This article reviews methods that use radiofrequency (RF) magnetic field gradients to encode magnetic resonance imaging (MRI), offering an alternative to the static magnetic field gradients conventionally used to form images. Since RF encoding was first demonstrated in the 1970s, several approaches have been developed, with potential advantages for both MRI and magnetic resonance spectroscopy (MRS), which measures the chemical composition of tissues. However, challenges remain in hardware design, integration of these techniques into imaging systems, and their use in living subjects. This review summarizes major RF encoding methods, their potential benefits and limitations, and future directions for their technical development, research use, and clinical application.

Fig. 1.��PubMed count of journal articles containing keywords relating to RF encoding (1979–2026). Year of reference to seminal work of surveyed methods (i.e. RFI , TRASE , Bloch–Siegert , SENF , FREE ) are indicated. See search terms at .

Multiple sclerosis (MS) lesions can interfere with automated MRI analysis, while limited labeled data make it difficult to train models that accurately detect these lesions. We developed MSRepaint, an AI-based imaging method that can both fill lesion-affected areas with realistic healthy-appearing tissue and generate synthetic lesions for training purposes. The method preserves surrounding brain anatomy and can work when some MRI inputs are unavailable. In testing, MSRepaint outperformed established lesion-filling methods and achieved accuracy comparable to a recent AI-based approach while processing images more than 20 times faster. Models trained using its synthetic lesions also performed better across multiple lesion-detection datasets than models trained with other synthetic or real training data. MSRepaint therefore provides a fast, flexible approach for improving MRI analysis and generating realistic training data for MS lesion detection.

Fig. 1.��Overview of MSRepaint training and inference. (a) Training data: A 2D multicontrast image 𝑥0 undergoes forward diffusion, generating noisy images 𝑥𝑡−1, 𝑥𝑡, and the final degraded image 𝑥𝑇, paired with its lesion mask 𝑀. (b) Training: The denoiser 𝜙𝑤 is trained to estimate ˆ𝑥𝑡−1 from 𝑥𝑡 with guidance from mask 𝑀. (c) Inference: The trained denoiser 𝜙𝑤∗ generates outputs guided by 𝑀target, while the repaint mask 𝑀repaint restricts updates to its region, preserving surrounding tissue. A repainting mechanism iteratively refines update boundaries to ensure seamless integration with the surrounding tissue. (d) Example tasks: Lesion filling sets 𝑀target=0 to replace lesions with healthy tissue, while lesion synthesis sets 𝑀target=𝑀repaint to generate new lesion patterns.

Wireless resonators can improve MRI performance, but they must safely deactivate, or “detune,” when exposed to the scanner’s transmitted radiofrequency power. We developed a simple method to determine the power level at which this detuning becomes reliable. Using MRI scans of test objects, the method compares signals near and farther from the resonator across different transmit-power levels to identify the detuning threshold. Testing with several resonator designs, MRI field strengths (3 and 7T), and scanners from different manufacturers produced consistent results. This approach provides a practical tool for testing and monitoring the safety of wireless resonators used in clinical and research MRI.

FIGURE 1

Passive detuning of wireless resonators. (A) Illustration of the induced current in the wireless resonator during RF transmission. (B) Crossed diodes in parallel with a distributed capacitor (capacitor shorted when diodes conduct). (C) Crossed diodes with a series inductor in parallel with a distributed capacitor (capacitor effectively open-circuited when diodes conduct). (D) Coaxial coil implementation with diodes placed across the gap of the open ends. The ellipses indicate the presence of multiple distributed capacitors along the conductor.

Carotid artery plaques can increase the risk of stroke, and understanding their structure and mechanical properties may help assess their stability. We developed a reproducible technique that combines ultrasound with magnetic resonance imaging (MRI) to evaluate carotid artery walls and plaques. The method aligns ultrasound and MRI images using anatomical landmarks, allowing the same areas of the artery to be located across both imaging methods. This alignment also guides the collection of additional ultrasound measurements, including blood flow and tissue stiffness data, from the same anatomical location. The technique improved consistency between MRI and ultrasound imaging and provides a practical framework for standardized, quantitative assessment of carotid plaques and artery wall properties.

