Global Shimming Research Articles

Hybrid active and passive local shimming (HAPLS) for two-region MRI.

An MRI scanner is equipped with global shim systems for shimming one region of interest (ROI) only. However, it often fails to reach state-of-the-art when shimming two isolated regions of interest simultaneously, even though the two-area shimming can be essential in scan scenarios, such as bilateral breasts or dyadic brains. To address these challenges, a hybrid active and passive local shimming technique is proposed to simultaneously shim two isolated region-of-interest areas within the whole FOV. A local passive shimming system is constructed by optimized bilateral ferromagnetic chip arrays to compensate for the magnet's significant high-order B0 inhomogeneities at the boundary of the manufacturer's specified homogeneous volume, thus locally improving the available FOV. The local active shimming consists of 40-channel DC loops powered by 64-channel current amplifiers. With the optimized current distribution, active shimming can correct the residual low-order B0 inhomogeneities and subject-specific field inhomogeneities. In addition, active shimming is used to homogenize the center frequencies of the two regions. With the implementation of the hybrid active and passive local shimming, the 95% peak-to-peak was reduced from 1.92 to 1.12ppm by 41.7%, and RMS decreased from 0.473 to 0.255ppm by 46.1% in a two-phantom experiment. The volume ratio containing MR voxels within a 0.5-ppm frequency span increased from 64.3% to 81.3% by 26.3%. The proposed hybrid active and passive local shimming technique uses both passive and active local shimming, and it can efficiently shim two areas simultaneously, which is an unmet need for a commercial MRI scanner.

Dynamic B0 shimming for multiband imaging using high order spherical harmonic shims.

To describe and implement a strategy for dynamic slice-by-slice and multiband B0 shimming using spherical harmonic shims in the human brain at 7T. For thin axial slices, spherical harmonic shims can be divided into pairs of shims (z-degenerate and non-z-degenerate) that are spatially degenerate, such that only ½ of the shims (non-z-degenerate) are required for single slice optimizations. However, when combined, the pairs of shims can be used to simultaneously generate the same in-plane symmetries but with different amplitudes as a function of their z location. This enables multiband shimming equivalent to that achievable by single slice-by-slice optimization. All data were acquired at 7T using a spherical harmonic shim insert enabling shimming up through 4th order with two additional 5th order shims (1st-4th+). Dynamic shim updating was achieved using a 10A shim power supply with 2 ms ramps and constrained optimizations to minimize eddy currents. In groups of eight subjects, we demonstrated that: 1) dynamic updating using 1st-4th+ order shims reduced the SD of the B0 field over the whole brain from 32.4 ± 2.6 and 24.9 ± 2 Hz with 1st-2nd and 1st-4th+ static global shimming to 15.1 ± 1.7 Hz; 2) near equivalent performance was achieved when dynamically updating only the non-z-degenerate shims (14.3 ± 1.5 Hz), or when a using multiband shim factor of 2, MBs = 2, and all shims (14.4 ± 2.0 Hz). High order spherical harmonics provide substantial improvements over static global shimming and enable dynamic multiband shimming with near equivalent performance to that of dynamic slice-by-slice shimming. This reduces distortion in echo planar imaging.

A 32-channel multi-coil setup optimized for human brain shimming at 9.4T.

A multi-coil shim setup is designed and optimized for human brain shimming. Here, the size and position of a set of square coils are optimized to improve the shim performance without increasing the number of local coils. Utilizing such a setup is especially beneficial at ultrahigh fields where B0 inhomogeneity in the human brain is more severe. The optimization started with a symmetric arrangement of 32 independent coils. Three parameters per coil were optimized in parallel, including angular and axial positions on a cylinder surface and size of the coil, which were constrained by cylinder size, construction consideration, and amplifiers specifications. B0 maps were acquired at 9.4T in 8 healthy volunteers for use as training data. The global and dynamic shimming performance of the optimized multi-coil were compared in simulations and measurements to a symmetric design and to the scanner's second-order shim setup, respectively. The optimized multi-coil performs better by 14.7% based on standard deviation (SD) improvement with constrained global shimming in comparison to the symmetric positioning of the coils. Global shimming performance was comparable with a symmetric 65-channel multi-coil and full fifth-order spherical harmonic shim coils. On average, an SD of 48.4 and 31.9 Hz was achieved for in vivo measurements after global and dynamic slice-wise shimming, respectively. An optimized multi-coil shim setup was designed and constructed for human whole-brain shimming. Similar performance of the multi-coils with many channels can be achieved with a fewer number of channels when the coils are optimally arranged around the target.

