Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Using Longitudinal EEG Connectivity and Power in Infancy to Predict Language Outcomes at Preschool and School Age
Poster Session B, Wednesday, September 30, 4:30 - 6:30 pm, Wangari Maathai
This poster is part of the Sandbox Series.
Anika van der Klis1, Karlijn Herijgers1, Elise Turk1, Caroline Junge1; 1Utrecht University
Early language development is supported by the maturation of neural systems involved in social attention and communication. Neural oscillations in the theta (3-6 Hz) and alpha (6-9 Hz) frequency bands can reliably be measured in infancy (van der Velde et al., 2019) and have been linked to social processing and later childhood intelligence (Jones et al., 2015, 2020; van der Velde et al., 2021). In particular, theta activity has been associated with social perception, while alpha suppression is thought to reflect increased cortical activation (Jones et al., 2015). Alpha connectivity within brain regions, especially in the left frontal area, has been related to language outcomes (Whedon et al., 2016). Currently, it remains unclear whether infant oscillatory power and connectivity in alpha and theta bands can predict long-term language development. This preregistered study uses data from the large-scale, longitudinal YOUth cohort to examine whether longitudinal EEG power and connectivity at 5 months and 10 months are associated with language skills at preschool age (2-4 years) and school age (3-7 years). We have already preprocessed EEG data of 2562 Dutch children of whom EEG was collected at 5 months and/or at 10 months. Continuous EEG was recorded while infants viewed social (videos of women singing nursery rhymes) and non-social (moving toys) stimuli. Each 1-minute video was repeated three times. Neural power and coherence – reflecting synchronized activity between brain regions – were computed in the theta (3-6 Hz) and alpha (6-9 Hz) bands. Language outcomes were collected for 750 children using the PPVT-III-NL receptive vocabulary test at 2–4 years. A subsample of 155 children also completed the pre-CELF assessment at 3–7 years for the BRAVECHILD add-on study. In the coming months, we will fit linear mixed-effects models predicting language outcomes using infant EEG power and coherence in theta and alpha bands at 5 months and 10 months, stimulus type (social vs. non-social), children’s age, and interactions between these variables, while accounting for repeated measures. Non-linear effects of age will be tested by including quadratic age terms. We expect that infant neural power and coherence will significantly predict future language skills. Based on prior literature, stronger theta connectivity and power across the brain are anticipated to be associated with higher language scores. We also expect that alpha suppression, particularly in the left frontal region, is related to better language outcomes. Additionally, we expect higher neural engagement (greater power and coherence) during the processing of social compared to non-social stimuli. The fitted mixed-effects models will show whether any predictive relationships between physiological activity and later language outcomes differ per frequency band (theta vs. alpha), and whether task-related (social) versus resting-state (non-social) activity provides stronger predictive values. The current study will provide new insight into how early neural markers of social attention relate to long-term language development. Findings may demonstrate that both power and connectivity in infancy serve as early predictors of long-term language ability. Furthermore, differences between social and non-social processing may help clarify the role of socially driven neural engagement in language acquisition.
Topic Areas: Language Development/Acquisition, Speech Perception