T Vaughan

Center for Magnetic Resonance Research Minnesota, USA, Minneapolis, Minnesota, United States

Are you T Vaughan?

Claim your profile

Publications (15)6.09 Total impact

  • [Show abstract] [Hide abstract] ABSTRACT: The objective of this study was to demonstrate the feasibility of simultaneous bilateral hip imaging at 7 Tesla. Hip joint MRI becomes clinically critical since recent advances have made hip arthroscopy an efficacious approach to treat a variety of early hip diseases. The success of these treatments requires a reliable and accurate diagnosis of intraarticular abnormalities at an early stage. Articular cartilage assessment is especially important to guide surgical decisions but is difficult to achieve with current MR methods. Because of gains in tissue contrast and spatial resolution reported at ultra high magnetic fields, there are strong expectations that imaging the hip joint at 7 Tesla will improve diagnostic accuracy. Furthermore, there is growing evidence that the majority of these hip abnormalities occur bilaterally, emphasizing the need for bilateral imaging. However, obtaining high quality images in the human torso, in particular of both hips simultaneously, must overcome a major challenge arising from the damped traveling wave behaviour of RF waves at 7 Tesla that leads to severe inhomogeneities in transmit B1 (B1+) phase and magnitude, typically resulting in areas of low signal and contrast, and consequently impairing use for clinical applications. To overcome this problem, a 16-channel stripline transceiver RF coil was used, together with a B1 shimming algorithm aiming at maximizing B1+ in six regions of interest over the hips that were identified on axial scout images. Our successful results demonstrate that this approach effectively reduces inhomogeneities observed before B1 shimming and provides high joint tissue contrast in both hips while reducing the required RF power. Critical to this success was a fast small flip angle B1+ calibration scan that permitted the computation of subject-specific B1 shimming solutions, a necessary step to account for large spatial variations in B1+ phase observed in different subjects. Copyright
    No preview · Article · Oct 2012 · NMR in Biomedicine
  • Source
    [Show abstract] [Hide abstract] ABSTRACT: We report the first comparison of cardiovascular magnetic resonance imaging (CMR) at 1.5 T, 3 T and 7 T field strengths using steady state free precession (SSFP) and fast low angle shot (FLASH) cine sequences. Cardiac volumes and mass measurements were assessed for feasibility, reproducibility and validity at each given field strength using FLASH and SSFP sequences. Ten healthy volunteers underwent retrospectively electrocardiogram (ECG) gated CMR at 1.5 T, 3 T and 7 T using FLASH and SSFP sequences. B1 and B0 shimming and frequency scouts were used to optimise image quality. Cardiac volume and mass measurements were not significantly affected by field strength when using the same imaging sequence (P > 0.05 for all parameters at 1.5 T, 3 T and 7 T). SSFP imaging returned larger end diastolic and end systolic volumes and smaller left ventricular masses than FLASH imaging at 7 T, and at the lower field strengths (P < 0.05 for each parameter). However, univariate general linear model analysis with fixed effects for sequence and field strengths found an interaction between imaging sequence and field strength (P = 0.03), with a smaller difference in volumes and mass measurements between SSFP and FLASH imaging at 7 T than 1.5 T and 3 T. SSFP and FLASH cine imaging at 7 T is technically feasible and provides valid assessment of cardiac volumes and mass compared with CMR imaging at 1.5 T and 3 T field strengths.
    Full-text · Article · Jan 2012 · NMR in Biomedicine
  • No preview · Conference Paper · Jan 2011
  • No preview · Conference Paper · Jan 2010
  • Source
    [Show abstract] [Hide abstract] ABSTRACT: There has been an increasing interest in constraining transmit B1 shimming with specific absorption rate (SAR) limits, especially at high magnetic field. Since most of the existing methods rely on solving a nonconvex optimization problem, they are typically faced with two difficulties: Only local optimum solutions are obtained, and they are susceptible to the chosen initial points for optimization. Here we introduce a two stage optimization method where a reliable initial point is acquired in the first stage by a convex semidefinite relaxation (SDR) approximation method. A high quality B1 shimmed map then can be obtained in the second stage optimization using the SDR initial points. The presented technique is verified with simulations for a 16-channel transmit coil array at 7T with a human head model.
    Full-text · Conference Paper · Jan 2008
  • No preview · Conference Paper · Jan 2007
  • Source
