---

Revision with unchanged content. Parallel Magnetic Resonance Imaging has had a significant impact in rapid and real time MR imaging. With multiple images captured within a fraction of a second, this method has facilitated real time cardiac imaging and intra operative imaging with ease. This method uses an array of coils with extremely localized sensitivities, thereby reducing the effective field of view that is unique to every coil element. Subsequent sub-sampling of the k space data obtained from each coil therefore results in a set of aliased images. Parallel imaging reconstruction then involves the restoration of the desired image by spatial re-arrangement of the aliased data in the image domain or by estimation of the full k space data with the help of the coil sensitivity profiles. Different reconstruction techniques have been proposed to regenerate the complete alias-free image. The reconstruction greatly depends on the coil sensitivity profiles and k space subsampling schemes. The results vary considerably according to the reconstruction method adopted. This work briefly describes the most popular image reconstruction techniques used in the field and provides a detailed analysis for the selection of the optimal method of image reconstruction based on various parallel imaging parameters and popular image quality markers such as SNR and artifact power.

---

Ph.D. Candidate, Department of Biomedical Engineering, Georgia Institute of Technology, Atlanta, GA, USA

---