Exp Brain Res (1992) 89:214-222
Experimental BrainResearch 9 Springer-Verlag1992
Dead zone for express saeeades Heike Weber 1, Franz Aiple 1, Burkhart Fischer l, and Alexander Latanov 2 1 Department of Neurophysiology, University of Freiburg, Hansastr. 9, W-7800 Freiburg, Federal Republic of Germany 2 Department of Supreme Neurological Activity, Lomonosow State University, Moscow, USSR Received September 23, 1991 / Accepted January 10, 1992
Summary. The saccadic eye movements of three humans and one n o n - h u m a n primate (a male rhesus monkey) have been measured for target eccentricities between 0.3 and 15 deg. With a gap task (fixation point offset precedes target onset by 200 ms) and a target at 4 deg, all subjects produced reasonable amounts of express saccades as indicated by a clear peak in the distribution of their saccadic reaction times (SRT): a b o u t 100 ms in h u m a n subjects and 70 ms in the monkey. This peak disappeared with decreasing target eccentricity below 2 deg, but saccades o f longer (regular) reaction times were still present. Thus it was found that there exists a dead zone for express saccades. In addition, small saccades have a much stronger tendency to overshoot the target and their velocity falls above the main sequence as defined by the least square fit of an exponential v = vo(1 - e x p ( - a/ao)) to the maximal velocity (v) versus amplitude (a) relationship (vo and ao are constants fitted). It is concluded that for small saccades the express way is blocked functionally or does not exist anatomically.
W y m a n and Steinman 1973a). It was found that these small saccades have long latencies up to 300 to 400 ms ( W y m a n and Steinman 1973b). In this study we raise again the question of a "dead zone" in relation to the reaction time of saccades, especially with respect to the short latency express saccades described in monkeys (Fischer and Boch 1983) and in m a n (Fischer and Ramsperger 1984). The express saccade uses a certain neural pathway, which m a y simply not exist for small saccades or which m a y be interrupted anatomically or blocked functionally at some point. On the other hand the conditions necessary for the execution of express saccades such as the disengagement of visual attention might not be available for small eccentricities (Mayfrank et al. 1986). Our results demonstrate quite clearly that h u m a n subjects and monkeys have a dead zone for express saccades, which is in the order o f one degree with individual variations depending on the subject. Within this dead zone even extensive practice and nonrandomized conditions do not lead to any appreciable a m o u n t of express saccades.
Key words: Express saccade - Reaction time - Attention - Fixation - Rhesus m o n k e y - H u m a n Methods Introduction In the early sixties Rashbass reported the absence of saccadic responses to targets appearing closer to the fovea than a b o u t 0.25 deg. He established the idea of a "dead zone" for visually guided saccadic eye movements (Rashbass 1961). Later on, however, this concept was questioned by Steinman and his group. They claimed that saccades as small as the miniature fixational saccades can be made, but only in response to visible targets (Timberlake et al. 1972; H a d d a d and Steinman 1973;
Offprint requests to: H. Weber
Three human subjects (all members of the department) and one rhesus monkey participated in this study. All were trained to make reasonable numbers of express saccades in the gap paradigm, when the fixation point is turned off 200 ms before the target appears at an eccentricity of 4 deg randomly to the right or left of the fixation point, as shown in Fig. 1 (left and upper part). We used only one monkey because his results were very clear and resembled closely those of our human subjects. Thus we felt that an additional animal experiment was not justified. The visual stimuli were generated by a Personal computer with a high resolution graphics interface (mirograph 510). They were presented on a Red-Green-Blue visual display unit (Nec Multisync II) with a frame rate of 80 Hz. The targets consisted of white squares 0.2~ in size and well above perceptual threshold. The fixation point was a small red spot (0.1~ in size) easily visible in the middle of the screen at a distance of 57 cm from the subjects' eyes. Gap and overlap trials were used as indicated in Fig. 1.
