Laboratory of Sensorimotor Research, National Eye Institute,
Bethesda, Maryland 20892
 |
INTRODUCTION |
In the previous paper, we described how rostral
superior colliculus (SC) neurons change their activity as the position
of a parafoveal target changes. By stepping a visual target to
locations around the fovea out to an eccentricity of 5°, we showed
that rostral SC neurons increase their activity with target locations in the contralateral hemifield. Like neurons further caudal in the SC,
rostral neurons are tuned for particular locations
showing peaks of
activity between 0.2 and 3° eccentricity. Moreover these neurons
cease to discharge if the target steps into the ipsilateral hemifield.
Rostral neurons have two additional properties. First, they behave
similarly whether monkeys remain fixating after the target steps or if
they make a saccade to the eccentric target. Second, rostral SC neurons
are active in a similar manner during smooth-pursuit eye movements. For
example, if a visual target moves smoothly into the contralateral
hemifield and monkeys are required to pursue the target, these neurons
increase their discharge rates. Moving a target slightly into the
ipsilateral hemifield results in decreased discharge rates of these
neurons. Thus rather than representing a static command to keep the
eyes still, we suggest that these neurons reflect a position-error
signal for movements of the eyes close to the fovea, including pursuit
(Krauzlis et al. 1997
, 2000). According to this
hypothesis, a signal reflecting the difference between eye and target
positionan error signalwould be shared by these oculomotor
subsystems, saccades, fixation, and pursuit.
These single neuron experiments, however, only
correlate the neuronal activity to smooth-pursuit behavior and do not
show that the activity contributes to the production of pursuit. For example, one way in which the neuronal activity might change with pursuit without contributing to pursuit generation would be for it to
change in relation to preparation of a saccade that is ultimately not
produced. Since the direction of the impending pursuit and the
impending saccade would both be related to the activity of the same SC
neurons, it would not be possible to separate activity into that
related to the impending pursuit and that related to the impending
saccade. Therefore to test directly the contribution of these rostral
SC neurons to pursuit, we altered their activity and measured the
effects on smooth-pursuit eye movements independent of saccades.
We performed two sets of experiments. First, we used electrical
stimulation of the rostral SC during pursuit initiation and maintenance
to determine if activation of these neurons influences smooth-pursuit
eye movements. Second, we inactivated the neurons using the GABA
agonist muscimol to determine whether a reduction of the activity of
these neurons also influences pursuit eye movements. The results of
these experiments show that changing the activity of rostral SC neurons
influences smooth-pursuit eye movements. Additionally, the results of
our experiments are consistent with the hypothesis that the SC may
provide a position signal for the smooth-pursuit system.
Brief reports of some of these experiments have been made previously
(Basso et al. 1997, 1998; Krauzlis et al.
1997b).
| |
METHODS |
Two monkeys were prepared for chronic
electrophysiological recording of single neurons, electrical
stimulation, and reversible lesions of the SC and recording of eye
movements. The surgical procedures were described in previous reports
on experiments in which the same monkeys were used (Basso and
Wurtz 1998; Krauzlis et al. 1997a-c). All
protocols were approved by the Institute Animal Care and Use Committee
and complied with the Public Health Service Policy on the humane care
and use of laboratory animals.
Behavioral paradigms
The general behavioral paradigms and storage of data
were identical to that described in the preceding paper
(Krauzlis et al. 2000).
Monkeys performed visually guided saccade and pursuit tasks. For the
saccade tasks, monkeys fixated a centrally located light-emitting diode
(LED) for a duration of 500 ms, after which time the fixation point
stepped to a peripheral location of 15° in either hemifield on the
horizontal meridian. Monkeys were required to make a saccadic eye
movement to the target quickly and accurately. Trials were aborted if
the monkeys failed to acquire the target within 500 ms or were not
accurate within 2° as measured by an electronic window. For pursuit
trials, monkeys fixated a centrally located LED for a variable duration
of 1,000-1,500 ms and with an accuracy of at least 2°. After this
time, the target either started to move away from the fovea at a
constant velocity (ramp trials) or stepped horizontally to a slightly
eccentric position and then moved back toward the fovea (step-ramp
trials). The amplitude of the step was adjusted to minimize the
occurrence of catch-up saccades (Rashbass 1961). Monkeys
were required to track the target with an accuracy of 3°. If the
monkeys failed to track the target, the trial was aborted by
extinguishing the target and delaying the onset of the next trial by
2 s.
During correct performance of trials in all experiments, monkeys were
rewarded with a drop of fruit juice or water. Monkeys worked daily
until satiated and were given supplemental fluid as required. The
monkeys' weight was monitored daily and they remained under the
supervision of the Institute veterinarian.
Electrical stimulation
Electrical stimulation was applied through tungsten
microelectrodes (Frederick Haer) with impedances between 0.3 and 1.0 M
measured at 1 kHz. Electrodes were aimed toward the SC through stainless steel guide tubes held in place by a plastic grid that was
secured to the recording chamber (Crist et al. 1988).
Electrical stimulation was applied as biphasic pulses. The parameters
of stimulation were adjusted on-line to produce the maximum effects on
saccades following the parameters used by Munoz and Wurtz
(1993b) and Paré and Guitton
(1994) for the fixation zone. The minimum pulse width was 100 µs and did not exceed 250 µs. The minimum frequency was 100 Hz and
did not exceed 250 Hz. These parameters are also consistent with
producing a sustained activity similar to that seen in buildup neurons
rather than the high-frequency bursting (>600 spikes/s) of the saccade
related burst neurons. Current intensity ranged between 9 and 35 µA.
The duration of stimulus train differed depending on the behavioral
paradigm as described in the following text.
During stimulation trials, saccades frequently were suppressed for the
duration of the presentation of the electrical stimulation train.
During these trials, the size of the windows was adjusted to avoid
deterring task performance. The electrical stimulation parameters were
adjusted at this time to maximize the effects on saccades. We also
interleaved trials in which electrical stimulation occurred while
monkeys attentively fixated at primary position without a visual
stimulus present to determine the amplitude of any evoked saccades,
thereby identifying the location of the stimulating electrode on the SC
map (Robinson 1972).
After the stimulation parameters were set for a given site, tests of
the effects of stimulation on pursuit commenced. Two basic experimental
manipulations were used, namely, stimulation during pursuit initiation
and stimulation during maintained pursuit (Fig.
1). We applied stimulation during pursuit
initiation simultaneously with the onset of the target motion (15°/s)
in the ramp trials and continued for 400 ms (Fig. 1, initiation). For a
number of sites, we varied the relative timing of the stimulus train
and the onset of the target motion in ramp trials. In these
experiments, two conditions were used. First, the stimulation occurred
simultaneously with the onset of the target motion (0 ms). In the
second condition, stimulation occurred 100 ms before the onset of
target motion (100 ms). In both conditions, the speed of target motion
was 15°/s. In another set of experiments, we varied the speed of the
target motion in step-ramp trials. The speeds tested were 2, 5, 10, and 15°/s. For these experiments, we started the stimulation
simultaneously with the onset of target motion. Finally, for the
experiments testing the effects of SC stimulation on maintained pursuit
(Fig. 1, maintained), a stimulus train of 300 ms occurred during
maintained pursuit in step-ramp trials, defined as 600 ms after the
onset of target motion. This period was typically well after any
catch-up saccades if they occurred and after pursuit had maintained a
constant speed approximating that of the target (either 5 or 15°/s).
