STAND-ALONE IMAGE RECONSTRUCTION FOR MULTI-SLICE ECHO-PLANAR IMAGING, WITH APPLICATIONS TO STUDY HUMAN BRAIN FUNCTIONS
Abstract
Optimizing the speed of image acquisition in magnetic resonance imaging (MRI) is a significant consideration to reduce patient examination time and/or to increase temporal resolution in dynamic studies. The advancement of simultaneous, multi-slice imaging increased the acquisition efficiency of MRI data. This technique for reducing scan time has opened a new door for functional MRI studies and diffusion-based fiber tractography to visualize the structural networks in the human brain [1]. The problem with the existing multi-slice image reconstruction algorithm using the MATLAB [2] program is that it is completely dependent on the MATLAB environment. In addition, the algorithm can be performed only on offline, preventing monitoring of subject motion and brain activation during scanning in order to adjust task presentation and for utilizing the brain signal to control other equipment and neurofeedback. To date, there is no stand-alone method for image reconstruction for multi-slice EPI data. To meet this need, I propose C/C++ programming language-based image reconstruction using the Slice-GRAPPA [3] algorithm for multi-slice acquisition and GRAPPA [4] algorithm for accelerating the image acquisition in the phase encoding direction. The main advantage of this reconstruction based on C/C++ is that it is stand-alone. In addition, optimizing the reconstruction program speed will enable it to be embedded into software to be applied in real time fMRI studies. This process was validated through matching the images from C/C++ language-based reconstruction with MATLAB environment-based reconstruction results. This thesis documents the process
used to determine the efficacy of the proposed methodology.
Collections
- OU - Theses [2093]