215 Gap t200
FP
I
Igap[
2000 40 I(
ST
FP 9
:
I
I I
I I
ST f
-
-
EM
120011
Lo )1
ST
Overlap
FP : f i x a t i o n point ST : stimulus EM : eye movement
I,
/
I I
I I
I I
,SRT,
Eye movements were measured using an infrared light reflexion method (Scalar Medical Iris eye movement measurement system) and stored on disc with a temporal resolution of 1 ms and a spatial resolution of 0.1 deg. Saccadic reaction time (SRT), amplitude and maximal velocity of one or two saccades following target onset were determined off line. For the monkey the stimulus display was the same as for the human subjects. The monkey, however, was first trained on a fixation task (Poggio and Fischer 1977): upon the appearance of the fixation point he had to pull a key. The key had to be released upon a dimming of the fixation point, which occurred after a randomly varying period between 2 and 6 s, in order to obtain a liquid reward. He was then trained in the gap paradigm (same time sequence as for the human subjects) to execute saccades. In this case the dimming occurred on the target and the monkey therefore learned to foveate the target by a saccade. During the measurements the monkey's head was fixed by a headholder. The eye movements were recorded by an infrared oculometer (Bach et al. 1983). The recording and analysis procedures were the same as for the human subjects.
Results
The absence of small express saccades Figure 2A shows the scatter plot of SRTs versus amplitudes for the three h u m a n subjects (HW, BF, MB). The data stem f r o m an experiment where the target position was randomized between 5 different locations ranging from 0.5 ~ to 6 ~. In all subjects the express saccades form an almost horizontal band at a b o u t 100 ms above the amplitude axis. This band becomes less prominent and finally disappears altogether at an amplitude of a b o u t 2 ~ in subject H W and BF and 0.5 ~ in subject MB. The regular saccades fall above the express band. At small eccentricities, where express saccades are already almost absent, saccadic reaction times increase in agreement with W y m a n and Steinman (1973). This is m o s t prominent in Subject HW. The series of histograms in part B of Fig. 2 present the same data in another way. They show the SRT-distribu-
h
FP
L
ST EM
Fig. 1. Gap paradigm (upper right diagram): the fixation point (FP) is extinguished 200 ms (gap) before the stimulus (ST) appeares. In the standard and training situation the ST appeares randomly 4 ~ at the left or right from the FP (left diagram). The saccadic reaction time (SRT) is measured from ST onset. In the overlap paradigm (lower right) the FP remains on until the end of the trial
tions of the saccades falling in the different ranges of amplitudes as indicated by the numbers at each histogram. One clearly sees the first peak at a b o u t 100 ms, representing the express saccades, disappear within the amplitude range 1-2 ~ in subject H W , below 1.5 ~ in subject BF and below 0.5 ~ in subject MB. N o n e of the saccades below these ranges is of the express type; instead, their reaction times fall in the range 120 to 200 ms (fast regular sacades) and in the range 200 to 300 ms (slow regular saccades). Part C of Fig. 2 shows the percentage of express saccades for each target position and part D for each range of amplitudes. These two graphs are not exactly the same for each of the subjects, because - as will be seen below - saccades made to targets at small eccentricities tend to overshoot. Figure 3 shows the result f r o m the same experiment in the monkey. Here the n u m b e r o f saccades for each eccentricity and the n u m b e r of target positions used was much larger. The disappearance of the horizontal express band (at a b o u t 70 ms) can be seen very clearly for amplitudes smaller than 2 ~ in part A. Since it is k n o w n that daily practice m a y increase the percentage o f express saccades, all subjects (humans and monkey) were trained in the gap task using a target position well inside the dead zone established by the first experiment (1 ~ for Subjects HW, BF and the monkey, 0.5 ~ for subject MB). Yet, even after 2000 trials m a d e on different days small express saccades were not found.
Anticipatory saccades Anticipatory saccades are saccades to expected target locations. Their parameters are p r o g r a m m e d before the actual appearance of the target. To elicit high numbers of anticipatory saccades we used again the gap paradigm, but the target was presented always on the right side f r o m the fixation point such that location and time of the stimulus' appearance could be predicted by fixation point
216
Absence of small express s a c c a d e s / S R T - h i s t o g r a m s Subject H W 400.
Subject BF
..
Subject MB
4OO
400
300-
300-
.sr 30O--"~#"
O~DO
, rr
,,'~,'.;. 9 -.
g 03
/
0
A
I "..
!
9 ""
..
;. ~ .
0
Amplitude [deg]
i
A
i
Amplitude [deg]
t
.....
o.o . . . .
03
0
A
Amplitude [deg]
0oi,5 ~ N=19 0,0O %
5~ N=23 5,0 %
0o.0
9, . t
1,0o.2,0 o
9,_ N=,S9
I
~ ~
115~-2#5~ N= 147
~
,2.2 ~
,.. 0,5~ N.~I
i :-
~*'~ "
. II. _ h_ ,
i
2,5o 3,5 o
i.
