The Effect of Acute Superior Oblique Palsy on Vertical Pursuit in Monkeys

Abstract

purpose. To investigate vertical smooth pursuit eye movements in monkeys with acute acquired superior oblique palsy (SOP).

methods. The trochlear nerve was severed intracranially in two rhesus monkeys. After surgery, the paretic eye was patched for 6 or 9 days, and then binocular viewing was allowed. Eye movements were measured with binocular, dual search coils, before and after surgery, under monocular viewing conditions. Vertical pursuit movements along the midline were elicited by using triangular-wave (20 deg/s, ±20°) or step-ramp (20 deg/s) stimuli at a distance of 66 cm.

results. During the early post-lesion period, before binocular viewing was allowed, pursuit velocity of the paretic eye during triangular-wave tracking was lower than that of the normal eye. When the viewing eye crossed straight ahead, the changes in pursuit velocity conjugacy were similar for upward and downward tracking. After habitual binocular viewing was allowed, differences between upward and downward pursuit emerged. When measured ∼30 days after lesioning, this directional asymmetry was less during the open-loop period of step-ramp tracking than during triangular-wave tracking.

conclusions. Rhesus monkeys with acute acquired SOP show characteristic changes in vertical pursuit, with deficits for both upward and downward tracking, and differences between the initiation of step-ramp pursuit and the sustained response during triangular-wave tracking. The habitual viewing condition (monocular versus binocular) also affected the pattern of deficit.

Most prior studies of superior oblique palsy (SOP) have focused on the static deviations of the eyes with the head upright or tilted. Although a few studies have reported dynamic changes during eye movements, including saccades, ^{1 2 3 4 5} smooth pursuit, ^{6 7} and the vestibulo-ocular reflex, ⁸ less is known about the acute effects of SOP on eye movement dynamics and how these changes evolve over time. In a recent series of papers, we discussed the effects of acute SOP in monkeys (created by surgical section of the trochlear nerve) on static alignment, saccades, and control of torsion. ^{9 10 11 12} The purpose of this article is to analyze the effects of SOP on vertical pursuit eye movements and to compare these effects with those on other eye movement subtypes in SOP. In this way, we can define further the ocular motor signature of acute SOP in the monkey, which will help provide a frame of reference for interpreting the ocular motor deficits in patients with vertical strabismus. 

Methods

General Experimental Procedures

The details of the experimental procedures have been described in the previous report that was based on the same two monkeys. ⁹ Dual search coils were implanted in each eye for measurement of three-dimensional eye positions using a search coil system with three magnetic fields. Signals from each coil were demodulated by frequency detectors, filtered in hardware with a bandwidth of 0 to 90 Hz, sampled at 1000 Hz, and stored on computer for later analysis. All experimental and surgical procedures, including anesthesia and postoperative analgesia, were performed according to a protocol that was approved by the Institutional Animal Care and Use Committee (IACUC) of The Johns Hopkins University, in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Trochlear Nerve Section and Experimental Protocol

After the animals were trained and the baseline data recorded, the trochlear nerve was sectioned intracranially as described in a previous report. ⁹ The left trochlear nerve was sectioned in M1 and the right trochlear nerve in M2. Before the animal recovered from anesthesia, the paretic eye was covered with an opaque patch to prevent binocular viewing. This patch remained on for 6 days in the case of M1 and 9 days in M2. Then, binocular viewing was allowed for the remainder of the study, except during testing with monocular viewing. Both animals developed vertical (10°–12° in adduction and down) and torsional (10°–12° in abduction and down) phorias (misalignment with one eye viewing) typical of SOP. ⁹  

Animals sat in a primate chair with the head fixed. The target was a small red spot (0.3° × 0.3° square) that was rear projected onto a tangent screen located 66 cm in front of the monkey. The experiment was performed in an otherwise dark room. There were two pursuit paradigms. Triangular-wave pursuit was elicited by a continuously moving, constant-velocity target (0.25 Hz, velocity ±20 deg/s, amplitude ±20°) in the vertical direction along the midline. For step-ramp pursuit, each trial began with fixation of a stationary center target. At a random time, the target stepped up or down (4.0°–5.2°, depending on target speed and the saccade latency of the monkey) and then moved at a constant speed of 20 deg/s in the opposite direction. The target continued to move for approximately 1 second after the eye started to move. The animal was rewarded for maintaining fixation of the target within a window of 4° for triangular-wave pursuit or 7° for step-ramp pursuit. Viewing was always monocular during the recording session. 

