Among many controversies in visual neuroscience is whether visual imag
ery of objects, scenes and living beings is based upon contributions o
i the early visual areas or depends on hierarchical higher visual area
s only and whether the cortical areas subserving visual imagery are id
entical to those underlying visual perception. These questions are imp
ortant for furthering our understanding of vision, since areas active
in visual imagery might tell us how the visual cortex represents objec
ts, scenes and living beings. Here, P. E. Roland and B. Gulyas present
their hypothesis, based on experimental evidence in man and primates
that the visual areas subserving visual imagery are parieto-occipital
and temporo-occipital visual association areas, and that these areas f
orm only a subset of the visual areas engaged in perception. This hypo
thesis is consistent with the view that objects, scenes and living bei
ngs are represented stored and re-evoked outside the domain of the pri
mary visual cortex and its immediate neighbours.