J
I
201
.111
2,0o-3,0o
I
N=129
4'0~176 N=114
1,00-2,0 =
I
i
t
| .
50 ~ 40 30
/
j
N=,e3
]
o;; oo
f
~
i
,
,
2,oo,oo
.=1,,
j
4,0 ~ -6,0 ~ N=73 97,3 %
10
0
B
40
80
120 160 200 240 280
SRT [ms]
0
B
40
80
120 160 200 240 280
SRT
[ms]
Fig. 2A-D. Gap paradigm, human data', three subjects: HW, BF, MB. The stimulus position is randomized between 5 different locations between 0.5 and 6~ at the right and 3* at the left side. Only data for right-directed saccades are shown. A Saccadic reaction time (SRT) versus amplitude of the first saccade (AMP) plots. B Distributions of the percentage of saccades (percent) versus the saccadic
offset. The three h u m a n subjects were tested for the 1~ and the 4 ~ target position. One result was that anticipations were less frequent towards the 1~ target position (about 30% of all saccades in subject BF and MB, only 4% in subject H W ) than to the 4 ~ position (about 70% in subject BF and MB, 45% in Subject HW). Furthermore, the amplitudes of the anticipatory saccades towards the 1~ target position overshooted the target by about 30% in all subjects, while longer latency saccades ( > 80 ms) hit the target correctly. Express saccades were again absent. Towards the 4 ~ target position, antici-
0
B
40
80
120 160 260 2~,0 2130
SRT [ms]
reaction time (SET) obtained within different ranges of saccadic amplitude. The numbers at each histogram give the amplitude range, the number of saccades included, and the percentage of express saccades. C Percentage of express saccades (ES) plotted against the target position. D Percentage of express saccades plotted against the saccadic amplitude
patory and longer latency saccades undershot the target by a b o u t 10% in all subjects.
Small saccades in the overlap task The absence of small express saccades in the gap task could be interpreted as the absence of the gap effect, i.e. the offset of the fixation point would no longer have a latency decreasing effect for small saccades as it does for larger saccades. To test this possibility we applied the
217 A b s e n c e of small express s a c c a d e s Subject BF Percent ES vs. target position
Subject HW 100 u) Q~ "0 8 0 r (3 o9 u}
u)
100
100 o3 8O
0
60
60 u~ o3 Q~ 4 0 CL X ,Q) 2O
4o X
Subject MB
20
ffl
60
20
o~ 0
0 1
0
2
3
4
Target position [deg]
C
0 0
C
1
2
3
1
4
Target position [deg]
co r "0
100~ u) Q~ "0 80 (.3
80
(3 o3 6 0 (,9 co O~ 4 0 X 0,)
1~176 t
40 o.. x
20
D
o~ ol '
0 0
1
2
3
4
Amplitude [deg]
5
0
6
O
_
201
o~ 0
4
!~176
60 (,9
20
3
l/ ,/'
Percent ES vs. saccade amplitude 100
2
Target position [deg]
C
1
2
3
4
Amplitude [deg]
overlap task for small saccades (within the dead zone for express saccades) and for larger saccades. The results are shown in Fig. 4 for the three human subjects. At the left small saccades are depicted, at the right larger (4 deg) saccades are shown. For each subject the upper two panels contain the SRT-distributions obtained in the gap task, the lower ones those obtained in the overlap task. First of all, one sees again the absence of the small express saccades (note: these are the results of a different experimental session than those shown in Fig. 2). Second, comparison of the upper and lower histograms from each subject clearly reveals that small saccades still show a gap effect: the mean latencies in the overlap task are longer than those in the gap task. This result is obtained for all three subjects. Applying the two-tailed T-test for comparison of the mean values of gap and overlap distributions obtained at small or larger eccentricities, respectively, resulted in highly significant differences (p < 0.01%) in all cases. This shows that the offset of the fixation point 200 ms before target onset fails to facilitate the express way for small saccades, fast regular saccades still being favoured. It is just when the fixation point remains visible that the number of fast regular saccades is reduced in favour of the slow regular saccades.
