Since FMRI can only detect differences, unless you have some rest (or other null task) periods in your experiment, you can only detect activation differences between your tasks (e.g., B-A, C-A, and D-A), rather than activation magnitudes.
However, assuming you have made separate activation maps for each task, your problem is now how to combine them into a single map that shows something useful. I suppose one useful thing would be an indication of which task(s) activated which voxels. To do this fully, you need 15 colors to indicate the various possible combinations (A only, A and B, A and C, ... , A and B and C and D). The main problem with this is that it is hard to come up with 15 distinct colors. If you do have 15 colors, then you could make a fim dataset with each voxel being binary-encoded for the presence or absence of A,B,C,D response, using 3dcalc. Then each voxel would have a value from 0 (no activation) up to 15 (activation from all tasks). Then you could display this in AFNI with the number of overlay colors set to 15.
bob cox