Maldevelopment of Convergence Eye Movements in Macaque Monkeys with Small- and Large-Angle Infantile Esotropia

Abstract

purpose. To describe symmetric convergence eye movements evoked by disparity and/or accommodative cues in esotropic macaque monkeys, with the goal of determining whether these animals have the vergence deficits found in humans with esotropia.

methods. Physical far and near targets were used to evoke large (∼8°) symmetric convergence eye movements in four adult macaque monkeys (two with strabismus, two normal), using positive-feedback rewards. One strabismic monkey had infantile-onset, small-angle esotropia (small-eso∼2°) induced by alternating occlusion from birth to age 9 months. The other strabismic monkey had naturally occurring, large-angle (∼25°) infantile-onset esotropia (large-eso). Visual acuity was normal in each eye as measured by spatial sweep visually evoked potentials (VEPs). Eye movements were recorded using magnetic search coils.

results. When viewing binocularly, both normal monkeys exhibited accurate, stereotyped symmetric convergence movements that achieved 87% to 96% of the required change in vergence angle by the end of the initial movement. In contrast, the small-eso monkey’s convergence response when viewing binocularly was variable, strikingly asymmetric, usually accompanied by a disjunctive saccade, and subnormal, achieving only 56% of required vergence. The convergence response of the large-eso monkey was also asymmetric and weak, achieving 18% of the required vergence and employing conjugate saccades to refixate the near target. Monocular viewing (i.e., accommodative vergence) caused substantial reductions in both convergence amplitudes and velocities in the normal monkeys, but had a minor effect on the vergence behavior of the strabismic animals.

conclusions. Monkeys with small- and large-angle infantile esotropia have striking maldevelopments of binocular (disparity-driven) convergence and use accommodative vergence and saccades to refixate near targets. Their vergence behavior resembles that in esotropic humans. The maldevelopment may be explained in large part by the paucity of binocular connections recently described in the visual cortex of esotropic macaques.

To shift gaze from a distant to a near target and maintain binocular fusion, the eyes must rotate horizontally in opposite directions (disjunctive eye movements), converging the visual axes of the eyes to align on the near target. Eye movement recordings in normal humans and monkeys have verified that the two major visual cues guiding vergence are image displacement on the two retinas (binocular disparity) and image defocusing (accommodative blur). ^{1 2 3} Vergence, accommodation, and disparity sensitivity are all immature at birth, ^{4 5 6 7 8 9} but develop rapidly in the first months of life to achieve adultlike levels of performance. If the sensory and/or motor pathways that drive vergence fail to develop properly, the visual axes become misaligned (developmental strabismus). Esotropia (convergent misalignment) represents more than 90% of all strabismus that appears in human infancy. ^{10 11} It also occurs in monkeys, either naturally ^{12 13} or as the result of experimental manipulations that perturb binocular input in the first weeks of life. ^{14 15 16}  

Many of the eye movement deficits of esotropic macaques are similar to those described in humans. In particular, directional deficits of pursuit eye movement and ocular fixation suggest that these animal are appropriate models for investigating the neural mechanisms of human strabismus. ^{17 18} Studies of esotropic macaques have shown, using photographic methods ¹⁹ or eye movement recording, ^{18 20} that the vergence error resembles that found in humans, in that it is uniform in different gaze positions (i.e., concomitant or nonparalytic). However, the dynamic properties of vergence in these monkeys are unknown. Eye movement recordings in esotropic humans have revealed major deficits of disparity-induced vergence, with preservation of accommodative vergence. ^{21 22} The human esotropes also show an abnormally high prevalence of disjunctive saccades during symmetric vergence trials. ^{21 23} The purpose of the present study was to describe symmetric convergence eye movements evoked by disparity and/or accommodative cues in esotropic macaque, with the goal of determining whether the vergence deficits in these animals mimic those in human esotropia. 

Methods

Animals

Four adult macaque monkeys, three Macaca mulatta (animals ZN, RH, CT) and one Macaca nemestrina (animal TM) were used in the experiments. Monkeys ZN and RH served as control animals and had normal eye alignment. Monkey CT had esotropic strabismus (Table 1) induced by alternating monocular occlusion from birth to age 9 months. ¹⁵ Monkey TM had naturally occurring esotropia with onset at age 4 to 6 weeks, as documented by Ronald G. Boothe at the Yerkes Primate Center (Atlanta, GA). At adult age, this animal was shipped to Washington University in St. Louis. Each of the monkeys was trained to fixate small tracking targets using a positive-feedback reward (a squirt of juice). The experimental protocol was approved by the Washington University Animal Studies Committee and adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Monocular visual acuity was measured awake using spatial sweep visually evoked potentials (VEPs) ^{24 25} (without correction for refractive error), documenting approximately equal vision in both eyes of the control and strabismic monkeys (Table 1 ; normal monocular spacial-sweep acuities in macaque are ≥15 cyc/deg). Cycloplegic refraction revealed mild anisometropia in normal monkey RH and esotropic monkey CT. Monocular VEPs obtained using a flash stimulus and an array of electrodes across the occiput showed no signs of albinism or achiasma in naturally strabismic monkey TM. ^{26 27 28} Funduscopic examination revealed normal pigmentation and no evidence of foveal hypoplasia or other stigmata of ocular albinism. Automated single-cover tests, using liquid crystal shutter goggles, documented a concomitant small-angle esotropia (small-eso) in monkey CT and a concomitant large-angle esotropia (large-eso) in monkey TM (Table 1) . ²⁰ Both strabismic monkeys preferred to fixate using the left eye. (They resembled humans with strabismus who often display a fixation preference or a dominant eye, but lack acuity decrements sufficient to label the nonpreferred eye amblyopic.) In addition to the strabismus, the esotropic monkeys displayed, under conditions of monocular viewing, the mild latent nystagmus (slow phase <1 deg/sec), the nasotemporal asymmetry of horizontal smooth pursuit, and the asymmetric horizontal motion VEPs that typify infantile strabismus. ¹¹  

