Language acquisition in the longitudinal BabyRhythm cohort

Abstract

Here we report preliminary analyses of the linguistic tasks selected for the Cambridge UK BabyRhythm project, data to be made available online via OSF. BabyRhythm is a study of 122 infants as they age from 2 – 30 months, investigating cortical tracking and sensorimotor synchronisation to acoustic and visual rhythm in relation to language acquisition. Participating infants attended for 8 brain recording sessions (EEG) between the ages of 2 – 11 months, and received language tasks beginning at 8 months. The tasks selected were both standardised and experimental, and were intended to sample 5 linguistic domains: semantics, phonology, grammar, rhythmic timing and gesture. For each task we here describe infant performance (mean, median, range), and we also report performance by sex and by monolingual (N=92) versus multilingual (N=30) status. We report relations between measures, and identify the most reliable language tasks that will be utilised in subsequent longitudinal brain-behaviour analyses.

Infant low-frequency EEG cortical power, cortical tracking and phase-amplitude coupling predicts language a year later.

“…Multivariate linear models were used to test whether the neural findings reported in 3.1 (for the full sample) were related to the different language measures outlined in the methods above. These language measures were selected as most reliable in previous analyses of the Cambridge UK BabyRhythm cohort, as they had the largest number of infants contributing data and scores were neither at floor nor ceiling (3). As noted, the parent-estimated measures were receptive and productive vocabulary scores from the UK-CDI (43) and the infant-led measures were (1) The ability to point at 12-months, (2) Word recognition at 18-months on a computerised comprehension test (CCT), and (3) Phonological processing at 24-months, as measured by non-word repetition (NWR).…”

“…These language measures were selected as most reliable in previous analyses of the Cambridge UK BabyRhythm cohort, as they had the largest number of infants contributing data and scores were neither at floor nor ceiling (3). As noted, the parent-estimated measures were receptive and productive vocabulary scores from the UK-CDI (43) and the infant-led measures were (1) The ability to point at 12-months, (2) Word recognition at 18-months on a computerised comprehension test (CCT), and (3) Phonological processing at 24-months, as measured by non-word repetition (NWR). Due to the inherent difference in parent-estimated and infant-led measures of language, separate multivariate linear models were run for each set of scores.…”

“…These were 1, gesture (adapted from (44)), in which pointing towards a present and absent target was used a pre-verbal measure of communicative intent at 12 months; 2, the computerized comprehension task (CCT), a touch screen-based measure of infant vocabulary in which the infant is asked to select a picture of a named item (45), and 3, a custom designed non-word repetition (NWR) task, which measured the child’s productive phonology. Three measures were taken from the NWR task administered at 24 months, (i) accurate reproduction of consonants, (ii) number of syllables reproduced, and (iii) stress patterns (canonical or non-canonical, see Rocha et al, preprint (3)). This NWR task measured phonological production of both nonsense words (e.g.…”

“…Here we contribute novel longitudinal neural data from a relatively large sample of 113 infants, utilising EEG recordings taken during the first year of life while infants listened to sung or chanted nursery rhymes. The neural measures are used to explore relations with subsequent language acquisition using a battery of standardised and experimental phonological, gestural and lexical measures taken from 12 to 24 months of age (3). The auditory neuroscience of adult speech processing has revealed that accurate encoding of amplitude modulations (AMs) in the speech signal via neuronal oscillations is fundamental to many different aspects of language comprehension.…”

“…Here we contribute novel longitudinal neural data from a relatively large sample of 113 infants, utilising EEG recordings taken during the first year of life while infants listened to sung or chanted nursery rhymes. The neural measures are used to explore relations with subsequent language acquisition using a battery of standardised and experimental phonological, gestural and lexical measures taken from 12 to 24 months of age (3).…”