Insects face a wide range of light inten­si­ties, which gradu­ally change through­out the daily cycle, and suddenly change between celes­tial condi­tions or habitat types. To under­stand how insects extract relevant infor­ma­tion from such dynamic visual scenes, it is neces­sary to study both sensory process­ing and behav­iour, which influ­ence each other recip­ro­cally. To disen­tan­gle this closed-loop, I am study­ing three key-stages: (i) adaptive behav­iour, (ii) natural inputs, and (iii) sensory processing.

Flying insects use visual infor­ma­tion to control their flight. In nature, the visual scenery is highly complex, which poses a challenge for insect brains to extract relevant and reliable infor­ma­tion. And yet, many animals master these challenges on a daily—and nightly—basis. Over the course of a single day, they face a wide range of light inten­si­ties from sunlight to starlight. Within a short moment, the light environ­ment can rapidly change between celes­tial condi­tions or habitat types – and at night with artifi­cial light. Insects are thus confronted with not only complex, but also dynam­i­cally chang­ing infor­ma­tion. How does the visual system adjust to these changes? What are the effects of light pollution?

To answer these questions, I study both sensory process­ing and behav­iour, which influ­ence each other recip­ro­cally. Using the noctur­nal elephant hawkmoth, I am disen­tan­gling this closed loop using three key-stages: (i) adaptive behav­iour, (ii) natural inputs, and (iii) sensory process­ing. The first stage will reveal natural flight dynam­ics of the moths in differ­ent light levels and forms the basis for imaging natural visual scenes. Here, I will quantify the dynam­ics of natural visual environ­ments from a flying moth’s perspec­tive, and then measure how early visual neurons adjust to these spatiotem­po­ral light varia­tions. Finally, to close the loop from sensing to behav­iour, I will use a compu­ta­tional model
to predict responses of downstream neurons that guide flight behaviour.