The design was to display four geometric shapes on a screen, each with a different color. The subjects viewed the screen without moving their eyes, and then the screen was blanked. Next, one of the items was redisplayed, but in a different location and rotation. The working memory task was to remember the original orientation and location. The objects were then withdrawn. Typically, subjects recalled well as long as the set size was no larger than four and as long as the probe object was not changed much. But precision of recall fell off drastically, even with only two or three objects, if the probe object was moved significantly.
Next, they wanted to explore the effects of moving the eyes across the stimulus field. Actually, this presents a major challenge to the brain, because more information has to be held in working memory. Even so, the results when eye movement was allowed and when it was not seemed about the same. The decline in accuracy depended mostly on the number of objects in the set. This suggests that the brain parcels out the task so that memory resources are assigned proportionately to each item. But eye movements are important. In another study, subjects made a series of eye movements to fixate on each item in a five-set display, and the screen was blanked just before the fifth item was fixated. Then subjects were tested to report the location and orientation of the objects. Accuracy was greatest for the fifth item, presumably because the brain had allocated more of its memory resources to registering the item the eyes were about to fixate. This, of course, is being done at the expense of remembering all of the previous four fixations.
These facts must have profound implications for reading comprehension. Certainly, these data seem to support the value of “whole reading” theory, which emphasizes fixing eyes on whole words, preferably where one eye fixation takes in several words.* The point is that if brain allocates more of its memory resources to each upcoming eye fixation, then reducing the number of fixations ought to increase comprehension, as indicated by working memory of what is seen with each fixation. In other words, the brain can process the meaning of multiple words in an upcoming fixation because the brain is devoting more memory resources to the next fixation task.
*This is not to denigrate phonics, for there is compelling evidence that phonics is the best way to learn words in the first place. But once learned, remembering the words you read is probably best achieved by reducing the number of eye fixations.
Source:Bays, P. M., and Husain, M. 2008. Dynamic shifts of limited working memory resources in human vision. Science. 321: 851-85