ON THE OPTOKINETIC RESPONSE DURING STEP-WISE CHANGES IN STIMULUS VELOCITY IN SQUIRREL MONKEYS
J. Kroller and F. Behrens
Institute of Physiology, Freie UniversiHit, Berlin, Germany Reprint address: J. Kroller, Institute of Physiology, Freie Universitat, Arnimallee 22, 14195 Berlin, Germany.
Tel: 49-30 838-2512; Fax: 49-30 838-2507
0 Abstract-In two awake untrained squirrel monkeys the horizontal optokinetic nystagmus (OKN) was studied. The goal was to quantify the buildup of the slow-phase eye movement velocity during the first two seconds and the eye movements after OKN interruption by a stationary surround. We intended to uncover possible effects of a 'charged' velocity storage on eye movements at a stationary surround. Using an optokinetic drum, a paradigm was designed to create sudden changes of the pattern (within 5 ms) between appearing to be stationary or rotating. Velocity steps from zero to 14 to 73°/s and back to zero could be achieved. OKN onset: 201 velocity trajectories were analyzed. The mean latency between the onset of pattern movement and the onset of slow-phase eye movements was 82.8 ± 16.5 ms. Over a limited period the initial increase in slow-phase velocity could be approximated by a straight line. The slope was on average 103 ± 67°/s2 and did not show a significant dependency on pattern movement velocity. Eye movement velocity at the end of the linear part increased linearly with drum velocity; the slope was 0.59. After the linear range, the slow-phase velocity increased further but at slower accelerations and usually reached the fina£ gain within the two seconds. The initial linear acceleration component is an open-loop reflex response and we conclude that closing the loop happens when about 60% of the stimulus velocity is reached.
OKN-offset: The influence of a fully charged OKN velocity storage mechanism on eye movements after a sudden exposure of a stationary surround was studied in 23 trials. After OKN interruption the velocity decay commenced after an interval of 83.5 ± 16.6 ms. On average the slope of
Most investigations of optokinetic nystagmus (OKN) have focussed on steady state conditions (1). In many of them, eye movements during the onset of optokinetic stimulation were also studied (2-9). Fewer reports presented a quantitative analysis of the OKN time course during the very onset of optokinetic stimulation: Collewijn (I 0) reported a minimum transport time of 7 5 ms be~ween illumination of a moving stimulus and the beginning of eye movements in the rabbit. He also reported an initial constant acceleration of the OKN slow-phase. In a report on rhesus monkeys (11) the linearly approximated acceleration during the first 100 ms of the OKN was measured. Also in experiments with monkeys, Miles and colleagues (12) and Kawano and Miles (13) found a very early, brief and transient velocity increase of 50 ms after the stimulus onset, followed by a second larger one. In the studies of OKN ter-
36
mination, the aftemystagmus (OKAN) was usually investigated (1,2,5,6,14,15,16) whereas eye movements after OKN in the presence of a stationary visible pattern (15, 17, 18) have been of less interest.
One goal of the present work was to study the OKN onset to a true stepwise increase in pattern velocity. As opposed to earlier reports (11,12), we left the moving stimulus on continuously and studied the slow-phase velocity buildup over 2s. To achieve fast stimulus velocity jumps, we designed a paradigm to switch between the drum appearing stationary or rotating and a physically stationary or rotating one. Secondly, we compared the velocity increase with the velocity decay after optokinetic stimulation. As OKAN represents a long-lasting open-loop condition, it was of greater interest to compare OKN onset to OKN termination at stepwise velocity reduction of a moving pattern to zero velocity. Therefore the return to holding phases after OKN (comparable to fixation suppression) was quantitatively analyzed. The comparison of the velocity increase and the offset response revealed strong similarities in the time courses of the slow phase eye velocity.
