Hand function after neonatal stroke: A graph model based on basal ganglia and thalami structure.

Area of Science:

• Neuroscience
• Developmental Biology
• Medical Imaging

Background:

• Neonatal arterial ischemic stroke (NAIS) is a key model for studying early brain injury and its long-term effects on neurodevelopment.
• Motor difficulties after NAIS often affect both the contra-lesioned (affected) and ipsilesional (unaffected) hands, complicating prediction.
• Predicting hand motor function from standard MRI is challenging, suggesting a need for advanced analytical approaches.

Purpose of the Study:

• To investigate the association between structural brain parameters in the basal ganglia (BG) and thalamus with hand motor function in children with NAIS.
• To develop a predictive model for hand motor function using macrostructural information from the BG and thalamus.

Main Methods:

• A cohort of 35 children with middle cerebral artery NAIS underwent structural T1-weighted 3D MRI and manual dexterity assessment (Box and Blocks Test).
• Graph representations of BG and thalamic structures (volumes, elongations, distances) were analyzed.
• A graph neural network (GNN) was employed for graph regression to predict hand motor function, using a BBT score ratio to account for behavioral and cognitive factors.

Main Results:

• The GNN model demonstrated a significant correlation between predicted and actual hand motor function score ratios for both hands (p < 0.05).
• Prediction accuracy was high, with a mean L1 distance < 0.03.
• Structural features influenced each hand differently: affected hand function correlated more with volume, while unaffected hand function correlated more with structural elongation.

Conclusions:

• Macrostructural characteristics of the basal ganglia and thalamus are significantly correlated with hand motor function in both hands post-NAIS.
• These findings highlight the importance of considering the overall structural organization of BG and thalamic networks, not just volume.
• MRI-based macrostructural features of the basal ganglia and thalamus show promise as early biomarkers for predicting hand motor outcomes after early brain injury.