Estimation of  in vivo γ-aminobutyric acid (GABA) levels in the neonatal brain
Moyoko Tomiyasu1,2, Noriko Aida3, Jun Shibasaki4, Katsutoshi Murata5, Keith Heberlein6, Mark A. Brown7, Eiji Shimizu2, Hiroshi Tsuji1, and Takayuki Obata1

1National Institute of Radiological Sciences, Chiba, Japan, 2Chiba University, Chiba, Japan, 3Department of Radiology, Kanagawa Children's Medical Center, Yokohama, Japan, 4Kanagawa Children's Medical Center, Yokohama, Japan, 5Siemens, Tokyo, Japan, 6Biomedical Imaging Technology Center, Burlington, MA, United States, 7University of Colorado, Cary, NC, United States


We examined in vivo brain γ-aminobutyric acid (GABA) levels of neonates and compared them with those of children. In this study, 32 normal neonates and 12 normal children (controls) had their brain GABA levels measured using clinical 3T edited-MRS. The neonates exhibited significantly lower GABA+ levels than the children in both the basal ganglia and cerebellum, which is consistent with previous in vitro data. While significantly higher GABA+/Cr levels were detected in the neonatal cerebellum, care should be taken when comparing GABA+/Cr levels between different ages. This is the first report about the in vivo brain GABA levels of neonates.


Gamma-aminobutyric acid (GABA) is considered to be the most prominent inhibitory neurotransmitter in the human brain. Although the in vivo brain GABA concentrations generated early in life are of great interest, they have not been elucidated. In vivo GABA signals can be detected using edited magnetic resonance spectroscopy (MRS), and creatine (Cr), N-acetylaspartate or water is usually used as a reference compound for GABA signal quantification [1]. However, the concentrations and T1/T2 values of these reference compounds change during development [2,3]. Therefore, such qualitative ways of determining neonatal GABA concentrations might provide inaccurate information.

The purpose of this study is to obtain accurate in vivo data about neonatal brain GABA levels in the basal ganglia (BG) and cerebellum, and to investigate the dependency of brain GABA levels on region and age. We demonstrate a method for obtaining data regarding neonatal brain GABA levels involving a clinical 3T MR scanner (Siemens, Germany), Gannet (a batch-analysis tool for GABA-edited MRS data), and in-house software. Since the GABA 3 ppm signal that appears in edited MRS spectra can contain macromolecule signals, it is referred to as GABA+. This is the first report about in vivo brain GABA levels of neonates.


Thirty-two normal neonates were enrolled (14 males; examined at a postconceptional age of 35-41-weeks), and 12 normal children (2 males; 6-16-years-old) were included as controls. The edited MRS data were obtained from the BG (2.6-7.6mL) and cerebellum (3.6-8.7mL) using the prototype MEGA-PRESS sequence (TE/TR: 69/1500ms; 128 averages (64x2)) [4]. Single-voxel MRS data were also obtained using the PRESS sequence (TE/TR: 30/5000ms) [5]. Using the Gannet software, the time domain data for GABA were processed into edited spectra. Then, using in-house software the peak areas of GABA+, Cr, and water were calculated by fitting Lorentzian line shapes to the data (Fig.1). To obtain normalized GABA+ levels, two different approaches were attempted: multiplying the peak area ratio of GABA+/water by the water concentration, and multiplying the peak area ratio of GABA+/Cr by the Cr concentration. The Cr concentrations were obtained from the PRESS MRS data based on the scaling of the unsuppressed water peak at water concentrations of 49.7M and 44.4M for neonates and children, respectively [6]. Corrections for T1 and T2 values were also applied. IBM SPSS statistics 23 (USA) was used for all statistical analyses. The Wilcoxon-signed-rank and Mann-Whitney U-test were used for the intra- and inter-group comparisons, respectively. P-values of less than 0.05 were considered to be significant.


