Language-Induced Frontal Gamma Activity Is Associated with Language Ability in 18-Month-Olds with High Familial Risk of Autism

Background: In autism spectrum disorder (ASD), deficits in language development are often observed with varied severity, leading to great heterogeneity among individuals. Language ability also serves as a predictor for further sociocognitive deficits in this population. Improved understanding of early language development in children with ASD is needed for the development of effective language therapies. Prior studies suggest that infants with high familial risk of ASD show different EEG patterns related to language as compared to low risk controls. In typically developing infants and toddlers, brain oscillations in the gamma frequency (~35-50Hz) are thought to reflect cognitive and perceptual processing, including language acquisition. Our lab has previously found that increased resting frontal gamma activity is negatively associated with language ability in high familial risk toddlers, but not low risk controls.

Objectives: To investigate whether induced gamma power during a passive language-based task is associated with language ability in 18-month-olds at high familial risk for autism.

Methods: This study analyzed 18-month EEG data and language measures collected as part of the longitudinal Infant Sibling Project, comparing infants by their risk level for ASD development. Infants with a sibling with ASD were assigned to high-risk group (HRA) and those without siblings or first degree relatives with ASD were assigned to low-risk group (LRC). Language abilities were measured using the Mullen Scales of Early Learning (MSEL). Two types of auditory stimuli were presented while EEG was continuously recorded; 20 words were identified as “familiar” while the other 20 were identified as “unfamiliar” based on typical language development standards of MacArthur-Bates Communicative Development Inventories (MBCDI). Linear regression models were used to examine group differences in the relationship between language-induced frontal gamma power and receptive language ability. Maternal education was included as a covariate.

Results: Participants included 36 LRC and 44 HRA toddlers; of the high-risk toddlers, 12 later met criteria for ASD (HRA+) and 32 did not (HRA-). There was no significant difference in language-induced gamma between groups. There was a significant negative relationship between language-induced low gamma power (40-45Hz) in response to familiar words in the HRA+ group (p=0.049), but no relationship in LRC group (p=0.94), and a marginally significant positive relationship in the HRA- group (p=0.058) were observed. In addition, the relationship between gamma and language in HRA+ group was significantly different from HRA- group (p<0.01) and marginally different from LRC group (p=0.07) in response to familiar stimulus. No significant relationship was observed in any groups between the gamma response to unfamiliar words and language ability.

Conclusions: This study demonstrates early differences in brain-language associations between high-risk 18-months-olds with and without ASD. Future directions include investigation of the phase-locked gamma response to language stimuli. In addition, further investigation of the longitudinal trajectory of language-induced EEG power as it related to language development may further inform development of language specific therapies for ASD.