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 1.
Timing of superior colliculus (SC) stimulation during pursuit
initiation and during maintained pursuit. Top: the
position of the target as a function of time. Middle: a
representative eye-position signal recorded in response to rightward
motion of a target at 15°/s. Up is right and down is left.
Bottom: an average of 10 differentiated eye-position
traces. In the eye-velocity traces, high-velocity saccades have been
removed. Traces are aligned on the onset of the target motion. The
black bars indicate the onset and duration of the train of electrical
stimulation pulses applied to the SC. The shaded regions on the
eye-velocity and eye-position traces indicate the open-loop period of
the pursuit response. The shaded region on the target position trace
indicates the visual input period driving the open-loop pursuit
response. For the pursuit-initiation experiments, a 400 ms duration,
150 Hz, 100 µs pulse width, biphasic stimulation train occurred
simultaneous with the onset of target motion. In the maintained pursuit
experiments, the same stimulus train for 300 ms duration occurred 600 ms after target motion onset.
|
|
We also tested the effects of SC stimulation on smooth eye movements
evoked by vestibular stimulation and compared these effects to those
obtained during visually driven smooth eye movements. To evoke
vestibular eye movements, monkeys experienced whole body passive
rotation achieved by mounting the primate chair on a rotating platform.
The monkeys were rotated sinusoidally at 0.4 Hz and a peak-to-peak
amplitude of 20° about a vertical axis that intersected the
interaural line, producing movement along the horizontal meridian. Monkeys were required to maintain eye position at a central location while they were rotated in complete darkness with no visible fixation stimulus present. In one monkey, we were unable to control completely for eye position, but the effects of the stimulation did not differ dramatically between the two monkeys. SC stimulation began during the
phase of the sinusoidal motion in which eye velocity was maximal for
each direction, ipsiversive or contraversive, and was maintained for
400 ms. For a direct comparison with visually driven smooth eye
movements, the monkeys remained stationary and a visual target moved
along the horizontal meridian sinusoidally with a frequency of 0.4 Hz
and a peak to peak amplitude of 20°. Identical to the vestibular
condition, SC stimulation began during the phase of the sinusoid in
which the eye velocity was maximal for each direction and was
maintained for 400 ms. For all experimental conditions, the stimulation
and no-stimulation trials were presented in an interleaved fashion
except for the vestibular and sinusoidal pursuit trials, which were
presented in separate, interleaved blocks.
Muscimol injections
We injected muscimol (Sigma) dissolved in saline in the
SC of two monkeys. The injection technique we used was originally described by Dias and Segraves (1997). Briefly,
a closed pressure system was used to inject small volumes of muscimol
into the SC. This system allows for precise control and measurement of
the injected volumes. The micropipette system was adapted for use with
the grid and guide-tube system described by Crist et al. (1988). Moreover, a fine wire inserted within the pipette
allowed us to electrically stimulate or record the neuronal activity
prior to making the injection. Once the location of a site was
identified by stimulation or by mapping the movement field, brief
pulses of air pressure of fixed intensity (30-80 psi) and duration
(4-20 ms) were applied by a picopump (World Precision Instruments) to release the muscimol contained within the pipette. The injected volumes
for each site in the two monkeys are listed in Table
1. We collected preinjection and
postinjection data on interleaved saccade and pursuit trials. Recovery
data were collected 24 h after the injections. Monkeys were
required to perform saccades along the horizontal meridian of 2, 5, 10, and 20° amplitude (8 trial conditions). Pursuit trials consisted of
ramps and step ramps of constant velocity targets moving at 15°/s
along the horizontal meridian in either direction. The location of the
step varied between 2 and 5° along the horizontal meridian in either
hemifield. Thus for pursuit trials, there were 10 possible conditions.
A step location either at 2 or 5° in the contralateral hemifield and
a subsequent ipsiversive or contraversive target motion (4 conditions),
or a step location of 2 or 5° in the ipsilateral hemifield and a
subsequent contraversive or ipsiversive direction of target motion (4).
Finally, ramp trials were those in which the target motion originated
at primary position and moved at a constant speed either ipsiversively
or contraversively (2).
Data analysis
Voltage signals proportional to the horizontal and vertical
components of eye position were filtered (6 pole Bessel, 
3 dB at 240 Hz) and then digitized at a resolution of 16 bits and sampled at 1 kHz.
The data were saved on disk for subsequent off-line analysis. An
interactive computer program was then used to filter, display, and
measure eye-position and eye-velocity signals. A signal encoding
horizontal eye velocity was obtained by applying a 29-point finite
impulse response filter (3 dB at 96 Hz) to the eye-position signal.
The high frequency was used to insure detection of small saccades
(compare Abel et al. 1979; Bahill et al.
1975; Breznen and Gnadt 1997). Also to maximize
detection of small saccades, we used stringent velocity (20°/s) and
acceleration (500-800°/s2) criteria. For the
pursuit-initiation measurements, data after the occurrence of detected
saccades were excluded from analysis as well as the saccades
themselves. Once individual eye-velocity records were obtained, the
computer program calculated average smooth eye velocity by aligning the
traces with respect to target motion onset and calculating the mean and
SD of the eye velocity for each millisecond of data. Measurements of
the data resulting from experiments manipulating pursuit speed or
pursuit maintenance, or during the muscimol experiments, were made on
the velocity traces in which each millisecond of the trace marked as a
saccadic velocity was excluded from the calculation of the average
smooth eye-velocity traces. These traces were defined as desaccaded
eye-velocity traces. For data passing normality tests run by
SigmaStat, we used Student's t-test for
statistical analysis; otherwise, we used the Mann-Whitney rank sum test.
| |
RESULTS |
Electrical activation of the rostral SC
SACCADES AND PURSUIT INITIATION.
In two monkeys, we electrically stimulated the SC at rostral sites
within the central 5° of the visual field. To verify that we were
stimulating at the rostral SC locations that have previously been shown
to alter saccades, we first determined the effect on saccadic eye
movements made to large target steps (10-15°). Ipsiversive saccades
were suppressed completely and typically for the duration of the
stimulus train (Fig. 2A).
Contraversive saccades (Fig. 2E) were also suppressed
somewhat, and frequently small contralateral saccades intruded during
the stimulation. As the stimulation site in the SC was moved further
from the foveal representation, stimulation was less effective in
suppressing large saccades, even large ipsilateral saccades and more
frequently evoked small contralateral saccades (not shown). These
results are consistent with those reported previously (e.g.,
Munoz and Wurtz 1993) and serve to confirm the location
of our stimulating electrode within the rostral pole of the SC map.
View larger version (38K):
[in this window]
[in a new window]
|
Fig. 2.