Methods
Experiments were conducted in two adult squirrel monkeys; in one animal with horizontal OKN to the right (M86) and in the other to the left (M103). The recordings were obtained during the initial experiments of a single-neuron study in which the animals were also confronted with a variety of visual stimuli other than the purely horizontal drum rotation to the side of penetrations. The animals were placed in a primate chair; the head was fixed in the upright position by means of an acrylic support screwed to the skull. They were not deprived of water and not rewarded or trained for any movement or behav.ior. About 20 min prior to the recording session D-amphetamine sulfate was administered orally (mixed with some juice) at dosages between 0.25 and 0.75 mg/kg to keep the monkeys awake. A horizontally rotating drum (diameter· 60 em) with vertical black and white
J. Kroller and F. Behrens
stripes was used to elicit OKN. The pattern periodicity was 2.37°; drum velocities were between 14 and 73°/s. Eye movements were recorded by means of the scleral search-coil technique (19). Technical details and the calibration procedures were described earlier (20,21). The sampling interval was 4 ms. Eye velocity was obtained by computing the slope of regression lines for each sampling point of the eye position trajectory. The linear approximation was computed over 2n + 1 data points; n points at each side of the respective sampling point. n was selected between 5 and 1 0; it was 8 in the trajectory samples of Figure 1. Before doing the analysis, we tested the influence of n
values up to 30. In doing so, we ensured that the amount of smoothing finally accepted in the analysis was clearly not critical to create distortions of the slow phase velocity component. The shape of the quick phases was slightly changed (reduced peak velocity and increased duration), which, however, could be neglected when focussing solely on the slow phase velocity.
OKN onset: Velocity steps in drum rotation started at zero and jumped to a given constantsp~ed drum rotation in the following way: Before recording, the drum was set into motion and simultaneously illuminated stroboscopically (flash duration 50 J.LS). Flash frequency and rotation speed were adjusted so that a flash appeared after the drum had rotated the distance of one black and white stripe. Therefore, the pattern appeared stationary and the monkeys performed not nystagmus but spontaneous eye movements, as they usually do when perceiving a stationary drum. When, in addition, constant illumination of the drum was switched on (shutter opening time 5 ms), the drum motion became visible and the monkey i~mediately began optokinetic eye movements. Thus, a true step response from zero velocity to a given value could be obtained. During the constant light illumination, the flashes were practically not visible and normal steady-state OKN responses were observed when comparing gain and stability of slow phases to pure constant light illumination. The periods of visible pattern movements were longer than 20 seconds and presented at random intertrial intervals between 5 and 30 seconds. OKN interruption: In some
OKN Onset and Termination
experiments OKN under flash illumination in the presence of a stationary drum could be elicited (22-24 ). Then sudden constant illumination let the monkeys change from OKN to fixation periods.
Time and amplitude measurements in the eye velocity trajectories were done graphically [25]. The plots could be enlarged to any desired amplification to achieve sufficient preciseness (the samples in Figure 1 were scaled at low magnitude to give an overview of the velocity time courses). For each trial of OKN onset, we tried to determine four parameters: (1) Latency between the onset of the constant light and the first clearly detectable increase or decrease in eye movement velocity. By inspection we determined the moment when the eye velocity clearly escaped from the random fluctuations of the holding phases. Faced with noisy velocity trajectories, not always monotonic velocity increases, a wide scatter of slopes of increase and superimpositions of rapid phases, we found this simple decision-based procedure sup~rior to any automatic onset detection. (2) Slope, with the initial part of the velocity increase linearly approximated. (3) The slow-phase velocity at the end of the linear portion (end velocity). (4) The latency from the stimulus onset to the end of this linear range (onset-termination interval).
Results
OKNOnset
Typical samples of optokinetic onset responses from both monkeys are shown in Figure 1. Left of the vertical line, the rotating drum appeared stationary. At the line, the constant light went on and the OKN commenced. Figure 1 shows responses in three trials during drum rotation to the right (M86) and in three with drum rotation to the left (M103). In total, 201 trials of onset responses were inspected, showing that after the drum rotation became visible, eye-movement velocity jumped to a given value. Eye movements during lower drum velocities could not be analyzed, because it was difficult to create the impression of a stationary drum when the flash frequency was too low.
37
Only the trials in which no saccade/quick-phase was present at least half a second before the moving stimulus onset were analyzed quantitatively. In most of the recordings, the initial velocity increase could be well approximated by a
~ ~
6
g] .....
(f) -0) Q)
1s "0
Figure 1. Six samples of eye movement initiation after the stripe pattern appeared rotating. At the vertical line the constant light was switched on. Horizontal lines mark zero velocity. Drum rotation as indicated in deg/s. Flash frequencies can be determined by drum velocity/2.37 deg (the periodicity of the stripe pattern). Trials from different experiments and monkeys: upper 3 trials, drum rotation to the right (M86); lower 3 trials, drum rotation to the left (M103). Velocity of saccades were clipped on the graphs. Latencies to movement onset of curve 1-5 were 84.7, 73.6, 59.0, 84.7, and 115.3 ms, respectively; it was not determined in curve 6. Vertical arrows mark the beginning and end of the assumed linear accelerations (see text). The estimated slopes were 164.5, 65.4, 150.4, 1 03.4, 79.6, and 98.7 deg!s2 from top to bottom. Note the different ordinate scalings.
