|Home | About | Journals | Submit | Contact Us | Français|
Saccades are fast eye movements that conjugately shift the point of fixation between distant features of interest in the visual environment. Several disorders, affecting sites from brainstem to extraocular muscle, may cause horizontal saccades to become disconjugate. Prior techniques for detection of saccadic disconjugacy, especially in internuclear ophthalmoparesis (INO), have compared only one point in abducting vs adducting saccades, such as peak velocity.
We applied a phase-plane technique that compared each eye’s velocity as a function of change in position (normalized displacement) in 22 patients with disease variously affecting the brainstem reticular formation, the abducens nucleus, the medial longitudinal fasciculus, the oculomotor nerve, the abducens nerve, the neuromuscular junction, or the extraocular muscles; 10 age-matched subjects served as controls.
We found three different patterns of disconjugacy throughout the course of horizontal saccades: early abnormal velocity disconjugacy during the first 10% of the displacement in patients with INO, oculomotor or abducens nerve palsy, and advanced extraocular muscle disease; late disconjugacy in patients with disease affecting the neuromuscular junction; and variable middle-course disconjugacy in patients with pontine lesions. When normal subjects made disconjugate saccades between two targets aligned on one eye, the initial part of the movement remained conjugate.
Along with conventional measures of saccades, such as peak velocity, phase planes provide a useful tool to determine the site, extent, and pathogenesis of disconjugacy. We hypothesize that the pale global extraocular muscle fibers, which drive the high-acceleration component of saccades, receive a neural command that ensures initial ocular conjugacy.
Saccades are fast eye movements that point the fovea of the retina at features of interest in our visual world.1,2 When we look between visual targets located at optical infinity, saccades are approximately conjugate—the eyes move together. Clinicians often examine the conjugacy of horizontal saccades when, for example, they test for internuclear ophthalmoparesis (INO) in patients with suspected multiple sclerosis (MS).3
Much is known about the brainstem mechanisms that generate conjugate horizontal saccades (figure 1). Premotor burst neurons, which lie in the paramedian pontine reticular formation (PPRF),4 generate a high-frequency burst of action potentials.5 Burst neurons are tonically inhibited by omnipause neurons located in the nucleus raphe interpositus, at the pontine midline, between the rootlets of the abducens nerves.6 When the need for a saccade arises, omnipause neurons cease to discharge and thereby promote the vigorous firing of burst neurons in the PPRF to generate the saccadic command.2,7 This pulse of innervation projects monosynaptically to the abducens nucleus, wherein lie two populations of neurons: abducens motoneurons and abducens internuclear neurons.1,3 The pulse of innervation travels from the abducens motoneurons on the abducens nerve to the ipsilateral lateral rectus muscle, which contracts to overcome the viscous resistance of the orbital tissues, accelerating the eye to high speeds. The abducens internuclear neurons also convey the pulse of innervation, via the medial longitudinal fasciculus (MLF), to medial rectus motoneurons in the oculomotor nucleus that, in turn, cause the contralateral medial rectus muscle to contract rapidly. Pale fibers in the global layers of the lateral and medial rectus muscles may play an important role in the initial, high-acceleration component of saccades.8–10
In normal subjects, the eyes turn quickly together as a “conjugate saccade.” Precise measurements of horizontal saccades show small amounts of dynamic disconjugacy.11 A disconjugacy arising within 5 msec of saccade onset can be accounted for by a 1-msec delay of onset of adduction, probably due to transmission in the extra neuron (abducens internuclear neuron and its axon in the MLF).12 A later disconjugacy, evident as greater peak velocity in the abducting eye, has been attributed to greater net forces due to the muscles and orbital tissues13; this causes a transient divergence.11 Nonetheless, the initial portion of horizontal saccades, corresponding to the saccadic pulse, is machine-like in normal subjects, and it is possible to define normal limits for several measures relating the speed of the abducting and adducting eyes.1
Disease acting at a variety of sites in this simple circuit can cause horizontal saccades to become disconjugate. The most well-recognized syndrome is INO, wherein slowing of the adducting eye is caused by inability of the MLF to conduct high-frequency signals.3,14,15 However, disease affecting the ocular motor nerves, the neuromuscular junction, or the extraocular muscles could also cause saccades to become disconjugate. For example, myasthenia gravis (MG) can imitate INO—“pseudo-INO.”16,17
Previous approaches to the study of conjugacy of horizontal saccades have compared the movement of abducting and adducting eyes, measuring, for example, the ratio of peak velocity or acceleration of each eye,18–20 or the ratio of the abducting to the adducting eye displacement when the abducting eye first reaches the position of an eccentric visual target.21 However, these methods measure only one point in the saccade and do not take into account the times of onset and the subsequent course of the saccade in each eye, for which normals show minor differences.1,11,13
