We describe the tracer kinetic analysis of [C-11]-labeled alpha-methyl-tryptophan (AMT), an analogue of tryptophan, which has been developed as a tracer for serotonin synthesis using positron emission tomography (PET) in human brain. Dynamic PET data were acquired from young healthy volunteers (n = 10) as a series of 22 scans covering a total of 60 minutes and analyzed by means of a three-compartment, four-parameter model. In addition, functional images of the K-complex were created using the Patlak-plot approach. The application of a three-compartment model resulted in low identifiability of individual k-values, especially that of k3. Model identifiability analysis using a singular value decomposition of the final sensitivity matrix showed parameter identifiability to increase by 50% when the Patlak-plot approach was used. K-complex values derived by the Patlak-plot approach overestimated the compartmental values by 10 to 20%, because of the violation of the dynamic equilibrium assumption. However, this bias was fairly constant in all structures of the brain. The rank order of K-complex values from different brain regions corresponded well to the regional concentrations of serotonin in human brain (P < 0.0001). These results indicate that the Patlak-plot method can be readily applied to [C-11]AMT data in order to create functional images of the K-complex, reflecting serotonin synthesis rate, within an acceptable error margin.