[P2] Poster: Day 2

[P2-30] Exploring the role of rhythmicity for infant word learning by entrainment of brain and behaviour in social contexts: A preliminary study

*Erica Flaten¹, Cristina Conati², Janet Werker¹ (1. Department of Psychology, University of British Columbia (Canada), 2. Department of Computer Science, University of British Columbia (Canada))

Keywords:

Word Learning,Infants,Neural Tracking,EEG,Eye-tracking

Infants learn words from caregivers labelling objects with their names. Such interactions involve infant-directed (ID) communication, which is inherently highly rhythmic. Infants attend more to, learn better from, and more strongly neurally track ID than adult-directed speech, however, whether this is specifically due to rhythmicity is currently unknown. We thus posit that rhythmicity in ID speech dynamically engages infants’ cognitive processes in real time (such as measured via eye-tracking), which enhances word learning, and that such learning processes are anchored by underlying neural activity. We are currently testing this hypothesis while developing machine learning (ML) techniques to utilize infants’ multiple signals together (e.g., eye-tracking and/or EEG, video of facial expressions) to predict learning outcomes. Specifically, 9- to 11-month-old infants (data collection is ongoing) from English-speaking homes were familiarized with two novel objects one at a time on a screen, each paired with a pseudoword (e.g., ‘Bap’ &‘Dit’). The word was spoken repeatedly over an intonation phrase, and these phrases were manipulated to be rhythmically regular (i.e., with regular inter-onset-intervals[IOIs] between word onsets) or irregular (e.g., jittered IOIs between words). During familiarization, infants’ visual and neural signals were measured using eye-tracking and EEG, respectively. Following this, infants’ associative word learning was then tested: infants heard the learned pseudowords one at a time while both objects appeared on the screen, and looking times to the correct vs. incorrect object was measured to index learning. We predict that infants’ brains will more strongly track the regular compared to irregular phrases, and that this pattern will predict their word learning outcomes. Additionally, ML models will predict which infants learned best using the eye-tracking (and in the future, EEG and facial expression) data from the familiarization phase. This project is the first to directly manipulate rhythmic regularity in ID speech to investigate word learning, and additionally, to employ ML techniques to extract features from infants’ multiple signals that predict learning outcomes. This work will better our understanding of the processes involved in early language acquisition.