Designing spontaneous behavioral switching via chaotic itinerancy.

Area of Science:

• Neuroscience
• Dynamical Systems Theory
• Computational Neuroscience

Background:

• Chaotic itinerancy, characterized by transitions among quasi-attractors, is observed in high-dimensional nonlinear systems.
• This phenomenon is crucial for spontaneous behavior generation in animal brains and is of interest for neurorobotics.
• Controlling high-dimensional nonlinear dynamical systems remains a significant challenge.

Purpose of the Study:

• To propose a reproducible method for implementing desired chaotic itinerancy.
• To enable straightforward design of quasi-attractor trajectories and transition rules.
• To facilitate autonomous behavioral control in neurorobotics.

Main Methods:

• Development of a novel method for reproducible chaotic itinerancy.
• Application to a high-dimensional chaotic neural network model.
• System parameter adjustment to control attractor dynamics.

Main Results:

• Successful implementation of reproducible chaotic itinerancy.
• Demonstrated ability to design quasi-attractor trajectories and transition rules.
• Leveraging intrinsic high-dimensional chaos for control.

Conclusions:

• The proposed method simplifies the design of chaotic itinerancy in complex neural networks.
• This facilitates the creation of autonomous behavioral controls for neurorobotic applications.
• Offers a pathway to better understand and engineer brain-like spontaneous behaviors.