Robust cortical thickness estimation in the presence of partial volumes using adaptive diffusion equation

Area of Science:

• Neuroimaging
• Computational Neuroscience
• Medical Image Analysis

Background:

• Automated cortical thickness estimation is vital for studying brain development, aging, and neurodegenerative diseases.
• Partial volume effects due to limited MRI spatial resolution can significantly impact thickness accuracy, especially with hard thresholding.
• The convoluted cortical structure exacerbates these partial volume challenges.

Purpose of the Study:

• To introduce a novel method for accurate cortical thickness estimation that explicitly addresses partial tissue volume effects.
• To develop a method robust to variations in MRI spatial resolution and scanner field strength.

Main Methods:

• A novel adaptive diffusion equation (ADE) method was developed for cortical thickness estimation.
• The ADE method incorporates gray matter fractions into the diffusivity term to account for partial tissue volumes.
• The proposed method was compared against established techniques including Laplace equation, linked distance metric, and FreeSurfer.

Main Results:

• The ADE method demonstrated robustness against finite voxel resolution and image blurring.
• Simulations, histological comparisons, and test-retest studies validated the method's accuracy and consistency.
• ADE showed superior consistency in multi-scanner test-retest studies compared to existing methods.

Conclusions:

• The novel adaptive diffusion equation (ADE) method provides accurate and histologically consistent cortical thickness estimates.
• ADE is robust to variations in image resolution and scanner field strength, outperforming existing methods.
• The method offers improved reliability for neuroanatomical studies across diverse imaging conditions.