A novel cross-modal perceptual effect: Sound-induced increase of the non-contact illusion

Recent research demonstrates that if two identical squares moving uniformly towards each other suddenly disappear when their leading edges contact, participants will have an illusory perception of the two squares as non-contact before their disappearance. Inspired by the "sound-induced visual segmentation" hypothesis that when two approaching 2D objects reach their closest state, a coincident sound will lead to an overestimation of their distance, the current study explored whether a task-irrelevant sound delivered when the two squares make contact can enhance this non-contact illusion and examined its underlying mechanisms. Across Experiments 1 and 2, the probability of the non-contact response was consistently higher when the sound was present than when it was absent. This sound-induced increase is inherently free from the collision-based response bias-if the sound had suggested a collision between two objects, it would have led to more contact, rather than non-contact, responses. Moreover, the non-contact response was never slower than the contact response in the sound-present condition, which contradicts the attentional distraction explanation because it would predict slower non-contact responses in that condition. Meanwhile, the high time-resolution event-related potential (ERP) data in Experiment 2 showed that the early cross-modal PD170 component (125-175 ms after sound onset) was significantly larger for the sound-present non-contact response than for the sound-present contact response, which indicates that low-level audiovisual cross-modal interactions underlie the sound-induced enhancement of the non-contact illusion, substantiating its perceptual nature. Together, the current findings establish this sound-induced enhancement as a promising new paradigm for studying multisensory processing. OPEN PRACTICES STATEMENT: The trial-level data, subject-level data, analysis scripts and experimental files for both experiments are available at https://osf.io/zt92d/overview . This study was not preregistered.