вход по аккаунту


Cortical control of double-step saccades Implications for spatial orientation.

код для вставкиСкачать
Cortical Control of Double-Step Saccades:
Implications for Spatial Orientation
W. Heide, MD, M. Blankenburg, E. Zimmermann, and D. Kompf, M D
To accurately localize a visual target in space despite eye movement-induced shifts of its retinal image, the brain
must take into account both its retinal location and information about current eye position or at least the preceding
eye displacement. We examined this ability with respect to saccadic eye movements by applying “double-step” stimuli,
where the locations of two sequentially flashed target lights have to be fixated by two successive saccades performed
after their disappearance. As the 2nd saccade will not start at the spatial location from which the 2nd target was seen,
a dissonance arises between its retinal coordinates and the motor coordinates of the required 2nd saccade. Nevertheless,
these saccades were performed quite accurately by 32 healthy human adults. To investigate the contribution of the
cerebral cortex, we recorded horizontal double-step saccades in 35 patients with focal unilateral hemispheric lesions.
Whereas frontal lesions impaired temporal properties, posterior parietal lesions caused spatial dysmetria or failure of
even ipsiversive 2nd saccades following contraversive 1st saccades. This reflects an inability to compensate for retinospatial dissonance by using nonretinal information (corollary discharge) about eye displacement associated with a
previous saccade into the contralesional hemifield. In conclusion, the parietal cortex is crucial for spatial constancy
across saccades.
Heide W, Blankenburg M, Zimmermann E, Kompf D. Cortical control of double-step saccades: implications
for spatial orientation. Ann Neurol 1995;38:739-748
When we move our eyes, the image of the visual world
shifts on our retina; but, nevertheless, we are able to
maintain a stable percept of visual space. To achieve
this spatial constancy and to accurately localize visual
targets in space, retinal information about a target’s
location is not sufficient, not even in connection with
proprioceptive information about eye position, as a
passive eye movement induced by a manual eye press
evokes the illusion that the world has moved, in contrast to an active eye movement associated with the
same inflow of afferent information (1, 21. Rather the
brain combines retinal signals with extraretinal information about active displacement of the eyes, probably
in terms of a neuronal signal (corollary discharge) that
monitors the motor command (efference copy) (31.
With respect to saccadic eye movements, evidence for
retinal and extraretinal signals has been found in the
superior colliculus [4],the frontal eye fields (FEFs) [ 5 ] ,
the supplementary motor area (SMA) [6], and the posterior parietal cortex (PPC) of monkeys 17-91. Particularly in the PPC, the visual receptive fields of neurons
are influenced by the direction of gaze (71 and prior
to an intended saccade [lo]. In humans, the role of
the PPC for the stability of visuo-spatial localization
across saccades has not yet been investigated.
This function can be tested with the “double-step”
saccadic paradigm 1111, which allows separation of a
From the Depurrnent of Neurology, Medical University at Lubeck,
Lubeck, Germany.
Received Jan 24, 1995. and
publication Jul 3, 1995.
I I ~revised
form May 22.Accepted for
target’s retinal vector from its saccadic motor vector;
i.e., two sequentially flashed visual targets have to be
fixated by two consecutive saccades. If both targets
have disappeared before the 1st saccade, the 2nd saccade will not start at the location from which the 2nd
target was seen, and thus a spatial dissonance is created
between the retinal coordinates of the 2nd target and
the motor coordinates of the necessary 2nd saccade
(“retino-spatial dissonance”). Nevertheless, normal
human subjects direct this saccade accurately to the
spatial location of the 2nd target and not to its retinal
location. This implies the use of extraretinal information about eye displacement associated with the 1st
saccade. To investigate the role of the human cerebral
cortex in this respect, we recorded double-step saccades in patients with focal unilateral hemispheric lesions, involving cortical areas that participate in the
control of saccades 1121, namely, the PPC, the FEF,
the SMA, or the dorsolateral prefrontal cortex (PFC)
C13f. Preliminary results have been published in abstract form { 1 4 ] .
