| In Magnetic Resonance Imaging (MRI), signal is complex valued and can be represented by a magnitude image and a phase map. Although magnitude images are used much more frequently in clinical diagnosis than phase maps, the latter should not be undervalued because a lot of valuable information is encoded only in phase, which has great potential in medical applications. My doctoral thesis focuses on phase information in MRI and its applications in the areas of fast imaging and Inversion Recovery (IR) imaging.;IR imaging is one of the most useful techniques in MRI contrast manipulation, but its use is limited because of the presence of phase errors. The thesis proposes a new method to correct phase errors that occur in IR imaging. The method models phase variation with a polynomial, whose coefficients are statistically determined by calculating relative vector rotations of complex fields after being shifted by n pixels. As discovered in this thesis, increasing the pixel shift effectively enhances phase signal without amplifying the corresponding noise, and thereby improves phase correction. The method has been successfully demonstrated with 2D in vivo IR imaging data.;Phase information has great potential also in fast imaging in MRI. For example, a recently published fast imaging method named "Skipped Phase Encoding and Edge Deghosting (SPEED)" conducts strategic spatial phase encoding and thereby accelerates MRI. SPEED is promising but is demonstrated so far only in 2D and only with a single coil. This thesis presents new developments based on the principle of SPEED: First, SPEED is extended from 2D to 3D to reduce scan time with more flexibility and efficiency; Second, SPEED is combined with Half-Fourier imaging to accelerate MRI further by a factor of nearly 2; Third, SPEED is simplified based on the sparseness of MR angiography data; Fourth, a new parallel imaging strategy named "SPEED with Array Coil Enhancement (SPEED-ACE)" is proposed, which extends SPEED from a single coil to multiple coils to further accelerate MRI. Finally, signal-to-noise ratio (SNR) in SPEED is studied, and a novel approach for SNR improvement is proposed and demonstrated with computer simulations and in vivo data. |