Scleral Coil and Head Restraint Surgery

After initial training, the monkeys were implanted with subconjunctival magnetic search coils in both eyes (a modification of the technique of Judge et al. ²⁹ ) and a custom-built polycarbonate head restraint device was attached to the skull. ³⁰ The surgical procedure was performed under aseptic conditions using general (intramuscular ketamine and isoflurane by endotracheal tube) and topical anesthesia (proparacaine drops). A 360° conjunctival peritomy was completed, and the coil was attached to the sclera 3 mm anterior to the extraocular muscle insertions, using biological adhesive (Vetbond; 3M Company, St. Paul, MN). The coil wires were looped laterally into a pocket created in the temporal quadrant of Tenon’s capsule, and the conjunctiva was reposited to its natural position at the limbus. The scalp was opened along a sagittal incision using a Bovie cutter and the skull was cleaned. The polycarbonate head restraint was attached using four slotted bolts drilled into the skull. The coil wires were tunneled from the lateral orbit beneath the scalp to connectors at the top of the head restraint. The wires, bolts, connectors, and base of the restraint were encased in dental acrylic, poured and shaped over the skull. An antibiotic-corticosteroid ointment (Cortisporin; GlaxoWellcome; Research Triangle Park, NC) was applied to the conjunctivae and the edges of the scalp wound. The animal was awakened and returned to its home cage. Postoperative analgesia was administered for 72 hours (buprenorphine intramuscularly every 8 hours). The coil wires were used for measurement of passive current only. Neither the coils nor head restraint caused chronic irritation or discomfort to the animals. 

Stimulus Presentation and Eye Movement Recording

Within a few days after the coil surgery each animal was returned to the primate chair for initial eye movement recording. The monkey sat with the head restraint secured in the middle of 3 × 3-ft field coils. Standard magnetic search coil techniques were used to record eye position. ^{31 32} Distant and near targets, subtending 0.5° of arc, were positioned in the midsagittal plane (cyclopean center) of the head. The distant target (located 60 cm from the eyes) was a red laser spot projected onto the back of a translucent screen, and the near target (15 cm from the eyes) was a red light-emitting diode (LED). These targets are known to elicit robust vergence responses in normal macaque monkeys, ³³ and stationary monochromatic light targets have been shown to be adequate stimuli for accommodation in humans (they are suboptimal when dynamic accommodation is needed to maintain focus on a moving target). ^{34 35}  

The average interpupillary distance (IPD) in the monkeys was 30 mm, requiring a convergence angle of 2.8° to bifixate the distant target and an angle of 10.4° to bifixate the near target. Thus, a step change in vergence from distant to near target demanded a +7.6° step change of vergence. The exact convergence angle demanded varied slightly (≤0.6°) as a function of each animal’s IPD. To facilitate analysis and simplify graphic interpretation of the results, the distant target was regarded as optical infinity (0° vergence) and all the graphs of vergence angle and eye position in the results have been adjusted accordingly. Eye position was calibrated by use of a calibration coil and by having the animal perform a lever-response task in which he had to detect 50% dimming of the target within 300 ms, while the target remained stationary at known horizontal and vertical positions. Experiments in normal and strabismic primates have confirmed that foveal fixation is necessary for accurate performance of this task. ³⁰ The calibration sequence was repeated separately for each eye under conditions of monocular viewing. After the initial session, calibration was rapid and remained stable from day to day. The lever was removed, and accurate fixation was encouraged thereafter by rewarding the animal for keeping the eye within a certain window-of-target position. 

Eye movements were recorded under conditions of binocular viewing in sessions lasting approximately 1 hour over a period of several weeks. The room was lit with dimmed background illumination, and the intensity of the laser spot and LED was 3.0 log units above our threshold for detection of a 100-ms flash. ³⁶ The targets were presented in repeated trials (typically 150 per session). To initiate a trial, the animal had to maintain eye position of the right or left eye within ±1.5° of the distant target for a randomized interval of 2 to 5 seconds. After which, the distant target was extinguished and the near target illuminated for an interval of 2 seconds. To receive a reward (a squirt of juice), the monkey had to respond within 1 second of onset of the near target by moving either eye to within a ±1.5° window of near-target position for an interval of 1 second or more. If the monkey failed to fixate the distant target for the full randomized interval or did not move the eye(s) to the near target to satisfy the specified temporal–spatial requirements, a buzzer sounded to provide negative auditory feedback, the trial was aborted, a time-out interval ensued, and no reward was delivered. 

The onset and offset of the targets was controlled by a computer (Macintosh; Apple Computer, Cupertino, CA, running Spike2 data acquisition/analysis software with a model 1401 digital–analog signal processor; Cambridge Electronic Design, Cambridge, UK). Voltages proportional to horizontal and vertical eye position were digitized at 500 Hz. To obtain eye velocity signals, the eye position signals were passed through a (DC to 90 Hz) finite-impulse response filter and differentiated. Angular resolution of the system was 10 minutes of arc (0.17 ± 0 0.08°) with a linearity ±1% over a range of ±40°. 

