Detection and correction of patient motion in magnetic resonance imaging
Abstract/Contents
- Abstract
- Due to the sequential nature of magnetic resonance imaging (MRI) data acquisition, correction of image artifacts originating from involuntary patient motion is essential for reliable diagnostic quality. Especially during MRI scans of certain patient populations such as children, elderly or patients with certain medical conditions (e.g. stroke, Parkinson's), motion correction methods must be incorporated into the MR imaging protocol for adequate image quality. With increased demands for higher resolution or time-resolved examinations (e.g. functional MRI), examination times also increase and even willing patients might have trouble staying still during the course of the examination. The first part of this thesis provides a method for retrospective correction of head motion artifacts using a multi-shot spiral-in \& out readout and parallel-imaging based iterative image reconstruction. The spiral-in part provided a low resolution image that was used for measurement of head motion. Due to rotational motion, locally undersampled areas appear in MR acquisition space (i.e., k-space), which violate the Nyquist theorem and cause artifacts even after motion correction. These artifacts were addressed through the data redundancy provided by multiple receiver channels that is present in modern receiver coils and an iterative conjugate-gradient based reconstruction. This method was then applied to diffusion tensor imaging (DTI) with multi-shot readout. Since DTI uses directional gradients to encode diffusion, rotational motion causes the image contrast to change, and it becomes incorrect to combine data with varying diffusion encodings on them. To address this issue, a non-linear conjugate gradient based reconstruction is presented and it is shown that this method provided more accurate description of white matter pathways compared to traditional methods. In the second part of this thesis, a prospective motion correction system using an optical tracking device is presented. Such systems are preferable compared to retrospective navigator-based methods due to various reasons, such as ability to perform motion correction independent of the MR data acquisition and immunity to spin history effects. The system proposed used a single camera mounted on the head coil and a self-encoded checkerboard marker mounted on the patient's forehead. Results on structural and diffusion imaging revealed that prospective motion correction outperforms retrospective navigator-based schemes. In the last part of the thesis, entropy-based retrospective autofocusing was used in combination with motion data obtained from prospective tracking to remove residual motion artifacts in the images. This method was especially useful for removing errors caused by inaccurate cross-calibration between the scanner and camera frame-of-references. It was also shown that prospective tracking can be the enabling technology for autofocusing in 3D acquisitions.
Description
| Type of resource | text |
|---|---|
| Form | electronic; electronic resource; remote |
| Extent | 1 online resource. |
| Publication date | 2010 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Associated with | Aksoy, Murat |
|---|---|
| Associated with | Stanford University, Department of Electrical Engineering |
| Primary advisor | Bammer, R. (Roland) |
| Primary advisor | Pauly, John (John M.) |
| Thesis advisor | Bammer, R. (Roland) |
| Thesis advisor | Pauly, John (John M.) |
| Thesis advisor | Moseley, Michael E. (Michael Eugene), 1951- |
| Advisor | Moseley, Michael E. (Michael Eugene), 1951- |
Subjects
| Genre | Theses |
|---|
Bibliographic information
| Statement of responsibility | Murat Aksoy. |
|---|---|
| Note | Submitted to the Department of Electrical Engineering. |
| Thesis | Thesis (Ph. D.)--Stanford University, 2010. |
| Location | electronic resource |
Access conditions
- Copyright
- © 2010 by Murat Aksoy
- License
- This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).
Also listed in
Loading usage metrics...