• Doctoral thesis (actual topic):
  "Minimally sparse but maximally versatile behavioural controllers"
  Main topics:
  
  
  • Motor continuity
  • Versatility of behavioural controllers
  • Hexapod locomotion
  
  For more details, see "Project".
  
  
• Doctoral thesis (former topic):
  "Bioinspired principles of complexity reduction for visual navigation"
  

  Topic of the doctoral project at the department for Cognitive Neurosciences was the question, which aspects of a visual stimulus shown at a decision point of a virtual maze can be used by rats to decide for a special direction. So the project should help to clear the question, in which way the complexity of visual information could be reduced to filter out information which can be used for navigational tasks. In cooperation with other projects it was intended, to use the found mechanisms of complexity reduction for artificial systems. A main target for the first year of the project was, to train rats in a virtual arena to enable them to conclude the direction to a reward out of a visual stimulus presented at the decision point. It was possible to show, that a rat in deed can be conditioned in this way.

• Diploma thesis: "Analysis of MMC-networks with non-linear characteristics"
  

  Target of the thesis was the development of MMC-networks for solving the homing-problem. That means the networks should be able to lead an agent to a target point. The target point itself was not visible to the agent, but only some orientation points. In this thesis, the two-dimensional case with three orientation points was analysed.

Animals show remarkably smooth and fast transitions between behavioural states without any sign of switching or even pausing: they show motor continuity. This phenomenon includes fast context-dependent adaptation of motion sequences, motor learning and/or adaptation to ontogenetic changes of the body mechanics (e.g., growth). Apart from such proximate mechanisms of adaptive behaviour, the underlying neural controllers have also adapted to phylo-genetic changes (evolution) in biomechanical constraints, but also in ecological requirements. The project addresses this wide range and immense versatility of behavioural controllers by investigating an unusually large and diverse data set by means of a hexapod locomotion database. The database will include considerably large samples of hexapod locomotory behaviour from several behavioural contexts (e.g., walking, climbing, searching, turning), different body sizes (larval stages and adults) and a range of different body mechanics (two to three species with different size, body mass, leg length, leg spacing, etc.) and load conditions (addition of external loads). By careful analysis of the associated changes in inter-limb and inter-joint coordination, the project will strive to identify a minimal set of mechanisms that can account for a large fraction of the observed behavioural variability across contexts, developmental stages and biomechanical constraints.