Speaker
Description
Different types of classical and operant conditioning experiments have shown that honeybees, as central place foragers, have a vast repertoire of stimuli that they can associate with the rewarding nectar and pollen of a flower. These associations can be unimodal, like the odour or visual component of a blossom, but they can also combine both modalities, and bees learn that only the olfactory-visual compound is rewarded and not its single elements.
By electrophysiologically recording mushroom body (MB) output neurons (MBONs), we focus on the sensory integration of both modalities. At this processing level, some subpopulations of MBONs respond to both modalities, whereas other subpopulations respond to either light or odour (Strube-Bloss and Rössler, Roy.Soc.OpSci 2018). Moreover, some MBONs send retrograde centrifugal feedback innervating the antennal lobe (AL). Thus, they may modulate olfactory processing within the AL. Before we address how cross-modal learning experiments will modulate single MBONs, we aim to understand how the general unimodal and multimodal MBONs code at this processing level. When we presented the bees with different lights, odours and the corresponding compound stimuli, we observed that superficial unimodal MBONs, which responded to only one modality (odour or light), responded differently when stimulated with the compound, suggesting a modulatory interaction.
Morphological studies have shown that olfactory projection neurons innervate the MB's lip and basal ring, while visual projection neurons innervate the calyx's collar and basal ring region.
We implemented these plausible connectivities into a rate-based model. We assumed that different populations of Kenyon Cells (KCs) relay the information of these regions in parallel, before MBONs integrate across different KC subpopulations. Unimodal MBONs have to be connected with the KC population of the individual single modality and to the basal ring. The latter multimodal pathway then dictates the modulatory effect. By combining modelling and previously known connectivity within the MB, we could predict the observed phenomena and infer plausible connectivity of MBONs with KCs.