Figure 2. Size histogram of 151 latencies between onset of pattern movement and clearly detectable increase in eye movement velocity. Mean 82.8 :::: 16.5 ms.
straight line (begin and end of the linear portion indicated by vertical arrows in Figure 1). After the linear increase, the velocity trajectory was quite variable among the different recordings. Usually the velocity increased in a curved manner until the stationary velocity was reached. No further attempts were made to describe this portion quantitatively. The data for the two monkeys did not deviate significantly, therefore in Figures 2-5 their results were pooled; the statistical analysis was done for the pooled and the separated data.
Because the monkeys were untrained and sometimes showed reduced attention or restlessness, some trials had to be left out of further analysis, showing no clearly measurable parameters at all. In other trials, not all parameters could be determined with certainty, for example, when a rapid eye movement spontaneously occurred shortly after constant light onset. In other cases, the termination of the linear part could be clearly discerned, and then the end velocity and the duration of the linear part were not accessible. Therefore, the number of data points in the statistical analysis of the four parameters is different, reflecting subpopulations of the total number of 201 trials.
Latencies. The distribution of latencies from both monkeys is shown in Figure 2. The aver-
age was 82.8 ± 16.5 ms for the pooled data of both monkeys; it was 85.3 ± 16.6 ms for the data from the animal M103 (n = 45) and 81.7 ± 16.5 for the other monkey M86 (n = 106). In the trials from one experiment, we found three clearly longer latencies within others of the usual length. The nature of these long values is unclear; blinks may have played a role. These long intervals were not used when determining mean and standard deviation. A linear regression analysis verified a slight increase of 0.08 ms/0 s- 1 between latency and drum velocity; the correlation coefficient, however was insignificantly small (0.084). The latency was also not correlated to the slope of the velocity increase. In the monkey M86, we also recorded a few responses to alternations from full darkness to the visible drum moving at velocities in the order of 30°/s. The latencies obtained were 79.0 ± 16.0 ms (n = 7).
Linear velocity increase. The mean acceleration ofthelinearportionwas 136.4 ± 64.4°/s2 (n =51) for M103 and 87.8 ± 62.3/s2 (n = 117) in M86; pooling the results from both monkeys provided 102.5 ± 66.6°/s2 . Thus, the velocity increase showed a wide scatter including a significant difference between the two animals (alpha < 0.001). Linear regression analysis did not reveal any correlation between drum velocity and lin-
:6KN Onset and Termination
60
"';;' 50 .......... (jl
! 40
£ 30 u 0
1i 20 -o ~ 10
0
0+-~~~~~~~~~~~~
10 20 30 40 50 60 70
drum velocity [deg/s]
Figure 3. Measuring the slow-phase eye velocity at the end of the linearly approximable range (ordinate) at different drum velocities (n =142). Regression line (shown): y = 0.59 x -1.9, r = 0.80. Regression line forced through the origin (not shown): y = 0.55 x.
ear eye acceleration. Similarly, the interval between pattern movement onset and termination of the linear range was widely scattered, however, this 'onset-termination interval' seemed to increase with drum velocity. The linear regression analysis yielded (pooled data): jnterval (ms) = 6.4 X drum velocity (degrees/per second) + 109.7 ms and a weak correlation (r=0.46). On average, the onset-termination interval was 300 ms. The slow-phase eye velocity at the end of the linear range, the 'end velocity', was more closely related to the stimulus velocity (Figure 3, r=0.80). The slope of 0.59 (0.68 in M103 and 0.55 in M86) indicates that, irrespective of the stimulus velocity, the slow-phase eye velocity increases linearly until 59% of the stimulus velocity is reached. When forcing the regression line through the coordinate system's origin, the slope of the regression line is 0.55. In the few trials with change from darkness to a visible moving pattern, the slope of the linear velocity increase was 145.1 :::'::: 29.5°/s2 (n=8).
