I was just reading a sci-fi novel yesterday and it included the concept of "subvocalization" as a way to control a mobile UI.  As I understand it, subvocalization is basically just mouthing the words without using your voice.  I have to say, I think if such a thing would work, it would be way less embarrassing than eg. talking to Google Glass in public.

Is there any research out there that says this sort of thing is plausible?