Patients and Methods
Subjects (Table 1) were 19 patients (9males, 10 females; age,
36-75 years; mean, 57.7 years) with unilateral circumscript
posterior parietal and 16 patients (6 males, 10 females; age,
Address correspondence to Dr Heide, Department of Neurology,
Medical University, Rarzeburger Allee 160, D-23538 Lubeck, Germany’
Copyright 0 1995 by the American Neurological Association
Table 1 . Subgrwps of subjei-t~
Lesion Site
Right PPC
Left PPC
Right FEF
Right PFC
Left PFC
Left SMA
2. s
3i, 1s
12 Gi269:
- 99;
13 :Ti
4 r;
(D - 4 D )
5 Tf
”bSignificanr difference between patients and controls (’p < 0.05; ‘p < 0.001).
Time = mean age of lesions; Vol = volume; Neg = mean neglect score; Rey = Rey score; Er = percentage of trials with erroneous 2nd
saccades, following a contralateral 1st saccade; Dis = with retino-spatial dissonance, QDis = without retino-spatial dissonance. D - ,$D =
difference Dis minus +Dis; PPC = posrerior parietal cortex: FEF = frontal eye field; PFC
prcfrontal cortex; SMA = supplementary motor
area; i = infarction; s = postsurgtry defect.
24-72 years; mean, 50.1 years) with frontal lobe lesions.
Thirty-two age-matched healthy adults ( 13 males, 19 females;
age, 27-80 years; mean, 52.4 years) served as control. Lesions were due to either ischemic infarction ( n = 25) in the
territory of the medial or anterior cerebral artery or to tumor
surgery (n = lo), 0.5 to 89 months (mean, 9.6 months) prior
to examination. All patients were alert and fully oriented and
were able to follow the instructions for the testing. Informed
consent was obtained from each patient or volunteer. Clinically, 2 frontal and 3 parietal patients had manifest contralateral hemiparesis, 6 parietal patients had moderate hemisensory deficits, 5 had mild optic ataxia for arm movements to
contralateral visual targets, and 13 patients (9 right, 2 left
parietal, 2 right FEF cases) showed signs of contralateral visual hemineglect that were quantified in terms of a neglect
score as explained in the next section. Visual acuity was at
least 0.8, and 3 patients had homonymous visual field defects
in the periphery of their contralateral lower quadrants that
did not impair their ability to see the test stimuli for eye
movements in both visual hemifields. Ocdomotor examination revealed impaired smooth pursuit and optokinetic nystagmus in most parietal and some frontal patients, more with
ipsiversive stimulation, but fixation was stable and ocular motility unrestricted. Brainstem functions were intact. Patients
with other neurological, psychiatric, or ophthalmological diseases were excluded.
Neuropsychological TestiTig
To relate the patients’ saccadic performance to visuo-spatial
orientation in general, we performed some neuropsychological tests that are regarded as standard in the literature 1151.
The score for copying Rey’s complex figure was significantly
reduced ( p < 0.005) only in the right PPC group, compared
with the control group (see Table 1); it ranged below the
5% percentile ( = 27.5) of normal subjects in 6 of 14 right
PPC patients, compared with 0 or 1 case in each of the other
groups. The severity of contralateral visual hemineglect was
assessed by the following tests r16): line bisection, line and
shape cancellation, visual extinction with bilateral stimulation, copying, reading. and writing tasks. A score of “0” was
given when no sign of neglect was present and when the
740 Annals of Neurology
performance was within 2.5 standard deviations of the normal control group. The worst possible performance was
scored by “5” in bisection and cancellation tasks and by “3”
in the remaining tests. The neglect score was defined as the
sum of these single scores, with a possible maximum of 30.
I t ranged between 0 and 18 in right parietal patients, between
0 and 6 in left parietal patients, between 0 and 8 in right
FEF cases, and between 0 and 1 in the other groups.
Anutoniirul Sites of Lesions
From horizontal high-resolution computed tomographic (CT)
or magnetic resonance (MR) images (slice distance was 7 or
8 mm), the well-demarcated lesions were reconstructed and
superimposed on CT-parallel anatomical templates, as was
suggested by Damasio and Damasio 177. The atlas of Duvernoy [ 181 was used as an additional anatomical reference for
the identification of landmarks, gyri, and sulci. Volumes of
lesions were estimated by the formula, 7r/6 x length x
width X height 1191, thus approximating the lesion to an
ellipsoid. The resulting volumes corresponded well to those
obtained with a computerized volumetric method E20). In
the single patients, volumes ranged between 7 and 109 cm’
with means around 40 cm’ (see Table 1) in each group, except for smaller lesions in the SMA group.
