PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Magn Reson Med. Author manuscript; available in PMC 2010 August 9.
Published in final edited form as:
PMCID: PMC2917980
NIHMSID: NIHMS217155

Metabolite Proton T2 Mapping In The Healthy Rhesus Macaque Brain At 3 T

Abstract

The structure and metabolism of the rhesus macaque brain, an advanced model for neurological diseases and their treatment response, is often studied non-invasively with MRI and 1H-MR spectroscopy (1H-MRS). Due to the shorter transverse relaxation time (T2) at the higher magnetic fields these studies favor, the echo-times used in 1H-MRS subject the metabolites to unknown T2-weighting, decreasing the accuracy of quantification which is key for inter- and intra-animal comparisons. To establish the “baseline” (healthy animal) T2 values we mapped them for the three main metabolites’ T2s at 3 T in four healthy rhesus macaques; and tested the hypotheses that their mean values are similar (i) among animals; and (ii) to analogs regions in the human brain. This was done with 3D multivoxel 1H-MRS at (0.6×0.6×0.5 cm)3=180μL spatial resolution over a 4.2×3.0×2.0=25 cm3 (~30%) of the macaque brain in a two point protocol that optimizes T2 precision per unit time. The estimated T2s in several gray and white matter regions are all within 10% of those reported in the human brain (mean±standard error of the mean): N-acetylaspartate=316±7, creatine=177±3 and choline=264±9 ms, with no significant gray versus white matter differences.

Keywords: Animal Models, Brain Metabolites, High Magnetic Field, MR Spectroscopic Imaging, Rhesus Macaque, Transverse Relaxation Time

Introduction

The biochemical, morphological and functional similarities between the human brain and its non-human primate counterpart has lead to extensive use of the latter as an advanced model for studies of neurological diseases and their treatment (14). Due to cost, complexity and frequent serial nature, these studies often favor non-destructive modalities: MRI for morphology and function and proton MR spectroscopy (1H-MRS) to probe the metabolism. The latter can monitor neuronal cells, cell energetics and membrane turnover through quantitative assessment of the N-acetylaspartate (NAA), creatine (Cr) and choline (Cho) - their surrogate markers (5,6).

Quantitative 1H-MRS requires accounting for various parameters that influence the signals of these metabolites. Some, e.g., the echo- and repetition-times (TE and TR) or the voxel size, are always known. Other instrumental factors, such as the static (B0) and radio-frequency (RF, B1) fields inhomogeneity can be estimated by field mapping (7) and internal water referencing (8). The molecular environment however, requires either knowledge of the longitudinal (T1) and transverse (T2) relaxation times (5); or to minimize their influence with long TR » T1 and short TE«T2. Since the flatter baseline, simpler peak structure and milder demands on the hardware of longer-TE spectra are often preferred (5), the ability to distinguish true metabolite levels from mere changes in relaxation times (quantification accuracy) depends on knowledge of T2 (9,10).

This knowledge is especially pertinent to animal models since the stronger B0s that enable the higher spatial resolution needed for their smaller brains come at a cost of shorter T2, exacerbating the weighting per given TE (11). To establish the “baseline” T2 values (of healthy animals) in order to correct for this weighting, we measured the T2s of NAA, Cho and Cr at 3 T in several brain regions of four healthy rhesus macaques in order to test the following two hypotheses: (i) That these regional T2s are similar among animals, i.e., are characteristic; and (ii) that they are also similar to analogous regions of the human brain, rendering macaques good model systems with respect to T2. To maximize spatial coverage and T2 precision this was done with three-dimensional (3D) 1H-MRS at 180μL spatial resolution in over 30% of that brain using a two-point protocol optimized for T2 precision per unit time (6,12).

Materials and Methods

Non Human Primates

Four healthy female adult (3 – 5 years old, 3.9 – 6 kg weight) rhesus macaques (Macaca mulatta) were studied. Each was tranquilized with 15 – 20 mg/kg intra-muscular ketamine hydrochloride and intubated to ensure a patent airway during the experiment (no mechanical ventilation was needed). Intravenous injection of 0.4 mg/kg atropine was administered to prevent bradycardia. Continuous infusion of 0.25 mg/kg/min propofol was maintained throughout via catheter in a saphenous vein. Heart rate, oxygen saturation, end-tidal CO2, respiratory rate and temperature were monitored continuously and a water blanket used to prevent hypothermia. Animals were under constant veterinary supervision and the protocol was approved by the Harvard Medical School and Massachusetts General Hospital IACUCs.

