Functional Connectivity during Language Processing in Infants with Familial Risk for ASD

Background: At the circuit-level, autism spectrum disorder (ASD) is characterized by disruptions in neural connectivity. Studies in early childhood have shown that resting-state frontal EEG power in the gamma (Benasich 2008, Gou 2011) and alpha bands (Levin 2017) relate to language ability. Infants are at elevated risk for developing ASD (familial-risk) if they have an older sibling with ASD; these infants are also at risk of language delay (Messinger 2015, Charman 2017). Given these elevated risks, neural connectivity during language processing in familial-risk infants warrants further investigation, particularly as it relates to language outcomes.

Objectives: We asked if: 1) there are differences in baseline EEG power and connectivity during a language processing task between 3-month old familial-risk and low-risk (no family history of ASD) infants, and 2) connectivity correlates with language ability and ASD symptoms at 18 months. Our overarching hypothesis is that neural connectivity is disrupted in infants with familial risk for ASD, with the largest differences found in infants with delayed or atypical development.

Methods: Participants included 33 familial-risk infants and 30 low-risk infants, from an ongoing longitudinal study at the UCLA Autism Center for Excellence. EEG was acquired at 3-month while each subject listened passively to a continuous stream of syllables for 2 minutes (EGI 128 channels). EEG was collected using NetStation, filtered at 1-50 Hz, and cleaned using EEGlab and independent component analysis. Relative spectral power was calculated using Welch’s method on average re-referenced data. Phase coherence was calculated using newcrossf on Laplacian filtered data. Frontal and temporal relative power and coherence were calculated in the theta (4-6 Hz), alpha (6-12 Hz), beta (12-30 Hz), and gamma (30-50 Hz) bands. Language ability and ASD symptoms were assessed at 18-month using the Mullen Scales of Early Learning (MSEL) and Autism Diagnostic Observation Schedule-Toddler Module (ADOS-T), respectively. Independent samples t-tests were used to compare group differences in power and coherence; Pearson’s correlations were used to relate coherence to MSEL verbal t-score (average of expressive and receptive language) and ADOS-T overall score.

Results: P-values from independent samples t-tests are in Table 1; significant behavioral correlations are in Table 2. 1) There were no significant group differences in frontal and temporal power (p > 0.05, Table 1). Across the frequency bands, familial-risk infants generally have hypoconnectivity in interhemispheric frontal and left frontal-right temporal connections; and hyperconnectivity in right-frontal and right frontal–left temporal connections. 2) Across the 50 subjects with 18-month behavioral data, right frontal-left temporal coherence in the beta band negatively correlated with 18-month MSEL verbal t-score (R=-0.283, p=0.046), and positively correlated with 18-month ADOS-T overall score (R=0.471, p=0.001).

Conclusions: Familial-risk infants show evidence of aberrant frontal-temporal connectivity at 3-month of age during a language processing task, and connectivity was associated with language ability and ASD symptoms at 18-month. Our findings suggest that neural connectivity during auditory language processing in early infancy is important for subsequent language development. We will expand our connectivity analysis to 6-12-month time points to examine potential divergent developmental trajectories in infants with atypical development.