Neonatal Brain Microstructure and Machine-Learning-Based Prediction of Early Language Development in Children Born Very Preterm

Rachel Vassar; Kornél Schadl; Katelyn Cahill-Rowley; Kristen Yeom; David Stevenson; Jessica Rose

doi:10.1016/j.pediatrneurol.2020.02.007

Neonatal Brain Microstructure and Machine-Learning-Based Prediction of Early Language Development in Children Born Very Preterm

Pediatr Neurol. 2020 Jul:108:86-92. doi: 10.1016/j.pediatrneurol.2020.02.007. Epub 2020 Mar 3.

Authors

Rachel Vassar¹, Kornél Schadl², Katelyn Cahill-Rowley³, Kristen Yeom⁴, David Stevenson⁵, Jessica Rose⁶

Affiliations

¹ Department of Orthopaedic Surgery, Stanford University School of Medicine, Stanford, California; Neonatal Neuroimaging Research Laboratory, Stanford University School of Medicine, Stanford, California; Division of Pediatric Neurology, Department of Neurology, University of California San Francisco, San Francisco, California. Electronic address: rvassar@stanford.edu.
² Department of Orthopaedic Surgery, Stanford University School of Medicine, Stanford, California; Neonatal Neuroimaging Research Laboratory, Stanford University School of Medicine, Stanford, California; Semmelweis University School of Medicine, Budapest, Hungary.
³ Department of Orthopaedic Surgery, Stanford University School of Medicine, Stanford, California; Neonatal Neuroimaging Research Laboratory, Stanford University School of Medicine, Stanford, California; Motion Analysis Laboratory, Lucile Packard Children's Hospital, Stanford, California; Department of Bioengineering, Stanford University, Stanford, California.
⁴ Division of Pediatric Neuroradiology, Department of Radiology, Stanford University, Stanford, California.
⁵ Neonatal Neuroimaging Research Laboratory, Stanford University School of Medicine, Stanford, California; Division of Pediatric Neonatology, Department of Pediatrics, Stanford University, Stanford, California.
⁶ Department of Orthopaedic Surgery, Stanford University School of Medicine, Stanford, California; Neonatal Neuroimaging Research Laboratory, Stanford University School of Medicine, Stanford, California; Motion Analysis Laboratory, Lucile Packard Children's Hospital, Stanford, California.

PMID: 32279900
DOI: 10.1016/j.pediatrneurol.2020.02.007

Abstract

Background: Very-low-birth-weight preterm infants have a higher rate of language impairments compared with children born full term. Early identification of preterm infants at risk for language delay is essential to guide early intervention at the time of optimal neuroplasticity. This study examined near-term structural brain magnetic resonance imaging (MRI) and white matter microstructure assessed on diffusion tensor imaging (DTI) in relation to early language development in children born very preterm.

Methods: A total of 102 very-low-birth-weight neonates (birthweight≤1500g, gestational age ≤32-weeks) were recruited to participate from 2010 to 2011. Near-term structural MRI was evaluated for white matter and cerebellar abnormalities. DTI fractional anisotropy, mean diffusivity, axial diffusivity, and radial diffusivity were assessed. Language development was assessed with Bayley Scales of Infant-Toddler Development-III at 18 to 22 months adjusted age. Multivariate models with leave-one-out cross-validation and exhaustive feature selection identified three brain regions most predictive of language function. Distinct logistic regression models predicted high-risk infants, defined by language scores >1 S.D. below average.

Results: Of 102 children, 92 returned for neurodevelopmental testing. Composite language score mean ± S.D. was 89.0 ± 16.0; 31 of 92 children scored <85, including 15 of 92 scoring <70, suggesting moderate-to-severe delay. Children with cerebellar asymmetry had lower receptive language subscores (P = 0.016). Infants at high risk for language impairments were predicted based on regional white matter microstructure on DTI with high accuracy (sensitivity, specificity) for composite (89%, 86%), expressive (100%, 90%), and receptive language (100%, 90%).

Conclusions: Multivariate models of near-term structural MRI and white matter microstructure on DTI may assist in identification of preterm infants at risk for language impairment, guiding early intervention.

Keywords: Diffusion tensor imaging; Early language development; Machine learning; Preterm infants.

Publication types

Research Support, N.I.H., Extramural
Research Support, Non-U.S. Gov't
Research Support, U.S. Gov't, Non-P.H.S.

MeSH terms

Cerebellum / diagnostic imaging*
Cerebellum / pathology
Diffusion Tensor Imaging
Female
Humans
Infant
Infant, Extremely Premature*
Infant, Very Low Birth Weight*
Language Development Disorders / diagnosis*
Language Development*
Language Tests
Machine Learning
Male
Neuropsychological Tests
Prognosis
Sensitivity and Specificity
White Matter / diagnostic imaging*
White Matter / pathology

Grants and funding

UL1 RR025744/RR/NCRR NIH HHS/United States