Effect of visual attention In the human subjects we also tested the possibility that the instruction to pay or not to pay attention to the fixation point could influence small saccades in a dif-
5
6
0
D
1
2
3
4
5
Amplitude [deg]
ferent way than larger saccades. However, this was not the case: the instruction to attend to the fixation point (overlap trials) decreased the number of fast regular small saccades and increased the number of slow regular small saccades such that the mean values were increased by a mean of about 30 to 40 ms (depending on the subject) very much the same way as it did for larger saccades. In order to separate the direction of gaze and the focus of directed visual attention we designed two experiments which were applied to the human subjects: In the first experiment the subjects were required to look straight ahead, while an attention target was presented at 4 ~ to the right (well outside the deadzone for express saccades). The saccade target appeared randomly 1~ left or right from the gaze direction. The time sequence of stimulus presentation for the attention target and the saccade target was the same as for the fixation point and the saccade target in the gap paradigm. Again no express saccades occurred to the 1~ target position. In the second experiment the subjects had to look again straight ahead, the attention target was presented at 3~ to the right, and the saccade target randomly at 4 ~ right or left from gaze direction. Thus the distance between attention target and saccade target was 1~ High numbers of express saccades were obtained in this condition. From these results we conclude that it is the size of the saccade itself that determines the deadzone for express saccades and not the distance between the saccade target and the focus of directed visual attention.
218
A b s e n c e of small e x p r e s s s a c c a d e s - m o n k e y / file" TA N = 2741
1 2 84
400
"
4 .
'.
300
9
.
,
'"
.
.
,.~.'.:" .:;.'~i"
" ' .
.
.
128-
100
4-
..... ::... ';:'. ',,'~':' : 4 ~ ' ~ O
1
A
.
'..
' ": .:': ,.
r ~. : :A .".," ". .. '.
2
... ,
3
4
,b _.1[. ~ 1 , .
Amplitude[deg]
.
.
.
8-
4-
.
1.5 ~ 9
492 ,
.
1.75 ~ -.
I[, ~IL .
739
~_
,
2'O~
.
,
,
4905
| | ,.
,
,-
128!
~ll.,~k
44
2.25~
440
167
30
12~ 8i 20-
4",_L
Q> 1 0
1
=
,
./,L
2 5 o / N = 362 , "~.~
12 0-
C
1
2
3
~
,
3.0 ~ / N= 381
~/
Target position [deg]
, - ' ,
,
,
"=?,--'r~
~ - - - , - - - ,
20 16
50(/1
12
40 "o
4.0 o / N =
1069
g_
30
0
B
20
x i1) 1 0
0
,.
16
o~
D
1264
/
8 4
"t~ 4 0 o~
o
1.00 / N =
I~ m~.
12
.
5 0 84
1/)
~[
..Ik
.
8
1299
8 4-
I-- 200 n" O3 ." ..'
0"5~
12-
.
g
9
A
100
200
300
400
SRT [ms] Fig. 3A-D. Data from monkey: the stimulus was presented at 9 different locations between 0.5 and 4~ at the right side. A-D same presentation as for the human data shown in Fig. 2
1
2
3
4
Amplitude [deg]
Whatever mechanism enables express saccades it is not active for small saccades. But, whatever mechanism enables fast and slow regular saccades it is active for small saccades as well as for larger saccades. It is however possible that very small saccades can be generated only as slow regular saccades.
Amplitude and velocity of small saccades To see whether small saccades differ from larger saccades in other respects than their latency distribution we looked at their spatial and dynamic parameters. Figure 5A shows scatter plots of (ai-at)/at (the relative error) for the three human subjects, where ai is the amplitude of the first saccade after target onset and at is the position o f the eye after a potential second saccade, which corrects an eventual error of the first saccade. The plots show that saccades below about 2 deg have a strong tendency to overshoot the target by higher relative
amounts than larger saccades, which have a stronger tendency to undershoot the targets. Furthermore we determined the main sequence, i.e. the maximal velocitiy (v) versus amplitude (a) relation, by fitting an exponential function, v = vo (1-exp- (a/ao)) to the data of each individual. We then plotted the mean values and standard deviations of (vi~v)/ai (the relative departure from the best exponential fit) as a function of ai, where vi and ai are the actually measured values of the maximal velocity and the amplitude of the saccade, respectively. The result is shown in part B of Fig. 5. Clearly, the values for small saccades fall above the main sequence. Very similar results were obtained for the monkey (Fig. 6). In the monkey many more saccades at small eccentricities have been measured (Fig. 6A, B, right panels, depict the express saccades only). Figure 6A (right panel) shows that the express saccades tend to overshoot the targets at small eccentricities. This is more evidence that it is difficult if not impossible to make small express
219
Absence
of small e x p r e s s s a c c a d e s / G A P
vs. O V L
Subject HW _+1 ~
+4 ~
I
301
N=450 141.5 +-17.7 file slhwsum
it
1
n~ 1 0
1 148 208.4 +-41.5 file solh1212
ot
N =
30-
I
E O r,,) a) 1 13.
