Flies, among many other species mainly rely on visual information to safely navigate through their environment. Behaviors like finding appropriate food sources or potential mates, avoiding obstacles and predators, or simply stabilizing their flight or walking trajectory are executed based on visual cues in their surroundings. Optic flow on the flies’ eyes for instance helps them to respond to disturbances caused by external distractors like a gust of wind. In order to do so, the animal has to extract relevant information about the produced optic flow. Calculating features, such as the speed and direction of motion is crucial for their successful compensation. A possible model for the computation of direction selectivity in Drosophila melanogaster is based on asymmetric temporal filtering of signals originating from two neighboring photoreceptors in the retina. Wide field neurons in the lobula plate of the fly’s optic lobes have long been known to respond in a direction-selective way. The computations in the upstream neuropils that lead to direction selectivity, however, have not yet been described. During my PhD-thesis, I characterized neurons presynaptic to directionselective lobula plate tangential cells by exploiting the genetic toolbox of the fruit fly in combination with physiological techniques. The use of genetically encoded calcium indicators and two-photon microscopy allowed me to directly investigate response properties of small columnar elements prior to the lobula plate. Expressing neuronal silencers in distinct subtypes of neurons and simultaneously recording from lobula plate tangential cells, I could further probe the contribution of these elements to the computation of direction-selective signals. Combining these techniques with several other methods available in our laboratory, including a behavioral readout of tethered walking flies, optogenetic activation, pharmacology, and modeling, resulted in a total of six publications that comprise this cumulative thesis.

Measuring calcium signals in T4 and T5 cells in the first study, established that both populations of neurons exhibit direction-selective response properties. Furthermore, T4 cells only respond to moving bright edges, whereas T5 cells encode exclusively dark motion. Silencing the synaptic output of T4 and T5 separately, we were able to determine that both lobula plate tangential cell responses as well as the turning behavior of walking flies were impaired only to bright or dark edges, respectively. We thus proposed that the detection of the direction of visual motion must happen either presynaptic to, or on the dendrites of T4 and T5 neurons, and that this computation takes place independently for brightness increments and decrements.