Eye movement in reading
Eye movement in reading involves the visual processing of written text. This was first described by the French ophthalmologist Louis Émile Javal in the late 19th century. He reported that eyes do not move continuously along a line of text, but make short, rapid movements (saccades) intermingled with short stops (fixations). Javal's observations were characterised by a reliance on naked-eye observation of eye movement in the absence of technology. From the late 19th to the mid-20th century, investigators used early tracking technologies to assist their observation, in a research climate that emphasised the measurement of human behaviour and skill for educational ends. Most basic knowledge about eye movement was obtained during this period. Since the mid-20th century, there have been three major changes: the development of non-invasive eye-movement tracking equipment; the introduction of computer technology to enhance the power of this equipment to pick up, record and process the huge volume of data that eye movement generates; and the emergence of cognitive psychology as a theoretical and methodological framework within which reading processes are examined. Sereno & Rayner (2003) believed that the best current approach to discover immediate signs of word recognition is through the recordings of eye movements and event-related potential.
- 1 Saccades
- 2 History
- 3 Cognitive psychology, infrared tracking and computer technology
- 4 Reading speed in dyslexia
- 5 Computer models of eye movement in reading
- 6 See also
- 7 Notes
- 8 References
- 9 External links
Skilled readers move their eyes during reading on the average of every quarter of a second. During the time that the eye is fixated, new information is brought into the processing system. Although the average fixation duration is 200–250 ms (thousandths of a second), the range is from 100 ms to over 500 ms. The distance the eye moves in each saccade (or short rapid movement) is between 1 and 20 characters with the average being 7–9 characters. The saccade lasts for 20–40 ms and during this time vision is suppressed so that no new information is acquired. There is considerable variability in fixations (the point at which a saccade jumps to) and saccades between readers and even for the same person reading a single passage of text. Skilled readers make regressions back to material already read about 15 percent of the time. The main difference between faster and slower readers is that the latter group consistently shows longer average fixation durations, shorter saccades, and more regressions. These basic facts about eye movement have been known for almost a hundred years, but only recently have researchers begun to look at eye movement behavior as a reflection of cognitive processing during reading.
If the perceptual span includes all or many of the words on a line of text, then eye movement measures would not likely reveal much information about cognitive processing; however, if the reader gains useful information only from the word directly focused on, then eye movement behavior could shed light on what role the eyes play in reading disorders such as dyslexia.
The lower line of text simulates the acuity of vision with the relative acuity percentages. The difficulty of recognizing text increases with the distance from the fixation point.
Unassisted observation, optical modelling and psychological introspection
Until the second half of the 19th century, researchers had at their disposal three methods of investigating eye movement. The first, unaided observation, yielded only small amounts of data that would be considered unreliable by today's scientific standards. This lack of reliability arises from the fact that eye movement occurs frequently, rapidly, and over small angles, to the extent that it is impossible for an experimenter to perceive and record the data fully and accurately without technological assistance. The other method was self-observation, now considered to be of doubtful status in a scientific context. Despite this, some knowledge appears to have been produced from introspection and naked-eye observation. For example, Ibn al Haytham, a medical man in 11th-century Egypt, is reported to have written of reading in terms of a series of quick movements and to have realised that readers use peripheral as well as central vision.
Leonardo da Vinci, (1452–1519) may have been the first in Europe to recognize certain special optical qualities of the eye. He derived his insights partly through introspection but mainly through a process that could be described as optical modelling. Based on dissection of the human eye he made experiments with water-filled crystal balls. He wrote "The function of the human eye, ... was described by a large number of authors in a certain way. But I found it to be completely different."
His main experimental finding was that there is only a distinct and clear vision at the "line-of-sight", the optical line that ends at the fovea. Although he did not use these words literally he actually is the father of the modern distinction between foveal vision (a more precise term for central vision) and peripheral vision. However, Leonardo did not know that the retina is the sensible layer, he still believed that the lens is the organ of vision.