Data Analysis

Eye movements were analyzed off-line with a modified version of data acquisition/analysis software (Spike2; Cambridge Electronic Design) and averages were plotted with graphics software (Igor Graphics; WaveMetrics, Lake Oswego, OR). Vergence angle was defined as the difference between horizontal left eye position and right eye position (LE minus RE), with eye positions to the right designated as positive values and to the left as negative values. Thus, convergence yielded positive values of vergence and divergence negative values. Unequal movements of the eyes were considered vergence, and equal movements were considered to be saccades. For calculating latencies and amplitudes, onset of eye movement was defined as eye velocity exceeding 10 deg/sec, and initial amplitude was eye position 250 ms after the onset of movement. Trials chosen for analysis were those in which the monkey fulfilled the criteria for stable fixation of the distant target and initiated a convergence or saccadic movement within 750 ms of the target step to the near position. Means ± SEM are reported in the results and were compared using the t-test with significance defined as the 5% confidence level. 

Results

Convergence Accuracy and Latency in Normal Monkeys

We recorded convergence eye movements in two normal monkeys to generate a normative range against which to compare the behavior of the strabismic animals. As shown in Figure 1A , normal monkey ZN executed accurate, stereotyped convergence eye movements in response to a change in the target position from far to near. The step-change in depth of the target required the monkey to converge through a total angle of 8° to achieve accurate bifixation on the near target. The animal converged an average of 7.7° ± 0.15°, or 96% of the near-target vergence angle. The near target in this and all experimental sessions was aligned as accurately as possible on the midsagittal plane (cyclopean center) of the head with the goal of requiring perfectly symmetric vergence in the two eyes. However, the implanted head-restraint device in each animal held the skull firmly but not rigidly, and owing to the powerful neck musculature of adult macaques, head translational displacements of approximately 1 mm to the right or left of the midsagittal plane were unavoidable during the course of a recording session. In the trials shown in Figure 1A , the near target was situated minimally to the right of the midsagittal plane, which required the left eye to converge (adduct) 0.7° more than the right. The right eye adducted −3.5° ± 0.16°, and the left eye +4.2° ± 0.15°. The latency from target step to onset of vergence in this animal was 180 ± 3.4 ms. The initial adduction movement of each eye undershot or overshot target position by as much as 1° (24%–29% of the final position) in approximately 40% of trials. These initial eye position errors were corrected to arrive at final eye position within approximately 400 ms of the onset of vergence. Note (in Fig. 1A ) that convergence was then sustained for the duration of the trial, which lasted several seconds. 

Normal monkey RH also executed robust, symmetric convergence, albeit not as systematically as normal monkey ZN. The convergence behavior of RH is displayed in the eye position records of Figure 1B . In the 10 trials shown, the near target was positioned exactly in the midsagittal plane of the skull, requiring each eye to adduct 3.5° for a total convergence angle of 7°. The initial vergence response of the monkey’s right eye was −3.1° ± 0.23° and of the left eye +3.0° ± 0.20°, for an initial convergence angle of 6.1° or 87% of the near-target angle. The latency from target step to onset of vergence was 190 ± 4.4 ms. At a latency of approximately 500 to 600 ms the monkey made smaller amplitude vergence and disjunctive saccadic eye movements to converge (mainly the left eye) an additional 1° to 1.5°. Monkey RH also exhibited quasisinusoidal, microoscillations (∼1°) of eye position while maintaining convergence in approximately 30% of trials. 

Convergence Variability in Small-Angle Esotropia

Monkey CT had small-angle (1.6° ± 0.7°) esotropia, and his convergence responses when viewing binocularly were of two types: (1) a sustained-convergence response, observed in 73% of trials, and (2) a pulsatile-convergence response, observed in the other 27% of trials. (Binocular viewing means that both eyes were open and uncovered; it does not mean that both eyes were aligned on the target or that the animal had normal binocular perception.) Representative trials of the preponderant, sustained-convergence type are shown in Figure 2A . Note that the monkey preferred to fixate using the left eye, and the angle of right-eye esotropia (i.e., the baseline vergence angle) fluctuated slightly from trial to trial. 

Both the amplitude and latency of evoked convergence in this animal were more variable and less accurate than that of the normal monkeys (compare Fig. 2A with Figs. 1A 1B ). The response variability was unrelated to initial vergence angle—that is, larger angles of esotropia at the start of a trial did not produce smaller convergence responses. Mean convergence amplitude was 4.5° ± 1.8° or 56% of the near target angle of 8° (i.e., the change in vergence from the nonzero value at the start of each trial to the vergence 250 ms after onset of eye movement). The subnormal amplitude of convergence in monkey CT did not help reduce the error between the deviated (right) eye and the target. He converted his strabismic deviation from an esotropia of 1.6° when fixing the distant target to an exotropia of 1.9° when fixing the near (1.6° distance eso +4.5° convergence = 6.1°; 8.0 target angle −6.1 = 1.9° exo error). A conspicuous feature of the animal’s response was a striking asymmetry in the amplitude of adduction of the preferred versus nonpreferred eye. The fixating left eye was responsible for the bulk (89%) of the vergence; the left eye adducted +4.0° ± 1.7° and the right eye −0.5° ± 0.9°. Convergence latency averaged 195 ± 46 ms, which was comparable to that of the two normal monkeys (t-test, P = 0.48). The variability of latency from trial to trial was, however, 10 times greater than in either normal animal. 

