Faculty Bibliography

We introduce new inductive, generative semisupervised mixtures with more finely grained class label generation mechanisms than in previous work. Our models combine advantages of semisupervised mixtures, which achieve label extrapolation over a component, and nearest-neighbor (NN)/nearest-prototype (NP) classification, which achieve accurate classification in the vicinity of labeled samples or prototypes. For our NN-based method, we propose a novel two-stage stochastic data generation, with all samples first generated using a standard finite mixture and then all class labels generated, conditioned on the samples and their components of origin. This mechanism entails an underlying Markov random field, specific to each mixture component or cluster. We invoke the pseudo-likelihood formulation, which forms the basis for an approximate generalized expectation-maximization model learning algorithm. Our NP-based model overcomes a problem with the NN-based model that manifests at very low labeled fractions. Both models are advantageous when within-component class proportions are not constant over the feature space region "owned by" a component. The practicality of this scenario is borne out by experiments on UC Irvine data sets, which demonstrate significant gains in classification accuracy over previous semisupervised mixtures and also overall gains, over KNN classification. Moreover, for very small labeled fractions, our methods overall outperform supervised linear and nonlinear kernel support vector machines.

Radiofrequency magnetic fields are critical to nuclear excitation and signal reception in magnetic resonance imaging. The interactions between these fields and human tissues in anatomical geometries results in a variety of effects regarding image integrity and safety of the human subject. In recent decades, numerical methods of calculation have been used increasingly to understand the effects of these interactions and aid in engineering better, faster, and safer equipment and methods. As magnetic resonance imaging techniques and technology have evolved through the years, so to have the requirements for meaningful interpretation of calculation results. Here, we review the basic physics of radiofrequency electromagnetics in magnetic resonance imaging and discuss a variety of ways radiofrequency field calculations are used in magnetic resonance imaging in engineering and safety assurance from simple systems and sequences through advanced methods of development for the future

Previous works have shown that placement of a high-dielectric pad can improve image intensity in a region adjacent to the pad, or that placement of dielectric pads around a large surface of the head can improve image homogeneity on an entire plane through the head in high-field MRI. Here, experimental results show that use of high-dielectric pads around the human head can reduce the required input radiofrequency power by 50% while enhancing image signal-to-noise ratio by 20-40% throughout the cerebrum at 3 T. Thus, dielectric pads may be used to provide a relatively simple and low-cost method for improving quality and safety of MRI in a variety of applications at 3 T

We describe the design and testing of a quadrature transmit, eight-channel receive array RF coil configuration for the acquisition of images of the entire human spinal column at 7 T. Imaging parameters were selected to enable data acquisition in a clinically relevant scan time. Large field-of-view (FOV) scanning enabled sagittal imaging of the spine in two or three-stations, depending upon the height of the volunteer, with a total scan time of between 10 and 15 min. A total of 10 volunteers have been scanned, with results presented for the three subjects spanning the range of heights and weights, namely one female (1.6 m, 50 kg), one average male (1.8 m, 70 kg), and one large male (1.9 m, 100 kg).

In calculations of temperature increase during MRI, it is typically assumed adequate to consider the Specific energy Absorption Rate (SAR) levels averaged over an entire repetition time (TR) rather than explicitly consider the heating (as it occurs in reality) during the RF pulses only. Here we investigate this assumption with numerical calculations of SAR and temperature increase for a human head in a volume coil at 64 MHz and 300 MHz during three very different pulse sequences, each having a TR of 200 ms and a time-average whole-head SAR of 3.0W/kg, as well as with semi-analytical calculations considering a gradient-echo sequence in a segment of tissue with SAR of 10W/kg delivered in a 1ms pulse with TR of up to 5000 ms. While it is possible to calculate a temporal effect of specific pulse sequence on temperature, the difference between pulse sequences is so small and so transient that it should typically be adequate to consider only the time-average SAR in each TR

