycle (V TCA), as well as the maximum apparent rate of forward glucose transport (T max) can be simultaneously imaged in the human brain through a single dynamic DMRSI measurement. The results clearly show that the glucose transport, neurotransmitter turnover, CMRGlc, and V TCA are significantly higher in gray matter than in white matter in the human brain; and the mean metabolic rates and their ratios measured in this study are consistent with the values reported in the literature. The HR dynamic DMRSI methodology presented herein is of great significance and value for the quantitative assessment of human brain glucose metabolism, aerobic glycolysis, and metabolic reprogramming under physiopathological conditions.

Iron accumulation and ferroptosis occur in the brain following ischemic stroke. However, the relationship between iron overload and cell type-specific fates remains largely unclear. Here, iron deposition and neuronal loss are reported within the perilesional cortex of three patients with ischemic stroke at both acute and subacute stages. It is identified that ischemia/reperfusion-induced iron overload triggers ferroptosis predominantly in neurons and to a lesser extent in astrocytes, whereas most astrocytes undergo reactive proliferation. Mechanistically, the reduced or elevated Nrf2/GPX4 and SLC7A11 levels in neurons or astrocytes, respectively, account for these distinct iron overload-induced cellular fates. Moreover, iron overload promotes astrogliosis by enhancing the transcriptional activities of several proliferation-related genes. Using mice with partial knockout of the transferrin receptor 1 (TfR1) gene Tfrc, astrocyte-specific Tfrc knockdown, and conditional astrocytic Cpt1a partial knockout (to induce fatty acid metabolism disorders), it is revealed that increased TfR1 palmitoylation and clathrin-mediated endocytosis drive astrocytic iron overload. Notably, ischemia/reperfusion-induced elevation of palmitic acid is associated with enhanced TfR1 palmitoylation. Treatment with antioxidants or iron chelators mitigates ischemic brain injury. Together, these findings provide a comprehensive framework linking ischemia/reperfusion-induced iron overload to cell type-specific fates. TfR1 palmitoylation emerges as a potential target for ischemic stroke therapy.

In volume-to-volume translations in medical images, existing models often struggle to capture the inherent volumetric distribution using 3D voxel-space representations, due to high computational dataset demands. We present Score-Fusion, a novel volumetric translation model that effectively learns 3D representations by ensembling perpendicularly trained 2D diffusion models in score function space. By carefully initializing our model to start with an average of 2D models as in existing models, we reduce 3D training to a fine-tuning process, mitigating computational and data demands. Furthermore, we explicitly design the 3D model's hierarchical layers to learn ensembles of 2D features, further enhancing efficiency and performance. Moreover, Score-Fusion naturally extends to multi-modality settings by fusing diffusion models conditioned on different inputs for flexible, accurate integration. We demonstrate that 3D representation is essential for better performance in downstream recognition tasks, such as tumor segmentation, where most segmentation models are based on 3D representation. Extensive experiments demonstrate that Score-Fusion achieves superior accuracy and volumetric fidelity in 3D medical image super-resolution and modality translation. Additionally, we extend Score-Fusion to video super-resolution by integrating 2D diffusion models on time-space slices with a spatial-temporal video diffusion backbone, highlighting its potential for general-purpose volume translation and providing broader insight into learning-based approaches for score function fusion.

PURPOSE:To develop an effective method for correcting head motion in high-resolution, non-water-suppressed MRSI. METHODS:MRSI scans are susceptible to subject motion due to the long data acquisition time required for sufficient spatial-spectral encodings. The problem is more serious in non-water-suppressed MRSI experiments since motion artifacts in the water and lipid signals make their removal even more challenging. To address this problem, we propose a novel motion correction method that detects and discards motion-corrupted k-space data using TR-wise linear navigators. The discarded data are replaced with reconstructed data obtained by reformulating motion correction as a missing data reconstruction problem using sensitivity encodings. Extrinsic priors for water and lipid signals and intrinsic priors for metabolite signals are incorporated in the motion correction process. We evaluated its performance on a spectroscopic phantom and three human groups: (1) five healthy adults performing three different voluntary motion patterns; (2) a healthy child; and (3) a cerebral hemorrhage patient with involuntary movements. RESULTS:In phantom and healthy subjects, the proposed method produced high-quality water images and metabolite maps that closely matched motion-free references, with 10%-20% quantitative gains in image/map quality (higher PSNR/SSIM, lower NRMSE). In the child and patient data, motion artifacts were noticeably reduced, with 20%-30% reductions in NAA linewidth and fitting error in the child, and ˜50% reductions in coefficient-of-variation in the patient. CONCLUSION:An effective motion-correction technique has been developed for high-resolution non-water-suppressed MRSI. This method has the potential to significantly enhance the robustness and clinical applicability of non-water-suppressed MRSI.

Magnetic resonance spectroscopic imaging has potential for non-invasive metabolic imaging of the human brain. Here we report a method that overcomes several long-standing technical barriers associated with clinical magnetic resonance spectroscopic imaging, including long data acquisition times, limited spatial coverage and poor spatial resolution. Our method achieves ultrafast data acquisition using an efficient approach to encode spatial, spectral and J-coupling information of multiple molecules. Physics-informed machine learning is synergistically integrated in data processing to enable reconstruction of high-quality molecular maps. We validated the proposed method through phantom experiments. We obtained high-resolution molecular maps from healthy participants, revealing metabolic heterogeneities in different brain regions. We also obtained high-resolution whole-brain molecular maps in regular clinical settings, revealing metabolic alterations in tumours and multiple sclerosis. This method has the potential to transform clinical metabolic imaging and provide a long-desired capability for non-invasive label-free metabolic imaging of brain function and diseases for both research and clinical applications. Ultrafast magnetic resonance spectroscopic imaging enables non-invasive high-resolution metabolic imaging of the whole human brain under healthy and diseased conditions.