    [Show abstract] [Hide abstract] ABSTRACT: Introduction: At very high frequencies and field strength the ability to influence the B 1 fields by means of unique coil layout [1-4] or a high number of independent RF transmit coils is crucial to improve image homogeneity and parallel imaging performance [5-8]. Recent numerical calculations of designs based on short transmission line coils with up to 80 individual elements seem to indicate great potential of such coils for RF shimming [9]. Here we describe a head coil design with 32 short transmission line elements, discuss practical considerations and present initial data acquired with a 32 channel transmit system when using equal RF amplitude per channel. Methods: The transmit/receive array consisted of two segments with sixteen concentrically arranged 8cm short resonance elements each (Fig.1A). The coil dimension was 21cm x25cm and a 12mm thick Teflon substrate between the conductors and shield was used. Decoupling capacitor networks [10, 11] between immediate neighboring elements were utilized to isolate the individual coils from each other (Fig. 1B). To avoid the potential for a signal loss at the seam between the two 16 element segments, the distance between the segments was minimized to 5mm. The ground of each array element was 5 cm wide and physically separated from neighboring elements. The length of each resonance element conductor strip was 8 cm. Gradient induced eddy currents in the ground plane were reduced by using a circuit board material with 5 µ m thin copper substrate. High voltage ceramic chip capacitors (ATC 100E and C series) were used to capacitively shorten the individual resonance elements towards ground and to achieve λ/2 resonance. A 7T Siemens whole body system based on TIM technology with 32 receive channels was used in conjunction with a 8 kW CPC RF amplifier (NY,USA), a 32 times equal amplitude RF splitter (Werlatone, NY, USA) and 32 T/R switches (16 Stark Contrast, Erlangen, Germany and 16 Varian Inc, CA). The transmit phase could be controlled by phase shifters. Results and Discussion: By building the transmission line coil with extensive decoupling capacitor networks (Fig.1B), we were able to reduce the average crosstalk between the 32 coil elements for the human head below -20dB (Fig.2). Initial experiments indicate that we can allow for fixed adjusted decoupling capacitors and even prefixed tune/match capacitors for similar head shapes. This reduces the setup time for such an array significantly. At 7 Tesla we were able to cover the whole brain with two segments of transmission line elements (Fig. 3). We attribute being able to avoid signal voids between the segments (Fig. 4) to the narrow gap of only 5mm. Conclusions: Our preliminary results indicate that coils based on short transmission lines can be built and decoupled at 7T. Such coils improve RF shimming ability and are expected to aid 3D parallel imaging performance. The presented coil for example allowed for two independent B 1 shim sets to be applied along the z direction. While certainly more careful studies will be needed to evaluate the true potential and the limitations of such coils the initial results are very encouraging. Fig.1 The coil and a schematic of the circuitry.
    Full-text · Article · Jan 2007
  • No preview · Conference Paper · Jan 2007
  • Source
    Full-text · Article · Jan 2005
  • Source
    [Show abstract] [Hide abstract] ABSTRACT: Synopsis: The objective of this study was to design, prototype and prove a new head coil that is both patient friendly and high in performance for head imaging with the new high field strength MR systems. An ideal clinical head coil would lend itself to the easiest and most comfortable patient positioning, leave the patient's face uncovered once the patient is in position, and would include these access and comfort features without compromising coil performance, safety and reliability. This coil has been achieved. Objectives: To design a head coil for patient accessibility, comfort, and high performance for high field MRI. Background: A survey of local radiological technicians advised above all else, that a new head coil should be both comfortable and accessible for the patient and easy to use for the technician. Top coil performance was assumed by this demanding group. The ideal coil should have a removable top as well, to allow for comfortable patient positioning in the coil. Furthermore, some commercial systems don't provide the space for a coil that slides over a patient's head. Accordingly, several commercial coil designs already incorporate this "pop top" feature. Investigators however pointed out that these commercial coils that separate into halves were not popular with some research applications such as fMRI. Apparently the electrical contacts that are broken and remade each time a new subject is loaded, become unstable over time due to wear and oxidation, resulting in noise "spikes" and temporal instabilities often seen in EPI images for example. These electrical contacts are required to complete the end ring current paths in birdcages and similar coil structures. While commercial coils must meet rigorous FDA safety criteria, it could be imagined that electrical contacts in a coil might possibly pose safety risks in certain situations, especially where electrolytic bodily fluids were present. Methods: The transmission line (TEM) resonator design was selected as a candidate to meet these challenging criteria. It performs well at high frequencies.