Electrical stimulation of superior colliculus affects strabismus angle in monkey models for strabismus

Electrical stimulation of the superior colliculus in strabismic monkeys results in a change in eye misalignment. These data support the notion of developmental disruption of vergence circuits leading to maintenance of eye misalignment.

Abstract

Disruption of binocular vision during the critical period for development leads to eye misalignment in humans and in monkey models. We have previously suggested that disruption within a vergence circuit could be the neural basis for strabismus. Electrical stimulation in the rostral superior colliculus (rSC) leads to vergence eye movements in normal monkeys. Therefore, the purpose of this study was to investigate the effect of SC stimulation on eye misalignment in strabismic monkeys. Electrical stimulation was delivered to 51 sites in the intermediate and deep layers of the SC (400 Hz, 0.5-s duration, 10–40 μA) in 3 adult optical prism-reared strabismic monkeys. Scleral search coils were used to measure movements of both eyes during a fixation task. Staircase saccades with horizontal and vertical components were elicited by stimulation as predicted from the SC topographic map. Electrical stimulation also resulted in significant changes in horizontal strabismus angle, i.e., a shift toward exotropia/esotropia depending on stimulation site. Electrically evoked saccade vector amplitude in the two eyes was not significantly different (P > 0.05; paired t-test) but saccade direction differed. However, saccade disconjugacy accounted for only ~50% of the change in horizontal misalignment while disconjugate postsaccadic movements accounted for the other ~50% of the change in misalignment due to electrical stimulation. In summary, our data suggest that electrical stimulation of the SC of strabismic monkeys produces a change in horizontal eye alignment that is due to a combination of disconjugate saccadic eye movements and disconjugate postsaccadic movements.

NEW & NOTEWORTHY Electrical stimulation of the superior colliculus in strabismic monkeys results in a change in eye misalignment. These data support the notion of developmental disruption of vergence circuits leading to maintenance of eye misalignment in strabismus.

strabismus is a misalignment of the visual axes of the two eyes with a worldwide prevalence in childhood of ~2–5% (Govindan et al. 2005; Greenberg et al. 2007; Lorenz 2002; Mohney 2007). A disruption in binocular vision during the critical period for visual and oculomotor development leads to developmental strabismus and this is also the basis for the induction of strabismus in animal models (Boothe et al. 1985; Harwerth et al. 1986; Kiorpes 2015). Recent studies have shown that neural activity within oculomotor and abducens motor nuclei innervating the extraocular muscle (EOM) drives the state of eye misalignment in strabismic monkeys (Das and Mustari 2007; Joshi and Das 2011; Walton et al. 2014a). In addition, neural recording of neurons in the midbrain supraoculomotor area (SOA) in strabismic monkeys has shown that activity in these cells is correlated with strabismus angle (Das 2012). Also, muscimol inactivation of the caudal fastigial nucleus (cFN) and posterior interposed nucleus (PIN) of the cerebellum in strabismic monkeys induces changes in eye misalignment (Joshi and Das 2013). These findings in different brain areas have led to a framework wherein disruption in oculomotor neural circuits related to vergence eye movements leads to eye misalignment.

The superior colliculus (SC) is a laminated midbrain structure known to be critical for saccadic eye movements (Gandhi and Katnani 2011). Superficial layers of the SC encode a retinotopic map of visual space, while the intermediate and deeper layers of the SC encode a motor map, i.e, encode a specific saccade vector from initial position (Cynader and Berman 1972; Gandhi and Katnani 2011; Goldberg and Wurtz 1972; Schiller and Stryker 1972; Wurtz and Goldberg 1971, 1972). The SC map is topographically organized such that saccadic amplitude increases from rostral to caudal and saccade direction changes from upward to downward along the mediolateral direction. The SC, especially the rostral part of the SC (rSC), is also involved in smooth pursuit, although its role in smooth pursuit appears not to be related to motor commands for smooth pursuit (smooth-pursuit eye movements are not elicited by rSC electrical stimulation during fixation) but rather to aspects of target selection/movement initiation (specifying an eye-movement goal) and perhaps providing a position signal to the smooth-pursuit system during ongoing pursuit (Basso et al. 2000; Gandhi and Katnani 2011; Krauzlis and Dill 2002; Krauzlis 2004).

