II. Eye Movements
- i. Ductions
- ii. Phorias and tropias (versions)
- iii. Ocular dynamic function
- 1. fixation
- 2. smooth pursuit
- 3. saccades
- 4. vestibulo-ocular response
- 5. nystagmus
- 6. synkinesis
Examination of the eye movements begins with observing the patient at rest. Head posture may be abnormal with either tilt or turn. These can be either pathologic (i.e. head tilt in the ocular tilt reaction) or compensatory (i.e. head tilt with IV nerve palsy). Eye alignment is noted as the patient looks straight ahead (primary gaze position). Sizable tropias will often be evident, though more subtle ones need more formal examination. Tendencies to gaze deviation and abnormal spontaneous eye movements should be apparent.
These characterize the movement range of each eye separately, with the other covered. The patient first makes saccades to right, left, up, and down (secondary gaze positions), then to the ‘corners’ for oblique (tertiary gaze) positions – i.e. right and up, left and up, right and down, left and down. Full horizontal excursion is denoted by the ability to hide the sclera behind the canthi. With full downgaze, the lower lid should cross the midsection of the pupil. Full upgaze lacks an objective clinical marker, and requires some experience with normal individuals. Upgaze range is also more restricted by normal aging.
This initial examination of duction range uses saccades to targets held by the examiner. If this is limited, then one must determine whether the limitation affects other eye movements. The patient follows a moving object as far peripherally as possible to test pursuit range. Of most importance is the VOR range. A dissociation between the range of ‘voluntary’ (saccadic and pursuit) eye movements and those generated by the VOR is best demonstrated by first having the patient look as far as possible in the direction of gaze limitation, using either saccadic or pursuit movements. While they fixate target in this position, the patient’s head is then moved smoothly in the opposite direction. With relative preservation of the VOR, the eyes will continue to fix the target and move further than they did with saccades or pursuit. This implies a supranuclear rather than a nuclear or infranuclear disorder.
Supranuclear upgaze palsy can also be confirmed by an intact Bell’s phenomenon. Slight downward adducting movements of the eyes usually accompany blinks (29) but forceful lid closure is accompanied in most by elevation – Bell’s phenomenon. The patient squeezes their eyes shut while the examiner keeps the upper lid open to view eye position.
ii. Phorias and tropias (versions).
The relative alignment of the eyes is then measured. With a tropic deviation, the two eyes are not aligned whether the patient is viewing with one eye or both eyes. Because the deviation persists with both eyes open, patients usually have diplopia. If not, either vision in one eye is poor or the image of one eye is suppressed, which can occur with chronic lesions. With a phoric deviation, the eyes are misaligned when either eye is viewing alone, but when both eyes are viewing the two eyes are aligned. Many normal subjects have a mild phoric deviation in primary position, usually an exophoria.
Orthotropia means binocular alignment with both eyes open, orthophoria means binocular alignment with either monocular or binocular viewing. Horizontal misalignments are either esotropias, in which the eyes are excessively converged or ‘cross-eyed’, or exotropias, in which the eyes are excessively diverged or ‘wall-eyed’. If it is already evident from ductions which eye is abnormal, a left/right prefix can be used. Hence a left VI nerve palsy with limited abduction causes a left esotropia. However, this should not detract from the fact that phorias and tropias describe the position of the two eyes relative to each other. In some cases, particularly with phorias, one cannot state that the abnormality originates in a certain eye, and it is best to omit the left/right designation.
With vertical tropias, a left/right designation is needed to indicate which eye is relatively higher or lower. When possible, these should be named according to the abnormal eye. Thus when the left eye cannot depress the abnormality is a left hypertropia; when the right eye cannot elevate it is a right hypotropia (in both cases the right eye is lower). When there is no abnormal eye, the default designation is to name the hypertropic eye. It is incorrect to name the tropia for which eye appears deviated in straight-ahead gaze: this may reflect interocular differences in vision or ocular preference rather than the side of dysfunction.
Examination for tropias and phorias begins in primary gaze position. The same measurements are then made in secondary gaze positions, and also at times in tertiary gaze positions, mainly if a vertical tropia is found or suspected. There are several techniques, all of which require conscious cooperation of the patient. Subjective techniques, which examine mainly for tropias, also require the patient to comment on their experience of diplopia. These include the red-glass test and the Hess screen. All use a target and colored lenses to distinguish one eye’s image from that of the other. By convention, a red lens is held before the right eye, and the left has either no lens or a green one.
