Horizontal movement detectors of honeybees: Directionally-selective visual neurons in the lobula and brain

Abstract

1. Intracellular recordings have been made from both input (ipsilateral) and output (contralateral) ends of directionally-selective optomotor neurons in the left lobulae of worker honeybees. Optical stimuli were two, monocular, horizontally-moving vertical gratings placed on opposite sides of the bee's head. The eyes could be stimulated separately or together. 2. Horizontal regressive motion-sensitive (HR) cells are strongly excited (and depolarized) by ipsilateral regressive motion. They are moderately excited by contralateral progressive motion but are inhibited by contralateral regressive motion. Ipsilateral progressive motion has little effect (Figs. 1 and 2). The HR cells (HRLeft and HRRight cells) are bilateral, reciprocal neurons which run from one lobula to the other via the anterior optic tracts, the optic tubercles, and the intertubicular tract. They have co-extensive dendrites and terminals in the distal lamellae of the lobulae (Figs. 4–6). HR cells are the bees' optomotor neurons, known hitherto from extracellular recordings. 3. Horizontal progressive-motion sensitive (HP) cells (HPRight and HPLeft cells) are excited and depolarized by ipsilateral progressive motion but are inhibited and hyperpolarized by ipsilateral regressive motion. It is uncertain if there are bilateral influences on HP cells (Figs. 7 and 8). No stained HP cells have been recovered. 4. Other directionally-selective cells with and without spikes have also been stained and reconstructed (Figs. 9 and 10). 5. A model network is proposed between HR and HP cells in which a) HR cells mutually inhibit each other; b) the ipsilateral, input end of each HR cell inhibits the HP cell originating in the same lobula; and c) the contralateral, output terminals of each HP cell excite the HR cell originating in the opposite lobula (Fig. 11 A). It is discussed how the model could explain the observed directional selectivities of cells (Figs. 11B, C and 12). Additionally, the possible role of such a model network in optomotor behavior is considered.