Figure 2B shows representative trials in which monkey CT responded with pulsatile rather than sustained convergence. The duration of the pulses ranged from 363 to 780 ms (mean, 488 ± 120 ms). The variability of latency and imprecision in amplitude for pulsatile responses was equivalent to that measured during sustained responses (mean latency pulsatile responses, 184 ± 82 ms; mean amplitude, 5.1° ± 1.6°). The pulsatile-convergence responses in monkey CT represent anomalous vergence behavior that cannot be explained as an artifact of the experimental paradigm or reward system. The strategy did not lead to a reward, because eye position of the preferred left eye did not remain within 1° of the near target angle for 1 second. The two normal monkeys exhibited convergence pulses of this general type in fewer than approximately 3% of all trials. 

Minimal Convergence in Large-Angle Esotropia

Monkey TM had large-angle (25° ± 2.1°) esotropia. He never generated convergence of more than approximately 2° in response to the near target. The trials displayed in Figure 2C show that when viewing binocularly the monkey preferred to fixate using the left eye and exhibited a right-eye esotropic deviation. A step change in position of the target to the near location evoked a relatively conjugate, rightward saccadic eye movement of approximately 4° to allow the preferred left eye to refixate the near target and the monkey to receive the reward (mean right eye amplitude +3.7° ± 0.6° and left eye amplitude +4.0° ± 0.4°). Minimal convergence during and after the saccades produced fluctuations in vergence angle and convergence of 0.5° or less or 18% of the near target angle of 8°. 

The mean latency of the saccadic-vergence response (510 ± 107 ms) was more than twice the delay recorded in the small-eso and normal monkeys. 

Convergence performance, measured in vergence position as a percentage of near target position, is summarized for the four monkeys in Figure 3 . The convergence accuracy of the two strabismic monkeys was significantly weaker than that of either normal animal (n = 50 trials, t-test, P < 0.01). The eye position tracings of Figures 1 and 2 also show that saccades during convergence occurred with much greater frequency in the strabismic compared with the normal monkeys. Saccades during the initial 200 ms of eye movement were uncommon in the normal monkeys (18% of trials in monkey ZN and 24% of trials in monkey RH), but occurred during most of the trials in the two esotropic animals (83% in small-eso monkey CT and 96% in large-eso monkey TM). 

Convergence Velocity Profiles and Interocular Asynchrony

Normal humans have been shown to exhibit small (i.e., <10%) asymmetries in the time taken for each eye to reach peak smooth eye velocity during symmetric disparity convergence movements that are free of any saccades. ³⁷ To determine whether strabismic monkeys have convergence asynchronies outside the normal range, we plotted convergence velocity profiles in the two control monkeys and in the small-eso monkey CT, using saccade-free responses. Because large-eso monkey TM did not show convincing convergence without saccades, he was excluded from the analysis. 

The average eye velocity profiles shown in Figure 4 were generated by having the computer first identify the peak eye velocity of each eye in each trial (n = 50), calculate the mean latency for all peaks in that eye, and align the velocity peaks at the mean latency for that eye to nullify variability in latency from trial to trial. The top panel of Figure 4 shows the averaged velocity responses of normal monkey ZN, with the left eye response (rightward velocity) upward and right eye response (leftward velocity) downward. The velocity profiles for the left and right eye were similar in shape but differed in magnitude; peak left eye velocity was +159.0 ± 3.2 deg/sec and peak right eye velocity was −93 ± 2.8 deg/sec (difference = 66 deg/sec or 42% lower in the right eye; the amplitude of adduction of the right eye in these trials was, on average, 17% lower than the amplitude of adduction in the left eye). Despite the sizable asymmetry in average peak velocity between the eyes, monkey ZN had interocular asynchrony (a delay in the right eye peak) of only 3.7 ± 2.5 ms (Fig. 5) . The interocular asynchrony in normal monkey RH was even smaller, measuring 2.3 ± 0.6 ms. Monkey RH had a peak left eye velocity of +101 ± 3.6 deg/sec and a peak right eye velocity of −72 ± 2.4 deg/sec (difference = 29 deg/sec or 29% lower in the right eye which adducted on average 3% more than the left eye). 

Small-eso monkey CT (Fig. 4 , bottom) had smaller average peak eye velocities, a larger interocular difference in peak eye velocity and greater interocular asynchrony than either normal monkey. Average eye velocity in the left eye peaked at +78 ± 6.3 deg/sec and in the right eye −39 ± 4.1 deg/sec (difference 39 deg/sec or 50% lower in the right eye which adducted on average 88% less than the left eye). In comparison to the two normal animals, monkey CT’s convergence eye velocities were 25% to 50% lower than normal. The lower velocities cannot be attributed to lower amplitudes of adduction, because the adduction amplitude of the dominant eye was comparable to that of the normal monkeys. The interocular asynchrony in monkey CT was 11.8 ± 1.0 ms (a lag of the left, nonpreferred eye peak), or three times the magnitude of the normal monkeys (Fig. 5) . The findings in these three animals suggest a systematic relationship between interocular equality of vergence velocity and interocular synchrony of peak velocity, with the greatest equality and synchrony in normal monkey (RH) and the least in the strabismic animal (CT). 

Accommodative/Monocular Viewing Convergence in Normal Monkeys

In the previous experiments the monkeys viewed binocularly and were able to use both binocular disparity cues and accommodative (blur) cues to drive convergence. To assess the relative contribution of these two cues in the normal and strabismic monkeys, we eliminated binocular disparity by recording vergence to the near target under conditions of monocular viewing. The near target, as in all previous experiments, remained positioned in the midsagittal plane of the head to evoke symmetric convergence. 