Rostral SC stimulation affects saccades and pursuit. All traces are
aligned on the onset of the target step or ramp (up arrow). Monkeys
made saccades to the target steps and a combination of saccades and
pursuit to the target ramps. The black bar indicates the onset and
duration of the electrical stimulation and the thick lines are traces
with stimulation. Dashed traces in C, D, G, and
H are SD of the control trials. Left:
trials during pursuit or saccades in the hemifield ipsilateral to the
site of the stimulated SC. Right: traces from
contraversive trials. A: eye-position traces in response
to target steps of 15°. B: eye-position traces in
response to target ramps of 15°/s. C: average
eye-velocity traces from a subset of the trials with saccades removed
(see METHODS). D: traces in C
are expanded to show the 1st 200 ms of data beginning at the time of
the onset of the target motion and the electrical stimulation. The
traces have been truncated at the time of the first saccade. Horizontal
dashed lines indicate 0°/s. Arrow in H identified the
trace obtained with electrical stimulation of the SC. The stimulation
site was the same for saccades and pursuit.
|
|
During pursuit initiation, when the target ramped away from the fovea
and the monkeys were required to track the target with a combination of
smooth-pursuit and saccadic eye movements, electrical stimulation of
the rostral SC dramatically reduced the smooth-pursuit response. As in
the case of saccades, the reduction was strongest when pursuit was
directed into the ipsilateral hemifield (Fig. 2B). During
contraversive pursuit, the stimulation had little effect on pursuit but
frequently evoked small, contraversive, catch-up saccades (Fig.
2F).
To reveal the dynamics of the pursuit response, we examined pursuit
eye-velocity traces (Fig. 2, C-H). Ipsiversive initial pursuit velocity and ipsiversive maintained pursuit velocity were both
reduced with stimulation (Fig. 2C). One factor that might explain the reduction of ipsiversive pursuit velocity during
stimulation is the suppression of the usual catch-up saccades,
resulting in pursuit of an eccentric target, which has been shown to be
slower than pursuit of centrally located targets (Lisberger and
Westbrook 1985). To eliminate this confounding factor, we
restricted our analysis of pursuit velocity to the initial 100 ms of
the pursuit response because pursuit is not influenced by feedback
during the 100 ms period before the eye starts to move. Therefore no visual feedback signals or catch-up saccades can contribute to this 100 ms open loop period of the pursuit response (see Figure 1
Lisberger et al. 1987). For ipsiversive pursuit during
the open loop period (Fig. 2D, shaded region), eye velocity
was reduced, ruling out an interpretation based on suppression of
catch-up saccades and parafoveal pursuit. Contraversive velocity
records also revealed a small but consistent effect on eye velocity
(Fig. 2G and expanded traces in H). The SC
stimulation slightly increased the initial velocity (Fig.
2H, shaded region) prior to the initiation of the first saccade.
Thus in general, smooth-pursuit eye movements to targets moving in the
ipsilateral hemifield are suppressed with electrical stimulation of the
rostral SC, whereas smooth-pursuit eye movements to targets moving in
the contralateral hemifield show little effect.
We performed a number of experiments that individually demonstrate a
contribution of SC neuronal activity to smooth-pursuit eye movements
and collectively suggest that the signal produced by the SC and used by
the pursuit system is a position signal. In Fig.
3, we present a schematic diagram that
represents a framework for describing the relation of the site of
stimulation within the SC to the effect on the direction of pursuit.
For example, if electrical stimulation is applied to the SC at a site
representing a position of 1.5°, as measured by the evoked saccade
amplitude, we would expect that when the motion signal (from the
target) and the SC position signal (from the stimulation) are in the
same direction, pursuit eye velocity should either be unaffected or it
should be enhanced (Fig. 3, rightward eye-velocity traces). In
contrast, when the SC stimulation and the target motion signal are
opposite, pursuit eye velocity should be reduced (Fig. 3, leftward
eye-velocity traces). Two points are important to note. First, the
effects should be asymmetricposition steps in the direction of motion
do not facilitate pursuit as much as position steps in the opposite
direction reduce it (Carl and Gellman 1987; Morris and Lisberger 1987). Second, the magnitude of the
reduction and facilitation will depend on the magnitude of the position signal that is imposedwithin limits, larger steps produce greater changes in pursuit (Carl and Gellman 1987; Morris
and Lisberger 1987). Therefore in the following experiments we
compare the effects of stimulation of the SC at different locations in
the SC map while monkeys make smooth-pursuit eye movements in both
directions. We also manipulate stimulation timing and target speed as
well as examine the effects of SC stimulation during maintained
pursuit.
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 3.
A schematic representation of what might be expected if the SC
stimulation generates a position signal that interacts with the visual
motion stimulus to drive the pursuit response. In A, the
ovals represent the SC map showing an example electrical stimulation
site of the right SC evoking a 2.0° leftward saccade. In
B, the vertical line indicates degrees of visual angle
with negative values indicating leftward positions represented in the
right SC. The gray bar indicates the onset and duration of the SC
stimulation and is plotted to represent the SC location. The effects on
pursuit are represented by the position traces with the thick lines
indicating stimulated trials and thin lines indicating control trials.
Upward traces are rightward eye movements. With a visual target moving
20°/s leftward (dotted line beginning at "target onset") and
stimulation of a 2.0° leftward site in the SC, we would expect to
modestly speed up leftward pursuit. In contrast, rightward pursuit
would be expected to slow down (see RESULTS). In
C, the same example is provided but plotted in the
velocity domain. Conventions are the same as in B.
|
|
EFFECTS OF STIMULATION AT DIFFERENT SC LOCATIONS.
Comparing the eye-velocity responses with and without electrical
stimulation at different sites within the SC revealed a dependence on
the SC site stimulated (Fig. 4). For a
site at the very rostral end of the SC, which virtually never evoked
saccades, pursuit eye velocity was slightly reduced during ipsiversive
pursuit and unaffected during contraversive pursuit (Fig. 4,
A and E). For a stimulation site slightly more
caudal (Fig. 4, B and F), ipsiversive pursuit
velocity was reduced more dramatically, whereas contraversive pursuit
velocity was not. For a stimulating electrode even more caudal (Fig. 4,
C and G) that evoked very small saccades during fixation, there was a reduction of ipsiversive pursuit eye velocity and
a modest increase in contraversive pursuit. A site further caudal,
which evoked on average a 4° saccade, produced no pursuit effect in
either direction (Fig. 4, D and H).
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 4.
Stimulation of different locations in the SC affects pursuit initiation
differently. Average eye velocity in response to target motion ramping
away from the fovea at 15°/s is plotted as a function of time. All
trials have been truncated at the time of the 1st saccade. Thick lines,
traces from stimulation trials; thin lines, traces from trials without
stimulation. Left: pursuit responses ipsiversive to the
SC stimulation site. Right: contraversive. Horizontal
dashed lines indicate 0°/s.
|
|
As is suggested by Figs. 2 and 4, the stimulation effects
appeared to occur later in the initial open-loop period of the
eye-velocity responses. Therefore to quantify the data across our
sample of stimulation sites, we divided the first 100 ms of pursuit eye velocity into two separate intervals, a first and second 50 ms. We
plotted the difference in the average eye speed of the stimulation and
no-stimulation trials as a function of SC stimulation site (Fig.
