Combining Multi-atlas Segmentation with Brain Surface Estimation.

Area of Science:

• Neuroimaging
• Computational Neuroscience
• Medical Image Analysis

Background:

• Whole brain segmentation and cortical surface reconstruction are crucial for human brain investigation.
• Current independent methods lead to spatial inconsistencies, hindering integrated cortical analyses.
• Existing strategies like FreeSurfer's "segmentation to surface to parcellation" have limitations.

Purpose of the Study:

• To develop a novel method, Multi-atlas CRUISE (MaCRUISE), for self-consistent whole brain segmentation and cortical surface reconstruction.
• To achieve reliable multi-atlas segmentation and labeling concurrently with accurate cortical surface reconstruction.
• To overcome limitations of existing independent segmentation and surface reconstruction methods.

Main Methods:

• Proposed a "multi-atlas segmentation to surface" approach, MaCRUISE, integrating multi-atlas segmentation with the CRUISE cortical reconstruction method.
• Obtained 132 cortical/subcortical labels simultaneously before reconstructing volume-consistent surfaces.
• Incorporated fuzzy tissue memberships to address partial volume effects and used sulci locations for topologically consistent surfaces.

Main Results:

• MaCRUISE achieves self-consistent whole brain segmentation and cortical reconstruction without compromising surface accuracy compared to CRUISE.
• Demonstrated comparable accuracy to FreeSurfer.
• Showcased greater robustness across an elderly population.

Conclusions:

• MaCRUISE is the first method to achieve reliable multi-atlas segmentation and labeling with accurate, consistent cortical surface reconstruction.
• The novel approach enhances spatial consistency between segmentation and surface data.
• MaCRUISE offers a robust and accurate solution for brain imaging analyses, particularly in diverse populations.