Detection of brain functional-connectivity difference in post-stroke patients using group-level covariance modeling

Area of Science:

• Neuroscience
• Medical Imaging
• Biostatistics

Background:

• Functional brain connectivity (fBC) derived from functional Magnetic Resonance Imaging (fMRI) signals shows promise for identifying brain pathologies.
• Probabilistic inter-subject comparisons are crucial for developing diagnostic biomarkers from fBC, but principled comparisons remain challenging.

Purpose of the Study:

• To develop a novel matrix-variate probabilistic model for comparing functional connectivity matrices on the Symmetric Positive Definite (SPD) manifold.
• To introduce a new algorithm for principled comparison of connectivity coefficients between brain regions.

Main Methods:

• A new matrix-variate probabilistic model was developed for inter-subject comparison of functional connectivity matrices.
• The model operates on the manifold of Symmetric Positive Definite (SPD) matrices.
• A novel algorithm was derived for comparing connectivity coefficients between pairs of regions.

Main Results:

• The model was applied to compare post-stroke patients with healthy controls.
• Neurologically-relevant differences in functional connectivity were identified.
• The proposed model demonstrated higher sensitivity compared to standard procedures.

Conclusions:

• This study presents the first report of functional connectivity differences between a single patient and a group.
• The developed model represents a significant advancement toward utilizing functional connectivity as a diagnostic tool in clinical practice.