We move our eyes for a variety of reasons. We may look upward as a surreptitious signal that we do not believe what someone else is saying, or avert our gaze out of embarrassment or modesty. But even when we are not sending or responding to social signals, our eyes are often on the move for a quite different reason — to satisfy the needs of vision. One might think that animals move their eyes in order to look from place to place, but oddly enough the oldest (in evolutionary terms) and commonest function of eye movements is to keep the image of the world stationary on the retina. The eyes sit in a mobile container (the head), itself attached to an often mobile body: if the eyes did not move to compensate for movements of the head, the image of the world would slither around on the retina and so be more difficult for the brain to analyse.
Some privileged animals, especially humans and other primates, can actively move their eyes to shift the image around on the retina. We do this because the central part of the retina, the fovea, is specialized for acute vision. When you try to recognize something, such as the printed letters on an optometrist's chart, you point your fovea directly at it and can then resolve gaps in the letters as small as 1 mm at a distance of 6 m (the usual distance for an optometrist's chart). If you were to look at the chart ‘out of the corner of your eye’ (with your far peripheral retina), the letters would have to be 100 times larger to be recognized. Because the fovea is so small, it is necessary to move the eyes in order to explore the scene at high resolution. To convince yourself how quickly visual resolution falls off in the peripheral retina, try to read the words at the edges of the page of a book while holding your eye fixedly on the middle of the page. Whatever the merits of ‘speed reading’, there is no possibility, as some claim, that every word can be recognized if the line of sight is simply moved down the middle of the page.
Fig. 1 The optokinetic response to rotation of the visual surround
A proper classification of human eye movements, still valid, was provided by the American psychologist Raymond Dodge, at the beginning of the twentieth century. He distinguished two types of compensatory movements: the vestibulo-ocular reflex (VOR) and the optokinetic response (OKR), both of which compensate for head movement. And he described three types of active, exploratory movement: (i) saccades — rapid shifts in the line of sight that are used to look at new targets; (ii) smooth pursuit — smooth movements of the eyes to follow a moving target; and (iii) vergence — alteration of the angle between the lines of sight of each eye, either increasing the angle (convergence) to look at close objects, or decreasing it (divergence), so as to look at things far away.
Stabilization movements
The most common cause of instability of the retinal image is one's own movement. If this is slow, then the whole image begins to slip across the retina; this overall movement is detected by the brain, which initiates a reflex, compensatory optokinetic eye movement. But when the head moves quickly, signals from the retina itself are too slow to drive useful compensatory movements: it takes 1/20 sec or so for visual signals to leave the retina, let alone travel through the various brain pathways to control eye movements. On the other hand, the organ of balance (the vestibular apparatus) in the inner ear, which detects both rapid rotations and linear motion of the head, has negligible internal delays. Its signals cause vestibulo-ocular reflex movements of the eyes, which compensate quite precisely for the head movement, even in complete darkness. Of course, if the head were to continue to rotate through a very large angle, uninterrupted counter-rotation of the eyes would carry them to the limit of travel in their sockets. In fact this does not happen because the smooth movement is frequently interrupted by rapid jerks of the eyes in the opposite direction, tending to re-centre the eyes in the orbits. This fast-slow-fast-slow, back and forth movement of the eyes is called nystagmus.
Fig. 2 Eye movements in reading
The VOR operates for rotations about all axes, even around the direction in which the eye is looking. Get someone to sit on an office chair that can turn, looking directly upward along the axis of the spindle of the chair. If the chair is rotated gently, a torsional VOR occurs: the eyes counter-rotate around the line of sight, flicking back from time to time.
While the VOR is remarkably accurate for fast head movements, even a small error will cause the image to slither across he retina during head movement. Such image movement activates a second stabilization response: the OKR. This visually-driven response may be demonstrated by placing a volunteer inside a rotating drum, the internal walls of which are covered with stripes, to create uniform motion of virtually the entire visual field, as shown in Fig. 1(b). The eye movements so elicited are called ‘optokinetic nystagmus’ (OKN): the eyes move smoothly in the direction of drum motion (slow phases) and intermittently flick rapidly back in the opposite direction (fast phases). (See Fig. 1(a).) Another interesting effect of sitting in such a rotating drum is that, after a few seconds, most observers feel that the drum has slowed down and even stopped, and that their body is rotating in the opposite direction. This is a sort of rotational analogue of the common experience, when sitting in a stationary train at the railway station, of mistaking the adjacent train pulling out for the departure of one's own. The rotation illusion is called circular vection and was studied by the nineteenth-century physicist Ernst Mach.
