Brain structure assessed using diffusion MRI and function assessed using quantitative EEG in very preterm infants and the ability to predict neurodevelopmental outcomes

Abstract

Background: Very preterm infants are at risk of neurodevelopmental impairments including visual and cognitive deficits. Early identification and prognostication of at-risk infants may enable early intervention and support, thus optimising outcomes. Electroencephalography (EEG) and Magnetic Resonance Imaging (MRI) are methods sensitive to aspects of brain maturation and may provide potential predictive biomarkers of deficits.Aims: To investigate the associations between EEG or MRI at term equivalent age (TEA) and: 1) visual function at TEA;2) visual function at 3 months corrected age (CA); and3) cognitive function at 12 months CA.Methods: Very preterm infants (n=146, born ≤30weeks gestational age, GA) without any major congenital or chromosomal abnormality were recruited. EEG was recorded at TEA and artefact-free data during active sleep were extracted for quantitative analyses. Power spectral, coherence and partial directed coherence (PDC) analyses were conducted in each frequency band at chosen electrodes. Frequency ranges analysed were: delta (0.5-2Hz), theta (2-6Hz), alpha (3-13Hz), and beta (13-30Hz). Parietal and occipital electrodes were chosen for vision-related analyses and frontal and central electrodes for cognitionrelated analyses.Brain MRI was carried out at TEA. Diffusion tensor model was used to estimate the fractional anisotropy (FA) and mean diffusivity (MD). Constrained spherical deconvolution and whole brain probabilistic tractography were computed. Tracts of interest were optic radiation (OR) for vision-related analyses and corpus callosum (CC) for cognition-related analyses. Mean FA and MD for each tract were calculated.