Psychiatric imaging, in particular functional imaging techniques such as functional magnetic resonance imaging (fMRI) are potentially powerful tools to explore the neurophysiological basis of the early stages of psychosis. Despite this impressive growth, neuroimaging has yet to become an established as diagnostic instrument this area, partly as a result of significant heterogeneity across the findings from research studies. The present review aims to: (i) assess the determinants of inconsistencies in the results from neuroimaging studies of the early stages of psychosis; and (ii) suggest approaches for future imaging research in this field that may reduce methodological differences between studies.