Atypical cortical encoding of the low-frequency temporal dynamics of natural speech identifies children with Developmental Language Disorder

Abstract

Developmental Language Disorder (DLD) is a neurodevelopmental condition that causes significant difficulties in understanding and using spoken language. Here we use electroencephalography (EEG) recorded during natural speech listening with 9-year-old children to identify dynamic neural processing patterns that characterize children with DLD compared to typically-developing age-matched controls. We applied Principal Components Analysis (PCA) to identify spatial ensembles of channels that represented distinct (uncorrelated) sources of cortical activity, and explored phase-amplitude coupling (PAC) between delta (0.5-4 Hz), theta (4-8 Hz) and low-gamma (25-40 Hz) oscillations. We then isolated EEG common spatial patterns (CSP) that identified children with DLD. The PAC analyses identified the delta band as a key source of group differences, and only delta-low gamma PAC differed significantly between participant groups. The CSP analyses also identified the delta band as a key mechanistic source of group differences. The findings are suggestive of distinct atypical low-frequency neural dynamics during speech encoding for children with DLD, which could be targeted by novel interventions.