Although the role of the SC in saccade and smooth-pursuit eye movements has been the focus of the majority of published studies, there is a fair bit of evidence supporting the role of the SC in vergence eye movements, an aspect that could be of specific interest for strabismus. Convergence related neurons have been identified in the cat rSC (Jiang et al. 1996). Also, electrical stimulation or pharmacological inactivation of the cat rSC produced changes in both accommodation and vergence (Ohtsuka and Sato 1996; Sawa and Ohtsuka 1994; Suzuki et al. 2004). Monkey studies also suggest a role of the SC in vergence, although perhaps interpretation is more complex. Billitz and Mays (1997) were not able to elicit vergence by electrical stimulation in the SC during far viewing but electrical stimulation during near viewing caused a relaxation of vergence. In another study, electrical stimulation in the SC interfered with vergence only if applied just before or during a vergence only movement or a combined saccade-vergence movement (Chaturvedi and van Gisbergen 1999, 2000). One hypothesis for the apparent lack of vergence changes due to electrical stimulation in normal monkeys is that a net “zero vergence” command is initiated because both convergence and divergence related neurons are activated. Lawler and Cowey (1986) performed ablations of the rSC in monkeys and suggested that there were problems with both disparity processing and eye alignment although eye movements were not explicitly recorded in this study. Neurons in the SC have been shown to receive monosynaptic projections from neurons in lateral intraparietal cortex (LIP) that also encoded depth information (Gnadt and Beyer 1998). Walton and Mays (2003) showed that saccade-related neurons in the caudal colliculus showed a weak relationship to vergence in that many burst neurons showed a reduction in saccade velocity sensitivity when looking at near targets compared with when looking at far targets. However they were unable to identify any systematic three-dimensional tuning of neurons. Interestingly, early studies investigating SOA vergence neurons also reported another population of vergence neurons located 4–5 mm dorsal and 2–3 mm lateral to the oculomotor nucleus (OMN) that they did not unequivocally localize using histological methods but could be in the rSC (Judge and Cumming 1986; Mays 1984; Mays et al. 1986). A recent study by Van Horn et al. (2013) indeed identified convergence and divergence neurons in the rSC that were modulated during slow vergence but not conjugate or fast vergence eye movements thereby postulating that the rSC only contributes to slow vergence. Also, electrical stimulation in this area produced vergence angle changes when looking at near targets.

Considering the evidence for SC involvement in vergence and the potential for disruption in vergence circuits as a neural substrate for strabismus, we decided to examine the role of the SC in maintaining the state of misalignment in strabismus monkey models. We used the strategy of electrical stimulation of the SC as it has commonly been used in this structure to elicit saccades and more recently also in the rSC to elicit vergence. Thus the specific goal of the study was to determine whether stimulation of the SC could influence the state of eye misalignment in juvenile strabismic monkeys. Some of these data have appeared before in abstract form (Upadhyaya et al. 2016).

METHODS

Subjects and surgical procedures.

Three adult strabismic monkeys (Macaca mulatta) were used in this study (monkeys E, L, and H; age: 6–7 yr; weight: 5–10 kg). These monkeys were previously reared using an optical prism-viewing paradigm. In this paradigm, infant monkeys viewed through a 20-prism-diopter Fresnel prism, oriented base-in and placed in front of one eye and another 20-prism-diopter Fresnel prism-oriented base-up and placed in front of the other eye. These horizontal and vertical Fresnel prisms were fitted in a lightweight helmet-like device that the animal wore for the first 4 mo of life starting from 1 to 2 days after birth. After the initial 4 mo of prism rearing, monkeys were reared for several years, in a normal visual environment at the University of Houston College of Optometry, before experiments were begun. Disruption of binocular vision due to prism viewing during this initial period leads to strabismus as it is the critical period for development of eye alignment, stereopsis, and binocular sensitivity.

Before experimentation, each juvenile monkey went through three sterile surgical procedures carried out under aseptic conditions under isoflurane anesthesia (1.25–2.5%). In the first surgery, a head stabilization post was implanted (Adams et al. 2007). In the second surgery, we stereotaxically implanted a 21-mm diameter titanium neural recording chamber along with a scleral search coil in one eye and in the third surgery we implanted a scleral search coil in the other eye (Judge et al. 1980). In monkeys L and H, the chamber was implanted at a stereotaxic location centered 3-mm anterior and 1-mm lateral to stereotaxic zero and tilted 20° dorsolateral to ventromedial in the coronal plane. In monkey E, the chamber was implanted in the mid-sagittal plane and centered at anterior-posterior 0 mm and tilted posteriorly by 38°. All surgical and experimental procedures were performed in strict compliance with National Institute of Health and The Association for Research in Vision and Ophthalmology guidelines and the protocols were reviewed and approved by the Institutional Animal Care and Use Committee of the University of Houston. After recovery from surgery and additional behavioral acclimatization to the laboratory environment, training in standard oculomotor tasks such as saccades, smooth pursuit, and fixation was undertaken.

Eye movement measurements and experimental procedures.

Binocular eye position was measured using the magnetic search coil method (Primelec Industries, Regensdorf, Switzerland). Eye coils were calibrated at the beginning of each experiment by rewarding the animals with small amounts of juice as they looked at a series of targets along the horizontal or vertical meridian projected onto a tangent screen at 57 cm. Calibration of each eye was performed independently during monocular viewing forced by occluding one or the other eye using liquid crystal shutter goggles (Citizen Fine Devices, Nagano, Japan) that were under computer control. Visual stimuli were generated using the BITS# visual stimulus generator (Cambridge Research Systems, Cambridge, UK) and Psychtoolbox 3 (Brainard 1997) operated under computer control and presented using a DepthQ projector running at a 120-Hz frame rate (Lightspeed Design, Bellevue, WA).

