Figure: This figure illustrates spike patterns from example flower- and bar- responsive units from the descending connective in hawk moths and corresponding tuning curves. The units shown do not peak in the red bar which is the range in which target selective descending neurons (TSDNs) from other insects show peak responses. Instead, they peak at intermediate size ranges which overlap with the angular extent that hawkmoth pollinator flower species would subtend on the moth’s visual field (blue bar)
Many agile hawkmoths use vision and other cues to hover feed and track flowers during foraging but do not engage in visual pursuit. The hawkmoth connective contains putative object-sensitive descending neurons (OSDNs) that are broadly directionally responsive. Some DNs show differential responses to 1-D wide field and target stimuli but have not been characterized to different 2-D target sizes and floral visual features. Target selectivity could be widely conserved even in non-pursuit species or OSDNs could be tuned for larger objects and floral features.
To test this, we collected 67 visually responsive units using a new neck connective multielectrode array preparation. Many units had temporal frequency tuning to wide field stimuli consistent with prior work. We next assessed receptive fields and found many units to be predominantly responsive on one side of the visual field but otherwise broadly responsive and typically centered close to the dorsoventral midline. We found no neurons tuned as small target detectors and only a subset (n = 41) that responded to bar stimuli of any size.
Most interesting, we used an unsupervised clustering algorithm and identified another class of units, flower responsive OSDNs (FR-OSDNs, n = 18, see Figure) that responded to both intermediate size bars and flowers with a broad tuning curve peaking between 20 and 60° in diameter. Another group was not responsive to narrow bars of any height but did respond to flower shaped stimuli (n = 13). This is consistent with our finding that M. sexta adopts a hovering position in the lab that makes the floral face subsume ~60-80° of the visual field. FR-OSDNs required contrast of the floral face on the background but were also responsive to different constituent parts of the flower-shaped stimuli. Taken together our results suggest that some DNs in the hawkmoth are tuned not for small, prey-like targets, but for intermediate sized and flower-like objects, reflecting their feeding behavior.