High-order resting-state functional connectivity network for MCI classification.

Area of Science:

• Neuroimaging
• Computational Neuroscience
• Biomarkers for Neurodegenerative Diseases

Background:

• Resting-state functional magnetic resonance imaging (RS-fMRI) is used to estimate brain functional connectivity (FC) networks for diagnosing neurodegenerative diseases.
• Conventional low-order FC networks, based on simple correlations, may lack sensitivity for effective disease biomarker identification.
• There is a need for more sophisticated network analysis to capture complex brain interactions.

Purpose of the Study:

• To propose and evaluate a novel method for extracting high-order functional connectivity (FC) correlations.
• To enhance the diagnostic accuracy for neurodegenerative diseases by utilizing these high-order FC features.
• To develop a pattern classifier integrating both low-order and high-order FC network features.

Main Methods:

• Utilized a sliding window approach on RS-fMRI time series to generate short, overlapping segments.
• Constructed low-order FC networks for each segment to capture short-term regional correlations and their dynamics.
• Developed a high-order FC network by clustering correlation time series and calculating inter-cluster correlations, followed by classifier design.

Main Results:

• Experimental results demonstrated the effectiveness of the proposed high-order FC network approach.
• The combined features of low-order and high-order FC networks improved disease diagnosis accuracy.
• High-order FC networks provide a more sensitive biomarker for neurodegenerative diseases.

Conclusions:

• High-order functional connectivity networks offer a promising advancement over traditional low-order methods for neuroimaging analysis.
• This novel approach enhances the potential of RS-fMRI for accurate and reliable diagnosis of neurodegenerative diseases.
• Integrating high-order FC features into pattern classifiers significantly improves diagnostic performance.