Magnetic Resonance Imaging is one of most important imaging technique, however, it still needs a lot of improvement to enhance the quality of the image and reduce time. In this report, the issue that has been focus on is who to reduce time of reconstruction while maintaining or improving the image feature.  There so many methods of reconstructing an image of MRI such as GRAPPA, SENSE and SMASH. Two of them were used to be improved are GRAPPA (Generalized Auto-calibrating Partially Parallel Acquisitions)  and SENSE(Sensitivity Encoding). Deep research on MRI has helped to develop various image reconstruction techniques with a goal of minimize the scan time while keeping or increasing the image quality. Some of these techniques are affected negatively by the technical and physiologic problems or artifacts that happen when magnetic field gradients are switched in a short time. For example, peripheral nerve stimulation can be caused by the switched magnetic field gradients. However, these limitations or problems can be solved by some of the proposed parallel imaging techniques. Parallel imaging is the most efficient and newest imaging technique and is counted to be the most effective. SENSE (Sensitivity Encoding), GRAPPA (Generalized Auto-calibrating Partially Parallel Acquisitions) and NL-GRAPPA (Nonlinear GRAPPA) are between many other common parallel imaging techniques. However, parallel imaging still has some problems when reconstructing images and therefore it requires more improvement and research.

This paper aims to study the image reconstruction of two of those common methods, NL-GRAPPA & SENSE, as images for both were produced and comparison to other methods. In this thesis we were able to successfully reconstruct NL-GRAPPA images and partially completing SENSE using MATLAB for both methods.