OKN Offset
In the experiments mentioned so far, drum rotation velocity and flash frequency were adjusted to create the impression of a stationary pattern. This was verified when the monkeys performed only spontaneous irregular eye movements as they did in the presence of a stationary, constantly illuminated drum. Under a different
A
constant light off on
8
40 80 120 latency [ms]
I 160
.. 39
Figure 4. (A) Example of OKN interruption when the apparent pattern movement was replaced by a stationary drum under constant light illumination (dashed line). Latency between stimulus movement interruption and beginning of eye movement velocity decay indicated by horizontal arrows. Eye movement deceleration is approximated by a straight line. Apparent drum velocity is 66.8 deg/s; gain = 0.8. (B) Offset latencies were determined in 17 trials (83.5 :t 16.6 ms).
condition, the monkey's perception of the same stimulus combination, however, can be different: When presenting an appropriate combination of rotation speed and flash frequency in the presence of ongoing optokinetic eye movements, the moving stroboscopically illuminated pattern could be perceived as moving and the nystagmus is continued. This phenomenon has described as sigma-OKN (22-24). It was further shown tha~ one can even le: the monkey perform OKN when the stroboscopic illuminated drum is stationary. We will call these eye movements 'apparent movement OKN' and simply use them as a tool to produce abrupt OKN terminations. The speed of the corresponding apparent drum motion can be calculated by multiplying pattern periodicity by flash frequency. In some experiments, we could elicit such an apparent movement OKN on a stationary drum. The OKN gain was similar to the gain during real pattern movement. When we could
40
achieve 2 to 3 minutes of stable apparent OKN, we switched on the constant light and then created a true step to zero velocity.
Figure 4A gives an example of the velocity trajectory after termination of an apparent movement OKN. The eye movement deceleration could be well described by two parameters. The first was the interval between the moment of constant illumination and the beginning of the decay; the second was the slope of decay. In two recordings, the edge between constant slow-phase velocity and deceleration was not sharp, and in four cases quick phases appeared at this moment. The mean of the remaining 17 latencies was 83.5 ::!: 16.6 ms (Figure 4 B). The data from the two monkeys did not vary significantly: M103 82.4 ::!: 17.0 ms (n = 8); M86 84.6 ::!: 17.1 ms (n = 9). In 22 of 23 trajectories, the velocity decrease could be well approximated by a straight line (Figure 4A). In some cases the linear decay did not immediately reach zero velocity. But in these cases the major part of decay was also linear. The slopes of linear decay (Figure 5) are only weakly dependent on drum velocity during OKN (r = 0.34). Averaged over all drum velocities, the slope of velocity decay was -195.4 ::!: 83.6°/s2 (n = 22). Separated for the two monkeys, the respective values were M103 183.0::!: 68.1 (n = 10), M86 205.8::!: 96.8 (n = 12). The lower dashed line shows that the slopes of decay were on average slightly higher
Figure 5. Relation between linear deceleration during fixation suppression and apparent drum velocity of the preceding OKN (dots). Dashed lines mark measurements of the slope of OKN buildup, i.e., linear acceleration: from Lisberger et al. (11) (upper line) and own measurements (lower line). Regression line fitted to the data points: y = 1.994 x + 86.8, r = 0.333.
J. Kr611er and F. Behrens
drum stationary I I I
f]W0~ flash on 1
constant light off I on
5s
Figure 6. Eye position trajectory during interruption of the apparent movement OKN to the right by sudden presentation of a stationary patter (dashed line). Apparent drum velocity 35.6 deg/s, OKN gain = 0.84. The OKN still has some impact on eye position. Arrow indicates the final cessation of low-velocity movements due to velocity storage.
than the slope of increasing velocity at OKN initiation. The upper dashed line was redrawn from Lis berger and colleagues, report ( 11) on early OKN initiation slopes after steps from darkness to a moving pattern.
Figure 6 shows that stored OKN-related neural activity is not totally erased by movement suppression during fixation. In all recordings, we found that after the sharp drop in slow-phase velocity, the eye movements during the fixation period were not stable as they usually were in the presence of an unmoved pattern. There was always a slow but constant drift in the direction of the previous OKN slow phase. The slow drift velocity never exceeded 10% of the OKN slowphase velocity. The drift was always interrupted by some saccades or quick phases. We interpreted this drift as indicating remaining activity of the velocity-storage mechanism. In 18 cases, from position or velocity trajectories the end of this slow component could be detected quite precisely as shown in Figure 6, although the amplitudes were low. The duration of the slow drift was 5.82 ::!: 0.98 seconds (M103 5.78 :±: 0.93, n = 6; M86 5.84 :±: 1.04, n = 12); it was not related significantly to the stimulus velocities applied.
