Plateau potentials are instructive signals for behavioral timescale synaptic plasticity in the neocortex

Abstract

Learning occurs via the adjustment of synaptic weights across a variety of timescales. The mechanisms supporting these processes, from single-shot to iterative learning, are unclear. The prevailing model in the neocortex, spike-timing dependent plasticity (STDP), requires many pairings of precisely coordinated activity. This is difficult to reconcile with single-shot learning in behaving animals. In hippocampus, an alternative form of plasticity driven by plateau potentials (behavioral timescale synaptic plasticity; BTSP) alters synaptic weights in a single trial to reorganize spatial representations. Here we show that layer 5 pyramidal neurons (L5 PNs) of mouse primary visual cortex (V1) exhibit highly prevalent plateau potentials that drive single-shot changes in sensory representations. Spontaneously occurring and experimentally induced plateaus rapidly and persistently modified L5 PN responses to visual stimuli. In acute slices, plateau potentials drove synapse-specific plasticity with few repetitions across seconds-long pairing intervals. Our results demonstrate that plateau potentials in the neocortex rapidly reshape neuronal representations through BTSP. This instructive form of plasticity provides a mechanism for dynamic adjustment of neocortical synaptic weights to support rapid learning.