Foveate animals have explicit ocular motor requirements for best vision. Images of objects of interest must be brought to the fovea, where visual acuity is highest, and kept there, relatively still, for a long enough period of time so that the brain can interpret what is happening in the visual scene. The map of the visual world upon the retina is, of course, two-dimensional, and hence to move the fovea to its center we rotate the globe around two axes (horizontal and vertical) that are orthogonal. But there is a third degree of freedom that allows for torsion — eye movements that rotate the globe around an axis that is roughly parallel to the line of sight. Torsion neither moves images away from the fovea nor increases, by much, motion of images that are already on the fovea. In this sense, a change in the orientation of the globe produced by torsion has little affect on foveal vision. But there are other reasons to have a mechanism that controls the torsional orientation of the eyes. Changes in torsion will affect the perception of objects with images that lie eccentric to the fovea. Such information from the peripheral retina, for example, contributes to the perception of the position of our head (actually the orbits) with respect to the veridical visual upright. Patients with an acute unilateral loss of labyrinthine function show torsion of the eyes such that the top pole rotates toward the side of the lesion (Curthoys et al., 1991).1 When asked to align the position of a vertical bar in an otherwise dark room to earth-vertical, patients tilt the image of the bar such that its top is rotated toward the side of the lesion (Curthoys and Wade, 1995).
Torsion also becomes important for optimal binocular visual function. The difference in the torsional positions of the two eyes — the angle of cyclovergence — determines the torsional disparity between the images of an object on the two eyes. Torsional disparity contributes to our perception of the shape, orientation, and location of objects in depth relative to the position of our heads. In normal subjects the angle of cyclovergence is tightly controlled (Van Rijn et al., 1994). The absolute torsional position of the eyes may fluctuate, but they do so together. As an example of a clinical disturbance that produces abnormal torsional alignment of the eyes, consider patients with a unilateral paralysis of the superior oblique muscle (Lindblom et al., 1997). In this case the palsied eye is relatively extorted (top pole rotated outward). Images of upright objects may appear slanted, with the top of the object seemingly closer than it really is. There also may be tor-sional diplopia, such that when viewing a horizontal bar, for example, the two images will be slanted with respect to each other, with the apparent intersection of the lines pointing toward the side of the weak, relatively extorted eye.
During rotation of the head around its naso-occipital (roll) axis the eyes must rotate around an axis parallel to that of the head to minimize motion of images in the retinal periphery. If the eyes are directed roughly straight ahead in the orbit, this compensatory motion of the eyes is also along the line of sight, and hence equivalent to eye torsion. Certainly some degree of "torsional slip" is tolerated naturally since even for normal subjects, the amplitude of "compensatory" eye movements in response to roll motion of the head is considerably less than that of the head (Tweed et al., 1994). The amount of compensation for roll movements of the head also varies with the nearness of the object of interest. For near objects the need for optimal foveal function — fine stereopsis and fusion — begin to
1 Actually, with an imbalance (physiological or pathological) in the vestibular inputs that rotate the head around the roll (naso-occipital) axis, the consequent rotation of the eye is a "counterroll" around a head-fixed, nasal-occipital axis. Hence, if the eyes are eccentric in the orbit, the compensatory counterroll will not be a torsional rotation around the line of sight. The caveat, of course, is that one must be clear about which coordinate frame — head fixed or eye fixed - one is using when discussing torsional rotation of the globe.
dominate over the need for stabilization in the retinal periphery (Misslisch et al., 2001).
Was this article helpful?