We identified the SC using a combination of neural activity recording and electrical stimulation methods. During neural recording (epoxy-coated tungsten electrodes; 1–5 MΩ; Frederik Haer, Brunswick, ME), as the electrode was advanced into the area of SC, we first encountered visual (superficial layers) and then saccade-related (intermediate and deeper layers) activity. SC locations were confirmed by electrical microstimulation methods wherein a train of low current cathodal pulses (10–40 μA, 400 Hz, 500 ms) delivered via the recording electrode elicited a staircase of saccades in a specific direction. After determination that a particular site yielded consistent staircase saccades at low currents, electrical stimulation of 500-ms duration was repeated multiple times with each eye viewing and these data were used for later analysis. Binocular eye position signals were processed with antialiasing filters (Krohn-Hite, Brockton, MA) at 400 Hz before digitization at 2.75 kHz with 12-bit precision (AlphaLab System; Alpha-Omega Engineering, Nazareth, Israel).

Data analysis.

The fundamental goal of the analysis was to examine changes in vertical and horizontal eye alignment during electrical stimulation and correlate them with amplitude and direction of saccades and postsaccadic movements elicited due to stimulation. Binocular eye position data from electrical stimulation sites that resulted in staircase saccades at low threshold currents were analyzed using custom MATLAB programs. Eye velocity was calculated from eye position data using a central difference algorithm. Eye position and velocity data were further filtered using an 80-point finite impulse response digital filter with a pass band of 0-80 Hz before analysis.

Each electrical stimulation trial was initially reviewed by the investigator to determine whether the stimulation evoked staircase saccades, an indication to accept the trial. Trials were rejected if the animal broke fixation just before the stimulation or if no saccades were initiated within 100 ms of stimulation. Once the trial was accepted, strabismus angle (difference in position of left and right eyes) was calculated just before stimulation and at the end of stimulation (i.e., after 500 ms) and a mean change in strabismus angle was calculated as the mean difference in strabismus angle before and after stimulation over the multiple stimulation trials at each site.

For analysis of electrically evoked saccades, saccade onset and offset were identified using a combination of velocity criterion of 30°/s and acceleration criterion of 3,000°/s². Although the staircase saccades following electrical stimulation were evident visually in the data records from all stimulation sites, the saccade detection criteria needed to be adjusted to reliably detect saccade onset and offset at a minority of stimulation sites (15/51). The adjustments to the criteria were relatively small and consisted of changing the velocity criteria between 20 and 60°/s and changing the acceleration criteria between 1,000 and 4,000°/s². At one site we used a velocity criterion of 100°/s. Generally, the higher threshold values were necessary to reliably detect saccades in the presence of significant postsaccadic movement and lower threshold values were necessary to identify smaller saccade amplitudes. All saccade onset and offset locations were visually verified before acceptance. Radial amplitude and direction of each saccade in each eye were calculated and a mean radial amplitude and mean direction for a single electrically elicited saccade at a particular SC site were calculated (mean of all electrically evoked saccades at a specific site). Saccades were grouped according to viewing eye (right eye viewing or left eye viewing) and also site of stimulation. Statistical analysis (SigmaPlot V12.5) focused on evaluating disconjugacy in radial saccade amplitude, disconjugacy in polar direction and correlating change in misalignment with saccade parameters.

RESULTS

Monkey E presented with an exotropia (XT; divergent misalignment) with either eye viewing of ~25° while the other two animals had strabismus that was variable and also changed depending on viewing eye (evidence for a dissociated horizontal deviation). Strabismus in monkey H ranged from 5° esotropia (ET; convergent misalignment) to 15° XT (median: 8° XT) during right eye viewing and 1° ET to 12° XT (median: 7° XT) during left eye viewing. Strabismus in monkey L ranged from 5° ET to 18° XT (median: 13° XT) during left eye viewing and 5° ET to 25° XT (median: 12° XT) during right eye viewing. The large variability in the strabismus angle was not on a moment-to-moment basis, i.e., was not related to fixation instability. Rather we observed variations over the several hours of a recording session and sometimes less exotropia during fixation compared with smooth pursuit in the same recording session. These relatively slow changes in strabismus angle may be the result of slow variations in the tonic accommodative state of the monkey. There were also variations that occurred over the several months during which data for this study were acquired, which is sometimes observed in human strabismus also. Note, however, that the current study focuses on changes in strabismus angle over a very short time period of electrical stimulation (500 ms), and therefore, innate variability in strabismus angle is not a factor. Monkeys also showed a vertical misalignment that was determined to be a dissociated vertical deviation (DVD) in monkeys E and L of ~0 to ~6° and a pure vertical strabismus in monkey H of ~0 to 2°. Alignment properties are also evident from the Hess screen chart representation of eye positions shown in Fig. 1, obtained as the animals’ performed smooth-pursuit tracking of a target moving along the horizontal or vertical meridian. Although each of the monkeys showed some evidence for pattern strabismus (variation of horizontal misalignment with vertical gaze position) and also variation of vertical misalignment with horizontal gaze position, these gaze-position dependent variations in strabismus angle were small (Fig. 1), as is typical of many humans with strabismus.

Fig. 1.

Plot showing horizontal and vertical position of right eye (red trace) and left eye (blue trace) as observed during monocular smooth-pursuit tracking (0.3 Hz, ± 15°) along the horizontal or vertical meridian in monkeys E, H, and L. Left: data acquired during left eye viewing and right column shows data acquired during right eye viewing. Positive values indicate rightward or upward positions and negative values indicate leftward or downward position. Traces are means developed from several cycles of desaccaded eye positions during smooth-pursuit tracking. All monkeys presented with varying degrees of exotropia and small vertical misalignment and pattern strabismus.