[1] It can be opened over the face by removing an element.[2] And because the TEM coil requires no end ring current path, this inductively coupled structure can be broken into two inductively coupled halves without making or breaking problematic electrical contacts. See Figure 1. To demonstrate these points, an open faced TEM head coil was constructed per Figure 1a. The test coil measured 26.5cm i.d. by 20cm long. The coil was then bench tested before, and again after electrical separation of the top and bottom halves. See Figure 1b. A multi-slice gradient echo data set was acquired from a volunteer for the image performance testing of the Figure 1 coil. See Figure 2. Results: Figure 2a records network S11 input power reflection measurements (top), and S12 transmission measurements (bottom) of the probed B 1 field for the undivided head coil of Figure 1a. The frequency measured marks the M=1, transverse B 1 field mode for the coil. The adjacent modes M=0 and M=2 are shown for reference. Figure 2b records the same measurements again, after the coil was electrically separated (but physically together) into two inductively coupled components. As observed in the network measurements in Figure 2, the process of cutting the coils' copper cavity into top and bottom pieces had no measurable effect on the TEM coil's performance. Q for the whole coil and the divided coil measured 816 and 822 respectively. The relative frequency shifted from 169.869 MHz to 169.762 MHz when the whole coil was divided. The relative B 1 field gain and B 1 uniformity differences between the unsplit and split coils were not measurable. The images in Figure 2c confirm the bench measurement findings. Image uniformity, SNR and RF power requirements for the coupled pair are identical to those reported for the conventional, physically continuous TEM head coil at 4T. [1] Conclusions: Demonstrated is a high performance, high field TEM head coil which uncovers the face (chin, mouth, nose, and eyes) for patient comfort during the scan and uncovers the whole head for the greatest access and ease of positioning during patient entry and exit. The ease of access and exposed face allow the coil to assume a smaller, closer contour about the head for further imaging performance increases. All of these features are achieved without making or breaking any electrical contacts, for added reliability and safety. Further work will include a more ergonomic contour for the coil package.
    Full-text · Article · Jan 2003
  • Source
    T Vaughan · C Snyder · G Adriany · L Bolinger · H Liu · A Stolpen · K Ugurbil
    [Show abstract] [Hide abstract] ABSTRACT: Synopsis: To make whole body imaging possible at 4T, new understanding, technology, and techniques have been developed. New understanding was required to update initial impressions that practical 4T body imaging was not feasible. RF technology including a body coil, front end, and specialty receiver technology was developed to make 4T body imaging possible. New techniques such as RF shimming were implemented to correct for artifacts not observed at lower field strengths. The result of bringing together new technology, techniques and understanding is a demonstration of the feasibility of 4T whole body MRI with examples from brain, breast, abdomen and heart. Objective: To investigate the feasibility of whole body imaging at 4T using a newly developed body coil and front end together with applications dedicated phased array and parallel array receivers. Background: Notably, 4T whole body imaging was performed nearly 15 years ago.[1,2] As with most "first" results, the images acquired were not the "best" results. Skepticism about imaging at 4T followed. RF power requirements were thought to be too high, and RF penetration too low for successful body imaging at 170 MHz. Recent success with body imaging has been demonstrated by the major MR system manufacturers at 3T,[3] though the limits of conventional coil technologies and RF power requirements are being challenged at this field. Is clinical quality body imaging yet possible at 4T and higher? By developing and applying new RF technology and techniques, 4T body imaging may now prove to be feasible. Methods and Materials: An active body coil system was developed for efficient head and body imaging at field strengths of 4T and higher. This complete RF front-end system employs an actively detuned TEM body coil for NMR signal excitation together with local receiver coils of the phased array or parallel array type.[4] Included in the system are the homogeneous transmit coil (a), the multi-channel receiver coil (b), the coil power supply and control unit (c), the optically triggered, nonmagnetic PIN diode driver unit (d), the non-magnetic, multi-channel preamplifier (e), and all of the necessary fiber optic control lines (f), power supply lines (g) and RF signal cables (h). Not shown are the high power, non-magnetic transmit / receive (TR) switches that complete the RF front end. The actively detuned body coil was used as a transmitter, together with applications specific receiver coils for better imaging of the regions of interest in the body. Figure 1. Body Coil System.
