Abstract
Efficient performance in visual detection tasks requires excluding signals from irrelevant spatial locations. Indeed, researchers investigating visual detection and search have found that performance in a variety of tasks involving multiple potential target locations can be explained simply by the uncertainty the added locations contribute to the task. Importantly, a similar type of location uncertainty may arise within the visual system itself. Converging evidence from hyperacuity and crowding studies shows that spatial localization of features declines in peripheral vision at a rate greater than predicted by the falloff in spatial resolution. This decline in localization ability should add inherent position uncertainty to detection tasks in much the same way as adding potential target locations. The current study used a modified detection task to measure how position uncertainty changes with eccentricity. Subjects judged whether a Gabor target appeared within a cued region of a noisy display. Both the eccentricity and size of the cued region varied across blocks. On trials when subjects reported detecting the target, they used a mouse to indicate the location of the target within the cued region. This allowed measurement of localization errors as well as detection errors. The Intrinsic Uncertainty Observer, an ideal observer degraded with internal response noise and position noise, accounted for both the detection and localization performance of the subjects. The results suggest that internal position uncertainty in normal peripheral vision grows linearly with eccentricity and is independent of target contrast. Internal position uncertainty appears to be a critical factor limiting search and detection performance.
Supported by NIH grant EY02688.