A neurophysiological explanation for biases in visual localization

详细信息    查看全文

• 作者：James C. Moreland ; Geoffrey M. Boynton
• 关键词：Summary statistics ; Localization ; Vision ; Mid ; level mechanisms ; Robust statistics
• 刊名：Attention, Perception, & Psychophysics
• 出版年：2017
• 出版时间：February 2017
• 年：2017
• 卷：79
• 期：2
• 页码：553-562
• 全文大小：
• 刊物主题：Cognitive Psychology;
• 出版者：Springer US
• ISSN：1943-393X
• 卷排序：79

文摘

Observers show small but systematic deviations from equal weighting of all elements when asked to localize the center of an array of dots. Counter-intuitively, with small numbers of dots drawn from a Gaussian distribution, this bias results in subjects overweighting the influence of outlier dots – inconsistent with traditional statistical estimators of central tendency. Here we show that this apparent statistical anomaly can be explained by the observation that outlier dots also lie in regions of lower dot density. Using a standard model of V1 processing, which includes spatial integration followed by a compressive static nonlinearity, we can successfully predict the finding that dots in less dense regions of an array have a relatively greater influence on the perceived center.