Another approach to analyzing dynamic features of motor systems is that of phase planes, which plot velocity as a function of change of position (displacement) for the entire movement, and thereby remove time and the effects of delays (latency). An example illustrating how phase-plane plots were derived from conventional time plots (figure 2) is shown in figure 3. Phase planes have been successfully used in eye movement research, including saccades,13 and forms of nystagmus,22 for which they have provided a number of insights. Thus, the goal of this study was to determine whether systematic application of binocular phase-plane analysis could detect dynamic abnormalities of saccades that were specific for diseases localized to the sites indicated in the circuit summarized in figure 1. We found that conjugate phase-plane analysis not only provided information about the likely site of lesion responsible for saccadic disconjugacy, but also supported the hypothesis that pale global “fast” fibers of the extraocular muscles initiate a conjugate high-acceleration component of saccades.17 Preliminary findings have been published as a short communication.23
We studied a group of 22 patients (age range 22–70 years, mean 55.8 years; 4 female) with abnormal horizontal saccades who had a range of disorders: saccadic palsy after cardiac surgery (2),24 horizontal saccadic gaze palsy from brainstem stroke (2), presumed abducens nucleus infarction (1),25 INO due to MS (7), INO due to brainstem stroke (2), oculomotor nerve palsy (1), abducens nerve palsy (2), myasthenia gravis (3), and chronic progressive external ophthalmoplegia (CPEO; 2). Patients and their clinical features are summarized in the table. We also studied 10 healthy control subjects (age range 30–60 years, mean 40.7 years; 3 female). All patients and control subjects gave informed written consent in accordance with the Declaration of Helsinki and the Institutional Review Board of the Cleveland Veterans Affairs Medical Center.
Patients and control subjects made horizontal saccades in response to 5- to 40-degree jumps of a visual target (laser spot) viewed binocularly on a tangent screen at a distance of 1.2 m. We also induced disjunctive saccades in control subjects by presenting two targets aligned on one (dominant) eye at distances of 15 cm (near) and 1.2 m (far).26 Horizontal and vertical eye positions were recorded using the magnetic search coil technique, which permits precise measurement even in patients who have limited ocular motor range.1 Coil signals were filtered (0–150 Hz) before digitization at 500 Hz; eye velocity and acceleration were computed as previously described.27
Details of the technique we used for constructing phase-plane plots are available in the appendix e-1 on the Neurology® Web site at www.neurology.org and are summarized in figure 3. For each saccade, the displacement (change in position) and velocity of each eye were normalized by assigning a value of 1.0 to the maximum displacement, and to the peak velocity, of the eye making the larger movement (the “strong eye”). Thus, the phase plane plotted the normalized velocity of each eye vs normalized displacement in 1% (0.01) position increments (figure 3C). The position axis of these plots was truncated at the end of movement of the eye making the smaller movement (the “weak eye”). Note that in such phase-plane plots, time is eliminated, and velocity is plotted as a function of eye displacement (position), not time. In this way, we were able to compare the velocity of each eye for the same normalized eye displacement during the entire movement—the velocity disconjugacy plot (figure 3, D and E). We applied this technique to 1,418 saccades from the 10 age-matched normal subjects and used linear regression to define 5% to 95% prediction intervals (PIs; figure 3F). We confirmed that normalized phase-plane analysis was independent of saccade amplitude in normal subjects (figure e-1). Finally, for each patient, the average velocity disconjugacy was calculated from at least 10 saccades and was then compared with the PI of the control subjects.
Patient (P)1 and P2 with post–cardiac surgery horizontal saccadic palsy due to presumed lesion of the brainstem burst and omnipause neurons (figure 1, lesion at sites 1 and 2)28 were found to have variably disconjugate saccades, especially between approximately 20% and 50% of eye displacement; P2 also showed some early disconjugacy. Disconjugacy is evident in a phase-plane plot (figure 4A) as the average velocity difference (velocity disconjugacy) of approximately 10 saccades, and on a representative time plot of a single saccade (figure 2B). Similarly, saccades from one (P4) of two patients with horizontal saccadic palsy due to pontine stroke (figure 1, lesion at site 2) were found to be disconjugate between 20% and 50% of eye displacement (figure 4B). Leftward saccades, from right gaze to center, made by P5 with a presumed acute left abducens nucleus lesion (figure 1, site 3),25 were very slow but did not show abnormal velocity disconjugacy (figure 4B).