Description of SC stimulation sites.

Electrical stimulation of SC was performed in 51 sites in the 3 animals of which 12 were in right SC and 39 in left SC. Of these 51 sites, 41 sites were tested under both right and left eye viewing conditions. Seven sites were tested under only left eye viewing and three other sites were tested under only right eye viewing. Stimulation at each of these 51 sites evoked staircase saccades of various amplitudes and directions. Figure 2 shows a polar plot of the evoked saccade vector in the viewing eye at each of these sites acquired during left eye viewing. Due to the previous finding of vergence neurons in rSC, our sample of stimulation sites was purposely skewed toward rSC, although we sampled at a minority of caudal SC sites as well. During left eye viewing, 36/48 stimulation sites yielded saccades of radial amplitude <5° (mean radial amplitude = 1.79°; range = 0.29–4.95°). Mean radial amplitude of electrically evoked saccades from the other 12 sites was 9.39° (range = 5.35–16.74°).

Fig. 2.

Polar plot showing amplitude and direction of left eye saccade vector in all animals evoked at each of the 48 electrical stimulation sites during left eye viewing.

Change in strabismus angle due to electrical stimulation of the SC.

Electrical stimulation of the SC invariably resulted in a change in strabismus angle. Figure 3, A–D, shows raw horizontal and vertical eye position data (Fig. 3, A and C) and horizontal and vertical strabismus angle Fig. 3, B and D; strabismus angle = left eye position minus right eye position) from seven superimposed trials following electrical stimulation of the left SC in monkey H. Eye movement traces are aligned on stimulation onset and the animal was fixating with the monkey’s left eye (right eye under cover) before stimulation. As seen in Fig. 3, electrical stimulation at this SC site results in a staircase of rightward and upward saccades with radial amplitude of ~2° (rostral site). Electrical stimulation for 500 ms at this site also results in a significant divergent change in horizontal misalignment of ~6° (more exotropic) and little change in vertical misalignment (Fig. 3, B and D).

Fig. 3.

Raw data showing eye movement responses following electrical stimulation in the left superior colliculus of monkey H (A–D) and monkey E (E–H) during monocular left eye viewing. A and E: multiple traces of horizontal eye position of the right (red) and left (blue) eyes. Data are aligned on start of stimulation. Contralateral staircase saccades are evoked during the period of electrical stimulation (shown by green bar). C and G: vertical eye position data of the right and left eyes during the same time period shows a staircase of upward saccades. B and F: traces of horizontal strabismus angle (left eye position-right eye position) show significant change in alignment (divergent change in monkey H and convergent change in monkey E) during stimulation. D and H: there is little change in vertical misalignment due to stimulation in monkey H and ~4° change in monkey E. In A–H, positive values indicate rightward or upward eye positions and negative values indicates leftward or downward eye positions.

Figure 3, E–H, shows data from another stimulation site in the left SC of monkey E wherein electrical stimulation evoked a convergent change in misalignment of ~2–6° (Fig. 3F, less exotropic) and also a significant change in vertical misalignment of ~2–4° (Fig. 3H). Radial amplitude of the electrically evoked saccade at this stimulation site was ~1.5°.

Stimulation at the different SC sites resulted in a varied amount of change in horizontal and vertical strabismus angle and these data are summarized in Fig. 4 (data are arranged in increasing order of convergence obtained during left eye viewing). During left eye viewing (Fig. 4A), 36/48 sites evoked a divergent change in strabismus angle (shift toward more exotropia or less esotropia; mean divergent change = 5.7°) and 12/48 sites evoked a convergent change in strabismus angle (shift toward less exotropia or more esotropia; mean convergent change = 1.8°). Similarly during right eye viewing (Fig. 4C), 27/44 sites evoked a divergent change and 17/44 evoked a convergent change (mean divergent change = 6.2°; mean convergent change = 1.4°). The amount of change in vertical strabismus angle (Fig. 4, B and D) was generally small across all sites. Magnitude (absolute value) of change in vertical strabismus angle during left eye view was 2.3° and during right eye view was 4.3°. Although in general electrical stimulation evoked a consistent divergent or convergent change during both right and left eye viewing, at 10 sites stimulation evoked a divergent change while one eye was viewing and a convergent change while the other eye was viewing. However, the magnitude of the convergent/divergent changes at these sites was generally small (mean divergent change across 10 sites = 2.6°; mean convergent change = 1.5°).

Fig. 4.

Change in horizontal and vertical strabismus angle (A and B: left eye viewing; C and D: right eye viewing) as a result of electrical stimulation. Data are arranged in increasing order of convergence obtained during left eye viewing and the same stimulation site index is used for A–D. Strabismus angle is calculated as the difference between left and right eye positions.

Analysis of saccade amplitude and direction disconjugacy.

