Sigmoidal synaptic learning produces mutual stabilization in chaotic FitzHugh-Nagumo model

Area of Science:

• Computational neuroscience
• Dynamical systems theory
• Neural modeling

Background:

• Investigates the interaction between coupled neurons at the end of a neural chain.
• Focuses on a bidirectional, two-cell FitzHugh-Nagumo neural model known for chaotic dynamics.

Purpose of the Study:

• To analyze how mutual stabilization of chaotic dynamics occurs in a neural model.
• To demonstrate the transition from chaotic to stable periodic behavior via synaptic learning.

Main Methods:

• Bifurcation analysis of an adapted FitzHugh-Nagumo model to identify periodic and chaotic regions.
• Dynamic adjustment of synaptic properties through neural learning to observe system evolution.

Main Results:

• Identified regions of periodic and chaotic behaviors in the neural model.
• Demonstrated that synaptic learning enables a transition from chaotic to stable periodic dynamics.
• Showcased mutual stabilization of chaotic dynamics between two coupled neurons.

Conclusions:

• Synaptic learning can serve as a mechanism for neurons to transition from chaotic to stable, ordered periodic states.
• Even simple systems of two coupled neurons can exhibit stabilization of chaotic behavior into sustained periodic activity.
• This stabilization occurs without the necessity of a large, complex neural network.