Integrated RF-shim coil allowing two degrees of freedom shim current.

High-quality magnetic resonance imaging and spectroscopic measurements require a highly homogeneous magnetic field. Different from global shimming, localized off-resonance can be corrected by using multi-coil shimming. Previously, integrated RF and shimming coils have been used to implement multi-coil shimming. Such coils share the same conductor for RF signal reception and shim field generation. Here we propose a new design of the integrated RF-shim coil at 3-tesla, where two independent shim current paths are allowed in each coil. This coil permits a higher degree of freedom in shim current distribution design. We use both phantom experiments and simulations to demonstrate the feasibility of this new design.

Optimization of magnetic resonance imaging high-order axial shim coils using boundary element method

In this paper, we present a novel nonlinear optimization algorithm for designing a shim coil system, especially a high-order axial shim coil, for a magnetic resonance imaging（MRI) system. In an MRI equipment, in order to eliminate higher-order harmonic components of the magnetic field within the volume of interest（VOI), passive shimming（PS) is adopted in traditional methods. However, such a method is suitable for the global shimming with low accuracy and poor target. Active shimming（AS) makes up for the shortcomings from PS with a set of shim coils which are designed to generate a specific magnetic fields to improve magnetic field homogeneity within the VOI. Because the complexity of wire pattern increases with the order of AS coil increasing, conventional optimization model cannot meet the design requirements for producing the complicated magnetic field. In this paper, we propose a nonlinear optimization method of designing the axial shim coils for an open-style bi-planar MRI system, based on boundary element method. The optimization model is built in light of influence extents of the various parameters on the coil characteristics for different shim coils. In such a new method, the field error between the magnetic field produced by designed shim coil and the desired target value is selected to be an optimal value subjected to some constraints including line spacing and coil radius, which makes it possible to realize the manufacture process. Meanwhile, the more design parameters, which involve not only the stream function values at each node, but also the compensation parameters and/or the number of grid nodes, are regarded as optimized variables to control the magnetic deviation and characteristics of designed coil. By using some designed shim coils for a 0.5 T open style bi-planar superconducting MRI, including Z1, Z2, Z3 and Z4, the efficiency of such a numerical design method is displayed. Especially for high-order shim coils, more optimized parameters are involved to control the magnetic deviation of the coils, thereby providing a more flexible and straightforward method of designing the axial shim coils.

Numerical investigation of a mono-tube damper with a shim stack

The objective of this paper is to investigate the operation of a mono-tube damper through the application of Computational Fluid Dynamics analysis to the piston and flows through a series of flexible shims which cover exits of the piston orifices. The shims and orifices combine to form a system of variable area flow paths of the damper in parallel with the permanent bleed orifices. Shim stack stiffness characteristics were obtained using experimental and Finite Element techniques. The deflection characteristics were non-linear and were highly dependent upon small gaps present between shims and the restraining bodies. With the nature of the shim deflection being highly complex the computational fluid dynamics models investigated the shim deflection using a global uniform displacement method and also a more representative displacement based upon the finite element shim modelling. It was observed that the global displacement models allowed radially inward flow to establish and also overpredicted pressure drops. Finite element analysis of the shims allowed accurate representation of flow paths to be simulated which closely matched experimental and mathematical predictions. The computational fluid dynamics analyses showed that the discharge velocity for the global shim offset is greater than that from a variable shim deflection calculation. The damper pressure drop is highly dependent upon the shape of the flow path formed by the shim deflection. The presence of sharp direction changes through the piston and valve assembly leads to increased damping rates and piston pressure drops.