In an attempt to further understand the effect of SC electrical stimulation on strabismus angle, we asked whether the changes in strabismus angle were due to disconjugate saccades (saccade amplitude disconjugacy or saccade direction disconjugacy). To analyze saccade amplitude disconjugacy, we calculated the mean radial amplitude of the electrically evoked saccade in the right and left eyes and these data are plotted in Fig. 5. The slope of the regression line of mean radial amplitude in right vs. left eyes during either eye viewing was 1.01 and statistical testing using a paired t-test showed radial amplitude of saccades was not significantly different in the two eyes (P > 0.05).

Fig. 5.

Correlation between mean radial amplitude of saccade evoked in the right and left eyes during both right (RE) and left eye (LE) viewing conditions (regression line slope = 1.01; r² = 0.90). Data from individual animals are color-coded using the same scheme as in Fig. 4.

Potentially, change in eye misalignment due to electrical stimulation could also be due to differing directions of left and right eye movements. Figure 6 is a polar plot showing the electrically evoked eye movements in the right and left eyes at two different SC sites. In each case, it is apparent that the eye movement directions in the two eyes are indeed different. Some of this eye movement direction difference could be due to a difference in the saccade direction. Saccade direction differences between right and left eyes are summarized in Fig. 7. The data show that saccade direction differences could be either positive or negative and at most sites the magnitude of the difference in saccade direction is <20°. In summary, analysis of saccade parameters in the two eyes showed that differences in saccade vector direction but not saccade vector amplitude could potentially contribute toward change in strabismus angle due to electrical stimulation.

Fig. 6.

Polar plots showing electrically evoked saccades in right eye (red traces) and left eye (blue traces) at 2 different sites in monkey E (A) and monkey H (B). In A, monkey is viewing with the right eye, and in B, monkey is viewing with the left eye. Plots show that the evoked movements are in different directions. A: RE direction ~25°, LE direction ~9°; B: RE direction ~27°, LE direction ~38°.

Fig. 7.

Summary of directional difference between left and right eyes electrically evoked saccades in the 3 monkeys obtained during left eye viewing conditions. Stimulation site index and color-coding scheme is the same as that used in Fig. 4.

Disconjugate saccadic eye movements do not fully account for the change in horizontal and vertical strabismus angle due to electrical stimulation.

We asked whether the observed change in horizontal and vertical strabismus angle was simply the cumulative consequence of several disconjugate saccades (due to saccade direction difference) that were evoked during the 500 ms of electrical stimulation. To perform this analysis, for each stimulation site, we calculated the contribution of horizontal (or vertical) saccade disconjugacy by multiplying the mean difference in horizontal (or vertical) saccade component of the two eyes due to a single electrically evoked saccade at that site with the number of saccades evoked in 500 ms at the same site. These data are plotted in Fig. 8, A and B, against the total change in misalignment over 500 ms at each stimulation site. The slope of the regression line in Fig. 8A (horizontal misalignment) is 0.49, which means that only ~50% of the change in horizontal strabismus angle upon electrical stimulation is due to horizontal saccade disconjugacy. The regression line slope in Fig. 8B (vertical misalignment) is 0.83, which means that most of the change in vertical strabismus angle due to stimulation is a result of vertical component of saccade disconjugacy.

Fig. 8.

A and B: contribution of horizontal (A) and vertical (B) saccadic disconjugacy (estimated from cumulative sum of disconjugacy due to each saccade in staircase) toward total change in horizontal or vertical strabismus angle over a 500-ms stimulation period at all stimulation sites during left eye viewing. C and D: contribution of horizontal (C) and vertical (D) slow postsaccadic movement disconjugacy toward total change in horizontal or vertical strabismus angle over a 500-ms stimulation period. These data show that horizontal saccade disconjugacy and slow postsaccadic movement disconjugacy accounts for ~50% each of total change in horizontal strabismus angle and that vertical saccade disconjugacy accounts for ~85% of total change in vertical strabismus angle due to electrical stimulation. Color coding scheme is same as in Fig. 4.

Disconjugacy in slow postsaccadic movements significantly contributes to change in misalignment during electrical stimulation.

The change in misalignment due to electrical stimulation is fundamentally a sum of disconjugacy due to saccades and the disconjugacy due to slow postsaccadic movement. Examination of the raw data plots in Fig. 3, indeed show significant postsaccadic movement disconjugacy. In Fig. 8, C and D, we plot the change in horizontal or vertical misalignment due to slow postsaccadic movement disconjugacy (derived as the difference between total change in misalignment and the contribution from saccade disconjugacy) against the total change in misalignment. These data highlight the significant contribution of slow postsaccadic movement disconjugacy toward the total change in misalignment. Fundamentally this analysis shows that during electrical stimulation, ~50% of change in horizontal misalignment is due to the slow postsaccadic movement disconjugacy. For vertical misalignment, there was no correlation between the change in vertical misalignment and the amount of postsaccadic movement disconjugacy (Fig. 8D), which might be expected if saccadic disconjugacy (Fig. 8B) can explain almost all of the change in vertical misalignment due to electrical stimulation.

Rostral-caudal influence on saccade disconjugacy and postsaccadic movement disconjugacy.

