Applying Joint Graph Embedding to Study Alzheimer's Neurodegeneration Patterns in Volumetric Data

Area of Science:

• Neuroimaging
• Network Neuroscience
• Biostatistics

Background:

• Volumetric MRI is a potential Alzheimer's Disease (AD) biomarker but lacks specificity.
• Analyzing whole-brain spatial patterns of neurodegeneration may enhance diagnostic utility.
• Current methods often focus on local changes, potentially missing broader network effects.

Purpose of the Study:

• To develop and validate a network-based analytical framework for studying neurodegeneration in Alzheimer's Disease.
• To extend graph embedding algorithms for analyzing morphometric connectivity from longitudinal structural MRI data.
• To identify novel network structures associated with Alzheimer's Disease progression.

Main Methods:

• Utilized a multiple random eigengraphs framework and a modified multigraph embedding algorithm.
• Estimated maximum likelihood edge probabilities from population-specific network modes and subject-specific loadings.
• Implemented a novel statistical testing procedure with permutation testing for group difference analysis and confounder control.

Main Results:

• The analysis revealed networks dominated by structures known to be associated with Alzheimer's Disease neurodegeneration.
• The proposed framework demonstrated promise in identifying AD-related neurodegenerative patterns.
• Novel network-structure tuples, not detectable by traditional methods, were identified.

Conclusions:

• The developed network analysis framework shows significant potential for advancing the study of Alzheimer's Disease.
• This approach offers a more specific and comprehensive method for identifying neurodegenerative biomarkers in AD.
• The findings suggest a promising new direction for Alzheimer's Disease research using advanced neuroimaging analysis.