Thesis Project Form

Extremely preterm infants (born before 28 weeks of gestational age) are at high risk of developing neurological complications due to incomplete brain development. Magnetic Resonance Imaging (MRI) is a key non-invasive tool for assessing neonatal brain maturation, but its use in early preterm infants remains understudied, particularly from a quantitative perspective. To date, no standardized or validated preprocessing and segmentation pipeline exists specifically for this gestational age group.

MRI Preprocessing:
Develop or adapt a robust preprocessing pipeline for neonatal brain MR images, including bias field correction, intensity normalization, and spatial registration, tailored to early preterm anatomy.

Automated Tissue Segmentation:
Implement and evaluate an automatic segmentation method to distinguish and quantify volumes of white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF). Both classical neuroimaging tools (e.g., SPM, FSL) and deep learning approaches (e.g., U-Net, nnU-Net) will be considered.

Volumetric Analysis:
Extract absolute and normalized brain tissue volumes (WM, GM, CSF), and assess inter- and intra-subject variability.

Clinical Correlation:
Analyze associations between brain tissue volumes and available clinical parameters (e.g., neurological scores, developmental outcomes, demographic or vital data) using statistical or machine learning techniques.

Gain expertise in medical image preprocessing, MRI brain tissue segmentation, and volumetric analysis using classical and deep learning methods. Learn to apply statistical and machine learning tools to correlate imaging biomarkers with clinical data in extremely preterm infants.