Discovery and visualization of structural biomarkers from MRI using transport-based morphometry.

Area of Science:

• Neuroimaging
• Computational Anatomy
• Biomedical Engineering

Background:

• Brain diseases often involve subtle changes missed by current imaging analysis.
• Existing computer-assisted methods rely on predefined features, potentially overlooking complex structural alterations.
• Magnetic resonance imaging (MRI) can reveal changes, but advanced analysis is needed for subtle alterations.

Purpose of the Study:

• To develop an automated pattern analysis technique for brain structure-phenotype relationships.
• To introduce transport-based morphometry (TBM) for lossless image transformation and enhanced separability.
• To validate TBM's ability to identify aging patterns and the role of aerobic fitness in brain health.

Main Methods:

• Developed transport-based morphometry (TBM), an image transformation technique.
• Applied TBM to structural brain images of healthy older adults.
• Utilized linear models for discrimination, regression, and blind source separation within the TBM framework.

Main Results:

• TBM successfully identified characteristic aging-related changes in brain structure.
• The method highlighted potential mechanisms of aerobic fitness in mediating brain health.
• TBM demonstrated the ability to visualize physically meaningful tissue distribution shifts via inverse transformation.

Conclusions:

• Transport-based morphometry (TBM) offers a powerful, feature-free approach to brain image analysis.
• TBM can uncover aging patterns and fitness-related brain health mechanisms.
• The TBM framework holds potential for elucidating genotype-structural-behavioral associations in various diseases.