Horizontal saccades from all nine patients with INO (P6 and P7 due to brainstem stroke, P8–P14 due to MS; figure 1, lesion at site 4) showed initial velocity disconjugacy, which remained abnormal throughout the whole eye displacement (figure 4C). A time plot of a representative saccade is shown in figure 2C.
Saccades from P15 with oculomotor nerve palsy (figure 1, lesion at site 5) and P16 and P17 with abducens nerve palsy (figure 1, lesion at site 6) were also found to be disconjugate from the onset of the movement (figure 4D). A representative time plot of a single leftward saccade from P17 with left abducens nerve palsy is shown in figure 2D, illustrating slowing of the abducting saccade.
The initial component of horizontal saccades from P18 through P20 with MG (figure 1, lesion at site 7) was found to be similar to that of controls. Thus, note on a representative time plot (figure 2E) how the peak velocities of abducting and adducting eyes are similar. Figure 4E displays the velocity disconjugacy plots from three patients with MG, two with pseudo-INO, and one with right pseudo–abducens nerve palsy. All these patients showed varying degrees of horizontal disconjugacy, but only later in the course of the saccades.
Saccades made by patients with mitochondrial CPEO (figure 1, lesion at site 8) were slow, but still conjugate in P21 with less advanced disease (see representative time plot in figure 2F). On phase-plane plots, P22 with long-standing CPEO showed early abnormal disconjugacy (figure 4F).
When normal subjects were induced to make disconjugate saccades between two targets aligned on the visual axis of one eye, no abnormal velocity disconjugacy could be identified for the first 10% of the eye displacement (figure 3F).
We applied a binocular phase-plane approach to analyze conjugate, horizontal saccades. Our goal was to determine whether lesions at specific sites in the pathway from the brainstem saccade generator to extraocular muscles caused specific patterns of disconjugacy. We found that disease affecting the MLF and the oculomotor or abducens nerve caused the velocity difference between abducting and adducting eyes to increase more than that of controls within the first 10% of the eye displacement. Conversely, disease affecting the neuromuscular junction and extraocular muscles only caused abnormally different velocities of the two eyes when eye displacement exceeded 10%. Thus, on the one hand, disease affecting the MLF (INO) or the oculomotor/abducens nerve as well as advanced CPEO caused substantial velocity difference between the eyes from the beginning of the saccade. On the other hand, saccades made by patients with myasthenia gravis and early CPEO did not show velocity differences between the eyes at the onset of the movement, although disconjugacy sometimes occurred with 20% or more of the eye displacement. In all these disorders—INO, oculomotor/abducens nerve palsy, MG, and CPEO—once velocity disconjugacy developed, it persisted throughout the course of the saccade. Conversely, patients with disease affecting the premotor saccadic-generator circuits of the PPRF showed variable velocity maximum disconjugacy within approximately 20% of eye displacement, which tended to decrease beyond 50% of displacement (figure 4, A and B).
In patients who develop selective saccadic gaze palsy after cardiac surgery, the clinical picture is presumed to be due to malfunction of both omnipause and burst neurons,24 a view supported by the finding of focal neuronal necrosis at the level of median and paramedian pons at autopsy.28 In our two patients, both horizontal and vertical saccades were slow and hypometric, suggesting involvement of the omnipause neurons, although burst neurons may also have been affected.24 In both patients, phase-plane analysis revealed horizontal saccades to be disconjugate (figure 4A). On the one hand, the pathologic process might not uniformly affect the saccadic burst neurons, so that projections to abducens motoneurons and internuclear neurons were affected to different extents; this might cause early disconjugacy, evident in P2. On the other hand, because both patients were studied more than a year after the onset of their saccadic palsy, compensatory mechanisms involving vergence eye movements might be contributing to the gaze shift.24 Phase-plane analysis seems to support this latter hypothesis, because our patients exhibited most prominent ocular disconjugacy at approximately 20% to 30% of eye displacement. On a representative time plot, transient vergence of the eyes is evident during the middle portion of the saccade (figure 2B). The same mechanisms may also account for the findings in one of the two patients with horizontal saccadic gaze palsy due to brainstem stroke, who was studied 20 days after the onset of symptoms and exhibited abnormal velocity disconjugacy between 20% and 50% of eye displacement (P4, figure 4B). In contrast, a patient with a presumed acute lesion affecting her left abducens nucleus25 showed conjugate saccades on phase planes, indicating commensurate involvement of abducens motoneurons and abducens internuclear neurons (figure 4B). In her case, slow leftward movements from right gaze back to the midline may simply have reflected viscoelastic forces applied by the orbital tissues after the sustained (step) innervation ended. Because the record was made during the acute phase of the illness, she may not have had time to develop compensatory mechanisms to induce gaze shifts. Because phase-plane plots were most variable in this small group of patients, more studies are required to clarify how much these or other mechanisms contribute to saccadic disconjugacy after brainstem lesions.