Exploratory movements
When you wish to scrutinize an object in the visual field, it is important to place its image on the fovea, where visual acuity is highest. Rapid saccadic eye movements do this. Even when you try your best to shift your gaze slowly and smoothly, for instance when reading, the eyes actually move in a series of small saccades (see Fig. 2). Saccades have fairly stereotyped time-courses, and the movements themselves last between only 1/40 sec, for the smallest saccades, and about a tenth of a second, for the largest. They are thus some of the fastest movements the body can produce.
Saccades are conjugate (i.e. matched in size in the two eyes) when gaze is transferred between two objects at the same distance from the viewer, but are disconjugate (different in size in the two eyes) when shifting gaze between two objects at different distances as well as in different directions in the field. If you are looking at something and another object of interest appears unexpectedly in your visual field, there is a delay of about quarter of a second before you make a saccade to look at the new object. But if the current target is suddenly removed before the second one appears, the delay in making the saccade to a new target can be as short as one tenth of a second. This suggests that the normal long delay is not entirely due to the time taken to process information about the new object in the brain; part of the delay is the time needed to disengage attention from what one was previously looking at.
Smooth pursuit eye movements are used to follow a moving target that one is looking at directly. For predictable target movements of modest velocity, eye velocity comes close to matching target velocity. Unlike saccades, smooth pursuit cannot easily be initiated voluntarily without a moving target to follow. When a target first starts to move, it shifts off the fovea before the smooth eye movement can start (because there is a delay of about one eighth of a second). Therefore there is often an initial ‘catch-up saccade’ mixed with the beginning of the smooth pursuit.
Dodge noticed and objectively documented the curious fact that patients in an asylum for the mentally ill (suffering from what we would now call schizophrenia) had imperfect smooth pursuit, in that they followed a swinging pendulum bob with a chain of tiny, rapid, saccadic eye movements rather than by continuous motion. Dodge's methods of recording eye movements were cumbersome and probably somewhat painful, but, contrary to what one might have expected, he found subjects in the asylum generally very co-operative — perhaps because they were otherwise so bored!
Vergence eye movements are the disconjugate movements made when shifting gaze from one distance to another. The ‘cross-eyed’ movement needed to look at close objects (for instance, at the end of your nose) is called convergence. The reverse movement is called divergence. Pure vergence movements, uncontaminated by saccades, occur when changing gaze between two targets, both lying straight ahead, but at different distances. These vergence movements are smooth, and approximately five times slower than conjugate saccades of a similar size. Convergence is normally closely associated with accommodation (focusing) of the eyes and constriction of the pupil, forming together the so-called ‘near response’ (because they all prepare the eye for looking at a near object).
Even when one does one's best to fixate a small object, tiny eye movements occur involuntarily. One of the most remarkable of all findings in this field is that the minute instability caused by these miniature eye movements is actually essential for vision. If the retinal image is artificially completely stabilized (for instance by monitoring eye position and using a feedback system to move the visual scene so as to exactly compensate for all eye movements), then vision fades out completely after a few seconds.
Eye muscles
Each eye is rotated by six extraocular muscles, which can achieve rotation about any axis, centred on a point approximately 13 mm behind the cornea. Rotations have awkward mathematical properties — e.g. one rotation followed by another about a different axis does not generally have the same effect as the two rotations applied in the opposite order. A complete description of eye movement needs advanced mathematics. In general the effect of contraction of any one eye muscle depends on the degree of contraction of all the others. This is just one of the complications that the brain must deal with as it programmes eye movements.
Why don't we see the world move when we move our eyes?
We have remarkably little awareness of our eye movements, which is perhaps just as well considering that we make several hundred saccades each minute! One issue that has attracted a great deal of attention is how a stable impression of the world is achieved from a series of ‘snapshots’ of information gained during each saccade. Passive displacement of the retinal image (by tapping the side of the eyeball gently, for example) produces a dramatic sense of movement of the visual environment, whereas the same retinal image displacement brought about by a voluntary saccade has no such consequence: the visual world is stable. It seems likely that the brain makes some sort of comparison between signals from the retina and internal messages from the eye movement control system, and ‘assumes’ that the world is stationary unless there is blatant contradiction between the two.
Very little is seen during saccadic movements themselves. This is mainly because the very rapid slipping of the image smears the detail. However, there is also a genuine decrease in sensitivity to light during a saccade — called saccadic suppression. This can be demonstrated by testing how bright a flash of light must be in order for it to be seen, either when the eye is still or during a saccade. (The flash must be so brief that it is not blurred by the movement.) The loss of sensitivity is very small, and this makes it unlikely that (as some argue) saccadic suppression is actively programmed to help to block signals to the brain during saccades. In any case, the resolution of this issue will not greatly advance the understanding of how the world comes to be perceived as stable, because no amount of alteration of visual sensitivity during saccades can alter the fact that before a saccade the retinal image is in one place and after it is in another.
— Stuart J. Judge
See also accommodation; eyes; gaze; vestibular system; vision.