With horizontal tropias, the patient sees a red and white light beside each other. Examination of horizontal tropias is aimed at localizing the weakness among the four different medial and lateral recti. If the red light is on the right, the right image belongs to the right eye – an uncrossed diplopia, indicating an esotropia. This is because the object’s image falls on a more leftward position in the right retina than in the left retina, and is interpreted as being further right in the visual field. Similarly, if the red light is left of the white one, it is a crossed diplopia, indicating an exotropia. The patient is asked to comment on the distance between the two lights as they look left and right. Patients with childhood strabismus and some with chronic neuropathic palsies usually note no change with lateral gaze. However, the diplopia of acute weakness varies with gaze. Paresis is always more evident when the muscle is in use: hence increase in uncrossed diplopia in left gaze indicates weakness of the left lateral rectus, whereas greater crossed diplopia in right gaze indicates left medial rectus weakness. A good rule is that the most lateral image in the direction of greatest image separation belongs to the weak eye.
The horizontal alignment should also be examined in up and down gaze. More convergent (less exotropia or more esotropia) or divergent positions in downgaze compared to upgaze – the so-called V and A patterns, respectively – are common in childhood strabismus. They also occur with some acquired tropias, for example the esotropia associated with bilateral IV nerve palsies. Last, a comparison of diplopia with far and near targets is useful, especially if no deviation has been uncovered at one distance in a patient complaining of diplopia. Esotropias from lateral rectus pareses are worse at far, and if subtle may only be present with a far target, because divergence requires use of these muscles, and whereas exotropias from medial rectus pareses are worse at near. Divergence and convergence insufficiencies have diplopia only at far or near respectively too.
With vertical tropias, one image is on top of the other. The higher image belongs to the lower eye. Examination of vertical tropias is aimed at localizing the weakness among the eight different vertical ocular muscles. The Bielschowsky or Parks three-step test is useful. The first step determines which eye is higher in primary gaze position. If there is a right hypertropia, then either the right depressors (inferior rectus, superior oblique) or the left elevators (superior rectus, inferior oblique) are weak. The second step determines whether the vertical separation is greater in right or left gaze. In left gaze, for example, the most active vertical muscles are the left superior and inferior recti and the right superior and inferior obliques. Because we know from step one that this is a right hypertropia, the combination of the two steps narrows the possibilities to the right superior oblique and left superior rectus. The third step determines whether the separation increases in right head tilt or left head tilt. Head tilt generates a compensatory counter-roll of the eyes. With right head tilt, this is an incyclotorsion of the right eye, by contraction of the superior rectus and superior oblique, and an excyclotorsion of the left eye, by contraction of the inferior rectus and inferior oblique. Greater separation of images in right head tilt therefore indicates a weakness of one of these four muscles. There is only muscle that fits the criteria of all three steps in this example, which is a right superior oblique palsy.
The three-step test is designed primarily to detect paresis of one cyclovertical muscle (30). The classic case is superior oblique palsy, which can be remembered by the ‘left-right-left’ or ‘right-left-right’ combination: i.e. a left hypertropia worse in right gaze and left head tilt. The third step of head tilt may be less impressive with weakness of the vertical recti, because their contribution to torsion is less than that of oblique muscles. In cases where the difference between right and left head tilt is unimpressive, a fourth step is to have the patient look in the horizontal direction where diplopia is maximal, and then compare whether the separation increases looking up or down in that position. The three-step test alone may lead to erroneous diagnosis in cases with restrictive rather than paretic dysfunction, or paresis of multiple cyclovertical muscles (31). Also, the differential diagnosis of skew deviation, myasthenia gravis, and thyroid ophthalmopathy, which are other common causes of vertical diplopia, must always be remembered. Hence the three-step test must be interpreted in the context of the rest of the examination and history.
Many cases of diplopia are oblique, with both vertical and horizontal displacements. Even cases of VI nerve palsies can have a small vertical component. In other cases there may be multiple muscles involved, as with a III nerve palsy. The examination of horizontal and vertical deviations must be conducted separately. Some patients have trouble limiting their attention to the separation in one direction alone, however. In these cases, one can substitute a Maddox rod for the red lens. The Maddox rod contains prisms that convert a point of light into a bar. With the prisms held horizontally, the red point becomes a red vertical line. By comparing the position of the white point with the red line, the horizontal tropia can be characterized. The vertical tropia can then be examined by turning the Maddox rod 90 degrees.
Without a red glass, the subjective determination of which image belongs to which eye can be done by simply covering one eye and asking the patient which of the two images disappear. This is repeated in the different gaze positions, and the patient reports on the change in the distance between the images in these positions.