Since vergence neurons were previously found within the rostral part of SC, we wondered whether there might be different influence on eye misalignment upon stimulation at rostral vs. caudal SC sites. To perform this analysis, we arbitrarily divided our data into stimulation sites that yielded saccades of radial amplitude <5° (36 rostral sites during left eye viewing) and sites that yielded saccades of radial amplitude >5° (12 caudal sites: site indexes 1, 3, 5, 6, 10, 11, 15, 17, 31, 38, 45, and 47). From Fig. 4, it is apparent that stimulation at either rostral or caudal sites can induce significant changes in horizontal or vertical strabismus angle.

Data in Fig. 8 show the contributions of saccade disconjugacy and postsaccadic movement disconjugacy toward change in misalignment at all stimulation sites. To establish possible rostro-caudal differences, we recalculated and plotted the contribution of saccade disconjugacy (Fig. 9, A and B) and postsaccadic movement disconjugacy (Fig. 9, C and D) separately for rostral and caudal sites. Note that the data points in Fig. 9, A–D, are identical to those in Fig. 8, A–D. In Fig. 9, A–D, black line fits are for rostral sites data and light gray line fits are for caudal sites data. Line fits are provided only when there is a correlation between x-axis and y-axis values, i.e., if the slope values of the regression are statistically significant. Slope values in Fig. 9A suggest that contribution of saccade disconjugacy toward change in horizontal misalignment is slightly greater at caudal sites (~54%) than rostral sites (~43%). Similarly slope values in Fig. 9C suggest that contribution of postsaccadic movement disconjugacy is larger at rostral sites (~57%) than at caudal sites (~46%). However a statistical comparison of slopes of rostral and caudal sites regression lines within Fig. 9, A and C, yielded no significant difference (t-test; P = 0.21). A similar consideration of vertical misalignment and rostro-caudal influences showed that although there was a linear relationship between saccade disconjugacy or postsaccadic movement disconjugacy and change in vertical misalignment for rostral sites data, there was no such relationship for the caudal site data. For the rostral sites, saccade disconjugacy and postsaccadic movement disconjugacy contribute ~50% each of the total change in vertical misalignment. Therefore, although the analysis of all sites (Fig. 8, B and D) suggested that vertical saccade disconjugacy is the primary contributor toward vertical change in misalignment, there might be some differences between rostral and caudal stimulation sites (Fig. 9, B and D).

Fig. 9.

Contribution of saccade disconjugacy (A and B) and postsaccadic movement disconjugacy (C and D) toward change in horizontal (A and C) and vertical misalignment (B and D) at rostral (black circles; black regression line) and caudal (gray stars; gray regression line) sites. Slopes of regression lines indicate relative contribution of each quantity toward change in horizontal or vertical misalignment. At caudal sites, there was no relationship between vertical saccade disconjugacy or vertical postsaccadic movement and vertical change in misalignment.

Could the differences in saccade direction and changes in eye misalignment be unrelated to electrical stimulation of the colliculus?

Eye movement disconjugacies have been described before in strabismic monkeys (Das et al. 2005; Fu et al. 2007; Walton et al. 2014b). It is possible that the saccade direction and eye misalignment changes that we observed are not specifically related to involvement of the colliculus and saccadic circuits but rather a function of other mechanisms, for example, disrupted downstream vergence circuits or possibly orbital factors that alter pulling directions of extraocular muscles. To assess the effect of these factors, we performed two control experiments. The first control experiment involved acquiring data during visually guided saccades that were matched in amplitude and direction with the previously collected electrically evoked saccades in the three monkeys. Although target movement was matched to the previously acquired electrically evoked saccade amplitude and direction, the actual amplitude and direction of the visually guided saccade were slightly different, perhaps due to small drift of the eye before saccade onset. Therefore, we paired visually guided saccades with electrically evoked saccades that were within 1° of amplitude and within 30° of saccade direction. For rostral sites, this criterion resulted in matching of one visually guided saccade with more than one electrically evoked saccade site. Figure 10 summarizes the comparisons between electrically evoked saccades (ES) and visually guided saccades (VGS) of matched amplitude and direction. Figure 10A compares difference in saccade direction of the right and left eyes between ES and VGS and shows no significant difference between these two quantities (paired signed-rank test for nonnormal data; P > 0.05; data points distributed around the unity line). The difference in right eye and left eye radial amplitude (Fig. 10B) was also not significantly different between ES and VGS (paired signed rank test; P > 0.05). To examine effects on misalignment (Fig. 10, C and D), we calculated a change in horizontal or vertical misalignment following a visually guided saccade (calculated at 200 ms after saccade offset) and compared this value to change in misalignment after a single electrically evoked saccade (total change in misalignment due to 500-ms stimulation divided by number of electrically evoked saccades in each staircase). Note that the change in misalignment in Fig. 10, C and D, includes both the effect of saccade disconjugacy and also postsaccadic movement disconjugacy for both ES and VGS. Change in horizontal misalignment due to ES was significantly different than VGS (Fig. 10C; paired signed rank test; P < 0.05). Note in Fig. 10C that although the data points appear above the unity line, the magnitude of change in horizontal misalignment is actually larger for ES as a majority of the values are more negative for ES than VGS. Change in vertical misalignment (Fig. 10D) due to ES was not significantly different from that during VGS (paired signed rank test; P > 0.05). In summary, these control data suggest that saccade disconjugacy (saccade amplitude and saccade direction) is not significantly different between ES and VGS but change in horizontal misalignment is indeed larger during ES than VGS. Thus electrical stimulation within the SC may indeed activate special circuits or cell populations, possibly vergence related, that causes changes in misalignment (especially horizontal misalignment) that are larger than during VGS.

