Memory States and Transitions between Them in Attractor Neural Networks

Area of Science:

• Cognitive Neuroscience
• Computational Neuroscience

Background:

• Human memory retrieval exhibits associative capabilities, linking similar memories.
• Existing models lack mechanisms for actively exploiting memory similarities during retrieval.
• Current attractor neural network models often require memory extinction for transitions.

Purpose of the Study:

• To introduce a novel neural mechanism for memory transitions.
• To demonstrate how neural dynamics can actively exploit similarities between memories.
• To propose a model where neural populations act as attractors.

Main Methods:

• Introduction of a novel mechanism in neural network dynamics.
• Modeling neural populations selective for multiple memories.
• Analysis of the role of excitation-inhibition balance control.

Main Results:

• A new mechanism allows transitions between memories by exploiting similarities.
• Neural populations selective for multiple memories function as attractors.
• The mechanism relies on controlling the excitation-inhibition balance within the neural network.

Conclusions:

• This work presents a novel mechanism for memory transitions.
• Active exploitation of memory similarities is possible through neural dynamics.
• The proposed model offers new insights into the neural basis of associative memory.