GABA and glutamate measurements in temporal cortex of autistic children

Abstract

Autism spectrum disorder (ASD) is a neurodevelopmental disorder and presents with challenges in social communication. A hypothesized underlying contributing mechanism is the imbalance in excitation and inhibition (E/I), partly influenced by the levels of excitatory neurotransmitter glutamate (Glu) and inhibitory neurotransmitter γ-aminobutyric acid (GABA) in the brain. Although many have reported the levels of GABA and Glu in the brain, only a few reports address the temporal cortex and then only with a small sample of autistic children, and often only in one hemisphere. We used a macromolecular suppressed edited-magnetic resonance spectroscopy (MRS) sequence to study GABA and Glu (as potential key players influencing E/I) in a large sample of children with ASD in the right and left temporal cortices of children with (N = 56) and without (N = 30) ASD (7-18 years). As a group, children with ASD exhibited no differences in the left hemisphere (GABA and Glu Cohen's |d|: 0.24 and 0.03), but the right hemisphere showed higher GABA and lower Glu concentrations (GABA and Glu Cohen's |d|: 0.53 and 0.65) compared to neurotypicals. Furthermore, a negative association was found between the right hemisphere Glu levels of the ASD group and a clinical assessment tool (r = -0.361, p = 0.022), reflecting autism trait severity (social responsiveness scale). In conclusion, we highlight the chemical abnormalities in children with ASD through a cross-sectional measurement. Longitudinal studies are warranted to determine whether these chemical levels persist or resolve over development.

Keywords: GABA; MEGA‐PRESS; MM‐suppressed; autism; glutamate.

FIGURE 1

Edited MRS voxel locations in the left and right auditory cortices. Every participant's voxel mask from both regions (separately) were transformed from their native space to MNI space and overlaid onto the Colin Holmes (CH) template. Light colors (yellow color) indicate areas of greater overlap across participants. The transformation was done for visual illustration only (not for metabolite concentration estimations).

FIGURE 2

In vivo difference (top) and sum spectra (bottom) acquired from the (a) left hemisphere (LH) and (b) right hemisphere (RH). The black continuous line within each spectrum is the overall average spectrum (ASD and TD combined) overlaid with ±1SD of the combined data (gray) and models (red). The difference spectra were scaled by a factor of 10 for visualization purposes only. ASD, autism spectrum disorder; LH, left hemisphere; RH, right hemisphere; TD, typically developing.

FIGURE 3

Raincloud plots of GABA (a and b), Glu (c and d), and GABA/Glu (e and f) in the left (left) and right (right) hemispheres. The statistical results are from running ANCOVA for every metabolite (separately) in each region with age and sex as covariates. Only GABA, Glu, and GABA/Glu in the right hemisphere are significantly different between the groups (p < 0.05). ASD, autism spectrum disorder; GS, group status; i.u., institutional units; LH, left hemisphere; RH, right hemisphere; TD, typically developing; p‐values for the GS, sex, and age effects are from ANCOVA. Right hemispheric results are highlighted to depict statistically significant group differences.

FIGURE 4

SRS correlations with Glu in the ASD and TD groups separately. ASD, autism spectrum disorder; i.u., institutional units; RH, right hemisphere; SRS, social responsiveness scale; TD, typically developing; p*: p‐value did not survive multiple comparison.