Search

A study explored the phenomenon of space constancy or visual stability of stimulus patterns across saccades (a series of small jerky movements of the eye) by making changes in natural, full-color pictures during selected saccades as observers (18 members of the University of Illinois community) examined them for 20 seconds in preparation for a recognition test. In the first experiment, the pictures were displaced up, down, left, or right by 0.3, 0.6, or 1.2 degrees. In the second experiment, the pictures were expanded or contracted in size by 10% or 20%. As a secondary task, subjects pressed a button whenever they detected a change in the picture. Three results from previous studies with simpler stimuli did not generalize to this situation. Evidence suggests that subjects' detection of image displacements as they examine complex pictures primarily involves the use of local information in the region of the saccade's landing position. Findings suggest support for a saccade target theory, which postulates that a shift of the stimulus is detected on the basis of the post-saccadic retinal location of the object to which the eyes are sent. Further research is needed to identify other sources of evidence for instability in the stimulus configuration from one fixation to the next. (Contains 42 references, 3 notes, 5 tables of data, and 4 figures of data.) (Author/RS)

In the study of perception during reading, the use of eye movement contingent control of the stimulus display has proved to be a useful research technique. With such a system, it is possible to experimentally manipulate, in real time, the characteristics of the stimulus display that is present on selected fixations as reading is in progress and to observe the effects of the manipulation on the eye movement pattern. The technique has been used in a situation where the display must be changed during the period of a saccadic eye movement and when it was necessary to make a change at some time after the beginning of a fixation. Among the instrumentation concerns that are important to consider when setting up a system to do this type of research are the following: (1) the eye movement equipment must have fast throughput and low noise, and yield new information with high frequency; (2) the program must sample the eye position with a high frequency, detect saccade onset and make the decision about whether to initiate a display change as early as the study permits, and minimize the amount of computation involved in changing the image; and (3) the display device must be capable of being rapidly refreshed and of initiating the presentation of a new image part way through the refresh cycle. (FL)

To examine the nature of forward saccadic eye movements in reading, eye movement records were collected from college students as they read a short passage. Forward saccades from this data set were analyzed to determine factors influencing the likelihood of any given letter in the text being the recipient of the next fixation. Data indicated that the likelihood of forward saccades taking the eyes to a particular letter position is a function not only of the distance of that position from the prior fixation, but also of the word length and the letter position in the word being read. The analysis suggests that in reading the eyes are simply sent to the next unidentified word with location preferences in the word being a complex function of length and distance. (Graphs of the data are included.) (MM)

A computer-based Eye-Movement Controlled Display System was developed for the study of perceptual processes in reading. Studies were conducted to identify the region from which skilled readers pick up various types of visual information during a fixation while reading. The results indicated that the subjects acquired word length pattern information at least 12 to 15 character positions to the right of the fixation point, and that this information primarily influenced saccade lengths. Specific letter and word shape information was acquired no further than 12 or 13 character positions to the right of the fixation point. It appeared that words were given a semantic interpretation only if they began no more than 5 or 6 character positions to the right of the fixation point. The perceptual span was found to be asymmetric, with the subjects acquiring information no further than 4 character positions to the left of the fixation point, and perhaps less. (Author/RB)

Three problems in the use of eye movement data for the study of language processing are discussed in this report: the perceptual span problem, the data summary problem, and the eye-mind lag problem. Recent research on perception during reading that bears on these problems is also described. Finally, a general approach to the use of eye movement data for studying language processing, based on present knowledge of perceptual processing and eye movement control during reading, is presented. (Author/FL)

Stressing that it is necessary to have information about the quality of eye movement data in order to judge the degree of confidence one should have in the results of an experiment using eye movement records as data, this report suggests ways for assessing and reporting such information. Specifically, the report deals with three areas: (1) characteristics of the eye movement signal, (2) algorithms used in reducing the data, and (3) accuracy of the eye position data. The paper argues that all studies involving eye movements data should report such information. Appendixes include linear interpolation algorithms for mapping the eye movement signal to stimulus space, and a way of obtaining an index of accuracy for each data point. (Author/FL)

