Running to fMRI in an attempt to prove a mechanism before anyone has even proven an effect is bad enough science anyways. But it looks like even that is useless.

http://www.pnas.org/content/113/28/7900.full