Humans can estimate the number of visually displayed items without counting. This capacity of numerosity perception has often been attributed to a dedicated system to estimate numerosity, or alternatively to the exploitation of various stimulus features, such as density, convex hull, the size of items and occupancy area. The distribution of the presented items is usually not varied with eccentricity in the visual field. However, our visual fields are highly asymmetric, and to date, it is unclear how inhomogeneities of the visual field impact numerosity perception. Besides eccentricity, a pronounced asymmetry is the radial-tangential anisotropy. For example, in crowding, radially placed flankers interfere more strongly with target perception than tangentially placed flankers. Similarly, in redundancy masking, the number of perceived items in repeating patterns is reduced when the items are arranged radially but not when they are arranged tangentially. Here, we investigated whether numerosity perception is subject to the radial-tangential anisotropy of spatial vision to shed light on the underlying topology of numerosity perception. Observers were presented with varying numbers of discs and asked to report the perceived number. There were two conditions. Discs were predominantly arranged radially in the “radial” condition and tangentially in the “tangential” condition. Additionally, the spacing between discs was scaled with eccentricity. Physical properties, such as average eccentricity, average spacing, convex hull, and density were kept as similar as possible in the two conditions. Radial arrangements were expected to yield underestimation compared to tangential arrangements. Consistent with the hypothesis, numerosity estimates in the radial condition were lower compared to the tangential condition. Magnitudes of radial alignment (as well as predicted crowding strength) correlated with the observed numerosity estimates. Our results demonstrate a robust radial-tangential anisotropy, suggesting that the topology of spatial vision determines numerosity estimation. We suggest that asymmetries of spatial vision should be taken into account when investigating numerosity estimation.
Predictive remapping of visual features beyond saccadic targets