Horizontal saccades of patients with INO or oculomotor or abducens nerve palsy consistently showed large velocity disconjugacy in the first 10% of eye displacement (figure 4C), which is expected with either loss or low-pass filtering of the high-frequency saccadic pulse command.15 This is in contrast to the effects of MG causing the appearance of either pseudo-INO or pseudo–abducens nerve palsy, for which saccades showed similar velocities for the initial part of eye displacement (figures 2E and 4E). Similarly, horizontal saccades of patients with CPEO may be conjugate during the first part of the movements (P21), unless the disease is advanced (P22), when factors such as fibrosis of the extraocular muscles may cause saccades to become disconjugate (figure 4F).
Some insights into factors influencing conjugacy of the initial component of saccades may be provided by results from our normal subjects as they made disjunctive gaze shifts between two targets aligned on one eye. In this case, phase-plane plots showed velocity differences between the two eyes exceeding the normal range for conjugate saccades only for displacements greater than 10% of the total (figure 3F). Taken with an earlier study,26 this suggests that the mechanism that generates the initial part of the saccade in these disjunctive gaze shifts is, in fact, conjugate. What could be the mechanism? Electromyographic studies of the extraocular muscles demonstrated a “division of labor” such that global fibers were more active during saccades and less active at the end of the saccade, when gaze was held steady.8 During such gaze holding, the orbital muscle layers were active. More recently, global layers have been shown to contain pale fibers, which seem well suited to rapid contraction.9 Furthermore, only these pale, global, fast fibers possess well-folded postjunction folds, making them less susceptible to fatigue in MG, so that saccade speed is preserved in the face of restricted range of motion.17 It has also been proposed that pale fibers are relatively spared in CPEO due to mitochondrial disorders, because the pale global fibers have relatively few mitochondria.29 Our current finding of relatively preserved saccadic velocity conjugacy in P21 with early CPEO (figure 4F) supports this proposal.
Taking this evidence together, it seems possible that a pathway from pre–motor burst neurons in the brainstem projecting to the pale global extraocular fibers ensures that the high-acceleration part of saccades is conjugate (Hering’s law of equal innervation). Thus, burst neurons in the PPRF may project to a subset of abducens motoneurons and internuclear neurons that innervate pale global fibers so as to produce the initial, conjugate portion of the saccade. Such a proposal is consistent with recent studies using rabies toxin as an anatomic tracer, which have demonstrated how subtypes of extraocular muscle fibers may be innervated by separate groups of motoneurons.10
Our binocular phase-plane approach promises to be a clinically useful technique in disorders that cause disjunctive movements of the eyes. In particular, we have shown it to be a sensitive method to detect even subtle disorders, such as mild INO, that are difficult to diagnose at the bedside,30 and to distinguish between central and peripheral disorders of ocular motility (i.e., true INO vs pseudo-INO) that can present with a similar clinical picture.17 Formal comparison with other methods that have been proposed to detect early INO using a single measurement of abducting and adducting eye movements18–21 seems indicated. More important, phase planes can provide useful information on the pathogenesis of various patterns of horizontal saccadic disconjugacy. For example, this approach may prove a useful research tool to tease out features of saccadic palsy due to brainstem lesions such a stroke. Thus, development of disconjugacy during recovery from brainstem stroke may provide insights into adaptive mechanisms, such as substitution of vergence for saccades. Finally, application of the binocular phase-plane approach to studies that correlate burst neuron activity to saccadic dynamics might help resolve a long-standing controversy regarding whether disjunctive gaze shifts are achieved by a superposition of saccadic and vergence commands (Hering’s law) or by programming of disjunctive saccades in the brainstem PPRF.31–33
The authors thank Dr. Robert Tomsak for referring patients, Dr. David Zee and Neil Miller for making available data from the patient with presumed abducens nucleus lesion, and Drs. David Zee and Matthew Thurtell for helpful discussions. A.S. thanks Professors Giulio Rosati, Isidoro Aiello, Stefano Sotgiu, GianPietro Sechi, Speranza Desole, and Bianca Marchetti for their support.
Address correspondence and reprint requests to Dr. R. John Leigh, Department of Neurology, 11100 Euclid Ave., Cleveland, OH 44106-5040 ude.esac@4ljr
Supplemental data at www.neurology.org
Supported by NIH grant EY06717, the Office of Research and Development, Medical Research Service, Department of Veterans Affairs, and the Evenor Armington Fund.
Disclosure: The authors report no disclosures.
Received May 12, 2008. Accepted in final form June 27, 2008.