Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Longitudinal Cortical Thickness Trajectories Predict Language Outcomes and Persistent Verbal Vulnerability in Late Talkers
Poster Session E, Friday, October 2, 11:00 am - 1:00 pm, Wangari Maathai
Julie Schneider1, Emil Cornea2, John Gilmore2, Tehila Nugiel3; 1University of California Los Angeles, 2University of North Carolina School of Medicine, 3Florida State University
Approximately 15% of toddlers exhibit delayed expressive vocabulary development by age two and are classified as late talkers (LTs; Rescorla, 2011), placing them at increased risk for persistent language difficulties (Rescorla, 2011; Singleton, 2018). However, outcomes among LTs are highly variable, and early predictors of persistent language vulnerability remain poorly understood. Prior neuroimaging work suggests that atypical cortical maturation in language-related regions may characterize children with language difficulties (Dynek et al., 2021), but most studies have focused on children after language delays have already emerged. The current study examined whether longitudinal cortical thickness (CT) development from birth to age two within language-related brain networks was associated with concurrent toddler language ability, later verbal outcomes, and persistent language vulnerability. Data from 544 children (390 families) came from the Early Brain Development Study, a longitudinal pediatric neuroimaging dataset collected at The University of North Carolina at Chapel Hill. CT trajectories from Years 0–2 were examined within language-related functional networks using linear mixed-effects models with piecewise developmental intervals (Wave1 = change from Year 0 to Year 1; Wave2 = change from Year 1 to Year 2). Models tested associations between CT development and expressive language outcomes at Year 2 (Mullen Expressive Language T score) as well as verbal abilities at Year 6 (Stanford-Binet Verbal IQ and Verbal Reasoning/Knowledge Composite), controlling for total brain volume, family, gestational age, and parental education. For Year-2, within the ventral language network, higher expressive language ability was associated with steeper cortical thinning (more negative CT) during the earlier developmental interval (Wave1; β = -0.04, p_FDR = .037) and reduced thinning during the later interval (Wave2; β = 0.05, p_FDR = .0001). Greater gains in expressive language from Year 1-2 were also associated with more positive CT Wave2 slopes in both the ventral language network (β = 0.034, p_FDR = .015) and the core expressive/phonological network (β = 0.028, p_FDR = .037). Stronger verbal reasoning and verbal knowledge abilities at Year 6 were associated with thinner cortex overall, regardless of timepoint, within the semantic language network (β = -0.020, p_FDR = .041). Significant Wave2 interactions also emerged within the language control network (β = -0.044, p_FDR = .037), indicating that children with stronger verbal abilities exhibited greater cortical thinning from Years 1–2. Follow-up classification analyses tested whether CT developmental trajectories could distinguish children with persistent language vulnerability. Cortical trajectories showed limited ability to classify Year-2 late-talker status (AUCs ≈ .54–.60), but ventral language network trajectories strongly discriminated children with low Year-6 verbal IQ (<85; AUC = .92). Greater cortical thinning across development was associated with reduced odds of later verbal impairment. Together, these findings suggest that developmental cortical thinning within language-related association cortex—particularly ventral language regions—may serve as a neurodevelopmental marker of persistent verbal-language vulnerability.
Topic Areas: Language Development/Acquisition, Disorders: Developmental