There appear to be no records of eye movement research until the early 19th century. At first, the chief concern was to describe the eye as a physiological and mechanical moving object, the most serious attempt being Hermann von Helmholtz's major work Handbook of physiological optics (1866). The physiological approach was gradually superseded by interest in the psychological aspects of visual input, in eye movement as a functional component of visual tasks. As early as the 1840s, there was speculation on the relationship between central and peripheral vision.
The subsequent decades saw more elaborate attempts to interpret eye movement, including a claim that meaningful text requires fewer fixations to read than random strings of letters. In 1879, the French ophthalmologist Louis Émile Javal used a mirror on one side of a page to observe eye movement in silent reading, and found that it involves a succession of discontinuous individual movements for which he coined the term saccades. In 1898, Erdmann & Dodge used a hand-mirror to estimate average fixation duration and saccade length with surprising accuracy.
Early tracking technology
Eye tracking device is a tool created to help measure eye and head movements. The first devices for tracking eye movement took two main forms: those that relied on a mechanical connection between participant and recording instrument, and those in which light or some other form of electromagnetic energy was directed at the participant's eyes and its reflection measured and recorded. In 1883, Lamare was the first to use a mechanical connection, by placing a blunt needle on the participant's upper eyelid. The needle picked up the sound produced by each saccade and transmitted it as a faint clicking to the experimenter's ear through an amplifying membrane and a rubber tube. The rationale behind this device was that saccades are easier to perceive and register aurally than visually. In 1889, Edmund B. Delabarre invented a system of recording eye movement directly onto a rotating drum by means of a stylus with a direct mechanical connection to the cornea. Other devices involving physical contact with the surface of the eyes were developed and used from the end of the 19th century until the late 1920s; these included such items as rubber balloons and eye caps.
Mechanical systems suffered three serious disadvantages: questionable accuracy due to slippage of the physical connection, the considerable discomfort caused to participants by the direct mechanical connection (and consequently great difficulty in persuading people to participate), and issues of ecological validity, since participants' experience of reading in trials was significantly different from the normal reading experience. Despite these drawbacks, mechanical devices were used in eye movement research well into the 20th century.
Attempts were soon made to overcome these problems. One solution was to use electromagnetic energy rather than a mechanical connection. In the "Dodge technique", a beam of light was directed at the cornea, focused by a system of lenses and then recorded on a moveable photographic plate. Erdmann & Dodge used this technique to claim that there is little or no perception during saccades, a finding that was later confirmed by Utall & Smith using more sophisticated equipment. The photographic plate in the Dodge technique was soon replaced with a film camera, but was still plagued by problems of accuracy, due to the difficulty of keeping all parts of the equipment perfectly aligned throughout a trial and accurately compensating for the distortion caused by the diffractive qualities of photographic lenses. In addition, it was usually necessary to restrain a participant's head by using an uncomfortable bite-bar or head-clamp.
In 1922, Schott pioneered a further advance called electro-oculography (EOG), a method of recording the electrical potential between the cornea and the retina. Electrodes may be covered with special contact paste before being placed on the skin. So, it is now unnecessary to make incisions in patient's skin. Common misconception about EOG is that measured potential is the electromyogram of extraocular muscles. In fact, it is only the projection of eye dipole to the skin, because higher frequencies, corresponding to EMG, are filtered out. EOG delivered considerable improvements in accuracy and reliability, which explain its continued use by experimentalists for many decades.
Cognitive psychology, infrared tracking and computer technology
There are 4 major cognitive systems involved in eye movement in reading: Language processing, attention, vision, and oculomotor control. Eye trackers bounce near infra-red light off the interior of the eyeball, and monitor the reflection on the eye to determine gaze location. With this technique, the exact position of eye fixation on a screen is determinable. Wang (2011) mentioned that a video-based eye-tracker which uses video cameras to record the eye position of human subjects, thereby recording pupil dilation and eye movements, can be used to examine how fixations, saccades, and pupil dilation responses are related to the information on the screen and behavioral choices during an experiment. According to Wang (2011:185), "understanding the relationship between these observables can help us to understand how human behavior in the economy can be affected by what information people acquire, where their attention is focused, what emotional state they are in, and even what brain activity they are engaged in. This is because fixations and saccades (matched with information shown on screen) indicate how people acquire information (and what they see), time lengths of fixations indicate attention, and pupil dilation responses indicate emotion, arousal, stress, pain, or cognitive load."