5). Negative values indicate a reduction
of eye velocity with stimulation. We did this for both intervals as
well as for both directions of pursuit. The SC site was determined by
the average saccade amplitude evoked by stimulation while the monkey fixated straight ahead with no visual stimulus present. In general, stimulation reduced ipsiversive pursuit velocity for most sites tested.
The number of statistically significant points was higher in the second
50-ms interval of the pursuit response (compare Fig. 5, A
and C). For contraversive pursuit, by about 1° in
amplitude, velocity was unaffected although for a couple of cases
beyond the 1° site, pursuit velocity was increased (Fig. 5,
B and D). Similar to ipsiversive pursuit, the
number of statistically significant points occurred more frequently in
the second 50 ms interval of the contraversive pursuit response.
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 5.
Effects of rostral SC stimulation across the sample of stimulation
sites. The difference in eye speed between the stimulated and
unstimulated trials (°/s) is plotted as a function of the site of
stimulation for the sample of 19 stimulation sites. The data are taken
from pursuit ramp trials with any portion of the eye velocity
containing saccades omitted by truncating the trace at the time of the
1st saccade. The 1st 100 ms of the stimulation trace during presaccadic
pursuitthe open-loop periodwas subtracted millisecond by
millisecond from the same time period of the no-stimulation trace. We
then took the average difference for 2 intervals; a 1st 50-ms interval
after the onset of target motion (A and
B) and a 2nd 50-ms interval (C and
D). Closed symbols indicate statistically significant
differences between stimulation and no-stimulation trials (see
METHODS). Missing symbols result from components of the
pursuit response that are missing such as would occur during a saccade.
Left: the difference in eye velocity between the
stimulation and no-stimulation trials is plotted for ramps ipsiversive
to the site of SC stimulation (A and C).
Right: the difference in eye velocity between
stimulation and no-stimulation trials is plotted for ramps
contraversive to the site of SC stimulation (B and
D). Points falling along the dashed line indicate no
difference between the stimulation and no-stimulation conditions.
Negative values indicate pursuit eye speed was suppressed and positive
values indicate eye speed was increased by the stimulation.
|
|
Note in Fig. 5 that for a few sites close to the fovea ipsiversive
pursuit velocity increased with stimulation as indicated by the points
above the line in Fig. 5C (and see example in Fig. 6A). One possibility for this
result is that the signal generated by the stimulation reflects the
tuning of the underlying neurons within the rostral SC that had unusual
properties; some neurons increase their discharge rates for in response
to targets located slightly ipsilateral to the fovea (Krauzlis
et al. 1997c, 2000). Preferential activation of these
ipsilateral neurons can explain the modest increases in ipsiversive
velocity seen occasionally. Stimulation of these sites close to the
foveal representation also frequently suppressed contraversive pursuit
eye velocity across the sample of sites (Fig. 5D).
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 6.
Effects of SC stimulation are evident in the early phase of pursuit
initiation. A-D: representative examples of pursuit eye
velocity with (thick lined traces) and without (thin lined traces)
rostral SC stimulation beginning 100 ms before the onset
of target motion. All target motions are directed ipsiversively.
E-H: trials when the stimulation and target motion
onset occurred simultaneously. All trials have been truncated at the
time of the 1st saccade. Thick lines, traces from stimulation trials;
thin lines, traces from trials without stimulation. Horizontal dashed
lines indicate 0°/s.
|
|
To confirm quantitatively the dependence of the stimulation effects on
SC site, we performed a two-way ANOVA comparing SC stimulation and
no-stimulation conditions with the location of the stimulating
electrode on the SC map. For this analysis, we were only interested in
the interaction term, namely, the comparison of eye velocity with and
without SC stimulation across SC location. We did one ANOVA for each
direction of pursuit. For ipsiversive pursuit there was a significant
interaction between the stimulation condition and the location on the
SC map [f(1,14) = 7.65; P < 0.001]. For contraversive pursuit, the differences failed to reach significance [f(1,14) = 1.27; P = 0.22].
Thus the effect of SC stimulation on pursuit varied with the location
of the stimulation within the SC with more rostral sites producing more
effective inhibition of ipsiversive pursuit.
EFFECTS OF SC STIMULATION TIMING ON PURSUIT.
As just described, the effects on pursuit frequently differed depending
on when in the initial 100 ms of open loop pursuit the eye
speed was measured. Specifically, the effect was largely restricted to
the later 50 ms of the open loop period (compare Fig. 5, A
and C and B and D).
This difference in the stimulation effect between the two intervals of
pursuit initiation may reflect the known differences in early and late
phases of pursuit initiation (e.g., Lisberger and Westbrook
1985). Alternatively the differences may reflect the timing
delays for the SC stimulation to reach the smooth-pursuit pathways. To
distinguish between these two possibilities, we manipulated the
interval between the onset of the SC stimulation and the onset of the
target motion. In one condition, we presented the 400 ms train of SC
stimulation simultaneous with the onset of target motion, examples of
which we have already presented. In the second condition, we presented
the onset of the SC stimulation train 100 ms before the onset of the
target motion. In both conditions, we compared only the ipsiversive eye
velocity with and without stimulation at different SC sites because of
the larger and more reliable effects on ipsiversive pursuit eye
velocity (Fig. 6). When stimulation preceded target motion onset, the
pursuit effects were evident immediately when the pursuit response was
initiated (Fig. 6, left). When the stimulation occurred
simultaneously with the onset of target motion, the effects were
delayed, as already demonstrated (Fig. 6, right). This
difference was particularly evident at a site representing about 0.8°
amplitude (compare Fig. 6, C and G).
This effect was evident across our sample of 14 sites. We measured the
latency of the stimulation effect by first taking the mean and standard
deviation of the first 100 ms of ipsiversive pursuit eye velocity in
the stimulation condition and then subtracting, millisecond by
millisecond, the no-stimulation velocity trace for each site. We
considered the onset of the stimulation effect as the time when the
difference trace changed from the mean eye-velocity trace in the
no-stimulation condition by two standard deviations for at least 5 ms.
When the stimulation onset preceded the onset of the target motion by
100 ms, the median latency of pursuit onset was 36 ms. In the
simultaneous stimulation and target motion condition, the median
latency of the effect of SC stimulation was 61 ms. These differences
were statistically significant across the sample of 14 sites
(P < 0.017).
Thus the effects of SC stimulation on pursuit initiation are not
restricted to the later 50-ms phase of pursuit initiation, demonstrating that the effects of stimulation are not a result of the
known differences in the early and late phases of pursuit initiation phases (e.g., Lisberger and Westbrook 1985)
but more likely result from the time it takes for the stimulation
effects to enter the smooth-pursuit pathways. Moreover the results
demonstrate that the signal generated by the SC has access to the
initial portion of pursuit.
STIMULATION OF SC AT DIFFERENT PURSUIT SPEEDS.
Up to this point, we have presented the results of SC stimulation on
pursuit eye movements of a single speed (15°/s). Past experiments
demonstrate that the pursuit system responds differently to target
perturbations depending on the speed of pursuit. For example, by
imposing small changes in the speed of a moving target that monkeys
were required to track, Schwartz and Lisberger (1991) demonstrated that the pursuit response to the speed perturbations increased as the target speed increased. Perturbations had a
dramatically smaller effect during fixation than during pursuit.
