Articles in PresS. J Neurophysiol (April 8, 2015). doi:10.1152/jn.00212.2015
1
2
ADAPTATION AND ADAPTATION TRANSFER CHARACTERISTICS OF
FIVE DIFFERENT SACCADE TYPES IN THE MONKEY
3
4
5
Yoshiko Kojima, Albert F. Fuchs and Robijanto Soetedjo
6
7
8
9
Department of Physiology and Biophysics and Washington National Primate
Research Center, University of Washington, Seattle, Washington 98195-7330,
USA.
10
11
Running Head: saccade adaptation characteristics in the monkey.
12
13
14
15
16
17
18
Contact information:
Yoshiko Kojima
1959 NE Pacific St. HSB I421, Washington National Primate Research Center,
Box 357330
University of Washington, Seattle, WA 98195-7330
Email: [email protected]
19
1
Copyright © 2015 by the American Physiological Society.
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Abstract
Shifts in the direction of gaze are accomplished by different kinds of
saccades, which are elicited under different circumstances. Saccade types
include targeting to simple jumping targets, delayed to visible targets after a
waiting period, memory-guided (MG) to remembered target locations, scanning
to stationary target arrays and express after very short latencies. Studies of
human cases and neurophysiological experiments on monkeys suggest that
separate pathways, which converge on a common locus that provides the motor
command, generate these different types of saccade. When behavioral
manipulations in humans cause targeting saccades to have persistent dysmetrias
as might occur naturally from growth, aging and injury, they gradually adapt to
reduce the dysmetria. Although results differ slightly between laboratories, this
adaptation generalizes or transfers to all the other saccade types mentioned
above. Also, when one of the other types of saccade undergoes adaptation, it
often transfers to another saccade type. Similar adaptation and transfer
experiments, which allow inferences to be drawn about the site(s) of adaptation
for different saccade types, have yet to be done on monkeys. Here we will show
that simian targeting and MG saccades adapt more than express, scanning and
delayed saccades. Adaptation of targeting saccades transfers to all the other
saccade types. However, the adaptation of MG saccades transfers only to
delayed saccades. These data suggest that adaptation of simian targeting
saccades occurs on the pathway common to all saccade types. In contrast, only
the delayed saccade command passes through the adaptation site of the MG
saccade.
44
45
46
47
Keywords
Saccade, Adaptation, Motor learning, Plasticity, Monkey.
48
2
49
Introduction
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
Plasticity is an important property of the nervous system. In the case of
movement, it helps to overcome persistent motor errors caused by growth,
injury and aging. Such motor plasticity or adaptation can be investigated quite
handily for saccadic eye movements, which shift the direction of gaze between
visual targets of interest. To cause adaptation, the target can be displaced during
the saccade so that the saccade appears to have the wrong amplitude or direction.
When faced with such repeated apparent errors, the oculomotor system
gradually adapts the size and/or direction of its saccades so that gaze eventually
lands near the displaced target (McLaughlin 1967). This adaptation paradigm
has been used in primates to infer the site(s) and mechanism of adaptation in the
saccade system.
Primate saccades can be elicited in a variety of situations. We consider
5 types here. The most natural type is the scanning saccade (S), which we use
to redirect our gaze from one interesting, usually stationary, object in space to
another, e.g., during reading. The remaining types are elicited primarily in the
laboratory in response to jumping spots, whose timing generally produces
different response latencies. Express saccades (E), which have the shortest
latencies (80-120 ms) (Fischer and Boch 1983; Fischer and Ramsperger 1984;
Fischer et al. 1993), occur if a target that is being fixated disappears before a
distant target appears. Typical latencies for targeting saccades (T) to
sequentially illuminated spots are ~150 ms or greater in highly trained monkeys.
Saccades to a distant target can be delayed (D) if the monkey is required to
maintain fixation until the fixation target is extinguished. Memory-guided
(MG) saccades occur when the distant target appears only briefly and the
monkey again must maintain fixation until the disappearance of the fixation
target instructs it to saccade to the remembered location of the extinguished
target. Although T saccades in the monkey have been shown to undergo
adaptation, adaptation of the other simian saccade types has not been studied
systematically. Therefore, the first goal in this paper is to remedy that gap in
our knowledge by documenting and comparing the characteristics of adaptation
for all 5 saccade types.
3
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
Knowledge of the pattern of adaptation transfer or generalization
between saccade types allows us to speculate whether sites of adaptation are
shared by different saccade types or whether an adaptation site is unique for a
particular type. Adaptation transfer between several saccade types has been
studied in humans. For example, although results differ somewhat between
laboratories, the adaptation of T saccades transfers to E, S, D and MG saccades
(Alahyane et al. 2007; Collins and Doré-Mazars 2006; Cotti et al. 2007; Deubel
1995; 1999; Fujita et al. 2002; Hopp and Fuchs 2010; 2002; Panouillères et al
2014; Zimmermann and Lappe 2009; for reviews see Table 1 in Hopp and
Fuchs 2004; Pelisson et al. 2010). However, the adaptation of MG saccades
transfers only to D saccades (Fujita et al. 2002; Deubel 1995; 1999; Hopp and
Fuchs 2010; 2002). Therefore, the adaptation of T and MG saccades occurs at
different neuronal loci. Although this information does not allow us to identify
specific neural adaptation sites, it suggests that the saccade commands for all of
the other four types pass through the T saccade adaptation site. In contrast, only
the D saccade command passes through the MG saccade adaptation site.
The only simian transfer study thus far revealed that T saccade
adaptation also transfers robustly to the other four types (Fuchs et al. 1996),
again suggesting that their saccadic commands pass through the T saccade
adaptation site. Neurophysiological studies indicate that T and D saccade
adaptation occurs downstream of the superior colliculus (T saccade: Edelman
and Goldberg 2002; Frens and Van Opstal 1997; Melis and van Gisbergen 1996;
cf. Takeichi et al. 2007, D saccade: Quessy et al. 2010). Neurophysiological
studies during adaptation suggests that one downstream structure that likely
participates in T saccade plasticity is the oculomotor vermis of the cerebellum
(Barash et al. 1999; Catz et al. 2008; 2005; Kojima et al. 2010b; 2011; Soetedjo
and Fuchs 2006; Soetedjo et al. 2008; Takagi et al. 1998). In contrast, nothing is
known about the sites or the neuronal substrate for the adaptation of any other
saccade type. To locate these different sites and to determine the mechanisms
that underlie the adaptations of the different saccade types, we require similar
transfer data in an experimental primate. Therefore, our second goal is to
determine how adaptation of each simian saccade type affects the other four
types. Such data provides a behavioral platform from which to launch the
4
115
116
various neurobiological techniques that can be brought to bear on the behaving
monkey.
117
5
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
Materials and Methods
Surgery and training
We measured eye movements in two rhesus monkeys (Macaca mulatta,
male, Monkey B: 7.4 and Monkey W: 5.0 kg) with the electromagnetic search
coil method (Fuchs and Robinson 1966; Judge et al. 1980; Robinson 1963).
Briefly, in an aseptic surgery, we implanted a three-turn coil of fine
Teflon-coated stainless steel wire on one eye of each monkey. The ends of the
wire terminated in a small connector fixed to the top of the skull with titanium
screws and dental acrylic. In the same surgery, we attached three receptacles to
the monkey's skull so we could stabilize the head during eye-movement
recordings. For more details see Fuchs et al. (1996).
After recovery from the surgery, each monkey was placed in a dimly lit
booth and trained to follow a small visual target with its eyes. The monkey sat
in a primate chair with its head restrained. We surrounded the animal with a
semicircular array of light-emitting diodes (LEDs), which were spaced every 1°
along the horizontal meridian and illuminated them sequentially to produce
target jumps. Each LED subtended 0.25°. The LEDs could be programmed to
emit either a red or a green light. We used the green LEDs only to elicit delayed
and MG saccades (see below). After the monkeys had been trained to fixate an
illuminated LED, i.e., the target, we trained them to make saccades between
sequentially stepping LED targets. If the monkey’s gaze arrived within ±2° of
the target within 0.8 sec of the target step and stayed there for at least 0.5 sec,
the monkey was continually rewarded every 1 – 1.2 sec. Note that 0.8 sec
allows the occurrence of a corrective saccade. We applied the same reward
schedule for all saccade types during pre-adaptation, adaptation and
post-adaptation sessions.
144
145
146
147
148
149
150
Saccade types and their adaptation
The data gathered in this study were obtained before we started the
cerebellar single unit recording experiments reported in Kojima et al. (2010b),
where the procedures for eliciting the 5 types of saccade are described in detail.
Briefly, after the monkeys made targeting saccades continually for at least 1
hour, we attempted to generate express saccades. For each animal, we
6
151
152
153
154
155
156
157
determined which target amplitude elicited a high incidence of saccades with
express latencies (for more detail, see EXPRESS SACCADES section later in
Materials and Methods). Monkey B made a high rate of express saccades to a
10° horizontal target step, whereas monkey W did to a 15° target step. For each
monkey, these same step sizes were used for adaptation and testing of all 5 types
of saccades. Once the animal reliably made saccades with express latencies, we
trained it on the delayed, MG and scanning saccade paradigms.
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
TARGETING SACCADES. After the monkey had fixated a red LED for at least 0.6
sec, another LED was illuminated within the next 0.7 to 0.9 sec and the first
went out. This sequence created a target step to either the right or left that
remained within ± 23° of straight-ahead.
