Program Information
Human Brain GABA J-Difference Editing Spectral Quantification with the Fast Pade Transform
J Zhang*, R Woods, J Alger, A Thomas, R Espinoza, K Narr, UCLA, Los Angeles, California
TU-G-134-9 Tuesday 4:30PM - 6:00PM Room: 134Purpose: To quantify in vivo GABA concentration levels using a novel parametric proton magnetic resonance signal quantification algorithm called the Fast Pade Transform.
Methods: Single voxel MEGA-PRESS data were collected on a Siemens 3T Tim Trio system using 12-channel phased-array head coil (TE=68 ms, TR=2000 ms, Voxel size=30x30x30 mm3, Vector size=1024, NEX=64). Both water reference and water suppressed spectra were obtained. A total of three phantoms (containing NAA, Cr, Cho, GABA, and Lac) were created with different GABA concentration levels of 1 mMol, 2 mMol and 3 mMol. In vivo data were collected from the anterior cingulate cortex region in 5 human volunteers. The fast Pade transform algorithm was implemented in MATLAB. Spectral data were processed with this FPT program to obtain parametric information such as frequency, line-width, phase and amplitude. GABA concentration levels were calculated using the ratio of metabolite and water amplitude information, since signal amplitudes are directly proportional to the amount of protons present in the acquisition volume.
Results: The results suggest that the Fast Pade Transform algorithm measured GABA concentrations in phantoms are very close to the actual known values. And this method is capable of quantifying in vivo MEGA-PRESS proton spectra in 5 human brain data with a mean measurement of 2.01 mMol, which agrees with literature reported human brain in vivo GABA concentration level.
Conclusion: This work demonstrates the potential of the fast Pade transform technique to quantify the GABA C-4 doublet resonance signals with the MEGA-PRESS acquisition. We plan to test this method with more in vivo data as well as with other MRS acquisition techniques in the future.
Funding Support, Disclosures, and Conflict of Interest: NIH funding - R01MH092301
Contact Email: