Group Similarity Constraint Functional Brain Network Estimation for Mild Cognitive Impairment Classification.

Area of Science:

• Neuroscience
• Medical Imaging
• Biomarker Discovery

Background:

• Functional brain networks (FBNs) are crucial biomarkers for brain function and neurodegenerative disease diagnosis.
• Accurate FBN estimation is challenging due to poor functional magnetic resonance imaging (fMRI) data quality and limited understanding of brain complexity.

Purpose of the Study:

• To propose a novel FBN estimation model incorporating group similarity prior.
• To enhance FBN modeling by extending it to tensor form with a tensor trace-norm regularizer for group similarity constraints.

Main Methods:

• Developed a novel tensor-based FBN estimation model.
• Incorporated a tensor trace-norm regularizer to enforce group similarity constraints.
• Applied the model to classify mild cognitive impairments (MCIs) from normal controls (NCs) using estimated FBNs.

Main Results:

• The proposed method effectively modeled FBNs.
• Achieved a classification accuracy of 91.97% for MCI detection, surpassing state-of-the-art methods.
• Post hoc analysis confirmed the generation of more biologically meaningful functional brain connections.

Conclusions:

• The novel tensor-based FBN estimation model with group similarity prior is effective for biomarker discovery.
• This method demonstrates significant potential for improving the diagnosis of neurodegenerative diseases like MCI.
• The enhanced FBNs provide more biologically meaningful insights into brain connectivity.