JOURNALOF NEUROPHYSIOLOGY Vol. 68. No. 4, October 1992. Printed

Dynamics and Efficacy of Saccade-Facilitated Vergence Eye Movements in Monkeys J. S. MAXWELL

AND

W. M. KING

Department ofPhysiology and Centerfir Rochester, New York 14642 SUMMARY

AND

Visual Science, University ofRochester Medical

CONCLUSIONS

INTRODUCTION

1. Four macaque monkeys were trained to fixate visual targets. Eye movements were recorded binocularly using the search coil technique. Saccades, vergence movements, and combinations of the two were elicited by training the monkeys to alternate the gaze between real visual targets that differed in viewing distance and eccentricity with respect to the monkeys’ heads. 2. When they shifted the gaze between targets that were at different viewing distances, the monkeys made vergence eye movements. For targets placed along the midsagittal plane, the monkeys often made binocularly symmetric vergence movements. The peak speed of symmetric divergence movements increased linearly with vergence amplitude by 5.7 deg/ s per degree of vergence. The peak speed of symmetric convergence movements increased linearly with vergence amplitude by 7.9 deg/s per degree of vergence. 3. For gaze shifts between targets placed eccentrically with respect to the midsagittal plane and at different viewing distances, the monkeys made saccades in combination with vergence eye movements. When a saccade occurred during a vergence movement, peak vergence eye speed increased abruptly and reached a peak that was proportional to the speed of the saccade. For four monkeys, peak divergence speed ranged from 242 to 3 15 deg/s and peak convergence speed ranged from 257 to 340 deg/s for 16-deg vergence and 20-deg saccadic eye movements. 4. For gaze shifts between far targets at the same viewing distance but different eccentricities, saccadic eye movements were transiently disjunctive even though there was no vergence requirement. Initially, the eyes diverged and then converged to restore fixation to the correct depth plane. Divergence was followed by convergence regardless of the direction of the saccade. 5. The presence of transient saccade-related disjunctive eye movements suggested that the abrupt increase in peak vergence speed during combined saccadic and vergence eye movements was produced by the linear addition of a vergence eye movement and the saccade-related transients. Consistent with this hypothesis, the rate of change in peak vergence speed during various-sized saccades between far targets (no vergence required) was similar to the rate of change in peak vergence speed during combined saccadic and vergence movements. However, the peak vergence speeds during the combined movements were higher than predicted by the linear addition hypothesis, suggesting the presence of an additional mechanism. 6. The saccade-related increase in peak vergence speed during combined saccades and vergences led to a significant decrease in the amount of time required to complete vergence movements. On average, divergence trials with 20-deg saccades were 39% shorter than comparable vergence trials not containing saccades. Convergence trials showed less of an effect and were 22% shorter than comparable saccade-free trials.

1248

Center,

When gaze is shifted between a near and far target in the midsagittal plane, the eyes must rotate in opposite directions (disjunctively) in order to visually acquire the new target and to avoid double vision. When gaze is shifted between two far targets at optical infinity, the eyes must rotate in the same direction through the same angle (conjugately). Disjunctive and conjugate eye movements are usually considered to be produced by independent neural subsystems.Indeed, the oculomotor system is often divided heuristically into five subsystems (Dodge 1902; Robinson 1981). According to this classification, smooth pursuit, saccadic, vestibuloocular, and optokinetic eye movements are produced by conjugate subsystemsand vergence is the sole disjunctive subsystem. The vergence subsystem was thought to operate independently from the conjugate oculomotor subsystems (Rashbass and Westheimer 196 lb), and vergence eye movements were characterized as “sluggish,” with maximum eye speedsof only a few tens of degreesper second (Rashbass and Westheimer 196la). Recent evidence, however, has challenged this traditional point of view. For example, in monkeys the vestibulo-ocular reflex (VOR) appears to generate rapid disjunctive eye movements to stabilize the retinal images of near targets on the foveae (nasooccipital VOR, Paige and Tomko 199 1: linear and angular VOR, Snyder and King 1992; Viirre et al. 1986). Similarly, smooth pursuit of moving nearby targets requires each eye to move at a different speed. It is not known where these disjunctive responsesarise, but for the VOR they anticipate changes in vergence eye position (Snyder et al. 1992), suggesting a central rather than a peripheral control mechanism. Vergence eye movements frequently occur in combination with saccadic eye movements when gaze is shifted between targets that differ in viewing distance and eccentricity with respect to the head. Several studies in humans have shown that saccadesduring vergence movements have different amplitudes in each eye (Enright 1984, 1986; Erkelens et al. 1989; Kenyon et al. 1980; Ono and Nakamizo 1978; Ono et al. 1978), and the difference in amplitude appears to be greater than would be predicted by a linear summation of the vergence and saccadic components in each eye. Erkelens et al. ( 1989) recently described a dramatic enhancement of vergence eye speed( 5200 deg/s) in human subjects when the vergence movement occurred in combination with a large saccade.For their subjects, 595%

0022-3077192 $2.00 Copyright 0 1992 The American Physiological Society

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 12, 2019.

VERGENCE

AND

SACCADE

of an 1 I-deg vergence movement was accomplished during the time course of a 4%deg saccade. They concluded that the saccadic system programs disjunctive saccades to create fast changes in vergence. To further examine the interaction of saccades and vergence eye movements, we presented monkeys with visual targets that required them to execute coordinated saccadic and vergence eye movements. We sought answers to the following questions: 1) Do saccades and vergence interact in monkeys as they do in humans? Cumming and Judge ( 1986) briefly reported high vergence speeds associated with saccades in rhesus monkeys, but did not systematically study this phenomenon. 2) Do enhanced vergence speeds during saccades shorten the overall time course of vergence movements? Past studies in humans (Enright 1984; Erkelens et al. 1989) calculated the percentage of the overall vergence movement that occurred during a saccade (intrasaccadic vergence) as a measure of the facilitation of vergence by saccades. We discuss why this procedure can be misleading, and we present an alternative analysis. 3) Are vergence movements centrally programmed by the saccadic system, as suggested by Erkelens et al.? We methodically examined combinations of vergence with varioussized saccades and conclude that some of the increase in vergence speed observed during saccades is not centrally programmed by the saccadic system. 4) Do saccades and vergence movements add linearly during combined saccade / vergence movements? In previous studies, vergence movements that occurred during saccade/ vergence interactions were compared with the sum of symmetric vergence movements and conjugate saccades. Most saccades, however, are not perfectly conjugate but contain a transient disjunctive component (Collewijn et al. 1988). We compared transient disjunctive eye movements during saccades between eccentric targets at the same viewing distance (no vergence required) to those occurring during trials between eccentric targets at various viewing distances (saccade plus vergence required). A preliminary report of this research has been published as an abstract (Maxwell and King 1990). METHODS

Subjects Three rhesus monkeys (Macaca mulatta, designated M 1, M2, and M3) and one pigtail macaque (Macaca nemestrina,designated M4) were used to study the interaction of vergence movements and saccades. All four monkeys had search coils (Robinson 1963) implanted in both eyes using the method of Judge et al. ( 1980). A stainless steel socket attached to the top of each monkey’s skull was used to stabilize its head during recording sessions. Surgeries were performed under aseptic conditions using gas anesthesia (isoflourane). An analgesic (Tylenol) was given postoperatively. Animal care and surgical procedures were in full accordance with the “Guiding Principles for Research Involving Animals and Human Beings” approved by the Council of the American Physiological Society.