Instrumentation and MRI

All experiments were done in a 3 T imager (Magnetom TIM Trio, Siemens AG, Erlangen, Germany) with its standard circularly-polarized transmit-receive knee coil. To guide the 1H-MRS volume of interest (VOI) sagittal, coronal and axial turbo spin echo MRI (TE/TR=15/10000 ms, 512×512 matrix, 140×140 mm2 field of view (FOV), 1 mm slice thickness) were acquired.

1H-MRS

A 4.2×3.0×2.0 cm3 anterior-posterior × left-right × inferior-superior (AP×LR×IS) VOI was image-guided onto the anatomy of interest and aligned along the splenium-genu axis of the corpus callosum, as shown in Fig. 1. The manufacturer’s automatic shim procedure adjusted the first and second order currents to consistent 26±1 Hz full-width-at-half-maximum (FWHM) VOI water linewidth. The VOI was excited using PRESS (TR=1820 ms, TEs see below) with two second-order Hadamard encoded slabs (4 slices) interleaved within every TR for optimal signal-to-noise-ratio (SNR) and duty-cycle (13). Interleaving also allowed us to apply a very strong, 18 mT/m, Hadamard slice select gradient to reduce the chemical shift displacement between NAA and Cho to 0.2 mm, 4% of the slice width (14). The 9.2 ms numerically-optimized PRESS 180°s were pulsed under 3.2 and 4.4 mT/m (AP×LR), leading to ~1 mm displacement (15%) at the VOI edge. The 4 slices were 16×16 chemical-shift-imaging (CSI) encoded over a 96×96 mm2 (AP×LR) FOV to yield 140 voxels (0.6×0.6×0.5 cm)3 each in the VOI. The MRS signals were digitized for 256 ms with 512 points at ±1 kHz bandwidth. The 16×16×4 scan took 15.4 min.

Fig. 1
Left: Axial (a) sagittal (b) and coronal (c) T2-weighted MRI and position of the 4.2×3.0×2.0 cm3 VOI (frame) and Hadamard slices (arrows) in the ~80 cm3 macaque brain.

Choice of TE and acquisition strategy

The conflicting desire for high spatial resolution and T2 precision in noisy MRS experiments makes for long acquisitions. To maximize their efficiency we employed two strategies: (i) 3D 1H-MRS to cover a larger volume at the same SNR per given time as single-voxel methods (13); and (ii) a two-point acquisition scheme that adjust the two TEs and the number of averages, N1 and N2, at each for the best T2 precision per unit-time (12,15). Using T20=185 as the “initial guess” for the T2s led to TE1=30 ms (minimum in our setup), N1=4, TE2 = TE1 + 1.25T×T20 = 260 ms, N2=12, resulting in a 4 hour protocol: 62 min. at TE1 and 184 at TE2. (12). The choice of T20 = 185 ms, just above the T2=172 ms obtained for Cr in the human brain with this method (16), was influenced by: (i) the error in the resultant T2s that are worse for values below than above it; and (ii) the error remains similar over −25% to +40% about T20, as shown in Fig. 2 of Ref. (9).

1H-MRS Post-processing

The 1H-MRS data were processed off-line with our in-house software. The data was voxel-shifted to align the CSI grid with the NAA VOI, Fourier transformed in the time, LR and AP directions and Hadamard transformed in the IS dimension. Automatic frequency and zero-order phase corrections were done in reference to the NAA peak in each voxel. The relative i-th metabolite (i=NAA, Cr or Cho) levels, S1i at TE1 and S2i at TE2 were estimated from their peaks’ area using the SITools-FITT parametric modeling and least-squares optimization software (17). The models comprised NAA, Cho, Cr, glutamate, glutamine, taurine and myo-inositol at TE1=30 ms, and just NAA, Cr and Cho at TE2=260 ms.

Metabolite T2 determination

Proton T2i relaxation times of the =NAA, Cr or Cho were assessed in vivo in each voxel using,

T2i=(TE2TE1)/ln(S1i/S2i),
[1]

Six different structures were examined: caudate, thalamus, putamen and cingulate gyrus in gray matter (GM); splenium of the corpus callosum and centrum semiovale in white matter (WM). Each was outlined manually on the axial MRI (cf. Fig. 2), our software then zero-filled the MRS to 256×256 and averaged the T2s in all voxels that fell entirely or partially within the outline. [Although zero-fill adds no new information to the data, it provides overlapping voxels that can increase the effective spatial resolution and reduce GM/WM partial volume (18)].

Fig. 2
Left: Axial T2-weighted images showing the regions where the voxels’ T2 were averaged (solid lines). In GM: cingulate gyrus (a) caudate (d), putamen (e) and thalamus (f); and in the WM: (b) centrum semiovale, (c) splenium of the corpus callosum. ...

