White Matter Brain Network Research in Alzheimer's Disease Using Persistent Features.

Area of Science:

• Neuroscience
• Medical Imaging
• Network Science

Background:

• Alzheimer's disease (AD) poses a significant social burden, yet effective disease-modifying drugs are lacking.
• Structural brain network analysis offers insights into the physiological changes associated with AD and its precursor, mild cognitive impairment (MCI).
• Persistent homology is an emerging tool for studying brain network dynamics and global organization, but its application to AD/MCI structural networks requires further investigation.

Purpose of the Study:

• To investigate how persistent homology measures reflect topological changes in white matter (WM) networks of individuals with AD and MCI.
• To compare the efficacy of persistent homology-based measures against traditional graph-theoretical measures in detecting network alterations in AD and MCI.
• To assess the potential of these topological measures for tracking AD progression.

Main Methods:

• Utilized diffusion tensor imaging (DTI) to construct WM networks from 150 subjects (AD, MCI, and normal controls - NC).
• Quantified network topology using previously developed persistent homology features and standard graph-theoretical metrics.
• Analyzed network properties across different brain parcellation schemes.

Main Results:

• Significant differences in topological network measures were observed among AD, MCI, and NC groups.
• Both AD and MCI groups exhibited decreased network integration and increased network segregation compared to NC.
• Persistent homology measures demonstrated superior statistical power and robustness over traditional graph-theoretic approaches.

Conclusions:

• Persistent homology provides a more sensitive method for detecting structural brain network alterations in AD and MCI.
• These findings enhance the understanding of the structural connectome in AD and offer a novel approach for monitoring disease progression.
• The study highlights the potential of advanced topological methods in neurodegenerative disease research.