Discussion
The aim of the present work was to describe the time course of the horizontal optokinetic
OKN Onset and Termination
nystagmus during the first 2 s and after steps from a given pattern movement velocity to zero in untrained monkeys.
Latencies
The mean delay of 83 ms that we found did not depend on stimulus velocity or properties of the consecutive eye movement. In experiments with stimulation-onset after darkness, Collewijn (10) found a comparable value in rabbits of 75 ms. Both latency values fit the latency of 60 ms for the activation of NOT neurons [15] after onset of oculomotor stimulation in the rabbit. A fast velocity rise requires an intact flocculus (26), and according to Waespe and colleagues (17), the initiation of flocculus neural activity in monkeys occurred after 70 ms. Therefore, our data support the idea that the direct OKN loop involving the flocculus could play a role in controlling the initial OKN onset. The latency from movement onset to movement initiation was very close to the latency from OKN interruption to the drop in eye movement velocity. This finding shows that the latency is also independent of the preexisting eye movement and the state of the velocity storage mechanism. Probably the same pathway in the brain controlled OKN onset and stable gaze holding phases after OKN. In the experiments of Waespe and colleagues (17), the OKN onset latency was longer (132 ms) than in the present results. Waespe and Schwarz (18) reported a latency of about 110 ms between OKAN interruption and onset of velocity decay in rhesus monkeys. This also is clearly longer than what we found.
In monkeys, Miles and colleagues (12) applied brief velocity steps ( 1 00 ms holding phase) of a large-field stimulus. They found clearly shorter latencies (50 ms) from the movement onset to a very early velocity peak, and a dependence of the amplitude of this peak and of the following eye acceleration on the pattern movement velocity. These findings are not in agreement with our data. An explanation for the different results can be that we did not couple the stimulus onset to preceding saccades as Miles and colleagues had done, since the early velocity peak depended on a postsaccadic en-
41
hancement ( 13). In our data, saccades or. quick phases were always more than 0.5 s prior to the onset of the pattern movement. Nevertheless, this is not a completely satisfactory explanation for the deviating results, since we recorded some trials in which the onset moment was less than 300 ms after a quick-phase. In none of these trials did we see latencies shorter than those shown in the data in Figure 2. In Robinson's early report on smooth pursuit in man (27), a comparable value for the initiation of eye movements was 130 ms. More recently, shorter latencies of 91 to 112 ms (28) and 93 ms (29) were found for smooth pursuit. These results are closer to the latencies we found in monkeys.
Linear Acceleration During 0 KN Onset
Applying a stimulus change from darkness to movement in monkeys, Kurzan and colleagues (30) reported initial eye movement acceleration values below 100°/52, while Lisberger and colleagues ( 11) found accelerations between 200 and 300°/s2 for drum velocities from 30 to 90°/s. We found values in between (mean 103°/s2). The monotonic increase of the linear slope with drum velocity in the data of Lisberger and colleagues does not correspond to our findings. The reason for the deviating results can be that we alternated from a visible stationary pattern to a moving one, while Lisberger and colleagues changed from darkness to a moving stimulus. In such a paradigm the change in stimulation can probably increase alertness.
Similar to the effect of body vibration in rabhits (3 a vadable attention level in our experiments can explain the wide scatter of the slope of the velocity increase. The wide scatter we found is in disagreement with the results of Miles and colleagues (12). They showed for both latency and time course of velocity increase an amount of intertrial variability that was an order of magnitude smaller than in our experiments. In their experiments, the animals were trained to fixate and were rewarded for saccades; the intertrial interval was 3 s at most, and the beginning of pattern movement was
42
triggered by saccades. In contrast, in our experiments the monkey's cooperation was not facilitated, which could explain the deviating results. A further comparison with Miles and colleagues' data is difficult because they finished the analysis 140 ms after pattern movement onset, whereas none of the linear acceleration we determined ended at or before 140 ms.
