Previous 
Up 
Next
A realistic model of the neural network generating locomotor
rhythmicity in the lamprey
was used as the basis for an extended model
including also the mechanical movements of the body
(Ekeberg 1993). The resulting combined neural and
mechanical model displays basic swimming capabilities, such as
swimming at different speeds, turning and even backward swimming.
One advantage with such an integrated neuro-mechanical simulation setup is that it makes it possible to analyze the role of sensory feedback. In this particular system, however, it turns out that incorporating the sensory feedback did not make much of a difference to the swimming pattern. This is not surprising since fictive swimming (i.e. neural swimming generation without the actual motion and, thus, without sensory feedback) is known to produce an output pattern similar to that of real swimming. The question then remains: what is the role of the sensory feedback when it only marginally affects the normal swimming pattern?
In order to investigate this, we have designed an experiment in which the simulated swimming pattern is perturbed. The model lamprey had to swim through a region of contrary water flow. This setup roughly mimics the natural situation where the lamprey, while swimming in shallow flowing water, has to pass a small region of less depth. When the water passes such a region it will locally have a higher velocity since the same amount of water per time-unit will have to pass through a smaller cross-section area. When the lamprey enters or exits this region it will have to adapt to the increased or decreased drag force respectively.
Without sensory feedback, the simulated lamprey had less chance of making it through the barrier (Ekeberg et al. 1995). However, it is still too early to draw any definite conclusions from these experiments regarding the role if the sensory feedback in the lamprey.
Next