N=174 121.0+-37.8 file hwhor4
GAP
1
rN=153 184.2+-47.5 file so4h1212 OVL
_
0
1O0
200
300
400
0
100
200
SRT [ms]
300
400
500
S R T [ms]
S u b j e c t BF +1 ~
~
+4 ~
. . .~ . . . .
sum ~
. . . . .
Tilesgb1f705
w GAP
20 t
"_
__
N~=203
I
1
N;12'503"29-6 file sobf1705
o
OVL
0
100
200
300
400
0
100
SRT [ms]
200
300
1 400
500
SRT [ms]
Subject MB +0.5 ~ 30-
+4 ~ N=277 181.8+-51.7 file rlO5mb
E 20-
N=165 103.8+-28.9 file rnbhor4
GAP O Q..
10,
oi-
r"
N= 200 264.9 +-82.2 file ovl05mb
30.
E o 20"t n 10.
r N=203 227.1 +-54.1 file sombrl4 OVL
0--
o
1O0
200
300
S R T [ms]
400
0
1O0
200
300
S R T [ms]
400
500
Fig. 4. Percent-distributions of the saccadic reaction times (SRT) for the three human subjects (HW, BF, MB). Data obtained with targets at small eccentricities (1 ~ 0.5 ~ are shown at the left, those to a larger eccentricity (4~) at the right side. Data obtained in the gap paradigm (GAP) are presented above those from the overlap paradigm (OVL) for each subject. The numbers at each histogram give the number of saccades included, the mean value of their reaction times, and the standard deviation
220 Small saccades A Relative error of amplitude Subject
Subject BF
HW
3-
Subject MB 3
3
2~
2 ....
~
~. . . .
.~
2. . . . 1"
............
No
0-
3 -2-3
-2 .
i
i
i 4 ai [degl
.
.
.
~
2
"0%•o
-1.
-3
o
........ o . . . . . . . . . . 9~, o~9 ? 9
-2.
.
-3
;
'
;
3 4 ai [deg]
0
i
2
3
i
6
ai [deg]
B Normalized velocity vs amplitude 25
15
24
-
20 1
0
.
;
1
6
-
~" 15 "3 > "N
5
"3
8
>
-~
5
G
0
0
I
2
3 ai
4
[degl
5
-5 .
0
,
1
2
3 ai
4
[deg]
5
6
-8
0
1
2
3
4
5
6
ai [deg]
Fig. 5A, B. Human data, three subjects (HW, BF, MB). Same experiment as in Fig. 2. A Relative error of amplitude ((ai-at)/at) of the first target directed saccade vs saccade amplitude (ai). Values falling below zero on the ordinate correspond to undershoots, those above zero to overshoots. Values on the zero line correspond to
saccades, which were not followed by a secondary (corrective) saccade. B Mean values and standard deviations of the normalized velocity ((vi-v)/ai) vs saccade amplitude (ai). The same amplitude ranges as in Fig. 2 were used for each subject
saccades. If a saccade to a target close to the fovea happens to be of the express type it has a tendency to overshoot and therefore to be of larger size. In contrast, with large eccentricities, express saccades usually have a stronger tendency than the regular saccades to undershoot the targets. In the monkey, small saccades show some deviation from the main sequence fit. This is similar to the results of the human subjects (see above), as can be seen in Fig. 6B.