These subjective tests do not reveal phorias and are difficult in patients who lack diplopia because they are suppressing one image. The cover test has the examiner watching for shifts of ocular fixation instead of asking the patient to report their experience. In the alternate cover test (Figure “alternate cover”, Video “alternate cover”) the right and left eyes are covered alternately while the patient views a small target, without allowing binocular vision. With monocular viewing, the eyes are not aligned if either a tropia or a phoria is present. When the cover is shifted to the other eye, the previously covered eye is not pointing at the target and must make a small saccade to look at it. Meanwhile, the other eye will have drifted off target, so that when the cover is shifted back, it too needs a small saccade to take up the target again. Thus, with an exodeviation, each eye when uncovered makes a small nasal-ward motion. With a right hypertropia, the right eye moves down when exposed and the left up. The size of the saccadic shift correlates with the distance between diplopic images in patients with tropias. Shift size is compared in the different gaze positions.
Both phorias and tropias cause saccadic shifts with the alternate cover test. The single cover test (Figure “single cover”) distinguishes the two. Only one eye is covered, then the cover is removed to permit binocular viewing. With one eye covered, the eyes are misaligned with either a tropia or a phoria. When the cover is removed and binocular viewing is allowed, the eyes will align with a phoria. This is evident as a slow shift of the previously covered eye while the eye never covered remains steadily on target. As the same eye is covered again, it drifts off fixation while the viewing eye remains steady. All four of these observations should be made. If there is a tropia, however, one of two things occur. The suddenly uncovered eye may shift to take up fixation, in which case the never-covered eye shifts off fixation: thus, there is a conjugate motion. With replacement of the cover, there is a conjugate motion in the reverse direction, as the never-covered eye must resume fixation of the target. However, if the patient prefers fixation with the never-covered eye, nothing may happen. The never-covered eye stays on target and the suddenly uncovered eye remains off-target, regardless of the presence of the cover. In this case, switching the single cover test to the other eye will reveal the conjugate motion.
Even so, there may be patients who are happy to maintain constant fixation with whichever eye happens to be the never-covered eye. This frustrating situation can be circumvented by the switching single cover test (Figure “switching cover”). Here one eye is covered, then the cover is removed allowing binocular viewing, then the other eye is covered, followed by binocular viewing again, and several repetitions of the cycle. With a tropia, the move from cover of one eye to binocular viewing may be accompanied by the conjugate shift as the previously covered eye takes up fixation, as above, in which case there will be no further shift when the other eye is covered next. However, if the shift to binocular viewing was not accompanied by the conjugate shift, this will inevitably occur when the other eye is covered.
Essentially, the key difference between phoria and tropia is that the opportunity for binocular fusion generates a mainly monocular vergence motion of the previously covered eye. The lack of fusional ability in a tropia is revealed by a conjugate shift at some point as one eye takes up fixation and the other moves off it.
Subjective and objective determinations of deviations can be quantified by use of prism bars. With the red lenses, the amount of prism needed to eliminate the distance between the images can be determined; with the cover tests, it is the amount needed to eliminate the ocular shifts. Prism measures are useful for following the course and for prescribing prisms to minimize diplopia.
The cover tests are least useful when the patient has additional motility problems like nystagmus and saccadic intrusions that confound the observation of the fixation shifts.
After determining range and alignment, one observes how the eyes move. First is observation of fixation as the patient maintains steady gaze on a distant target with the eyes in primary position. This is best done by occluding one eye, especially in patients with a tropia, to ensure that small movements aren’t due to fixation shifts between the eyes. Saccadic intrusions and nystagmus are noted. Small movements may be noticeable only with fundoscopy while the patient attempts to maintain steady fixation. Nystagmus may not be apparent until fixation is suppressed, either by 20-diopter Frenzl goggles, or by doing fundoscopy of one eye while the other eye is occluded. Saccadic intrusions include those with a brief interval between the saccadic components (square wave jerks, macrosaccadic oscillations – Figures “square wave jerks”, “macrosaccadic oscillations”) and those without (ocular flutter, opsoclonus, voluntary nystagmus – Figures “opsoclonus”, “ocular flutter”). This interval can be appreciated clinically with experience, but may require eye movement recordings for verification. The inability to maintain steady fixation while objects appear in peripheral vision, as during confrontation testing of visual fields, is a form of motor impersistence that is caused by frontal lesions (32).
