David Zola

Integrating information across eye movements

Abstract A series of experiments was conducted in which a word initially appeared in parafoveal vision, followed by the subjects eye movement to the stimulus. During the eye movement, the initially displayed word was replaced by a word which the subject read. Under certain conditions, the prior parafoveal word facilitated naming the foveal word. Three alternative hypotheses were explored concerning the nature of the facilitation. The verbalization hypothesis suggests that information acquired from the parafoveal word permits the subject to begin to form the speech musculature properly for saying the word. The visual features integration hypothesis suggests that visual information obtained from the parafoveal word is integrated with foveal information after the saccade. The preliminary letter identification hypothesis suggests that some abstract code about the letters of the parafoveal word is stored and integrated with information available in the fovea after the saccade. The results of the experiments supported the latter hypothesis in that information about the beginning letters of words was facilitatory in the task. The other two hypotheses were disconfirmed by the results of the experiments.

Eye movement control during reading : I. the location of initial eye fixations on words

Sixty-six college students read two chapters from a contemporary novel while their eye movements were monitored. The eye movement data were analyzed to identify factors that influence the location of a readers initial eye fixation on a word. When the data were partitioned according to the location of the prior fixation (i.e. launch site), the distribution of fixation locations on the word (i.e. landing site distribution) was highly constrained, normal in shape, and not influenced by word length. The locations of initial fixations on words can be accounted for on the basis of five principles of perceptuo-oculomotor control: a word-object has a specific functional target location, a saccadic range error occurs that produces a systematic deviation of landing sites from the functional target location, the saccadic range error is reduced somewhat for saccades that follow longer eye fixations, there exists perceptuo-oculomotor variability that is a second, nonsystematic source of variation in landing sites, and the perceptuo-oculomotor variability increases with distance of the launch site from the target.

Is visual information integrated across successive fixations in reading

College students read a passage presented in AlTeRnAtInG cAsE on a CRT while their eye movements were monitored. During certain saccades, the case of every letter was changed (a became A, B became b). This change was not perceived and had no effect on eye movements. Apparently visual features of the type which specify the difference between upper- and lowercase letters are not integrated across fixations during reading.

Eye movement control during reading: II. Frequency of refixating a word.

An analysis of over 40,000 eye fixations made by college students during reading indicates that the frequency of immediately refixating a word following an initial eye fixation on it varies with the location ofthat fixation. The refixation frequency is lowest near the center of the word, posi-tively accelerating with distance from the center. The data are well fit by a parabolic function. Assuming that refixation frequency is related to the frequency of successful word identification, the observed curvilinear relation results naturally from models that postulate a linear decrease in visual information with retinal eccentricity. A single letter difference in fixation location in a word can make a sizeable difference in the likelihood of refixating that word. The effects of word length and cultural frequency on the frequency of refixating are also examined.

Redundancy and word perception during reading.

The experimental investigation reported in this article deals with the process of extracting visual information during reading. Subjects read texts that manipulated the predictability of a target word through the choice of an immediately preceding word. Spelling errors were also introduced into some of the target words. A detailed examination was made of the subjects’ eye-movement patterns. Several aspects of eye behavior were analyzed to determine if contextual constraint and misspelling influenced perception during reading. Subjects exhibited no differences in the frequency of fixating the target words in the high-constraint and low-constraint conditions. However, the fixation of the target words was shorter in the high-constraint condition. In addition, fixation durations and regression probabilities associated with misspelled words were also significantly inflated. The eye-movement patterns showed that minimal spelling errors often disrupted reading, even when the misspelled words were highly predictable. These results suggest that language constraint does expedite processing during reading; furthermore, they suggest that such facilitation does not necessarily occur through a reduction in the visual analysis of the text.

Time course of visual information utilization during fixations in reading

College students read short texts from a cathode-ray tube as their eye movements were being monitored. During selected fixations, the text was briefly masked and then it reappeared with one word changed. Subjects often were unaware that the word had changed. Sometimes they reported seeing the first presented word, sometimes the second presented word, and sometimes both. When only one word was reported, two factors were found to determine which one it was: the length of time a word was present during the fixation and the predictability of a word in its context. The results suggested that visual information is utilized for reading at a crucial period during the fixation and that this crucial period can occur at different times on different fixations. The pattern of responses suggested that the first letter of a word is not utilized before other letters and that letters are not scanned from left to right during a fixation.

Some temporal characteristics of processing during reading.

College students read passages displayed on a cathode-ray tube as their eye movements were being monitored. During occasional fixations, all letters to the left of the directly fixated letter or all letters more than four to the right of the fixated letter were replaced by other letters. This replacement occurred either for only the first 100 ms of the fixation or only after the first 100 ms of the fixation. The eye movement data indicated that the eyes can respond to change in the visual stimulus within less than 100 ms and to orthographic irregularity in the text within less than 160 ms. No evidence was found for a left-to-right attentional scan during a fixation. The results were interpreted within the framework of a chronology of processing events occurring during a fixation in reading.

Perceiving words during reading: Lack of facilitation from prior peripheral exposure

As their eye movements were being monitored, college students read short texts displayed on a cathode-ray tube. As they read, the contents of certain word locations changed from fixation to fixation, alternating between two words differing in two letters. This manipulation had no effect on reading unless the subjects happened to regress to or reread the word later. The results indicated that these words, which were low in contextual constraint, were read only when directly fixated, and that there was no facilitation from prior peripherally obtained information about the words.

Toward the use of eye movements in the study of language processing

Three problems in the use of eye movement data for the study of language processing are discussed: the perceptual span problem, the data summary problem, and the eye‐mind lag problem. Recent research on perception during reading is described which bears on these problems. Finally, a general approach to the use of eye movement data for studying language processing is presented, based on present knowledge of perceptual processing and eye movement control during reading.

Chapter 5 – About Regressive Saccades in Reading and Their Relation to Word Identification

Publisher Summary This chapter illustrates regressive saccades in reading and their relation to word identification. While reading, the most common pattern is that of the eyes moving forward from one word to the next, with fixations lasting for about 250 ms. Variations in this general pattern are also observed, such as the eyes skipping the next word, or making an additional fixation on the word before moving on to the next word(s). Occasionally, the eyes also return to previously read portions of the text, which correspond to what is commonly called “regressive saccades” in opposition to the more frequent “progressive saccades”. Research has been conducted to determine the extent to which this variability in ocular behavior, and particularly that associated with progressive movements, relates to ongoing processing of the encountered words and sentences. A single mechanism cannot be responsible for the variability of the ocular behavior in reading. Both higher level lexical and linguistic characteristics of the text and lower level “visuomotor” constraints and predetermined “oculomotor” scanning strategies contribute to this variability. A “regressive saccade”, on the other hand, is initiated, if a review of previously read text is necessitated to “re encode” it or further process it to deeper levels.