Neural selectivity for communicative auditory signals in Phelan-McDermid syndrome

Abstract

Background: Phelan-McDermid syndrome (PMS), a neurodevelopmental disorder caused by deletion or mutation in the SHANK3 gene, is one of the more common single-locus causes of autism spectrum disorder (ASD). PMS is characterized by global developmental delay, hypotonia, delayed or absent speech, increased risk of seizures, and minor dysmorphic features. Impairments in language and communication are one of the most consistent characteristics of PMS. Although there is considerable overlap in the social communicative deficits associated with PMS and ASD, there is a dearth of data on underlying abnormalities at the level of neural systems in PMS. No controlled neuroimaging studies of PMS have been reported to date. The goal of this study was to examine the neural circuitry supporting the perception of auditory communicative signals in children with PMS as compared to idiopathic ASD (iASD).

Methods: Eleven children with PMS and nine comparison children with iASD were scanned using functional magnetic resonance imaging (fMRI) under light sedation. The fMRI paradigm was a previously validated passive auditory task, which presented communicative (e.g., speech, sounds of agreement, disgust) and non-communicative vocalizations (e.g., sneezing, coughing, yawning).

Results: Previous research has shown that the superior temporal gyrus (STG) responds selectively to communicative vocal signals in typically developing children and adults. Here, selective activity for communicative relative to non-communicative vocalizations was detected in the right STG in the PMS group, but not in the iASD group. The PMS group also showed preferential activity for communicative vocalizations in a range of other brain regions associated with social cognition, such as the medial prefrontal cortex (MPFC), insula, and inferior frontal gyrus. Interestingly, better orienting toward social sounds was positively correlated with selective activity in the STG and other "social brain" regions, including the MPFC, in the PMS group. Finally, selective MPFC activity for communicative sounds was associated with receptive language level in the PMS group and expressive language in the iASD group.

Conclusions: Despite shared behavioral features, children with PMS differed from children with iASD in their neural response to communicative vocal sounds and showed relative strengths in this area. Furthermore, the relationship between clinical characteristics and neural selectivity also differed between the two groups, suggesting that shared ASD features may partially reflect different neurofunctional abnormalities due to differing etiologies.

Fig. 1

Differential activity in the STG elicited by communicative vs. non-communicative vocalizations in the PMS group (a), but not the iASD group (b) (p < 0.05, k > 194, corrected)

Fig. 2

Selective activity for communicative vs. non-communicative sounds as a function of social orienting abilities in PMS. A positive correlation was observed between differential “social brain” activity (e.g., cortex along the STS, MPFC) and social orienting skills in the PMS group (p < 0.05, k > 194, corrected)

Fig. 3

Selective activity for communicative vs. non-communicative sounds as a function of autism symptom severity in the social communication domain. A negative correlation was found in children with PMS between preferential activity in social brain networks and social symptom severity as measured by the social affect subscale of the ADOS-2 (p < 0.05, k > 194, corrected)

Fig. 4

Selective activity for communicative vs. non-communicative sounds as a function of repetitive behaviors. A negative correlation was found in the iASD group between selective activity in the STG and MPFC and the repetitive behavior subscale of the ADOS (p < 0.05, k > 194, corrected)