Supervised contrastive learning enhances graph convolutional networks for predicting neurodevelopmental deficits in
Hailong Li1, Junqi Wang2, Zhiyuan Li3
1Imaging Research Center, Department of Radiology, Cincinnati Children's Hospital Medical Center, Cincinnati, OH, USA; Neurodevelopmental Disorders Prevention Center, Perinatal Institute, Cincinnati Children's Hospital Medical Center, Cincinnati, OH, USA; Artificial Intelligence Imaging Research Center, Cincinnati Children's Hospital Medical Center, Cincinnati, OH, USA; Department of Radiology, University of Cincinnati College of Medicine, Cincinnati, OH, USA.
Insights
Early prediction of neurodevelopmental deficits in very preterm infants is crucial. A novel graph convolutional network (GCN) model with supervised contrastive learning (SCL) accurately predicts these deficits using brain structural connectomes, enabling timely interventions.
Area of Science:
- Neuroscience
- Artificial Intelligence
- Pediatrics
Background:
- Very preterm (VPT) infants face high risks of neurodevelopmental deficits, often diagnosed late.
- Early prediction is vital for timely interventions, but current methods are limited.
- Brain structural connectome (SC) analysis shows promise but requires advanced modeling.
Purpose of the Study:
- To develop and evaluate deep learning models for early prediction of neurodevelopmental deficits in VPT infants.
- To leverage graph convolutional networks (GCNs) for analyzing graph-structured SC data.
- To enhance GCN model performance using supervised contrastive learning (SCL) for data scarcity.
Main Methods:
- Applied GCN models to brain SC data from VPT infants at term-equivalent age.
- Utilized SCL to improve model robustness and handle limited data.
- Trained and validated models on a cohort of approximately 280 VPT infants from the CINEPS study.
- Assessed neurodevelopmental outcomes (cognition, language, motor skills) at 2 years corrected age.
Main Results:
- The SCL-enhanced GCN model achieved AUCs of 0.72–0.75 for predicting neurodevelopmental deficits.
- This performance surpassed several competing models, demonstrating the effectiveness of the approach.
- The results validate the hypothesis that SCL improves GCN-based prediction accuracy.
Conclusions:
- SCL significantly enhances GCN models for predicting neurodevelopmental deficits in VPT infants.
- This approach offers a promising tool for early identification and intervention.
- Accurate, early prediction of neurodevelopmental outcomes is achievable using advanced AI on brain SC data.
Abstract:
Very preterm (VPT) infants (born at less than 32 weeks gestational age) are at high risk for various adverse neurodevelopmental deficits. Unfortunately, most of these deficits cannot be accurately diagnosed until the age of 2-5 years old. Given the benefits of early interventions, accurate diagnosis and prediction soon after birth are urgently needed for VPT infants. Previous studies have applied deep learning models to learn the brain structural connectome (SC) to predict neurodevelopmental deficits in the preterm population. However, none of these models are specifically designed for graph-structured data, and thus may potentially miss certain topological information conveyed in the brain SC. In this study, we aim to develop deep learning models to learn the SC acquired at term-equivalent age for early prediction of neurodevelopmental deficits at 2 years corrected age in VPT infants. We directly treated the brain SC as a graph, and applied graph convolutional network (GCN) models to capture complex topological information of the SC. In addition, we applied the supervised contrastive learning (SCL) technique to mitigate the effects of the data scarcity problem, and enable robust training of GCN models. We hypothesize that SCL will enhance GCN models for early prediction of neurodevelopmental deficits in VPT infants using the SC. We used a regional prospective cohort of ∼280 VPT infants who underwent MRI examinations at term-equivalent age from the Cincinnati Infant Neurodevelopment Early Prediction Study (CINEPS). These VPT infants completed neurodevelopmental assessment at 2 years corrected age to evaluate cognition, language, and motor skills. Using the SCL technique, the GCN model achieved mean areas under the receiver operating characteristic curve (AUCs) in the range of 0.72∼0.75 for predicting three neurodevelopmental deficits, outperforming several competing models. Our results support our hypothesis that the SCL technique is able to enhance the GCN model in our prediction tasks.
More Related Videos
11:14A Novel Experimental and Analytical Approach to the Multimodal Neural Decoding of Intent During Social Interaction in Freely-behaving Human Infants
Published on: October 4, 2015
08:51Statistical Modelling of Cortical Connectivity Using Non-invasive Electroencephalograms
Published on: November 1, 2019