The most frequently lesioned area (Jamaged in 7 1-100%
of the cases, see Fig 1) was selective for each subgroup of
patients. In the PPC it was located in the inferior parietal
lobule, along the border between the angular and supramarginal gyrus (Brodmann’s areas 39 and 40), extending cranially
toward the intraparietal sulcus, caudally to the temporoparietal junction, and posteriorly into the angular gyrus. In
part of the frontal cases, it involved the assumed location of
the right FEF (n = 4 ) in the premvtor region around the
middle portion of the precentral sulcus and the adjacent precentral gyrus [2 1,221, with some lesions extending also more
anteriorly into the prefronral cortex. Other lesions (five left,
three right) overlapped only in the dorso-lateral PFC (area
46).anterior to the FEF, or (n = 4 ) in the anterior portion
of the SMA (the assumed location of the supplementary eye
field) in the dorso-medial part of the superior frontal gyrus
Vol 38 No 5 November 1995
F i g I . Superimposed lesions of frontal and parietal patients. delineated on computed tomographic (CT)-parallel anatomical templates. at an angle of 36' to the inferior orbitomeatal line. according to Damasio and Damasio { I 7}. (A)Lesions of right or
left prefrontal cortex and posterior parietal cortex. (B) Right
frontal eye field and ldt supplementay motor area lesions. Black
areas are involiied in 86 t o loo%,, hatched areas in 71 t o SSgO
of the cases within each group.
Eye Mwement Recordings
Horizontal saccadic eye movements were recorded using infrared reflection oculography with photo diode arrays [24].
The recording system was mounted on a table where the
subject's head was positioned on a chin rest and fixed by a
head holder. The center of the pupil was coded as horizontal
eye position, with linearity being maintained up to 20" of
eccentricity. The analog signal was digitized, sampled at 200
H t , and stored on a hard disk. For visual stimulation, a red
laser spot of 0.5" was projected on a tangent white screen,
116 cm away from the subject's eyes. Target movements
were controlled by the PC, target positions were sampled,
and stored together with the eye position signals.
at 6"/6"(Figs 2 and 3A) or 6"/8". These four different combinations of double steps for each set of amplitudes and flash
durations were presented in pseudorandom order so that
subjects could not know the target location for a given trial
in advance. Following the 2nd target, the stimulus was extinguished and reappeared at the fixation point 1 second later
so that at least the 2nd saccade was performed while the
screen was blank, thus being never visually guided. Subjects
were instructed to fixate the rwo target locations in the order
of their appearance as exactly as possible. Recording time of
single trials was 20 seconds, containing four or eight double
steps. The analyses were predominantly based on trials with
flash durations of 140/100 msec (eight to 12 trials for each
pair of targets), because this stimulus was easier to be attended to than the faster one (100/80 msec) and yielded
more saccades with retino-spatial dissonance (see next section) than the slowest stimulus (1801100 msec); so only two
to six trials were performed with these latter stimuli. In addition, single-step trials of visually guided saccades were performed with 12 target steps of 10" but random frequency
and direction.
