Integrated cortical structural marker for Alzheimer's disease.

Area of Science:

• Neuroimaging
• Neurology
• Biomedical Engineering

Background:

• Early diagnosis of Alzheimer's disease (AD) is crucial for effective management.
• Distinguishing very mild AD from normal cognition using neuroimaging is challenging.

Purpose of the Study:

• To develop and validate an approach integrating multiple cortical morphology measures for improved discrimination of very mild AD.
• To assess the combined diagnostic power of gray matter thickness and white matter geometric measures.

Main Methods:

• Applied FreeSurfer to MRI scans from 83 individuals with very mild AD and 124 controls.
• Generated measures of cortex thickness, white matter convexity (sulcal depth), and surface metric distortion.
• Utilized Principal Component Analysis (PCA), stepwise logistic regression, and 10-fold cross-validation for analysis.

Main Results:

• Cortical thickness, convexity, and metric distortion showed complementary patterns reflecting gray and white matter changes.
• The integrated classifier significantly outperformed classifiers based on single measures.
• Generated global and surface-based AD likelihood maps.

Conclusions:

• Integrating diverse cortical morphology measures enhances the ability to detect very mild Alzheimer's disease.
• PCA-based integration offers a powerful framework for high-dimensional data analysis in early disease diagnosis.
• This approach holds promise for future early diagnosis of neurodegenerative diseases with abnormal brain surface patterns.