Reduced GABAergic Action in the Autistic Brain
Summary
An imbalance between excitatory/inhibitory neurotransmission has been posited as a central characteristic of the neurobiology of autism [11], inspired in part by the striking prevalence of seizures among individuals with the disorder [22]. Evidence supporting this hypothesis has specifically implicated the signaling pathway of the inhibitory neurotransmitter, γ-aminobutyric acid (GABA), in this putative imbalance: GABA receptor genes have been associated with autism in linkage and copy number variation studies [3, 4, 5, 6, 7], fewer GABA receptor subunits have been observed in the post-mortem tissue of autistic individuals [8, 9], and GABAergic signaling is disrupted across heterogeneous mouse models of autism [1010]. Yet, empirical evidence supporting this hypothesis in humans is lacking, leaving a gulf between animal and human studies of the condition. Here, we present a direct link between GABA signaling and autistic perceptual symptomatology. We first demonstrate a robust, replicated autistic deficit in binocular rivalry [1111], a basic visual function that is thought to rely on the balance of excitation/inhibition in visual cortex [12, 13, 14, 15]. Then, using magnetic resonance spectroscopy, we demonstrate a tight linkage between binocular rivalry dynamics in typical participants and both GABA and glutamate levels in the visual cortex. Finally, we show that the link between GABA and binocular rivalry dynamics is completely and specifically absent in autism. These results suggest a disruption in inhibitory signaling in the autistic brain and forge a translational path between animal and human models of the condition.
Results and Discussion
During binocular rivalry, two images, one presented to each eye, vie for perceptual dominance as neuronal populations that are selective for each eye’s input suppress each other in alternation [16, 17]. The strength of perceptual suppression during rivalry is thought to depend on the balance of inhibitory and excitatory cortical dynamics [12, 13, 14, 15] and may serve as a non-invasive perceptual marker of the putative perturbation in inhibitory signaling thought to characterize the autistic brain.
We therefore measured the dynamics of binocular rivalry in individuals with and without a diagnosis of autism (41 individuals, 20 with autism). As predicted, individuals with autism demonstrated a slower rate of binocular rivalry (switches per trial: controls = 8.68, autism = 4.19; F(1,37) = 16.52, ηp2 = 0.311, p = 0.001; Figure 1Figure 1A), which was marked by reduced periods of perceptual suppression (proportion of each trial spent viewing a dominant percept, (dominant percept durations)/(dominant + mixed percept durations): controls = 0.69; autism = 0.55; F(1,36) = 7.27, ηp2 = 0.172, p = 0.011; Figure 1Figure 1B). The strength of perceptual suppression inversely predicted clinical measures of autistic symptomatology (Autism Diagnostic Observation Schedule [ADOS]: Rs = −0.39, p = 0.027; Figure 1Figure 1) and showed high test-retest reliability in a control experiment (R = 0.94, p < 0.001; see Supplemental Experimental ProceduresSupplemental Experimental Procedures and also [1818]). These results replicate our previous findings in an independent sample of autistic individuals [1111] and confirm rivalry disruptions as a robust behavioral marker of autism.
Download (3.06 MB )
Atypical Dynamics of Binocular Rivalry in Autism
(A) Schematic of perceptual experience during binocular rivalry. Two images, one presented to each of an individual’s eyes, oscillate back and forth in perceptual awareness as each is suppressed in turn. Throughout a run, participants were instructed to continuously report their perceived image (red, green, mixed) through button press (right, left, up).
(B) Slower rate of binocular rivalry alternations in autism. Individuals with autism demonstrated fewer perceptual switches between the inputs to their left and right eyes, compared with control participants (ηp2 = 0.311, p = 0.001).
(C) Reduced proportion of perceptual suppression in autism. Individuals with autism demonstrated a reduced strength of perceptual suppression, periods of time during which one image is fully suppressed from visual awareness (ηp2 = 0.172, p = 0.01).
(D) The strength perceptual suppression during binocular rivalry inversely predicted autistic symptom severity, measured using the ADOS, a clinical measure of autistic symptomatology.
In all plots, error bars represent 1 SEM. ∗∗p ≤ 0.01, ∗∗∗p ≤ 0.001 difference between the two groups. See also Table S1Table S1.