neural structures and their connections involved in the generation of saccades: on the afferent side, striate cortex is involved in addition to the retina and the lateral geniculate nucleus (not shown). Centrally, the superior colliculus (in loop I), the frontal eye field (in loop II), and the parietal cortex (in loop Ill), are supposed to govern the generation of saccades. According to Schiller (Schiller et al. 1987), ablation of the superior colliculus abolishes both express saccades and fast regular saccades. Our results have shown that small fast regular saccades can still be made to eccentricities where express saccades are already absent. This indicates that the fast regular pathway via the superior colliculus can be used effectively by small saccades. Furthermore, small fast regular saccades are favoured by the gap condition as compared to the overlap condition. It is thus very unlikely that the signal of disengagement of visual attention, originating presumably from the parietal cortex, is not available for small saccades. Therefore one is left with the conclusion that the afferent projection to the superior colliculus does not exist or that the intrinsic tectal connection from the superficial layers to the deep layers (Moschovakis et al. 1988a); (Moschovakis et al. 1988b) is not effective or does not exist for small saccades. This interpretation would imply a lack of the connection between line (a) and (c) in loop I (thick line) of Fig. 7, whereas the pathways marked by (b) are taken to be effective for saecades of
C o n c l u s i o n and discussion
This study has shown that in man and monkey there exists a zone around the fovea in which a visual target cannot be reached by express saccades. Saccades of longer latencies (fast and slow regular saccades), however, are still able to reach targets within this zone. This suggests that a facilitating signal, which enables the express saccade, is not available or does not even exist for small saccades. Alternatively, the neural pathway mediating express saccades may simply not exist or may be interrupted for small saccades. To understand the (at least) three different modes in the latency distributions of saccades, a three loop model for the initiation of saccadic eye movements has been proposed (Fischer 1987). Figure 7 shows schematically some of the most important
221
Small saccades- monkey A Relative error of amplitude express and regular ':
..,
.
express only
9 .
9 .
.
3
.
*9 , e , , , t _ ~ ~ g * , , %,,o~ ' , ~, 9.
~'~
,
..
,,
, 9
~ ~-
,
, :,
1
r~
9
,
:%
"., ,~...~..:'.?:4.'................................... . .? r .~:"'.': i, :':
-..
"~o i
" "
:C !.::.,".,:" ,:~," :."
*%,
-2
-2
-3
-3 0
1
2
3
4
5
0
1
2
ai [ d e g ]
3
4
5
ai [ d e g ]
B Normalized velocity vs amplitude express and regular
express only
20
20
"-- 10 .-%
y"--- 10
._
>
"g
~
0
> i
J
-~.
0
v
-10
-10
0
1
2
3
4
5
ai [ d e g ]
1
2
3
4
5
ai [ d e g ]
Fig. 6A, B. Monkey data. A Same presentation as in Fig. 4A. The left panel gives the values of all (express and regular) saccades, the right panel only those of the express saccades. B Same presentation as in Fig. 4B. Again the left panel contains the values of express and regular saccades, the right only those of express saccades
all sizes. Whether the direct projection (a) - > (c) is never present or never in action for small saccades, or whether this projection is lost during the development of the fovea, remains an open question. The fact that children and dyslexics (age 9-11) make more express saccades than adults or normal readers (Fischer and Weber 1990) suggests that the express way may exist from birth as a way of reacting to visual stimuli in a reflex-like manner. Recently it has been shown in the monkey that inactivation of the fixation neurons in the superior colliculus leads to express saccades in response to a suddenly occurring stimulus eventhough the animal was supposed to maintain fixation at a central fixation point (Munoz and Wurtz 1991). This opens the possibility that parafoveal stimuli activate these cells leading to an increased fixation activity (Munoz et al. 1991) thus preventing express saccades. The fact that small saccades are relatively too fast in their velocity and that they tend to overshoot remaines to be explained. It is generally believed that a retinal error
signal arrives at the brain stem and is converted there into saccade size and velocity9 Thus it is likely that it is in the efferent part of the optomotor system, where these parameters may be mismatched for small saccades. On the other hand it cannot be excluded that the retinal error signal for small saccades is already wrong when leaving the visual cortex and/or passing the superior colliculus. In this context, it is also important to remember that small saccades in contrast to larger saccades need a visual target (Haddad and Steinman, 1973) and that voluntary saccades aimed at parafoveal locations where there is no visible target are too large. This may indicate that indeed small saccades are generated by a different mechanism than larger saccades. Examples for saccades to "nontarget" locations are remembered, anticipatory, and antisaccades. These three types of saccadic eye movements all need an intact frontal cortex (Deng et al. 1986; Pierrot-Deseitligny et al. 1991; Braun et al. 1992; Guitton et al. 1985). It may therefore be speculated that the generation of small saccades is not supported by the
222
Fig. 7. Schematic diagram of some neural structures involved in saccade generation, and their connections. V1 =striate cortex, Pre=prestriate cortex, Par=parietal cortex, FEF=frontal eye field, Nc = nucleus caudatus, Sn = substantia nigra, SC = superior colliculus, EM = efferent eye movement structures in the brain stem. For further explanation see text. Thick line: loop I, double line: loop II, triple line: loop III
f r o n t a l eye fields in the same way as t h a t o f larger saccades. T h e m a n central c o n t r i b u t i o n to small saccades could originate f r o m the parietal cortex, where a high p r o p o r t i o n of visual cells have large receptive fields sparing the fovea ( M o t t e r a n d M o u n t c a s t l e 1981).