Figure 1

Example placement of points used for alignment in the axial plane for a representative subject. Points are placed at the wall regions of the ICA and ECA as well as the bifurcation region for both the MRI image (A) and ultrasound image (B).

Prenatal alcohol exposure (PAE) can affect how the brain develops, but its effects on the shape and structure of the brain’s outer layer, or cortex, are not fully understood. Using MRI scans from 121 people aged 7–21 years, including 56 with PAE, we examined several features of cortical structure and how they changed with age. Youth with PAE showed differences in brain folding, groove depth, and structural complexity in specific brain regions compared with unexposed youth. Some of these features also developed differently with age in the PAE group. Overall, the findings suggest that measures of brain folding, particularly gyrification, may be more sensitive to the effects of prenatal alcohol exposure than commonly studied measures such as cortical thickness.

Fig. 1. A: Distribution of ages for participants with PAE (blue) or unexposed (salmon). MRI scans are represented by dots with horizontal lines connecting scans of the same participant. There are 163 scans (73 PAE) from 123 participants (58 PAE). 84 participants had 1 scan, 38 participants had 2, and 1 participant had 3 scans. Figure was created using ggplot2 in R (). B: Flowchart depicting the number of scans (and participants) at each stage of data acquisition and processing. Note that there were no significant group differences in the quality of T1-weighted data or MaCRUISE segmentations.

Learning through language requires the brain to rapidly transform words into meaningful concepts, but how this process unfolds over time remains unclear. Using a combination of functional MRI and electroencephalography (EEG), we tracked brain activity in healthy adults as they processed language. We identified a sequence of brain networks that became active within about 700 milliseconds, progressing from visual word recognition to retrieving meaning, interpreting context, integrating sentence structure and meaning, and ultimately forming a coherent concept. Transitions between these stages involved shared brain regions, particularly the left temporoparietal junction. People with stronger language comprehension showed greater involvement of later processes related to interpreting and integrating information. These findings reveal how different brain networks work together over fractions of a second to transform language into conceptual knowledge.

Figure 1

Spatial and temporal perspectives on language processing. Current neuroimaging approaches examine language processing through separate methodological silos, limiting our understanding of how distributed brain networks coordinate dynamically in real time. a) Spatial perspective: brain systems supporting language comprehension, after Fedorenko et al. (2024), showing the distributed architecture of perception of linguistic input (green), core language processing (red), motor planning (navy), and higher-order cognition (purple). Regions are rendered on an inflated left-hemisphere cortical surface from probabilistic functional atlases for the language, multiple demand (light purple), and default mode (dark purple) networks (Lipkin et al. 2022; Yeo et al. 2011), and Harvard-Oxford for the perception and motor areas. While fMRI reveals the spatial organization of these networks, it lacks the temporal resolution to capture their millisecond-scale coordination. b) Temporal perspective: a schematic ERP waveform illustrating the N400 (semantic processing, ~400 ms) and P600 (syntactic/integrative processing, ~600 ms) components that reflect the rapid timecourse of language comprehension. c) Topographical maps illustrating the spatial distribution of ERP effects, reproduced from Schacht et al. (2014, Fig 1B and 1C): the main effect of semantics in the N400 window (top) and the grammaticality-violation effect in the P600 window (bottom). Licensed under CC BY; cropped from the original. While EEG provides precise temporal information about language processing stages, it offers limited spatial resolution for localizing these processes within the distributed brain networks shown in (a). This methodological divide has hindered our understanding of how spatially distributed language networks achieve their rapid, coordinated function during real-world comprehension.

]]> /valiant/2026/08/24/a-comparative-evaluation-of-multiple-enlarged-perivascular-space-segmentation-tools/ Mon, 24 Aug 2026 19:28:42 +0000 /valiant/?p=7380 LeFevre, James D.; Robb, W. Hudson; Liu, Dandan; Jackson, T. Bryan; Pechman, Kimberly R.; Shashikumar, Niranjana; Vyas, Yukti; Landman, Bennett A.; Davis, L. Taylor; Hohman, Timothy J.; Jefferson, Angela L. (2026). . Magnetic Resonance Imaging, 134, 110749.