Figure 6A shows representative convergence responses in normal monkey ZN viewing monocularly with the left eye. When his responses while viewing binocularly (Fig. 1A) were compared with those when viewing monocularly (Fig. 5A) , the monocular responses were typically lower in amplitude, contained more saccades and were less systematic than the binocular. Specifically, monocular viewing (Fig. 5A) caused lower amplitudes of adduction of the occluded right eye, versional abducting microsaccades of the right eye approximately 40 ms after the onset of convergence, and combined vergence and versional saccadic oscillations of both eyes for several hundred milliseconds after the onset of the response. The amplitude of the evoked convergence in these trials was 1.7° ± 0.4° less (20% less) than the amplitude recorded when viewing binocularly (t-test, P < 0.01). Mean latency when viewing monocularly (175 ± 8.1 ms) was comparable to that recorded when viewing binocularly (t-test, P = 0.24). The monocular-viewing vergence responses of normal monkey RH (not shown) were similar. In both normal monkeys, the amplitude of initial convergence when viewing monocularly was 20% to 31% lower, and the prevalence of saccades during vergence three to four-fold greater, than when viewing binocularly: percent trials with saccades when viewing was monocular 82% versus binocular 18% in monkey ZN, and monocular 87% versus binocular 24% in monkey RH (t-test, P ≤ 0.01). 

Accommodative/Monocular Viewing Convergence in Esotropic Monkeys

Monocular viewing in small-eso monkey CT (occlusion of the right eye, Fig. 6B ) caused a substantial increase in the angle of the strabismus, from esotropia of 1.6° ± 0.7° viewing binocularly to esotropia of 6.1° ± 1.9° viewing monocularly—that is, a latent, additional eso-deviation of 4.5°. A comparison of convergence when viewing binocularly (Fig. 2A) versus monocularly (Fig. 6B) revealed a greater convergence response and more variability of initial right eye position when viewing was monocular. Under both viewing conditions monkey CT responded predominantly by adducting the fixating left eye to the near target angle (target angle from left eye, 4°). The left eye adducted +4.3° ± 0.8° when viewing monocularly compared with +4.0° ± 1.7° when viewing binocularly. However, monocular viewing led to a fivefold greater adduction response of the right eye: binocular viewing, −0.5° ± 0.9°, and monocular viewing, = −2.7° ± 1.6°. The overall effect was that monocular viewing boosted the near-target convergence response of CT from subnormal (+4.5° binocular) to normal (+7.0° monocular). Mean latency viewing monocularly (184 ± 13 ms) was equivalent to that viewing binocularly (195 ± 46 ms; t-test, P = 0.15). 

The monocular viewing responses of large-eso monkey TM (Fig. 6C) were poor and equivalent to those during binocular viewing (Fig. 2C) , in both amplitude of convergence and latency (t-test, P = 0.42 and 0.25). Unlike the two normal monkeys, the frequency of saccadic vergence in the strabismic monkeys was the same for monocular and binocular viewing (t-test, P = 0.22 monkey CT and 0.12 monkey TM). Disconjugate saccades occurred during convergence in approximately 75% of trials for small-eso monkey CT and approximately 95% of trials for large-eso monkey TM, whether they viewed monocularly or binocularly. 

Accommodative-Convergence/Accommodation Ratios

The synkinetic relationship between accommodative convergence and accommodation is expressed clinically as a ratio (AC/A, in prism diopters/sphere diopters with 1 prism diopter = 0.57°). ³⁸ A stimulus AC/A ratio is measured in the typical clinical setting with the assumption that lens accommodation matches the accommodative demand of the near target to eliminate all blur. Measurement of an actual response AC/A requires recording with an optometer. ²³ We measured a stimulus AC/A ratio in each of our four monkeys with the assumption that they performed the 5.0 D of accommodation demanded for the near target (the distant target [0.6 m] demanded 1.6 D of accommodation and the near target [0.15 m] demanded 6.6 D, where D = 1/target distance in meters). The stimulus AC/A ratios in normal monkeys ZN and RH measured 2.66 and 2.54, respectively. These values agree with response AC/A ratios (2.5–3.0) calculated from the data of Cumming and Judge ³ (their Fig. 3A , monkey H), who used an infrared optometer and eye coils to measure lens accommodation and vergence simultaneously in a normal monkey. The stimulus AC/A ratio in small-eso monkey CT was also normal: 2.80. The very weak accommodative vergence response of large-eso monkey TM yielded a very low stimulus AC/A ratio: 0.52. The older age of monkey TM (19 years compared with 5 years for monkeys ZN and CT) may account in part for his low ratio, as stimulus AC/A ratios fall considerably in normal humans with onset of presbyopia. ^{23 39}  

Convergence When Viewing Binocularly Versus Monocularly

To reveal with greater clarity the degree to which binocular disparity cues enhanced convergence, average convergence velocity and position responses were plotted (Figures 7 8 9) in normal monkey ZN and the two esotropic monkeys for binocular versus monocular viewing (n = 20 trials). In describing these results, we refer to binocular viewing as disparity vergence and monocular viewing as accommodative vergence though obviously the monkeys had access to both disparity and accommodative cues when viewing binocularly. 

Figure 7 shows average convergence velocity profiles in these three animals. In normal monkey ZN (Fig. 7A) , disparity vergence velocity exceeded at its peak 200°/sec, or two-and-one-half-times the peak velocity of accommodative vergence (79°/sec). In contrast, small-eso monkey CT had accommodative and disparity peak vergence velocities (Fig. 7B) that were equivalent, and 36% and 75% lower than the peak velocity of accommodative and disparity vergence, respectively, in the normal monkey (t-test, P ≤ 0.01). Large-eso monkey TM’s (Fig. 7C) accommodative and disparity vergence velocities were also equivalent, but substantially weaker than either the small-eso or the normal monkey’s responses (89% and 88% lower than the disparity and accommodation velocities, respectively, of the normal monkey). 