Studying smooth pursuit in monkey experiment, Lisberger and Westbrook (32) measured eye movements during steps to different velocities while starting from zero and found much larger values for the acceleration under foveal onset of stimulus movement than we found for the OKN. A separation of the acceleration into a portion before 20 ms and a longer one, as was possible in their data, does not correspond to our data. In their data, the second acceleration component ( 60 to 80 ms) depended on both the speed of the fixation point and the starting position of a step-ramp paradigm. The early acceleration component at 20 ms after movement onset vanished when the step amplitude became zero, indicating similarities between foveal onset of smooth pursuit and the present OKN data. Smooth pursuit in humans provided some lower values for the average acceleration (time window below 100 ms (28)). From Figure 4 of Behrens and colleagues (29), we evaluated the acceleration during smooth pursuit in the order of 40°/s2.
Taken together, two velocity peaks have been described in related experiments that do not exist under the present stimulation conditions: (a) the early postsaccadic enhancement in a large-field stimulation commencing 50 ms after optokinetic stimulus onset, and (b) the velocity peak 20 ms after movement onset in a smooth pursuit task. Interestingly, the early velocity peak during smooth pursuit (28) and the linear acceleration phase in our data have a similarity in that neither depends on stimulus velocity. This finding provides evidence that the initial OKN velocity increase stems from the smooth pursuit system.
In our experiments, at least the first 83 ms of the eye movement have to be under open loop conditions. Using continuous open loop conditions (stabilized retinal image on the stimulated eye (33,34 ), Dubois and Collewijn (35) showed
J. Kroller and F. Behrens
an eye movement recording during optokinetic stimulation that displayed a slow-phase velocity increase interrupted by three quick phases within the first 2 seconds after stimulus onset. However, the slow-phase velocity increase was much slower than the linear slope that we found. From a recording of Behrens and Grtisser (36), one can estimate an acceleration on the order of 4°/s2 under open loop obtained by retinal image stabilization. Thus, there is a clear discrepancy in velocity increase between the data from stabilized retinal images and the transient open loop at OKN onset.
In our results the slope of the linear portion seems to be an independent variable, probably influenced by attention, but not stimulus velocities between 14 and 73°/s. The transition from the initial linear to the following curved and slower increase takes place when about 50% to 60% of the required eye movement velocity is obtained by the reflective component (assuming a final gain of about 0.9). Since the uniform initial velocity rise can continue unchanged up to several hundred milliseconds, the movement remains under open loop, in all likelihood, until the linear velocity increase ends. Thus, the velocity increase of the direct OKN component can be divided into two portions: a fast reflexive acceleration, followed by a slower one that is probably feedback controlled. With the stimulation velocities of the present report, the final eye velocity was always reached within 2 seconds and afterwards remained constant. The even slower velocity increase from the indirect OKN component at higher drum velocities is not considered in this report.
Gaze-Holding Phase after OKN
With the present work, we intended also to observe the influence of velocity storage on eye movements after the change from a moving to a stationary stimulus (corresponding to fixation suppression). The velocity changes of the present decay were between the results for the increasing velocities and clearly not lower than the increase in velocity in the data from our laboratory (Figure 5). As a necessary prerequisite for any statement on velocity storage impact on
OKN Onset and Termination
non-nystagmic eye movements, we found that after a long lasting apparent movement OKN, the afternystagmus in darkness was well developed. This is in accordance with earlier reports (22,24,37). Therefore, we could be sure that the velocity storage mechanism was maximally 'loaded' when the animals were suddenly faced with the stationary pattern. The 'loaded' storage mechanism should be able to continue moving the eyes in the direction of the previous OKN. However, other mechanisms had to be considered: ( 1) Frictional forces could facilitate the velocity decay compared to acceleration. (2) After the stationary pattern became visible and before the eye velocity reached zero, the retinal slip changed to the opposite direction. Its velocity activated the contralateral pathways from retina and nucleus of the optic tract [38-40] and could actively decelerate the eye until zero velocity was reached.
Thus, during OKN termination, two neuronal mechanisms, velocity storage and retinal slip velocity, seem to counteract. The onset .. velocity increase stems from the direct 0 KN component and retinal slip velocity, but is not influenced by the velocity-storage mechanism. Thus, from similar slopes of OKN onset and offset in our results, it is clear that there is no impact of the fully loaded velocity storage on the slope of the velocity decay. One might consider 'discharging' of the loaded storage mechanisms to explain the missing effect on the velocity decay. However, this disfacilitation seems not fast enough (18) to explain the missing impact that occurs at the shortest latency after OKN interruption. Instead, this fast interruption of any storage impact on eye movements could indicate that, in squirrel monkeys,. 2 third mechanism controls the velocity storage impact on eye movements. Because of corresponding latencies, the pursuit system and the ipsilateral flocculus (Waespe and colleagues ( 17,41) could serve this purpose. Only supplying the storage mechanism with a stationary visual pattern over about 6 seconds was sufficient to set its activity to zero. The interval agrees with earlier time measurements of Waespe and Schwarz ( 18), who determined the duration of OKN effects from the amount of 'dumping' of the OKAN slow phase velocity in BOG recordings. The
43
persisting slow nystagmus beats during the fixation -phase after OKN show that the velocitystorage activity is turned down only by about 90%. In the current study, the coil technique enabled us to determine the small nystagmus beats during the fixation period fairly precisely (Figure 6) without relying on the afternystagmus.