Results

The SNR and spectral resolution obtained from the 180μL voxels at both TEs, are shown in Fig. 1. Voxel SNRs (defined as peak-height divided by the root-mean-square of the noise) were 29.7±4.0, 23.4±4.0 and 14.8±4.8 for NAA, Cr and Cho [mean±standard deviation (SD)] at TE1. The metabolites’ linewidth, Δω, of 4.0±0.2 Hz at TE1 and 3.9±0.1 Hz FWHM at TE2 represent T2*s (=1/πΔω assuming Lorentzian shapes) of 80±4 and 82±2 ms. The scaling of these Δω from the 26±1 Hz in the VOI approximately as the cube root of its ratio to the voxel volume (25 cm3/180μL) indicates that macroscopic susceptibility dominates T2* in the former (19).

Examples of the short and long TE spectra shown in Figs. 1 and and22 demonstrate the substantial T2-weighting incurred at longer TEs. The mean± standard error of the mean (SEM) global T2s among in the 560 voxels in the VOIs of the four macaques were: NAA=316±7, Cr=177±3 and Cho=264±9 ms. These are all within 10% of the average corresponding values reported recently in the human brain (9,10,16). Their histograms from all 140 voxels in each of the four macaques, are shown in Fig. 3. The histograms overall similar shape characterized by peak position and half-width-at-half-height (mean±SEM): NAA=79±2, Cr=32±1 and Cho=71±2 ms, indicate excellent inter-animal reproducibility. The NAA, Cr and Cho T2s from the regions shown in Fig. 2 are compiled in Table 1, showing similar GM and WM values in these structures.

Fig. 3
Histograms of the NAA, Cr and Cho T2s from all 140 voxels in the VOI for each of the four macaques. Note the inter-animal similarity of the histograms for each metabolite, indicating that: (i) a global T2 value for each would lead to less than 15% variation ...
Table 1
Mean±SEM proton T2 relaxation times (milliseconds) at 3 T of the N-acetylaspartate (NAA), creatine (Cr) and choline (Cho) in the various GM and WM brain regions shown in Fig. 2, from the N=4 macaques studied.

Discussion

Metabolite T2s are essential to determine optimal acquisition protocols (13), and for reliable quantification at intermediate and long TEs (5,10). It is surprising, therefore, that despite studies showing in vivo T2s to shorten as B0 is increased (11,20), only few report their values or spatial variation at 3 T, the current de facto clinical high field. Moreover, despite the growing use of non-human primates as models for neurological disorders and treatment (14), their metabolites T2s or their inter- and intra-animal variations at 3 T, were hereto unknown. This limited both the accuracy of quantification as well as optimal comparisons between human and macaque 1H-MRS observable metabolism - the underlying motivation for the use of animal models.

Combining higher field 3D 1H-MRS with an efficient two-point protocol enabled us to map the T2s at 180μL spatial resolution and at the optimal precision for the given measurement time over an extensive, 25 cm3 (≈30%) portion of the ~80 cm3 macaque brain in vivo (6,12). This allowed us to establish the global and regional T2s values in several animals in order to ascertain: (i) the inter-animal reproducibility and spatial variations; and (ii) their similarity to the respective values in analogous regions of the human brain at the same magnetic field strength.

The inter-animal reproducibility is reflected by the similar histograms’ peak positions and shape (see Results and Fig. 3) as well as the less than 5% SEM of the T2s for each metabolite. They combine to support the first hypothesis: that these mean T2s may be characteristic of healthy macaques. The less than 10% variation in T2 among different GM or WM regions for NAA, Cr and Cho (cf. Table 1) will lead to under 7% metabolite level variations when obtained with short or intermediate (TEs 144 ms) 1H-MRS at this B0 [See Appendix in Ref. (9)]. It is noteworthy that even for larger variations, of the order of the ~20% half-widths of the histograms in Fig. 3, use of the mean T2 would lead to a quantification error of less than 15% for Cr and under 7% for NAA at TE=144 ms. These variations will decrease proportionally at shorter TEs.

The few percent differences between the T2s estimated here and the 265–343, 152–172 and 200–248 ms ranges reported for NAA, Cr and Cho in the human brain at 3 T (9,10,16,21,22), also validate the second hypothesis. They establish a “threshold” necessary to consider the healthy macaque brain as a model for its human counterpart with respect to this “metabolic environment” parameter. In addition, the less than 10% T2 variations between WM and GM in these animals indicates that for the purpose of metabolic quantification use of the mean value for each metabolite is sufficient as long as TE < T2.

