Multidimensional motion correction in MRI

点击翻译按钮获取中文摘要

BACKGROUND: The availability of high field strength magnetic resonance imaging (MRI) scanners is driving resolution to higher levels. However subject motion can often negate the benefits of increasing field strength by blurring images or introducing artifacts. Degradation in image quality is a result of the motion itself and the change in local magnetic field uniformity resulting from the motion. Therefore methods for rapid measurement of motion in six-degrees of freedom (DoF) and rapid generation of a magnetic field (B0) map, which can be used for active shimming, are required. An approach to detect bulk motion is the navigator-echo method, which measures motion from the MR signal acquired in k-space. The geometric features of a navigator k-space trajectory and the motion extraction algorithms are two key aspects in developing navigator-based techniques. We have developed acquisition and computational algorithms to determine motion in 6 DoF using a spherical navigator (SNAV) rapidly, by registering the navigator data acquired at the moved position with navigator templates collected at a reference position. For real time shimming following the detection of large motion, a need exists for rapid methods to measure the local magnetic field in 3D. At high field strength such measurements are often difficult because of numerous wraps of the phase images used to generate the field map. We have developed robust and computationally efficient phase unwrapping and B0 mapping methods, enabling accurate and rapid B0 mapping. OBJECTIVE: We propose to develop efficient and robust implementations of our SNAV technique to provide 6 DoF motion determination during MR imaging at high field strengths along with real time shimming. Technical developments and experimental evaluation will focus on combining the novel navigator technique with real time field mapping to compensate for involuntary and physiological motions at 3T and 7T. APPROACH: The program will focus on: • Optimization of the accelerated SNAV trajectories • Implementation of real-time B0 field mapping technique • Modeling of the changes of B0 for different head motions and incorporating these models, along with the real time maps, into an active shimming algorithm • Integration of the novel shimming and motion correction techniques within ultra-high resolution pulse sequences for brain imaging and optimization of timing parameters to allow on-the fly motion correction • Optimization of analysis techniques to minimize processing time, while expanding the range of motion that can be tracked • Evaluation in purpose-built dynamic models SIGNIFICANCE: The success of this work will lead to the ability to monitor and correct for 6 DoF motion in real time using an MR-based technique, while also performing active shimming based on the motion measurements made. The proposed technical developments promises to have immediate impact on MR imaging.