First off, this is brilliant use of eye tracking. I could see eye tracking being so useful for learning, such as tracking attention, knowing when to need a break, etc. Eye tracking is super useful in user testing - for example I was watching some user testing at a major corporation and we used eye tracking. We instructed the user to minimize the app and then open it again, and we noticed that when it was in the systray, people really got confused. It was really easy to see because we could see where their eyes were scanning. Using eye tracking for people who can't speak (for any reason, maybe physical or mental injury or disability, and here age / language barrier) could be huge.
The attention differences are certainly interesting, although I wonder how much of it might be training to use touchscreens, or the idea of a screen that it might change at any time, without related motion. This is completely unnatural, as you would normally think something will change when something walks over to it, or moves slightly. Fixed objects tend to fade more into the background of our mental landscape, which is the basis of a lot of sleight of hand. Like removing something when you can't see it. It'd be interesting to compare against something that doesn't change when you interact with it, like television screens. (is it the touch part, or the screen part?)
I also wonder how touchscreens are sometimes differently interpreted by animals, kind of like mirrors are. For example, some cats seem to think that the little toy apps are under them, or inside of them, or are just generally confused. It's really quite fascinating.