Abstract
Past work has reported inverted-U relationships between arousal and auditory task performance, but the underlying neural network mechanisms remain unclear. To make progress, we recorded auditory cortex activity from behaving mice during passive tone presentation and simultaneously monitored pupil-indexed arousal. In these experiments, the neural discriminability of tones was maximized at intermediate arousal, revealing a neural correlate of the inverted-U. We explained this arousal-dependent sound processing using a spiking model with clusters. In the model, stimulus discriminability peaked as the network transitioned from a multi-attractor phase exhibiting slow switching between metastable cluster activations (low arousal) to a single-attractor phase with uniform activity (high arousal). This transition also qualitatively captured arousal-induced reductions of neural variability observed in the data. Altogether, this study elucidates computational principles to explain interactions between arousal, neural discriminability, and variability and suggests that transitions in the dynamical regime of cortical networks could underlie nonlinear modulations of sensory processing.