Three-Dimensional Muscle Motion Reconstruction Using Tagged MR Images

Xiaofeng Liu, Johns Hopkins University

For clinical and scientific studies, it is important to understand the internal muscle motion of tissues such as the tongue during speech and the heart in its beating cycle. Magnetic resonance (MR) tagging places non-invasive and temporary markers (tags) inside the soft tissues in a pre-specified pattern, yielding images that carry information about motion in the tagging direction. These images can be processing using harmonic phase (HARP) method to compute the in-plane motion. The dissertation studies the three-dimensional (3D) muscle motion using MR tagging with a focus on tongue imaging, and addresses the technical challenges in both 2D and 3D motion estimation.

In the dissertation, we developed HARP tracking refinement methods to reliably and automatically track the whole tissue from tagged MRI even when traditional HARP tracking fails. We measured 3D tongue motion during speech by re-implementing and optimizing the zHARP imaging sequence, and using a specialized MR triggering and vocal repetition method. We developed a thin plate spline based 3D tongue motion tracking method using tagged MR images by extending the 3D-HARP method for cardiac motion tracking. We developed a method to reconstruct a 3D, dense, incompressible deformation field from tagged MR images based on the divergence-free vector spline with incomplete data samples, and applied it to both tongue and cardiac motion reconstruction. Finally, we performed preliminary studies of the internal tongue motion pattern and muscle mechanisms of glossectomy patient and compared them with normal speakers.