Adaptation of targeting saccades was induced by a modified version of
the intra-saccadic step (ISS) paradigm (McLaughlin 1967). In the McLaughlin
paradigm, the target jumped along the horizontal meridian, and then, during a
targeting saccade, jumped backward so the eye appeared to overshoot the target.
The subject then made a corrective saccade backward to acquire the target. This
procedure, when repeated over several hundred trials, caused a gradual decrease
in the size of the initial saccade. In this study, we did not examine the gradual
gain increases that are produced by a forward ISS. In our modified version of
the paradigm, we caused the error of the targeting saccade to remain a constant
percentage of its amplitude. We employed this adaptation strategy to provide
the same percentage ISS to all 5 saccade types, which had different accuracies.
For example, MG saccades were less accurate than targeting saccades (White et
al. 1994). A computer detected the saccade end position (when the eye velocity
decelerated to ≤ 20°/s) and moved the target back 25% of the target step
amplitude from the saccade end position. Because the LEDs were placed every
1°, the target position was rounded down to the nearest integer. The schematic
at the top of Fig. 1A shows the schematic target movement and elicited saccade
that occurs in the “Targeting” saccade paradigm.
181
182
Figure 1 near here
183
7
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
EXPRESS SACCADES. To elicit express saccades, we extinguished the fixated red
LED briefly (gap paradigm) before illuminating a second LED at a distance of
either 10° for monkey B or 15° for monkey W (Fischer and Boch 1983; Fischer
and Ramsperger 1984; Fischer et al. 1993). We varied the duration of the
extinction gap in 10 ms steps from 80 to 200 ms to determine which duration
elicited the highest incidence of saccades with express latencies. The most
effective gap was 100 ms for monkey B and 160 ms for monkey W. With these
gaps, the distribution of latencies was bimodal with the trough occurring at
~100 ms for monkey B and ~110 ms for monkey W (Fig. 1B, black bars). We
considered as express saccades only those with latencies from 50 to 100 ms for
monkey B and 50 to 110 ms for monkey W.
To adapt express saccades, we measured saccade latencies online and
applied an adapting step only if the latency was <100 ms for monkey B and
<110 ms for monkey W (Fig. 1A “Express”). The target was stepped backward
25% of the saccade size from the saccade end position. If a saccade latency was
greater than the maximum accepted express value, there was no back-step and
instead the target was turned off for 300 ms. We used this short duration here to
gather the maximum number of saccades before the monkey became fatigued.
202
203
204
205
206
207
208
209
210
211
212
213
214
215
SCANNING SACCADES. To elicit scanning saccades, we simultaneously
illuminated 3 red LEDs, each separated by either 10° (monkey B) or 15°
(monkey W) along the horizontal meridian and rewarded the monkey whenever
it made, at its own pace, more than 3 consecutive saccades between any of the
three LEDs. Monkeys most often made a saccade to the adjacent LED, i.e., 10°
(monkey B) or 15° (monkey W).
To adapt scanning saccades, all three illuminated LEDs were
extinguished at the end of the saccade and 3 other LEDs located short of each
LED target were illuminated to produce a backward-step of the entire 3 LED
array by 25% of the saccade amplitude from the saccade end position (Fig. 1A
“Scanning”). The occasional (~1%) saccades that were made between the two
most eccentric LEDs were not analyzed. However, the back step also occurred
for those saccades.
216
8
217
218
219
220
221
222
223
224
225
226
227
228
DELAYED SACCADES. To elicit a delayed saccade, we required the monkey to
continue to fixate an illuminated red LED for 0.3 sec before a peripheral green
LED was illuminated continuously to show the destination for the future
saccade. We required the monkey to delay its saccade to the green target by
continuing to fixate the red LED for an additional 0.3 sec. When the fixated red
LED went off, it signaled the animal to make a saccade to the peripheral,
illuminated green LED. After the monkey had fixated the green LED for 0.6
sec, the red LED was turned on in place of the green LED to serve as the initial
fixation spot for the next trial. We used short fixation durations here to
maximize the number of saccades before the monkey became fatigued.
To elicit adaptation, the green LED was stepped backward by 25% of
the saccade amplitude from the saccade end position (Fig. 1A “Delayed”).
229
230
231
232
233
234
235
236
237
238
239
240
MG SACCADES. To elicit a MG saccade, we required the monkey to fixate a red
LED for 0.3 sec and then we turned on a green LED briefly for 0.2 sec at the
next target location. After the green LED was extinguished 0.3 sec later, the red
fixation LED was also turned off, serving as a signal for the monkey to saccade
to the remembered location of the green LED. When the saccade landed at the
remembered target location, the green LED came on at the actual location.
After the monkey had fixated the green LED for 0.6 sec, the red LED was
turned on in its place to serve as the initial fixation LED for the next trial.
To produce MG saccade adaptation, the green LED was stepped
backward by 25% of the saccade amplitude from the saccade end position (Fig.
1A “Memory guided”).
241
242
243
244
245
246
247
248
249
Experimental conditions
Once the monkey had demonstrated its ability to perform the various
saccade tasks and adapt all types of saccade, we began the transfer experiments.
In each experiment, we induced adaptation of one saccade type (the adapt
saccade) and tested whether that adaptation transferred to one or more (up to
four) of the other types (the test saccades).
Figure 1C illustrates the time course of an experiment in which one
saccade type was adapted and 3 other types were tested. Before adaptation (Fig.
9
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
1C, Pre), we collected ~30 of the saccades to be adapted and ~30 of each of the
3 saccade types to be tested. We then reduced the amplitude of the adapted
saccade by jumping the targets backward as the monkey made saccades toward
them. The backward-step was applied for saccades in both the left- and
rightward directions. After each ~ 10 min segment of adaptation, we collected
~20 of one of the three test saccade types so that by the end of the entire
adaptation, each test saccade had been sampled three times. During a ~10 min
segment of adaptation, a monkey adapted approximately 370 targeting saccades,
180 express saccades, 280 delayed or MG saccades and 550 scanning saccades
in each direction. The backward-step was not applied during the collection of
test saccades.
After the ~20 test saccades, the amplitude of the adapt saccade had
shown little or no recovery. Across both monkeys, the average gain recovery
was 0.020 ± 0.028, 0.014 ± 0.016, 0.012 ± 0.024, 0.017 ± 0.015, and 0.022 ±
0.026 for targeting, express, scanning, delayed and MG saccade adaptation,
respectively. These gain recoveries did not differ significantly between saccade
types (unpaired t test, p > 0.05 with Bonferroni correction). Therefore, we
ignored it in further analyses.
Finally, we discontinued the back-step and monitored the recovery of
the adapted saccade gain (Fig. 1C, Recovery). For most experiments in monkey
B, we were able to test four types of saccade in each adaptation experiment. For
most experiments in monkey W, who did not work as long, we could test only
two types of saccade in each experiment.
273
274
275
276
277
278
279
280
281
282
Data analysis
We did the initial data acquisition and the on-line data processing as
described in Kojima et al. (2010a). Briefly, eye and target position signals were
digitized at 1 kHz and stored on a hard disk through an interface (Power 1401,
Cambridge Electronic Design). During the experiment, we used a custom
program running in Spike 2 (CED) to control the data acquisition box (Power
1401, CED) and to analyze the incoming data on-line. The data were saved
after the experiment and analyzed off-line on a computer using custom
programs that ran analysis software (Spike 2, CED). Saccade onset and end
10
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
were identified by an eye velocity criterion of 20°/s. Saccades elicited by the
initial (not intra-saccadic) target jump were selected for analysis. Parameters of
saccades and target jumps, e.g., latency, onset, offset, amplitude, duration, peak
velocity, etc., and the saccade traces for every trial were exported to other
programs (Matlab, MathWorks) for subsequent analysis. We eliminated
saccades with a vertical component that was ≥ 3 standard deviations from the
mean and if the horizontal and vertical initial eye positions differed from the
initial target positions by ≥ 1°. These restrictions eliminated < 5% of the trials.
We calculated gain as [saccade size / target step size]. Before adaptation,
the gain of some saccade types in both monkeys was idiosyncratically different.
In monkey B, scanning saccades showed significantly higher gains than the
other 4 types (unpaired t test, p < 0.05 with Bonferroni correction). In monkey
W, delayed saccades showed significantly higher gains than express saccades
(unpaired t test, p < 0.05 with Bonferroni correction, here and for all other t
tests). Therefore, in each monkey, we normalized the gain of each saccade type
as [saccade gain / mean pre-adaptation gain].
The gain change of the adapted saccade was calculated as [mean
normalized gain of the first 15 saccades of adaptation (black rectangle in Fig.
1C “first15”) – mean normalized gain of the last 15 saccades (black rectangle in
Fig. 1C “last15”)]. To assess whether the adaptation was significant, we
compared the mean normalized gain of the first and last 15 adapt saccades
(unpaired t test, p < 0.05). Saccades in all of the data sets used in this study
showed a significant gain reduction during adaptation (unpaired t test, p < 0.05).
We treated rightward and leftward saccades as independent data sets because
adaptation in one horizontal direction does not produce adaptation in the
opposite direction (human: Deubel et al. 1986; Miller et al. 1981; Semmlow et
al. 1989, cf., Rolfs et al. 2010, or monkeys: Straube et al. 1997; Scudder et al.
1998; Noto et al. 1999), so each adapted direction provides a separate data set.
The average number of rightward saccades during the adaptation session was
827±321, 593±292, 945±584, 411±108, 429±103 for targeting, express,
scanning, delayed and MG saccades, respectively. For leftward saccades, it was
767±216, 569±232, 945±598, 435±104, 482±105 for targeting, express,
scanning, delayed and MG saccades, respectively.