Data Analysis

Data analysis was performed using custom software developed in a commercial program (MatLab; The Mathworks, Natick, MA). Raw eye coil signals were converted to rotation vectors and angular velocity vectors, as described previously. ¹³ Because rotation vectors are analyzed using the right-hand rule, we inverted the signs of the horizontal and vertical components to make them consistent with the clinical definition of eye movement directions. Thus, in the following results, rightward, upward rotations of the eye, from the perspective of the animal, are positive. To compare the results more easily between the two monkeys, the data were mirrored between the two eyes for M2, who had a right SOP. Thus, for both monkeys the left eye was depicted as if it was paretic and the right eye as if it was normal. 

An interactive program was used to select periods of smooth tracking for analysis. First, all eye positions within ±12.5° were extracted along each ramp. Periods of tracking contaminated by saccades and blinks were removed by visual inspection. For triangular-wave tracking, the ratio of the paretic eye velocity to the normal eye velocity (pursuit velocity ratio), averaged over 1° intervals, was calculated for 25 equally spaced viewing eye positions over the central 25° (±12.5°). For step-ramp pursuit, the onset of pursuit was taken when smooth eye velocity reached 2 deg/s in the direction of the target ramp. ¹⁴ Our analyses of step-ramp tracking focused on the open-loop interval during the first 80 ms of pursuit that occurs before visual feedback is possible. In this way, we were able to analyze the response of the pursuit system to the muscle palsy before visual feedback could be used to improve tracking and obscure any deficit. The open-loop response was characterized as the average acceleration of the eye over the first 80 ms of tracking. Statistical comparisons were performed with one- or two-way ANOVA. 

Results

General Vertical Pursuit Behavior after SOP

Figure 1shows a representative example with normal eye (NE) viewing from M2 of tracking of a target moving vertically at 20 deg/s in a triangular-wave trajectory, before and after the lesion (post-lesion day 8 during habitual NE viewing). Before the lesion (left) the profiles of eye position and eye velocity during pursuit were normal in both eyes. After the lesion (right, top trace), the paretic eye (PE) was higher than the NE (as was the case during steady fixation), and this misalignment was greatest in down gaze. Pursuit velocity of the PE was reduced for both upward and downward movements (right, bottom trace, note the increase in catch-up saccades and the separation of the velocity traces for both directions of tracking after the lesion). There was also a small, ∼3° to 4° horizontal, exodeviation (the eyes are relatively diverged) after the lesion. The pattern of change in pursuit was qualitatively similar in M1, though before lesioning, this monkey showed a small up/down asymmetry with a preference for downward tracking. 

Changes in Steady State Pursuit Velocity

As shown in Figure 1 , right, post-lesion pursuit eye movements by the PE were slower than those made by the NE. To examine the conjugacy of the movements, we calculated a pursuit velocity ratio between the PE and the NE, which allowed us to examine the effect of the lesion without contamination from any spontaneous variability in pursuit velocity commands. In Figure 2 , the velocity ratios of pursuit were plotted as a function of the NE position in the orbit when the NE was viewing, before and after surgery. Before the lesion, in both monkeys, the ratios were close to 1.0 and were little affected by vertical eye position. During the initial period after the lesion, while the PE was habitually patched, both animals showed a similar pattern of disconjugacy, with lower velocities by the PE for both directions. For upward tracking, the velocity ratios decreased as the eyes moved upward (Fig. 2) . For downward tracking, the velocity ratios were stable over the central 10° but increased gradually in both the upper and lower positions (Fig. 2) . 