Results from studies of language identification in noise and of word identification from tachistoscopic presentation clearly indicate that contextual information can be used to facilitate word identification under inadequate stimulus conditions. But these results do not provide strong evidence that such an interaction is occurring during normal reading. Other forms of evidence using tasks more similar to reading (proofreading errors and errors in reading aloud) produce the same dilemma. Three aspects of perception have been identified that might be affected by contextual information: where the eyes are sent, the visual region attended to, and the visual information within that region that is used for reading. But studies that have investigated perception during reading are not definitive on these issues; and there currently appears to be no clear evidence that the contextual information environment exerts control over what visual information is used in reading, that is, over the functional stimulus. In fact, subjects appear to be responding to considerable visual detail of words that are almost completely constrained by their prior context. From present evidence, then, it seems quite possible that contextual facilitation is not achieved by reducing the amount of visual information a reader acquires from individual words. (RL)

This report argues that the Mean Gaze Durations or eye movement records used by M. A. Just and P. A. Carpenter to develop a model of reading comprehension are an inappropriate measure of processing time. An alternative approach called Read to Right of Gaze (RRG-1), which assumes that more than one word can be read during a fixation, is recommended instead. Research results are reported that also support the assumption that the perceptual span is asymmetrical to the right, allowing skipped words to be read on preceding fixations; therefore, the processing time for the majority of words in a passage is quite different from the time recorded using Mean Gaze Durations. (AEA)

While most present research suggests that visual information acquired from peripheral visual areas on one fixation during reading facilitates the identification of words available on the next fixation, some researchers hold with the "word unit hypothesis," which suggests that information gained peripherally from a word on one fixation does not facilitate its processing on the next. In a study conducted to discover whether peripheral facilitation was taking place during reading, 16 college students read short texts on a cathode-ray tube as their eye movements were monitored. As they read, the contents of certain word locations changed from fixation to fixation, alternating between two words differing in two letters. The results indicated that the words were read only when directly fixated, and that there was no facilitation from prior peripherally obtained information about the words, thus supporting the word unit hypothesis of information processing. (FL)

The application of computer technology to reading research has made it possible to study the perceptual processes that take place as people read with a level of precision never before attainable. This computer assisted research has revealed that where the eyes go for fixations is of significance to normal reading and that during each fixation only a word or two is being read. In addition, it has shown that when during the fixation a word is perceived the length of fixation varies, probably depending upon when the mind is ready to deal with that information. Eye movement research techniques are likely to be useful in studying higher mental processes occurring during reading and will probably be important in the future for diagnosing specific reading difficulties. (FL)

The research reviewed in this paper concerns the visual perceptual processes occurring as people are engaged in the act of reading. The issues that are examined include the control of eye movements, perception during a fixation, and perception across successive fixations. Consideration also is given to problems related to obtaining information about higher mental processes from eye movement data. (Author/RL)

At some time during every fixation a decision is made to move the eyes, directing them to a new location in the stimulus array. To understand the eye movement control processes, three general hypotheses concerning the cognitive basis for deciding to move the eyes were investigated: the saccade (movement) initiation time is determined only on the basis of information obtained prior to the current fixation; the saccade initiation occurs only after all processing of the presently viewed information has been completed; and the saccade initiation occurs when some processing event takes place before the processing of current information is completed. A review of current research indicates that (1) the time of initiating a saccade can be influenced by visual information acquired on the fixation which terminates in that saccade; (2) processing of the information is not necessarily completed by the time of the decision to move the eyes; (3) with fixations too short to process information, the next saccade must be based entirely on prior information; and (4) part of the variability in the saccade onset time may be associated with the amount of processing time elapsing before the information is utilized. (MM)

A software package is described that collects and reduces eye behavior data (eye position and pupil size) using an IBM-PC compatible computer. Written in C language for speed and portability, it includes several features: (1) data can be simultaneously collected from other sources (such as electroencephalography and electromyography); (2) logically defined events can be detected in real time on any data channel; and (3) either of two types of data matrix can be produced. Data reduction algorithms and data structures are described. (Five tables of data describing aspects of the software package are included.) (Author/RS)