Training Each of the monkeys had at least several months of experience tracking visual targets before the start of these experiments. They

INTERACTIONS

1249

were trained to visually acquire and fixate red light-emitting diodes (LEDs) and a red laser-generated target spot on a tangent screen. To obtain a reward (a drop of juice), the LED or laser target had to be fixated by both eyes within a 2.0-deg window until the target was extinguished. A trial was aborted if the monkey did not acquire the target within 500-750 ms. This time was empirically chosen to compel the monkey to work diligently but still allow it enough time to complete trials easily.

Paradigms The purpose of these experiments was to examine the interaction between saccades and vergence, and in particular to determine the degree to which centrally programmed disjunctive saccades might contribute to vergence changes. We required monkeys to change fixation between two visual targets that differed in horizontal eccentricity with respect to the monkey’s head and viewing distance with respect to a point centered between the two eyes, so that a vergence movement and saccade were required to fixate the new target. The monkeys viewed LEDs or laser targets in a room illuminated by amber safelights. The LEDs, their cowlings, the tangent screen, and other objects in the room were clearly visible and provided multiple cues for target distance. The monkeys alternately fixated near and far targets. The nearest target was placed - 10 cm in front of the eye and required a vergence angle (the angle made by the lines of sight of the 2 eyes) of - 16 deg. The exact vergence angle depended on the monkeys’ intraocular distances and the target’s location. LEDs were also placed at distances that required vergence changes of - 12, 8, and 4 deg. Another LED (far target) was 220 cm from the front of the eyes and required a vergence angle of -0.8 deg. The laser spot was also used as a far visual target and required a vergence angle of - 1.8 deg. Figure 1 shows the arrangement of the targets for three paradigms. In Symmetric Vergence trials (Fig. 1A), the near and far LEDs were aligned in the monkey’s midsagittal plane. In this configuration, there was no stimulus for a conjugate horizontal eye movement, and the monkeys could change fixation using a pure vergence movement. In Saccade trials (Fig. 1 B), the monkeys shifted their gaze between two far targets. In this configuration, there was no stimulus for a vergence eye movement, and the monkeys could change fixation with a pure saccade. In Saccade + Vergence trials (Fig. 1C), the monkeys alternated fixation between the near LED and the laser spot. The laser spot was offset horizontally from the midsagittal plane either 10 or 20 deg so that fixation of the spot required both a saccade and a vergence eye movement. There was no attempt to control vertical movement of the eyes during the gaze shift, but during the fixation interval, vertical eye position was required to be 52.0 deg of the target.

Data acquisition The eye coil system was calibrated at the beginning of each recording session by having the monkey fixate far LEDs at known eccentricities. Separate windows specifying acceptable eye position errors were calculated for each eye because the lines of sight subtended different angles when the eyes were converged. A computer program calculated the required eye position on the basis of empirically determined values of target distance, target eccentricity, and intraocular distance (see Snyder and King 1992 for details). This procedure allowed the fixation of near targets to be tightly controlled. Different paradigms were used to test the monkeys as follows: Symmetric Vergence trials requiring 16, 12, 8, or 4 deg of vergence; Saccade trials requiring 5- to 20-deg saccades with no vergence requirement; and Saccade + Vergence trials requiring a 16deg vergence movement combined with a lo- or 20-deg saccade.

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 12, 2019.

1250

J. S. MAXWELL

A SYMMETRIC

VERGENCE

AND W. M. KING

B SACCADE ONLY

0F-F

C SACCADE+VERGENCE

0

FIG. 1. Experimental paradigms used to study saccade and vergence interactions. A : Symmetric Vergence trials elicited vergence movements between a near (N) and a far target (F). B: Saccade Only trials elicited saccades between far (F) equidistant targets. C: Saccade + Vergence trials elicited combinations of saccades and vergence movements.

For one monkey, the Saccade + Vergence paradigm was also run with a vergence requirement of 4 deg. An additional set of data was collected for another monkey using slightly different paradigms. These data were used to compare the increases in peak vergence speed that accompanied saccades during Saccade + Vergence trials with the peak speed of the transient disjunctive eye movements that occurred during Saccade trials. In Saccade trials, the monkey fixated a far central target (a spot of light on the tangent screen). After 1 s, the target light was extinguished and a new target appeared on the screen to the left or right of midline (up to t20 deg in 5-deg increments). The direction and amplitude of the target locations were randomized. In divergence Saccade + Vergence trials, the monkey fixated a central near target (a red LED) that required 9 deg of convergence. The near target was extinguished, and a new far target appeared on the tangent screen (~20 deg left or right of midline in 5-deg increments). In convergence Saccade + Vergence trials, the monkey fixated a far target on the tangent screen (220 degrees left or right of midline in 5-deg increments). This target was extinguished, and the near target LED was lit. Divergence and convergence Saccade + Vergence trials were randomly interleaved. Two hundred Saccade and 200 Saccade + Vergence trials were collected. The data were collected in groups of 10 or 20 trials per paradigm (for example, 10 convergent Symmetric Vergence trials followed by 10 divergent Symmetric Vergence trials) for all the paradigms, and the entire protocol was repeated until 240 trials per paradigm per monkey was obtained. The session was terminated if at any point the monkey’s performance deteriorated. The experiments were controlled by an LSI 11/23+ computer. Horizontal and vertical eye position for both eyes (filtered at 80 Hz) were sampled at a rate of 1 kHz, displayed on a video monitor, and written onto digital tape for subsequent off-line analysis.

Data analysis Vergence angle was defined as the horizontal difference between right eye position (RE) and left eye position (LE). Conjugate eye position was calculated as the average of the two eye positions, i.e., (LE + RE)/2. By these definitions, unequal movements of the eyes were considered vergence movements and equal movements were considered conjugate movements, regardless of how they were generated by the CNS.

