GABA estimation in the brains of children on the autism spectrum: measurement precision and regional cortical variation

Abstract

(1)H magnetic resonance spectroscopy ((1)H MRS) and spectral editing methods, such as MEGA-PRESS, allow researchers to investigate metabolite and neurotransmitter concentrations in-vivo. Here we address the utilization of (1)H MRS for the investigation of GABA concentrations in the ASD brain, in three locations; motor, visual and auditory areas. An initial repeatability study (5 subjects, 5 repeated measures separated by ~5days on average) indicated no significant effect of reference metabolite choice on GABA quantitation (p>0.6). Coefficients of variation for GABA+/NAA, GABA+/Cr and GABA+/Glx were all of the order of 9-11%. Based on these findings, we investigated creatine-normalized GABA+ ratios (GABA+/Cr) in a group of (N=17) children with autism spectrum disorder (ASD) and (N=17) typically developing children (TD) for Motor, Auditory and Visual regions of interest (ROIs). Linear regression analysis of gray matter (GM) volume changes (known to occur with development) revealed a significant decrease of GM volume with Age for Motor (F(1,30)=17.92; p<0.001) and Visual F(1,16)=14.41; p<0.005 but not the Auditory ROI (p=0.55). Inspection of GABA+/Cr changes with Age revealed a marginally significant change for the Motor ROI only (F(1,30)=4.11; p=0.054). Subsequent analyses were thus conducted for each ROI separately using Age and GM volume as covariates. No group differences in GABA+/Cr were observed for the Visual ROI between TD vs. ASD children. However, the Motor and Auditory ROI showed significantly reduced GABA+/Cr in ASD (Motor p<0.05; Auditory p<0.01). The mean deficiency in GABA+/Cr from the Motor ROI was approximately 11% and Auditory ROI was approximately 22%. Our novel findings support the model of regional differences in GABA+/Cr in the ASD brain, primarily in Auditory and to a lesser extent Motor but not Visual areas.

Keywords: ASD; Autism spectrum disorder; GABA; MEGA-PRESS; MRS; Spectroscopy; γ-Aminobutyric acid.

Figure 1

Spatial overlap of the MRS voxel locations, each acquired during separate scanning sessions, for a representative subject of the repeatability study. The large spatial overlap illustrates the high degree of repeatability of the MRS voxel placement procedures.

Figure 2

Population analysis of GABA+/Cr ratios in ASD. A) The age and GM covaried ratio of GABA+ to Cr was investigated in each of the three MRS voxels. Population level decreases were not observed for the Visual ROI. Motor ROI GABA+/Cr was significantly reduced in ASD (p<0.05), as well as Auditory ROI (p < 0.01 corrected).

Figure 3

To demonstrate generalized robustness, stacked MEGA-PRESS spectra are shown for all Auditory ROI measures (N=13 ASD, N=11 TD). The GABA peak at 3 ppm is indicated in grey. The approximate location and regional overlap for Auditory ROI placement is shown in Figure 4.

Figure 4

Relationship between GABA+/Cr ratio and grey matter volume. A) The spatial overlap of the auditory MRS voxel generated from the combined TD and ASD subjects is shown. B) No correlation between Age corrected auditory GABA+/Cr and the volume of grey matter within the auditory MRS voxel was found for either the TD (R² = 0.04) or the ASD (R² = 0.09) subjects. The high degree of spatial overlap and low correlation between GM volume and GABA+/Cr suggest that the small variations in voxel placement and individual anatomy do not affect the GABA+/Cr levels. Importantly, Auditory GABA+/Cr estimates alone may offer partial discriminatory resolution of ASD from TD.