The difference in amplitude of convergence for the two viewing conditions is shown in Figure 8 . Two hundred milliseconds after the onset of the response, disparity vergence in normal monkey ZN (Fig. 8A) exceeded accommodative vergence by approximately 20% (1.7°). In small-eso monkey CT (Fig. 8B) , the inverse was true: accommodative vergence exceeded disparity vergence by approximately 33%, or 2.5°. To facilitate comparison, the starting point of the monocular response (6.1° convergences) in Figure 8B has been rescaled to the starting point of the binocular response (1.9°). In large-eso monkey TM (Fig. 8C) , weak accommodative vergence exceeded very weak disparity vergence by approximately 0.5° at a latency of approximately 400 ms. 

The plots in Figure 9 show convergence velocity as a function of vergence position rather than time (i.e., phase plane plots). These profiles reinforce the point that disparity vergence—throughout the course of the response—was substantially more robust than accommodative vergence in the normal monkey, whereas in the strabismic animals the two types of vergence were equivalent and, in the case of disparity, much weaker than normal. Taken together, the data of Figures 7 8 9 imply that convergence in the esotropic monkeys was driven largely by accommodative cues, with little input from binocular disparity. 

The convergence behavior of small-eso monkey CT, however, revealed some sensitivity to binocular disparity. Disparity cues appear to have provided weak negative feedback to CT’s fusional vergence system. The negative feedback was evident as a smaller angle of esotropia under conditions of binocular viewing (compare Figs. 2A and 6B ) and a smaller amplitude of convergence to the near target when viewing binocularly versus monocularly (Fig. 8B) . 

Discussion

The main purpose of our study was to determine whether convergence eye movements in strabismic monkeys were maldeveloped and whether the maldevelopments were similar to those described in strabismic humans. Our results show that monkeys with esotropic have major deficits of convergence when viewing binocularly. The deficits appear to be remarkably similar to those reported using eye movement recordings in humans who have strabismus. ^{21 23} The deficits in the strabismic monkeys varied in magnitude and latency, with the most profound deficits in the animal with the largest angle of infantile esotropia. 

Similarity between Strabismic Monkeys and Humans

Kenyon et al. ²¹ and Ciuffreda and Kenyon ²³ provided the first comprehensive description of abnormalities of vergence eye movements in strabismic humans using an infrared-reflection recording technique. The stimuli were physical far and near targets placed in the subjects’ midline so as to evoke symmetric vergence movements of each eye to target disparities of approximately 4° to 6°. Normal humans responded predominantly with symmetric vergence. The strabismic patients (constant or intermittent heterotropia, with or without amblyopia) responded predominantly with asymmetric vergence accompanied by a disjunctive saccade. The asymmetry of vergence amplitude between the two eyes in their normal subjects was 10%, compared with an average asymmetry of 88% in the strabismic subjects. The prevalence of saccadic vergence in the normal human was less than 10%, compared with a prevalence of 85% in the strabismic patients. Both studies also compared vergence behavior in normal and strabismic subjects under conditions of binocular versus monocular viewing. The vergence behavior of the patients with strabismus did not change, whereas the vergence of normal subjects became remarkably strabismus-like when they viewed monocularly, typified by a four- to fivefold increase in the prevalence of asymmetric-saccadic vergence. The authors concluded from these findings that vergence in humans with strabismus when viewing binocularly was a monocular, accommodative vergence response driven chiefly by visual inputs to the dominant eye. ^{21 23}  

Our results in strabismic macaque are compatible with the findings in humans and the conclusions of Kenyon et al. ²¹ and Cuiffreda and Kenyon. ²³ When viewing binocularly, the normal monkeys in the present report made remarkably symmetric, smooth vergence movements, with a prevalence of accompanying saccades on the order of approximately 20%. In contrast, the prevalence of saccadic vergence in the two strabismic monkeys exceeded 80%, and the prevalence of asymmetric vergence approached 100%, whether viewing binocularly or monocularly. When the normal monkeys were required to view monocularly, their vergence behavior became remarkably strabismus-like, dominated by asymmetric, saccadic convergence responses. 

Symmetric Convergence in Normal Primate

Most previous studies of vergence in monkeys have used haploscopic viewing to manipulate accommodative blur and binocular disparity independently, ^{3 40 41} or have aligned the targets along the visual axis of one eye to evoke asymmetrical accommodative vergence. ^{3 42} Maxwell and King ³³ used a stimulus setup similar to that which we used, placing physical (LED) far and near targets in the midline to evoke natural, symmetrical accommodative-disparity vergence. The average vergence latencies and velocities recorded in our control monkeys are comparable to those reported in normal macaque by Maxwell and King, ³³ Cumming and Judge, ³ and Mays et al. ⁴¹ The latencies for normal monkeys in our study and in previous studies are also comparable to those reported for normal adult humans—that is, approximately 160 to 200 ms. ^{1 43 44} The convergence velocities in normal monkeys, including ours, are, however, higher on average than those reported in human. Erkelens et al. ⁴⁴ measured peak vergence speeds of approximately 4 deg/sec per degree of convergence for symmetric vergence trials in humans, ⁴⁴ compared with approximately 13 to 25 deg/sec per degree of convergence in normal monkeys (pooling our data with that of Maxwell and King ³³ ). It is possible that the higher velocities in our control monkeys occurred in part because we used only one distance for the near target, which may have enhanced performance by allowing the animals to predict accurately the near target distance and preprogram execution of a stereotyped, convergence response. 