Reading speed in dyslexia
People with dyslexia generally suffer from a decreased reading speed, which can be caused by many different variables. There are many remedies to try and combat these deficits, depending on what biological theory of dyslexia they are based on. One such idea is based on magnocellular deficit, where magnocellular pathways are uncoordinated, causing the skipping or re-reading of lines.
Computer models of eye movement in reading
In recent years researchers have developed various eye-movement simulation models in order to support, explain, and guide empirical research (the studies that collect data from experiments on human reading behaviours). Computational models can help test existing hypothesis, and also generate further predictions for empirical research to investigate. Building a reading model based on empirical data allows researchers to inspect and manipulate various assumptions of cognitive processes within the model, and see how the interactions between these elements influence the simulated reading behaviours, and hence informs further empirical studies. On the other hand, empirical research in eye-movements using eye-tracking is one of the most delicate measurement of human behaviours, which serves as a productive test bed for computer models of cognition. Below is a brief summary of these reading models as classified by Nuthmann (2014).
Reading models can be classified by whether they emphasize on lexical processes (Reader, EMMA, E-Z Reader, SWIFT) or oculomotor control processes (Competition-Interaction Theory, SERIF). Models in the first group focus on the effect of relatively high-level cognitive processes like those on word frequency, word parsing or word predictability, while models in the second group focus on the more primitive mechanisms of oculomotor processes in reading such as how word length of currently fixated word and its neighbour words affect saccade amplitude and latency (or fixation duration).
Those models that emphasize on lexical processes can be further divided into two groups based on how attention is allocated in the reading model. The main stream model E-Z Reader (Reichle, Rayner, & Pollatsek, 2003), along with older ones like Reader (Just & Carpenter, 1980) and EMMA (Salvucci, 2001) assumes that attention is allocated to only one word sequentially. On the contrary, the SWIFT model assumes that attention in reading is distributed among the currently fixated word and its neighbours. Since attention allocation determines the input stream in these models, difference in attention allocation results in fundamental difference in the subsequent processing and simulation output. E-Z Reader incorporates various lexical properties into a complex structure of interaction between lexical processing, saccade programming and attention shifting; it reproduces the widest range of phenomena in reading behaviours and can be generalised to reading behaviours in elder people, in different languages and in different types of text presentation (Reichle, 2011). In contrast, SWIFT resorts to a dynamic field of parallel activation that stochastically triggers saccade programming, yielding parsimonious simulations of refixation, regression (fixate back to a previous word), and neighbouring effects (how neighbouring words affect the current fixation) (Engbert & Kliegl, 2011). The debate is still on whether attention in reading should be relaxed to more than one word.
For those models that emphasize on oculomotor processes, the aim is not to offer an overarching model that simulates reading behaviours comprehensively, but to investigate how some basic oculomotor processes may account qualitatively for some phenomena in reading behaviours. Along this line, Yang and McConkie (2001) use modeling to investigate the competition between signals that trigger and inhibit saccade initiation, concluding that only saccades after long saccade latency (>225 ms ) are affected by cognition. Another model SERIF (McDonald, Carpenter, & Shillcock, 2005) studies the effect of fovea split on reading behaviours; the model extends fovea split to hemispheres difference and competition, and explains the IOVP effect (Vitu et al., 2001) which is regarded by E-Z Reader as an artifact of misallocated fixations. The advantage of this approach resides in that the oculomotor processes have material basis and may undercut some of those assumptions of lexical processes made by E-Z Reader and SWIFT.
