Fmri Reveals That Toddlers with an ASD Respond Abnormally in the Superior Temporal Sulcus to Social Orienting Stimuli during Natural Sleep

Background:  Unusual social responding, such as a failure to respond to name, is a red flag for autism spectrum disorders (ASD) in infants and toddlers, but the neural systems that underlie this deficit are under-studied. The superior temporal sulcus (STS) has been dubbed the “chameleon of the human brain” because of its involvement in many social tasks. It is unknown whether its dysfunction underlies social responding deficits at the time of first clinical signs of autism, and could thus serve as an early biomarker of risk for ASD.

Objectives:  We aimed to test whether there were early socially-relevant neural differences between infants and toddlers later diagnosed with ASD and those who develop typically by comparing functional magnetic resonance (MR) brain response to social auditory stimuli compared to language and non-social sounds.

Methods:  A population-based screening method and community referral identified toddlers at risk for an ASD as young as 12 months, and children were followed until their third birthday for final diagnostic judgment. ASD and typically developing (TD) comparison children (12-48 months old) were scanned with MR imaging during natural sleep. Blood oxygen level dependent (BOLD) images were acquired during auditory stimulation with social (calling the child’s own name), language (short, monotone phrases), and non-social (alerting or orienting sounds) stimuli (Figure 1) in 61 ASD and 57 TD sleeping participants. The number of voxels with significant BOLD response (p < 0.01) to each condition was measured within each participant’s left and right STS as determined by hand tracing on the anatomical MR image. The proportion of activated voxels within these two regions between groups and conditions was compared using a general linear model.

Results:  By design, groups were well-matched on age and gender. In the left STS, response was related to condition across all participants (F(1,116) = 5.3, p = 0.02), individuals with ASD showed reduced responsiveness across conditions (F(1,116) = 6.3, p = 0.01), and there was a significantly different pattern of response across conditions in the two groups (F(1,116): 5.9, p=0.02). The interaction is illustrated in Figure 2: Whereas the TD group showed a stair-step reduction in response from social to language to non-social stimuli, the ASD response was not differential across conditions. For the right STS, there was no linear effect of condition on brain response (F(1,116) = 1.96, p = 0.17), but across conditions the ASD group had less response (F(1,116) = 7.7, p= 0.006). The pattern of response to the conditions was not significantly different between groups (F(1,116) = 3.0, p = 0.08).

Conclusions:  Infants and toddlers who go on to be diagnosed with an ASD show reduced responsiveness in the left and right STS to auditory stimuli during natural sleep. Furthermore, in the left STS, the typical preferential responsiveness to social compared to language and non-social stimuli is not present among ASD infants and toddlers. Failure to demonstrate a graded response of the left STS to socially-meaningful stimuli may be an early neural marker of ASD that portends future behavioral deficits.