Cumming and Judge ³ compared vergence velocities in normal monkeys when viewing binocularly versus monocularly and found that disparity was a much more powerful stimulus to vergence than accommodative blur. ³ Convergence velocities when the monkeys viewed binocularly were on average three times greater than those achieved when they viewed monocularly. Our results (Fig. 9) in normal monkeys ZN and RH were similar, with binocular viewing evoking convergence velocities 2.5 to 3.0 times greater than monocular viewing. 

Vergence System Noise in Strabismic Monkey

Several features of abnormal vergence in the esotropic monkeys can be interpreted as excess noise in the CNS pathways that drive vergence. The most fundamental of these features is variability in position of the nonpreferred eye (and the angle of esotropia) from trial to trial under conditions of binocular viewing (see Fig. 2 ). This variability was labeled “scatter” by Quick et al., ¹⁹ who used still photographs (i.e., a modified Hirshberg technique) to measure eye alignment in strabismic macaques at different horizontal, vertical, and near-far gaze positions. They noted scatter of eye position in the nonfixating eye only, which was on average two times greater than that measured in the fixating eye. Quick et al. ascribed the scatter to a noise generator at an as yet unspecified location in the vergence neural pathways. The precision afforded by scleral search coil recordings in the present report reveal with greater clarity not only the magnitude of scatter in the nonfixating eye of strabismic macaque, but also scatter in the fixating eye. On average, our strabismic monkeys displayed scatter in the fixating eye that was 2 times greater, and in the nonfixating eye 10 times greater than that observed in the control animals. Fixation scatter of this magnitude has also been documented in earlier reports from our laboratory of other esotropic monkeys recorded using binocular search coils. ^{20 30}  

The notion of abnormal vergence system noise is supported by the finding of both pulsatile and sustained vergence responses in small-eso monkey CT (Fig. 2) . The unpredictable profiles of the pulsatile responses, which varied widely in latency and amplitude, coupled with the fact that they appeared mainly during binocular viewing, imply that the monkey retained a weak capacity for detection of binocular disparity. Residual disparity sensitivity, though usually subthreshold, could occasionally trigger but not sustain convergence. Weak disparity sensitivity also explains the finding of a smaller angle of esotropia when the animal viewed binocularly. Weak disparity sensitivity provided negative feedback to the vergence system, reducing the angle of strabismus. 

The sustained convergence responses in monkey CT conformed to the pattern of accommodative vergence responses in the two control animals. However, even the sustained responses of the strabismic monkey indicate processing contaminated by temporal–spatial noise, in that the variance of the sustained responses (in latency and amplitude) was at least five times greater than that measured in either normal monkey. 

Small-Angle Esotropia in Humans and Monkeys

Small-eso monkey CT displayed vergence behaviors remarkably similar to those of patients who have small-angle esotropia (microstrabismus). The humans, like monkey CT, usually have larger angles of strabismus when one eye is covered, show a normal stimulus AC/A ratio, and demonstrate a residual capacity to drive vergence using binocular disparity cues (clinically termed intact fusional vergence amplitudes). ^{10 45} Harweth et al. ⁴⁶ used psychophysical methods rather than vergence eye movement recordings to reveal subnormal disparity sensitivity in microstrabismic macaques. The microstrabismus was inferred from the fact that the animals initially had large angles of esotropia produced in infancy by eye muscle surgery. Within 1 year of the procedure, the animals spontaneously recovered (by inspection) eye alignment and showed varying degrees of stereopsis as well as psychophysically measured fixation disparities. 

Visual Cortex and Vergence Maldevelopment

The esotropic monkeys were able to generate accommodative vergence and normal saccadic eye movements, yet displayed striking maldevelopments of binocular-disparity induced vergence. These results imply that the neural mechanism of the disparity vergence deficit lies not in the extraocular muscles, motor nuclei, or convergence-related neurons of the midbrain, but rather in the visual cortical areas that process disparity early in the vergence sensorimotor pathway. Disparity-selective neurons implicated in the control of vergence have been found in many cortical visual areas in monkey and fall generally into near and far subtypes, sensitive to crossed and uncrossed disparities, respectively, that could drive convergence and divergence. ^{47 48 49 50 51} Masson et al. ⁵² have shown that patterns containing binocular disparity act as a powerful stimulus to vergence in both humans and monkeys, even if the patterns cannot evoke a percept of binocular fusion because they are composed of anticorrelated dots. Taken together, these findings imply that vergence is driven in part by disparity-sensitive neurons at the earliest (i.e., prestereoscopic) stage of binocular processing—visual area V1. Esotropic monkeys have a paucity of V1 horizontal connections joining ocular dominance columns of opposite ocularity, both in lamina 2/3 and in lamina 4B. ^{18 53 54} The output from lamina 4B provides a major projection to extrastriate areas MT/MST, implicated in the perception of stereopsis and the control of vergence. ^{51 55} These binocular connections would be important for processing of binocular disparity, and monkeys with infantile strabismus have been shown to lack normal disparity sensitivity. ^{15 46 56} The deficit of binocular V1 connections may also account for the abnormal disparity–vergence behavior we have described. 

 Supported by National Eye Institute Grant R01 EY 10214-01A3 (LT), and in part by Grant EY02687 and an unrestricted grant to the Department of Ophthalmology and Visual Sciences from Research to Prevent Blindness, Inc.

