Little is known about how the brain binds together signals from multiple sensory modalities to produce unified percepts of objects and events in the external world. Using event-related functional magnetic resonance imaging (fMRI) in humans, we measured transient brain responses to auditory/visual binding, as evidenced by a sound-induced change in visual motion perception. Identical auditory and visual stimuli were presented in all trials, but in some trials they were perceived to be bound together and in others they were perceived as unbound unimodal events. Cross-modal binding was associated with higher activity in multimodal areas, but lower activity in predominantly unimodal areas. This activation pattern suggests that a reciprocal and 'competitive' interaction between multimodal and unimodal areas underlies the perceptual interpretation of simultaneous signals from multiple sensory modalities.