Smooth pursuit is best assessed as the patient follows an object moving approximately sinusoidally left and right, then up and down. Most people can accurately and smoothly follow an object that takes a second to move from left to right or vice versa (2 seconds to complete one full cycle – 0.5 Hz). Pursuit requires attention. Bored subjects will either lag behind the target or make small saccades to jump ahead of the target, anticipating its destination. In both cases the patient must be urged to concentrate on keeping their gaze locked to the target. If pursuit is perfect, then eye speed matches target speed. ‘Pursuit gain’ is the ratio of eye speed over target speed, which equals one if perfect. Inadequate pursuit is betrayed by the interruption of the smooth following eye movement by small ‘catch-up’ saccades, which result when the eye falls too far behind the target, and the difference between eye and target position triggers a corrective saccade (Video “impaired pursuit”). Poor smooth pursuit is sometimes referred to as ‘saccadic pursuit’. However the saccades are corrective rather than abnormal, and are a sign of ‘decreased smooth pursuit gain’, which is the better term. With more rapid target oscillations and higher velocities, all subjects use a mixture of smooth pursuit and small saccades to track targets. A more unusual abnormality is excessive pursuit gain, in which the smooth pursuit eye speed is greater than the target speed, causing the eye to overtake the target, and generating ‘back-up saccades’ in the opposite direction to target and smooth eye motion (33).
Optokinetic responses (OKR or OKN) are smooth eye movements generated by motion of large portions of the visual field, rather than the small targets that generate pursuit responses. This is usually tested with large striped or textured objects, such as an ‘OKN drum’. AS the stripes moves across the patient’s vision, they are asked to “watch each stripe as it goes by”. This generates a jerk nystagmus with the slow phase in the direction of stripe motion. If the patient cannot generate the slow phase the eyes remain stationary. If the slow phase is generated but there is a defect in generating quick phases and saccades, the eyes deviate smoothly in the direction of stripe motion and do not return to midposition. In humans, it is unusual to see deficits in optokinetic slow phases without abnormalities in smooth pursuit. However, asymmetries in these smooth eye movements are sometimes more easily appreciated with OKR than sinusoidal smooth pursuit.
Normally the VOR keeps the eyes stable in space despite head motion. However, if the head is moving in concert with a moving target, it may be desirable to keep the eye stabilized in the orbit to follow the target. This is VOR-cancellation. It can be tested with the patient in a swivel chair. They place their outstretched arms together and are told to watch their thumbs while they rotate en bloc with the chair. The eye should remain stationary in the orbit. Impairment causes the eye to lag behind the target, with catch-up saccades eventually generated, creating an appearance of nystagmus (Video “impaired VOR cancellation”).
Saccades are quick eye movements that shift gaze from one object to another. The examiner holds up two objects and asked the patient to look at one and then the other. Horizontal and vertical saccades are tested separately. Three different aspects are noted. First, the latency of response is the time between the command to move the eyes and when the saccade begins. Prolonged latency is usually difficult to appreciate clinically, and requires eye movement recordings. However, gross difficulties in initiating saccades can be observed. Some patients cannot generate saccades to command: if so, they should be observed to see whether they can saccade to suddenly appearing objects, such as someone passing in the hallway, or unexpected sounds. This dissociation between command and ‘reflex’ saccades is typical of ocular motor apraxia (34,35).
Second, the accuracy of the saccade is noted. Most large saccades slightly undershoot the target by 10%(36) and are followed by a single secondary saccade that lands the eye on the target. Greater undershoot, sometimes with multiple secondary saccades, indicate a pathologic hypometria (Figure “saccadic hypometria”, video “saccadic hypometria”). Hypermetria is an excessively large saccade to the target, which is betrayed by a corrective secondary saccade in the reverse direction (Figure “myasthenic edrophonium test”, Video “saccadic hypermetria”). If the secondary saccade is also too large, then another corrective saccade is generated in the original direction, creating a saccadic oscillation around the target position. Hypermetria implies a cerebellar disturbance (37), whereas hypometria is less specific. Certain combinations of hypermetria and hypometria can be localizing. Hypermetria in one horizontal direction and hypometria in the other is saccadic lateropulsion (Video “saccadic lateropulsion”). This includes ipsipulsion (hypermetric towards the side of the lesion) with lateral medullary syndromes (38,39) and contrapulsion (hypermetric away from the side of the lesion) with lesions of the brachium conjunctivum in the midbrain, which is associated with ipsilateral ataxia (40). In centripetal hypermetria, saccades away from primary position are hypometric but those back to primary position are hypermetric (41). This indicates a cerebellar failure to adjust saccadic force for the difference in orbital resistance and elasticity encountered by centrifugal and centripetal motions.
