Predicting Outcome After Newborn Stroke: A Lesion Network Mapping Study Leveraging Large-Scale Data

Predicting the development of cerebral palsy after neonatal stroke remains challenging. This study aimed to identify novel acute brain functional connectome-based correlates of cerebral palsy after neonatal stroke.

Stroke lesions were segmented from routine clinical diffusion images of a cohort of term-born neonates with symptomatic arterial ischemic stroke, recruited to Swiss (from 2000 to 2013) and Australian (from 2003 to 2014) pediatric stroke registries. Lesions and 3-Tesla resting-state functional magnetic resonance imaging scans of term-born newborns from the developing Human Connectome Project were coregistered to a template. A neonatal stroke functional connectome was created by computing voxel-wise correlations between lesions and gray matter regions. Linear regressions compared functional connections to lesions between participants who did and did not develop cerebral palsy.

From the total N=199 recruited participants, 85 newborns with stroke were included (65% male; median age at magnetic resonance imaging of 4 days), of which 33% developed cerebral palsy at a median age of 2.1 years. Multiple gray matter regions were more highly functionally correlated to lesions in participants who developed cerebral palsy (1721 voxels; t: 5.4-7.4; all P<0.05, family-wise error rate-corrected). These regions included the basal ganglia, thalamus, cerebellum, frontal regions (inferior and orbital frontal and superior frontal), temporal regions (pole, superior, and mesial temporal, including hippocampus and amygdala), and the insula.

This study identified functional networks related to the development of cerebral palsy after neonatal stroke. Building on prior individual lesion-based studies, this work suggests that the development of cerebral palsy after neonatal stroke is related to disruptions of broader functional networks involving motor and extramotor regions, as opposed to only lesions in motor regions.