Related Experiment Videos
Associative memory retrieval induced by fluctuations in a pulsed neural network
1RCAST, The University of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo 153-8904, Japan.
Summary
This study explores associative memory retrieval in pulsed neural networks using FitzHugh-Nagumo neurons. The research demonstrates the network
Area of Science:
- Computational neuroscience
- Complex systems
- Artificial intelligence
Background:
- FitzHugh-Nagumo neurons are a simplified model of neuron dynamics.
- Pulsed neural networks utilize discrete-time signaling for information processing.
- Associative memory relies on recalling stored information from partial cues.
Purpose of the Study:
- To investigate associative memory retrieval in a pulsed neural network.
- To explore the role of spatio-temporal firing patterns in memory representation.
- To analyze the impact of system fluctuations on memory recall.
Main Methods:
- Numerical simulations of a pulsed neural network model.
- Utilizing FitzHugh-Nagumo neuron dynamics.
- Analyzing spatio-temporal firing patterns and system fluctuations.
Main Results:
- Memory is encoded in the spatio-temporal firing patterns of neurons.
- System fluctuations facilitate memory retrieval.
- The network demonstrates the capacity for alternate retrieval of two distinct memory patterns.
- Storage capacity of the network is numerically investigated.
Conclusions:
- Pulsed neural networks based on FitzHugh-Nagumo neurons can perform associative memory retrieval.
- Spatio-temporal dynamics and system fluctuations are key mechanisms for memory recall.
- The network shows potential for storing and retrieving multiple memory patterns.