In this study, 39 MR spectra for a total of 32 neonates were analyzed. The GABA+/water ratios of the neonates were significantly lower than those of the children in both the BG and cerebellum (p< 0.05, Fig.2-a1). Similarly, the GABA+/Cr ratios of the neonates were significantly lower than those of the children in the BG (p< 0.01). However, the GABA+/Cr ratios of the neonates were significantly higher than those of the children in the cerebellum (p< 0.02, Fig.2-b1). The neonates (8.8mM) and children (9.0mM) exhibited similar Cr concentrations in the BG. However, their cerebellar Cr concentrations differed significantly (p< 0.01; 7.1mM and 11.8mM in the neonates and children, respectively; Table1). The neonates' normalized GABA+ levels were significantly lower than those of the children in both brain regions, regardless of the method employed (p< 0.05, Fig.2-a2, b2).


The most important findings of this study were that neonates displayed significantly lower in vivo brain GABA+ levels than children in both the BG and cerebellum. Previous studies that examined the GABA concentrations of cerebrospinal fluid and postmortem brains also detected significant increases in GABA concentrations with age [7, 8]. In the present study, comparisons of the subjects' GABA/water and GABA+ levels detected similar percentage differences between the neonates and children (Fig.2-a). This is because the higher water concentrations and longer water T1/T2 values of the neonates, which are used as correction factors when calculating GABA+ levels, canceled each other out. As for the higher GABA+/Cr ratios detected in the neonatal cerebellum, the Cr concentrations in neonates are about 60% of those observed in children; therefore, dividing neonatal GABA+ concentrations by such lower Cr values might lead to higher GABA+/Cr levels (Fig.2-b).


Data regarding in vivo neonatal brain GABA+ levels were obtained using edited MRS. In both the BG and cerebellum, the GABA+ levels of neonates were significantly lower than those of children, which is consistent with previous in vitro data. When comparing GABA+/Cr levels between different age groups, care should be taken regarding the subjects' Cr concentrations. Further studies of age-related macromolecule contribution rates are required.


This work was supported by Japan Society for the Promotion of Science (JSPS) KAKENHI Grant Number 26461843.

The authors appreciate the help of Mr. M. Sato, Mr. K. Kusagiri, Y. Muramoto, and Y Suzuki, who acquired the MRI and MRS data, and thank Ms. A. Suzuki, Ms. K Sato, and Ms. H. Kamada for their assistance.


1. Puts, N.A. and R.A. Edden, In vivo magnetic resonance spectroscopy of GABA: a methodological review. Prog Nucl Magn Reson Spectrosc, 2012. 60: p. 29-41.

2. Kreis, R., et al., Brain metabolite composition during early human brain development as measured by quantitative in vivo 1H magnetic resonance spectroscopy. Magn Reson Med, 2002. 48: p. 949-958.

3. Tomiyasu, M., et al., Neonatal Brain Metabolite Concentrations: an In Vivo Magnetic Resonance Spectroscopy Study with a Clinical MR System at 3 Tesla. PLoS One, 2013. 8(11): p. e82746.

4. Mescher, M., et al., Simultaneous in vivo spectral editing and water suppression. NMR Biomed, 1998. 11(6): p. 266-72.

5. Bottomley, P.A., Spatial localization in NMR spectroscopy in vivo. Ann NY Acad Sci., 1987. 508: p. 333-348.

6. Williams, L.A., et al., Neonatal brain: regional variability of in vivo MR imaging relaxation rates at 3.0T - initial experience. Radiology, 2005. 235: p. 595-603.

7. Brooksbank, B.W., D.J. Atkinson, and R. Balazs, Biochemical development of the human brain. III. Benzodiazepine receptors, free gamma-aminobutyrate (GABA) and other amino acids. J Neurosci Res, 1982. 8(4): p. 581-94.

8. Casado, M., et al., Analysis of cerebrospinal fluid gamma-aminobutyric acid by capillary electrophoresis with laser-induced fluorescence detection. Electrophoresis, 2014. 35(8): p. 1181-7.


Figure 1. MRS peak quantification using in-house software.

During the calculation of the full width at half maximum (FWHM) of the Lorentzian line, the use of T2, magnetic field inhomogeneity (Inhomo), and an exponential line at 3Hz (as a window function) were considered (a). For the peak fitting, a first-order slope and the baseline offset were also considered (b, c).

Figure 2. Boxplots of in vivo brain GABA+ levels in the basal ganglia (BG) and cerebellum.

The GABA+ levels were normalized to GABA+/water (a1 to a2) or GABA+/Cr (b1 to b2). * < 0.05, ** < 0.01

Table 1. Cr concentrations.

Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)