Simultaneous and consistent labeling of longitudinal dynamic developing cortical surfaces in infants.

Area of Science:

• Neuroimaging
• Developmental Neuroscience
• Computational Anatomy

Background:

• The first two years of life are critical for human cerebral cortex development.
• Accurate longitudinal parcellation of infant cortical surfaces is vital for studying brain development in healthy and high-risk infants.
• Existing cross-sectional methods lack longitudinal consistency, leading to inaccurate developmental measurements.

Purpose of the Study:

• To develop a novel method for accurate, consistent, and simultaneous labeling of longitudinal cortical surfaces in serial infant brain MRI scans.
• To improve the understanding of early cortical development by addressing limitations in existing parcellation techniques.

Main Methods:

• A new method formulated as an energy minimization problem with data fitting, spatial smoothness, and temporal consistency terms.
• Data fitting integrates multi-atlas contributions adaptively based on cortical folding similarity.
• Spatial and temporal consistency terms adaptively encourage label smoothness and vertex correspondence based on local cortical folding geometries.
• Efficient minimization achieved using a graph cuts method.

Main Results:

• The method was applied to longitudinal cortical surface parcellation of 13 healthy infants (0-18 months).
• Evaluations demonstrated high accuracy and longitudinal consistency of the proposed method.
• The study revealed novel properties of dynamic and regionally heterogeneous cortical surface area development.

Conclusions:

• The proposed method provides accurate and longitudinally consistent parcellation of infant cortical surfaces.
• This advancement enables a deeper understanding of early brain development and its regional variations.
• The findings offer new insights into the dynamic nature of cortical development in the first 18 months of life.