There are also a comment in that thread on Sean Carroll's site from someone who seems offended at the idea that hallucinations could be something other than contact with a deeper reality. (And, anyway, if the research is correct, in a sense they are: they're a direct perception of the action of some nifty mathematics within the structure of your own head. If Riemann was inspired by them, that makes perfect sense, because some of the math involved is exactly what Riemann was working on.)