The change over time of the velocity ratios, measured as the viewing eye crossed 0° for both NE and PE viewing, is shown in Figure 3 . In the immediate period after the SOP was created, but before binocular viewing was allowed (post1), the velocity ratios dropped for both downward and upward pursuit to about the same degree. In each monkey, statistical analysis of the velocity ratios with a two-way ANOVA showed a significant difference between before and after lesioning (P < 0.001 for M1 and M2), but no difference between upward and downward pursuit (P = 0.47 for M1, P = 0.19 for M2). Once binocular viewing was allowed (Fig. 3 , top, post2 and post3), a considerable difference between upward and downward pursuit emerged over time, with a further decrease in the ratio during downward pursuit in both monkeys (P < 0.001). Approximately 30 days after lesioning with NE viewing, the velocity ratio for upward pursuit was 16% and 18% higher than that for downward pursuit in M1 and M2, respectively. 

When tested with the PE viewing (Fig. 3 , bottom, post2 and post3), the velocity ratio decreased further over time in the downward direction for M1 (P < 0.001), but increased toward normal in the upward direction for M2 (P < 0.001). Overall, the velocity ratio with PE viewing was higher (closer to normal) than with NE viewing (P < 0.02 for both upward and downward) in M2 (who preferred to fix with her PE). In contrast, in M1 (who preferred to fix with her NE), the velocity ratio with PE viewing was lower than that with NE viewing for upward pursuit (P < 0.02). There was no significant difference for downward pursuit (P > 0.1). 

We next examined the relationship between the pursuit velocity ratios and static phorias, ⁹ measuring with the NE at straight-ahead gaze in both monkeys. Figure 4depicts velocity ratios in the downward direction as a function of vertical deviations before and after SOP. In both monkeys, the velocity ratios showed a significant correlation with static phorias (r = −0.73 for M1, r = −0.96 for M2). 

Changes in the Initiation of Pursuit

To examine the open-loop, initial period of pursuit, before visual feedback could be used to improve the pursuit response, we used the step-ramp stimulus, both to measure the first 80 ms of tracking and to eliminate the early catch-up saccades that usually occur during the onset of tracking. Figure 5shows the pre- and post-lesion responses for M2 for downward tracking at 20 deg/s. Responses were aligned on the target step. After the lesion, the PE was higher than the NE, and the pursuit acceleration and velocity of the PE were reduced. 

Acceleration ratios between the PE and the NE in the open-loop period (0- to 80-ms interval) are shown in Figure 6 . Before lesioning with NE viewing, there was no difference in the acceleration ratio between upward and downward direction (P > 0.5 for M1 and M2). After lesioning (about the same time as post3 data in Fig 3 ) with NE viewing, the ratio had decreased significantly in both monkeys (P < 0.001), slightly more for downward than for upward tracking. The ratio for upward pursuit was 11% higher than that for downward pursuit in M1 (P < 0.001) and 6% in M2 (P = 0.046). This asymmetry, however, was less than that during steady state tracking. Note that for downward pursuit during both NE viewing and PE viewing, the decrease in the acceleration ratio for the open-loop period was less than in the velocity ratio during steady state tracking, especially in M2 (26% vs. 36%, NE viewing; 12% vs. 27%, PE viewing, respectively). 

Discussion

The main result of this study is that rhesus monkeys with induced SOP showed bidirectional and orbital position-dependent changes in vertical pursuit. Tracking was affected toward targets moving both in and opposite the direction of the action of the paralyzed muscle. Immediately after the lesion, during the period when the paretic eye (PE) was habitually patched, pursuit became disconjugate, and the degree, as reflected in the pursuit velocity ratio, depended on vertical eye position. When the viewing eye crossed straight ahead, the changes in velocity ratio were similar for upward and downward tracking. After habitual binocular viewing was allowed, however, direction-dependent differences between upward and downward pursuit emerged with a further relative decrease for downward tracking. There were also differences between the open-loop portion of step-ramp tracking and the sustained response during triangular-wave tracking. We will discuss these findings in turn. 

