Abstract 109: Towards Individualised Prediction of Neurodevelopmental Outcomes After Neonatal Stroke: A Brain Lesion Network Mapping Study Leveraging Large-Scale Datasets

Abstract

Introduction: Prediction of adverse neurodevelopmental outcomes after neonatal arterial ischemic stroke (NAIS) is challenging. We identified brain regions functionally connected to stroke lesions (lesion networks), and investigated whether such lesion networks are related to the development of cerebral palsy (CP) after NAIS. Methods: We performed a novel lesion network mapping approach, involving: 1. Creation of a normative connectome using the developing Human Connectome Project (dHCP) dataset. Resting state functional magnetic resonance imaging (MRI) scans of 518 neonates born at term (>36 weeks) from the dHCP were aligned to a template, built from T 2 -weighted scans of the same neonates; 2. Manual segmentation of lesions using diffusion MRI scans of 85 term-born neonates with NAIS, who were identified from Australian and Swiss pediatric stroke registries. Lesion masks were aligned to the template; 3. Construction of lesion network maps, by computing functional correlations between lesion masks and other brain grey matter regions, making use of the normative connectome; 4. Investigation of relationships between lesion networks and CP, diagnosed by clinical examination at least 18 months after NAIS. Lesion network maps were compared between those who did (n=28) and did not (n=57) develop CP after NAIS, with two sample t-tests. Results: In participants who developed CP after NAIS, lesions were more strongly functionally connected to the following regions: frontal (inferior and orbital frontal), temporal (pole, superior, and mesial temporal), insula, basal ganglia, thalamus and cerebellum (Figure 1). Conclusions: Leveraging large-scale datasets and innovative connectomic methods, we uncovered brain networks related to adverse outcomes after NAIS. This provided novel findings compared to past methods such as voxel-based lesion-symptom mapping. Our results point to differences in network vulnerabilities after stroke in neonates compared to adults.