Multimodal Imaging and AI-driven Diagnosis of Preterm Infant Brain Injury: A Comprehensive Review

Preterm brain injury critically impacts neurodevelopment. Imaging-such as cranial ultrasound, structural and diffusion magnetic resonance imaging (MRI), and metabolic/functional modalities-is essential for early detection, grading, and prognosis. However, neonatal imaging faces challenges including rapid development, motion, noise, and protocol variability. Recent advances in machine learning and deep learning enable automated analysis, multimodal feature extraction, and predictive modeling, even with limited or multi-center data. This review organizes the field into a clinical-imagingcomputational perspective, summarizes methods, addresses data and evaluation issues, and discusses future directions toward robust, interpretable, and clinically integrated artificial intelligence systems.