RE, LE, vergence angle, and conjugate eye position were digitally differentiated. The resulting eye speeds were smoothed with a low-pass filter (transition band 60-90 Hz), which did not appreciably change the amplitude of the speeds. Saccade onset time was defined when conjugate eye speed first exceeded 15 deg/s; saccade offset time was defined when conjugate eye speed first fell below 15 deg/s. Similarly, vergence onset time was defined when vergence speed first exceeded. 11 deg/ s. Because vergence speed usually approached zero gradually, it was difficult to unambiguously determine when the movement first reached its final value. For this reason, vergence offset time was defined when the vergence angle achieved 95% of its final value. Although this procedure may have underestimated vergence duration, it was systematic and did not affect comparisons between trials. Residual vergence error was defined as the difference between the vergence angle at saccade offset time and the vergence angle at vergence offset time. For this measure, vergence offset time was calculated when eye position reached final value. Statistical differences were computed using a t test or an F test on the coefficients of linear regressions (Snedecor and Cochran 1974). Usually, only one saccade occurred during Symmetric Vergence trials and divergence Saccade + Vergence trials, but multiple saccades occurred regularly during convergence Saccade + Vergence trials (see RESULTS). If a second saccade occurred, it was invariably near the end of the vergence movement and had the appearance of a small corrective saccade. Vergence changes during these second saccades were usually small (< 1 deg) relative to the required vergence change and were not deleted from the trials in which they occurred. RESULTS

The purpose of these experiments was to examine the effect of saccades on vergence movements. We will briefly describe the eye movements recorded with Symmetric Vergence trials during which no saccade was required, with Saccade trials during which no vergence change was required, and last, with Saccade + Vergence trials during which vergence movements and saccades occurred concurrently. Divergence movements during Saccade + Vergence trials had distinctly different time courses than convergence movements during Saccade + Vergence trials. Conse-

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 12, 2019.

VERGENCE

AND

SACCADE

INTERACTIONS

1251

quently, divergence and convergence movements were ana- deg to the right of the far target (C). Figure 3A is comparalyzed separately. ble to Fig. 211 in that it illustrates a trial in which there was Figure 2 shows an example of a 16-deg Symmetric Diverno conjugate component and the vergence movement was gence trial (A ) , a leftward Saccade trial (B), and a 10-deg symmetric. In Fig. 3 B, transient changes in vergence (), p) Saccade + Vergence trial (C). In the Symmetric Vergence occurred during a Saccade trial. Note that the saccade in trial, (Fig. 2A), the trajectories of the two eyes were equal in Fig. 2 B was to the left and the saccade in Fig. 3 B was to the magnitude but opposite in direction, and therefore conjuright, but both saccades initiated a transient divergence folgate eye position (defined as the average of left and right eye lowed by a transient convergence. These saccade-related position) was approximately unchanged throughout the disjunctive transient eye movements will be considered in trial (- - -). more detail below. Similar to the divergence movement ilThe 16-deg leftward saccade illustrated in Fig. 2 B was lustrated in Fig. 2C, in Fig. 3C an abrupt peak in vergence made between two far targets so there was no apparent stimspeed occurred during the saccade. Saccades nearly always ulus for a vergence eye movement. Nevertheless, a transient produced an abrupt increase in peak vergence speed that 1.4-deg divergence with a peak speed of 85 deg/s occurred was in the same direction as the required vergence move(Fig. 2 B, )) . The divergent transient was immediately fol- ment. lowed by a convergent transient (p) that brought the eyes back toward the target. Symmetric Vevgence trials Figure 2C shows an example of a Saccade + Vergence trial in which the far target was 10 deg to the left of the near In Symmetric Vergence trials, the far LED and each of target. A shift in gaze from the near target to the far target the four near LEDs were aligned along the monkey’s midrequired a 16-deg divergence movement and a 10-deg left- sagittal plane. There was no stimulus for a horizontal conward saccade. The left and right eye speeds and amplitudes jugate eye movement with this target arrangement, yet the were clearly not equal during the saccade, but this does not monkeys often made small ( l-3 deg) saccades away from necessarily indicate that the saccades made by the two eyes midline during vergence movements (Fig. 4) followed eiwere not equal. Some inequality in the movements of the ther by a saccade back to the target near the end of the two eyes would have been expected if a binocularly symmetvergence movement (Fig. 4) or by a slow return to the tarric saccade was linearly added to a symmetric vergence get over the time course of the vergence movement. Sacmovement. The vergence movement and the saccade were cades during convergence movements were more common in the same direction for the left eye (and would have than saccades during divergence movements for Symmetric added) but were in the opposite direction for the right eye Vergence trials. For example, for vergence changes of 16 (and would have subtracted). The presence of an abrupt deg, 72% of the convergence trials, but only 17% of the increase in peak vergence speed coincident with the sac- divergence trials, contained saccades. cade, however, suggests that the right and left eye saccades Figure 5, A and B, shows plots of peak vergence speed as were not symmetric (because vergence was defined as the a function of vergence amplitude during Symmetric diverportion of the eye movement that is not conjugate). It gent and convergent gaze shifts, respectively. These plots should be noted that saccades always produced an increase show that vergence speed for symmetric divergence movein peak vergence speed, and high vergence speeds such as ments was linearly related to vergence amplitude, with an those illustrated in Fig. 2C never occurred in the absence of average sensitivity for the four monkeys of 5.7 deg/ s per a saccade. degree of vergence ( range 5.4-6.0 deg ) . Vergence speed was Figure 3 shows an example of a 16-deg Symmetric Con- also linearly related to vergence amplitude for convergence vergence trial (A), a rightward Saccade trial (B), and a movements, with an average sensitivity of 7.9 deg/s per Saccade + Vergence trial in which the near target was 10 degree of vergence (range 6.2- 10.2 deg). Fig. 5 also shows ”

A SYMMETRIC CONJ

VERG

VERGENCE

B SACCADE

ONLY

\

”

“/

”

C SACCADE+VERGENCE

SPEED

SPEED

VERG

POS

REYE

POS

FIG. 2. Examples of divergent eye movements elicited by the 3 experimental paradigms. A: saccade-free Symmetric Divergence movement. B: leftward saccade between far targets. Note the disjunctive transient eye movement that accompanied the saccade. C: typical Saccade + Vergence trial in which a 16-deg divergent movement was combined with a IO-deg leftward saccade. CONJ SPEED, conjugate horizontal eye speed; VERG SPEED, vergence speed; VERG POS, vergence position; LEYE POS, left eye position; CONJ POS, conjugate horizontal eye position; REYE POS, right eye position.

100

msec

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 12, 2019.

1252

J. S. MAXWELL

A SYMMETRIC

VERGENCE

VERG

CONJ

B SACCADE

ONLY

C SACCADE+VERGENCE

SPEED

SPEED

LEYE

POS

REYE

POS

AND W. M. KING

FIG. 3. Examples of convergent eye movements elicited by the 3 experimental paradigms. A: saccade-free Symmetric Convergence movement. B: rightward Saccade Only trial between far targets. A disjunctive transient eye movement accompanied the saccade. C: typical Saccade + Vergence trial in which a 16-deg convergent movement was combined with a IO-deg rightward saccade. The slow postsaccadic tail (-+ ) shown in C was characteristic of Saccade + Convergence trials.

