The Deployment of Visual Attention
Abstract
This is the final report summarizing research on the deployment of attention in visual search. In visual search, observers look for targets among distractor items. Models of search had assumed that items were sampled without replacement. items, rejected as distractors, would not be revisited during the search. Research reported here falsifies that hypothesis. Our data are consistent with sampling with replacement - no use of information about rejected distractors. The data do not reject models that posit small amounts of memory for rejected distractors (e.g. don't revisit the last N rejected items; N < 7). why is search so apparently random? Observers could have the benefits of memory for rejected distractors if they searched in an orderly manner (e.g. "reading" a display from left to right). Our second line of experiments shows that the temporal costs of such a strategy are too high. Left to its own devices, covert visual search proceeds at about 20-30 items/second. Commanded shifts of attention proceed at only 3-5 shifts/second. For the tasks used here, anarchic covert search will be faster than commanded, orderly search. Our Guided Search model is adapted readily to search with replacement. Nature may have stumbled on a way to perform sophisticated search tasks with simple tools.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 2003
- Accession Number
- ADA416391
Entities
People
- Jeremy M. Wolfe
Organizations
- Brigham and Women's Hospital