11
316
317
To compare adaptation characteristics across saccade types, we also fit
the course of adaptation with an exponential function:
318
319
f(x) = a*exp(-b*x) + c
(eq. 1)
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
Because all adaptations had at least 300 adapt trials, we determined the gain
change after 250 saccades as [mean normalized gain of the first 15 saccades of
adaptation – f(250) of the exponential fit]. Finally, we calculated the initial
jump of the adaptation as [mean normalized gain of the last 15 saccades of
pre-adaptation - mean normalized gain of the first 15 saccade of adaptation] and
the recovery jump at the onset of recovery as [mean normalized gain of the first
15 saccades of recovery - mean normalized gain of the last 15 saccades of
adaptation].
To assess the adaptation transfer to the test saccades, we applied an
unpaired t-test to the last 15 test saccades during pre-adaptation and the first 15
test saccades of the 3rd test saccade block (see Fig. 1C). If the normalized gain
of saccades in the third test block was not significantly different than the
normalized gain of pre-adaptation test saccades (unpaired t test, p > 0.05), we
considered that adaptation had not transferred to the test saccade, i.e., the
percentage transfer was 0% (Fuchs et al. 1996). On the other hand, if the
normalized gain of the test saccades was smaller and the normalized gain
changes of the adapted and test saccades were not significantly different
(unpaired t test, p > 0.05), we considered that adaptation transfer was complete,
i.e., the percentage transfer is 100%. If the latter t-test was significant (unpaired
t test, p < 0.05), we then calculated the percentage transfer as:
341
342
Percentage transfer = [(GpreT - GadaT) / (GfirstA – GlastA)] * 100
(eq. 2)
343
344
345
346
347
348
In this equation, GpreT is the mean normalized gain of the last 15 test saccades of
preadaptation, GadaT is the mean normalized gain of the first 15 test saccades of
the 3rd test, GfirstA is the mean normalized gain of the first 15 saccades of
adaptation, and GlastA is the mean normalized gain of the last 15 adapted
saccades before the 3rd test (black box #1, 2 or 3 “for Transfer test”). We chose
12
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
these gain measures so that the target steps were always the same for the test
saccades (GpreT and GadaT, without ISS) and the adapted saccades (GfirstA and
GlastA, with ISS).
As will be seen in the Results, there often is an immediate drop in gain at
the beginning of adaptation. Therefore, we also calculated the percentage
transfer in a second way, which took the drop into account by changing the
denominator of eq. 2 to (GpreA – GlastA), where GpreA is the mean normalized
gain of the last 15 adapt saccades of preadaptation (use of this modified eq. 3
produced the grey bars in Fig. 5). In this case, the target condition differs
between GpreA (without ISS) and GlastA (with ISS).
To avoid drawing conclusions based only on non-significant results (see
the description of eq. 2), we used the same equation to calculate the actual
percent transfer without applying the adjustment for no (0%) or complete
(100%) transfer (use of this modified eq. 4 produced the striped bars in Fig. 5).
These two calculations did not change the conclusions of our results.
Because the transfer of adaptation varied from experiment to
experiment (Fuchs et al. 1996), we collected 10 data sets for each pair of adapt
and test saccade types for each monkey so that the total permutation of the
transfer data sets was 10 (data sets) * 5 (adaptation saccade types) * 4 (other test
saccade types) * 2 (monkeys) = 400. To infer whether the adaptation
transferred to a particular test saccade on average, we tested the mean
percentage transfer obtained from the 2 monkeys (n = 20 datasets, t test). If it
was significantly different from 0 (unpaired t test, p < 0.05 with the Bonferroni
correction), we inferred that the adaptation had transferred to the test saccade
type. Because the number of saccade types tested in each adaptation
experiment varied, the number of adaptation data sets for each saccade type did
as well. For monkey B, the sets numbered 20, 18, 24, 10, and 10 for targeting,
express, scanning, delay and MG saccades, respectively. For monkey W, they
numbered 22, 20, 20, 20, and 24, respectively.
All experiments were performed in accordance with the Guide for the
Care and Use of Laboratory Animals (1996) and exceeded the minimal
requirements recommended by the Institute of Laboratory Animal Resources
and the Association for Assessment and Accreditation of Laboratory Animal
13
382
383
384
Care International. All the procedures were evaluated and approved by the
local Animal Care and Use Committee of the University of Washington
(ACC#2602-01).
385
14
386
387
388
389
390
Results
We induced gain decrease adaptation of targeting, express, scanning,
delayed or MG saccades and tested the transfer of that adaptation to one or more
of the other 4 types. We first will describe the adaptation characteristics of each
type of saccade and then the transfer of that adaptation.
391
392
The course of adaptation of different saccade types
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
All types of saccade showed significant adaptation. However, the
amount of gain change during adaptation varied from one saccade type to
another. Figure 2A-E shows representative courses of adaptation and recovery
of the five types of saccades for monkey B. Note that although other types of
saccade were inserted and tested during the adaptation (see Fig. 1C), we have
plotted only the adapted saccades. Also recall that the 20 test saccades inserted
at various times during adaptation caused very little recovery of the adapted
saccade (see Materials and Methods).
As described in (Straube et al. 1997), the course of targeting saccade
adaptation (Fig. 2A) was nicely fit with an exponential function (grey line, r =
0.73, rate constant = 646). There was an initial jump in gain when we stopped
the target back-step (vertical dashed line between “Adapt” and “Recovery”),
and the subsequent recovery was gradual. The adaptation courses of express,
scanning, delayed and MG saccades (Figs. 2B-E) were qualitatively similar but
exhibited some quantitative differences. For example, express, scanning and
delayed saccades adapted less than targeting saccades, whereas MG saccades
adapted the same amount as targeting saccades but in fewer trials. Also,
scanning saccades showed an initial jump decrease in gain at the onset of
adaptation, whereas the other types did not.
Fig. 2F shows representative courses of adaptation for each saccade type
in monkey W. They were qualitatively similar to their counterparts from
Monkey B.
416
417
Figure 2 near here
418
15
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
We compared five measures that describe adaptation for the five
different types of saccade. The mean gain decrease at the 250th saccade of
adaptation was greater for targeting and MG saccades than for the other 3 types
in both monkeys (Fig. 3A white bars). For the average from both monkeys, MG
saccades exhibited a significantly greater reduction than did scanning saccades
(unpaired t test, p<0.05 with Bonferroni correction, Fig. 3A black bars). In
monkey B, the initial jump at the onset of adaptation was larger for scanning
saccades than for the other 4 types; however, it was the smallest among all test
saccades in monkey W (Fig. 3B white bars). Finally, the jump at the beginning
of recovery was significantly larger for MG saccades than for delayed, scanning
and express saccades in both monkeys (Fig. 3C).
To analyze the time course of adaptation for the different saccade types,
we compared the rate constants (1/b in eq. 1) and the absolute gain changes (a
in eq. 1) of the exponential fits of adaptation. Because we did not always collect
adaptation data until the gain change had reached a plateau, rate constants and
the absolute gain changes might be slightly inaccurate in some experiments.
Fig. 3D shows the distribution of the width of 95% confidence interval for the
rate constants of the exponential fits of all adaptations. Because the width of the
95% confidence interval was <1000 in most data sets, we excluded those with
larger widths in our consideration of rate constants and absolute gain changes.
The number of experiments remaining after this exclusion was 32, 18, 33, 18,
27 for targeting, express, scanning, delayed, and MG saccade adaptation,
respectively. Across saccade types, the rate constant was significantly greater
for targeting saccades, and smaller for MG saccades than for the other 4 saccade
types (unpaired t test, p<0.05 with Bonferroni correction, Fig. 3E black bars).
The absolute gain changes for targeting and MG saccades were larger than those
of the other saccade types (Fig. 3F black bars). In summary, targeting and MG
saccades exhibited the greatest amount of adaptation, but MG saccade
adaptation was faster and showed a larger recovery jump. Other saccade types
were not significantly different from each other.
449
450
Figure 3 near here
451
16
452
The transfer of adaptation from one saccade type to another
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
ADAPTATION TRANSFER FROM TARGETING TO OTHER SACCADE TYPES
Figure 4 shows data on the transfer from adapted saccades to test
saccades from representative experiments in monkey B. Although we tested
three types of saccade during every experiment (see Materials and Methods),
each panel here shows the data associated with only the adapted saccade (gray)
and one type of test saccade examined three times during the course of
adaptation (black). For example in column “T” (Fig. 4), the plots in rows E, D,
and MG are from the same experiment, i.e., B627. Because test saccade #1 was
MG, #2 was delayed, and #3 was express (see Fig. 1C), the number of adapt
saccades is smallest for the first-tested MG saccades and largest for the
last-tested express saccades.
As the gain of targeting saccades decreased (Fig. 4, column T), the gains
of all types of test saccades exhibited a gradual decrease as well, albeit smaller.
The mean percent gain decrease of the adapted targeting saccades calculated
from data taken just before and just after the 3rd set of data from the test saccade
(see Materials and Methods) ranged from 21.6 ± 7.7 to 27.0 ± 11.3% (4 data sets
in column T, bar graphs, grey bars) and was always highly significant (p<0.05,
unpaired t-test, * in grey bars). In contrast, the mean gain decrease for tested
express, scanning, delayed and MG saccades was only 16.1 ± 9.0, 19.5 ± 10.3,
8.4 ± 6.6 and 13.5 ± 6.6%, respectively (bar graphs, black bars); however, each
decrease was significant (p<0.05, * in black bars). The percent gain decreases
of all test saccade types were significantly less than those of the adapted
targeting saccade (* above bar graphs). For these exemplar experiments, the
percent gain transfer from targeting saccades was 70.9, 72.2, 38.9 and 57.9% to
express, scanning, delayed and MG saccades, respectively.
