It's worth noting that just about everything you do from day to day is taken care of by your muscle memory. Walking, eating, drinking, typing, driving, and even holding conversations and playing instruments are mostly taken care of without having to focus too much thought on them. "Wait, don't you have to process what the person you're talking to is saying, and think about your response when conversing?" Well, yes and no, which is why I said "mostly". Granted that if the person you're talking to is presenting you with a decent amount of new information, you'll have to reason through it before you can say something intelligent in response. But I've noticed many times that when someone says "So, I saw Movie X this weekend" or "How was your day?", words like "Oh yeah, how was that?" and "Eh, alright" kind of fall out of my mouth before I know it.

With regard to playing an instrument, the patterns of finger movements (and probably breathing movements too, for wind players) that make up scales, riffs and songs seem to get embedded in your muscle memory somewhere between 50 and 1000 repetitions. Which is why they say "Practice makes perfect." Indeed, many times after I've committed a song to mental memory, I'll still make technical mistakes while playing it. But, if I stop playing the song for a few days and let it drift out of my short-term memory, I find myself able to pick up the instrument and play the song almost to perfection, letting my fingers do the work and my mind drift elsewhere.

A friend of mine told me once that humans resemble machines much more than we're aware. Apparently, humans remain motionless until they're presented with a situation they've been programmed for (ie, learned how to react to). When this happens, they make the appropriate movements with their body until the situation is "dealt with", and then proceed to do nothing until another stimulus comes. Aside from the fact that humans tend to experiment and learn how to deal with new situations (which could be explained by this theory as being the reaction to both boredom and new circumstances or objects), my main argument with this idea is that it discounts brain activity as non-motion. Superficially, though, the model does work well for explaining how muscle memory works. The body remains still until it gets hungry, and then it finds a food source, does the requisite killing or preparing or paying, eats, and then goes off to find a comfortable place to sit until it has to take care of something else.