i'm not sure that i explained what is going on terribly clearly above, but the outline seems to me correct.
the cog-sci lab that issued this statement has a website here:
Laboratory for Cognitive Brain Mapping, RIKEN BSI
if you tool around on it, you can find links to a considerable series of papers. read around and it becomes more obvious what is and is not happening with these experiments. the article that seems most germaine is the short piece from nature neuroscience, which outlines what appears to be this experiment, but with an emphasis more on the mapping procedures than on the reversal of this maps (a map is reversed in a sense when its status is shifted from representation to template).
the interesting part of this seems to be the extent to which these folk have adapted newer mri technologies to this kind of research. the scans they are using are quite long, so they rigged up a lovely head-immobilization apparatus that features a bite bar...but anyway, they've been able to generate quite detailed maps of physiological correlates to some aspects of visual perception. in the longer pieces, these folk are much more explicit about exactly what they're doing---their work is framed as a partial view of some aspects of the complexity of visual processing.
the maps themselves sound quite cool though, regardless of whether the claims made about them, once reversed, are legit.
plus these folk use cool graphics to present their results.
that is a good thing.