Single photon emission computed tomography (SPECT) and positron emission tomography (PET) have advanced our understanding of the biological underpinnings of depression. There is, however, considerable variability in the literature. Depression is a complex disorder with marked heterogeneity in diagnosis and treatment. There is also evidence of heterogeneity in pathophysiology. In addition, the literature is marked by inconsistencies in the use of imaging techniques and data-analytical procedures. In this review we have attempted to focus on the SPECT studies that have used more refined methodologies and more homogenous clinical sub-groups of patients. We have focused on the main diagnostic sub-types of depression and on specific issues such as treatment response, correlates of neuroimaging abnormalities in depression, and so-called 'emotional circuitry' - the connectivity of regions implicated in depression. The future of molecular imaging in depression will be determined by the pace of the development of useful ligands and the exciting opportunities emerging in the field of imaging genomics. Future studies must attend to several key confounds including clinical heterogeneity, medication and the problems surrounding recruitment of drug-naive patients. It remains the case that longitudinal studies are the design of choice if questions relating to state and trait are to be addressed. Molecular imaging will be used increasingly to quantify neuroreceptor and transporter binding, and the activity of neurtransmitters, allowing the neurochemistry of this complex condition to be explored.