Fig. 10.

Comparison of electrically evoked saccades (ES) and visually guided saccades (VGS) of similar amplitude and direction in all animals. A: comparison of difference in saccade direction of right and left eyes during ES or VGS shows no significant differences (points distributed around the dotted unity line). B: difference in radial saccade amplitude of right and left eyes during ES or VGS showed no significant difference. C: comparison of the change in horizontal strabismus angle due to a single ES or VGS shows larger (more negative; majority data points above the unity line) effects due to electrical stimulation than visually-guided movements. D: Comparison of the change in vertical strabismus angle due to a single ES or VGS shows no significant difference between electrical stimulation and visually guided movements.

In a second control, we analyzed eye movement direction differences between the two eyes during smooth-pursuit eye movements in the horizontal and vertical directions. Similar direction differences during smooth pursuit and electrically evoked saccades would suggest that areas downstream of the colliculus, perhaps related to generation of pattern strabismus, are responsible for the stimulation effects observed in this study. Direction differences during smooth pursuit were calculated from the desaccaded eye movements shown in Fig. 1. In each monkey, there was indeed difference in smooth-pursuit direction between the two eyes but the difference was constant and relatively small (right eye viewing: monkey E = ~5°, monkey H = ~4°, monkey L = ~6°; left eye viewing: monkey E = ~8°, monkey H = ~4°, monkey L = ~5°). In contrast, saccade direction differences and misalignment changes due to electrical stimulation of the SC are either positive or negative as shown in Figs. 4 and 7 and are larger in magnitude. To illustrate these differences more clearly, Fig. 11 plots the position of the right eye vs. the left eye for both smooth pursuit (after desaccading) and a single electrically evoked saccade staircase in monkey H. The slope of the regression line is 0.94 for smooth pursuit and 2.13 for the electrically evoked saccade train. Therefore, this second set of control data shows that direction differences and misalignment changes due to SC electrical stimulation are significantly greater than might be predicted due to smooth-pursuit disconjugacy and are therefore likely a product of activation of specific circuits or cell types within the SC.

Fig. 11.

Comparing horizontal component of right and left eye movement during horizontal visually guided smooth pursuit (red trace; solid line regression fit) and electrically evoked saccades (blue trace; dotted line regression fit) in monkey H. Disconjugacy and change in alignment is significantly larger during electrical stimulation (saccade slope = 2.13; smooth-pursuit slope = 0.94).

DISCUSSION

In this study, we have used the strategy of electrical stimulation to study the role of the SC in driving eye misalignment. The main findings in our study were 1) electrical stimulation of the SC induced significant changes in eye misalignment that could be either divergent or convergent depending on stimulation site. 2) The electrically evoked saccades in the two eyes were of similar radial amplitude but differed in saccade direction. 3) The change in misalignment evoked by electrical stimulation was due to both disconjugate saccades and disconjugate postsaccadic movement. These results advance our understanding of strabismus in primates in that it implicates the SC as being part of the circuit that serves as a neural substrate for steady-state ocular misalignment. Additionally, the mechanism by which the SC participates in strabismus is consistent with our previously proposed idea of a disrupted vergence circuit. Below we discuss the implications of these findings in the context of saccadic circuitry and maintenance of the state of misalignment in strabismus.

Disconjugate saccade behavior following electrical stimulation of the SC.

Disconjugate saccade behavior was primarily due to saccade direction differences. The differences in saccade directions in the two eyes was quite varied but most frequently were under 20°. Differences in saccade direction in strabismic monkeys have been observed before. While examining oblique saccade disconjugacy, Walton et al. (2014b) observed that for the most part only saccade direction was affected. They also showed that electrical stimulation of the paramedian pontine reticular formation (PPRF), a structure that receives projections from the SC, often evoked saccades of varying directions (Walton et al. 2013) and that direction preference of PPRF was disrupted in comparison to the normal (Walton and Mustari 2015). In a recent study, Fleuriet et al. (2016) also reported horizontal and vertical saccade disconjugacies following SC stimulation in strabismic monkeys and proposed that stimulation in strabismic monkeys activates different desired displacements for the two eyes. In other words, the saccade disconjugacy observed with SC stimulation is due to separate right eye and left eye maps within the colliculus. However, Economides et al. (2016) found some amplitude and direction differences upon electrical stimulation of caudal colliculus in a cohort of surgically induced strabismic monkeys but concluded that tectal maps were not different in the two eyes.

Our analysis showed that only part of the change in horizontal misalignment (~50%) is due to disconjugate horizontal saccades (Figs. 6–9). It is unlikely that disconjugacy is driven by initial eye position differences between the two eyes because two of the animals in the study (monkeys L and H) presented with small misalignment but still showed significant changes in misalignment following stimulation. In their study of strabismic monkeys, Fleuriet et al. (2016) also did not find a significant effect of eye position upon saccade disconjugacy due to SC stimulation. Also, the small amount of pattern strabismus in our monkeys (Fig. 1) was insufficient to account for horizontal or vertical disconjugacy during stimulation. However, the fact that saccade disconjugacy during visually guided saccades was not significantly different from saccade disconjugacy during electrically evoked saccades (Fig. 10, A and B) suggests that there was nothing special about electrical stimulation vis-à-vis saccade disconjugacy. We suggest that disruption in saccadic circuitry either within or downstream from the SC was responsible for saccadic disconjugacy.