AGiL

that peak vergence speeds during Symmetric Vergence trials were increased when saccades occurred. The open squares plotted in Fig. 5 represent peak vergence speed during trials that contained small saccades (~3 deg), and these points tended to fall above the regression lines for saccadefree vergence movements (crosses and solid lines).

deg/s). For every monkey, however, peak vergence speed increased monotonically with saccade speed for either divergence or convergence eye movements. The correlation coefficients of divergence trials for each monkey were 0.95, 0.99,0.97, and 0.98. The correlation coefficients of convergence trials were 0.77, 0.66, 0.75, and 0.92. All of these correlations were statistically significant (P 5 0.00 1). Relative to vergence trials that did not contain saccades, Saccade -I- Vergence trials peak vergence speeds for trials that contained saccades were The effect of saccades on vergence speed was systematiremarkably high. The average peak divergence speed meacally examined using Saccade + Vergence trials that re- sured for the four monkeys was 280 deg/s during 20-deg quired the monkey to make gaze shifts that combined sac- Saccade + Vergence trials ( 16-deg vergence change), and cades with 16-deg vergence movements. Figure 6 shows the average peak convergence speed was 298 deg/s during that peak vergence speed was related to the peak speed 20-deg Saccade + Vergence. For trials not containing sac(and, therefore, size) of the accompanying saccade. The cades, however, the average peak speeds were 9 1 deg/s and Symmetric Vergence trials that contained small saccades 120 deg/ s, respectively, for a 16-deg vergence change. were included in Fig. 6 along with the data from the SacThese results demonstrate that saccades occurring during cade + Vergence trials. The vergence requirement for all of a vergence movement can sharply increase peak vergence these trials was constant ( - 16 deg), so one would expect speed. It should not be concluded from these data alone, peak vergence speed to be fairly constant as well ( - 100 however, that saccades are programmed to be asymmetric, or, for that matter, that vergence movements containing saccades have shorter durations than vergence movements without saccades. Comparing Fig. 3,A and C, for example, peak vergence speed is much higher for the Saccade + Vergence trial, but it is not immediately evident that the near target was acquired more rapidly. We examined the effect of saccades on vergence movements in several different ways in the analysis that follows. Figure 7 allows a visual comparison of the effect of saccades on the time course of vergence movements. Ten sequential Symmetric Vergence trials were superimposed on 10 sequential 20-deg Saccade + Vergence trials for one of the monkeys to allow a trial-by-trial comparison. For clarity, vergence onsets were aligned in time. For both divergence movements (Fig. 74 and convergence movements (Fig. 7 B), saccade onset nearly coincided with vergence onset, and peak vergence speed was much faster in trials FIG. 4. Example of a Symmetric Vergence trial containing small saccades away and then back to the target. The final portion of vergence was with saccades. The divergent movements illustrated in Fig. 7A appear to be completed more rapidly when the trials completed during the 2nd saccade. Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 12, 2019.

VERGENCE

AND

SACCADE

1253

INTERACTIONS

B CONVERGENCE

A DIVERGENCE

R 0 0

qo

0

2

4

6

8

IO

12

VERGENCE AMPLITUDE

14

16

0

2

4

6

8

IO

VERGENCE AMPLITUDE

(de@

12

14

16

(de@

FIG. 5. Peak vergence speed was related to vergence amplitude in Symmetric Vergence trials. The plotted points represent data from a single monkey (M3). Open squares: data from Symmetric Vergence trials that contained small saccades. Crosses: data from trials that did not contain saccades. Solid lines are the linear regressions calculated for trials that were saccade-free. Dotted lines are the regressions calculated for comparable data from 3 other monkeys (points not shown for clarity). Trials containing saccades (open squares) had larger peak speeds than trials without saccades (crosses and regression lines). A : divergent eye movements. B: convergent eye movements.

contained saccades, but a visual comparison of convergent Saccade + Vergence and Symmetric Vergence trials is made more difficult by the occurrence of small saccades in many of the convergent Symmetric Vergence trials. To quantitatively evaluate the effect of saccades on the time course of vergence movements, the relationship between saccade speed and vergence duration was examined (Fig. 8). Because the vergence requirement was the same

for all of the trials ( - 16 deg), one would expect vergence duration to be fairly constant if vergence duration were independent of saccade speed. However, there was a significant correlation (P 5 0.001) between saccade speed and vergence duration for divergence movements for all four monkeys (correlation coefficients for the 4 monkeys 0.95, 0.92, 0.40, and 0.88). There was a statistically significant correlation between saccade speed and vergence duration

A DIVERGENCE

B CONVERGENCE

P 200

400

600

800

1000

PEAK SACCADE SPEED (deg/sec)

0

I

200

I

400

I

600

1

800

1000

PEAK SACCADE SPEED (deg/sec)

FIG. 6. Peak vergence speed was related to peak saccade speed during Symmetric Vergence trials (o ) and Saccade + Vergence trials (A: lo-deg Saccade + Vergence trials; +: 20-deg Saccade + Vergence trials). A: divergent movements. B: convergent movements. Plotted points: data from a single monkey. Solid line: regression line through those points. Dashed lines: regression lines from 3 other monkeys.

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 12, 2019.

1254

J. S. MAXWELL A

I3

DIVERGENCE

AND W. M. KING

CONVERGENCE

SYMMETRIC

VERG

FIG. 7. Ten sequential Symmetric Vergence trials (-+, labels) without saccades were superimposed on 10 sequential Saccade + Vergence trials (not labeled) to compare the time courses of vergence eye movements. Divergent movements (A ) often had shorter durations when combined with saccades. Convergent movements (B) had more variable trajectories and durations when combined with saccades.

SPEED

SYMMETRIC

VERG

VERGENCE

VERGENCE

POS SYMMETRIC

VERGENCE

10 deg SYMMETRIC

150 degkec

VERGENCE

100 msec

for convergence movements for three of the four monkeys (correlation coefficients 0.01, 0.74, 0.25, and 0.45). The slopes of the regression lines were -0.34, -0.33, -0.07, and -0.23 for divergence movements and 0.00, -0.27, -0.07, and -0.10 ms deg-’ s-l for convergence movements. The effect of saccades on vergence duration was also tested by comparing the vergence durations of Saccade + Vergence trials containing 20-deg saccades (the largest saccades used in these experiments) with the durations of Symmetric Vergence trials that contained no saccades. The vergence durations of the Saccade + Vergence divergence trials for Ml, M2, M3, and M4 were on average 0.38,0.68,0.88, and 0.52 those of Symmetric Vergence trials not containing saccades. These decreases were all statistically significant (t test, P < 0.0 1). For convergence trials, Saccade + Vergence durations were 0.90, 0.68, and 0.76 those of saccade-free l

l

Symmetric Vergence trials for Ml, M2, and M4, respectively. These decreases were statistically significant for two of the monkeys (M2 and M4). This comparison could not be made for the fourth monkey (M3) because there was only one Symmetric Convergence trial that did not contain a saccade. These tests indicate that the increased peak vergence speed that occurs during saccades reduces the amount of time required to complete a vergence movement, particularly for divergence movements. The preceding analysis indicates that saccades had a smaller and more variable effect on the duration of convergent movements, as well as Figs. 3C and 7 B, suggest that the slowness of convergence during the postsaccadic interval may have been responsible. The residual vergence error (the amount of vergence left to complete after the saccade) was similar for divergent and convergent eye movements.