Similarly, Komatsu and Wurtz (1989) demonstrated that
the effect of electrical stimulation of middle temporal/medial superior
temporal (MT/MST) depended on the speed of pursuit and was not
influential during fixation. Both sets of results were interpreted as
acting on the visual input to the pursuit system, which, in the case of
the stimulation, was subsequently combined with the information
obtained from the speed of the moving visual target. We tested whether the effects of SC stimulation also depended on pursuit speed to determine whether the SC signal could be combined with the visual input
to the pursuit system.
At 15 stimulation sites, we varied the target speed among 2, 5, 10, and
15°/s and measured the effects of SC stimulation on the pursuit
responses. To determine the influence of pursuit speed, we plotted the
difference in eye speed during the stimulation and no-stimulation
conditions against target speed for each of the 15 stimulation sites
(Fig. 7). The eye speed used for the calculation was the second 50 ms of the open-loop pursuit response. During this period of the pursuit response, the effect of SC
stimulation depended on pursuit eye speed for many of the sites tested.
Across the sample of stimulation sites, indicated by different symbols in Fig. 7, the difference in eye speed becomes increasingly negative as
the speed increases, indicating that the suppression of pursuit eye
speed was greater for faster speeds. Across the sample, there were
significant differences between the eye speeds in the stimulation and
no-stimulation conditions for ipsiversive pursuit [Kruskal-Wallis; H = 7.829(3); P = 0.050]. The pairwise
comparisons (Dunn's) indicated that the statistical significance
resulted from the differences between the 2 and 15°/s conditions. The
differences for contraversive pursuit across the sample of sites failed
to reach significance [Kruskal-Wallis: H = 7.308 (3);
P = 0.063] for any of the speeds despite the tendency
for the stimulation to have a greater effect on faster pursuit speeds.
This finding is consistent with both behavioral and stimulation
experiments in that greater effects are seen for faster speeds,
indicating that the SC stimulation interacts with the visual inputs for
pursuit eye movements.
View larger version (12K):
[in this window]
[in a new window]
|
Fig. 7.
SC stimulation effects are dependent on target speed. The difference in
eye speed between the stimulation and no-stimulation trials for the 2nd
50 ms of the open-loop pursuit period is plotted as a function of
target speed for each site of stimulation. Negative values indicate
that pursuit speed was reduced in the stimulation trials compared with
the no-stimulation trials.
|
|
The dependence of the stimulation effects on speed varied with the site
of the stimulation of the SC map. For example, a stimulation site close
to the fovea tended to have greater effects on faster speeds than
slower speeds. The effects at this site also tended to occur sooner for
the faster speeds of target motion [Fig.
8A; t(19) = 15.93; P < 0.001]. Similarly, this site reliably,
albeit very modestly, increased contraversive pursuit at the slowest speed (P = not significant) and reduced ipsiversive
pursuit at the highest speed but only at a later pursuit period [Fig.
8E; t(36) = 2.098; P = 0.04]. Additionally, the effects of stimulation at sites further from
the fovea differed. A site evoking a saccade amplitude of 0.2°
suppressed pursuit maximally for both directions of pursuit at the
fastest speed [ipsilateral t(84) = 3.73;
P < 0.001 and contralateral t(96) = 4.023; P < 0.001] and was slightly better at
suppressing the slower speed than the 0° site (Fig. 8B).
Finally, a stimulation site of 4.0° had a negligible effect on
pursuit except for a modest effect on the later phase of the 15°/s
ipsiversive pursuit (ipsilateral and contralateral, P = not significant).
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 8.
Stimulation of the SC affects pursuit at different speeds. Two speeds,
5 and 15°/s, with and without SC stimulation are superimposed.
Portions of the eye-velocity traces containing saccades have been
removed (see METHODS). Thick lines are traces from
stimulation trials and thin lines are traces from trials without
stimulation. Left: pursuit responses ipsiversive to the
SC stimulation site. Right: contraversive. Horizontal
dashed lines indicate 0°/s. Note that the same stimulation site and
parameters had different effects on pursuit depending on the speed of
target motion.
|
|
Finally, as a correlate to the behavioral experiments where speed
perturbations were imposed during both pursuit and fixation, we tried
to stimulate the SC during fixation to see if we could evoke
smooth-pursuit eye movements under similar conditions. At sites where
stimulation affected both saccades and pursuit, presentation of
electrical stimulation during fixation never evoked a smooth-pursuit response, even at currents exceeding threshold for producing effects on
saccades (not shown).
In summary, the effects of SC stimulation depend on the speed of
pursuit, consistent with the SC stimulation affecting the visual
processing stages of smooth pursuit rather than later stages of the
pursuit pathway. The effects are not exclusively motor either since
electrical stimulation of the SC did not evoke smooth-pursuit eye
movements during fixation such as is seen in the frontal eye field
smooth-pursuit region (Gottlieb et al. 1993).
Additionally, as seen with a single target speed, the effects of
stimulation depended on the location of the stimulation on the SC map
with maximal effects seen at locations representing target positions just slightly off the fovea. Taken together, these results
support the hypothesis that the SC provides a signal that can influence smooth-pursuit eye movements.
SC STIMULATION AND PURSUIT MAINTENANCE.
Stimulation of the rostral SC affected the monkeys' ability to
maintain smooth-pursuit eye movements as well as to initiate them (Fig.
2C). Therefore we specifically tested the effects of rostral
SC stimulation on maintained pursuit by presenting 300 ms trains of
stimulation at 600 ms after the onset of target motion. By 600 ms, the
initial catch up saccades, if any, had already occurred, and in most
cases, pursuit attained a constant velocity approaching that of the
target (as illustrated in Fig. 1). The effects on maintained pursuit
with rostral SC stimulation mirrored the effects seen for pursuit
initiation. For example, at a site close to the fovea, stimulation of
the SC resulted in suppression of maintained pursuit at 15°/s
ipsiversively (Fig. 9A). For
this site, contraversive pursuit was suppressed also (Fig.
9B). We tested the effects of SC stimulation on pursuit eye
velocity maintained at 5°/s as well and found that for most cases the
trend was similar (not shown).
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 9.
Stimulation of the SC influences maintained smooth pursuit. Average
eye-velocity traces are shown for stimulation and no-stimulation trials
for ipsiversive (A) and contraversive (B)
directions of pursuit. Portions of the eye-velocity traces containing
saccades have been removed (see METHODS). The black bars
indicate the onset and duration of the SC stimulation. Horizontal
dashed lines indicate 0°/s. Thick lined traces, with stimulation;
thin lined traces, without stimulation.
|
|
To quantify the effects of the SC stimulation across our sample of 15 sites, we calculated the difference in eye speed between the
stimulation and the no-stimulation conditions for a 200 ms interval
beginning 100 ms after the onset of the SC stimulus train. We then
plotted this difference for ipsiversive and contraversive pursuit
directions at two speeds, 5 and 15°/s, as a function of SC
stimulation site (Fig. 10). Most of
these sites were within 1° of the foveal representation, and many
suppressed maintained pursuit in both directions. Some sites resulted
in a slight facilitation of contraversive pursuit (Fig. 10B,
filled circles above dashed line). However, the predominant effect was
a reduction in the average pursuit eye speed.
