Homeodynamic feedback inhibition control in whole-brain simulations

Area of Science:

• Computational Neuroscience
• Neuroimaging
• Mathematical Modeling

Background:

• Large-scale brain simulations often suffer from overexcitation due to network structures, yielding unrealistic results.
• Existing models struggle to achieve biologically plausible dynamic regimes and accurately fit empirical data.

Purpose of the Study:

• To implement a homeodynamic plasticity mechanism, dynamic Feedback Inhibitory Control (dFIC), within The Virtual Brain (TVB) simulation framework.
• To enable precise control over brain activity levels in simulations for enhanced biological realism.
• To improve the fitting of computational brain models to functional magnetic resonance imaging (fMRI) data.

Main Methods:

• Integration of a homeodynamic plasticity mechanism into the Jansen-Rit neural mass model.
• Heterogeneous adjustment of inhibitory coupling weights to achieve target dynamic regimes.
• Analytical and simulation-based derivation of convergence conditions for simulated brain activity.

Main Results:

• Demonstrated control over target activity levels, leading to biologically plausible simulations (e.g., post-synaptic potentials, fMRI BOLD signals).
• The dynamic Feedback Inhibitory Control (dFIC) method allows for increased variability in model dynamics.
• Successful application of dFIC in fitting TVB models to static and dynamic fMRI measures, considering global brain synchronization.

Conclusions:

• The novel dynamic Feedback Inhibitory Control (dFIC) method offers computational neuroscientists a tool to tune brain models to specific dynamical regimes.
• This approach enhances biological realism in brain network models and aids in understanding their behavior.
• dFIC facilitates better model fitting to empirical fMRI data, improving the utility of The Virtual Brain (TVB) framework.