A DIVERGENCE

B CONVERGENCE +

I

0

200

400

PEAK SACCADE

600

800

SPEED (de&ec)

1000

0

200

I

400

PEAK SACCADE

I

600

I

800

1000

SPEED (deg/sec)

FIG. 8. Vergence duration was related to peak saccade speed for Symmetric Vergence ( q ) and Saccade + Vergence trials ( n, + ) . For the monkey whose data points are illustrated, vergence duration decreased appreciably for divergent movements that contained saccades (A) but not for convergent movements that contained saccades (B) .

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 12, 2019.

VERGENCE

AND

SACCADE

For example, the mean residual vergence errors remaining for IO-deg divergent and convergent Saccade + Vergence trials were 3.1 t 1.7 and 3.4 t 1.7 (SD) deg, respectively. The mean residual vergence errors remaining for 20-deg divergent and convergent Saccade + Vergence trials were 2 . 7 t 2.1 and 4.0 t 1.7 deg, respectively. There was no statistical difference between residual vergence error for convergence and divergence movements for any of the monkeys, implying that the eyes had attained the same proximity to the target by the end of the saccade for both convergent and divergent movements. This finding suggests that the relatively slow postsaccadic vergence speed was responsible for the overall smaller effect of saccades on the duration of convergence movements.

Smaller vergence movements All the Saccade + Vergence trials described thus far required fairly large vergence changes ( - 16 deg). We also tested one of the monkeys on a 20-deg Saccade + Vergence paradigm requiring only 4 deg of vergence. The durations of these vergence movements (calculated using the 95% completion criterion) were dramatically reduced when saccades occurred. Figure 9A shows the average of 10 Symmetric Divergence trials superimposed on the average of 10 Saccade + Vergence trials. The mean peak vergence speed of the Symmetric Divergence trials was 38 t 5 (SD) deg/s, but the mean peak speed of the 20-deg Saccade + Vergence divergence trials was 184 t 14 deg/s. Vergence duration was shortened from 2 18 t 30 to 66 t 11 ms, essentially the same duration as the average 20-deg saccade for these trials (63 t 6 ms). The time courses of convergence trials were different (Fig. 9B). Each convergence trial began with a transient divergence so that the initial vergence change was in the wrong direction. A period of rapid convergence ensued during the saccade, and then vergence speed slowed so that the trajectories of trials with and without saccades overlapped during the final 150 ms. The average peak Symmetric Convergence speed was 30 t 5 deg/s, and the average peak Saccade + Vergence speed was 176 t 28 deg/ s. Despite the increase in peak vergence speed, the duration of convergent movements were statistically the same for trials with and without saccades (257 t 4 1 and 246 t 48 ms, respectively). A

DIVERGENCE

Relation

VERGPOS

ofintrasaccadic

vergence to saccade duration

It has been suggested that the relative contribution of a saccade to vergence depends on the relative sizes of the vergence and versional components (Enright 1984). However, a distinction should be made between the vergence that occurs concurrently with a saccade and the vergence that is produced by the saccade. Large saccades have longer durations than small saccades, so more concurrent vergence movement should occur during a large saccade than during a small saccade simply because of the longer time course of the saccade. Therefore a large amount of intrasaccadic vergence is not evidence for a programmed disjunctive saccadic contribution to vergence. The observation, however, that some of the increase in intrasaccadic vergence could be accounted for by saccade duration does not mean that saccades cannot contribute to vergence changes. In fact, data such as those presented in Fig. 7A indicate that they do. It does mean, however, that the effect of saccade duration should be controlled when calculating the amount of vergence produced by disjunctive saccades. In the following analysis, we assumed that vergence was composed of two components, rapid vergence produced by disjunctive saccades and slow vergence produced by the “classical” vergence system. We estimated the amount of slow vergence that would occur during each saccade by multiplying saccade duration by the peak vergence speed of Symmetric Vergence trials not containing saccades (obtained from the regressions illustrated in Fig. 5). (This calculation assumes that the amount of slow vergence is the same as would have occurred had there been no saccade, which might not be true; see DISCUSSION.) This value was then divided by the total change in vergence amplitude for that trial to compute the fraction of vergence completed during the saccade due to slow vergence. For example, the average 20-deg saccadein this study had a duration of -60 ms. The average 16-deg Symmetric Divergence trial (without a saccade) was characterized by a peak vergence speed of - 90 deg/s. Therefore, on average, 34% of a 16-deg vergence movement would be completed during a saccadelasting 60 ms [ ( 100 x .060 s X 90 deg/ s)/ 16 deg] . These values overestimate the effect of saccade duration because we estimated the slow vergence component using peak ver-

I5 CONVERGENCE

SYMMETRIC

VERGSPEED

1255

INTERACTIONS

VERGENCE

FIG. 9. Saccades significantly effect the time course of small vergence changes. The average of 10 sequential Symmetric Vergence trials (+, labels) is superimposed on the average of 10 sequential Saccade + Vergence trials (not labeled) for relatively small vergence changes ( -3 deg). A : divergence trials. B: convergence trials. Note that during convergence Saccade + Vergence trials, the vergence movements were initially in the wrong direction.

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 12, 2019.

J. S. MAXWELL

AND

W. M.

KING

keys the lines were not parallel, suggesting that only part of the change in intrasaccadic vergence was accounted for by changes in saccade duration. For 20-deg Saccade + Vergence trials, we found that an increase in saccade duration alone accounted for 67% of the increase in intrasaccadic divergence and 83% of the increase for intrasaccadic convergence on average for all monkeys. As mentioned above, these values overestimate the effect of saccade duration, so the results imply that minimally, 33% of the intrasaccadic divergence and 17% of the intrasaccadic convergence during 20-deg Saccade + Vergence trials was mediated by disjunctive saccadic components.

Disjunctive 1

0

I

200 PEAK

I

400 SACCADE

I

600 SPEED

I

800 (deg/sec)

1 1000

FIG. 10. The amount of intrasaccadic vergence increased with peak saccade speed. Squares: Symmetric Vergence trials. Triangles: lo-deg Saccade + Vergence trials. Crosses: 20-deg Saccade + Vergence trials.

gence speed obtained during saccade-free Symmetric Vergence trials. This procedure generated an upper bound rather than a less accurate estimate based on an assumed “average” slow vergence speed during Saccade + Vergence trials. The calculation was performed on the data presented in Fig. 10. The dashed regression line is for the calculated values. The slopes of the regression lines representing the fraction of the total vergence completed during the saccade and the fraction due to slow vergence were similar for these data, suggesting that changes in saccade duration alone could account for the increase in intrasaccadic vergence and saccadic speed. It should be noted that although the increase in intrasaccadic vergence is consistent with the increase in saccade duration, the calculated value is not as high as the empirical value. In addition, for the other mon-

A DIVERGENCE

PEAK SACCADE

SPEED (deg/sec)

Peak vergence speed was a function of peak saccade FIG. 1 1. Vergence trials ( •I ) for divergent (A ) and convergent eye movements movements were combined with saccades of different amplitudes targets required saccades of O-20 deg but no change in vergence.