Saccadic Stimuli
In double-step trials, a central fixation point disappeared after
an unpredictable delay from stimulus onset, and the two targets were flashed successively for 140 msec (1st step) and
100 msec (2nd step). Alternative presentation times were
100/80 or 180/100 msec. Each pair of targets was located
either in the same hemifield at horizontal eccentricities of
10" and 5" (see Fig 3A) or 14"/7" or in different hemifields
Data Analysis
Data were analyzed off-line. After elimination of blinks or
other artifacts, a program searched for saccades. When eye
velocity changed by 40"/sec or more, a saccade was detected
and displayed on the monitor, together with the target position, so that it could be classified visually as primary saccade
in response to the 1st or 2nd target step or as corrective
saccade (secondary or tertiary saccade toward the remem-
Heide et al: Double-Step Saccades and Cortex
A) Double-step stimulus
normal subject
right parietal patient
, , , ,
1st saccade +
742 Annals of Neurology
Vol 38 No 5
bered location of the target, with a latency of less than 400
msec after the previous saccade). Saccadic latencies, amplitudes, and maximal velocities were calculated by the program
and stored for statistics. Furthermore. we assessed the duration of the intersaccadic interval, eye position after primary
saccades in response to the 1st and 2nd target, and final eye
position (FEP) after the last corrective saccade to the location
of the 2nd target. When the r!nd target had disappeared before the end of the 1st saccade, the condition of retinospatial dissonance was considered to be given, and these double-step responses entered statistics separately from those
without spatial dissonance. Thus, the occurrence of retinospatial dissonance in any particular trial depended on the
response latency of the 1st saccade plus its duration, which
had to be longer than the total presentation time of both
targets. Consequently, the frequency of saccades with dissonance (ranging between 40Tf and 80% in normal subjects)
was highest with the shortest double-step stimulus ( 100180
msec) and in patients with i:he longest response latencies
(PPC patients). In the following cases, responses were excluded from further analysis: ( 1 ) when the 1st target was
omitted (only one saccade being performed directly to the
2nd target), (2) when the response was anticipatory (latency
of less than 80 msecj, ( 3 )when the FP was not fixated during
rarget presentation. Second saccades were labeled as "erroneous" either when there was no saccade t o the 2nd target or
2nd saccade
Fig 2. EsampIeJ of double-step smades in a normal subjeit ( A )
and a patient with a right parietal lesion ( B ) 3the two targets
being presented at 6" iii opposite hemifields. The trares in each
panel 5 b O N ' tbe horizontul position (R = rightward, L = leftward) of the right eye (solid linei and the target (broken
line). Retino-Jpatiul diisonance omirved ajier the 1st saiiadt,
ai the 2nd target was not iisible anji more at that time. The location of the blanked 2nd target i.r markcd by the fine dashed
line. Ideallji it shorild be met bji the final eye position, the error
ofw~hiihi.r indicated Ly thin biheaded arrows. It i j rather
small in the normal subjeit. bzrt large in the patient who erroneously perfrms the 2nd saicade uccording t o the retinal iuctor of
the 2nd target. indicated b.y thick vertical arrows.
I Retino-spatial dissonance
final eye position
B) Frequency of erroneous 2nd saccades
"1 d-
I) R - R
2) R - L
3) L - L
4) L - R
3) L - L
4) L - R
\Noretino-spatial dissonance
1) R - R
2) R - L
stimulus condition
Oleft SMA
Fig 3. ( A )The four different types of double-J.tep stimuli in the
right ( = R) or ldt I = L) hemifield. ( B ) Perientage of erroneous
(aborted or dysmetric) 2nd saccades with (centralpanel) and
without retbo-spatial dissonanie ibottom pane4 in each subgroup of subjects. plotted separately according to the four stimulus conditions of ( A ) .In all figures. asterisks indiiate a sign$cant differencefrom the control group ('p < 0.0s: **p < 0.01:
or '**p < 0.005). Although right prefrontal patients shou'ed
most errors with crorsed A-timuli(conditions 2 and 4). this was
independnt of whether dissoname ociurred. whereas parietal pat i ~ n t shad eletiated error rats speiifically in case of dissonance.
iuhen the 1st target was located in the iontralesional hemifield
(condition 1 for left and ionditions .3 and 4 for right parietal
when it was performed into the wrong direction or into the
wrong visual hemifield.
The U test was applied to analyze these parameters statistically for group differences, whereas the x2 test was used to
compare frequencies of certain events between groups, such
as the occurrence of erroneous 2nd saccades. Correlations
between clinical, visuo-spatial, and saccade-related parameters were assessed with Spearman's rank correlation test. The
level of statistical significance was 5 T or less.
Visually Guided Single-Step Saccades
In accordance with previous reports C12, 251, simple
visually guided saccades to targets in the contralesional
November 1095
hemifield were hypometric (gain 8376 versus 94 ?
8% in normal subjects; p < 0.01) and had significantly
prolonged latencies (291 msec in right and 240 msec
in left PPC lesions versus 190 ? 42 msec in normals;
p < 0.005 and p < 0.05, respectively) in both parietal
groups; whereas this was not the case in the frontal
patients, except for a tendency ( p < 0.1) of prolonged
latencies (229 msec) in right FEF cases. Contralateral
saccadic deficits correlated with the neglect score (latencies: r = 0.71, p < 0.005; gain: r = -0.48; p
< 0.05). In right PPC lesions, latencies of ipsilesional
saccades were prolonged (241 msec; p < 0.02), too,
but to a lesser extent than contralesional saccades.