    Full-text · Article · Jan 2003
  • T. Vaughan · M. Garwood · K. Ugurbil
    [Show abstract] [Hide abstract] ABSTRACT: These are exciting times for biomedical MRI. The new decade has arrived with 3 T magnets in the clinics, with 4 T, 7 T, and 8 T magnets in the laboratories and 9.4 T magnets on the order books. The main reason for these high field magnets, now reaching the limits of niobium-titanium superconducting technology, is to gain higher signal-to-noise ratio (SNR) and the imaging speed, resolution and contrast that come with it. Magnetic field strength however, is only one of several instrument parameters affecting the SNR. Our objective is to design a more efficient RF volume coil for human head and body MRI at the highest fields available. A new TEM volume coil is reported which improves on performance, patient access, and magnet bore space requirements compared to conventional coils. The solutions explored for the head coil in this study carry to body coils with arm slots and to the extremity coils for the foot, shoulder, elbow etc
    No preview · Conference Paper · Feb 2001
  • Source
    X Zhang · X Zhu · H Qiao · H Liu · T Vaughan · K Ugurbil · W Chen
    [Show abstract] [Hide abstract] ABSTRACT: SYNOPSIS A double-tuned human head TEM coil is designed for in vivo proton and 31 P MR applications at 7T. 3D 31 P CSI and high-resolution human head image acquired from the whole brain using this coil are presented. A superior SNR of 91:1 of 31 P spectra from 15.5-cc voxels in the human brain is achieved within 18-minute data collection in vivo. The results of in vivo MR experiments and bench test indicate i) excellent performance of this double-tuned large volume coil design for human proton and 31 P MR studies; ii) great advantages of 31 P MRS at ultra-high fields. INTRODUCTION The TEM volume coil was introduced by Vaughan et al originally for human head MR applications at 4T (1,2). Recent studies further demonstrated the superior performance of the TEM coil design for proton MRI at 7T (3) and 8T (4). In this work, we present a double-tuned, circular-polarized TEM volume coil for human head proton and 31 P MRI/MRS studies at 7 Tesla. A much shorter coil length, compared with commonly used single-tuned coil, was applied in this coil design, for restricting coil's B 1 field to the human brain for higher sensitivity of brain 31 P spectrum detection. METHOD The circular-polarized double-tuned TEM volume coil was built in our lab to operate at 296.05 MHz and 119.85 MHz, Larmor frequencies of proton and 31 P at 7T. The method of design and construction of this coil was similar to the previous work (3). 8 resonant elements of proton channel and 8 resonant elements of 31 P channel were alternatively positioned in a cylindrical cavity made from 18-µm thick copper foil, with dimensions of 34-cm in diameter and 15.3-cm in length, as indicated in Figure 1. The coil diameter was measured to be 28 cm. The inner conductors of the coaxial resonant elements were 0.64-cm diameter solid copper rods for both proton and 31 P channels. Adhesive-backed copper foils (3M, St Paul, MN) with 35µm thickness were directly wrapped on the dielectric material of the struts to simply form the outer conductors of the resonant elements. The diameters of these outer conductors were 1.47-cm for proton channel and 0.84-cm for 31 P channel, respectively. For 31 P channel, the much thinner dielectric of ~0.1-cm combined with a 30-pF capacitor connected on one end of each resonant element of 31 P channel was applied to ensure the desired resonant mode operating at a relatively low frequency of 120 MHz, i.e. Larmor frequency of 31 P at 7T. The dielectric materials between the inner conductors and outer conductors for both channels were PTFE that has a low loss property that is important for RF coil designs, especially high frequency coil designs. The coil was built with no end-cap that made the coil friendlier to patients and easy to use. The coil was driven quadraturely with two KDI 3dB-90 0 -hybrids model QH-23 for 1 H and QH-21 for 31 P, respectively. All MR experiments were performed on a 7T/90cm magnet interfaced to the Varian INOVA console. RESULTS The prototype coil as shown in Fig. 2 was optimized for human head proton and 31 P studies at 7T. The proton and 31 P channels were tuned to 296.05 MHz and 119.85 MHz, respectively. The S 21 measurements performed on the network analyzer showed that the isolation between the proton channel and 31 P channel was better than –30 dB while isolation between the two quadrature-ports of proton channel was greater than –35 dB and the quadrature port isolation of 31 P channel had the same level as that of proton channel. Loaded and unloaded Q measured 195/90 for proton channel at 296.05 MHz and 455/120 for 31 P channel at 119.85 MHz. The coil can be easily matched to system 50 Ohm for both proton and 31 P channels. Fig.3 shows the experiments with a mineral oil phantom. The results show this circular-polarized coil was well behaved