High-throughput hyperpolarized 13C metabolic investigations using a multi-channel acquisition system

Magnetic resonance imaging and spectroscopy of hyperpolarized (HP) compounds such as [1-13C]-pyruvate have shown tremendous potential for offering new insight into disease and response to therapy. New applications of this technology in clinical research and care will require extensive validation in cells and animal models, a process that may be limited by the high cost and modest throughput associated with dynamic nuclear polarization. Relatively wide spectral separation between [1-13C]-pyruvate and its chemical endpoints in vivo are conducive to simultaneous multi-sample measurements, even in the presence of a suboptimal global shim. Multi-channel acquisitions could conserve costs and accelerate experiments by allowing acquisition from multiple independent samples following a single dissolution. Unfortunately, many existing preclinical MRI systems are equipped with only a single channel for broadband acquisitions. In this work, we examine the feasibility of this concept using a broadband multi-channel digital receiver extension and detector arrays that allow concurrent measurement of dynamic spectroscopic data from ex vivo enzyme phantoms, in vitro anaplastic thyroid carcinoma cells, and in vivo in tumor-bearing mice. Throughput and the cost of consumables were improved by up to a factor of four. These preliminary results demonstrate the potential for efficient multi-sample studies employing hyperpolarized agents.

High resolution T1ρ mapping of in vivo human knee cartilage at 7T.

PurposeSpin lattice relaxation time in rotating frame (T1ρ) mapping of human knee cartilage has shown promise in detecting biochemical changes during osteoarthritis. Due to higher field strength, MRI at 7T has advantages in term of SNR compared to clinical MR scanners and this can be used to increase in image resolution. Objective of current study was to evaluate the feasibility of high resolution T1ρ mapping of in vivo human knee cartilage at 7T MR scanner.Materials and MethodsIn this study we have used a T1ρ prepared GRE pulse sequence for obtaining high resolution (in plan resolution = 0.2 mm2) T1ρ MRI of human knee cartilage at 7T. The effect of a global and localized reference frequency and reference voltage setting on B0, B1 and T1ρ maps in cartilage was evaluated. Test-retest reliability results of T1ρ values from asymptomatic subjects as well as T1ρ maps from abnormal cartilage of two human subjects are presented. These results are compared with T1ρ MRI data obtained from 3T.ResultsOur approach enabled acquisition of 3D-T1ρ data within allowed SAR limits at 7T. SNR of cartilage on T1ρ weighted images was greater than 90. Off-resonance effects present in the cartilage B0, B1 and T1ρ maps obtained using global shim and reference frequency and voltage setting, were reduced by the proposed localized reference frequency and voltage setting. T1ρ values of cartilage obtained with the localized approach were reproducible. Abnormal knee cartilage showed elevated T1ρ values in affected regions. T1ρ values at 7T were significantly lower (p<0.05) compared to those obtained at 3T.ConclusionIn summary, by using proposed localized frequency and voltage setting approach, high-resolution 3D-T1ρ maps of in vivo human knee cartilage can be obtained in clinically acceptable scan times (<30 min) and SAR constraints, which provides the ability to characterize cartilage molecular integrity.

Numerical and experimental evaluation of RF shimming in the human brain at 9.4 T using a dual-row transmit array

To provide a numerical and experimental investigation of the static RF shimming capabilities in the human brain at 9.4 T using a dual-row transmit array. A detailed numerical model of an existing 16-channel, inductively decoupled dual-row array was constructed using time-domain software together with circuit co-simulation. Experiments were conducted on a 9.4 T scanner. Investigation of RF shimming focused on B1(+) homogeneity, efficiency and local specific absorption rate (SAR) when applied to large brain volumes and on a slice-by-slice basis. Numerical results were consistent with experiments regarding component values, S-parameters and B1(+) pattern, though the B1(+) field was about 25% weaker in measurements than simulations. Global shim settings were able to prevent B1(+) field voids across the entire brain but the capability to simultaneously reduce inhomogeneities was limited. On a slice-by-slice basis, B1(+) standard deviations of below 10% without field dropouts could be achieved in axial, sagittal and coronal orientations across the brain, even with phase-only shimming, but decreased B1(+) efficiency and SAR limitations must be considered. Dual-row transmit arrays facilitate flexible 3D RF management across the entire brain at 9.4 T in order to trade off B1(+) homogeneity against power-efficiency and local SAR.

3D phase unwrapping using global expected phase as a reference: Application to MRI global shimming

MRI phase data often suffers from phase wrapping (i.e., phase may be discontinuous by 2π jumps). Numerous MRI phase unwrapping strategies were developed in the past using a criterion based on phase information of local or neighboring voxels. In this study, an alternative and novel three dimensional phase unwrapping strategy is introduced. This method considers the global character of the phase distribution and utilizes continuous trigonometric functions to construct an expected phase map as an unwrapping reference, which is then used to guide the phase correction of every individual voxel. The original phase is estimated by analyzing the derivative of the wrapped phase image. Simulations of various phase wrapped situations were performed and this new method was also used for an in vivo application (i.e., MRI automatic global shimming). Both simulated and experimental results demonstrate that our proposed method is more reliable and robust than traditional algorithms at obtaining correct phase maps, especially in regions of low-signal and air cavities, such as the abdomen and pelvis.

