Automatic whole brain MRI segmentation of the developing neonatal brain

Area of Science:

• Neuroimaging
• Developmental Neuroscience
• Medical Image Analysis

Background:

• Neonatal brain Magnetic Resonance (MR) imaging is crucial for assessing infant development.
• Automatic segmentation of neonatal brain MR images is challenging due to rapid developmental changes, low signal-to-noise ratio, and partial volume effects.
• Accurate segmentation is vital for quantitative analysis of brain growth and development.

Purpose of the Study:

• To develop and validate a novel framework for accurate intensity-based segmentation of the developing neonatal brain.
• To segment the neonatal brain into 50 distinct regions from early preterm to term-equivalent age.
• To improve upon existing atlas-based segmentation techniques for neonatal MR imaging.

Main Methods:

• A novel intensity-based segmentation algorithm incorporating a structural hierarchy and anatomical constraints was developed.
• The algorithm models brain intensities across the entire neonatal brain volume.
• The proposed method was compared against standard atlas-based techniques.

Main Results:

• The proposed segmentation method demonstrated improved label overlap compared to manual reference segmentations.
• The technique achieved highly accurate segmentation results across a wide range of gestational ages (24 weeks to term-equivalent age).
• The algorithm proved robust in segmenting the developing neonatal brain.

Conclusions:

• The novel segmentation framework provides accurate and robust analysis of neonatal brain development using MR imaging.
• This method advances the quantitative assessment of brain growth in infants, particularly during the critical preterm to term-equivalent period.
• The technique offers a significant improvement over standard methods for neonatal brain MR image segmentation.