Motor symphony and central pattern generation create a whole new view on behavior. They say that for many types of behavior it is not necessary to have any visual or other input, and furthermore, the organism (nervous system) can create the behavior on its own. The phenomena that are directly related to these concepts are the phantom limbs, cat and cockroach walking without higher nervous system, headless rooster running, tongue rolling, etc.

The phenomenon of the phantom limbs, described in the article written by Ronald Melzach, shows that "we do not to have a body to feel a body". People that can feel phantom limbs, their movement and pain, probably, experience the action of the proprieceptors that bring the information from the "remaining neurons in the slump" to the nervous system. This is what I think the author has meant by the "neuromatrix"--the network of proprieceptors that carry the information from the body to the central nervous system.

Also, the experiments on cats and cockroaches that were described in "The Control of Walking" show that having proprieceptors carry the information to the brain is not that important, because some of the information about walking lies in the nervous system itself, and it can be "replayed" without any input to the brain from the outside. The same fact is illustrated by the fact that if the rooster's head were cut off, for the first number of minutes the rooster will still be running around.

I think that this notion of the stored motor sympony and CPG that allows one to be able to perform some actions without any input from the body and environment shows that it this "stored motor symphony and CPG" have to be genetic. I remember the experiments we did in General Biology to show that tongue rolling is genetic. I think this phenomena of tongue rolling depends on whether the "motor symphony" for the tongue rolling is present among the person's genes.

Since the "motor symphony and the CPG" are genetic phenomena, the experiments that were made to show that motor symphony can be learned only show that motor symphony is active. For example, in one's learning how to dance, type, etc. one just adjusts his/her general stored motor symphony to the new task, as opposed to creating a new motor symphony.

Tongue rolling interesting, may indeed be a difference in genetic information and their role in CPG creation. Worth looking into more. Be careful, though. The running rooster still has lots of input, so that doesn't quite show the existence of a CPG (though it does say some important things about how they are controlled). And CPG's needn't be genetic, can be learned through experience. AND adjusted while being played using sensory input. PG