The Role of Inhibitory Neurons in Deviance Sound Detection in Regular and Random Statistical Contexts

Detecting statistical regularities in sound and responding to violations of these patterns, termed deviance detection, is a core function of the auditory system. In the human brain, studies have shown that deviance responses are enhanced in regular compared with random auditory contexts, but the underlying neuronal circuit mechanisms remain unclear. Here, we examined how inhibitory neurons contribute to context-dependent deviance responses in the mouse auditory cortex (AC). Using two-photon calcium imaging in AC of awake head-fixed male and female mice, we recorded neuronal activity during presentation of spectrotemporally rich moving ripple sounds, with deviant ripples embedded in either regular or random ripple sequences. AC neurons exhibited stronger responses to deviant sounds in regular contexts compared with random ones. To identify circuit mechanisms, we selectively inactivated parvalbumin (PV), somatostatin (SST), or vasoactive intestinal polypeptide (VIP) inhibitory neurons during the deviant stimulus presentation. Inactivation of PV and SST neurons broadly increased deviance responses in both contexts. In contrast, VIP inactivation selectively reduced responses to deviant stimuli in the regular, but not random, context, decreasing the context-dependent deviance signal enhancement. At the population level, inactivating all three neuronal subtypes increased detectability of the deviant stimulus, but the effects were context-dependent only for VIP inactivation. These findings reveal the distinct role of VIP neurons in modulating deviance signals based on context regularity, identifying a specific inhibitory neuron type that is critical for context-sensitive auditory processing and predictive coding.