 Submitted for publication July 10, 2002; revised January 16, 2003; accepted February 17, 2003.

 Disclosure: L. Tychsen, None; C. Scott, 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.

 

Table 1.

 

Visual and Ocular Motor History and Findings in Two Normal and Two Strabismic Monkeys

Animal/Age	History	Eye Alignment	Visual Acuity (SSVEP-cpd)	Refractive Error	Other Ocular Motor Findings
ZN/8 y	Normal	Ortho	RE 15.3; LE 17.4	RE+0.50; LE+1.00	Stable fixation; normal pursuit/OKN
RH/5 y	Normal	Ortho	RE 20.4; LE 17.4	RE−1.00+ 1.00 × 090; LE−2.50+ 3.00× 070	Stable fixation: normal pursuit/OKN
CT/5 y	Alternate occulsion; birth to 9 mo	Small-angle esotropia; RET ∼2°	RE 15.9; LE 18.1	RE −0.75+ 2.00× 110; LE−0.75+ 1.75× 080	Latent nystagmus; pursuit/OKN asymmetry
TM/19 y	Naturally occuring infantile esotropia	Large-angle esotropia; RET ∼25°	RE 14.5; LE 15.0	RE+0.50; LE+0.50	Latent nystagmus; pursuit/OKN asymmetry

Figure 1.

 

Multiple trials of symmetric convergence eye movements in two normal monkeys evoked by a step change in target position from far to near. (A) Responses of monkey ZN. (B) Responses of monkey RH. Targets were positioned in each animals’ midline (0° eccentricity) and stepped to the near target position at the beginning of each raw data tracing. Viewing was binocular. Upward–positive values of vergence = convergence, and downward–negative values = divergence. Dotted line: near target vergence angle. Upward–positive values of horizontal right and left eye positions = rightward and downward–negative values = leftward.

Figure 2.

 

Trials of convergence in two esotropic monkeys in response to change in target position from far to near. (A) Asymmetric, variable responses of the sustained-convergence type in small-eso monkey CT. Bulk of vergence response was due to adduction of left eye. The monkey preferred to fixate using the left eye and had right eye esotropia averaging 1.6°. (B) Responses of unsustained, pulsatile convergence type in monkey CT. (C) Large-eso monkey TM responded using conjugate saccades and exhibited minimal convergence. TM preferred to fixate using the left eye and had right eye esotropia averaging 25°. Note striking variability in latency of response from trial to trial in the strabismic monkeys. Stimulus paradigm and conventions are the same as in Figure 1 .

Figure 3.

 

Convergence amplitude 250 ms after the onset of vergence in the four monkeys. Average amplitude as a function of target amplitude. Initial convergence in the two strabismic monkeys was significantly weaker than normal.

Figure 4.

 

Average right eye and left eye velocity profiles for symmetric convergence trials. Data from three monkeys (two normal, one esotropic). Greatest interocular asymmetry (degrees/second) and interocular asynchrony (ms) in velocity peaks was measured in strabismic monkey CT. Solid line: left eye velocity; dashed line: right eye velocity. Upward = rightward; downward = leftward.

Figure 5.

 

Interocular asynchrony for symmetric convergence trials in the three monkeys shown in Figure 4 . Asynchrony is the difference in latency of average right eye versus left eye velocity peak. Asynchrony in esotropic monkey CT was significantly greater than normal.

Figure 6.

 

Multiple trials of accommodative (i.e., monocular viewing) symmetric convergence eye movements in a normal monkey and two esotropic animals. Viewing using the left eye (right eye occluded). (A) Responses of normal monkey ZN were asymmetric, with the bulk of the response due to adduction of left eye. Right eye movements showed disjunctive saccades. (B) Strabismic monkey CT. Angle of left eye esotropia increased an average of 4.5° with monocular viewing (compare with Fig. 2A ) and the position of the left eye was more variable. (C) Strabismic monkey TM. Angle of esotropia and responses during monocular viewing were comparable to those during binocular viewing (Fig. 2C) .

Figure 7.

 

Convergence velocity during binocular viewing (disparity) versus monocular viewing (accommodative) in a normal monkey and the two strabismic animals. (A) Normal monkey. Disparity vergence velocity was 2.5 times greater than accommodative. (B) Esotropic monkey CT. Disparity and accommodative vergence velocities were equivalent to each other and similar to accommodative response in normal monkey. (C) Esotropic monkey TM. Disparity and accommodative responses were equally weak. Averages for 20 trials each condition in each animal.

Figure 8.

 

Convergence amplitude during binocular viewing (disparity) versus monocular viewing (accommodative) in a normal monkey and the two strabismic animals. (A) Normal monkey. Disparity response exceeded accommodative response by ∼20%. (B) Esotropic monkey CT. The slow- but normal-amplitude accommodative response exceeded the disparity response by ∼33%. The starting point of the accommodative response (dashed line) has been adjusted from 6.1° to 1.9° to facilitate comparison with the disparity response. (C) Esotropic monkey TM. Accommodative responses were minimally greater than negligible disparity response.

Figure 9.

 

Convergence velocity plotted as a function of vergence angle, in a normal monkey and two strabismic animals. (A) Normal monkey ZN′s disparity response was symmetric and robust, achieving 2.5 times the velocity of accommodative response. (B) Esotropic monkey CT’s disparity and accommodative responses were equivalent. Velocity profiles became irregular as convergence angle exceeded 5°. (C) Esotropic monkey TM. Response profiles were irregular and exceedingly weak.