Sleep promotes phonological learning in children (Knowland et al., 2019)

Purpose: Establishing stable and flexible phonological representations is a key component of language development and 1 which is thought to vary across children with neurodevelopmental disorders affecting language acquisition. Sleep is understood to support the learning and generalization of new phonological mappings in adults, but this remains to be examined in children. This study therefore explored the time course of phonological learning in childhood and how it varies by structural language and autism symptomatology.
Method: Seventy-seven 7- to 13-year-old children, 30 with high autism symptomatology, were included in the study; structural language ability varied across the sample. Children learned new phonological mappings based on synthesized speech tokens in the morning; performance was then charted via repetition (without feedback) over 24 hr and followed up 4 weeks later. On the night following learning, children’s sleep was monitored with polysomnography.
Results: A period of sleep but not wake was associated with improvement on the phonological learning task in childhood. Sleep was associated with improved performance for both trained items and novel items. Structural language ability predicted overall task performance, though language ability did not predict degree of change from 1 session to the next. By contrast, autism symptomatology did not explain task performance. With respect to sleep architecture, rapid eye movement features were associated with greater phonological generalization.
Conclusions: Children’s sleep was associated with improvement in performance on both trained and novel items. Phonological generalization was associated with brain activity during rapid eye movement sleep. This study furthers our understanding of individual differences in the acquisition of new phonological mappings and the role of sleep in this process over childhood.

Supplemental Material S1. Interaction shown in the Follow-up model between Language Composite, AI Group and Condition. Data points shown as triangles were excluded from the model on the grounds of being overly influential according to DFBETAS; these participants are not included in the figure summary statistics.
Supplemental Material S2. Performance pre and post training at the Day1:AM session (Training model); summary statistics shown for all participants (error bars give standard error), while individual points are only shown for those participants in the High AI group with no diagnosis or pending diagnosis of ASD.

Supplemental Material S3. Performance at sessions Day1:AM-post, Day1:PM and Day2:AM (TimeCourse model); summary statistics shown for all participants (error bars give standard error), while individual points are only shown for those participants in the High AI group with no diagnosis or pending diagnosis of ASD.

Knowland, V. C. P., Fletcher, F., Henderson, L.-M., Walker, S., Norbury, C. F., & Gaskell, M. G. (2019). Sleep promotes phonological learning in children across language and autism spectra. Journal of Speech, Language, and Hearing Research. Advance online publication. https://doi.org/10.1044/2019_JSLHR-S-19-0098