Acknowledgments-We thank L. Weiss for helping with the data acquisition and Karla Steingraber for correcting the English manuscript.
REFERENCES
1. Collewijn H. The optokinetic contribution. In: Carpenter RHS, ed. Vision and visual dysfunction; vol 8; Eye movements. London: Macmillan; 1991:45-70.
3. Griittner R. Experimentelle Untersuchungen tiber den optokinetischen Nystagmus. Zeitschrift fiir Sinnesphysiologie 1939;68:2-47.
4. Collewijn H. Optokinetic eye movements in the rabbit: input-output relations. Vision Res 1969;9: 117-32.
5. Cohen B, Matsuo V, Raphan T. Quantitative analysis of the velocity characteristics of optokinetic nystagmus and optokinetic after-nystagmus. J Physiol 1977;270: 321-44.
6. Cohen B, Henn V, Raphan T, Dennett D. Velocity storage, nystagmus, and visual-vestibular interactions in humans. Ann NY Acad Sci 1981;374:421-33.
7. Baloh RW, Derner JL. Optokinetic-vestibular interaction in patients with increased gain of the vestibulaocular reflex. Exp Brain Res 1993;97:334-42.
8. Yokota J-I, Reisine H, Cohen B. Nystagmus· induced by electrical stimulation of the vestibular and prepositus hypoglossi nuclei in the monkey: evidence for the site of induction of velocity storage. Exp Brain Res 1992;92: 123-38.
9. Tan HS, Collewijn H, Van der Steen J. Optokinetic nystagmus in the rabbit and its modulation by bilateral microinjection of carbachol in the cerebellar flocculus. Exp Brain Res 1992;90:456-68.
10. Collewijn H. Latency and gain of the rabbit's optokinetic reactions to small movements. Brain Res 1972:36:59-70. Lisberger SG. Miles FA. Optican LM. Eighmy BB. Optokinetic response in monkey: underlying mechanisms and their sensitivity to long-term adaptive changes in vestibuloocular reflex. J Neurophysiol1981 :45:869-90.
12. Miles FA, Kawano K, Optican LM. Short-latency ocular following responses of monkey; 1: Dependence on temporospatial properties of visual input. J Neurophysiol1986;56:1321-54.
13. Kawano K, Miles FA. Short-latency ocular following responses of monkey; 2: Dependence on a prior saccade eye movement. J Neurophysiol1986;56: 1355-80.
14. Waespe W, Henn V. Vestibular nuclei activity during optokinetic after-nystagmus (OKAN) in the alert monkey. Exp Brain Res 1977;30:323-30.
15. Collewijn H. The oculomotor system of the rabbit and its plasticity. Berlin/Heidelberg/New York: Springer; 1981:240.
44
16. Raphan T, Cohen B. Organizational principle of velocity storage in three dimensions: the effect of gravity on cross-coupling of optokinetic after-nystagmus. Ann N Y Acad Sci 1988;545:74-92.
17. Waespe W, Rudinger D, Wolfenberger M. Purkinje cell activity in the flocculus of vestib:.;lar neurectomized and normal monkeys during optokinetic nystagmus (OKN) and smooth pursuit eye movements. Exp Brain Res 1985 ;60:243-62.
18. Waespe W, Schwarz U. Characteristics of eye velocity storage during periods suppression and reversal of eye velocity in monkeys. Exp Brain Res 1986;65:49-58.
19. Robinson DA. A method of measuring eye movement using a scleral search coil in <.1 magnetic field. IEEE Trans Biomed Eng 1963;10: 137-45.
20. Behrens F. An improved calibration procedure for eye movement recordings in animals using Robinson's electromagnetic scleral search-coil technique. J Neurosci Meth 1989;29: 195-9.
