Mid-Level Audiovisual Crossmodal Correspondences: A Narrative Review

This paper critically reviews the literature on mid-level audiovisual crossmodal correspondences, that is, those associations that emerge between structured, often dynamic stimuli in vision and audition. Unlike basic correspondences (involving perceptually simple, or unitary, features) or complex ones (involving semantically rich combinations of stimuli), mid-level correspondences occur between temporally and/or spatially patterned stimuli that are perceptually structured but are typically not inherently meaningful (e.g., melodic contours and moving shapes). Taken together, the literature published to date suggests that such correspondences often rely on structural or analogical mappings, reflecting shared spatiotemporal organization across the senses rather than the direct similarity of low-level features or emotional content. Drawing on evidence from developmental, comparative, and experimental studies, we discuss the possible mechanisms underpinning these mappings - including perceptual scaffolding, amodal dimensions, and metaphorical mediation - and outline open questions regarding their perceptual, cognitive, and neural bases. We also evaluate key methodological approaches and provide suggestions for future research aiming to understand the hierarchy of crossmodal correspondences across levels of perceived stimulus complexity. Besides advancing theoretical models, our paper offers practical insights for domains such as multimedia design and crossmodal art.