eye movement

transients during saccades

During Saccade trials there were no stimuli for vergence movements, yet disjunctive transient eye movements accompanied all saccades (Figs. 2 B and 3 B). For these saccades, an initial divergent transient was followed rapidly by a convergent transient that restored fixation to the correct plane of depth. The presence of disjunctive transient eye movements during Saccade trials suggested that the abrupt increase in vergence speed during Saccade + Vergence trials could have been due to the summation of saccade-related disjunctive transients with slow vergence movements. We examined this idea by comparing the peak vergence speeds during the disjunctive transients that occurred during Saccade trials with the peak vergence speeds that occurred during Saccade + Vergence trials. One of the monkeys executed a series of Saccade trials in which the target was stepped from straight ahead to an eccentric position between t_20 deg to the right or left of midline in 5-deg increments. The amplitude of the target step and its direction were randomized. The monkey also executed Saccade + Vergence trials between a near target (an LED requiring 9 deg of convergence) and far targets that were presented as they were for

B CONVERGENCE

PEAK SACCADE

SPEED (deg/sec)

speed during Saccade trials (a) and during Saccade + (B) . During Saccade + Vergence trials, 9-deg vergence (O-20 deg). During Saccade trials, gaze shifts between

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 12, 2019.

VERGENCE

AND

SACCADE

the Saccade trials. Convergence and divergence trials were randomly interleaved. Figure 11 suggests that most of the increase in peak vergence speed observed when saccades occurred during vergence trials was due to the disjunctive transient eye movements that always accompanied saccades. In Fig. 11 A, the peak speeds of divergence transients of Saccade trials (A) and the peak vergence speeds of Saccade + Vergence trials (Cl) were plotted as a function of saccade speed. The slopes of the regression lines for Saccade and Saccade + Vergence trials were equal (P 5 0.00 1) for trials containing leftward saccades (0.2 1 and 0.23, respectively) and for trials containing rightward saccades (0.11 and 0.17, respectively, P 5 0.05). The similarlity of the slopes for Saccade and Saccade + Vergence trials suggests that the relationship between peak saccade speed and peak vergence speed for Saccade + Vergence trials presented in Fig. 6 was due largely to the relationship between saccade speed and the speed of the divergence transient. Whereas the slopes of the two functions in Fig. 11A were the same, the peak vergence speeds for Saccade + Vergence trials were higher than for Saccade trials. This offset occurred partly because the speeds of Saccade + Vergence trials also included a component related to the peak speed of the slow 9-deg vergence change. However, the offset between the two regressions was greater than one would have predicted from the divergence amplitude-speed relationship for this monkey (Fig. 5A). The amplitudespeed relationship would predict a peak vergence speed of 32 deg/s for a 9-deg vergence movement, but the difference in y intercepts of the Saccade and Saccade + Vergence trials regression lines was - 85 deg/ s. This discrepancy might reflect the involvement of a saccade-related mechanism other than that which produced the saccade-related vergence transients. These results should be confirmed, however, using a range of vergence angles. For convergence trials, the slopes of the regression lines

100

msec

FIG. 12. The amplitude of saccade-related transient disjunctive eye movements depended on the relative timing of the saccade with respect to the vergence eye movement. A : when the saccade (- - -: saccade speed) occurred late in the vergence movement (--: vergence speed), the vergence change appeared similar to that observed in Saccade Only trials (e.g., Figs. 2 B and 3 B). B and C: convergent portion of the transient ( upward deflection) decreased in amplitude when the saccade occurred earlier in the movement.

VERG

INTERACTIONS

SPEED

VERGENCE

HORlZ

1257

50 Deg/s

POS

EYE POS

FIG. 13. Superimposed examples of transient disjunctive eye movements associated with 8-deg vertical saccades. Note that horizontal eye position changes co.25 deg during the vertical saccade. No net vergence change results from the saccade-related transient. VERG SPEED, vergence speed; VERGENCE POS, vergence position; HORIZ EYE POS, horizontal eye position; VERTICAL EYE POS, vertical eye position.

of Saccade and Saccade + Vergence trials are not equal (Fig. 11 B), implying that peak vergence speed did not result from linear addition of saccade-related convergence transients with slow vergence. The data suggest either that there was a vergence speed component in addition to the saccade-related convergent transient or that convergent transients during Saccade + Vergence trials were larger than those measured during Saccade trials. In other words, the amplitude of the convergence transient might be larger when a convergent movement was executed. We were unable to test this idea methodically, but a visual inspection of individual trials suggested that when a saccade occurred during a planned divergence movement, the convergent transient was suppressed unless the saccade occurred near the end of the vergence movement. For example, Fig. 12A shows that a convergent transient was present during a divergence movement when the saccade occurred late in the trial. However, the convergent transient was partially suppressed when the saccade occurred earlier in the trial (Fig. 12 B) and was completely suppressed when the saccade occurred during the interval of peak vergence speed (Fig. 12C). In contrast, the initial divergent transient occurred even when it opposed convergence movements. Saccade + Vergence trials always began with a brief period of divergence for three of the four monkeys (as in Fig. 9). For these monkeys, the divergent transient was never fully suppressed. The relative timing of the saccade and vergence movements may be important for these differences. Large saccades usually occurred at the beginning of convergent movements and during the middle of divergent movements. For convergence trials, therefore, the divergent transient occurred early with respect to the onset of the vergence and was essentially unopposed. During divergence trials,

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 12, 2019.

1258

J. S. MAXWELL

the saccades occurred later, so that a portion of the converIsent transient was opposed by the d ivergent eye movement.

Vertical saccades Sharp peaks in horizontal vergence speed were also associated with vertical saccades. For example, Fig. 13 shows 10 trials where the monkey made - 8-deg upward saccades. There were no associated horizo ntal saccades, bu t disjunctive transient eye movements occ urred con sistently in the horizontal eye mo vement reco rds. For these data, there was no significant net change in vergence position. The direction of the vertical saccade-induced disjunctive transients was idiosyncratic (King, unpublished observations). For this monkey, upward saccades were associated with con vergent-divergent transi ent waveforms and downward sactransients cades were associated with d ivergent-convergent (not shown). For other mon keys, these direction s were different. DISCUSSION

