Multiple Connection Pattern Combination From Single-Mode Data for Mild Cognitive Impairment Identification.

Area of Science:

• Neuroscience
• Medical Imaging
• Artificial Intelligence

Background:

• Mild cognitive impairment (MCI) is a critical predictor of Alzheimer's disease (AD) progression.
• Brain connectivity (BC) from fMRI is a validated biomarker for MCI, but single-pattern analyses are often insufficient.
• Identifying underlying neural changes requires integrating multiple connectivity patterns.

Purpose of the Study:

• To develop and validate a novel Multiple Connection Pattern Combination (MCPC) approach for improved MCI diagnosis.
• To enhance the identification of differences between MCI patients and normal controls (NCs).
• To overcome the limitations of single-pattern analyses in neuro-disease diagnosis.

Main Methods:

• A novel Multiple Connection Pattern Combination (MCPC) approach was developed.
• The MCPC method utilizes a kernel combination trick to integrate diverse connection patterns.
• The approach was validated using fMRI data from 63 MCI cases and 64 NC cases from the ADNI dataset.

Main Results:

• The proposed MCPC method achieved a classification accuracy of 87.40%.
• The MCPC approach significantly outperformed methods relying on single connection patterns.
• This demonstrates the efficacy of combining multiple patterns for MCI detection.

Conclusions:

• The MCPC approach offers a more effective strategy for diagnosing MCI by integrating multiple brain connectivity patterns.
• This method shows significant potential for early detection and prediction of Alzheimer's disease.
• Combining functional and effective connectivity patterns improves diagnostic performance in neurodegenerative disease research.