479
480
Figure 4 near here
481
482
483
484
Figure 5 shows the percent transfer from adapted targeting saccades to
the various test saccades. In column T, each white bar represents data
calculated using eq. 2 and averaged across the 10 experiments performed on
17
485
486
487
488
489
490
491
492
493
each monkey and the black bar represents the data averaged across both
monkeys. Across the monkeys, the average percentage gain transfer from
adapted targeting to express, scanning, delayed and MG saccades was 73.6 ±
18.1, 58.1 ± 16.7, 63.8 ± 20.5 and 34.6 ± 16.7% (black bars), respectively, and
was significantly different from zero (* in black bars). Thus, adaptation of
targeting saccades transferred significantly to each of the other 4 types of
saccade. The thin grey and striped bars represent the percent transfer evaluated
by the modified eq. 3 and modified eq. 4, respectively (see Materials and
Methods - Data analysis).
494
495
Figure 5 near here
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
ADAPTATION TRANSFER FROM EXPRESS TO OTHER SACCADE TYPES
Figure 4, column E, shows data for the transfer of adaptation from
express saccades to the four other saccade types taken from an exemplar
experiment in monkey B. The mean gain decreases of the adapted express
saccades before the 3rd tests of the different types were ~20% and always
significant (p<0.05, unpaired t-test, grey bars). Although targeting saccades
showed a significant gain decrease, scanning, delayed and MG saccades did not
(black bars). The gain changes of all the test saccades differed significantly
from that of the adapted express saccade; the gain transfer was 25.6, 0, 0 and 0%
for targeting, scanning, delayed and MG saccades, respectively.
Across both monkeys (Fig. 5, column E), the average percentage gain
transfer from adapted express saccades to targeting, scanning, delayed and MG
saccades was 71.5 ± 22.3, 22.1 ± 19.4, 23.3 ± 20.6 and 18.8 ± 19.9%,
respectively, but the transfer was significantly different from zero only for
targeting saccades. Thus, adaptation of express saccades transfers only to
targeting saccades.
513
514
515
516
517
ADAPTATION TRANSFER FROM SCANNING TO OTHER SACCADE TYPES
Figure 4, column S, shows data for the transfer of adaptation from
scanning saccades to the four other saccade types taken from representative
experiments in monkey B. The mean gain decrease of the adapted scanning
18
518
519
520
521
522
523
524
525
526
527
528
529
saccades was always significant (grey bars). Targeting and express test
saccades also showed significant gain decreases, but delayed and MG saccades
did not (black bars). The gain decreases of targeting and express saccades were
not significantly different from that of the adapted scanning saccade. Therefore,
the percentage gain transfer was 100, 100, 0 and 0% for targeting, express,
delayed and MG saccades, respectively.
Across both monkeys (Fig. 5 column S), the average percentage gain
transfer from adapted scanning saccades to targeting, express, delayed and MG
saccades was 67.1 ± 21.5, 68.6 ± 27.0, 33.0 ± 22.8 and 22.8 ± 20.5%,
respectively; however, the transfer was significantly different from zero only
for targeting and express saccades. Thus, adaptation of scanning saccades
transferred only to targeting and express saccades.
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
ADAPTATION TRANSFER FROM DELAYED TO OTHER SACCADE TYPES
Figure 4, column D, shows data for the transfer of adaptation from
delayed saccades to the four other saccade types taken from a representative
experiment in monkey B. The mean gain decrease of the adapted delayed
saccades was always significant (grey bars) as was the mean gain decrease of all
types of test saccades (black bars). The gain decreases of targeting and
scanning saccades were significantly less than those of the adapted delayed
saccade (* above bar graphs). Therefore, the percentage gain transfer was 48,
100, 72 and 100% for targeting, express, scanning and MG saccades,
respectively.
The average percentage gain transfers from adapted delayed saccades to
other test saccades across both monkeys were 68.8 ± 20.7, 80.2 ± 17.5, 43.6 ±
23.1 and 38.9 ± 24.1% for targeting, express, scanning and MG saccades,
respectively (Fig. 5 column D). The transfer to each type of test saccade was
significantly different from 0. Thus, adaptation of delayed saccades transferred
to all types of saccades.
547
548
549
550
ADAPTATION TRANSFER FROM MG TO OTHER SACCADE TYPES
Figure 4, column MG, shows data for the transfer of adaptation from
MG saccades to the four other saccade types taken from representative
19
551
552
553
554
555
556
557
558
559
560
561
562
experiments in monkey B. The mean gain decrease of the adapted MG saccades
was always significant (grey bars). The tested targeting, express and scanning
saccades showed no significant gain decrease but delayed saccades did (black
bars). The gain decrease for each type of test saccade was significantly
different from that of the adapted saccade, and the percentage gain transfer was
0, 0, 0 and 80.1% for targeting, express, scanning and delayed saccades,
respectively.
In summary (Fig. 5, column MG), the average percentage gain transfer
from adapted MG to targeting, express, scanning and delayed saccades was
14.0 ± 15.8, 15.5 ± 15.9, 18.8 ± 16.4 and 40.3 ± 24.3%, respectively, but was
significantly different from 0 only for delayed saccades. Thus, adaptation of
MG saccades transfers only to delayed saccades.
563
564
20
565
Discussion
566
567
568
569
570
571
572
573
574
575
The objective of this study was to test the adaptation characteristics of
different types of saccades in the monkey and to determine the degree to which
the adaptation of one type generalized to other types. First, we will compare the
adaptation characteristics such as gain decrease, initial and recovery jumps and
the rate constant for the 5 types of saccades we examined. Then we will
consider the adaptation transfers between the different types and draw
inferences as to the existence of common and distributed adaptation sites.
Finally, we conclude by comparing our data with similar data collected in
non-human and human primates by our lab and others.
576
577
ADAPTATION CHARACTERISTICS OF THE DIFFERENT SACCADE TYPES
578
The intrasaccadic step adaptation paradigm, which has been used to
reduce the amplitude of simian targeting saccades, also produced a significant
adaptation of memory-guided (MG), delayed, scanning and express saccades
(Fig. 2). However, the adaptation characteristics of the different saccade types
often were not the same. In particular, MG saccades showed the fastest and
greatest amount of adaptation and the greatest recovery jump (Fig. 3). These
results suggest that MG saccades are able to undergo the greatest adaptation and
to switch rapidly back to the un-adapted state when the adaptation paradigm is
discontinued.
It has been suggested that saccade adaptation consists of both a rapid
and a slow process (Ethier et al. 2008; Smith et al. 2006; for review Tian et al.
2009). Smith at al (2006) showed that it is the jump in gain after error-clamps
or a break that is a marker of the fast process. Unfortunately, our experimental
protocol does not allow a direct evaluation of these two processes. If, however,
the recovery jumps in gain (Fig. 4) reflect the rapid components and the
subsequent gradual gain changes represent the slow component, MG saccades
might have adaptation mechanisms with especially robust rapid and slow
components because they exhibited a greater recovery jump and faster gradual
gain change.
The differences in adaptation characteristics suggest that the neuronal
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
21
598
599
600
601
602
mechanisms that produce adaptation of the various saccade types could be
different. Moreover, the mechanisms might, at least in part, lie in different
pathways within the saccade circuitry. Whether an adaptation pathway might
be specific to a single saccade type or be shared by several types can be inferred
from the transfer of adaptation among different saccade types.
603
604
ADAPTATION TRANSFER BETWEEN DIFFERENT SACCADE TYPES
605
Adaptation transfer depended strongly on which saccade type was
adapted. In agreement with a previous study on monkeys (Fuchs et al. 1996),
our study shows that the adaptation of targeting saccades transfers to express,
scanning, delayed and MG saccades (Fig. 6A, column “T”, the area above the
diagonal in each square). In addition, the current study showed that adapted
delayed saccades also transferred to all other types of saccades. Adaptation of
express saccades transferred only to targeting saccades, whereas scanning
saccade adaptation transferred to both targeting and express saccades.
Adaptation of MG saccades transferred only to delayed saccades.
606
607
608
609
610
611
612
613
614
615
Figure 6 near here
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
Our transfer data suggest that the adaptation sites of some types of
saccade are shared with other types, whereas some sites are not. Fig. 6B shows
how these sites might be configured based on our data. Each saccade type can
be adapted and therefore each has an adaptation site (a color coded spot)
through which its saccade signal flows (color coded arrow). If adaptation of
one kind of saccade transfers to another, the tested saccade pathway traverses
the adaptation site of the adapted saccade. For example, the adaptation of MG
saccades transferred to delayed saccades so the delayed saccade pathway
(brown arrow) traverses the MG adaptation site (red filled spot). Similar logic
can be applied for the remaining transfers of saccade adaptation. Because
adaptation of both targeting and delayed saccades transferred to all other
saccade types (Fig. 6A), the saccade pathway for the other types must traverse
the adaptation site for both targeting and delayed saccades, suggesting that they
have a common adaptation site, i.e., the brown/green spot. Indeed, when either
22
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
targeting (Frens and Van Opstal 1997; cf. Takeichi et al. 2007) or delayed
(Quessy et al. 2010) saccades undergo adaptation, the neuronal activity in the
superior colliculus does not change, indicating that both targeting and delayed
saccades use an adaptation site that is downstream of the superior colliculus.