Steady State Tracking

After surgery, with the normal eye (NE) viewing, vertical pursuit by the PE became slower, as measured in the changes in the velocity ratios that reflected the degree of pursuit disconjugacy. During the immediate post-lesion period, when the PE of the monkey was habitually patched, both animals showed a reduction of velocity ratios for both upward and downward pursuit (up to 19% in M1 and 28% in M2 at straight ahead). Because the secondary action of the superior oblique muscle (SOM) is infraduction, a decrease of the velocity ratio in the contraction direction (downward) of the SOM is expected after the SOP. Seemingly paradoxical, however, was a similar decrease in the velocity ratio during relaxation (upward) of SOM. There are several reasons that may account for the abnormality in upward tracking. The contribution of the SOM to tracking not only consists of an increase in active force in the agonist during downward tracking but also a decrease in active force when it is the antagonist muscle (during upward tracking). When the paretic muscle is the antagonist, this decrement in active force is missing, so the contribution to the forces that rotate the eye upward is less. Another explanation, not necessarily mutually exclusive, is that adaptive mechanisms contribute to the decrease in upward tracking when the paretic muscle is acting as an antagonist, to prevent the misalignment from increasing as the eyes move upward in the orbit. 

We also found that the pursuit velocity ratio depended on vertical eye position. Since the relative contribution of contraction of the agonist and relaxation of the antagonist varies with orbital position, a paresis of SOM in an agonist-antagonist pair also leads to an orbital-position-dependent imbalance. During downward tracking, we found that the velocity ratios were stable over the central positions (implying a relatively constant effect of the loss of agonist forces on tracking), but in up-gaze positions, the velocity ratios decreased gradually as the eye moved downward. This result is compatible with an increase in the relative contribution of the SOM to rotation of the globe downward as it moves into infraduction. There was, however, also a slight increase in the velocity ratio as the eyes moved far down in the lower positions, suggesting that inferior rectus forces become more important here, lessening the need for contribution of the (paralyzed) SOM. During upward tracking, however, the velocity ratios decreased (moved farther from 1.0) as the eyes moved up from eccentric downward positions. This seems counterintuitive, but has to be the case, because the vertical phorias decreased as the eyes move upward. 

After binocular viewing was allowed, both animals showed a large, further decrease of velocity ratios in the downward direction when measured at the end of the 30-day post-lesion period. The reason for this further increase in disconjugacy for downward pursuit is uncertain, but it may reflect the same mechanisms associated with increased static misalignment after the restoration of habitual both eyes viewing. ⁹ Schor et al. ¹⁵ studied the interactions between static vertical phoria adaptation and nonconjugate adaptation of vertical pursuits in normal humans. They suggested that vertical phoria and nonconjugate pursuit adaptation share a common mechanism. Our results are compatible with this idea. Over the 30-day period of study, we found a strong correlation between the pursuit velocity ratios and static phorias in both monkeys. Whether these late changes are due to central adaptive mechanisms or to changes in the mechanics of the denervated and lengthened SOM (or other vertical muscles) or both, remains to be shown. 

When viewing with the PE, the two monkeys showed a different pattern of change in the velocity ratios over time. M1 had a further decrease in the downward direction, but M2 had an increase in the upward direction. When compared with the results with NE viewing, the velocity ratio with the PE viewing was significantly higher for both directions in M2 but not in M1. The reason for these differences could be that two monkeys adopted different patterns of fixation during habitual binocular viewing. M2 preferred to fix with her paretic eye and M1 with her normal eye. If the paretic eye was preferentially used for fixating, some degree of conjugate adaptation would likely be invoked in the attempt to optimize tracking for the paretic eye. ¹⁶  

Open-Loop Response

When viewed with the NE, the effect of the SOP on the open-loop period of step-ramp pursuit was qualitatively similar to that seen with triangular-wave pursuit. At the end of the 30-day post-lesion period, there was a reduction in the acceleration ratio, more so for downward than upward tracking. The asymmetry during open-loop pursuit, however, was less than that during steady state tracking in both animals, especially for M2. In addition, pursuit in the open-loop period was less disconjugate than during sustained tracking. These differences suggest that the adaptive responses to SOP—be they disconjugate or conjugate (especially in M2)—affect open-loop and sustained pursuit tracking differently. 

Comparison with Studies in Patients

There has been one prior study of the open-loop response of smooth pursuit in patients with abducens nerve palsy before and after 1 week of habitual monocular experience with the paretic eye. ¹⁶ Only horizontal step-ramp pursuit was studied, and it was found that the acceleration of the normal eye during the initial period of tracking increased considerably. Because only the normal eye was measured in their study, these adaptive changes cannot be compared easily with our results (ratios between the PE and the NE), although even in this study the changes for tracking in and opposite the direction of the paralysis were almost symmetric. 