Dynamic B0 shimming at 7 T

Dynamic slice-wise shimming improves B0 field homogeneity by updating shim coil currents for every slice in a multislice acquisition, producing better field homogeneity over a volume than can be obtained by a single static global shim. The first aim of this work was to evaluate the performance of slice-wise field-map-based second-order dynamic shimming in a human high-field 7 T clinical scanner vis-à-vis image based second order static global shimming. Another goal was to characterize eddy currents induced by second and third order shim switching. A final aim was to compare global and dynamic shimming through shim orders to elucidate the relative benefits of going to higher orders and to dynamic shim updating from a static shimming regime. An external hardware module was used to store and dynamically update slice-optimized shim values during multislice data acquisition. High-bandwidth multislice gradient echo scans with B0 field mapping and low-bandwidth single-shot echo planar scans were performed on phantoms and humans using second-order dynamic and static global shims. For the measurement of second and third order shim induced eddy currents, step response temporal phase changes of individual shims were measured and fit to shim harmonics spatially and to multiexponential decay functions temporally. Finally, an order-wise field-map-based comparison was performed with first, second and third order global static shimming, first and second order dynamic shimming, as well as combined second or third order global and first order dynamic shim. Dynamic shimming considerably improved B0 homogeneity compared to static global shimming both in phantoms and in human subjects, reducing image distortion and signal dropout. The unshielded second and third order shims generated strong B0 and self and cross-term eddy fields, with multiple time constants ranging from milliseconds to seconds. Field homogeneity improved with increasing order of shim, with dynamic shimming performing better than global shimming. Hybrid global and dynamic shimming approach yielded field homogeneity better than global static shims but worse than dynamic shims.

Software compensation of Eddy current fields in multislice high order dynamic shimming

Dynamic B 0 shimming (DS) can produce better field homogeneity than static global shimming by dynamically updating slicewise shim values in a multislice acquisition. The performance of DS however is limited by eddy current fields produced by the switching of 2nd and 3rd order unshielded shims. In this work, we present a novel method of eddy field compensation (EFC) applied to higher order shim induced eddy current fields in multislice DS. This method does not require shim shielding, extra hardware for eddy current compensation or subject specific prescanning. The interactions between shim harmonics are modeled assuming steady state of the medium and long time constant, cross and self term eddy fields in a DS experiment and ‘correction factors’ characterizing the entire set of shim interactions are derived. The correction factors for a given time between shim switches are shown to be invariable with object scanned, shim switching pattern and actual shim values, allowing for their generalized prospective use. Phantom and human head, 2nd and 3rd order DS experiments performed without any hardware eddy current compensation using the technique show large reductions in field gradients and offsets leading to significant improvements in image quality. This method holds promise as an alternative to expensive hardware based eddy current compensation required in 2nd and 3rd order DS.

Dynamic shimming of the human brain at 7 T

Dynamic shim updating (DSU) of the zero- to second-order spherical harmonic field terms has previously been shown to improve the magnetic field homogeneity in the human brain at 4 Tesla. The increased magnetic field inhomogeneity at 7 Tesla can benefit from inclusion of third-order shims during DSU. However, pulsed higher-order shims can generate a multitude of temporally varying magnetic fields arising from eddy-currents that can strongly degrade the magnetic field homogeneity.The first realization of zero- to third-order DSU with full preemphasis and B(0) compensation enabled improved shimming of the human brain at 7 Tesla not only in comparison with global (i.e. static) shimming, but also when compared to state-of-the-art zero- to second-order DSU. Temporal shim-to-shim interactions were measured for each of the 16 zero- to third-order shim coils along 1D column projections on a spherical phantom. The decomposition into up to 3 exponentials allowed full preemphasis and B(0) compensation of all 16 shims covering 67 potential shim-to-shim interactions. Despite the significant improvements achievable with DSU, the magnetic field homogeneity is still not perfect even when updating all zero- through third-order shims. This is because DSU is still inherently limited by the shallowness of the low order spherical harmonic fields and their inability to compensate the higher-order inhomogeneities encountered in vivo. However, DSU maximizes the usefulness of conventional shim coil systems and provides magnetic field homogeneity that is adequate for a wide range of applications.

