The aim of this functional magnetic resonance imaging (fMRI) study was to investigate the neural correlates associated with the generation of general (i.e., prototypical) and specific (i.e., exemplar) visual mental images from concrete nouns. The fMRI paradigm included a non-imagery baseline, and two activation conditions requiring the generation of either general or specific images. Image generation times and brain activation were recorded. Analysis of the behavioral results showed that generating general images took less than the specific ones. The comparison of each activation condition with the baseline showed significant increase in brain activation in left frontal areas in both kinds of images, with the additional involvement of the posterior cingulate cortex during the generation of specific images. When the two activation conditions were contrasted with each other and masked for their respective comparison with baseline, significant activation was found in right frontal areas for general mental images, whereas a significant increase in activation in the left superior frontal region and the right thalamus was detected during the generation of specific mental images. These findings suggest that general and specific mental images are generated with the support of two different neural pathways. The generation of general images seems to involve brain areas associated with the formation of global gestalt-like images (areas in the right hemisphere), while the generation of specific mental images appears to require additional support from areas involved in the retrieval of visual details (i.e., the right thalamus).