Racemic fluoxetine consists of R- and S-fluoxetine, which are metabolized to R- and S-norfluoxetine, respectively. This study was designed to compare brain levels achieved with R-fluoxetine to those achieved with racemic fluoxetine in healthy subjects using fluorine-19 (19-F) magnetic resonance spectroscopy (MRS). In all, 13 healthy volunteers received study drug for 5 weeks using a dosing schedule designed to achieve steady state for 20 mg/day racemic fluoxetine, 80 mg/day R-fluoxetine, or 120 mg/day R-fluoxetine. The resulting brain drug levels were measured using 19-F MRS. At 5 weeks, the racemate, 80 and 120 mg/day R-fluoxetine groups had mean brain levels of 25.5, 34.9, and 41.4 microM, respectively. In the serum, R-norfluoxetine, which is thought to be an inactive metabolite, accounted for 17, 71, and 63% of the fluoxetine/norfluoxetine concentration, respectively. When the relative proportion of active to total species in serum are taken into account, the data suggest that doses of R-fluoxetine greater than 120 mg/day would be needed to achieve brain levels of active drug comparable to 20 mg/day of racemate. The 120 mg/day R-fluoxetine group experienced a mean increase in QTc interval of 44 ms, with one individual having an increase of 89 ms, which suggests that higher doses may not be tolerable. While these data support the use of MRS to aid in defining the therapeutic dose range for drug development, they also highlight the need for additional studies with concurrent animal models to establish the validity of using serum drug/metabolite ratios to interpret MRS determined brain drug levels.