Third, saccadic velocity is noted. Slow saccades occur with many disorders, but also in some normal states such as fatigue. Myopathies in particular are characterized by slow saccades. Subtle decreases in velocity may be difficult to appreciate if conjugate, but unilateral slowing can be detected by comparing the two eyes, noting the time at which the saccade ends in one eye versus the other. A lag in which the adducting eye is still seen to be sliding to a halt after the other eye has stopped is an important sign of a subtle internuclear ophthalmoplegia (INO) (42,43).
Other observations of specific saccadic abnormalities can be made too. In myasthenia gravis, saccades sometimes start with good speed, only to fade in the middle of the movement, completing the saccade with a much slower drift to the final position. This is intra-saccadic fatigue(44,45) (Figures “myasthenic sleep test”, “myasthenic edrophonium test”). Saccadic intrusions such as macrosaccadic oscillations (Video “macrosaccadic oscillations”) and ocular flutter are sometimes triggered by saccades (46). Ocular flutter that occurs at the termination of a saccade is termed ‘flutter dysmetria’(47) (Figure “ocular flutter”).
The vestibulo-ocular responses (VOR) are more difficult to observe clinically. Slow head turns may generate compensatory (Doll’s eye) eye movements that reveal these responses when pursuit and saccadic systems are impaired. This maneuver can demonstrate the integrity of nuclear and infranuclear ocular motor structures in comatose patients (48). Normal conscious subjects can also create a normal Doll’s eye response if they are instructed to keep looking straight ahead, but they may use pursuit and OKR to do this if the head turn is slow enough. Rather, rapid head movements are needed to isolate the VOR. Pursuit and OKR are ineffective at high accelerations and frequencies, where the VOR performs well. Head-shaking visual acuity can assess the integrity of the entire VOR system. The patient reads the Snellen chart first with the head still, and then while the examiner rapidly shakes the head at a frequency of 2 Hz or more. Visual acuity should only fall by one to three lines. Greater drop implies deficient VOR gain, though some cases of cerebellar disease have hyperactive VOR (49).
The head thrust maneuver tests each horizontal semicircular canal separately (Video “head thrust test”). As the patient fixates a distant target, the head is rapidly turned to one side, for instance the left. If the left horizontal canal is destroyed, its normal increase in excitation will be absent. While a decrease in the spontaneous firing rate of the right canal will still give information about slow left head turns, a fast turn will drive the right firing rate down to zero, so that the rate no longer correlates with head velocity. The resulting VOR of the eye to the right will be inadequate, and betrayed by a small rightward saccade to return to the target at the end of the head turn.
Another test for lateralized vestibular damage is post-head shaking nystagmus(50,51). The head is rapidly shaken left and right for 30 seconds, with Frenzl goggles on. During this time information about head velocity from both sides is being stored in the nodulus. If there is unilateral vestibular damage, this stored information will be asymmetric. When the shaking stops, the stored information decays slowly over a few seconds: during this time, asymmetry will be manifest as jerk nystagmus in the direction of the better ear.
The chief observations are first, whether the waveform is pendular (Video “pendular nystagmus”) or jerk (with a fast component in one direction and a slow component in the other). Second, the direction of nystagmus is noted: horizontal, vertical or torsional. Third, nystagmus may be monocular or binocular, and if binocular it may be symmetric or asymmetric. These observations should be made with the eyes in primary position and then in secondary gaze positions (Video “gaze evoked nystagmus”). They can be repeated with Frenzl lens to remove the suppressive effect of vision, which is prominent with peripheral vestibular lesions. Observations during fundoscopy may be the only means of detecting a small amplitude nystagmus. Fixation can also be removed by covering the other eye during fundoscopy. Hyperventilation can accentuate or reveal nystagmus of either peripheral or central origin (52,53).
The effects of body and head position on nystagmus are examined next. The Dix-Hallpike maneuver involves extending the neck and turning the head laterally, then rapidly placing the patient supine with the head hanging below the body. The effects of static positions can also be noted by having the patient lie on one side then the other.
A last point of dynamic function is the search for aberrant movements. These are usually associated with trauma, compression, or congenital nerve palsies. On occasion, there is no history of prior palsy, and in such cases a compressive lesion must be sought. In addition to the various misdirections from one III nerve branch to another described under III nerve palsies, there are occasional descriptions of misdirections between the ocular motor nerves and the trigeminal motor branch, such that involuntary jaw motion is precipitated by eye movements (54).