The spread of attentional processing represents a possible mechanism through which discrete features, represented within different cortical regions, are bound together. A central tenet of the influential biased-competition model of visual attention suggests that focal spatial selection results in the preferential processing of an entire object (Desimone & Duncan, 2005; Duncan, 2006). Indeed, experimental evidence has shown that attention spreads both within the visual boundaries of an object and to an object’s task-irrelevant visual features, seemingly confirming that there is an inherent bias to process objects as wholes (O’Craven et al., 1999, Schoenfeld et al., 2003, Martinez et al., 2006). Recent evidence further suggests that attention spreads across sensory boundaries to encompass the multisensory features of an object. A task-irrelevant sound, for example, receives enhanced processing when it is simultaneously presented with an attended visual stimulus, relative to when the same task-irrelevant sound is presented alone (Busse et al., 2005; Talsma et al., 2007). In comparison, Talsma and colleagues (2007) showed no auditory-to-visual spread of attention during attend-auditory task, possibly reflecting the dominance of vision in object processing. We used event-related potentials to investigate whether a well-established manipulation of visual object processing modulates the object-based, cross-modal spread of attention. Participants viewed illusory contour (IC) and non-illusory contour stimuli (N-IC) presented at central fixation, and performed a target-detection task that was independent of the configuration of the inducers (i.e., the pacman-like elements that constitute the IC and N-IC stimuli). The visual stimuli were presented either alone or paired with a task-irrelevant tone. Our results demonstrate that positioning the inducers to form the apices of a gestalt object (i.e., an IC stimulus) not only leads to enhanced visual object processing but also to enhanced processing of the object’s multisensory features (i.e., the simultaneously presented task-irrelevant tone). Segmentation of a visual scene, like what occurs during the formation of illusory contours, represents an essential first step in the formation of distinct objects and subsequent feature integration (Driver et al., 2001). We therefore propose a model for feature integration where (1) the visual boundaries of an object are established through processing in the occipitotemporal cortex, and (2) attention then spreads through interactions with the occipitoparietal cortex to cortical areas that process features that fall within the object’s visual boundaries, including task-irrelevant multisensory features. Our data strongly establish a link between visual object processing and the cross-modal spread of attention, and offer further evidence that the spread of attentional processing plays an important role in feature binding.