(A) Schematic of perceptual experience during binocular rivalry. Two images, one presented to each of an individual’s eyes, oscillate back and forth in perceptual awareness as each is suppressed in turn. Throughout a run, participants were instructed to continuously report their perceived image (red, green, mixed) through button press (right, left, up).
(B) Slower rate of binocular rivalry alternations in autism. Individuals with autism demonstrated fewer perceptual switches between the inputs to their left and right eyes, compared with control participants (ηp2 = 0.311, p = 0.001).
(C) Reduced proportion of perceptual suppression in autism. Individuals with autism demonstrated a reduced strength of perceptual suppression, periods of time during which one image is fully suppressed from visual awareness (ηp2 = 0.172, p = 0.01).
(D) The strength perceptual suppression during binocular rivalry inversely predicted autistic symptom severity, measured using the ADOS, a clinical measure of autistic symptomatology.
In all plots, error bars represent 1 SEM. ∗∗p ≤ 0.01, ∗∗∗p ≤ 0.001 difference between the two groups. See also Table S1Table S1.
To test whether altered binocular rivalry dynamics in autism are linked to the reduced action of inhibitory (γ-aminobutyric acid [GABA]) or excitatory (glutamate [Glx]) neurotransmitters in the brain, we measured the concentration of these neurotransmitters in visual cortex using magnetic resonance spectroscopy (MRS). Spectra were acquired from an occipital voxel (2.5 × 2.5 × 3 cm), centered bilaterally on the calcarine sulcus (Figure 2Figure 2A).
Reduced GABAergic, but Conserved Glutamatergic, Action in the Autistic Visual Cortex
(A) Magnetic resonance spectra were acquired from individuals with and without autism: one example control (left) and autistic (right) participant shown here, with the acquisition region (visual cortex) shown in color (80% probability of VOI placement). Spectra included the neurotransmitters predicted to govern binocular rivalry dynamics from computational models, GABA and glutamate, as well as control metabolites TNAA, TCho, TCr, and Ins.(B) GABA strongly predicted the strength of perceptual suppression during rivalry in control individuals (Rs = 0.062, p = 0.002, top left), but this relationship was absent in autism (Rs = 0.02, p = 0.473, top right). However, glutamate strongly predicted binocular rivalry dynamics in both controls (rho = 0.40, p = 0.031, bottom left) and autism (Rs = 0.60, p = 0.004, bottom right).
(C) GABA was the only molecule to show a significantly stronger effect on rivalry dynamics in controls, as compared to autism, demonstrating a selective disruption in GABAergic action in the autistic brain (ηp2 = 0.167, p = 0.013). In all plots, medicated individuals are labeled with unfilled circles.
Error bars represent 1 SEM. See also Figures S1Figures S1 and S2S2 and Table S2Table S2.
GABA and glutamate are predicted to contribute to different aspects of binocular rivalry dynamics: mutual inhibition between (GABA) and recurrent excitation within (glutamate) populations of neurons coding for the two oscillating percepts [1414]. Specifically, in classic models of binocular rivalry, a period of perceptual suppression is maintained through excitation within the neuronal population selective for the dominant image, as well as cross-inhibition of the neuronal population selective for the non-dominant image [12, 19, 20]. Critically, reducing either mutual inhibition or recurrent excitation is predicted to reduce the strength of perceptual suppression during rivalry in one implementation of this model [1414], mirroring the dynamics we observed in autism. We therefore considered each neurotransmitter separately to test whether inhibitory or excitatory signaling was selectively disrupted in the autistic brain.
Given prior evidence from genetic, animal, and post-mortem studies, we hypothesized that inhibitory signaling may be affected in the autistic brain. This disruption could take the form of reduced levels of GABA or glutamate. However, reports of perturbations of key components of the GABA signaling pathway, such as receptors [3, 4, 5, 6, 7, 8, 9] and inhibitory neuronal density [1010], suggest that GABA levels themselves may not be altered in autism but instead may be less predictive of rivalry dynamics.
As predicted by models of binocular rivalry, GABA concentrations in visual cortex strongly predicted rivalry dynamics in controls, where more GABA corresponded to longer periods of perceptual suppression (Rs = 0.62, p = 0.002; Figure 2Figure 2B). However, this relationship was strikingly absent in individuals with autism (Rs = 0.02, p = 0.473; Figure 2Figure 2B).
