Encoding the expectation of a sensory stimulus

Most organisms possess an ability to differentiate unexpected or surprising sensory stimuli from those that are repeatedly encountered. How is this sensory computation performed? We examined this issue in the locust olfactory system. We found that odor-evoked responses in the antennal lobe (downstream to sensory neurons) systematically reduced upon repeated encounters of a temporally discontinuous stimulus. Rather than confounding information about stimulus identity and intensity, neural representations were optimized to encode equivalent stimulus-specific information with fewer spikes. Further, spontaneous activity of the antennal lobe network also changed systematically and became negatively correlated with the response elicited by the repetitive stimulus (i.e. ‘a negative image’). Notably, while response to the repetitive stimulus reduced, exposure to an unexpected/deviant cue generated undamped and even exaggerated spiking responses in several neurons. In sum, our results reveal how expectation regarding a stimulus is encoded in a neural circuit to allow response optimization and preferential filtering.