TY - JOUR TI - Multisensory expectations shape olfactory input to the brain DO - https://doi.org/doi:10.7282/T37P91V6 PY - 2017 AB - Mammals spontaneously learn contingencies among sensory stimuli including across sensory modalities. Stimulus recognition is faster and more accurate when cross-modal cues are congruent with previous experience (e.g. Gottfried & Dolan, 2003). This suggests that information from multiple sensory modalities could converge in early sensory processing regions in the brain. In the olfactory system, the olfactory bulb glomerulus receives heavy anatomical top-down projections from brain regions that might contain such information. Using wide-field in vivo imaging of awake head-fixed mice expressing the calcium indicator GCaMP in GABAergic periglomerular (PG) interneurons in the olfactory bulb, neural activity can be evoked not only by odors but also by lights, tones, and whisker deflections. Anesthesia eliminates responses to non- olfactory stimuli and tracheotomy demonstrates that these signals are not driven by respiratory changes. Non-olfactory stimuli were most effective when presented at long inter-stimulus intervals (e.g. > 60 sec), but evoked observable bulbar activity at all intervals tested. To test whether non-olfactory stimuli convey odor-predictive information to the olfactory system, mice were presented with 13 presentations of a light-tone-odor sequence to establish an expectation about the odor. This expectation was subsequently violated by omitting the expected tone while presenting the light and odor as usual. There was an increase in GABAergic interneuron activity during the odor presentation when its preceding tone cue was omitted. Because these interneurons are responsible for presynaptic inhibition of olfactory sensory neurons (OSNs), the primary sensory neurons of the olfactory system, the experiment was repeated in separate animals expressing the exocytosis indicator synaptopHluorin (spH) in OSNs. This revealed a complementary suppression of odorant-evoked neurotransmitter release from OSNs during the odorant presentation in which the tone was unexpectedly omitted. This effect was not observed if mice were anesthetized or if the absence of the tone was unsurprising. Imaging of sniff-by-sniff calcium dynamics in OSN presynaptic terminals revealed that suppression of activity is present on the first inhalation of odorant during the surprising trial. This reduction in presynaptic calcium suggested a GABAB receptor- mediated presynaptic inhibition. Blocking GABAB receptors with CGP35348 abolished the effect of tone omission on odorant-evoked neurotransmitter release. To test whether PG interneurons are sensitive to olfactory expectations in the absence of bottom-up odor information, mice were again presented with the light-tone-odor sequence, but the odor was subsequently omitted. On this odor-omitted trial, the magnitude of GCaMP signals during the time of the expected odor presentation was not significantly different from the previous odor-present trial, but the spatial pattern of activity was diffuse and not obviously glomerular. In a modified paradigm, two different tones were presented with two different odors, with each pairing delivered on 13 interleaved trials. When the odor was then switched such that the tone was unexpectedly followed by the other odorant, the amplitudes of the odor-evoked responses became more similar to each other. These experiments show that non-olfactory stimuli can drive activity in the olfactory bulb and that expectations and violations thereof can shape sensory processing as early as the primary input into the brain. KW - Psychology KW - Senses and sensation KW - Olfactory sensors KW - Smell--Psychological aspects LA - eng ER -