Background
Extremely preterm infants spend their first three months of life on the neonatal intensive care unit. They are at high risk of later neurodevelopmental conditions like ADHD, but this is difficult to predict at an individual level. Prognostic models often utilise ‘static’ predictors like their weight at birth. However, their intensive care admission is associated to rich longitudinal data, which could characterise the non-linear ‘bumpiness’ of their clinical trajectory. For example, they may gradually get well, but with two episodes of sepsis which decelerate their progress each time.
Novelty & Importance
To understand neurodevelopmental risk in extremely preterm infants requires to model the dynamism of their clinical course. In a highly novel piece of work, we have previously shown the feasibility of mining their history of painful procedures, which is considered very important by these infants’ families, not least because these disturb sleep. Here you will extend this innovative approach by integrating multiple other relevant longitudinal data, such as body weight. A second novel aspect is our research group’s interest in biomedical data across multiple nested time scales, including a strand of research on physiological monitoring of signals such as the electroencephalograph (EEG), which vary millisecond by millisecond.
Aims & Objectives
You will test the feasibility of retrieving longitudinal data like body weight from routine health records, integrating these with single in-depth measures like EEG, and whether applying machine learning to these metrics to create individualised predictive models offers improved prognostic value.
References
[1] Laudiano-Dray et al. Quantification of neonatal procedural pain severity: a platform for estimating total pain burden in individual infants. 2020. Pain.
[2] Georgoulas et al. Sleep-wake regulation in preterm and term infants. 2021. Sleep.
[3] Whitehead. Co-developing sleep-wake and sensory foundations for cognition in the human fetus and newborn. 2025. Developmental Cognitive Neuroscience.
[4] Gelegen et al. Metabolic expenditure, neurodevelopment, and weight gain into early childhood after fetal growth restriction. 2025. Preprint at medRxiv: https://doi.org/10.1101/2025.03.03.25323221
[5] Whitehead. Families and patient involvement in designing a project to analyse routine clinical paediatric EEG recordings for research purposes. https://osf.io/epk4g/
[6] Golmohammadi and Whitehead. A co-produced guide for how to explain and use bioelectrical health data for research purposes. 2025. https://osf.io/b5kw3/
[7] Salazar de Pablo et al. Individualized prediction models in ADHD: A systematic review and meta-regression. 2024. Molecular Psychiatry.