21. Behrens F, Weiss L-R. An algorithm separating saccadic eye movements automatically by use of the acceleration signal. Vision Res 1992;32:9889-93.
22. Grtisser 0-J, Pause M, Schreiter U. Three methods to elicit sigma-optokinetic nystagmus in Java Monkeys. Exp Brain Res 1979;35:519-26.
23. Behrens F, Grtisser 0-J, Smooth pursuit eye movements and optokinetic nystagmus elicited by intermittently illuminated stationary patterns. Exp Brain Res 1979;37:315-36.
24. Adler B, Collewijn H, Curio G, Grtisser 0-J, Pause M, Schreiter U, Weiss L. Sigma-movement and sigma-nystagmus: a new tool to investigate the gaze-pursuit system and visual-movement perception in man and monkey. Ann NY Acad Sci 1981;374:284-302.
25. Kroller J, Behrens F. Investigation of the horizontal, vertical and oblique optokinetic nystagmus and afternystagmus in squirrel monkeys. J V estib Res 1995;5:171-86.
26. Zee DS, Yamazaki A, Butler PH, Gucer G. Effects of ablation of flocculus and paraflocculus on eye movements in primate. J Neurophysiol 1981 ;46:878-99.
27. Robinson DA. The mechanics of human smooth pursuit eye movement. J Physiol1965;180:569-91.
28. Tychsen L, Lisberger SG. Visual motion processing for
J. Kroller and F. Behrens
the initiation of smooth-pursuit eye movements in humans. J Neurophysiol1986;56:953-68.
29. Behrens F, Collewijn H, Gri.isser 0-J. Velocity step responses of the human gaze pursuit system: Experiments with sigma-movement. Vision Res 1985;25:893-905.
30. Kurzan R, Straube A, Buttner U. The effect of muscimol micro injections into the fastigial nucleus on the optokinetic response and the vestibula-ocular ret1ex in the alert monkey. Exp Brain Res 1993;94:252-60.
31. Magnusson M. Effect of alertness on the vestibula-ocular ret1ex and on the slow rise in optokinetic nystagmus in rabbits. Am I Otolaryngol 1986;7:353-9.
32. Lisberger SG, Westbrook LE. properties of visual input that initiate horizontal smooth pursuit eye movements in monkeys J Neurosci 1985;5: 1662-73.
33. Koerner F. Schiller PH. The optokinetic response under open and closed loop conditions in the monkey. Exp Brain Res 1972;14:318-30.
34. Leigh RJ, Newman SA, Zee DS, Miller NR. Visual following during stimulation of an immobile eye (the open loop condition). Vision Res 1982;22: 1193-7.
35. Dubois MFW, Collewijn H. Optokinetic reactions in man elicited by localized retinal motion stimuli. Vision Res 1979;19:1105-15.
36. Behrens F, Grosser 0-J. The effect of monocular pattern deprivation and open-loop stimulation on optokinetic nystagmus in squirrel monkeys (saimiri sciureus). In Flohr H, ed. Post-lesion neural plasticity. Berlin/ Heidelberg: Springer-Verlag; 1988;455-72.
37. Pasik P, Pasik T. A comparison between two types of visually-evoked nystagmus in the monkey. Acta Otolaryngol Suppl (Stockh) 1975 ;330:30-7.
38. Collewijn H. Direction-selective units in the rabbit's nu~leus of the optic tract. Brain Res 1975;100:489-508.
39. Hofmann KP, Schoppmann A. Retinal input to direction selective cells in the nucleus tratus opticus of the cat. Brain Res 1975;90:359-66.
40. Mustari MJ, Fuchs AF. Discharge patterns of neurons in the pretectal nucleus of the optic tract (NOT) in the behaving primate. J Neurophysiol1990;64:77-90.
41. Waespe W, Cohen C, Raphan T. Role of flocculus and paraflocculus in optokinetic nystagmus and visual-vestibular interactions: effects of lesions. Exp Brain Res 1983;50:9-33.
![Seamless Interaction with Scrolling Contents on …1]Conference Papers...Seamless Interaction with Scrolling Contents on Eyewear Computers Using Optokinetic Nystagmus Eye Movements](https://static.documents.pub/doc/80x56/5e57ca0a3fac3a2a276ea982/seamless-interaction-with-scrolling-contents-on-1conference-papers-seamless.jpg)