Acknowledgements. We thank Ruth Lieberth for technical assistance in monkey training and data collection, and Dr. Monica Biscaldi for participating as subject. This work was supported by the Deutsche Forschungsgemeinschaft, SFB 325, Tp C5.
References Bach M, Bouis D, Fischer B (1983) An accurate and linear infrared oculometer. J Neurosci Methods 9:9-14 Braun D, Weber H, Mergner T, Schulte-M6nting J (1992) Saccadic reaction times in patients with frontal and parietal lesions. Brain (in press)
Deng SY, Goldberg ME, Segraves MA, Ungerleider LG, Mishkin M (1986) The effect of unilateral ablation of the frontal eye fields on saccadic performance in the monkey. In: Keller B, Zee D (eds) Adaptive processes in visual and oculomotor systems. Pergamon Press, Oxford, pp 201-208 Fischer B (1987) The preparation of visually guided saccades. Rev Physiol Biochem Pharmacol 106:1-35 Fischer B, Boch R (1983) Saccadic eye movements after extremely short reaction times in the monkey. Brain Res 260:21-26 Fischer B, Ramsperger E (1984) Human express saccades: extremely short reaction times of goal directed eye movements. Exp Brain Res 57:191-195 Fischer B, Weber H (1990) Saccadic reaction times of dyslexic and age-matched normal subjects. Perception 19 : 805-818 Guitton D, Buchtel HA, Douglas RM (1985) Frontal lobe lesions in man cause difficulties in suppressing reflexive glances and in generating goal-directed saccades. Exp Brain Res 58:455-472 Haddad GM, Steinman RM (1973) The smallest voluntary saceade: implications for fixation. Vision Res 13:1075-1086 Mayfrank L, Mobashery M, Kimmig H, Fischer B (1986) The role of fixation and visual attention in the occurrence of express saccades in man. Eur Arch Psychiat Neurol Sci 235:269-275 Mosehovakis A, Karabelas A, Highstein S (1988a) Structure function relationships in the primate superior colliculus. I. Morphological classification of efferent neurons. J Neurophysiol 60: 232-262 Moschovakis AK, Karabelas AB, Highstein SM (1988b) Structurefunction relationships in the primate superior colliculus. II. Morphological identity of presaccadic neurons. J Neurophysiol 60: 263-302 Motter BC, Mountcastle VB (1981) The functional properties of the light-sensitive neurons of the posterior parietal cortex studied in waking monkeys: foveal sparing and opponent vector organization. J Neurosci 1: 3-26 Munoz DP, Wurtz RH (1991) Disruption of visual fixation following injection of gabaergic drugs into the fixation zone of the primate superior colliculus. Soc Neurosci Abstr 17 : 544 Munoz DP, Pelisson D, Guitton D (1991) Movement of neural activity on the superior colliculus motor map. Science 25:1358-1360 Pierrot-Deseilligny C, Rivaud S, Gaymard B, Agid Y (1991) Cortical control of memory-guided saccades in man. Exp Brain Res 83 : 607-617 Poggio GF, Fischer B (1977) Binocular interaction and depth sensitivity in striate and prestriate cortex of behaving rhesus monkey. J Neurophysiol 40:1392-1405 Rashbass C (1961) The relationship between saccadic and smooth tracking eye movements. J Physiol 159 : 326-338 Schiller PH, Sandell JH, Maunsell JH (1987) The effect of frontal eye field and superior colliculus lesions on saccadic latencies in the rhesus monkey. J Neurophysiol 57:1033-1049 Timberlake GT, Wyman D, Skavenski AA, Steinman RM (1972) Letter to the editors: the oculomotor error signal in the fovea. Vision Res 12:105%1064 Wyman D, Steinman RM (1973a) Small step tracking; implications for the oculomotor "dead zone". Vision Res 13 : 2165-2172 Wyman D, Steinman RM (1973b) Latency characteristics of small saccades. Vision Res 13:2173-2175
