Pattern recall in networks of chaotic neurons

Area of Science:

• Computational neuroscience
• Complex systems dynamics

Background:

• Neural information processing relies on complex dynamics.
• Spiking neural networks (SNNs) are computational models inspired by biological neurons.
• Chaotic dynamics offer potential for rich information processing.

Purpose of the Study:

• To investigate the utility of chaotic dynamics in neural information processing.
• To present a novel chaotic spiking neural network model.
• To extend the model's capability to recall previously stabilized activity patterns.

Main Methods:

• Development of a novel chaotic spiking neural network model.
• Utilizing non-linear dynamic state (NDS) neurons with non-linear equations.
• Implementing a threshold-based spike output mechanism.
• Enabling time-delayed self-connections for network stabilization.

Main Results:

• The network generates a vast number of periodic activation patterns due to chaotic dynamics.
• The extended model successfully recalls specific activity patterns from a set of stabilized patterns.
• Demonstrated the potential for complex pattern recall in chaotic neural networks.

Conclusions:

• Chaotic dynamics are a powerful tool for neural information processing.
• The novel NDS neuron model provides a robust platform for generating and recalling complex neural activity patterns.
• This research opens new avenues for understanding and engineering brain-like computation.