Enlarged perivascular spaces (ePVS) are fluid-filled spaces around small blood vessels in the brain that can become more visible with aging and small vessel disease and may reflect reduced clearance of waste from the brain. Measuring ePVS manually on MRI scans is time-consuming and impractical for large studies. To address this, researchers developed DORES, a deep learning tool that automatically identifies and measures ePVS using two types of brain MRI images. The model was developed using data from the Vanderbilt Memory and Aging Project and evaluated against expert manual measurements and three other automated tools. DORES showed good performance in identifying ePVS in both white matter and the basal ganglia, a group of structures deep within the brain, and its estimates of ePVS number and volume agreed well with expert measurements. Testing on an independent Alzheimer’s disease imaging dataset showed somewhat lower performance, as was also observed with the other automated methods. Results also varied depending on the type of MRI scanner used, suggesting that scanner differences can affect measurement consistency. Overall, DORES provides a promising automated approach for measuring ePVS in older adults, although scanner-related differences should be considered when applying the method across multiple research sites.

Fig. 1.��Representative White Matter ePVS Segmentations of DORES Performance in VMAP.

T₁-weighted axial scans were selected to illustrate DORES performance at the 25th (top row; Dice = 0.55), 50th (middle row; Dice = 0.63), and 75th (bottom row; Dice = 0.73) percentiles of white matter regional Dice scores. The first column displays the raw, skull-stripped images in the white matter. The second column displays the corresponding segmentation overlays. Manual segmentations are shown in blue, whereas DORES predictions are shown in red. Voxels where manual and DORES segmentations overlap appear purple. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

]]> /valiant/2026/08/24/diffusion-mri-with-a-multi-shot-rosette-readout/ Mon, 24 Aug 2026 19:23:05 +0000 /valiant/?p=7361 Harkins, Kevin D.; Lee, Tzu-Wei; Martin, Jonathan B.; Alderson, Hannah E.; Gore, John C.; Does, Mark D. (2026). . Journal of Magnetic Resonance, 391, 108132.

Diffusion-weighted imaging (DWI) is an MRI technique that measures the movement of water in tissue, but its image quality can be affected by magnetic field variations, electrical currents generated during scanning, and differences between repeated measurements. This study tested whether a spiral-like scanning pattern called a rosette readout trajectory could reduce these common artifacts. Multi-shot rosette DWI, which collects an image over several measurements, was evaluated in test objects and an isolated ferret brain at 7 Tesla, as well as in the spinal cords of living mice at 15.2 Tesla. The rosette approach produced consistently high-quality images at both magnetic field strengths. Correcting differences between repeated measurements also prevented unwanted loss of the diffusion-weighted signal. Overall, multi-shot rosette DWI effectively measured and corrected several important sources of MRI artifacts, demonstrating its potential for producing more reliable, high-quality diffusion images.

Fig. 1.��A rosette readout can be divided into multiple shots, here comprised of all 5 leafs of a single rosette acquisition window, or multiple segments. Each segment consists of portions from all shots that cross the center of k-space at the same time in the acquisition window. Images reconstructed from individual segments can be used to estimate the background static  field, while images reconstructed from individual shots can be used to estimate shot-to-shot phase variations. This example 5 leaf rosette contains 6 segments, as the first inout segment and the last outin segment are considered individually.

