University of Illinois Urbana-Champaign

Reconstruction of out-of-FOV lipid signals from sparse data for partial-brain proton magnetic resonance spectroscopic imaging

Xiong, Jiahui

This item is only available for download by members of the University of Illinois community. Students, faculty, and staff at the U of I may log in with your NetID and password to view the item. If you are trying to access an Illinois-restricted dissertation or thesis, you can request a copy through your library's Inter-Library Loan office or purchase a copy directly from ProQuest.

Permalink

https://hdl.handle.net/2142/104047

Description

Title
Reconstruction of out-of-FOV lipid signals from sparse data for partial-brain proton magnetic resonance spectroscopic imaging
Author(s)
Xiong, Jiahui
Contributor(s)
Liang, Zhi-Pei
Issue Date
2019-05
Keyword(s)
  • 1H spectroscopic imaging
  • out-of-FOV lipid removal
  • sparse sampling
  • partial separability
Abstract
Motivation— Acquisition of partial-brain 1H-MRSI dataset requires the application of slice-selective RF pulses, which are never perfect in practice. As a result, brain regions outside the region of interest, such as the lipid-rich subcutaneous layer around the skull, can be excited. Consequently, for a k-space sampling scheme that produces the desired FOV, any out-of-FOV lipid signals can alias back into the FOV and contaminate signals within the FOV. Since 1H-MRSI deals with low-concentration metabolites, these aliased lipids can dominate over metabolite signals of interest and severely interfere with spectral quantification of metabolite levels. One method that is currently used is to do an oversampling along the aliased direction to double the FOV so that accidentally excited lipids do not fold back. This approach, however, necessarily lengthens the scan time and thus is not desirable for clinical applications. Therefore, the goal of this work is to address the lipid aliasing problem from a post-processing point of view, reconstructing and then removing out-of-FOV lipids from a partial-brain 1H-MRSI dataset to make spectral quantification of brain metabolite levels more reliable. Due to time limitation, however, this thesis only focuses on the reconstruction stage of the project. Approach— Instead of trying to reconstruct any folded out-of-FOV lipids directly from the partial-brain dataset, we propose to acquire an auxiliary dataset to help us with the process. To achieve the desired spatiotemporal resolution for this auxiliary dataset but with a trivial addition on the total scan time, we further propose to utilize a partial separability model that allows us to reconstruct the auxiliary signals from highly sparsely-sampled (𝒌, 𝑡)-space data. Results— The proposed method to reconstruct out-of-FOV lipids for the auxiliary dataset from sparsely sampled (𝒌, 𝑡)-space has been evaluated by a simulation study. The results show that the proposed approach can reconstruct lipid signals that have been down-sampled by a factor of nearly eight.
Type of Resource
text
Language
en
Permalink
http://hdl.handle.net/2142/104047

Owning Collections

Senior Theses - Electrical and Computer Engineering PRIMARY
The best of ECE undergraduate research