Introduction

The inhibitory γ-aminobutyric acid (GABA) and the excitatory glutamate (Glu) are the two major neurotransmitters in the human brain. Both are essential for normal physiological and mental functionality [1] and altered levels of GABA have been associated with numerous neurological and neuropsychiatric disorders [2]. MR spectroscopy offers the only non-invasive technique for measuring in vivo GABA and Glu levels. MEGA-editing is the most popular method for detecting GABA and, to date, studies have mainly been conducted in single voxels within the brain, with cortical regions being prominent regions of interest [3]. MRSI reports are still uncommon [4,5], and all of them share the same limitations, namely, large voxel sizes, which hamper the assessment of anatomically resolved GABA levels, and also restricted rectangular target volumes to overcome extracranial lipid artifacts, which is problematic for studies of cortical regions. We propose a MEGA-edited acquisition scheme that exploits the advantages of edited-MRSI at 7T, while tackling challenges that arise with ultra-high-field. Adiabatic slice selection and MM-suppressed MEGA-editing, combined with fast non-Cartesian readout and real-time motion and B₀ corrections ensure stability against both volunteer-related and typical ultra-high-field artifacts.

Methods

Phantom and in vivo measurements (n=5) were performed on a Siemens Magnetom 7T whole-body MR scanner using a 32-channel receive coil. Spatial localization was achieved by a B₁⁺- and CSDE-insensitive 1D-semiLASER sequence [6] employing a 900us slice-selective SINC excitation pulse and one pair of adiabatic GOIA pulses (W16,4 modulation, 8ms duration, 10kHz bandwidth) refocusing a 16mm slice (Fig.1a). Hamming density-weighted, concentric ring trajectories [7] were used for data sampling thereby optimizing the SNR per unit time and mitigating extracranial lipid artifacts by improving the point-spread-function. The FOV of 220x220mm² was subdivided into 32x32 voxels. MRSI data were obtained with 2778Hz spectral bandwidth, 2 averages, 2-step phase cycling, TE/TR 69/2800ms, and TA 24:12min. To account for strong B₀ and B₁⁺ inhomogeneities at 7T across the slice, we introduce an adiabatic MEGA-editing scheme where conventional frequency-selective Gaussian editing pulses are replaced by broadband asymmetric hyperbolic secant pulses (combination of one-half of a 32-ms-long HS-1 pulse and the other half of an 8-ms-long HS-4, see Fig.1b). The narrow transition band and flattop were 162Hz and 534Hz, respectively. MM-suppressed GABA-editing was achieved by a global inversion-preparation that nulled co-edited MM signals at 2.99ppm using a 100‐ms‐long, 40th‐order WURST pulse. Real-time motion and scanner instability correction was performed via 3D echo planar imaging navigators (vNav) combined with selective reacquisition of corrupted data (i.e., >0.4mm translation or >0.4° rotation) [6]. Differences in metabolite concentrations between GM and WM were assessed via correlation analysis between metabolic ratio maps and GM/WM tissue composition.

Results

Fig.2a illustrates the simulated and measured inversion profiles for the adiabatic HS pulse and a conventional (FWHM=120Hz) Gaussian pulse. Simulations of the ⁴CH₂-GABA 3.01ppm resonance in Fig.2b showed high editing stability using the asymmetric HS editing pulse, with less than 1.8% integrated GABA signal amplitude variation, even in the presence of a ±30% B₁⁺ offset. Fitted spectra (DIFF and EDIT-OFF) from different brain regions are shown in Fig.3, including voxels with typically good spectral quality and voxels where fitting is challenging due to baseline distortions and B₀ inhomogeneities (frontal lobe). Applying IR reduced the mean SNR calculated over all volunteers and all voxels by about 50% from 241.1±99.0 to 119.6±47.1 (p<0.005). The mean GABA CRLBs increased from 4.2±2.8 to 8.0±2.3 when applying IR (p<0.005), while for Glx the CRLBs increased from 3.2±2.2 to 3.6±2.6 (p=0.015). Fig.4 depicts metabolic ratio maps (i.e., GABA/tNAA, GABA⁺/tNAA, Glx^DIFF/tNAA, and tCr/tNAA) of one volunteer, as well as the pure GABA maps (with IR-ON) obtained for all five volunteers. The linear regression result of GABA/tNAA and Glx/tNAA versus the GM fraction is provided in Fig.5a. GM/WM ratios calculated from linear regression over all volunteers for all metabolic maps are listed in Fig.5b. GABA/tNAA showed a 2.15-fold increase in GM/WM contrast when applying IR, while Glx/tNAA experienced a slight contrast reduction of about 17%.

Discussion/Conclusion

An adiabatic MEGA-editing scheme was developed and incorporated into a B₁⁺-insensitive MRSI sequence to account for the severe B₀ and B₁⁺ variations frequently encountered in whole-slice MRSI at ultra-high field. Integrating time-efficient, concentric ring encoding, and real-time motion and scanner instability correction makes our approach promising for whole-slice metabolic imaging of neurotransmitters (but also antioxidants, and onco-metabolites) in the human brain. Metabolic maps showed distinct anatomical contrast and were used to calculate metabolite abundance differences in GM and WM. Without IR, GABA⁺/tNAA showed only little GM/WM contrast, but this contrast increased 2.15-fold when IR (i.e., GABA/tNAA) was enabled. We attribute this to an elevated abundance of MM_2.99ppm in WM compared to GM, which is in accordance with previous reports [8].

Acknowledgements

This study received support from the Austrian Science Fund (FWF): KLI 718, KLI 646, P 30701 and J 4124; the FFG Bridge Early Stage Grant #846505.

References

[1] Agarwal et al., 2012, Am. J. Neuroradiol. 33, 595–602; [2] Cawley et al., 2015, Brain 138, 2584–95; [3] Brennan et al., 2017, Psychiatry Res. Neuroimaging 269, 9–16; [4] Bogner et al., 2014. Neuroimage 103, 290–302; [5] Jensen et al., 2005, NMR Biomed. 18, 570–576; [6] Moser et al., ISMRM 2018, No. 1287; [7] Hingerl et al., 2018, Magn Reson Med. 79(6):2874-2885; [8] Považan et al., ISMRM 2017, No. 1058