Improved shimming for fMRI specifically optimizing the local BOLD sensitivity

In functional MRI, magnetic field inhomogeneities due to air-tissue susceptibility differences may lead to severe signal dropouts and geometric distortions in echo-planar images. Therefore, the inhomogeneities in the field are routinely minimized by shimming prior to imaging. However in fMRI, the Blood Oxygen Level Dependent (BOLD) effect is the measure of interest, so the BOLD sensitivity (BS) should be optimized rather than the magnetic field homogeneity. The analytical expression for an estimate of the BOLD sensitivity has been recently developed, allowing for the computation of BOLD sensitivity maps from echo-planar images and field maps. This report describes a novel shimming procedure that optimizes the local BOLD sensitivity over a region of interest. The method is applied in vivo and compared to a standard global shimming procedure. A breath-holding experiment was carried out and demonstrated that the BS-based shimming significantly improved the detection of activation in a target region of interest, the medial orbitofrontal cortex.

SU‐FF‐I‐131: High Spatial Resolution 1H Magnetic Resonance Spectroscopic Imaging Of Human Breast Cancer At 3T

Purpose: To demonstrate the feasibility of high spatial resolution 1H magnetic resonance spectroscopic imaging (MRSI) of human breast cancer in a clinical setting at 3T. Method and Materials: The pulse sequence consists: outer volume pre‐saturation (OVP), independent CHESS pulses for water and lipid suppression, and standard PRESS pulse sequence with an elliptical weighted k‐space sampling scheme. Ten patient studies were performed in a 3.0 Tesla whole‐body clinical imager (Siemens Trio). A four‐channel breast coil (two for each breast) was used for both MRI and MRS. Patients lay prone with their breasts in the coil wells; when possible, gentle breast compression was applied to minimize motion. After global shimming, clinical DCE‐MRI was performed which defined the volume of interest for MRSI. Data processing consisted of line broadening, standard fast Fourier transformations, and phasing. Results: All ten studies were technically successful. The spectra were acquired with FOV = 120×120mm2; TR of 1500ms, TE of 80ms, slice thickness of 14mm, and 24×24 elliptical weighted k‐space sampling with 2 averages for total acquisition of 11.9minutes. All spectra are good in spectral quality, and Cho signals are clearly visible in the MRI‐lesion area, consistent with malignancy, while there was no detectable Cho in the control area. Interestingly, the distribution of Cho signal was nonuniform across the MRI lesion. Conclusion: To our knowledge, this is the first 2D/3D MRSI study of human breast cancer with short TE (less than 135ms) at 3T and highest spatial resolution (up to 0.25cm3) to date. In summary, we have presented a robust technique for high spatial resolution in vivo 1H MRSI of human breast cancer that uses the combined advantages of high field, short TE, multi‐voxel, and high spatial resolution itself to overcome shimming and lipid suppression difficulty, and demonstrated its potential for routine clinical examination.

Volume parcellation for improved dynamic shimming

The need for a homogeneous magnetic field in magnetic resonance imaging is well established, especially at high static magnetic field strengths where susceptibility-induced image distortions and signal losses become excessively large. Dynamic shim updating, where the optimal set of shim currents is applied for each slice during a multi-slice acquisition, has been shown to improve magnetic field homogeneity to a greater extent than conventional global shimming. Here, in an initial feasibility study, we show via simulation that improved efficacy of shimming can be achieved by using the novel parcellated dynamic shimming method. The results of these simulations indicate that parcellated dynamic shimming based on just linear shim terms can perform approximately as well as slice-based dynamic shimming with up to third-order shim terms. This work shows that the effective magnetic field inhomogeneity can be further reduced if shimming and image data acquisition are sequentially performed over a series of compact, cuboidal sub-volumes rather than planes. Further work is needed to develop an imaging approach that can be used for the optimal implementation of parcellated dynamic shimming.