Our results, like those of Ono et al. ( 1978)) Enright ( 1984) and Erkelens et al. ( 1989) show that vergence movements are not exclusively generated by an independent vergence subsystem with slow dynamics. The monkeys in the present study exhibited significant increases in vergence speeds when saccades occurred during vergence movements. Peak vergence speeds during binocularly symmetric vergence movements did not usually exceed 100 deg/s, whereas peak vergence speeds during eye movements in which saccades and vergence movements were combined often exceeded 300 deg/s. A saccade was always associated with high vergence speeds, and peak verge&e speed increased as a function of saccade speed (Fig. 6). Our data suggest that the increase in peak vergence speed is partially due to the addition of a disj unctive transient eye movement that norm ally accompanies saccades. , Enhanced peak vergence speeds often translated into faster overall vergence movements. The durati ons of divergent movements were significantly reduced by saccades for all four monkeys (Fgi . 8A ), and the durations of convergent movements were reduced significantly for three of the monkeys (Fig. 8 B). This suggests that the oculomotor system can utilize disjunctive saccades to facilitate vergence movements. It is surprising, perhaps, that vergence movements are not even faster. Complex saccadic m echanisms have evolved to ensure rapid conjugate eye movements, yet shifts in gaze between targets differing in viewing distance and eccentricity are limited in speed by the vergence component. For example, the average duration for a 15deg vergence movemen t combined with a 20-deg saccade for one of the monkeys was 290 ms, whereas the duration of the saccade itself was - 60 ms. There are several reasons why changes in vergence may not be completed within the time course of the saccade. First, most saccades fall short of the target, so that a corrective saccade is usually needed to complete the gaze shift. Given the latency ( - 150 ms) and duration of a corrective saccade, a conjugate change in fixation actually requires -200-250 ms, which is comparable to the duration of a vergence movement plus a saccade. Second, our analysis may have overestimated vergence du ration because it was

AND

W. M.

KING

defined relative to the motor output of the system (i.e., vergence angle) rather than to the sensory input. We operationally defined the end of vergence as the time when it attained 95% of its final steady state value. We do not know how much useful vision occurs during vergence eye movements, especially during the slow post-saccadic tail. Third, accommodation has a slow time course consistent with slow fusional vergence. If vergence were as rapid as saccades, changes in accommodation changes would lag behind limiting useful vision.

Possible art$acts Several possible recording artifacts might cause intrasaccadic vergence transients. First, disjunctive eye movement records would be expected if the gain of one eye coil was tot high. However, a divergent transient would be expected with saccades in one direction and a convergent transient with saccades in the other direction. Saccades in this study initially elicited a divergent movement regardless of the direction of the saccade. Second, the leads from the search coils exited the orbit temporally and might have selectively impeded adduction. We cannot rule this possibility out, but it is highly unlikely because disjunctive saccades (during vergence) in humans have been observed using optical (Ono et al. 1978) and video (Enright 1984) techniques where there were no conceivable external mechanical restrictions on eye movements. In addition, Collewijn et al. ( 1988 ) found disjunctive transient eye movements in human subjects using a coil system in which the leads of the search coils exited nasally. Third, blinks elicit co-contraction of the extraocular muscles (Collewijn et al. 1985; Evinger et al. 1984). The saccade-related disjunctive transients we observed were not caused by overt blinks (closure of the eyelid). Direct observation of one of the monkeys showed that it did not blink during saccades, yet all of its saccades elicited vergence transients.

LX@rences between convergence and divergence There was a conspicuous difference in the trajectories of divergent movements and convergent movements. Typically, divergence speed decreased more rapidly than convergence speed after a saccade. Postsaccadic convergence speed was often constant until interrupted by a corrective saccade that brought the eyes onto target. The differences between convergence and divergence may reflect different neural control mechanisms. A much greater number of convergence-related neurons than divergence-related neurons have been reported in the midbrain (Mays 1984). It is, of course, possible that there was a sampling bias or that divergence neurons were located elsewhere in the brain stem, but the preponderance of convergence-related cells may also reflect different mechanisms for generating divergence and convergence. Passive forces in the orbits resulting from the elasticity of the muscles and orbital fascia tend to keep the eyes centered in the orbits. These forces should assist divergent movements and oppose convergent movements.

Comparison

to human data

The present data are similar to the human data of Erkelens et al. ( 1989), who also used relatively large changes in

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 12, 2019.

VERGENCE

AND

SACCADE

vergence ( - 11 deg) and version (45 deg) in their “symmetric vergence” (Symmetric Vergence) and “asymmetric vergence” (Saccade + Vergence) paradigms. The subjects in their study attained peak vergence speeds of -4 deg/ s per degree of vergence, which included trials with and without small saccades. For comparable trials in monkeys, peak vergence speed increased 5 8 and 12.7 deg/ s per degree of vergence for divergence and convergence Symmetric Vergence trials, respectively. Convergence increased more rapidly, partly because convergent movements more frequently contained saccades than did divergent movements. In Symmetric Vergence trials not containing saccades, peak vergence speed rose 5.7 and 7.9 deg/ s per degree of vergence for divergence and convergence trials, respectively. The highest peak speed given by Erkelens et al. ( 1989) for human subjects during an asymmetric vergence trial was 180 deg/s for a 45-deg saccade combined with an 1 I-deg vergence movement. Monkeys often had peak vergence speeds >300 deg/ s when 20-deg saccades were combined with 15-deg vergence movements. Erkelens et al. ( 1989) reported that, on average, their subjects completed 75% of the required vergence change during a saccade for changes in vergence angle of 11 deg combined with 45-deg saccades when the target was stepped inward (convergence) and 95% when stepped outward (divergence). This is not the same as saying that 75 or 95% of the vergence was mediated by a saccade. At least some of the intrasaccadic change in vergence would have occurred even if the saccade had been symmetric and, therefore, not contributed to vergence. For example, 75% of an 11 -deg vergence movement could be completed in 130 ms (the duration of a 45-deg saccade) if the average (slow) vergence speed was just 63 deg/ s during the saccade. Not all of the intrasaccadic vergence in our study could be accounted for in this way. It was estimated that, maximally, 67% of the change in divergence and 83% of the change in convergence would have occurred in the interval of time spanned by the saccade for a 15-deg vergence movement combined with a 20-deg saccade. This implies that, minimally, 33% of the change in divergence and 17% of the change in convergence was mediated by disjunctive saccades. This is probably an underestimate, because the mean peak vergence speed (for Symmetric Vergence trials without saccades) was used for the calculation and not the average vergence speed during the interval of time over which the saccade occurred. We agree with Erkelens et al. ( 1989) that a significant portion of a vergence movement can be mediated by a saccade-related process. Those authors concluded that the “effect was so large, and the performance of (their) subjects so reliable” that the asymmetric saccades must have been programmed by the oculomotor system. We conclude, however, that some of the added vergence was the result of disjunctive transient eye movements that occur even when a vergence movement is not called for and that are unlikely to be centrally programmed. Disjunctive transient eye movements during saccades Collewijn et al. ( 1988) reported that divergent transient eye movements normally occur in humans during saccades between eccentrically displaced far targets located equidistant from the head. We observed transients in monkeys as well. In monkeys, the divergent transient was followed rap-