PURPOSE: To improve the homogeneity of transmit volume coils at high magnetic fields (> or =4 T). Due to radiofrequency (RF) field/tissue interactions at high fields, 4 T to 8 T, the transmit profile from head-sized volume coils shows a distinctive pattern with relatively strong RF magnetic field B(1) in the center of the brain. MATERIALS AND METHODS: In contrast to conventional volume coils at high field strengths, surface coil phased arrays can provide increased RF field strength peripherally. In theory, simultaneous transmission from these two devices could produce a more homogeneous transmission field. To minimize interactions between the phased array and the volume coil, counter rotating current (CRC) surface coils consisting of two parallel rings carrying opposite currents were used for the phased array. RESULTS: Numerical simulations and experimental data demonstrate that substantial improvements in transmit field homogeneity can be obtained. CONCLUSION: We have demonstrated the feasibility of using simultaneous transmission with human head-sized volume coils and CRC phased arrays to improve homogeneity of the transmit RF B(1) field for high-field MRI systems

Using a small resonant loop to produce a longitudinal traveling wave on a human 7-T system allows MR to be performed over the entire volume of the human leg. We have used this capability to perform localized proton MR spectroscopy of the lipid composition of muscle in volunteers with a coil placed approximately 30 cm away from the region of interest. Spectra with a reasonable signal-to-noise ratio can be acquired in a clinically relevant data acquisition time of less than 5 min using the loop in transmit/receive mode, maintaining the full flexibility to acquire spectra from any part of the calf and/or thigh. If a local receive coil is used in combination with the remote transmit coil, then the signal-to-noise improves significantly, as expected

A comparison of methods to decrease RF power dissipation and related heating in conductive samples using passive conductors surrounding a sample in a solenoid coil is presented. Full-Maxwell finite difference time domain numerical calculations were performed to evaluate the effect of the passive conductors by calculating conservative and magnetically-induced electric field and magnetic field distributions. To validate the simulation method, experimental measurements of temperature increase were conducted using a solenoidal coil (diameter 3 mm), a saline sample (10 mM NaCl) and passive copper shielding wires (50 microm diameter). The temperature increase was 58% lower with the copper wires present for several different input powers to the coil. This was in good agreement with simulation for the same geometry, which indicated 57% lower power dissipated in the sample with conductors present. Simulations indicate that some designs should be capable of reducing temperature increase by more than 85%

It is important to accurately characterize the heating of tissues due to the radiofrequency energy applied during MRI. This has led to an increase in the use of numerical methods to predict specific energy absorption rate distributions for safety assurance in MRI. To ensure these methods are accurate for actual MRI coils, however, it is necessary to compare to experimental results. Here, we report results of some recent efforts to experimentally map temperature change and specific energy absorption rate in a phantom and in vivo where the only source of heat is the radiofrequency fields produced by the imaging coil. Results in a phantom match numerical simulation well, and preliminary results in vivo show measurable temperature increase. With further development, similar methods may be useful for verifying numerical methods for predicting specific energy absorption rate distributions and in some cases for directly measuring temperature changes and specific energy absorption rate induced by the radiofrequency fields in MRI experiments

The major radiofrequency engineering challenges of high-field MR neuroimaging are as follows: (1) to produce a strong, homogeneous transmit B, field, while remaining within regulatory guidelines for tissue power deposition and (2) to receive the signal with the maximum signal-to-noise and the greatest flexibility in terms of utilizing the benefits of parallel imaging. Borrowing from developments in electromagnetic hyperthermia, the first challenge has been met by the use of transmit arrays, in which the input power to each element of the array can be varied in terms of magnitude and phase. Optimization of these parameters, as well as the form of the applied RF pulse, leads to very homogeneous B, fields throughout the brain. The design of large receive arrays, using impedance-mismatched preamplifiers and geometrical overlap for interelement isolation, has resulted in significant sensitivity improvements as well as large acceleration factors in parallel imaging. 0 2010 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 20, 2-13, 2010; Published online in Wiley InterScience (wwwintersciencemiley.com). DOI 10.1002/ima.20219