Published ahead of print on April 8, 2009
doi: 10.3174/ajnr.A1557
Challenges of Using MR Spectroscopy to Detect Neural Progenitor Cells In Vivo
Z. Donga,b,
W. Dreherc,d,
D. Leibfritzc,d and
B.S. Petersona,b
a Department of Psychiatry, Columbia University, New York, NY
b MRI Unit, New York State Psychiatric Institute, New York, NY
c Fachbereich 2, Universität Bremen, Bremen, Germany
d Center for Advanced Imaging, Bremen, Germany
View larger version (18K):
[in this window]
[in a new window]
|
Fig 1. Synthesized spectra of proton MR spectroscopy with typical line shapes as seen in vivo and their spectral fitting estimated by using the method of SVD. The same noise was added to the synthesized signals in the time domain. A, An apparent single spectral peak (black) consisting of 3 Lorentzian peaks with zero phase (red, green, and blue) is estimated correctly when the 3 fitted components have identical phases. The complex combination of these components (red) coincides perfectly with the synthesized peak in black, as is true for all the other fittings (B–D). B, A single peak with a Gaussian line shape (black) is decomposed into 4 components, each with differing phases. Unlike the peaks identified in A, where the phases are identical, these components cannot be interpreted as valid peaks even though their combination constitutes a "perfect" fit. C, A single peak with an asymmetric line shape (black), simulating the distortion caused by inhomogeneity of the magnetic field, is decomposed into 5 components (colored), each with differing phases. None of the components can be regarded as a valid peak. D, The synthesized signal (black) consists of 2 components, 1 the same as that in C and the other its duplicate but reduced in amplitude by a factor of 5 and shifted in frequency by 10 Hz. The signal is decomposed into 6 components, each with differing phases (colored), none of which can be interpreted as valid. Freq indicates frequency.
| |
View larger version (30K):
[in this window]
[in a new window]
|
Fig 2. Shown are the synthesized in vivo 1H-MR spectroscopy spectrum (solid black), the overall fitting provided by SVD (dashed red), the individual components provided by SVD (solid color), and the residue (blue). The metabolites used in the simulation include N-acetylaspartate (NAA), N-acetyl aspartylglutamate (NAAG), total creatine (tCr), total choline (tCh), GABA, taurine (Tau), lactate (Lac), glutamine/glutamate (Glx), and myo-inositol (mI). The chemical shifts and J-couplings were taken from Govindaraju et al.20 Simulation parameters include the following: system frequency = 200 MHz, spectral width = 4006 Hz, TE = 20 ms, Lorentzian line width = 10 Hz, number of data points = 4096. The first 1024 data points of the FID were used in the SVD procedure. The number of spectral components was 24 (19 were in the range of 0–5 ppm), which was enforced by the user as prior knowledge. The overlapping lines of Cr and PCr, Ch and PCh, NAA and NAAG, and the doublets of Lac cannot be resolved and are instead detected as singlets. The phase error in the NAA singlet at 2.02 ppm is obviously caused by the spurious broad component in its vicinity. Two components with negative phases are detected around 2.3 ppm, clearly indicating that these components are not interpretable and other components in this region are not reliable. Increasing or reducing the number of spectral components as prior knowledge did alter the overall fitting and did not ameliorate the behavior of spectral components. Freq indicates frequency.
| |