The difference between the two correlations was significant (ηp2 = 0.167, p = 0.013; Figure 2Figure 2C), indicating a reduced impact of GABA on perceptual suppression in the autistic brain.
Importantly, this finding was specific to GABA: glutamate strongly predicted the dynamics of binocular rivalry in autism (Rs = 0.60, p = 0.004; Figure 2Figure 2B), to the same degree as that found in controls (Rs = 0.40, p = 0.031 in controls; p = 0.71 for the difference of this effect between autism and controls; Figure 2Figure 2C). The absence of a group difference in the concentrations of GABA and Glx (both p > 0.32; Figure S1Figure S1; Table S2Table S2) indicate that GABA levels themselves are not altered in the autistic visual cortex [2121], despite a specific reduction in the effect of GABA on autistic visual behavior. These findings suggest that alterations in the GABAergic signaling pathway may characterize autistic neurobiology. Consistent with prior evidence from animal and post-mortem studies, such dysfunction may arise from perturbations in key components of the GABAergic pathway beyond GABA levels, such as receptors [3, 4, 5, 6, 7, 8, 9] and inhibitory neuronal density [1010].
These results demonstrate reduced GABAergic, but conserved glutamatergic, action in the autistic visual system. No other metabolite measured predicted rivalry dynamics in either group (all p > 0.11 for TNAA, TCho, TCr, and Ins), and GABA was the only metabolite to show a markedly reduced impact on the dynamics of binocular rivalry in autism (all other effects between autism and controls: p > 0.29; Figure 2Figure 2C). Furthermore, this relationship was specific to GABA measured in the visual cortex: binocular rivalry dynamics were not related to any metabolites measured in a control region of interest, the motor cortex, in either group (all Rs < 0.17, all p > 0.25).
The specificity of this effect to GABA and not glutamate argues against the possibility that these findings are driven by differences in the spectral quality between the two groups. Further, spectral fit errors (an indication of data quality) and frequency drift (an indication of subject motion) were within the expected ranges and matched between groups (all p > 0.46; Figure S1Figure S1; Table S2Table S2). Repeatability control experiments indicated that our metabolite measurements were highly reliable (coefficient of variation [CVs] < 8%) with a high reproducibility relative to the effect range of our study (SDs < 13% of the group range for each metabolite) (see Supplemental Experimental ProceduresSupplemental Experimental Procedures). These results confirm previous reports of high test-retest reliability of in vivo MRS measurements [2222] and add to a growing literature of replicated relationships between GABA concentration and psychophysical performance, including sensory sensitivity thresholds [23, 24], motor response inhibition [2525], and binocular rivalry dynamics [15, 26].
Atypical sensory perception has been noted since the earliest reports of autism [2727], although little is known about the neural underpinnings of these symptoms. Interestingly, recent reports of reduced spatial suppression [2828], decreased global motion perception [29, 30], larger population receptive fields [3131], and more variable evoked responses [32, 33] in autism mirror the perceptual consequences of reduced GABAergic inhibition in animal studies [34, 35]. Computational accounts suggest that reducing inhibition may account for a wide range of autistic perceptual symptoms [3636]. One such account predicted reduced perceptual suppression during rivalry in autism but found only a trend toward the finding ([3737], but see [3838]). Further work is needed to test whether this litany of visual disruptions indeed results from reduced GABAergic signaling in the autistic brain, as we demonstrate here for one robust autistic visual symptom.
Our findings provide evidence for an empirical link between a specific neurotransmitter measured in the brains of individuals with autism and an autistic behavioral symptom. Although perceptual in nature, this symptom strongly predicts higher-order clinical measures of autistic symptomatology and may be well poised to serve as a behavioral marker of a perturbation in GABAergic signaling in the autistic brain. Together with the pivotal roles of GABA in canonical cortical computations [3939] and neurodevelopment [4040], these findings point to the GABAergic signaling pathway as a prime suspect in the neurobiology of this pervasive developmental disorder [4141]. Future work should replicate these findings and investigate GABAergic signaling in autism at multiple developmental time points to establish whether GABAergic perturbations predict, or arise in response to, the developmental onset of autistic symptomatology.
No hay comentarios:
Publicar un comentario
Aportes