MOTION CORRECTION
This project involves the development of techniques for measuring and correcting the effects of motion which can occur in many anatomic regions during the application of an MR imaging sequence. Typically, special "navigator echo" signals are used to monitor the motion. Various corrections can be performed using the navigator echo information. Post imaging corrections using algorithmic iterative corrections can also be performed with or without using navigator echo information. These techniques are being applied to a variety of anatomic regions, in particular to thoracoabdominal and musculoskeletal areas.
Navigator Echo Correction Technique
Shoulder imaging is a common application of musculoskeletal MR imaging, primarily based on its efficacy for evaluating rotator cuff disease. These examinations require high spatial resolution in order to adequately image structures such as the rotator cuff. Unfortunately, degredation of image quality due to motion is common. Adaptive motion correction using navigator echoes is effective in improving image quality in approximately three quarters of cases where motion is present.
Autocorrection without Navigator Echoes
In routine clinical practice, most MR imaging techniques do not employ special pulse sequences or reconstruction algorithms to correct for motion. As a consequence, motion present during imaging becomes an inherent feature of the imaging process. Autocorrection is a post processing algorithm developed to retrospectively remove motion from high-resolution shoulder images without the need for special pulse sequences.
Autocorrection using Prescan Navigator Echoes
One limitation of autocorrection is the large number of iterations required before an acceptable improvement in image quality can be acheived. A method for reducing the number of iterations of the autocorrection algorithm has been developed using the central slice theorem and modification to the imaging pulse sequence. A limited number of navigator echoes are acquired before the imaging sequence to obtain a "snap-shot" of the object, minimzing the influence of motion. An estimate is made of the motion induced phase shifts in the image, followed by autocorrection to eliminate errors in low signal to noise regions. This technique has resulted in as much as a six fold reduction in computation time compared to autocorrection alone. All detectable motion artifacts were eliminated using this technique on phantom and in-vivo data.
Autocorrection for MR Angiograms
MR angiography allows three-dimensional information to be acquired. However, these exams are prone to motion induced artifacts arising from the loss of the breath-hold and slow diaphram drift. Further study will investigate the feasibility of using autocorrection techniques to reduce motion artifacts.
Information on this page is taken from graduate work done by Kiaran McGee, Ph.D.
