A dynamic attractor network model of memory formation, reinforcement and forgetting

Area of Science:

• Neuroscience
• Computational Neuroscience

Background:

• Empirical evidence links frequent memory revisitation to enhanced recall.
• Familiar items correlate with larger hippocampal neural representations.

Purpose of the Study:

• To provide a mechanistic understanding of how hippocampal neural assemblies change based on stimulus presentation frequency.
• To model the dynamic evolution of memory representations.

Main Methods:

• Developed a rate attractor network model incorporating an online Hebbian learning rule.
• Included background firing activity, neural adaptation, and heterosynaptic plasticity.
• Simulated memory pattern presentation frequency to observe assembly dynamics.

Main Results:

• Demonstrated that Hebbian learning and background firing activity explain the relationship between memory assembly size and stimulation frequency.
• Showed that frequently stimulated assemblies grow independently, forming orthogonal representations.
• Identified labile connections in unstimulated assemblies, allowing neuron recruitment and facilitating forgetting.

Conclusions:

• The model provides a mechanistic explanation for how memory representation size and persistence are modulated by stimulation frequency.
• Highlights the role of neural plasticity and network dynamics in memory formation, consolidation, and forgetting.
• Suggests a neuronal basis for forgetting through the dynamic reallocation of neural resources.