Wissenschaftliche Nachwuchsgruppe "Drosophila-Laufverhalten: Neurogenetische Aufklärung und Übertragung auf die Robotik" (Strauß R.)
Am Hubland, 97074 Würzburg
Research foci (and basic equipment-based research projects):
It is the ultimate goal of this group to understand the higher control of walking by the insect brain and to use the gained insight for the construction of robots with insect-like high manoeuvrability. Within the five-year granting period the BioFuture research group is focusing on the functions of the central complex, a conspicuous neuropil present in all insects which interconnects the protocerebral brain hemispheres. Our previous and present research has proven the importance of this brain region for valuating sensory information and bringing about coherent and autonomous behavior. Our research aims to fill the broad gap which opens between our quite detailed knowledge about the peripheral processing of sensory information on the one hand side and the well studied processes of leg coordination in the thoracic centers of insects on the other side. The functions of the latter centers are remotely comparable to those of the spinal cord in walking vertebrates.
As the basis of this project we have isolated 230 mutant lines with congenital walking defects. Also, we have established quantitative behavioral paradigms. The combination allows us to map and analyse modules of walking control at all levels of the nervous system.
(1) Stepping control is a modular system. Specific mutants show, that different parameters of walking like step length, stepping frequency, swing speed of legs, or swing phase duration are all controlled by separable subsystems. Step length is optimized by the protocerebral bridge within the central complex of the brain.
(2) Leg placement on rough terrain is predominantly under tactile control. A newly developed stepping-stone paradigm allowed us to identify error tolerant leg stroke trajectories with repeat functions, which optimize leg placement without resorting to visual information. Visual influences on leg placement are weak. Specific eye regions minimize some stepping error types of particular leg pairs. But vision influences step length via visual far-field orientation. Orientation toward a visual target increases the steps and thereby the walking speed by at least 20%. Leg placement information is conveyed from anterior to posterior legs which consideraby improves their success rate for placements on first attempt. In contrast to present walking robots fly legs can cling to the side surfaces of stepping stones by that saving subsequent correcting strokes.
(3) Visual orientation control can be broken up into modules. For the landmark orientation behavior of flies we were able to identify specific brain regions for aspects of choice behavior, for the persistence of a chosen course, and for the adaptive retreat from landmarks which turned out to be inaccessible. (3a) Landmark choice: Normal flies prefer nearer objects over further away objects. They avoid rechoices after they had already explored a landmark. The protocerebral bridge is part of a motivational network, which keeps up interest in landmarks in general. (3b) Persistence of a chosen course: Normal flies keep their bearings even if the visual target becomes invisible during their approach. Respective analyses in many mutant lines with different structural central complex defects showed, that two of its four neuropilar regions are involved in this ability of the flies. With the help of toxigenetic methods we finally identified the ellipsoid body of the central complex as a key player bringing about after-fixation. It might be the seat of a simple representation of the visual surroundings. (3c) Adaptive retreat: Specifically the ablation of the mushroom bodies destroys the ability to immediately leave a chosen target after it has turned out to be inaccessible.
(4) The central complex balances existing asymmetries between body or brain sides by that enabling flies to walk in a straight line. This function has been unraveled by the analysis of three lines with structural central complex defects which can only walk in circles. All three genes are presently studied also with respect to their molecular properties.
(5) Turning behavior has been described kinematically. Twenty mutant lines with congenital defects in turning behavior were isolated for further analyses.
(6) Applicability of biological results is put to the test in three different robotic projects. (6a) The visual orientation behavior of flies is tested on a wheeled robot equipped with panoramic camera vision. (6b) In an collaborative effort with the groups of Prof. Cruse (Uni. Bielefeld) and Prof. Frik (Uni. Duisburg) a six-legged walking robot has been developed. The modified copy of a Tarry II robot possesses a basal leg coordination according to the model of stick insects. Principles of the higher control of walking found in the fly will now be implemented.
(6c) Principles of leg placement and specific designs of the end parts of robot legs are being tested on a single leg test stand.
Rewards and prizes:
Independent research group for five years awarded in the contest "BioFuture" of the German Federal Ministry of Research in 1999 (Strauss).
"Förderpreis der Würzburger Universitätskorporationen" in 2001 for excellence in research (Strauss).
"Best-Poster Award" of the 9th European Symposium on Drosophila Neurobiology, Dijon 2002 (Mronz, Strauss).
We are equipped for behavioral physiology, molecular biology, and for brain anatomy in Drosophila and for building and testing model robots.