Release probability distribution modulates attractor state stability and mobility in continuous-attractor neural networks

We investigate how the distribution of synaptic release probabilities influences neural network dynamics using a continuous-attractor neural network (CANN) model with short-term synaptic depression. Motivated by experimental findings showing that astrocytic NMDA receptor blockade narrows neurotransmitter release probability distributions, we incorporate gamma-distributed release probabilities into our model. Our analysis reveals that the variation of release probabilities modulates attractor state stability within the network. Specifically, release probability variation narrowing increases the threshold for translational instability and reduces the network’s diffusion constant under noise perturbations, while the variation widening affects the network oppositely. These findings suggest a novel mechanism by which astrocytic NMDA receptors may modulate neural circuit dynamics and information processing through the regulation of synaptic release probability distributions.<br/><br/>© 2025 The Authors