Abstract
A synaptic network model is proposed to characterize both short-term and long-term memory. The model describes the calcium concentration-dependent bistability of CaMKII phosphorylation levels and the transition of CaMKII between DOWN and UP states under different stimulation protocols. Numerical results demonstrate that short-term memory can be characterized by a trajectory tending to zero, while long-term memory can be characterized by a heteroclinic trajectory based on the principle of winnerless competition. Additionally, the results suggest that the model forms memories more efficiently with faster and longer training.
Original language | English |
---|---|
Pages (from-to) | 7499-7525 |
Number of pages | 27 |
Journal | Nonlinear Dynamics |
Volume | 112 |
Issue number | 9 |
DOIs | |
State | Published - May 2024 |
Keywords
- AMPARs
- Hippocampus
- Memory
- Network