This could be done on a "generic" (the Talairach-Tournoux) atlas brain by using the T-T atlas database set of voxels for each brain region. Once the numbers are loaded into 3D volumes, then display via AFNI (say) is relatively straightforward.
If you want to use a specific subject's brain, then you would have to create a 3D volume (from the 2D slices) and manually draw the sub-volumes that define each anatomical region (this can be done in AFNI, usually by a patient postdoc).
The 2 paragraphs above are just quick outlines of the needed steps. More discussion of your data and goals would be needed to give more details. Also, some indication of your experience with FMRI data processing and visualization ("your" meaning "everyone you can con into helping you", not just you personally).
I'm curious how you obtained the tissue samples needed for the gene expression analysis. From living subjects? This is quite interesting.
bob cox