Although time and cost preclude a test-retest study to quantify intra-animal T2 reproducibility (precision), their coefficient of variation (CV)=SD/mean can nevertheless be estimated from the intra-animal performance. Given the 3.6/SNR16 theoretical limit on the optimal CV obtainable with this approach [achieved when T2[equivalent]T20, as shown in Fig. 2 of (9); and SNR16 is that obtained if all the time is spent for 4+12=16 averages at TE1], places a lower limit on the instrumental noise contribution, CVinst (9,12). Since our T20=185 ms is close to the Cr T2, its SNR4=23.4 at TE1 extrapolates to SNR16=SNR4×(16/4)½=46.8 that limits its CVinst to 3.6/SNR16≈8%.

The actual upper limit on CVinst and the variations that are due to the biological noise, CVbiol, combined, CVtot, is estimated from the histograms in Fig. 3 at ~20%. Assuming that these two noise sources are independent and therefore, add in quadrature:

CV2tot=CV2biol+CV2inst,
[2]

leads to an intra-animal voxel T2 CVbiol of ~18%. Averaging adjacent voxels can improve CVtot via two mechanisms: First, it will increase the SNR (compare the voxel spectra in Fig. 1 with regional averages in Fig. 2), i.e., reduce CVinst. Second, the regional variations in CVbiol will be smaller than in the whole VOI. The confluence of both effects is reflected in the 3 – 7% regional CVtot in Table 1, that are sufficiently small for the T2s to be considered characteristic and suggest that more T2 measurements on healthy animals, e.g., at each study’s inception, are unnecessary in that they are unlikely to yield significantly different results.

Finally, we note that although more metabolites can be quantified at the short TE, e.g., the myo-inositol, glutamine and glutamate (see Fig. 2), we focused only on the T2s of the three major singlets. This is due to our choice of T20=185 ms for optimal precision over the 140–260 ms range (that suit Cr and Cho best) but which also requires a 230 ms delay between TE1 and TE2. During the consequent TE=260 ms, all shorter T2 (and J-coupled) species signals decay (or dephase) severely, as shown in Fig. 2 (23). The relationship between the precision and SNR (12), and an already long protocol motivated us to optimize for just one cluster of T2 values.

Conclusion

Combining 3D 1H-MRS with an optimized two-point acquisition protocol makes for the most efficient use of time to map the T2 distribution of some of the observable neuro-metabolites in rhesus macaques. These regional T2 values are similar among animals indicating that they may be characteristic and are also in excellent agreement with reported values in humans at 3 T, as expected from a model-system. The T2s and their variations between WM and GM structures indicate that for the purpose of metabolic quantification at 3 T the use of one T2 value for each metabolite is sufficient for intermediate or long TE 1H-MRS across both humans and macaques.

Acknowledgments

We thank Dr. Joanne Morris and Ms. Shannon Luboyeski for animal veterinary care. This work was supported by NIH Grants EB01015, NS050520, NS050041, NS051129, NS059331, AI028691 and NS040237. The MGH A.A. Martinos Center for Biomedical Imaging is also supported by National Center for Research Resources grant number P41RR14075 and the Mental Illness and Neuroscience Discovery (MIND) Institute.