There are problems that may hinder the development the modeling of reading behaviours. First, computer models in reading tend to make assumptions of direct links between lexical properties and saccade attributes (by putting a formula linking the two into the model). These links are descriptive generalizations of existing empirical data and normally do not survive new data. In search for new data to support their models, researcher might have to keep adding and adjusting lexical process assumptions and are likely to ignore possible parsimonious process that underlies these assumptions. Caution should be taken when adding assumptions of how lexical processes affect saccade programming because reading is not what humans are evolved to do. The modeling of eye-movements in reading should treat reading as special case of visual perception and take into account more basic oculomotor properties(e.g., split fovea, depth perception) before making assumptions about the link between lexical processes and eye movements. Another problem is that there has been a lack of comparison between models on the same data set. Researcher teams tend to focus on the development of their own model and might not be able to fully appreciate the merits of other models due to the technical difficulties in sharing programs. In the future we might see unifying model that both explains complex reading behaviors well and is parsimonious enough to be generalizable to other eye-tracking fields such as visual search, scene viewing.
- Biological theories of dyslexia
- Eye movement (sensory)
- Eye movement in music reading
- Gaze-contingency paradigm
- Fixation (visual)
- Eye tracking
- Eye tracking device
- Sereno & Rayner (2003).
- Rayner, Foorman, Perfetti, Pesetsky, & Seidenberg (2001).
- Rayner, Slattery, Belanger (2010).
- Rayner K (1975).
- Hans-Werner Hunziker, (2006) Im Auge des Lesers: foveale und periphere Wahrnehmung – vom Buchstabieren zur Lesefreude [In the eye of the reader: foveal and peripheral perception – from letter recognition to the joy of reading] Transmedia Stäubli Verlag Zürich 2006, ISBN 978-3-7266-0068-6.
- Heller (1988:39).
- Leonardo da Vinci (1955), das Lebensbild eines Genies, Emil Vollmer Verlag, Wiesbaden Berlin. Dokumentation der Davinci Ausstellung in Mailand 1938, p. 430; cited in 'Hans-Werner Hunziker, (2006) Im Auge des Lesers: foveale und periphere Wahrnehmung – vom Buchstabieren zur Lesefreude' [In the eye of the reader: foveal and peripheral perception – from letter recognition to the joy of reading] Transmedia Stäubli Verlag Zürich 2006 ISBN 978-3-7266-0068-6.
- Hueck (1840), Weber (1846).
- Cattell (1885, 1886).
- Rayner, Pollatsek, & Alexander (2005).
- Lamare (1893).
- Delabarre (1898).
- Erdmann B & Dodge R (1898).
- Schott E (1922).
- Finocchio, Preston, & Fuchs (1990).
- Liu, Zhou, Hu (2011).
- Tecce, Pok, Consiglio, O'Neil (2005).
- "Eye Movements During Reading". Pitt.edu. Retrieved 2014-04-14.
- wang (2011).
- Stein J., The magnocellular theory of developmental dyslexia Dyslexia. 2001 Jan–Mar;7(1):12–36.
- Abadi, R. V. (2006). Vision and eye movements. Clinical and Experimental Optometry, 55–56.
- Delabarre E.B. (1898) A method of recording eye-movements, Psychological Review 8, 572–74.
- Engbert, R. & Kliegl, R. (2011) Parallel graded attention models of reading. The Oxford handbook of eye movements. Liversedge, S., Gilchrist, I., & Everling, S. (Eds.) Oxford University Press.
- Erdmann B & Dodge R (1898) Psychologische Untersuchung über das Lesen auf experimenteller Grundlage, Niemeyer: Halle.
- Finocchio, Dom; Preston, Karen L; Fuchs, Albert F. (1990). "Obtaining a quantitative measure of eye movements in human infants: A method of calibrating the electrooculogram". Vision Research 30(8): 1119–28. doi:10.1016/0042-6989(90)90169-L.
- Heller D (1988) "On the history of eye movement recording" in Eye movement research: physiological and psychological aspects, Toronto: CJ Hogrefe, 37–51.
- Helmholtz H (1866) Handbuch der physiologischen Optik, Voss: Hamburg.
