Neonatal brain-age models in full- and preterm infants.

Area of Science:

• Neuroimaging
• Developmental Neuroscience
• Biomarker Discovery

Background:

• Prematurity poses significant risks to infant brain development and neurodevelopmental outcomes.
• Current biomarkers for identifying at-risk infants are limited.
• Diffusion magnetic resonance imaging (dMRI) offers insights into white matter development.

Purpose of the Study:

• To develop and validate brain-age prediction models using white matter features from dMRI.
• To assess the model's performance in both healthy and high-risk preterm infant cohorts.
• To determine if brain-age gap predicts health complications in preterm infants.

Main Methods:

• Tractometry-based brain-age models were constructed using white matter features from dMRI data.
• Models were trained and validated on two datasets: the developing Human Connectome Project (dHCP) and a clinical cohort (LPCH).
• Model accuracy was evaluated by comparing predicted brain age to chronological age, and the brain-age gap was assessed against clinical complication scores.

Main Results:

• White matter features demonstrated high predictive accuracy for brain age in both dHCP (within 3.9 days) and LPCH (within 6.6 days) datasets.
• The brain-age gap, derived from white matter maturation, showed no significant association with a composite score of common prematurity complications.
• Tractometry-derived brain-age models accurately reflect neonatal brain maturation but appear limited in detecting clinical complications.

Conclusions:

• Brain-age models based on white matter maturation are effective in characterizing neonatal brain development.
• Current white matter maturation metrics may lack the sensitivity to capture the cumulative impact of prematurity-related health issues.
• Multimodal or longitudinal biomarkers may be necessary to better assess health outcomes in high-risk preterm infants.