A Novel Eigen-Volume-based Co-Activation Pattern Framework for Dynamic Functional Biomarkers of Multiple Sclerosis.

Area of Science:

• Neuroimaging
• Biomarker Discovery
• Neurology

Background:

• Multiple sclerosis (MS) monitoring relies on imaging biomarkers.
• Resting-state functional MRI (rs-fMRI) provides functional insights complementing structural imaging.
• Dynamic brain activity patterns may offer novel diagnostic and prognostic information in MS.

Purpose of the Study:

• To investigate a novel co-activation pattern (CAP) approach for dynamic rs-fMRI as a dual-purpose biomarker in MS.
• To assess the diagnostic utility of dynamic CAP features for distinguishing MS patients from healthy controls (HCs).
• To evaluate the correlation between dynamic CAP features and clinical disability (EDSS) in MS patients.

Main Methods:

• rs-fMRI data from 25 MS patients and 41 HCs were analyzed using a novel CAP-based dynamic approach.
• Dynamic CAP features included dwell time, persistence, and transition probabilities.
• LASSO regression was used for severity prediction, and performance was benchmarked against standard CAP and sliding-window (SW) methods.

Main Results:

• Significant differences in brain activity were found between MS patients and HCs within the default mode, sensorimotor, and language networks (p < 0.05).
• Transition probabilities demonstrated strong correlations with EDSS (r > 0.75) and superior classification performance compared to standard CAP and SW approaches.
• Network shifts, particularly transitions toward sensory, motor, and executive networks, correlated with disease severity and compensatory recruitment.

Conclusions:

• Dynamic brain activity patterns are altered in MS and are linked to clinical disability.
• The proposed dynamic CAP approach enhances the distinction between MS patients and HCs, aiding clinical monitoring.
• Transition probabilities show promise as a biomarker for tracking MS progression and understanding adaptive reorganization in response to disease severity.