INTERACTIONS

1259

idly by a convergent transient that returned fixation to the correct plane of depth. Because the characteristics of these disjunctive transient eye movements were similar to the characteristics of intrasaccadic vergence changes elicited during Saccade + Vergence trials, we hypothesized that some of the increase in peak vergence speed observed with saccades during Saccade + Vergence trials was due to the addition of intrasaccadic disjunctive transients to slow vergence, and our results were consistent with this hypothesis (Fig. 11). Are saccade-related peaks in vergence speed programmed? The fastest way to make a combined vergence eye movement and saccade would be for the saccadic system to correct retinal error independently for each eye. This does not appear to occur. If saccades were programmed independently for each eye, asymmetric vergence movements would be expected to have the same durations as saccades. Whereas smaller vergence movements (~5 deg) were sometimes completed within the time course of a saccade (Fig. 9), larger vergence movements were almost never completed during that short of a period. Even the fastest 15-deg vergence movements took 200-300 ms to complete, whereas saccades of this size took ~50 ms. In addition, during trials in which the near and far targets were aligned with one eye, saccades occurred in both eyes even though a saccade was required only for the nonaligned eye. For these reasons, it seems unlikely that vergence movements are programmed directly by the saccadic system. Our data suggest that some of the change in vergence angle during saccades is mediated by saccade-related disjunctive eye movement transients. The saccade-related transient consisted of a brief divergence followed by convergence and occurred during all saccades, even those between equidistant far targets. It seems unlikely that the transients were programmed by the saccadic system to assist vergence because, first, they occurred when no vergence was required and, second, horizontal transients occurred during vertical saccades (Fig. 13). It is likely that the transients are the result of some feature common to both horizontal and vertical systems. For example, a simple model of the oculomotor system generates saccade-related disjunctive transients if one assumes there is a small delay between activation of the medial and lateral rectus muscles such as might occur through the internuclear pathway (King, unpublished observations). The oculomotor system is evidently able to utilize these disjunctive transient changes to shorten the length of time required to complete vergence movements. We acknowledge the assistance of D. Carson during surgery and in caring for the monkeys. L. Snyder, G. Paige, and C. Schor provided helpful comments on preliminary drafts of this manuscript. This work was supported by National Institutes of Health grants EY04045, EY-06632, and RR-05403 (to Dr. W. M. King) and EY-013 19 (Center for Visual Science Center Grant). Address for reprint requests: W. Michael King, University of Rochester Medical Center, Box 642, 60 1 Elmwood Ave., Rochester, NY 14642. Received

18 February

1992; accepted

in final

form

3 June

1992.

REFERENCES COLLEWIJN, H., ERKELENS,~. nation of human horizontal 404: 157-182, 1988.

J., ANDSTEINMAN, R.M.Binocularco-ordisaccadic eye movements. J. Physiol.

Lond.

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 12, 2019.

1260

J. S. MAXWELL

H., VAN DER STEEN, J., AND STEINMAN, R. M. Human eye movements associated with blinks and prolonged eyelid closure. J. Neurophysiol. 54: 1 l-27, 1985. CUMMING, B. G. AND JUDGE, S. J. Disparity-induced and blur-induced convergence eye movement and accommodation in the monkey. J. Neurophysiol. 55: 896-914, 1986. DODGE, R. Five types of eye movements in the horizontal meridian plane of the field of regard. Am. J. Physiol. 8: 307-329, 1902. ENRIGHT, J. T. Changes in vergence mediated by saccades. J. Physiol. Lond. 350: 9-3 1, 1984. ENRIGHT, J. T. Facilitation of vergence changes by saccades: influences of misfocused images and of disparity stimuli in man. J. Physiol. Lond. 371: 69-87, 1986. ENRIGHT, J. T. Convergence during human vertical saccades: probable causes and perceptual consequences. J. Physiol. Lond. 410: 45-65, 1989. ERKELENS, C. J., STEINMAN, R. M., AND COLLEWIJN, H. Ocular vergence under natural conditions. II. Gaze shifts between real targets differing in distance and direction. Proc. R. Sot. Lond. Ser. B Biol. Sci. 236: 44 l465, 1989. EVINGER, C., SHAW, M. D., PECK, C. K., MANNING, K. A., AND BAKER, R. Blinking and associated eye movements in humans, guinea pigs, and rabbits. J. Neurophysiol. 52: 323-349, 1984. JUDGE, S. J., RICHMOND, B. J., AND CHU, F. C. Implantation of magnetic search coils for measurements of eye position: an improved method. Vision Res. 20: 535-538, 1980. KENYON, R. V., CIUFFREDA, K. J., AND STARK, L. Unequal saccades during vergence. Am. J. Optom. Physiol. Opt. 57: 586-594, 1980. LEIGH, R. J., NEWMAN, S. A., FOLSTEIN, S. E., LASHER, A. G., AND JENSEN, B. A. Abnormal ocular motor control in Huntington’s disease. Neurology 33: 1268-1275, 1983. MANNING, K. A., EVINGER, C., AND SIBONY, P. A. Interaction between eye and lid movement: blinks encoded with voluntary gaze shifts. Sot. Neurosci. Abstr. 16: 90 1, 1990. COLLEWIJN,

AND W. M. KING J. S. AND KING, W. M. Saccades contribute to vergence in rhesus monkeys. Sot. Neurosci. Abstr. 16: 90 1, 1990. MAYS, L. E. Neural control of vergence eye movements: convergence and divergence neurons in midbrain. J. Neurophysiol. 5 1: 109 1- 1108, 1984. ONO, H. AND NAKAMIZO, S. Changing fixation in the transverse plane at eye level and Hering’s law of equal innervation. Vision Res. 18: 5 ll519, 1978. ONO, H., NAKAMIZO, S., AND STEINBACH, M. J. Nonadditivity of vergence and saccadic eye movement. Vision Rex 18: 735-739, 1978. PAIGE, G. D. AND TOMKO, D. L. Eye movement responses to linear head motion in the squirrel monkey. II. Visual-vestibular interactions and kinematic considerations. J. Neurophysiol. 65: 1183- 1196, 199 1. RASHBASS, C. AND WESTHEIMER, G. Disjunctive eye movements. J. Physiol. Lond. 159: 339-360, 196 la. RASHBASS, C. AND WESTHEIMER, G. Independence of conjugate and disjunctive eye movements. J. Physiol. Lond. 159: 36 l-364, 196 1b. ROBINSON, D. A. A method of measuring eye movement using a scleral search coil in a magnetic field. IEEE Trans. Biomed. Eng. 10: 137- 145, 1963. ROBINSON, D. A. Control of eye movements. In: Handbook OfPhysiology. The Nervous System. Motor Control. Washington, DC: Am. Physiol. Sot., 198 1, sect. 1, vol. II, part 2, p. 1275- 1320. SNEDECOR, G. W. AND COCHRAN, W. G. Statistical Methods. Ames, IA: Iowa State Univ. Press, 1974. SNYDER, L. H. AND KING, W. M. The effect of viewing distance and the location of the axis of rotation on the monkey’s vestibulo-ocular reflex (VOR). I. Eye movement responses. J. Neurophysiol. 67: 86 l-874, 1992. SNYDER, L. H., LAWRENCE, D. M., AND KING, W. M. Changes in vestibulo-ocular reflex (VOR) anticipate changes in vergence angle in monkey. Vision Res. 32: 569-575, 1992. VIIRRE, E., TWEED, D., MILNER, K., AND VILIS, T. A reexamination of the gain of the vestibuloocular reflex. J. Neurophysiol. 56: 439-450, 1986. MAXWELL,

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (130.070.008.131) on January 12, 2019.