Dynamic shim updating on the human brain

Dynamic alteration of shim settings during a multi-slice imaging experiment can improve static magnetic-field homogeneity over extended volumes. In this report, a pre-emphasized dynamic shim updating (DSU) system capable of rapidly updating all non-degenerate zeroth through second-order shims is presented and applied to high-field multi-slice imaging studies on the human brain. DSU is utilized in both non-oblique and oblique slicing geometries while updating in-plane and through-slice shims. Image-based magnetic-field maps are used to quantify homogeneity improvements and comparisons are made on a slice-specific basis between static global shimming and increasing orders of shim inclusion utilized DSU. The influence of oblique slicing geometry on DSU-utilized global homogeneity is also quantified computationally. Finally, the effect of DSU on susceptibility artifact reduction in single-shot axial-sliced EPI is analyzed using experimental acquisitions.

Interleaved snapshot echo planar imaging of mouse brain at 7.0 T

Single-shot echo planar imaging (EPI) of a mouse brain at high field is very challenging. Large susceptibility-induced gradients affect much of the brain volume, causing severe image deformations and signal loss. Segmented EPI and other conventional multi-shot approaches alleviate these problems but suffer from lower temporal resolution and motion artifacts. We demonstrate that interleaved snapshot EPI represents a simple and robust alternative approach and one that is particularly suitable for high-field T2*-weighted functional imaging of a mouse brain. Similarly to segmented multi-shot techniques, it significantly reduces the susceptibility-related artifacts. At the same time, it preserves the high temporal resolution and the snapshot capability of a conventional EPI by acquiring entire image within a single TR period. We discuss implementation details of the interleaved snapshot EPI sequence and the trade-offs involved between the imaging efficiency, the number of interleaved excitation-acquisition blocks and the artifact reduction. To document the sequence utility, murine brain in vivo imaging with the interleaved snapshot EPI method was compared with a conventional EPI. We found that at least five interleaved blocks were necessary to restore the signal in most cortical areas. We also show that a standard global shimming procedure provides sufficient homogeneity for multi-slice interleaved snapshot EPI acquisition. In contrast, the conventional EPI of comparable image quality would be limited to a single slice with highly optimized local shim. Finally, an in vitro comparison with turbo FLASH acquisition shows the interleaved snapshot EPI to have superior time resolution and signal-to-noise ratio.

High-order MR shimming: a simulation study of the effectiveness of competing methods, using an established susceptibility model of the human head

The first step in the process of shimming a magnetic field is to characterize it by obtaining a field map and decomposing that map into a convenient set of basis functions. The strength of each member of the set is then calculated. Finally, a set of correction elements which generate fields corresponding to the same spatial distribution as the basis functions is energized so that the sum of their fields and the error fields is substantially zero. The basis functions used typically are solutions to Laplace’s equation and have been shown to be very effective when the region of interest is substantially free space. This paper addresses issues associated with shimming the magnetic field in a region in which there is a distribution of materials with different susceptibilities and which therefore is not free space. In such a region, Laplace’s equation is no longer valid and in principle cannot be used to describe the magnetic field there. It is demonstrated that in spite of this, the same set of basis functions suffices for analyzing the field and the same set of elements suffices for correcting the field. The motivation for this study stems from the need to improve the magnetic field homogeneity when biological specimens are being imaged by magnetic resonance. In particular, this paper describes a study carried out by various simulated shimming strategies to improve the uniformity of the magnetic field over a multitissue model of susceptibility of the human head. The topics of magnetic susceptibility, the effect of shimming on MR images, shimming hardware and shimming methods are briefly reviewed. Two slices of the human head model were selected for detailed study, both offset inferior to the origin and including the base of the brain and the anterior sinus. The results of the study include comparisons between the strategies of global shimming, local slice-selective shimming and combinations of the two; the effects of shimming to various orders of spherical harmonics; and the effects of rotation and displacement of the head with respect to the shim frame of reference.

Functional MRI in Epilepsy

*Department of Neurology, Medical College of Wisconsin, Milwaukee, Wisconsin, U.S.A.; †University Hospital Gent, Gent, Belgium; ‡Departments of Diagnostic Radiology and Neurosurgery, Yale University, New Haven, Connecticut; §Department of Neurology, University of Pennsylvania, Philadelphia, Pennsylvania; Clinical Epilepsy Section, National Institutes of Health, Bethesda, Maryland; ¶Department of Radiology, Mayo Clinic, Rochester, Minnesota; and **Department of Neurology, Medical College of Georgia, Augusta, Georgia, U.S.A.