All ETDs from UAB

Advisory Committee Chair

Donald B Twieg

Advisory Committee Members

Allan C Dobbins

Rotem A Elgavish

Stanley J Reeves

Ernest M Stokely

Xincheng Yao

Document Type

Dissertation

Date of Award

2010

Degree Name by School

Doctor of Philosophy (PhD) School of Engineering

Abstract

Rapid and accurate mapping of relaxation rates applies to a number of neuroimaging and functional magnetic resonance imaging (fMRI) studies. Previously developed relaxation rate mapping methods are based mostly on non-linear fitting of intensity images. For the first time, this study developed an ultrafast and direct transverse relaxation rate mapping technique, Spin Echo Single-Shot Parameter Assessment by Retrieval from Single Encoding (SE-SS-PARSE). Four useful imaging parameters, local transverse magnetization magnitude (Mxy0), frequency (f), reversible and irreversible transverse re-laxation rate (R2' and R2), can be estimated using an iterative searching algorithm simul-taneously and quantitatively. In addition, the SE-SS-PARSE technique is free of geome-tric errors and blurring artifacts that commonly occur in the other SS techniques, such as echo planar imaging (EPI). Specifically, this work developed the signal model of SE-SS-PARSE and tested this technique with computer simulations, experimental phantoms and biological samples. In simulations, even at a noise level with low signal-to-noise ratio (SNR) of 20dB, SE-SS-PARSE is able to generate accurate local magnetization parameter maps subject to statis-tical comparison with actual parameter values. After successful testing with a numerical phantom, the SE-SS-PARSE technique was performed on a realistic four-tube phantom, which was designed to produce in-vivo-like and different R2 and R2' values in each tube. The estimated relaxation rates from SE-SS-PARSE were highly correlated with relaxation rates computed from slower "gold standard" conventional MRI methods (correlation coefficients r1=0.9636 for R2', r2=0.9788 for R2). The ultrafast SE-SS-PARSE tech-nique was also compared with a widely used T2-weighted imaging MRI technique, fast spin echo (FSE), in mapping R2 values. Four-tube phantom studies indicate that SE-SS-PARSE produces results with accuracy equivalent to FSE, but with a much shorter acqui-sition time. In addition, biological sample studies show that SE-SS-PARSE with fat satu-ration pulse produces more accurate R2 estimates than used without fat saturation when fat is present. In conclusion, SE-SS-PARSE is an ultrafast MRI method that can produce reliable R2 and R2' mapping. It is expected that SE-SS-PARSE will find important appli-cations in neuroimaging and fMRI studies, especially in recording rapidly changing phy-siological phenomena.

Included in

Engineering Commons

Share

COinS