Abstract We combined complementary non-invasive brain imaging techniques with behavioural measures to investigate the anatomy and time course of brain activity associated with semantic priming in a lexical-decision task. Participants viewed pairs of stimuli, and decided whether the second item was a real word or not. There were two variables, the semantic relationship between the prime and the target (related or unrelated) and the interval between the onset of prime and target (200 or 1000 ms), to vary the degree of semantic expectancy that was possible during task performance. Behavioural results replicated the well-established finding that identification of the target is facilitated by a preceding semantically related prime. Event-related functional magnetic resonance imaging (efMRI) identified two brain areas involved in the semantic-priming effect. Activity in the anterior medial temporal cortex was diminished when target words were primed by semantically related words, suggesting involvement of this brain region during active semantic association or integration. In contrast, activity in the left supramarginal gyrus in the temporal-parietal junction was enhanced for target words primed by semantically related words. Brain areas influenced by the interval between prime and target words, and by the interaction between word interval and semantic priming were also identified. A parallel experiment using event-related potentials (ERPs) unveiled a striking difference in the time course of semantic priming as a function of expectancy. In line with previous reports, the primary effect of semantic priming on ERPs was the attenuation of the N400 component, in both short- and long-interval conditions. However, the priming effect started significantly earlier in the long-interval condition. Activity in the anterior medial temporal cortex has previously been shown to contribute to the N400 component, a finding that links the priming results obtained with efMRI and ERP methods.