During the second year of the project, we have developed a sampling technique by introducing the variable repetition time approach and combine it with the variable averaging approach we introduced previously. In this past year, we have introduced a method that utilized echo-planar encoding (a very fast spatial encoding technique) to sample the outer portions of the k-space. This method has been demonstrated in phantoms studies and has made the uneven sampling scheme we introduced previously more efficient. In addition to the development of techniques for spatial localization based on CSI, we have also worked on another aspect of metabolite imaging, spectral fitting. Current techniques for spectral fitting are not very robust when the data are low in signal-to-noise ratio and are corrupted by other distortions. Our recent work has adapted a genetic algorithm for spectral fitting which works consistently well. This technique has the advantage of flexibility in line shape modeling, ease of incorporating prior information, and the ability to achieve global optimization. Metabolite images were generated with this technique.
Showing the most recent 10 out of 493 publications
