The brain is able to create coherent percepts from multisensory input. This phenomenon, known as multisensory integration (MSI), is a ubiquitous feature of everyday life and has been found to be essential for a reliable interaction with the environment. Recent functional neuroimaging studies suggest that several different networks are engaged in various forms of MSI depending on the nature of information being integrated. However, little is known about the neural basis of a fundamental form of MSI in natural conditions; integration of common auditory and visual objects which are conceptually related, such as when we look at a cat and hear a meowing sound. Here we used event-related fMRI to compare the brain response to conceptually related and unrelated pairs of audio-visual stimuli denoting common objects. Our protocol was designed to preclude contamination of the results by cognitive processes additional to those needed for MSI. The results indicate that higher-order temporal and occipital areas respond to coincident sounds and pictures regardless of their semantic relationship; whereas, the right claustrum/insula region is differentially activated in association with multisensory integration of conceptually related common objects. This observation has important implications for understanding how multimodal information about common objects is represented in the brain.