The short latency express saccade is often considered to be a reactive
saccade like a targeting saccade (for reviews Hopp and Fuchs 2004; Pelisson et
al. 2010). Therefore, it might be expected to be one of the last types to join the
common final saccadic pathway. However, because we show here that
scanning saccades transfer to both targeting and express saccades, targeting and
express saccade commands must traverse the site for scanning saccade
adaptation (Fig. 6B). Similarly, delayed saccades are considered to be
voluntary saccades that require cortical processing (for reviews Hopp and
Fuchs 2004; Pelisson et al. 2010). However, because delayed saccades
transferred to all other saccade types, the site for delayed saccade adaptation
must lie on the saccade command pathway for all other saccade types. Just
because delayed saccades apparently involve cortical processing, their
adaptation site does not have to lie in the cortex. Based on our transfer data, it
seems plausible that the adaptation sites for both the reactive targeting saccade
and the voluntary delayed saccade could lie downstream of the adaptation site
for the reactive express saccade.
From this simple schematic, we suggest that the adaptation sites of
targeting and delayed saccades are shared and likely are located the farthest
downstream along the saccade command pathway. In contrast, the adaptation
sites for express, scanning, and MG saccades are likely to be separate and
located more upstream. Please note that the adaptation sites depicted in the
signal flow schematic of Fig. 6B do not necessarily refer to specific structures in
the brain; they could represent different groups of cells in the same neural
structure or different synapses on the same cell. Moreover, adaptation transfer
between any of the saccade types was never 100%, suggesting that, in addition
to the site indicated in Fig. 6B, each type also may have a site specific to its own
adaptation.
662
663
COULD EXPERIMENTAL CONDITIONS ACCOUNT FOR ANY OF THE
23
664
DIFFERENCES IN TRANSFER BETWEEN SACCADE TYPES?
665
We will consider three possible factors here. Although there may be
others, these have been explicitly suggested in the review process.
In humans, the amount of forward but not backward adaptation transfer
between saccade types depends on the preview duration of the saccade target
(Zimmermann 2013). In our study of backward adaptation, the target preview
duration for scanning saccades was the inter-saccadic interval (~1 sec), for
delayed saccades was the delay + the saccade latency (300+~150 = ~450 ms),
for MG saccades was 300 ms and for targeting and express saccades was the
saccade latency (~150 and ~80-120 ms, respectively). Although the preview
duration for targeting saccades was twice that for express saccades, they
transferred to each other. Also, although the preview duration for targeting
saccades was less than half that for the delayed saccade, they still transferred to
each other. Thus, preview duration also is not a factor in the transfer of
backward adaptation in the monkey.
Does the amount of transfer depend on the velocities of the different
saccade types? Across both monkeys, the average pre-adaptation peak velocity
was fastest for scanning saccades (~370 °/s) and slowest for MG saccades
(~280 °/s). Targeting, express and delayed saccades had similar, intermediate
peak velocities (~340 °/s). If the transfer were dictated by the similarity of
saccade peak velocity, express saccade adaptation would have transferred to
delayed saccades, but it didn’t (Fig. 6A). Also, if transfer preferentially occurs
from faster to slower saccades (or vice versa), scanning saccade adaptation
would have transferred to MG saccades (or MG saccade adaptation would have
transferred to the scanning saccades); again, this was not the case.
The characteristics of the adaptation process itself also did not appear to
be a factor in the adaptation transfer between saccade types. For example,
transfer did not always occur between saccade types with similar adaptation
speeds, e.g., express to scanning, Fig. 3A, E. Nor did transfer always occur
from faster to slower adapting saccades, e.g., MG to express, or from slower to
faster adapting saccades, e.g., express to MG (Fig. 3).
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
COMPARISON WITH PREVIOUS STUDIES
24
697
NON-HUMAN PRIMATES. Our lab’s earlier study on the transfer of T saccade
698
adaptation in the monkey also demonstrated substantial, but somewhat greater,
transfer to all the other saccade types tested here (Fuchs et al. 1996). One
possible explanation for the modest quantitative difference between the studies
is that Fuchs et al. (1996) used an adapt step that was a fixed percentage of the
target step, whereas this study employed an adapt step that was a fixed
percentage of the actual saccade (see Materials and Methods). A study in
humans suggested that the adaptation process indeed might be different when
elicited by different back step paradigms (Zimmermann and Lappe 2010). A
direct comparison of the results of these adaptation paradigms in the monkey
might shed some light on the adaptation process, itself.
699
700
701
702
703
704
705
706
707
708
709
HUMAN PRIMATES. In Fig. 6A, we compare the transfer between pairs of
710
saccade types in the monkey (the area above the diagonal in each square) with
the transfer between the same two pairs in humans (the area below the diagonal
in each square). As Fig. 6A shows, there are mostly similarities but some
differences.
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
Targeting saccade adaptation. In both monkeys and humans, adapted targeting
saccades transferred to express saccades (Fig. 6A, column T, upper area in each
square). Adapted targeting saccades also transferred to scanning saccades in the
monkey, but not consistently in humans. A possible extenuating factor might
have been the different numbers of fixed target spots used to elicit scanning
saccades in the two species. When either 7 (Alahyane et al. 2007), 6 (Deubel
1995; 1999), or 4 (Zimmermann and Lappe 2009) spots were used with humans,
there was no transfer; however, when 2 spots were used with either humans
(Cotti et al. 2007) or monkeys (Fuchs et al. 1996) or 3 were used with monkeys
(this study), transfer occurred. However, one study on humans (Panouillères et
al. 2014) produced no transfer when three spots were used (only supplementary
data).
In the monkey, adaptation of targeting saccades transferred to both MG
and delayed saccades, whereas transfer was inconsistent in the human. For
human MG saccades, transfer depended on the delay before the saccade “go”
25
730
731
732
733
734
735
736
737
signal. If MG saccades were executed after delays >1 sec (Deubel 1995; 1999),
no transfer occurred; however, after shorter delays, transfer did occur to MG
saccades (Fujita et al. 2002; Hopp and Fuchs 2010). Our delay (0.3 sec) was
similar to that of these last two human studies, so the transfer to MG saccades
was consistent between the two primate species when similar conditions were
used. For human delayed saccades, targeting saccade adaptation produced no
transfer whether the delay was long (>1 sec, Deubel 1995; 1999) or short (0.3
ms, Collins and Doré-Mazars 2006).
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
Express saccade adaptation. Express saccade adaptation transferred to
targeting saccades both in monkeys (Fig. 6A, column E) and humans (Deubel
1995; 1999; Hopp and Fuchs 2010; 2002). To the best of our knowledge,
transfers of express saccade adaptation to scanning or delayed saccades have
not been investigated in humans. Express saccade adaptation did not transfer to
MG saccades in this study, but there was significant transfer in humans (Hopp
and Fuchs 2010).
There are several differences between this monkey study and the human
studies. In our study, we used shorter gap durations to elicit express saccades.
Second, the distribution of latencies in our study was bimodal, but unimodal in
the human study (Hopp and Fuchs 2010). Therefore, in our study, fewer total
saccades were in the latency range of express saccades so the average gain
change was smaller. However, when we selected the subset of our experiments
with the largest gain changes (16 ± 3.6%, n = 3; compared with 14 ± 0.7% in
Hopp and Fuchs 2010), there still was no significant transfer to MG saccades;
therefore, the lower average gain change in monkeys did not explain the lack of
transfer.
756
757
758
759
760
761
762
Scanning saccade adaptation. Scanning saccade adaptation transferred to
targeting saccades in both monkeys (this study) and humans (Alahyane et al.
2007; Cotti et al. 2007; Deubel 1995; 1999; Erkelens and Hulleman 1993;
Panouillères et al. 2014; Zimmermann and Lappe 2009) (Fig. 6A, column S).
Scanning saccade adaptation also transferred to monkey express saccades, but
transfer to human express saccades has not been tested. Scanning saccade
26
763
764
765
766
767
768
adaptation transferred to neither delayed nor to MG saccades in monkeys and
showed only a weak transfer to those saccade types in humans (Deubel 1995;
1999). In order to provide a data set with gain changes more comparable with
those of Deubel (1999), we selected the subset of our experiments with the
largest gain changes (12.3 ± 1.3%, n=4; compared with 20.5% in Deubel 1999);
however, there still was no transfer (0 ± 0%) from scanning to MG saccades.
769
770
771
772
773
774
775
Delayed saccade adaptation. Delayed saccade adaptation transferred to all
other types in the monkey (Fig. 6A, column D). In humans, no data are
available for the transfer of delayed saccades to express or scanning saccades.
Adapted delayed saccades transferred to targeting and MG saccades both in
monkeys (this study) and humans (targeting: Collins and Doré-Mazars 2006;
MG: Fujita et al. 2002; Hopp and Fuchs 2010).
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
MG saccade adaptation. Adaptation of MG simian saccades transferred only to
delayed saccades (Fig. 6A, column MG). In humans, the transfer of MG
saccade adaptation to other types was rather modest, if present at all. There was
either weak transfer to targeting, express and delayed saccades (Fujita et al.
2002; Hopp and Fuchs 2010) or none at all (Deubel 1995; 1999). There was no
transfer to S saccades in either monkeys (this study) or humans (Deubel 1995;
1999).
In summary, there is substantial agreement regarding the transfer of
adapted saccades to test saccades across non-human and human primates (Fig.
