Using a randomized, placebo-controlled, crossover study, this study will evaluate fMRI as a
pharmacodynamic biomarker of opioid antagonism in adolescents with eating disorders. The
hypothesis is that fMRI will be able to detect acute reward pathway modulation by naltrexone
(an opioid antagonist) in pre-defined regions of interest (anterior cingulate cortex, nucleus
accumbens, dorsolateral prefrontal cortex).