SBE-UKRI: "Looked but failed to see" errors in human vision.

点击翻译按钮获取中文摘要

It often happens that we fail to notice something that is in plain view. Sometimes this can even be something that we are actively looking for, like typos in our writing. These “looked but failed to see” (LBFTS) errors can be potentially dangerous, as when a radiologist fails to see signs of cancer in a lung x-ray or a driver fails to see someone crossing the road. The aim of this project is to understand the source of these LBFTS errors and to investigate possible ways to reduce them.There are a variety of different causes that might lead someone to fail to report something that is clearly visible when pointed out. For example, some LBFTS errors can be thought of as a version of bad luck: You can't process everything, so only some stimuli get processed adequately. The experiments in Aim 1 investigate whether LBFTS errors in some tasks are random or whether there are attributes of the missed stimuli that make them less likely to be found. A second possibility, explored in Aim 2, is that there are perceptual biases within the functional visual field or useful field of view, defined as the region around the point where you are looking, within which you could find your target (even if you don’t). For example, if you are reading English, while you are looking at one word, you are biased to the next words to the right rather than to the words directly above or below where you are fixating. When reading Arabic or Hebrew, you would be biased to words on the left of your fixation point. The two principal investigators on this project have different ideas about where the biases in the functional visual field might come from and they exploit an "adversarial collaboration" in order to resolve these differences. Finally, Aim 3 is devoted to LBFTS errors that occur when doing one task but still "keeping an eye out" for other stimuli (e.g. when a doctor is examining a chest x-ray, it may show more than just the fractured rib about which the patient is complaining, such as a suspicious sign of cancer). The investigators will try to develop ways to help observers do their primary task while staying attuned to the possible occurrence of important but non-related events. If the sources of LBFTS errors can be identified and reduced, it could have a major impact on important societal arenas, such as medical image perception and traffic accidents.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.