We review reports published over the past 5 years on positron emission tomography (PET) of neurotransmission in depressive disorders. The molecular tools of PET neuroimaging are compounds labeled with a positron-emitting nuclide. PET radioligands have been used in recent years to study several aspects of monoaminergic and cholinergic neurotransmission in the brain of depressed subjects and healthy controls. The value of kinetic parameters of certain PET radioligands has often been reported to be lower in depressed subjects than in healthy ones, but there is usually no reliable relationship between the binding potential of the neuroreceptor or transporter and the clinical condition of depressed subject. In addition, many recent PET studies have noted either higher binding potentials in depressed subjects or no difference between binding potentials of depressed and healthy subjects. In our view, recent research has neither proved nor refuted the idea that neuromolecular processes that can be assessed by the radioligands currently available for PET studies of humans are causally related to depressive disorders. The future success of PET research for understanding molecular mechanisms in depressive disorders may therefore require the invention and development of further molecular tools for studying a wider range of neuronal events in the living human brain.