6C). (1) In both species, adapted targeting saccades transferred to express
saccades, to MG saccades elicited with short delays and to scanning saccades
elicited with only a few target spots and to delayed saccades in monkeys but not
humans. (2) Adapted express saccades transferred to targeting saccades in both
species and to MG saccades in humans but not monkeys; other saccade types
were not tested in humans. (3) Adapted scanning saccades transferred to
targeting saccades in both species and showed little if any transfer to delayed or
MG saccades in either species. In monkeys, there also was a transfer to express
saccades, but this transfer has not been tested in humans. (4) Adapted delayed
saccades transferred to all saccade types in the monkey; transfer to express and
27
796
797
798
scanning saccades has not been tested in humans. (5) Adapted MG saccades
showed little if any transfer to targeting, express, or scanning saccades, but
transferred to delayed saccades in both primate species.
799
800
SPECULATION ON THE POSSIBLE NEURONAL SITES OF ADAPTATION
801
Because the generation of different saccade types is thought to involve
different brain regions (for reviews Gaymard et al. 1998; Hopp and Fuchs 2004;
Lynch and Tian 2006; Pelisson et al. 2010; Pierrot-Deseilligny et al. 2004), it
seems logical to expect that the sites where they undergo adaptation may also be
distributed.
Considerable progress has been made in understanding the adaptation
mechanisms for the extensively documented simian targeting saccade. Studies
using a variety of techniques have placed the site of adaptation of targeting
saccades downstream of the superior colliculus (SC). First, SC saccade-related
activity does not change during adaptation of targeting saccades (Frens and Van
Opstal 1997; Takeichi et al. 2007). Second, there is significant adaptation
transfer from targeting saccades to saccades evoked by low current electrical
stimulation of the SC (Edelman and Goldberg 2002; Melis and van Gisbergen
1996). Finally, stimulation of the SC to act as a surrogate error signal induces
saccade adaptation artificially (Kaku et al. 2009; Soetedjo et al. 2009).
Unit recording together with inactivation, lesion and anatomical
evidence have identified the oculomotor vermis (OMV) as an important
component of the downstream adaptation pathway. The gradual change in
amplitude that occurs during saccade adaptation is associated with gradual
changes in the discharge patterns of neurons of the OMV (Catz et al. 2008;
Kojima et al. 2010b), its output, the caudal fastigial nucleus (Inaba et al. 2003;
Scudder and McGee 2003) and the brain-stem burst generator to which cFN
neurons project (Kojima et al. 2008). Moreover, the error signal for targeting
saccade adaptation reaches the OMV over climbing fiber inputs to OMV
Purkinje (P)-cells that originate in the SC (Catz et al. 2005; Kaku et al. 2009;
Soetedjo and Fuchs 2006; Soetedjo et al. 2009; Soetedjo et al. 2008). However,
even targeting saccade adaptation apparently has a cortical component because
it can be influenced or driven by cognitive processes, such as reinforcement
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
28
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
(Madelain et al. 2011), attention (Khan et al. 2014) or the nature of a perceptual
task (Schutz et al. 2014).
There are a few scattered studies concerning possible sites of adaptation
for delayed, MG, and scanning saccades. When delayed saccades are adapted,
SC saccade-related activity also does not change (Quessy et al. 2010),
indicating that the adaptation site for delayed saccades also is downstream of
the SC. Because OMV P-cells discharge for all the saccade types examined in
this current study (Kojima et al. 2010a), it is possible that the OMV could
constitute at least part of the final common adaptation site of targeting and
delayed saccades hypothesized in Fig. 6B. Neurons in the lateral intraparietal
(LIP) of the simian cortex show no change in activity during adaptation of
either MG or targeting saccades in the monkey (Steenrod et al. 2013).
Therefore, the site of MG saccade adaptation is likely downstream of the LIP.
Finally, human studies have suggested that the adaptation of scanning saccades
involves some part of the parietal cortex (Cotti et al. 2007; Cotti et al. 2009;
Gerardin et al. 2012; Panouillères et al. 2014) and the lateral cerebellum
(Alahyane et al. 2008; Panouillères et al. 2012; 2013).
846
847
CONCLUSION
848
Prior to this paper, the detailed characteristics of adaptation in the
monkey had been extensively documented only for targeting saccades. With
the help of that behavioral data, significant progress has been made during the
last 20 years in understanding the neuronal basis of targeting saccade adaptation.
In this paper we provide comprehensive adaptation data for four other types of
simian saccades. Equipped with this foundation, we can expect similar
progress to be made in understanding the neuronal basis of adaptation of
Express, Scanning, Delayed and Memory-guided saccades. A logical place to
begin such studies would be the oculomotor vermis, whose P-cells discharge for
all 5 saccade types.
Our second contribution documents for the first time the transfer of
adaptation between 5 types of saccade in the monkey. These data allow us to
create a simple hierarchical model of how adaptation sites might be roughly
configured for the different saccade types. For some of the reasons mentioned
849
850
851
852
853
854
855
856
857
858
859
860
861
29
862
863
864
865
866
867
earlier in the Discussion, it currently is premature to equate the adaptation sites
in Fig. 6B with actual neuronal structures, with the possible exception of the
most downstream site that probably involves the cerebellum. The rough
correspondence of the adaptation transfer data in monkeys and man provide
some optimism that the neuronal mechanisms revealed in the monkey may have
relevance when extrapolated to humans.
868
30
869
Figure Legends
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
Figure 1
Schematic illustration of the experimental protocol. A, The task used to elicit
the 5 types of saccade examined in this study. Thick grey lines indicate both
target position and LED ON/OFF state, thin black lines indicate eye position. B,
Distributions of saccade reaction times from representative express saccade
experiments for monkeys B and W. The bimodal distributions were elicited by
gaps of 100 and 160 ms, respectively. Reaction times between 50-100 and
50-110 ms in moneys B and W, respectively, are express saccades (black bars).
C, Time course of an experiment. In this example, one saccade type (grey) is
adapted and 3 other types (red, blue, and green) are tested. We collect ~30
saccades of all 4 types during the pre-adaptation session (“Pre”). We then
reduce the amplitude of the adapting saccade (“Adapt”). During adaptation, we
collect ~20 trials each of the 3 test saccades three times without backward-steps.
Finally, we discontinue the backward-step and monitor the recovery of the
adapted saccade gain (“Recovery”).
886
887
888
889
890
891
Figure 2
Time course of adaptation measured as normalized saccade gain (black dots)
and recovery (grey dots) of each of the five types of saccade for both monkeys.
Each dot indicates one saccade trial. Grey line on the black dots indicates the
exponential fit of saccade gain during adaptation.
892
893
894
895
896
897
898
899
900
901
Figure 3
Comparison of adaptation characteristics of the five different types of saccade.
A, Average amount of gain decrease at saccade number 250. B, Initial gain
decrease following the start of the backward-step task. C, Initial gain increase
at the start of recovery. D, Distribution of the width of 95% confidence interval
of rate constants of each exponential fit. E, Rate constant of the exponential fit.
F, Absolute gain change of the exponential fit. For this figure and the rest, T, E,
S, D and MG labels indicate targeting, express, scanning, delay and
memory-guided saccades, respectively. Error bars are 1SD. * indicates
31
902
903
unpaired t test p < 0.05 with Bonferroni correction. White bars show data from
each monkey. Black bars are averages of the data from both monkeys.
904
905
906
907
908
909
910
911
912
Figure 4
A representative transfer matrix from adapted saccade to test saccades. T, E, S,
D, and MG see Fig. 3. Grey dots and lines from the adapted saccade, black dots
and lines from the tested saccade. Label B### indicates the experiment number.
Vertical bars indicate mean saccade gain ± 1SD of the pre-adaptation data and
one of the three tests during adaptation. Bar graphs compare the gain change of
adapted (grey) and test (black) saccades (see Methods for details). * indicates
unpaired t test with p < 0.05. ns indicates unpaired t test with p > 0.05.
913
914
915
916
917
918
919
920
Figure 5
Summary of adaptation transfer among T, E, S, D and MG saccades for
monkeys B and W. White bars, percent transfer for each monkey using eq. 2.
Black bars, average percent transfers for both monkeys. Grey bars, average
percent transfers for both monkeys using the modified eq. 3; striped bars are
generated from modified eq. 4 (see Materials and Methods). Error bars are 1SD.
* indicates unpaired t test with p < 0.05.
921
922
923
924
925
926
927
928
929
930
931
932
933
Figure 6
Summary and interpretation of adaptation transfer. A, Summary of transfer
experiments. Area above the diagonal in each square indicates the monkey
result obtained in this study and (Fuchs et al. 1996). Area below indicates the
human result from Alahyane et al. (2007) (“A”), Cotti et al. (2007) (“C”),
Collins and Doré-Mazars (2006) (“O”), Deubel (1995; 1999) (“D”), Erkelens
and Hulleman (1993) (“E”), Fujita et al. (2002) (“F”), Hopp and Fuchs (2010;
2002) (“H”), Panouillères et al. (2014) (“P”), and Zimmermann and Lappe
(2009),(“Z”). B, Circuit implications based on adaptation transfer data. Arrows
indicate the route for a hypothetical saccade signal that drives each saccade type.
Circles indicate the adaptation site for each saccade type. C, Distilled version
of A (see text).
934
935
32
936
937
938
939
940
941
942
Acknowledgements
This study was supported by National Institute of Health (NIH) grants
EY019258 (RS), EY023277 (YK), EY00745(AFF) and RR00166 from the
National Center for Research Resources (NCRR), a component of the NIH. The
contents of this paper are solely the responsibility of the authors and do not
necessarily represent the official views of NCRR or NIH.
943
944
References
945
Alahyane N, Fonteille V, Urquizar C, Salemme R, Nighoghossian N, Pelisson D, and Tilikete C.
