Alterations in dynamic connectivity in Alzheimer's disease: Network changes and improved multi-stage classification

Area of Science:

• Neuroimaging
• Neurodegenerative Diseases
• Brain Connectivity

Background:

• Alzheimer's disease (AD) involves progressive network disruption.
• Static functional connectivity (sFC) is well-studied, but dynamic functional connectivity (dFC) in AD remains unclear.

Purpose of the Study:

• To investigate group differences in sFC and dFC across the Alzheimer's disease spectrum.
• To assess the discriminative value of dFC for classifying cognitive impairment stages.
• To explore associations between functional connectivity metrics and cognitive function.

Main Methods:

• Resting-state fMRI data from 174 participants (CN, SMC, EMCI, LMCI, AD) were analyzed.
• Group differences in sFC, dFC, and graph-theoretical metrics were assessed.
• A BrainNetCNN model evaluated classification performance of sFC and dFC features.

Main Results:

• sFC decreased in MCI but increased in AD; dFC variability reduced in pre-dementia groups and increased in AD.
• dFC showed higher classification accuracy than sFC for binary and five-class tasks.
• Combined sFC and dFC features achieved the highest diagnostic accuracy (e.g., 82.7% for five-class).

Conclusions:

• dFC provides complementary information to sFC for characterizing AD-related network alterations.
• dFC metrics are valuable for differentiating diagnostic groups across the Alzheimer's disease spectrum.
• Dynamic functional connectivity analysis holds potential for improved AD diagnosis and staging.