Bupropion is an atypical antidepressant and smoking cessation aid; its use is associated with wide intersubject variability in efficacy and safety. Knowledge of the brain pharmacokinetics of bupropion and its pharmacologically active metabolites is considered important for understanding the cause-effect relationships driving this variability. Brain concentrations from rats administered a 10 mg/kg subcutaneous dose of racemic bupropion were analyzed using a stereoselective LC/MS-MS method. A 2 mg/kg dose of (S,S)-hydroxybupropion, which has comparable pharmacologic potency to bupropion, was administered to a separate group of rats. Plasma exposure and unbound concentrations in both matrices from companion equilibrium dialysis experiments were determined to assess potential carrier-mediated transport at the blood-brain barrier. Exposures to unbound forms of bupropion enantiomers were similar in plasma; this was also true in brain. This trend held for reductive diastereomer metabolite pairs in the two matrices. Unbound (R,R)-hydroxybupropion exposure was 1.5-fold higher than (S,S)-hydroxybupropion exposure in plasma and brain following bupropion administration. Unbound concentration ratios (Kp,uu) of a given molecular form decreased over time: between 4 and 6 h, these were < 1 for the two bupropion enantiomers, and they were ~ 1 for metabolites that formed. Administration of preformed (S,S)-hydroxybupropion also demonstrated a declining Kp,uu. The temporal shift in Kp,uu among the different molecular forms provides evidence regarding the operation of carrier-mediated transport and/or within-brain metabolism of bupropion, and, thereby, fresh insight regarding the causes of intersubject variability in the safety and efficacy of bupropion therapy.