at 296.05 MHz. Fig. 4 shows a high quality T 1 weighted human head image acquired using the coil with a slice thickness of 2 mm. In 31 P MRS study, an excellent SNR of 91:1 from a 15.5-cc voxel inside human brain was achieved within 18 minutes in vivo, as shown in Fig.5. Such superior sensitivity made it possible to reliably detect a small amount of uridine diphosphate (UDP) resonance in the deepest brain region as shown in Fig 5. The average RF power in a duty cycle used to generate such a 3D 31 P chemical shift imaging was 0.05 W/kg with a pulse width of 500 µs. CONCLUSION A circular-polarized doubly tuned TEM coil is successfully designed and constructed for human head 1 H and 31 P MRI/MRS at 7T. Although only 8 struts were used for each channel with a short coil length, the coil still presents an excellent performance for both proton and 31 P in human brain studies at 7T. The success of this coil will certainly aid ultra-high field human 31 P MRS studies.
    Preview · Article ·
  • Source
    H Liu · C Snyder · T Vaughan · K Ugurbil
    [Show abstract] [Hide abstract] ABSTRACT: Generally, sodium is the second most abundant element in the biological system after hydrogen. Under normal conditions the concentration of sodium in human tissue is approximately 15mM intracelluarily, while 145mM extracellularily. And, this large ionic concentration gradient across the cellular membrane is maintained dynamically by the Na-K pump on the cell membrane that consumes ATP. The failure of the Na-K pump as result of any impaired ATP production during an ischemia event will inevitably lead to a significant increase in the intracelluar sodium concentration and then total tissue sodium content locally. For the same reason, the tissue sodium concentration can be an important indicator of its pathological status as well as a potential predictor of its fate during an ischemic event or degeneration or other pathological development (1-2). In the case of myocardial infraction (MI) or brain stroke, a quantitative sodium imaging can provide an important piece of information regarding to the physiology of an underlying pathology, which can be more specific and quantitative in monitoring its evolution than the routine MRI examination. We have implemented a sodium imaging technique at 7 Tesla, which was based on a volumetric gradient-echo (GE) acquisition scheme. And, in the sequence a short RF pulse was used to minimize unwanted magnetization relaxation processes during imaging. Furthermore, an optimized k-space averaging scheme was used to shorten total data acquisition time. Also, a circularly polarized TEM cavity resonator (tuned to 78.5MHz with lamped capacitors) was used as a volume probe in imaging studies. An efficient head-insert gradient coil set was used. We have obtained brain sodium images from human subjects with the following scan parameters: TR/TE=50/1.5msec, flip angle<90, NSA=8. The spatial resolution parameters of the volumetric image set were: matrix of 64x48x16 corresponding to field of views of 25.6x19.2x12.8cm3. All imaging experiments were performed on a 7T whole body scanner with Varian console, Siemens gradient amplifier and 4kW CPC RF amplifier. 3D sodium MR images of the brain of human subject were obtained at 7 Tesla in about 8 min. Representative brain sodium images are shown below. Also, the transverse relaxation time (T2) was measured using a multiple echo version of the same imaging sequence. Since the sodium magnetization relaxation times of the brain tissue are significantly shorter than those of proton, its imaging can be performed more rapidly than the hydrogen imaging. However, caution should be taken when sequence repetition time is short. When TR was 50msec, the worst case SAR was found to be 2.225w/kg. This experimental result clearly suggests the potential of 7Tesla MR scanner for high field imaging applications using nuclei with relatively low gyromagnetic ratio. Furthermore, the added SNR at 7 Tesla allows a possibility of achieving high resolution imaging with voxel volume as small as 0.128 cc in reasonable scan duration. The sodium MR imaging provides information on total tissue sodium concentration, which is a physiological parameter for characterizing pathological tissue in a disease process. Considering the potential usefulness and the quality of images and the scan time required, the non-invasive sodium imaging can be both feasible and practical for diagnostic imaging applications of human at 7T. Acknowledgments This work was partially supported by NIH grants P41 RR08079, NS38070, and the grants from WM Keck foundation.
    Preview · Article ·
  • No preview · Article ·

Publication Stats

105 Citations
6.09 Total Impact Points

Top Journals


  • 2001-2012
    • Center for Magnetic Resonance Research Minnesota, USA
      Minneapolis, Minnesota, United States
  • 2011
    • Optech Montréal
      Montréal, Quebec, Canada
  • 2010
    • City of Stockholm
      Tukholma, Stockholm, Sweden
  • 2007
    • Humboldt-Universität zu Berlin
      Berlín, Berlin, Germany
  • 2005
    • University of Minnesota Duluth
      Duluth, Minnesota, United States