Switches to slow rhythmic neuronal activity lead to a plasticity-induced reset in synaptic weights

Abstract

AbstractMemory formation relies on changes in synaptic connections via synaptic plasticity, which is driven by neuronal activity. Simultaneously, behavior is controlled by switches in brain states: quiet waking is characterized by large low-frequency fluctuations in cortical activity that are strongly suppressed during active waking. We use computational modeling to study how switches in brain states interact with memory formation. We show that the low-frequency rhythmic bursting activity reminiscent of quiet waking induces a reset in synaptic weights. This reset occurs regardless of the implemented synaptic plasticity rule, whether phenomenological or biophysical, and is robust to biological variability and network heterogeneity. Mathematical analysis shows that the reset is rooted in the endogenous nature of the low-frequency quiet waking rhythmic activity. This brain state-dependent reset in synaptic weights restores the ability to learn through synaptic plasticity rules between periods of active waking, which could be a key component of memory homeostasis.