Consistent reconstruction of cortical surfaces from longitudinal brain MR images

Area of Science:

• Neuroimaging
• Medical Image Analysis
• Computational Anatomy

Background:

• Longitudinal studies of the human brain require accurate reconstruction of cortical surfaces.
• Subtle changes in the cerebral cortex over time are crucial for understanding neurological development and disease progression.
• Existing methods may lack the consistency and accuracy needed for detailed longitudinal analysis.

Purpose of the Study:

• To develop a novel deformable surface method for accurate and consistent reconstruction of inner, central, and outer cortical surfaces from longitudinal brain MR images.
• To enable precise measurement of longitudinal changes in cortical thickness.
• To assess the utility of these measurements in distinguishing between normal and diseased clinical groups.

Main Methods:

• A deformable surface method driven by a force derived from Laplace's equation was used.
• Cortical surfaces were initially reconstructed from a group-mean image of aligned longitudinal scans.
• Longitudinal surfaces were then consistently reconstructed by jointly deforming the group-mean surfaces to all individual time points.

Main Results:

• The method was successfully applied to two sets of longitudinal human brain MR images.
• Qualitative and quantitative results confirmed the accuracy and consistency of the reconstructed cortical surfaces.
• The reconstructed surfaces enabled clear observation of the overall decline trend in cortical thickness over time.

Conclusions:

• The proposed deformable surface method provides accurate and consistent reconstruction of longitudinal cortical surfaces.
• This approach effectively measures longitudinal changes in cortical thickness.
• The method shows potential for differentiating clinical groups based on cortical thickness trajectories.