Subcortical coding of predictable and unsupervised sound-context associations

Abstract

The acoustic context is ubiquitous, highly informative and often predictable. Yet, we know little about where and how contextual sounds are coded in the brain. We have some understanding about the coding of reinforced sounds in the auditory cortex and subcortical midbrain. Contextual sounds, however, are learned in an unsupervised (non-reinforced) manner. Here we measured plasticity in the auditory midbrain, a hub of incoming and feedback auditory input, and found that it reflected learning of contextual information in a manner that depended on the predictability of the sound-context association and not on reinforcement. We found broad frequency-specific shifts in tuning in auditory midbrain neurons. These shifts were paralleled by an increase in response gain and correlated with an increase in neuronal frequency discrimination. The auditory midbrain, thus, codes for predictable and unsupervised sound-context associations. This points to a subcortical engagement in the detection of contextual sounds, a detection that facilitates the processing of relevant information described to occur in cortical auditory structures.