946
Separate neural substrates in the human cerebellum for sensory-motor adaptation of reactive
947
and of scanning voluntary saccades. Cerebellum (London, England) 7: 595-601, 2008.
948
Alahyane N, Salemme R, Urquizar C, Cotti J, Guillaume A, Vercher JL, and Pelisson D.
949
Oculomotor plasticity: are mechanisms of adaptation for reactive and voluntary saccades
950
separate? Brain research 1135: 107-121, 2007.
951
Barash S, Melikyan A, Sivakov A, Zhang M, Glickstein M, and Thier P. Saccadic dysmetria and
952
adaptation after lesions of the cerebellar cortex. The Journal of neuroscience : the official journal
953
of the Society for Neuroscience 19: 10931-10939, 1999.
954
Catz N, Dicke PW, and Thier P. Cerebellar-dependent motor learning is based on pruning a
955
Purkinje cell population response. Proceedings of the National Academy of Sciences of the
956
United States of America 105: 7309-7314, 2008.
957
Catz N, Dicke PW, and Thier P. Cerebellar complex spike firing is suitable to induce as well as to
958
stabilize motor learning. Current biology : CB 15: 2179-2189, 2005.
959
Collins T, and Dore-Mazars K. Eye movement signals influence perception: evidence from the
960
adaptation of reactive and volitional saccades. Vision research 46: 3659-3673, 2006.
961
Cotti J, Guillaume A, Alahyane N, Pelisson D, and Vercher JL. Adaptation of voluntary saccades,
962
but not of reactive saccades, transfers to hand pointing movements. Journal of neurophysiology
963
98: 602-612, 2007.
964
Cotti J, Panouilleres M, Munoz DP, Vercher JL, Pelisson D, and Guillaume A. Adaptation of
965
reactive and voluntary saccades: different patterns of adaptation revealed in the antisaccade
966
task. The Journal of physiology 587: 127-138, 2009.
967
Deubel H. Separate adaptive mechanisms for the control of reactive and volitional saccadic eye
968
movements. Vision research 35: 3529-3540, 1995.
33
969
Deubel H. Separate mechanisms for the adaptive control of reactive, volitional, and
970
memory-guided saccadic eye movements. . In: Gopher D, Koriat A (eds) Attention and
971
performance XVII cognitive regulation of performance: interaction of theory and application MIT
972
Press, Cambridge, pp 697–721 697–721, 1999.
973
Deubel H, Wolf W, and Hauske G. Adaptive gain control of saccadic eye movements. Human
974
neurobiology 5: 245-253, 1986.
975
Edelman JA, and Goldberg ME. Effect of short-term saccadic adaptation on saccades evoked by
976
electrical stimulation in the primate superior colliculus. Journal of neurophysiology 87:
977
1915-1923, 2002.
978
Erkelens CJ, and Hulleman J. Selective adaptation of internally triggered saccades made to
979
visual targets. Experimental brain research 93: 157-164, 1993.
980
Ethier V, Zee DS, and Shadmehr R. Spontaneous recovery of motor memory during saccade
981
adaptation. Journal of neurophysiology 99: 2577-2583, 2008.
982
Fischer B, and Boch R. Saccadic eye movements after extremely short reaction times in the
983
monkey. Brain research 260: 21-26, 1983.
984
Fischer B, and Ramsperger E. Human express saccades: extremely short reaction times of goal
985
directed eye movements. Experimental brain research 57: 191-195, 1984.
986
Fischer B, Weber H, Biscaldi M, Aiple F, Otto P, and Stuhr V. Separate populations of visually
987
guided saccades in humans: reaction times and amplitudes. Experimental brain research 92:
988
528-541, 1993.
989
Frens MA, and Van Opstal AJ. Monkey superior colliculus activity during short-term saccadic
990
adaptation. Brain research bulletin 43: 473-483, 1997.
991
Fuchs AF, Reiner D, and Pong M. Transfer of gain changes from targeting to other types of
992
saccade in the monkey: constraints on possible sites of saccadic gain adaptation. Journal of
993
neurophysiology 76: 2522-2535, 1996.
994
Fuchs AF, and Robinson DA. A method for measuring horizontal and vertical eye movement
995
chronically in the monkey. Journal of applied physiology 21: 1068-1070, 1966.
996
Fujita M, Amagai A, Minakawa F, and Aoki M. Selective and delay adaptation of human
997
saccades. Brain research Cognitive brain research 13: 41-52, 2002.
998
Gaymard B, Ploner CJ, Rivaud S, Vermersch AI, and Pierrot-Deseilligny C. Cortical control of
999
saccades. Experimental brain research 123: 159-163, 1998.
1000
Gerardin P, Miquee A, Urquizar C, and Pelisson D. Functional activation of the cerebral cortex
1001
related to sensorimotor adaptation of reactive and voluntary saccades. NeuroImage 61:
34
1002
1100-1112, 2012.
1003
Hopp JJ, and Fuchs AF. The characteristics and neuronal substrate of saccadic eye movement
1004
plasticity. Progress in neurobiology 72: 27-53, 2004.
1005
Hopp JJ, and Fuchs AF. Identifying sites of saccade amplitude plasticity in humans: transfer of
1006
adaptation between different types of saccade. Experimental brain research 202: 129-145, 2010.
1007
Hopp JJ, and Fuchs AF. Investigating the site of human saccadic adaptation with express and
1008
targeting saccades. Experimental brain research 144: 538-548, 2002.
1009
Inaba N, Iwamoto Y, and Yoshida K. Changes in cerebellar fastigial burst activity related to
1010
saccadic gain adaptation in the monkey. Neuroscience research 46: 359-368, 2003.
1011
Judge SJ, Richmond BJ, and Chu FC. Implantation of magnetic search coils for measurement of
1012
eye position: an improved method. Vision research 20: 535-538, 1980.
1013
Kaku Y, Yoshida K, and Iwamoto Y. Learning signals from the superior colliculus for adaptation
1014
of saccadic eye movements in the monkey. The Journal of neuroscience : the official journal of the
1015
Society for Neuroscience 29: 5266-5275, 2009.
1016
Khan A, McFadden SA, Harwood M, and Wallman J. Salient distractors can induce saccade
1017
adaptation. Journal of ophthalmology 2014: 585792, 2014.
1018
Kojima Y, Iwamoto Y, Robinson FR, Noto CT, and Yoshida K. Premotor inhibitory neurons carry
1019
signals related to saccade adaptation in the monkey. Journal of neurophysiology 99: 220-230,
1020
2008.
1021
Kojima Y, Soetedjo R, and Fuchs AF. Behavior of the oculomotor vermis for five different types of
1022
saccade. Journal of neurophysiology 104: 3667-3676, 2010a.
1023
Kojima Y, Soetedjo R, and Fuchs AF. Changes in simple spike activity of some Purkinje cells in
1024
the oculomotor vermis during saccade adaptation are appropriate to participate in motor
1025
learning. The Journal of neuroscience : the official journal of the Society for Neuroscience 30:
1026
3715-3727, 2010b.
1027
Kojima Y, Soetedjo R, and Fuchs AF. Effect of inactivation and disinhibition of the oculomotor
1028
vermis on saccade adaptation. Brain research 1401: 30-39, 2011.
1029
Lynch JC, and Tian JR. Cortico-cortical networks and cortico-subcortical loops for the higher
1030
control of eye movements. Progress in brain research 151: 461-501, 2006.
1031
Madelain L, Paeye C, and Wallman J. Modification of saccadic gain by reinforcement. Journal of
1032
neurophysiology 106: 219-232, 2011.
1033
McLaughlin S. Parametric adjustment in saccadic eye movements. . Perception and
1034
Psychophysics 2 359-362, 1967.
35
1035
Melis BJ, and van Gisbergen JA. Short-term adaptation of electrically induced saccades in
1036
monkey superior colliculus. Journal of neurophysiology 76: 1744-1758, 1996.
1037
Miller JM, Anstis T, and Templeton WB. Saccadic plasticity: parametric adaptive control by
1038
retinal feedback. Journal of experimental psychology Human perception and performance 7:
1039
356-366, 1981.
1040
Noto CT, Watanabe S, and Fuchs AF. Characteristics of simian adaptation fields produced by
1041
behavioral changes in saccade size and direction. Journal of neurophysiology 81: 2798-2813,
1042
1999.
1043
Panouilleres M, Alahyane N, Urquizar C, Salemme R, Nighoghossian N, Gaymard B, Tilikete C,
1044
and Pelisson D. Effects of structural and functional cerebellar lesions on sensorimotor
1045
adaptation of saccades. Experimental brain research 231: 1-11, 2013.
1046
Panouilleres M, Habchi O, Gerardin P, Salemme R, Urquizar C, Farne A, and Pelisson D. A role
1047
for the parietal cortex in sensorimotor adaptation of saccades. Cerebral cortex (New York, NY :
1048
1991) 24: 304-314, 2014.
1049
Panouilleres M, Neggers SF, Gutteling TP, Salemme R, van der Stigchel S, van der Geest JN,
1050
Frens MA, and Pelisson D. Transcranial magnetic stimulation and motor plasticity in human
1051
lateral cerebellum: dual effect on saccadic adaptation. Human brain mapping 33: 1512-1525,
1052
2012.
1053
Pelisson D, Alahyane N, Panouilleres M, and Tilikete C. Sensorimotor adaptation of saccadic eye
1054
movements. Neuroscience and biobehavioral reviews 34: 1103-1120, 2010.
1055
Pierrot-Deseilligny C, Milea D, and Muri RM. Eye movement control by the cerebral cortex.
