ABSTRACT
In the course of evolution, a large variety of bodies and nervous systems have come into existence, each of them adapted to and optimized for its own natural environment, or ecological niche. As different environmental conditions necessitate or favour different ways of moving, the range of motor systems evolved is as vast as the range of environmental conditions that can be encountered on earth. Locomotion of an organism that merely performs a small number of behaviours, like moving from one place to the other to forage and flee from enemies, might be based on a very simple control system. Biologists have hypothesized that the walking behaviour of an insect could easily be controlled by a system that lacks a central control unit or brain that mediates the synchronization, provided each of the six legs is controlled by a local module (see Cruse et al., 2004). Instead, the direct coupling of the legs through the body and the environment, as well as their interconnection through some local coordination influences, is sufficient to cause stable and adaptive gaits. With such a simple control system, the animal would be able to walk and master obstacles of notable size (Bläsing, 2006; Bläsing & Cruse, 2004). It could live successfully as long as its environment did not provide too many changes that called for intensive problem solving. Building a computer simulation of such a system is a complicated, yet solvable, task, as has been demonstrated by Cruse and Schilling (Chapter 3, this volume). Similar results have been obtained by scientists who have built models of the motor systems of amphibians, such as salamanders (Ijspeert, 2001), which basically consist of different sets of neural oscillators for cyclic movements coupled with each other in such a way that different types of locomotion emerge – as in the salamander, swimming and walking. Such basic types of locomotion are essential prerequisites for more complex behaviours (see Chapter 3 by Cruse and Schilling, this volume). In the real world, even amphibians do more than move from one place to another. If we regard, for example, a frog catching
a fly, it becomes clear that cyclic movements are not sufficient for such behaviour. In this case, discrete, goal-directed movements are needed that are controlled by a system that monitors the stimulus, in this case, the fly (see Arbib & Liaw, 1995). The endeavour to model such movements has given birth to the idea of schemata, conceptualized as functional units of motor behaviour and corresponding perceptual processes, including their neural correlates (see, e. g., Arbib, Conklin, & Hill, 1987).
