Systems and Computational Neurobiology
Research overview
Whenever a seeing organism moves through the environment, it creates, through its own movement, a shift of the images of the environment on its retina. The concomitant distribution of motion vectors on the retina is called ‘optic flow’. The optic flow is a rich source of information about the animals own movement and, hence, is used extensively for visual course control. This is particularly pronounced in fast flying animals such as flies which are specialized in motion vision.
We are interested in how motion information from the changing retinal images is computed in the fly visual system and how this information is decoded for flight control. In general, this processing is done in two steps: In the first step, local motion vectors are calculated from local changes in retinal brightness. This calculation is done according to the Reichardt model of local motion detection, most likely in the fly's medulla. From the resulting vector fields (the ‘neural optic flow’), important course control parameters are extracted in a second step. This is realized in the fly visual system at the level of the lobula plate. Here, tangential cells integrate, by their large dendrites, the output signals of retinotopically arranged local motion-sensitive neurons and, in addition, interact amongst each other. Postsynaptic to these tangential cells, descending neurons become further selective for specific optic flow fields and transmit this information to the neck motor neurons or, via the cervical connective, to motor centers for legs and wings in the thorax.
Our work combines experimental and theoretical analysis ranging from visual responses, membrane properties and pharmacology of individual neurons up to network responses to natural image sequences created by the fly’s own flight maneuvers. Our experimental animals are the blow fly Calliphora vicina (BigFly) and the fruitfly Drosophila melanogaster (LittleFly). While the first species allows for intracellular and optical recording from individual neurons, the latter provides in addition a wealth of genetic techniques including tissue specific expression of genetically encoded indicators and blockers of nervous activity. In collaboration with Winfried Denk (Structural Neurobiology), we also try to fully reconstruct important parts of the optic lobes of both species at the ultrastructural level using his recently developed Serial Block Face Scanning Electron Microscope (BlueFly). Biophysically realistic compartmental models of individual neurons obtained from 2P-image stacks allow us to reconstitute the network of motion processing neurons in computer simulations (ModelFly). As a joint project with Martin Buss and Kolja Kuehnlenz (TUM, Munich, sponsored by the BMBF within the excellence cluster CoTeSys), our knowledge about the fly motion vision system goes into the development of miniature airborne vehicles (RoboFly). The research in our department is structured into the above mentioned five project areas and one independent Research Group (Oliver Griesbeck). The department is closely collaborating with the Emeritus Group of Bert Sakmann (Cortical column in silico).
