Machine learning techniques for predicting neurodevelopmental impairments in premature infants: a systematic review
Arantxa Ortega-Leon1, Daniel Urda2, Ignacio J Turias1
1Intelligent Modelling of Systems Research Group, Department of Computer Science Engineering, Algeciras School of Engineering and Technology (ASET), University of Cádiz, Algeciras, Spain.
Frontiers in Artificial Intelligence
|February 5, 2025
Summary
Machine learning (ML) models can predict Neurodevelopmental Impairments (NDIs) in very preterm infants. This review analyzes ML applications, suggesting multimodal data for improved early identification and intervention strategies.
Area of Science:
- Neonatal Medicine
- Artificial Intelligence
- Neuroscience
Background:
- Very preterm infants face high risks of Neurodevelopmental Impairments (NDIs), affecting cognitive, motor, and language skills.
- Early identification of NDIs is crucial for timely intervention in this vulnerable population.
Purpose of the Study:
- To systematically review the application of Machine Learning (ML) techniques for predicting NDIs in premature infants.
- To provide a comparative analysis of ML studies in this field from 2018-2023.
Main Methods:
- Systematic review of scientific literature published between January 2018 and December 2023.
- Analysis of 26 selected studies focusing on ML algorithms and data sources for NDI prediction.
- Exploration of common data types including clinical, neuroimaging, and omics data.
Main Results:
- Identified 26 relevant studies applying ML to predict NDIs in preterm infants.
- Commonly utilized data sources include clinical and neuroimaging data.
- Omics data is proposed as a valuable addition for enhanced predictive accuracy.
Conclusions:
- Multimodal data models are significant for overcoming limitations in current NDI prediction.
- Insights guide researchers and clinicians in developing improved early identification and intervention strategies.
- Enhanced ML approaches promise better outcomes for preterm infants at risk of NDIs.
