Neural response was measured using fMRI in six healthy volunteers, performing a mental imagery task, using verbal cues exclusively. They listened to a list of animal names from which to generate a mental image, and listened passively to a list of abstract words. They were tested twice, using the same protocol. SPM99-processed results showed for both sessions activation in the calcarine sulcus and local activation foci, mainly in the occipito-parietal region. Other studies involving figurative mental imagery using verbal cues, have shown activation in the occipito-temporal area, but none in the calcarine sulcus or in the dorsal route. We account for the discrepancies relative to previous mental imagery studies using verbal cues, in terms of differences in the experimental conditions. In our opinion, restricting the stimuli to a single semantic category (animals) and increasing the time dedicated to the production of MI, may have enhanced the components of the pictures. This mental imagery generation protocol shows the importance of the design of experimental tasks on anatomo-functional responses.