Role of the SC in eye misalignment: postsaccadic movement disconjugacy.

In our study, ~50% of the change in horizontal misalignment is due to differential postsaccadic movement in the two eyes, i.e., a slow vergence drift. It is unlikely that these postsaccadic movements are due to faulty neural integration (postsaccadic drift due to pulse-step mismatch) because one could reasonably expect that the neural integrator functions similarly during smooth pursuit, visually guided saccades, and electrically evoked saccades, but the changes in horizontal misalignment are greater during electrical stimulation than during either pursuit or visually guided saccades (Figs. 10C and 11). Almost none of the change in vertical misalignment evoked by electrical stimulation is due to vertical postsaccadic movement disconjugacy (Fig. 8D). It should however be noted that vertical misalignment is usually significantly smaller than horizontal misalignment and that there were differences identified between rostral and caudal stimulation sites (discussed later, Fig. 9, B and D).

The afferent and efferent anatomical projections of the SC are diverse (May 2006); focusing on potential “vergence” pathways, the SC receives extensive projections from cortical areas encoding disparity information including LIP and frontal eye field (FEF). It also receives projections from the cFN and the PIN (May et al. 1990), areas that we showed are related to the strabismus state (Joshi and Das 2013). There are major reciprocal connections between the SC and the central mesencephalic reticular formation (cMRF) (lateral to OMN) and studies have identified connections between the cMRF and the SOA (related to vergence), Edinger-Westphal nucleus (related to accommodation), oculomotor, and abducens nuclei (Bohlen et al. 2016; Büttner-Ennever et al. 2001; May et al. 2015; Ugolini et al. 2006). These structures could be the basis for a ‘slow vergence’ circuit involved in static alignment and potentially disrupted in strabismus. As part of the horizontal saccade circuitry, the SC and the cMRF connect to the PPRF and electrical stimulation and neural recording within the PPRF in strabismic monkeys have suggested that saccade disconjugacy in strabismus may be a consequence of disruption within this circuit (Walton and Mustari 2015; Walton et al. 2013). It is possible that the circuit that mediates saccade disconjugacy is separate from the aforementioned slow vergence circuit. The difference between electrically evoked saccades and similar amplitude and direction visually guided saccades (Fig. 10) also suggests that specific vergence circuits or cell population within the SC that affect alignment are being recruited by stimulation.

Role of the SC in eye misalignment: rostral vs. caudal influences.

Ohtsuka et al. (2002) reported a patient with a focal lesion in the rSC, who showed a deficit in convergence and a static exotropia. In the normal monkey, Van Horn et al. (2013) found vergence cells while specifically targeting the rSC and also found that electrical stimulation induced divergence when looking at near targets. It was therefore possible that, in the strabismic monkey, the changes in misalignment due to stimulation might have been restricted to rostral areas in the SC. However, we found that stimulation in the caudal colliculus (defined arbitrarily as sites yielding radial saccade amplitude >5°) also induced changes in misalignment (Figs. 3 and 9). Although their studies were primarily focused on investigating saccadic tectal maps in strabismus, examination of Fig. 1 from Fleuriet et al. (2016) and Fig. 2 from Economides et al. (2016) also suggests that stimulation at some caudal colliculus sites can result in disconjugate postsaccadic movements. Comparing relative contributions of saccade disconjugacy and slow postsaccadic movement disconjugacy toward total change in horizontal misalignment at rostral and caudal sites yielded no significant differences (Fig. 9; ~50% each), although the trend was for increased contribution of slow postsaccadic movement in rostral SC and increased contribution of saccadic disconjugacy in caudal SC. Evidence points to a continuum along the rostro-caudal axis for saccade amplitude representation and similarly horizontal vergence cells may also be spread throughout the SC.

Relative contributions of vertical saccade disconjugacy and vertical postsaccadic movement disconjugacy to change in vertical misalignment were similar to each other (~50% each) at rostral sites. However, at caudal stimulation sites, there was no correlation between vertical saccade disconjugacy or vertical postsaccadic movement disconjugacy and vertical change in misalignment for stimulation suggesting that there might be some sort of saturation effect on vertical disconjugacy components upon stimulation at caudal SC. Identifying neural substrates for vertical vergence might help interpreting this data.

GRANTS

This work was supported by National Eye Institute Grant R01-EY-026568 and University of Houston College of Optometry Core Grant P30-EY-07551.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

S.U. and V.E.D. conceived and designed research; S.U. and H.M. performed experiments; S.U. analyzed data; S.U., H.M., and V.E.D. interpreted results of experiments; S.U. prepared figures; S.U. and V.E.D. drafted manuscript; S.U., H.M., and V.E.D. edited and revised manuscript; S.U., H.M., and V.E.D. approved final version of manuscript.