MRI scanning is often limited for patients with deep brain stimulation (DBS) implants because radiofrequency (RF) energy used during MRI can heat the implanted leads and potentially cause tissue injury. This study developed and tested a six-channel head coil designed to control the distribution of the MRI electric field and reduce heating near DBS implants. Using computer-based electromagnetic simulations, the researchers adjusted individual coil elements to reshape the transmitted RF field. The approach was tested first with a simple conductive wire and then with four realistic DBS lead models. The coil could be optimized either to reduce the electric field at a specific location, such as the tip of a DBS lead, or to reduce the maximum specific absorption rate (SAR), a measure of RF energy absorbed by tissue, across the entire head. Depending on the optimization method, the electric field at a targeted location was reduced by 45.9% to 68.3%, while whole-head SAR was reduced by an average of 72.09% across the four DBS models. These findings demonstrate that the proposed coil can flexibly reshape the MRI transmit field and substantially reduce simulated RF-related heating around DBS implants, providing a promising approach for improving MRI safety in patients with these devices.

FIGURE 1.��

[A] Simulation model of a 3 T body transmit (Tx) coil combined with a decoupled six-channel wireless head coil array (blue). [B] Zoomed-in view of the 6-channel wireless resonator array with the birdcage coil model hidden. The image combines the human head and the simplified DBS lead model. The wireless resonator conductors are shown in blue, and the DBS electrode is shown in red. [C] Combined human head and DBS lead model. All DBS lead positions are shown together for illustration purposes; however, in the simulations, each DBS configuration was evaluated independently.

]]> /valiant/2026/08/24/is-correction-for-gradient-nonlinearity-necessary-in-a-brain-diffusion-tensor-mri-clinical-study/ Mon, 24 Aug 2026 18:54:20 +0000 /valiant/?p=7352 Kanakaraj, Praitayini; Yao, Tianyuan; Li, Zhiyuan; Newlin, Nancy R.; Kim, Michael E.; Gao, Chenyu; Yu, Tian; Krishnan, Aravind; Rogers, Baxter P.; Hohman, Tim; Jefferson, Angela L.; Shashikumar, Niranjana; Pechman, Kimberly R.; Davis, L. Taylor; Moyer, Daniel; Schilling, Kurt G.; Archer, Derek; Anderson, Adam; Landman, Bennett A. (2026). . PLOS ONE, 21(7), e0350808.

Diffusion tensor imaging (DTI) is an MRI technique used to study the structure of brain tissue, particularly white matter. However, imperfections in the MRI system’s magnetic field gradients can slightly alter diffusion measurements across different parts of the brain and may bias results if not corrected. This study examined the effects of correcting these gradient nonlinearities (GNL) and whether they could influence conclusions about aging and neurological conditions. We analyzed 948 imaging sessions from the Vanderbilt Memory & Aging Project, including measurements of white and gray matter. GNL correction produced changes of about 1% in fractional anisotropy and 3.3% in mean diffusivity, two measures of brain tissue microstructure, as well as changes of about 5 degrees in the estimated direction of nerve fibers, affecting at least 20% of the brain. Some brain regions, particularly superior, occipital, and parietal areas, were more affected, while larger-scale structural measurements changed by up to 12%. Although the effects were generally small at the individual level, they could become statistically important in large studies, particularly those involving multiple MRI scanners or sites. These findings suggest that GNL effects should be considered and, when possible, corrected or measured in large brain imaging studies to improve the reliability of comparisons across individuals, scanners, and clinical groups.

Fig 1.��(a) The MCI clinical cohort comprises 327 participants, each undergoing up to four diffusion MRI sessions acquired on Scanner A (blue) or Scanner B (green).

(b) Maps of the diagonal elements of the gradient nonlinearity tensor, L_xx, L_yy, and L_zz, estimated from empirical phantom field maps for each scanner (FOV 384 × 384 × 384 mm). These entries describe the local scaling of nominal gradients applied along the x, y, and z axes and illustrate scanner-dependent spatial variation in effective diffusion weighting. (c) At a representative brain location (marked in panel b), the sphere plot shows the angular deviation between nominal and GNL-corrected b-vectors (example shown for Scanner B), and the line plot shows the corresponding effective b-values across diffusion volumes for Scanner A (blue) and Scanner B (green) relative to the nominal b-value (orange). Together, these panels summarize the cohort, the underlying gradient nonlinearity fields, and their impact on the effective diffusion encoding used in our analyses. Our study analyzes the effects of GNL on this clinical cohort.