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 10.
Effects of stimulation on maintained pursuit for all sites. The
difference in average eye speed with and without stimulation is plotted
for 16 sites of SC stimulation site in 2 monkeys. The arrangement of
this figure is the same as in Fig. 4. Points falling along the dashed
line indicate no difference between the 2 conditions. Points falling
below the line indicate suppression and points falling above the line
indicate facilitation of maintained pursuit. We quantified the
maintained pursuit effects by calculating the difference between the
stimulation and the no-stimulation eye-velocity traces millisecond by
millisecond for a 20 ms interval beginning 100 ms after the onset of
the SC stimulus train. The duration of the SC stimulation was 300 ms.
We then calculated the average difference between the 2 traces. This
was done for both speeds of pursuit, 5 and 15°/s, as well as both
directions of pursuit. We then plotted the difference in the average
eye speed as a function of the location of the site of stimulation
within the SC.
|
|
In sum, like the effects on pursuit initiation, maintained pursuit was
reduced in both directions and was greater for faster pursuit speeds
when the stimulation site in the SC represented small locations
ipsilateral to the pursuit target position, very close to the fovea.
Indeed, there were occasional increases in contralateral pursuit as well.
SC STIMULATION AND THE VOR.
We next tested whether rostral SC stimulation affected smooth eye
movements simply because the stimulation affected all smooth eye
movements or because the effects of stimulation were specific for
smooth-pursuit eye movements. We did this by comparing the effects of
stimulation during smooth eye movements evoked by head rotationsthe
vestibular-ocular reflex (VOR)to the effects of stimulation obtained
during smooth-pursuit eye movements. For pursuit in these
experiments, we presented monkeys with a target moving sinusoidally at
0.4 Hz and a peak-to-peak amplitude of 20°. To generate the VOR,
monkeys fixated straight ahead in the dark while the chair was rotated
sinusoidally at 0.4 Hz with a peak-to-peak amplitude of 20°. On
interleaved trials, SC stimulation was introduced for 400 ms beginning
during the phase in which the ipsiversive and contraversive eye
velocity was maximal. Stimulation within the SC during sinusoidal
pursuit when the peak eye velocity was maximal in the direction
ipsilateral to the stimulation suppressed smooth-pursuit eye velocity
(Fig. 11A; note that when
the velocity trace begins to turn downward, this reflects the time when
the eye velocity peaks and then declines indicating the turn around of
the sinusoidal movement). At the same SC site and stimulation conditions but during the VOR, there was only a modest effect on smooth
eye velocity (Fig. 11B). There were no effects on
contraversive pursuit or VOR at this site (not shown). Across our
sample of seven sites, six showed significantly different effects of
stimulation in the ipsiversive sinusoidal pursuit and VOR conditions
(Fig. 11C). Thus the effects of electrical stimulation
differ for pursuit and VOR eye movements and argue against the
possibility that the SC stimulation is a generic inhibitory signal that
affects all smooth eye movements similarly.
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 11.
Differential effects of SC stimulation on vestibular-ocular
reflex (VOR) in the dark and sinusoidal pursuit. A: eye
velocity during sinusoidal smooth pursuit (0.4 Hz, 20° peak-to-peak
amplitude) ipsiversive to the site of SC stimulation is plotted with
and without stimulation. B: ipsiversive eye velocity in
response to whole body passive rotation in the dark (0.4 Hz, 20°
peak-to-peak amplitude). Thick lined traces, with stimulation; thin
lined traces, without stimulation. Saccades have been removed (see
METHODS) Horizontal dashed lines indicate 0°/s. Filled
bars indicate the duration of the SC stimulation. C: the
average difference in ipsiversive pursuit eye velocity between
stimulation and no-stimulation trials measured over the 150-ms interval
of SC stimulation beginning 100 ms after the onset of the stimulation
is plotted against the average eye velocity over the same interval
during VOR. Six of the 7 cases had significantly different effects of
stimulation between pursuit and VOR. In all cases, the VOR effect was
smaller. For this example case, there was a statistically significant
difference in eye velocity with SC stimulation (P < 0.014). None of the other 6 cases was statistically significant.
|
|
STIMULATION OF SUPERFICIAL SC LAYERS.
Since our sites of stimulation in the SC were at locations where
low-threshold stimuli evoked saccades, it seems likely that the pursuit
effects resulted from activation of the premotor neuronal elements
within the intermediate and deep layers. Despite this, it is possible
that the effects on pursuit resulted from activation of the overlying
visual neurons within the superficial layers of the SC as well. The
significance of stimulating the superficial rather than the
intermediate layers is particularly important because it is known that
cortical areas encoding visual motion signals project directly to the
SC superficial layers (Ungerleider et al. 1984) and that
superficial layer visual neurons themselves convey motion signals
(Davidson and Bender 1991).
To explicitly test the hypothesis that the pursuit effects resulted
from activation of the overlying visual neurons in the SC, we measured
smooth-pursuit eye movements in response to target ramps of 15°/s in
two conditions. In the first condition, monkeys pursued a moving
target, and on interleaved trials, we presented a train of stimulation
to the intermediate layers of the SC at sites where saccades could be
affected at low thresholds. In the second condition, we moved the
electrode dorsally, into the superficial layers, confirmed that the
electrode was located within the same region of the visual field
representation by recording multiunit activity, and then interleaved
stimulation trials using the same parameters. In both conditions, the
stimulation was presented simultaneously with the onset of the target motion.
Comparing smooth-pursuit eye speed with and without stimulation of the
SC in a single penetration revealed different effects on the eye
movement depending on whether the stimulation occurred within the
superficial layers or the intermediate and deep layers of the SC (Fig.
12). Stimulation of the superficial
layers of the SC at a site very close to the representation of the
fovea, increased ipsiversive pursuit speed (Fig. 12A,
top left), whereas contraversive pursuit was relatively
unaffected (Fig. 12A, top right). In marked contrast, stimulation right below this site in the intermediate and
deep layers suppressed pursuit eye speed in both directions (Fig.
12A, bottom). Thus for this site, the effect of
superficial layer stimulation on ipsiversive pursuit was
opposite that seen for the same direction of pursuit with
intermediate and deep layer stimulation, and there was no effect on
contraversive pursuit with stimulation of the superficial layers.
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 12.
Differential effects of intermediate and superficial layer stimulation
of the rostral SC on pursuit eye velocity. A: pursuit
eye velocity with stimulation (thick lined traces) and without
stimulation (thin lined traces) for both directions of pursuit. The
filled bars indicate the duration of the SC stimulation.
Top: traces are with intermediate layer stimulation;
bottom: traces are with superficial layer stimulation.
B: the average eye speed in the 2nd 50 ms of pursuit
initiation is plotted as a function of SC stimulation site.
