Single input optimal control for globally coupled neuron networks.

Area of Science:

• Computational Neuroscience
• Systems Neuroscience
• Nonlinear Dynamics

Background:

• Synchronized neural activity is fundamental to brain function but can lead to pathological states.
• Controlling large-scale neural network dynamics, particularly desynchronization, remains a significant challenge.
• Existing methods often require complex interventions or lack precise control over network states.

Purpose of the Study:

• To develop and apply a novel computational method for desynchronizing globally coupled neural networks.
• To investigate the efficacy of controlling neural synchrony via targeted input to a single neuron.
• To assess the applicability of this control strategy across different neural models, including the Kuramoto and Hodgkin-Huxley models.

Main Methods:

• Application of discrete-time dynamic programming to reduced phase models of neural populations.
• Numerical minimization of a cost function across the entire state space to determine optimal control inputs.
• Testing the method on the Kuramoto model and a reduced phase model of Hodgkin-Huxley neurons with electrotonic coupling.

Main Results:

• The dynamic programming approach successfully generated control inputs to desynchronize neural networks.
• Effectiveness was evaluated by averaging control inputs across various coupling strengths for robustness.
• The method demonstrated applicability to Hodgkin-Huxley models, even when driven by multiplicative stimuli.

Conclusions:

• Discrete-time dynamic programming provides an effective strategy for desynchronizing coupled neural networks.
• Targeting a single neuron with precisely calculated inputs can disrupt population-level synchrony.
• This computational approach offers a promising tool for understanding and manipulating neural dynamics.