A study of eye movements was conducted to determine whether the visual regions are perceived in their entirety on each fixation or whether the actual region perceived varies from fixation to fixation. The resulting data were used in a frequency of effects analysis. The frequency of effects problem arises once it has been established that an experimental manipulation has produced an effect, and consists of trying to determine the frequency and size of the effect. Text was displayed on a CRT and refreshed every three milliseconds (msec). The subjects' eyes were monitored during reading, with sampling of eye position every msec. During selected fixations, the letters at certain locations were replaced by other letters, thus resulting in erroneous letters being present at specific retinal locations on those fixations. The eye movement data were then analyzed to determine whether errors produced an effect, and if they did, to estimate the frequency with which this occurred. Subjects were 12 college students, each of whom read 16 short passages, which yielded over 1,000 fixations in the control and two experimental conditions. The results indicated that in one instance, a manipulation that produced a 21-msec increase in fixation duration was actually producing a 151-msec increase in only 21% of the instances and was having no effect in the remaining 79% of the cases. The findings suggest that studies of eye movements in reading should use the frequency of effects analysis procedure to determine more accurately the effects of a given manipulation. (HTH)

Research involving eye movement monitoring can help in understanding the nature of the mental processes involved in reading, how these develop as one learns to read, and what processing strategies or characteristics are more common in those children who fail to show normal progress in learning to read. First, eye movement records show that the eyes pause longer on some words than on others, move various distances and directions between pauses, and exhibit patterns of movements indicating what is being attended to at different moments, what kinds of processing are taking place, and when processing difficulties arise. Second, eye movement data are useful in analyzing simultaneously collected data, such as brain wave records or oral reading protocols. Finally, eye movement data can be used for controlling experimental manipulations during ongoing reading. The ability to control the nature of the stimulus pattern that is present on any given fixation during reading provides a powerful technique for studying the perceptual processes that are taking place. Research techniques involving the monitoring of eye movements provide powerful new ways of studying the characteristics of the reading process, even when eye movements themselves are not involved in the etiology of the reading disorder. (HOD)

To investigate the relationship between the location of the words being read and the location of the eyes in the text, three experiments were conducted using the Disappearing Text Technique with college students. This was done by occasionally removing the text during reading and having the reader report the last word that had been read. Distributions of the locations of the last word read, in relation to the location of the last word on which the eyes were centered prior to the text's disappearance, served as data and were compared for different conditions. The first study compared these distributions when the text disappeared during rather than following a fixation and when the text was simply blanked out rather than being replaced by a masking pattern. The second study manipulated the time during the fixation at which the text disappeared, and the third study replicated results from the earlier studies. Results showed that, on the one hand, the last word reported as having been read was most commonly the word fixated during the last fixation, or the word to the right of it. On the other hand, the responses were not restricted to these two word locations. Thus, the data did not support an anticipation model of reading nor the acquisition of peripheral cues concerning upcoming words. (HOD)

Eye movement monitoring is useful both in the control of experiments on reading and as a source of data. Experiments using eye monitoring techniques have helped develop the following conclusions about the reading process: the region of text read during a fixation is quite small and asymmetric to the right of the center of vision, successive fixations are essentially nonoverlapping, the durations of fixations reflect cognitive processing occurring at that point in the text, and the eye is directed to specific locations in the text during reading. To continue exploring the reading process with eye movement techniques, a monitoring system was developed and tested using a Biometrics eye movement monitor connected to a PDP-11/40 computer. As the eye monitor reported eye positions 1,000 times per second, the computer controlled text display on a cathode ray tube (CRT) and recorded all data for analysis. Printouts of the data showed how eyes moved across a line of text, eye fixations and their lengths, movements between fixations, and changes in text display. Since the computer can be programed to make adjustments in the text display while eye movements are in progress, it provides greater control over experimental data and a faster means of verifying and refining previous research. (RL)