1056
Current opinion in neurology 17: 17-25, 2004.
1057
Quessy S, Quinet J, and Freedman EG. The locus of motor activity in the superior colliculus of
1058
the rhesus monkey is unaltered during saccadic adaptation. The Journal of neuroscience : the
1059
official journal of the Society for Neuroscience 30: 14235-14244, 2010.
1060
Robinson DA. A METHOD OF MEASURING EYE MOVEMENT USING A SCLERAL SEARCH
1061
COIL IN A MAGNETIC FIELD. IEEE transactions on bio-medical engineering 10: 137-145,
1062
1963.
1063
Rolfs M, Knapen T, and Cavanagh P. Global saccadic adaptation. Vision research 50: 1882-1890,
1064
2010.
1065
Schutz AC, Kerzel D, and Souto D. Saccadic adaptation induced by a perceptual task. Journal of
1066
vision 14: 4, 2014.
1067
Scudder CA, Batourina EY, and Tunder GS. Comparison of two methods of producing adaptation
36
1068
of saccade size and implications for the site of plasticity. Journal of neurophysiology 79: 704-715,
1069
1998.
1070
Scudder CA, and McGee DM. Adaptive modification of saccade size produces correlated changes
1071
in the discharges of fastigial nucleus neurons. Journal of neurophysiology 90: 1011-1026, 2003.
1072
Semmlow JL, Gauthier GM, and Vercher JL. Mechanisms of short-term saccadic adaptation.
1073
Journal of experimental psychology Human perception and performance 15: 249-258, 1989.
1074
Smith MA, Ghazizadeh A, and Shadmehr R. Interacting adaptive processes with different
1075
timescales underlie short-term motor learning. PLoS biology 4: e179, 2006.
1076
Soetedjo R, and Fuchs AF. Complex spike activity of purkinje cells in the oculomotor vermis
1077
during behavioral adaptation of monkey saccades. The Journal of neuroscience : the official
1078
journal of the Society for Neuroscience 26: 7741-7755, 2006.
1079
Soetedjo R, Fuchs AF, and Kojima Y. Subthreshold activation of the superior colliculus drives
1080
saccade motor learning. The Journal of neuroscience : the official journal of the Society for
1081
Neuroscience 29: 15213-15222, 2009.
1082
Soetedjo R, Kojima Y, and Fuchs AF. Complex spike activity in the oculomotor vermis of the
1083
cerebellum: a vectorial error signal for saccade motor learning? Journal of neurophysiology 100:
1084
1949-1966, 2008.
1085
Steenrod SC, Phillips MH, and Goldberg ME. The lateral intraparietal area codes the location of
1086
saccade targets and not the dimension of the saccades that will be made to acquire them. Journal
1087
of neurophysiology 109: 2596-2605, 2013.
1088
Straube A, Fuchs AF, Usher S, and Robinson FR. Characteristics of saccadic gain adaptation in
1089
rhesus macaques. Journal of neurophysiology 77: 874-895, 1997.
1090
Takagi M, Zee DS, and Tamargo RJ. Effects of lesions of the oculomotor vermis on eye
1091
movements in primate: saccades. Journal of neurophysiology 80: 1911-1931, 1998.
1092
Takeichi N, Kaneko CR, and Fuchs AF. Activity changes in monkey superior colliculus during
1093
saccade adaptation. Journal of neurophysiology 97: 4096-4107, 2007.
1094
Tian J, Ethier V, Shadmehr R, Fujita M, and Zee DS. Some perspectives on saccade adaptation.
1095
Annals of the New York Academy of Sciences 1164: 166-172, 2009.
1096
White JM, Sparks DL, and Stanford TR. Saccades to remembered target locations: an analysis of
1097
systematic and variable errors. Vision research 34: 79-92, 1994.
1098
Zimmermann E. The reference frames in saccade adaptation. Journal of neurophysiology 109:
1099
1815-1823, 2013.
1100
Zimmermann E, and Lappe M. Mislocalization of flashed and stationary visual stimuli after
37
1101
adaptation of reactive and scanning saccades. The Journal of neuroscience : the official journal of
1102
the Society for Neuroscience 29: 11055-11064, 2009.
1103
38
Targeting
B
Express
% of Saccade
A
Monkey-B
Monkey-W
20
20
10
10
0
Gap
0
0
100
200
0
100
200
Saccade Latency (ms)
Scanning
C
1st test
2nd test
First15
Delayed
Pre
Adapt
adapt saccade
3rd test
Last15
#1 #2 #3
for Transfer test
test saccade 1
test saccade 2
Recovery
test saccade 3
Note: Number of test saccade types varied between experiment
Memory guided
Target &
Eye Position
Time
2 columns (18cm)
Fig. 1
Monkey-B
Normalized Saccade Gain
A
Targeting
Pre
Monkey-W
Recovery
B
C
Express
D
Scanning
E Memory guided
Delayed
1.2
1.2
1.2
1.2
1.2
1
1
1
1
1
0.8
0.8
0.8
0.8
0.8
0.6
0.6
0.6
0.6
0
F
Normalized Saccade Gain
Adapt
500 1000 1500
Saccade Number
Pre
Adapt
2000
0
500
1000
1500
0
500
1000
1500
2000
0.6
0
500
1000
0
1.2
1.2
1.2
1.2
1
1
1
1
1
0.8
0.8
0.8
0.8
0.8
0.6
0.6
500
Saccade Number
Fig. 2
1000
Recovery
1.2
0
500
1000
0.6
0
500
0.6
0.6
0
500
1000
0
500
0
500
2 columns (18cm)
Initial Jump
B
Recovery Jump
C
0.2
0.2
0.1
0.1
0.1
0
0
0
0.2
0.2
0.2
0.1
0.1
0.1
0
0
0
0.3
0.3
0.3
0.2
0.2
0.2
0.1
0.1
0.1
0
T E S D MG
Saccade type
0
T E S D MG
Saccade type
0
∗
n. experiment
0.2
D
Average for
2 monkeys
Monkey-W
30
20
10
0
0
1,000
3,000
5,000 20,000
Width of 95% confidence interval of rate constant exponential fit
E
Rate constant
Gain Decrease
@250
Monkey-B
F
∗
T
E S D MG
Saccade type
Absolute gain change
A
Monkey-B
Monkey-W
∗
500
500
500
0
0
0
0.3
0.3
0.3
0.2
0.2
0.2
0.1
0.1
0.1
0
T E S D MG
Saccade type
0
T E S D MG
Saccade type
Fig. 3
0
Average for
2 monkeys
∗
∗
T
∗
E S D MG
Saccade type
2 columns (18cm)
Adapt saccade
T
E
B629 40
S
∗
B608
D
40
ns
B720 40
MG
∗
B626
40
1.0
1.0
T
20
0.8
0
∗∗
0.8
20
1.0
0
∗ ∗ 0.8
1.0
20
0
20
∗∗
0.8
0
∗
40
1.0
Gain change
Normalized Saccade Gain
B627
0.8
∗
0.6
0
1000
E T
0
2000
S T
B619
40
∗∗
1000
0.8
0.6
T E
40
∗
0.8
0
B629
500
0
D T
B720 40
20
1.0
0
∗ ∗ 0.8
∗∗
0.8
T S
0.6
0
S E
2000
ns
500
M T
B723
1.0
20
0
1000
B627
40
∗
1.0
0
20
∗∗
0.8
0
∗
0.8
0
∗∗
∗
∗
1000
T D
B627
0
20
0.8
0
∗∗
40
0
500
B726
M E
40
1.0
20
20
0.8
0.6
E S
1000
B629
40
B628
1.0
20
0.8
0
0
∗∗
0.8
0
∗
∗
0
500
B723
M S
40
∗
1.0
20
20
0 ∗
0.8
0
D S
500
40
0.8
0
∗∗
0.6
0.6
0
1000
E D
B629
40
1.0
∗
D E
1.0
0.6
0
500
0.6
1.0
20
0
B720
0.6
0
0.6
40
20
1.0
20
D
0
0.6
0
1.0
MG
0.6
ns
40
1.0
20
0
0.6
0.6
B522
S
Test saccade
E
0.6
0.6
0
40
1.0
20
0.8
0
2000
S D
B626
∗
40
1.0
20
0.8
0
B720
40
1.0
20
∗
0.8
0
S M
0.6
ns
0
500
M D
Saccade Number
∗∗
0.6
0
1000
Saccade Number
Fig. 4
T M
0.6
0
1000
Saccade Number
E M
0.6
0
2000
Saccade Number
0
500
Saccade Number
D M
2 columns (18cm)
Adapt saccade
T
E
S
D
MG
T
% transfer
100
50
∗
0
% transfer
∗
S
∗
∗
∗
50
∗
∗
0
100
50
∗
∗
0
100
MG
∗
50
0
100
D
Test saccade
E
100
W
B W Ba &
vg.
50
0
∗
B W B&W
a vg.
B W B&W
a vg.
∗
B W B&W
a vg.
B W B&W
a vg.
Fig. 5
1.5 column (13.5cm)
A
T ested Saccade
Human
Adapted Saccade
T
E
S
D
MG
D MG
A
CD
E
PZ
DH
T
E
S
D
FH
O
DH
H
AC
DPZ
D
DO
D
FH
H
D
Saccade
Signal
Fig. 6
FH
C
B
MG
D
T
E
S
Transferred
Weak transfer
No transfer
Not consistent
Not tested
DH
D
Monkey
Adaptation
Site
Human
T ested Saccade
Monkey
Adapted Saccade
T
E
S
D MG
T
E
S
D
MG
1 column (8.9cm)