Top: ipsiversive pursuit; bottom:
contraversive pursuit. Eight of 8 ipsiversive cases had significantly
different effects on the eye movement between the superficial and
intermediate layer stimulation. Seven of 8 cases differed significantly
in the contraversive direction.
|
|
We replicated this experiment at eight sites in two monkeys and found
similar effects in all cases. We plotted the difference in eye speed
between the stimulation and no-stimulation conditions for each of the
paired sites, one in the superficial layers and one in the intermediate
and deep layers (Fig. 12B). For ipsiversive pursuit, all
eight sites had statistically significant differences in eye speed
between stimulation in the intermediate and deep layers compared with
the superficial layers. For contraversive pursuit, seven of the eight
sites were significantly different in the two conditions. Thus the
pursuit effects we report cannot be explained by the activation of
superficial layer visual neurons.
Muscimol inactivation of the SC
In contrast to electrical stimulation, which activates the
underlying neuronal elements, muscimol injections should reduce the
activity of the underlying neurons and produce opposite behavioral effects. Therefore if the signal generated by SC neuronal activity is
used by the smooth-pursuit system, we expect unilateral inactivation of
the SC to produce a reduced smooth eye velocity when the pursuit target
is located within the region of visual space represented by the
inactivated part of the SC (Fig.
13). To determine this, we inactivated
the SC while monkeys performed smooth-pursuit eye movements in
both directions. For these experiments, we switched to the step-ramp
paradigm so that we could independently assess the interactions between
the location of the target step and the direction of target motion with
the location of the SC map inactivated.
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 13.
A schematic representation of what would be expected of pursuit if the
signal generated in the SC were removed with muscimol.
A: the ovals represent the SC map showing an example
injection site in the right SC where stimulation would evoke a 2.0°
leftward saccade. This is a schematic injection and does not indicate
the actual spread of muscimol, particularly since injections centered
on this location affected saccades as large as 10° (see
RESULTS). B: the vertical line indicates the
degrees of visual angle with negative values indicating leftward
positions represented in the right SC. The effect on pursuit is
represented by the position traces; thick lines, eye position expected
after the injection; thin lines, the eye position before the injection.
Upward traces are rightward eye movements. With a visual target moving
20°/s leftward (dotted line beginning at "target onset") and
inactivation of 2.0° leftward site in the SC, we would expect to slow
down leftward pursuit. In contrast, rightward pursuit would remain
unchanged. C: the same example is provided but in the
velocity domain. The conventions are the same as in B.
|
|
SACCADES.
As we did with electrical stimulation, we first confirmed that our
muscimol injections affected saccade generation as has been reported
previously (e.g., Aizawa and Wurtz 1998; Quaia et al. 1998). To do this, we interleaved with the pursuit trials target step trials with stimuli located at 2, 5, 10, and 20° along the horizontal meridian in both hemifields and measured the saccades made to the target steps before and after the injections. We observed effects on saccades similar to those reported previously. First, contralateral saccades were hypometric, had reduced velocities and
increased latencies. Second, for many sites, even though the injection
was centered on a site evoking a 1° saccade with stimulation, saccades as large as 10° were affected by the muscimol. Third, in
some cases, ipsilateral saccades occurred with a slightly shorter latency (data not shown). Finally, shortly after the injection (within
5 min), a static fixation offset into the ipsilateral hemifield
developed (see also Hikosaka and Wurtz 1983).
PURSUIT.
We made eight unilateral SC injections in two monkeys. Seven of the
injections were between 0.2° and 3.5° sites. One injection was made
at an 8.8° site (see Table 1). Eye-velocity traces before and
after a single muscimol injection into the rostral SC are shown in Fig.
14 and reflect a typical finding. For
contraversive pursuit initiation, reduced eye velocity was observed,
sometimes even when the target started slightly in the ipsilateral
hemifield (Fig. 14G). For large steps into the ipsilateral
visual field, pursuit was frequently enhanced (Fig. 14F). In
contrast, an injection made more caudally had no effect and an
injection made in the superficial layers had effects of the opposite
sign (not shown).
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 14.
Muscimol inactivation of the SC influences smooth pursuit. The
schematic depicts the location of the target onset and direction of
target motion. Negative values indicate target steps into the
ipsilateral hemifield and positive values indicate target steps into
the contralateral hemifield. Thick lines, eye-velocity traces after the
injection; thin lines, before the injection. Left:
ipsiversive pursuit velocities; right: contraversive eye
velocities. The traces were truncated at the time of saccade
occurrence. Horizontal dashed lines indicate 0°/s.
|
|
To quantify the effects of inactivation of smooth-pursuit eye
movements, we measured pursuit eye velocity in the second 50 ms of the
initial 100 ms, open-loop pursuit response. We made this measurement
before and after the injections of muscimol and plotted the average eye
speed after the injection against the average eye speed before the
injection. We divided the data points into four categories
corresponding to the four plots in Fig.
15. The categories were: data in which
the target stepped into the contralateral hemifield and moved
contraversively (Fig. 15A), data in which the target stepped
into the contralateral hemifield and moved ipsiversively (Fig.
15B), data in which the target stepped into the ipsilateral
hemifield and moved ipsiversively (Fig. 15C), and data in
which the target stepped into the ipsilateral hemifield and moved
contraversively (Fig. 15D). This allowed us to determine whether the effects of rostral SC inactivation depended on the direction of pursuit or the location of the target step, similar to but
not identical to, the directional and retinotopic distinctions found
with MT and MST lesions (e.g., Dürsteler and Wurtz
1988; Dürsteler et al. 1987).
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 15.
Muscimol effects depended mostly on the location of
the target step rather than the direction of pursuit. Average eye
speed, taken from desaccaded traces, in the 1st 100 ms of pursuit after
the muscimol injection is plotted against the same interval before the
muscimol injection. Each plot contains 3 points per injection site. Two
points are for the step-ramp conditions in which the target stepped
into either hemifield, and 1 point is for the ramp condition when the
target motion began at the fovea. The data are taken from 7 injections
in 2 monkeys. The injection site at 8.8° on the SC map is not
indicated. A and C: data from ipsiversive
pursuit trials; B and D: data from
contraversive pursuit trials.
|
|
The reduction of pursuit eye velocity when monkeys tracked a target
moving in the contralateral hemifield was evident in virtually every
case (Fig. 15A). Furthermore many of the injections (not in
the example case shown) revealed that eye speed was enhanced during
ipsiversive pursuit (Fig. 15C), suggestive of a directional deficit in smooth-pursuit eye movements. However, comparing the cases
when the target stepped into the contralateral hemifield and moved
ipsiversively revealed a trend toward a reduction in pursuit eye speed,
indicating that step location was important (Fig. 15B). When
the target stepped into the ipsilateral hemifield and moved
contraversively, a trend toward enhanced pursuit eye speed was evident
(Fig. 15D). In sum, these results suggest that direction and
location are both affected by SC inactivation. However, the direct
comparison of both variables reveals a dominant effect from the
location of the step (cf. Fig. 15, A and D).
There are a few points in the plots inconsistent with that
interpretation; however, consideration of additional facts reveals most
are consistent. First, one statistically significant point in the
contralateral-contraversive condition falls above the line, indicating
that the eye speed was enhanced (Fig. 15A). Additionally, in
the ipsilateral-ipsiversive condition one point, which was statistically significant, falls below the line, indicating pursuit eye
speed was reduced. These points come from the same injection that was
made slightly more dorsal to the site influencing saccades. This result
may have occurred due to inactivation of superficial layer neurons.
