Machine Learning and Prediction in Fetal, Infant, and Toddler Neuroimaging: A Review and Primer

Brain-phenotype predictions of language and executive function can survive across diverse real-world data: Dataset shifts in developmental populations

Predictive modeling potentially increases the reproducibility and generalizability of neuroimaging brain-phenotype associations. Yet, the evaluation of a model in another dataset is underutilized. Among studies that undertake external validation, there is a notable lack of attention to generalization across dataset-specific idiosyncrasies (i.e., dataset shifts). Research settings, by design, remove the between-site variations that real-world and, eventually, clinical applications demand. Here, we rigorously test the ability of a range of predictive models to generalize across three diverse, unharmonized developmental samples: the Philadelphia Neurodevelopmental Cohort (n=1291), the Healthy Brain Network (n=1110), and the Human Connectome Project in Development (n=428). These datasets have high inter-dataset heterogeneity, encompassing substantial variations in age distribution, sex, racial and ethnic minority representation, recruitment geography, clinical symptom burdens, fMRI tasks, sequences, and behavioral measures. Through advanced methodological approaches, we demonstrate that reproducible and generalizable brain-behavior associations can be realized across diverse dataset features. Results indicate the potential of functional connectome-based predictive models to be robust despite substantial inter-dataset variability. Notably, for the HCPD and HBN datasets, the best predictions were not from training and testing in the same dataset (i.e., cross-validation) but across datasets. This result suggests that training on diverse data may improve prediction in specific cases. Overall, this work provides a critical foundation for future work evaluating the generalizability of brain-phenotype associations in real-world scenarios and clinical settings.

Brain age prediction and deviations from normative trajectories in the neonatal connectome

Structural and functional connectomes undergo rapid changes during the third trimester and the first month of postnatal life. Despite progress, our understanding of the developmental trajectories of the connectome in the perinatal period remains incomplete. Brain age prediction uses machine learning to estimate the brain's maturity relative to normative data. The difference between the individual's predicted and chronological age-or brain age gap (BAG)-represents the deviation from these normative trajectories. Here, we assess brain age prediction and BAGs using structural and functional connectomes for infants in the first month of life. We use resting-state fMRI and DTI data from 611 infants (174 preterm; 437 term) from the Developing Human Connectome Project (dHCP) and connectome-based predictive modeling to predict postmenstrual age (PMA). Structural and functional connectomes accurately predict PMA for term and preterm infants. Predicted ages from each modality are correlated. At the network level, nearly all canonical brain networks-even putatively later developing ones-generate accurate PMA prediction. Additionally, BAGs are associated with perinatal exposures and toddler behavioral outcomes. Overall, our results underscore the importance of normative modeling and deviations from these models during the perinatal period.

Development of neonatal connectome dynamics and its prediction for cognitive and language outcomes at age 2

The functional brain connectome is highly dynamic over time. However, how brain connectome dynamics evolves during the third trimester of pregnancy and is associated with later cognitive growth remains unknown. Here, we use resting-state functional Magnetic Resonance Imaging (MRI) data from 39 newborns aged 32 to 42 postmenstrual weeks to investigate the maturation process of connectome dynamics and its role in predicting neurocognitive outcomes at 2 years of age. Neonatal brain dynamics is assessed using a multilayer network model. Network dynamics decreases globally but increases in both modularity and diversity with development. Regionally, module switching decreases with development primarily in the lateral precentral gyrus, medial temporal lobe, and subcortical areas, with a higher growth rate in primary regions than in association regions. Support vector regression reveals that neonatal connectome dynamics is predictive of individual cognitive and language abilities at 2  years of age. Our findings highlight network-level neural substrates underlying early cognitive development.

Brain age prediction and deviations from normative trajectories in the neonatal connectome

Structural and functional connectomes undergo rapid changes during the third trimester and the first month of postnatal life. Despite progress, our understanding of the developmental trajectories of the connectome in the perinatal period remains incomplete. Brain age prediction uses machine learning to estimate the brain's maturity relative to normative data. The difference between the individual's predicted and chronological age-or brain age gap (BAG)-represents the deviation from these normative trajectories. Here, we assess brain age prediction and BAGs using structural and functional connectomes for infants in the first month of life. We used resting-state fMRI and DTI data from 611 infants (174 preterm; 437 term) from the Developing Human Connectome Project (dHCP) and connectome-based predictive modeling to predict postmenstrual age (PMA). Structural and functional connectomes accurately predicted PMA for term and preterm infants. Predicted ages from each modality were correlated. At the network level, nearly all canonical brain networks-even putatively later developing ones-generated accurate PMA prediction. Additionally, BAGs were associated with perinatal exposures and toddler behavioral outcomes. Overall, our results underscore the importance of normative modeling and deviations from these models during the perinatal period.

Automated neonatal nnU-Net brain MRI extractor trained on a large multi-institutional dataset

Brain extraction, or skull-stripping, is an essential data preprocessing step for machine learning approaches to brain MRI analysis. Currently, there are limited extraction algorithms for the neonatal brain. We aim to adapt an established deep learning algorithm for the automatic segmentation of neonatal brains from MRI, trained on a large multi-institutional dataset for improved generalizability across image acquisition parameters. Our model, ANUBEX (automated neonatal nnU-Net brain MRI extractor), was designed using nnU-Net and was trained on a subset of participants (N = 433) enrolled in the High-dose Erythropoietin for Asphyxia and Encephalopathy (HEAL) study. We compared the performance of our model to five publicly available models (BET, BSE, CABINET, iBEATv2, ROBEX) across conventional and machine learning methods, tested on two public datasets (NIH and dHCP). We found that our model had a significantly higher Dice score on the aggregate of both data sets and comparable or significantly higher Dice scores on the NIH (low-resolution) and dHCP (high-resolution) datasets independently. ANUBEX performs similarly when trained on sequence-agnostic or motion-degraded MRI, but slightly worse on preterm brains. In conclusion, we created an automatic deep learning-based neonatal brain extraction algorithm that demonstrates accurate performance with both high- and low-resolution MRIs with fast computation time.

Brain-phenotype predictions can survive across diverse real-world data

Recent work suggests that machine learning models predicting psychiatric treatment outcomes based on clinical data may fail when applied to unharmonized samples. Neuroimaging predictive models offer the opportunity to incorporate neurobiological information, which may be more robust to dataset shifts. Yet, among the minority of neuroimaging studies that undertake any form of external validation, there is a notable lack of attention to generalization across dataset-specific idiosyncrasies. Research settings, by design, remove the between-site variations that real-world and, eventually, clinical applications demand. Here, we rigorously test the ability of a range of predictive models to generalize across three diverse, unharmonized samples: the Philadelphia Neurodevelopmental Cohort (n=1291), the Healthy Brain Network (n=1110), and the Human Connectome Project in Development (n=428). These datasets have high inter-dataset heterogeneity, encompassing substantial variations in age distribution, sex, racial and ethnic minority representation, recruitment geography, clinical symptom burdens, fMRI tasks, sequences, and behavioral measures. We demonstrate that reproducible and generalizable brain-behavior associations can be realized across diverse dataset features with sample sizes in the hundreds. Results indicate the potential of functional connectivity-based predictive models to be robust despite substantial inter-dataset variability. Notably, for the HCPD and HBN datasets, the best predictions were not from training and testing in the same dataset (i.e., cross-validation) but across datasets. This result suggests that training on diverse data may improve prediction in specific cases. Overall, this work provides a critical foundation for future work evaluating the generalizability of neuroimaging predictive models in real-world scenarios and clinical settings.