typical models of plasticity suggest that synapses shape the dynamics of neurons by stabilizing connections between those two neurons (think hebbian plasticity). alternatively, freezable chaos considers synapses and neurons as separate but coupled. it suggests that the default state of a network is chaos — unstructured and spontaneous. plasticity’s role is to destabilize the the network allowing it to remain as dynamic as possible. the mechanism for working memory in this new framework is by stopping synaptic plasticity (freezing the chaos). this freezable chaos framework has advantages over traditional synaptic facilitation models
fig7
