Functional Mechanisms of Recovery after Chronic Stroke: Modeling with the Virtual Brain

Area of Science:

• Neuroscience
• Computational Neuroscience
• Medical Informatics

Background:

• Stroke recovery mechanisms require better translation from basic science to individualized therapies.
• Current approaches lack effective links between big data and personalized treatment strategies.

Purpose of the Study:

• To present an approach using The Virtual Brain (TVB) platform for modeling brain dynamics post-stroke.
• To simulate functional magnetic resonance imaging (fMRI) signals and identify potential biomarkers for stroke recovery.

Main Methods:

• Utilized rest fMRI, T1w, and diffusion tensor imaging (DTI) data from 20 stroke patients and 11 controls.
• Developed individual brain models in TVB using structural connectomes from DTI.
• Optimized local and global brain parameters, simulated blood-oxygen-level-dependent (BOLD) signals, and validated against empirical data.
• Performed regression analysis linking model parameters with clinical motor performance.

Main Results:

• Stroke patients showed reduced conduction velocity, increased local dynamics, and decreased local inhibitory coupling compared to controls.
• A negative correlation was observed between local excitation and motor recovery.
• Positive correlation found between local dynamics and motor recovery outcomes.

Conclusions:

• TVB modeling reveals a post-stroke brain state characterized by excitation-over-inhibition and local-over-global dynamics.
• The findings suggest TVB's potential for identifying individualized biomarkers to predict stroke recovery.
• This neuroinformatics approach offers a promising avenue for personalized stroke rehabilitation strategies.