- Hunziker, H. (2006). Im Auge des Lesers: foveale und periphere Wahrnehmung – vom Buchstabieren zur Lesefreude [In the eye of the reader: foveal and peripheral perception – from letter recognition to the joy of reading] Transmedia Stäubli Verlag Zürich 2006, ISBN 978-3-7266-0068-6.
- Javal, E. (1878) "Essai sur la physiologie de la lecture", in Annales d'ocullistique 80, 61–73.
- Just, M.A., & Carpenter, P.A. (1980). A theory of reading: from eye fixations to comprehension. Psychological review, 87(4), 329.
- Lamare, M. (1893) Des mouvements des yeux pendants la lecture, Comptes rendus de la société française d'ophthalmologie, 35–64.
- Liu, Y.; Zhou, Z.; Hu, D. (2011). "Gaze independent brain-computer speller with covert visual search tasks". Clinical Neurophysiology 122(6): 1127–36. doi:10.1016/j.clinph.2010.10.049. Retrieved 1 November 2011.
- McDonald, S. A., Carpenter, R. H. S., & Shillcock, R. C. (2005). An anatomically constrained, stochastic model of eye movement control in reading. Psychological review, 112(4), 814.
- Nuthmann, A. (2014, September). Eye movements and visual cognition lecture 2 (University of Edinburgh, UK).
- Rayner, K.; Foorman, B.; Perfetti, C.; Pesetsky, D. & Seidenberg, M. (2001). How psychological science informs the teaching of reading. Psychological Science in the Public Interest 2(2): 31–74.
- Rayner, K.; Slattery, Timothy J; Belanger, Nathalie N. (2010). Eye movements, the perceptual span, and reading speed. Psychonomic Bulletin & Review 17(6): 834–39. doi:10.3758/PBR.17.6.834. Retrieved 1 November 2011.
- Rayner K. (1975). Eye movements, perceptual span, and reading disability, Annals of Dyslexia, 33(1), 163–73. doi:10.1007/BF02648003
- Rayner; K.; Pollatsek, J.; Alexander, B.(2005). Eye movements during reading. The science of reading: A handbook. [1-4051-1488-6 ]: Blackwell Publishing. pp. 79–97. ISBN 1-4051-1488-6 (Hardcover); 978-1-4051-1488-2.
- Reichle, E. (2011). Serial-attention models of reading. The Oxford handbook of eye movements. Liversedge, S., Gilchrist, I., & Everling, S. (Eds.) Oxford University Press.
- Reichle, E.D., Rayner, K., & Pollatsek, A. (2003). The EZ Reader model of eye-movement control in reading: comparisons to other models. Behavioral and brain sciences, 26(04), 445–76.
- Salvucci, D.D. (2001). An integrated model of eye movements and visual encoding. Cognitive Systems Research, 1(4), 201-220.
- Schott E (1922) Über die Registrierung des Nystagmus und anderer Augenbewegungen vermittels des Saitengalvanometers, Deutsches Archiv für klinisches Medizin 140, 79–90.
- Sereno, S.; Rayner, K. (2003). Measuring word recognition in reading: eye movements and event-related potentials. Trends in Cognitive Science, 7(11): 489–93.
- Tecce, J.; Pok, L.J.; Consiglio, M.R.; O'Neil, J.L. (2005). Attention impairment in electrooculographic control of computer functions. International Journal of Psychophysiology, 55(2): 159–63. doi:10.1016/j.ijpsycho.2004.07.002. Retrieved 1 November 2011.
- Vitu, F., McConkie, G.W., Kerr, P., & O'Regan, J.K. (2001). Fixation location effects on fixation durations during reading: An inverted optimal viewing position effect. Vision research, 41(25), 3513–33.
- Wang, J. (2011). "Pupil dilation and eye-tracking." A handbook of process tracing methods for decision research: a critical review and user's guide: Society for Judgment and Decision Making Series. pp. 185–204. ISBN 1-84872-864-6.
- Yang, S.-N., & McConkie, G.W. (2001). Eye movements during reading: a theory of saccade initiation times. Vision Research, 41, 3567–85.