Language Associated Neural Responses of Infants at High Risk for ASD and Nonsyndromic Craniosynostosis

Background: Literature suggests that infants at high risk for autism spectrum disorder (HR-ASD) and nonsyndromic craniosynostosis (NSC), early cranial suture fusion, have similar language acquisition profiles. Essential in language remediation for these cohorts is early detection and prevention. Auditory event-related potentials (ERPs) measure passive neurological responses to speech sounds, suggesting a promising avenue for studying infant speech development. In particular, the mismatch negativity (MMN) quantifies perceptual narrowing and can predict future language development. The MMN component has demonstrated attenuated responses in HR-ASD infants compared to typically-developing (TD) controls. Previous work has established differences between the auditory ERPs of HR-ASD and those of metopic synostosis (MSO) and sagittal synostosis (SSO) infants. However, comparisons have not been made with coronal synostosis (CSO) or across groups of NSC.

Objectives: To characterize specificity of language acquisition in HR-ASD by comparing MMN responses to speech sounds between HR-ASD, CSO, MSO, SSO, and TD controls.

Methods: HR-ASD, NSC, and TD infants were recruited from the Yale Autism Program and Yale Craniofacial Clinic. Participants were presented with a non-native phoneme discrimination paradigm involving the Hindi retroflex phoneme /da/ and the dental phoneme /da/ in random order. Auditory stimuli were set at 80 dB, and EEG was recorded at 250 Hz with a 128-channel HydroCel Geodesic Sensor Net. Analysis focused on four electrode clusters: left and right frontal, and left and right central electrodes. The MMN component was calculated as the largest negative amplitude in the difference wave between 80-300ms after stimulus presentation. Statistical comparisons were performed with ANOVA and studentized T-tests.

Results: 12 HR-ASD, 5 CSO, 14 SSO, 14 MSO, and 34 age-matched TD infants were included in analyses. The MMN amplitudes measured in left frontal clusters were statistically different between cohorts (p=0.043). HR-ASD, SSO, and MSO infants produced attenuated left frontal responses compared to TDs (p=0.025, p=0.001, p=0.003). Between NSC cohorts, SSO and MSO infants had attenuated MMN responses in the left frontal clusters compared to CSO (p=0.022, p=0.04). MMN amplitudes in the left central clusters were also significantly different between cohorts (p=0.006). SSO and MSO infants demonstrated attenuation in the left central clusters compared to TDs (p=0.015). In comparison to HR-ASD, SSO produced attenuated right frontal and central clusters (p=0.031, p=0.008), while MSO demonstrated attenuated right central clusters (p=0.005). There were not differences between CSO infants and TD or HR-ASD infants.

Conclusions: This represents the largest ERP comparison of ASD with different subtypes of NSC. Results replicate earlier findings that HR-ASD infants respond with lower MMN amplitudes than TD infants. In comparison with SSO and MSO infants, our findings suggest that HR-ASD infants may manifest cerebral attenuation focal to the left frontal brain, the hemisphere responsible for language production. Importantly, CSO infants may represent an intermediary phenotype between TD and HR-ASD infants. This study begins to develop the early language profile of HR-ASD infants within the clinical context of different suture fusion synostosis. Results warrant future studies comparing language acquisition in HR-ASD and NSC.