The brain, rather than integrate sensory inputs and concentrate them into concepts as currently believed, appears to increase the complexity from the perceived object to the idea of it. Topological models predict indeed an increase in dimensions and symmetries from the environment to the higher activities of the brain. Models predict that informational entropy in the primary sensory areas must be lower than in the higher associative ones. In order to demonstrate the novel hypothesis, we introduce a method for the measurement of information in fMRI neuroimages, i.e., nucleus clustering's Renyi entropy derived from strong proximities in feature-based Voronoi tessellations, e.g., maximal nucleus clustering. The technique facilitates the objective detection of entropy/information in zones of fMRI images generally not taken into account. We found that the Renyi entropy is higher in associative cortices than in the visual primary ones. It suggests that the brain lies in higher dimensions than the environment and that it does not concentrate, but rather dilutes the message coming from external inputs.