Predicting the Conversion from Mild Cognitive Impairment to Alzheimer's Disease Using Graph Frequency Bands and...

Area of Science:

• Neuroscience
• Medical Imaging
• Computational Biology

Background:

• Accurate prediction of mild cognitive impairment (MCI) progression to Alzheimer's disease (AD) is critical for effective patient management.
• Machine learning (ML) and resting-state functional magnetic resonance imaging (rs-fMRI) show promise in classifying AD and MCI.
• Identifying individuals with progressive MCI (pMCI) versus stable MCI (sMCI) is key for timely interventions.

Purpose of the Study:

• To develop and validate an ML-based framework for predicting MCI to AD progression using rs-fMRI data.
• To identify a parsimonious set of connectivity-based features for accurate classification.
• To enhance the precision of early AD risk assessment.

Main Methods:

• Utilized three years of rs-fMRI data from 142 sMCI and 136 pMCI patients in the ADNI cohort.
• Applied graph signal processing to filter rs-fMRI data into low, middle, and high frequency bands.
• Extracted connectivity-based features, performed feature selection using particle swarm optimization (PSO) and simulated annealing (SA), and classified using support vector machine (SVM) with radial basis function (RBF) kernel and 10-fold cross-validation.

Main Results:

• The proposed framework achieved optimal accuracy with minimal feature utilization.
• Using PSO-selected features, the SVM model demonstrated 77% accuracy, 70% specificity, and 83% sensitivity.
• Key predictive features included graph metrics (clustering coefficient, strength, eccentricity, modularity) across different frequency bands.

Conclusions:

• The study highlights the efficacy of the proposed framework in identifying individuals at risk of AD development.
• A parsimonious feature set derived from filtered rs-fMRI data can accurately predict MCI to AD progression.
• This approach offers a promising tool for advancing precision in early AD diagnosis and intervention.