Investigation Of Several Methods In Magnetic Resonance Imaging

Magnetic resonance imaging (MRI) is the most flexible diagnostic imaging modalities, possessing the ability to characterize a wide range of parameters in the living subject and provide exquisite spatial resolution. MRI has also been extensively used in cognitive neuroscience and for the mapping of human brain function. However, MRI still meets challenges in some applications, such as low field permanent system and in vivo liver MRS. So some solutions were proposed in this thesis and the main contents include:1. Multi-shot diffusion weighted PROPELLER on 0.35T open MRI systemThe purpose of this chapter was to develop diffusion weighted periodically rotated overlapping parallel lines with enhanced reconstruction (DW-PROPELLER) on low-field open MR systems. The DW-PROPELLER was implemented on a 0.35T open MR system, and an auto pre-scan technique was used to mitigate non-Carr-Purcell-Meiboom-Gill (CPMG) artifacts. High quality diffusion weighted images and associated ADC maps were obtained. Multi-shot DW-PROPELLER is feasible for clinical use in the diagnosis of acute cerebral infarction on low-field open systems.2. Image correction during large and rapid BO variations in an open MRI system with permanent magnets using navigator echoes and phase compensationAn open permanent magnet system with vertical BO field and without self-shielding can be quite susceptible to perturbations from external magnetic sources. BO variation in such a system located close to a subway station was measured to be greater than 0.7 u T by both MRI and a fluxgate magnetometer. This BO variation caused image artifacts. A navigator echo approach that monitored and compensated the view-toview,variation in magnetic resonance signal phase was developed to correct for image artifacts. Human brain imaging experiments using a multislice gradient-echo sequence demonstrated that the ghosting and blurring artifacts associated with BO variations were effectively removed using the navigator method. 3. Motion artifacts reduction in in vivo 1H-MRS of liver(1) In vivo 1H-MRS of liver:The effects of frequency correctionThe aim of this work was to evaluate the effect of frequency correction on reducing the motion artifacts of in vivo 1H-MRS of liver. Single voxel spectra of middle region in the human liver were acquired using frame-by-frame PRESS 1H-MRS at 1.5 Tesla. The frame-by-frame variations of the frequency of the residual water-signal were analyzed and the variations were corrected retrospectively by frequency shifting individual spectra prior to averaging. Comparison of the summed spectra prior to frequency correction, the summed spectra with frequency correction showed increased SNR and narrower linewidth of both lipid(-(CH2)n-) and residual water signals. This work demonstrated that the frequency correction effectively improved the spectra of in vivo 1H-MRS of liver.(2) Navigator gated breathhold liver proton MR spectroscopy.Multiple breathhold summation was suggested to reduce motion artifacts in liver 1H magnetic resonance spectroscopy, but there could be substantial misregistration among breathholds. We propose a navigator gated breathhold method that can be easily implemented on a standard scanner. The diaphragm navigator echo is used to monitor respiration during the entire scan. A retrospective gating is used to exclude misregistred breathholds. This navigator gating substantially improved the spectral width and peak.4. Benefits of 32-channel over 12-channel coils for BOLD-fMRI at 3.0T MRIThe purpose of this study was to investigate how much the 32-channel 3T phased-array head coil can benefit the BOLD-fMRI, as exemplified in the visual cortex at the brain surface and the medial temporal lobe at the brain center. Ten subjects participated in the visual stimulus study and eight subjects in the memory encoding study on a 3T MRI system using both the 32-channel and 12-channel head coil. For each coil, single-shot gradient-echo EPI data at two different spatial resolutions (2×2×2 mm3 and 3×3×3 mm3) were collected. Statistical analysis was performed at the individual level using both the unsmoothed and smoothed functional data. In agreement with prior studies, the time-course SNR reaches an asymptotic limit due to effect of physiological noise as the image SNR is increased with the 32-channel coil or large effective voxels. The high image SNR from a 32-channel head coil at 3T MRI system is beneficial for both high and low resolution BOLD-fMRI in the medial temporal lobe, but may benefit only high resolution BOLD-fMRI in the visual cortex, which is close to the coil elements.