Tegetmeyer et al. ^{6 7} recently quantified the amplitude and gain of smooth pursuit in patients with ocular muscle weakness, though the duration of palsy was at least 3 months at the time the patients were tested. They found that, in monocular viewing conditions, the gain of pursuit of the covered paretic eye decreased for both tracking directions. A similar symmetric decrease in the gain of the vestibulo-ocular reflex has also been reported in patients with SOP. ⁸ In contrast, we observed changes to be symmetrical in our monkeys only during the immediate post-lesion period when the paretic eye was habitually patched. Changes became asymmetrical after habitual binocular viewing was allowed. The reasons for the difference remain to be shown, but could include factors such as differences in site and size of the lesion among studies, and long-term effects related to central adaptive processes and the mechanical consequences of a sustained ocular deviation. 

 Supported by National Eye Institute Grants R01-EY001849 (DSZ), K12EY015025 (HSY), and K23 EY00400 (MFW); an Abe Pollin Scholarship (MFW); the Arnold-Chiari Foundation (MFW); and the Albert Pennick Fund (MFW).

 Submitted for publication January 7, 2008; revised April 9, 2008; accepted June 23, 2008.

 Disclosure: J. Tian, None; X. Shan, None; H.S. Ying, None; M.F. Walker, None; R.J. Tamargo, None; D.S. Zee, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Pursuit position and velocity traces with NE viewing in M2, before and 8 days after SOP during habitual monocular viewing. The target moved vertically along the midline in a triangular-wave trajectory (0.25 Hz, ±20°). In the eye velocity trace, large rapid deflections were caused by saccades. These have been clipped for clarity. Positive positions and velocities are upward and rightward. Note the separation of the velocity traces of the PE and the NE after surgery.

Figure 2.

 

Pursuit velocity ratios (PE/NE velocity) as a function of vertical position of the viewing eye during NE viewing (NEV) before and after SOP. Post-lesion data were taken at 4 (M1) and 8 (M2) days after lesioning (while the PE was patched). Positive positions are for up and negative for down. Error bars, SD.

Figure 3.

 

Pursuit velocity ratios (PE/NE velocity) as the viewing eye crossed 0° during tracking of a target moving in a triangular-wave trajectory. Error bars indicate one SD. NEV, NE viewing (top); PEV, PE viewing (bottom). Post1 data were taken at 4 and 8 days while patched, post2 data on the day (6 and 9) when the patch was removed (about 4 hours after binocular viewing was allowed), and post3 data at 29 and 27 days after the lesion, for M1 and M2, respectively. *Post1 values that were significantly decreased compared with before lesioning; ns, no difference between upward and downward (two-way ANOVA); #post3 values that were significantly decreased compared with post2; +, post3 values that were significantly increased compared with post2 (one-way ANOVA). The most striking change was the increase in the difference between the velocity ratios for upward and for downward tracking over time.

Figure 4.

 

Pursuit velocity ratios in the downward direction as a function of vertical deviations with the NE viewing at straight ahead before and after SOP. Data are fit by a linear regression. Data points are taken from multiple days after the SOP.

Figure 5.

 

Step-ramp pursuit position and velocity traces with NE viewing for M2, before and 25 days after the SOP. After fixation with the viewing eye at a straight-ahead target, the target stepped to an up position (4.4°) and started moving downward at a constant velocity (20 deg/s). Ten individual trials are overlaid.

Figure 6.

 

Acceleration ratios in the open-loop period for M1 and M2 in the first 80 ms of step-ramp tracking. Error bars, SD. NEV, NE viewing (top); PEV, PE viewing (bottom). Post-lesion data were taken at 24 (M1) and 25 (M2) days after the lesion, which is at about the same time as post3 data in Figure 3 . The differences between pre and post and upward and downward tracking were statistically significant (one-way ANOVA).

The authors thank Corena Bridges, Dale Roberts, and Adrian Lasker for technical support.