Words and Meters: Neural Evidence for a Connection Between Individual Differences in Statistical Learning and Rhythmic Ability in Infancy

Abstract

Music and language are both hierarchically structured: syllables combine into words, and meters are superimposed on musical beats. Statistical Learning (SL) is a learning mechanism that drives speech segmentation and word learning through computation of transitional probabilities between syllables. Individual differences in SL ability are predictive of further language development, and the current study investigated whether Rhythmic Ability (RA; beat/meter perception in rhythmic stimuli) is a mechanism correlating with SL in six- to nine-month-old infants. We further explored whether RA of the parents predicted infant RA and/or SL. We used EEG to measure infants’ neural entrainment in two conditions: (1) a speech segmentation SL condition in which transitional probabilities between syllables were the only cue to segment the speech stream into words; (2) a RA condition exposing infants to a syncopated rhythm, yielding a 4/4 meter. We correlated neural entrainment to the words in the SL condition with the meter in the RA condition. Parents completed standardized tasks measuring RA and were asked about their solo and joint (parent-child) engagement with music. Results revealed a relationship between neural entrainment indexing SL and RA in infants, which was unaffected by infant age. This relationship was specific to neural entrainment to words and meters, suggesting overlapping neural mechanisms for processing items at hierarchically similar levels. We found no evidence that parental RA predicted infant RA or SL. However, frequency of parent-child joint musical engagement appears to have a positive effect on infant RA, paving the way for further studies on this topic.