Second, in the contralateral-ipsiversive plot (Fig. 15B),
there is one statistically significant point above the unity line,
indicating pursuit eye speed was enhanced. This injection site was
centered on a location very close to the fovea (0.2°) and may reflect
an inactivation of neurons coding slightly ipsilateral target
locations, tuning that is seen in some rostral SC neurons (Krauzlis et al. 1997c, 2000). Alternatively, since the
injection site was so close to zero and the step location was large for this point, the result may reflect a disinhibition of SC locations beyond the inactivated region (Munoz and Istvan
1998). Finally, there are two statistically significant points
falling below the unity line in the ipsilateral-contraversive plot
(Fig. 15D), demonstrating that pursuit eye speed was
reduced. Both of these points are from two different cases in the
condition in which the target step was 2°. With the static fixation
offset resulting from the injection averaged across trials, the target
step was actually located at 1° ipsilateral. Except for the fact that
the target would soon be in the contralateral hemifield, it is unclear
why this resulted in a reduction in pursuit eye speed.
In sum, after inactivation of the rostral SC, contraversive pursuit is
reduced and ipsiversive pursuit is frequently enhanced. For most
points, the location of the target step was a better predictor of the
pursuit deficit than was the direction of pursuit. An injection made
more dorsally increasing the likelihood of affecting superficial layer
neurons, affected pursuit in a manner predicted by the stimulation
results namely, ipsiversive pursuit was suppressed and
contraversive pursuit was enhanced. Finally, an injection located in the more peripheral representation was ineffective at
influencing pursuit eye movements, also consistent with the stimulation
results and indicating that the inactivation is not a nonspecific
effect of reduced SC activity. Inactivation of the rostral SC
influences pursuit in a manner opposite the effects resulting
from electrical stimulation, namely pursuit in the ipsilateral hemifield is reduced with electrical stimulation and pursuit in the
contralateral hemifield is occasionally enhanced with electrical stimulation. Thus we conclude that the activity of the rostral SC
influences smooth-pursuit eye movements.
| |
DISCUSSION |
The results of our experiments demonstrate that the rostral SC
influences smooth-pursuit eye movements. Activation of the rostral SC
with electrical stimulation modifies pursuit in a manner similar to
that for saccades (Fig. 2). Such stimulation reduces smooth-pursuit eye
velocity ipsilateral to the site of stimulation and in some cases, also
contralateral. Depending on the site of the stimulation, contraversive
smooth eye velocity also could be slightly enhanced. We demonstrated a
number of stimulation effects: the stimulation effects vary with the
location of the stimulation on the SC map with most rostral sites
producing the most effective suppression of ipsiversive pursuit (Figs.
4 and 5); the effects can occur immediately with the onset of pursuit (Figs. 6); pursuit eye movements are more affected as pursuit speed
increases and are not evoked by stimulation during fixation (Figs. 7
and 8); the stimulation effects are evident for both pursuit initiation
and maintenance (Figs. 9 and 10); the effects are not due to a
generalized suppression because smooth eye movements evoked by head
rotations are largely unaffected (Fig. 11); and it is the intermediate
and deep layers of the SC that are responsible for the effects since
superficial layer stimulation had either the opposite or no effect on
pursuit (Fig. 12). In contrast to the stimulation, reduction in the
activity of the rostral SC by injection of muscimol affects pursuit in
an opposite manner; such injections reduce contraversive pursuit
velocity and ipsiversive pursuit velocity is frequently enhanced (Figs.
14 and 15). These results of activating and inactivating the neurons in
the rostral SC indicate that the activity of SC neurons contributes to
smooth-pursuit eye movements. Moreover the results are consistent with
the interpretation that the SC provides a position signal that can be
used to drive the pursuit response.
The relationship of the results to the predictions of the position
signal hypothesis outlined in the previous paper is not always obvious
so we will consider the relationship in several steps. First, we will
compare the effects of the rostral SC alterations on pursuit with those
we obtained for saccades to compare the deficits of the pursuit system
to those of the saccadic system, which is known to be driven by
position signals. Next we will consider the extent to which the results
of the pursuit experiments fit with the position signal hypothesis.
Finally, we consider the possibility that the signal we are altering in
the SC is a motion signal rather than a position signal.
Pursuit and saccade effects are similar
The first indication that effects on pursuit result from SC
activity related to position is the similarity of activation and inactivation of the rostral SC on pursuit and saccades. For pursuit of
targets moving in the ipsiversive direction, activation of the rostral
SC slows pursuit of moving targets just as such stimulation suppresses
saccades to ipsilateral stationary targets. In contrast, for pursuit of
targets moving in the contraversive direction, activation of the
rostral SC either increases pursuit speed or has little effect just as
such stimulation has less effect on saccades made to large,
contralateral stationary targets. When the SC is inactivated by
muscimol injections, the results are largely opposite the stimulation
results for both pursuit and saccades. Inactivation of the rostral SC
leaves pursuit of ipsiversive targets unaffected or speeds it. Saccades
to stationary, ipsilateral targets are largely unaffected or occur with
a slightly shorter latency (see also, Aizawa and Wurtz
1998; Quaia et al. 1998). Inactivation decreases
pursuit speed to contraversive targets and small, contralateral
saccades are hypometric. Our experiments show that activation and
inactivation have opposite effects on pursuit, that these effects are
different for ipsiversive and contraversive movements, and that the
type of change is similar for both pursuit and saccades. Thus the
current results provide evidence that the activity in rostral SC
influences pursuit comparable to the previous evidence that this region
influences saccades. Moreover these results suggest that the two types
of eye movements are influenced by the same signal coded by rostral SC neurons.
The nature of the activity in the rostral SC fixation neurons has been
proposed to signal attentive fixation (Munoz and Wurtz 1993a,b). The tonic activity of rostral neurons during
fixation, the pause in activity of these neurons during saccades, the
suppression of saccadic eye movements with electrical stimulation, and
the shorter latency of large saccadic eye movements resulting from temporary inactivation of these neurons support that interpretation. Our results on pursuit are very similar to those for saccades, and it is therefore possible that the effects we see on pursuit result
from a general signal that acts to suppress eye movementspursuit as
well as saccadesand produce fixation. In the present experiments, when the stimulation was as close to the representation of 0° eccentricity on the SC map as was possible, we also found suppression of pursuit and saccades both ipsiversively and contraversively. We
think this activity reflects small target positions adjacent to the
fovea as described in our previous papers (Krauzlis et al. 1997,
2000) rather than a static fixation command for three reasons.
First, despite a suppression of pursuit in both directions, the effects
were asymmetricalstimulation reliably suppressed ipsiversive pursuit
more than contraversive pursuit. Second, for some stimulation sites
where fixation neurons could be recorded, stimulation suppressed
ipsiversive saccades and pursuit but facilitated contraversive pursuit,
albeit modestly. Thus one stimulation site both increased and decreased
pursuit. Third, smooth eye movements produced by head rotations were
largely
TOP
ABSTRACT
INTRODUCTION
