thanks Daniel. I was actually interested in implementing some of my own image processing algorithms...for example, lets say i wanted to take a T1 scan and get some kind of measure of how flat the cortex is at certain points under the skull...i would have to then create tangent planes to every mm of cortical tissue, and then plot the normal vector at that point and perhaps write out a new dataset with all the tangent vectors...do you think this is something that could be done from within afni, or would i be best off looking to access the images directly voxel by voxel?
Russell