This shit doesn't work. (Score -1, Troll)

Trust me. Stick to Perl+C. Oops! Forgot this is related to brain science.. fMRI, AFAIK, has so far been successful in producing partially convincing results only for very basic sensory+motor activities (I know some people will claim otherwise, but that's more related to the economics of the enterprise than science). Besides, the signal-noise ratio is so low (of the order of a percent? I forget) that anything of a higher order like personal preferences will either be indiscriminable from noise or will be an artefact of your brain responding simply to the stimulus. AKA, there is not a simple way of controlling the stimulus related effects. E.g., my bottle is flashier than yours and therefore it's going to produce more brain activity. That doesn't mean that my bottle is more appealing than yours or the fluid in my bottle is tastier etc.. So yeah. Whatever. Looks like another way to justify a million dollar machine that only re-confirms what's already been observed in monkeys. That is not an entirely useless thing to do, though.