Inhibition of Return and Refixation in Visual Search

Abstract

Search is efficient only if observers know when to leave a location and when to inspect it again. Inhibition of return (IOR) is usually inferred from slower responses at a previously cued or inspected location and is often taken as a mechanism that favors novel locations. The inference is incomplete. A cost measured with an experimental probe need not predict the next voluntary fixation, while returns are frequent in natural viewing. This review places classic cueing and visual-search studies alongside work on refixation, virtual and three-dimensional search, and real-world eye movements. The evidence separates three questions that are often conflated: how an old location is processed, whether it is selected again, and whether returning benefits the task. Laboratory results reliably demonstrate conditional costs, but their expression varies with response requirements and scene continuity. Studies of active viewing show that incomplete encoding, comparison demands, environmental change, and movement economy can make a return worthwhile. IOR is thus treated here as one bias in fixation priority, not a ban on looking back. The proposed return-control account predicts that revisiting will be favored when the expected benefit of another look exceeds the recency-related disadvantage and the cost of moving elsewhere.