Abstract
Observers often incorrectly localize the starting position of a translating target in the direction of the target's motion (Froehlich, 1923). Recent evidence suggests the source of the error is an absence of spatial attention (Muesseler & Aschersleben, 1998) and/or a degradation in visibility due to lateral inhibition (e.g., Geer & Schmidt, 2006; Kirschfeld & Kammer, 1999). A positive correlation between error and target velocity supports the former; but velocity is typically confounded with the spatial displacement of contours between animation frames. We investigated whether a displacement exists that produces a local maximum localization error within an inverse-U shaped function. Such a relationship would obviate a critical role for attention. We presented normal observers with either a large square or a narrow bar target animated with displacements of 7.5 to 120 min/arc at a variety of eccentricities and measured localization performance using a mouse-driven cursor. The results show that an intermediate displacement, which roughly scales with RF size, elicits the greatest error. The results converge with detection data, appear consistent with a rapid long-range center-surround inhibitory mechanism (see, for example, Bair et al., 2003), and have implications for theories of target localization and metacontrast.