References

1. Lentz MR, Westmoreland SV, Lee V, Ratai EM, Halpern EF, Gonzalez RG. Metabolic markers of neuronal injury correlate with SIV CNS disease severity and inoculum in the macaque model of neuroAIDS. Magn Reson Med. 2008;59(3):475–484. [PMC free article] [PubMed]
2. Liu Y, D’Arceuil H, He J, Duggan M, Gonzalez G, Pryor J, de Crespigny A. MRI of spontaneous fluctuations after acute cerebral ischemia in nonhuman primates. J Magn Reson Imaging. 2007;26(4):1112–1116. [PubMed]
3. Garcia-Cabezas MA, Rico B, Sanchez-Gonzalez MA, Cavada C. Distribution of the dopamine innervation in the macaque and human thalamus. Neuroimage. 2007;34(3):965–984. [PubMed]
4. Yang SH, Cheng PH, Banta H, Piotrowska-Nitsche K, Yang JJ, Cheng EC, Snyder B, Larkin K, Liu J, Orkin J, Fang ZH, Smith Y, Bachevalier J, Zola SM, Li SH, Li XJ, Chan AW. Towards a transgenic model of Huntington’s disease in a non-human primate. Nature. 2008;453(7197):921–924. [PMC free article] [PubMed]
5. Jansen JF, Backes WH, Nicolay K, Kooi ME. 1H MR spectroscopy of the brain: absolute quantification of metabolites. Radiology. 2006;240(2):318–332. [PubMed]
6. Juchem C, Merkle H, Schick F, Logothetis N, Pfeuffer J. Region and volume dependencies in spectral line width assessed by 1H 2D MR chemical shift imaging in the monkey brain at 7 T. Magn Reson Imaging. 2004;22(10):1373–1383. [PubMed]
7. Gruetter R, Tkac I. Field mapping without reference scan using asymmetric echo-planar techniques. Magn Reson Med. 2000;43(2):319–323. [PubMed]
8. Dong Z, Dreher W, Leibfritz D. Toward quantitative short-echo-time in vivo proton MR spectroscopy without water suppression. Magn Reson Med. 2006;55(6):1441–1446. [PubMed]
9. Zaaraoui W, Fleysher L, Fleysher R, Liu S, Soher BJ, Gonen O. Human brain-structure resolved T(2) relaxation times of proton metabolites at 3 Tesla. Magn Reson Med. 2007;57(6):983–989. [PubMed]
10. Tsai SY, Posse S, Lin YR, Ko CW, Otazo R, Chung HW, Lin FH. Fast mapping of the T2 relaxation time of cerebral metabolites using proton echo-planar spectroscopic imaging (PEPSI) Magn Reson Med. 2007;57(5):859–865. [PubMed]
11. Michaeli S, Garwood M, Zhu XH, DelaBarre L, Andersen P, Adriany G, Merkle H, Ugurbil K, Chen W. Proton T2 relaxation study of water, N-acetylaspartate, and creatine in human brain using Hahn and Carr-Purcell spin echoes at 4T and 7T. Magn Reson Med. 2002;47(4):629–633. [PubMed]
12. Fleysher L, Fleysher R, Liu S, Zaaraoui W, Gonen O. Optimizing the precision-per-unit-time of quantitative MR metrics: Examples for T(1), T(2), and DTI. Magn Reson Med. 2007;57(2):380–387. [PubMed]
13. Goelman G, Liu S, Hess D, Gonen O. Optimizing the efficiency of high-field multivoxel spectroscopic imaging by multiplexing in space and time. Magn Reson Med. 2006;56(1):34–40. [PubMed]
14. Goelman G, Liu S, Fleysher R, Fleysher L, Grossman RI, Gonen O. Chemical-shift artifact reduction in Hadamard-encoded MR spectroscopic imaging at high (3T and 7T) magnetic fields. Magn Reson Med. 2007;58(1):167–173. [PubMed]
15. Fleysher R, Fleysher L, Gonen O. The optimal MR acquisition strategy for exponential decay constants estimation. Magn Reson Imaging. 2008;26(3):433–435. [PMC free article] [PubMed]
16. Kirov I, Fleysher L, Fleysher R, Patil V, Liu S, Gonen O. Age dependence of regional proton metabolites T2 relaxation times in the human brain at 3 T. Magn Reson Med. 2008;60(4):790–795. [PMC free article] [PubMed]
17. Soher BJ, Young K, Govindaraju V, Maudsley AA. Automated spectral analysis III: application to in vivo proton MR spectroscopy and spectroscopic imaging. Magn Reson Med. 1998;40(6):822–831. [PubMed]
18. Bernstein MA, Fain SB, Riederer SJ. Effect of windowing and zero-filled reconstruction of MRI data on spatial resolution and acquisition strategy. J Magn Reson Imaging. 2001;14(3):270–280. [PubMed]
19. Li BS, Regal J, Gonen O. SNR versus resolution in 3D 1H MRS of the human brain at high magnetic fields. Magn Reson Med. 2001;46(6):1049–1053. [PubMed]
20. Otazo R, Mueller B, Ugurbil K, Wald L, Posse S. Signal-to-noise ratio and spectral linewidth improvements between 1.5 and 7 Tesla in proton echo-planar spectroscopic imaging. Magn Reson Med. 2006;56(6):1200–1210. [PubMed]
21. Traber F, Block W, Lamerichs R, Gieseke J, Schild HH. 1H metabolite relaxation times at 3.0 tesla: Measurements of T1 and T2 values in normal brain and determination of regional differences in transverse relaxation. J Magn Reson Imaging. 2004;19(5):537–545. [PubMed]
22. Mlynarik V, Gruber S, Moser E. Proton T (1) and T (2) relaxation times of human brain metabolites at 3 Tesla. NMR Biomed. 2001;14(5):325–331. [PubMed]
23. Choi C, Coupland NJ, Bhardwaj PP, Kalra S, Casault CA, Reid K, Allen PS. T(2) measurement and quantification of glutamate in human brain in vivo. Magn Reson Med. 2006;56(5):971–977. [PubMed]