RF resonators, also known as RF coils, RF antennas, and electromagnetic probes, are crucial devices for obtaining high-quality magnetic resonance images for clinical diagnosis. In this project, we develop numerical methods for analysis and design of such resonators for MRI applications. Specific mathematical models will be developed for low- and high-field MRI systems, which may include high-frequency phase variation and bioeffect dosimetry for RF fields.

The mathematical basis of tomographic imaging is conventionally rooted in the well-established Fourier or radon transform theories, so that image quality is mainly dependent on how the data space is sampled. In practice, physical and temporal constraints often prevent a sufficient coverage of the data space, resulting in various image artifacts, such as Gibbs ringing, resolution degradation, and various motion effects. This project is aimed at overcoming these problems by developing new model-based imaging techniques that can effectively incorporate information a priori into the imaging process. Application of these techniques to cardiac imaging and functional brain mapping is also addressed.

In many imaging modalities, physical quantities of interest are often transduced to the phase of a complex signal. The main difficulty in extracting this phase function from the measured data lies in the fact that the phase of a complex function is uniquely defined only in the principal value range (-pi, pi), and any value outside this interval will be wrapped around to produce a so-called wrapped phase. In this project, we are developing new algorithms to reconstruct the unwrapped phase directly from the wrapped phase or from the original complex signal.

Conventional magnetic resonance imaging methods collect and process data based on the classical Fourier transform theory. Consequently, they suffer an important trade-off between spatial resolution, temporal resolution, and signal-to-noise ratio. In this project, we are developing a novel model-based imaging method to enable high-resolution dynamic image sequences to be obtained in real time.

Magnetic resonance spectroscopic imaging promises to provide an entirely new way to examine the dynamics of human biochemical processes in vivo noninvasively. However, its practical applications have been thus far rather limited because of low sensitivity and long imaging time. The primary objective of this research is to develop mathematical methods to effectively utilize the readily available anatomical information to constrain the spectral distribution so that we can reduce imaging time without compromising spatial resolution.

Neural networks are computational paradigms motivated by neurobiological principles. They are endowed with several unique attributes that promise to provide a new approach to solving many difficult problems in machine vision, robot control, and speech recognition. In this research project, we will strive to find answers to two fundamental questions: (1) What computational characteristics should a formal neuron possess? and (2) What are the organization principles for building large-scale networks based on these elements. We anticipate that this research will have many applications, including automatic image feature extraction and human-machine intelligent interaction.

We have modified a fiber optic fluorometer to measure fluorescent signal intensities across an epithelium barrier. As a medically relevant example, we measured the pH of the effusion formed during infection of the middle ear (otitis media) of the chinchilla. Because the choice of antibiotic to be used in clincial therapy is dependent on the pH of the effusion, a noninvasive method of measuring the pH is highly desirable. We were able to detect changes of 0.2 pH units with the fluorescent pH probe carboxy-seminaptharhodafluor. The development and resolution of the otitis media was followed with magnetic resonance imaging, since this is a noninvasive monitoring method that will not perturb the progression of the infection.

The goal of this project is to develop new miniature radio frequency coils that will extend the application of NMR microscopy to smaller fields of view. Such coils will provide a new window for the observation of microscopic biological phenomena. Existing NMR microscopy coils do not possess sufficient sensitivity for signal detection from cellular sample volumes and do not provide adequate signal-to-noise for high-resolution NMR imaging at these dimensions. We propose to develop new coils with diameters as small as 10 µm, which would be two orders of magnitude smaller than those presently used.

Functional magnetic resonance (MR) techniques provide physiological and biochemical information in a noninvasive fashion. Unlike proton relaxational agents, which act by altering tissue relaxation times (T₁ and T₂), functional agents encode physiological information directly in the chemical shift of the NMR active nuclei. We are investigating new functional MR agents to map the partial pressure of oxygen, temperature, or pH in tissues to MR images. The goal is to obtain accurate, localized information in vivo within reasonable imaging times. This technology is relevant to radiation therapy and hyperthermia treatment of tumors, where knowledge of local oxygen pressure, temperature, and pH can provide valuable information for optimizing treatment protocols.

The objective of this research is to improve the performance of nuclear magnetic resonance (NMR) microscopy systems by fabricating the radio-frequency (RF) detection coil and preamplifier on a single monolithic GaAs substrate. The RF microcoil will be formed from 4 to 7 turns of a gold conductor in a planar geometry. A simple tuning and impedance matching network will link the RF coil to a single stage, common source configuration, GaAs metal semiconductor field effect transistor (MESFET). The amplifier will be designed for high gain and low noise over the relatively narrow bandwidth needed for NMR signal detection. Gallium arsenide MESFET amplifiers are ideal for this purpose.

We are using MRI to assess the efficacy of new macrolide drugs in curing middle ear disease in the animal models of the human disease. Research areas include the construction of specialized high sensitivity radio-frequency coils, and the determination of minimum drug doses that are efficacious in treatment.

Fast T₁ mapping techniques can be used to assess the temperature profiles within agricultural products during heating. This can be used as a noninvasive method for assessing efficacy in bacterial kill-rate in foods.

Conventional MRI provides anatomical and spatial information in both medical and nonmedical scanning regimes. Using newly developed agents, we are able to determine noninvasively such measurements as temperature, oxygen levels, and pH. Areas of investigation are chemical synthesis and theoretical modeling of new compounds, optimization of RF coil geometries, and development of post-processing algorithms to increase sensitivity.

The overall aim of this research area is to increase the mass sensitivity of nuclear magnetic resonance by two to three orders of magnitude using extreme miniaturization of the radio-frequency coil used as the detector. Susceptibility matching schemes are being developed to ensure that high-resolution spectra are acquired.

Fluorine- and proton-based phase-transition agents are being synthesized for in vivo temperature mapping using magnetic resonance imaging. Applications to hyperthermia treatment of cancer are being investigated.