fMRI studies aren't exactly junk science, but they do tend to 'explain' too much. Economists are quick to grab neuroeconomics papers and say 'aha - so that part of brain lights up explaining our behavior!'. it is not nearly so simple. (though the studies on oxytocin and cooperation/trust are pretty compelling). I just don't think we have the resources to map out all the complex variables involved in decisionmaking, brain science etc. But habiutation via unexpected reward is a key driver. And if the baseline level increases over time, we need either more, or more novel stimuli to get the same buzz..

Regarding dopamine, too many of our activities in this culture promote/end up with a release of dopamine as an 'end'. I also think our chronic cultural depressed serotonin levels are linked to dopamine (and other) cravings.

Low serotonin linked to carbohydrate craving, obesity and depression (Wurtman 1986, 1995)

Diets high in sugar will cause release of dopamine in the pleasure center of the brain (Hoebel 2005)

Sugar is a ‘gateway’ substance that increases likelihood of addiction to other substances, e.g. amphetamines (Hoebel 2003)

Areas in brain activated in food craving are also activated in drug craving (Raglund et al 2004)

I did not intend to imply that fMRI (functional MRI) is junk science. It is actually an incredible scientific technique. However some of the papers presenting findings from the study of fMRI images may be from researchers a little too anxious to join the ranks of the published. And there may be "peer reviewed" journals that are a little too anxious to publish the latest hot topic. Does anyone have a clue as to the number of medical journals in the US or the World who feel the need to fill their pages with something interesting?
--Conventional MRI gives exquisite detail. Functional MRI gives poor resolution as it is only showing gross blood flow. fMRI papers may be extremely useful but should be evaluated cautiously.