Cortex. 2025 Nov 15;194:35-49. doi: 10.1016/j.cortex.2025.11.003. Online ahead of print.
ABSTRACT
Children’s ability to extract statistical regularities from speech is considered fundamental to lexical, syntactic, and grammatical development. However, the neural oscillatory mechanisms supporting this process in childhood remains poorly understood. While beta-band oscillations have been linked to statistical learning in visual and motor domains, it is unclear whether similar dynamics support auditory statistical learning in children. In this study, we recorded electroencephalography (EEG) from children aged 8-12 years as they listened to a continuous stream of trisyllabic nonwords (e.g., dapiku), where syllable order within each nonword was fixed (high predictability), but transitions between nonwords were variable (low predictability). Beta power was significantly lower for the more predictable second and third syllables relative to the less predictable first syllable. This effect emerged only after repeated exposure and was localised to left prefrontal electrodes. Beta power also correlated with post-exposure recognition accuracy. Additional learning-related modulations were observed in the theta-alpha and delta-theta bands, suggesting broader oscillatory engagement. These findings indicate that auditory statistical learning in middle childhood engages frequency-specific neural dynamics, with beta power modulations showing parallel effects to those observed in other modalities.
PMID:41330012 | DOI:10.1016/j.cortex.2025.11.003