Failure of Double-Step Saccadic Perf omance
In the double-step saccadic paradigm, some of the
frontal lesions often caused failure of the 2nd saccade
whether there was retino-spatial dissonance or not. In
contrast, parietal lesions did so only in case of dissonance, if the 1st saccade had been directed into the
hemifield contralateral to the side of the lesion. This
was most obvious in patients with right PPC lesions
and happened even when the 2nd target had appeared
in the healthy ipsilesional (right) hemifield, as illustrated in Figure 2: In comparison with the normal subject (A), the parietal patient’s (B) 1st saccade was delayed, but not dysmetric; whereas his 2nd saccade was
markedly dysmetric, with a large error of FEP. It did
not even reach the right hemifield where the 2nd target
had been presented; but it was directed back toward
the invisible fixation point. Obviously it was performed
according to the initial retinal vector of the 2nd target
(black vertical arrow in Fig 2), thus not taking into
account extraretinal information about the contralateral displacement of the eyes associated with the 1st
The specificity of this disorder as a parietal deficit of
the spatial programming of saccades is evident from
Figure 3B, where the percentage of failure or gross
dysmetria of 2nd saccades is plotted for each group of
subjects. Exclusively in parietal patients, it was elevated
above the level of the control group only in case of
retino-spatial dissonance (upper panel), but not without dissonance (bottom panel). This difference was
most marked when the 1st saccade had been directed
into the contralateral hemifield, being highly significant
in the right PPC group ( p < 0.001, in conditions 3
and 4 of the stimuli plotted in Fig 3A), and weaker in
the left PPC group ( p < 0.05, only in condition 1). It
was, however, not significant in the frontal groups
where the percentage of erroneous 2nd saccades was
either within normal limits (left SMA, left PFC) or
elevated both with and without retino-spatial dissonance (right PFC, right FEF). Consequently, when the
percentage of erroneous 2nd saccades without dissonance was subtracted from the percentage with disso-
nance (last column of Table l),the resulting difference
was significantly elevated only in parietal patients,
amounting to 48% in right and to 13% in left PPC
cases, but to 3% or less in all other groups.
In the mirror-symmetrical condition, when the 1st
target appeared in the ipsilateral and the 2nd target in
the contralateral hemifield (condition 2 in right and 4
in left hemispheric lesions), the percentage of aborted
2nd saccades was also elevated in both parietal groups
( p < 0.01), further in right FEF and in right PFC cases
( p < 0.01), with and without retino-spatial dissonance.
In parietal and right FEF patients, it correlated with the
neglect score ( r = 0.61; p < 0.03),obviously reflecting
hemineglect (or extinction) of a contralateral visual target. Correspondingly, when the two targets were presented in reverse temporal order, the contralateral 1st
target was neglected by parietal and right FEF patients
(in about 4 5 2 1versus l o g in normal subjects), so that
only one saccade was performed directly to the 2nd
Patients with PFC lesions, who had no visual hemineglect, often aborted the 2nd saccade in conditions 2
and 4 (see Fig 3) when it had to cross the vertical
meridian, independent of the hemifield where the 2nd
target had appeared and independent of retino-spatial
dissonance. Further, patients with PFC or right FEF
lesions and, to a lesser extent, patients with SMA lesions generally tended to execute the saccade to the
location of the 1st target after the saccade to the 2nd
target, thus in the reverse temporal order, irrespective
of the targets’ retinotopic locations. Obviously frontal
lesions generally impair the temporal control and triggering of voluntary double-step saccadic sequences, independent of the influence of visual hemineglect or
retino-spatial dissonance.
Spatial Accuracy of Double-Step Saccades
In normal subjects, the mean amplitude gain of the 1st
saccade (reflected by P I in Fig 4) was similar to that
of single-step saccades when both targets were located
in opposite hemifields (93 -+ 185%) and was slightly
lower (85 2 12%))in the other conditions (1 and 3).
With increasing response latencies, eye position after
the 1st saccade was influenced more by the location of
the 2nd target, according to the amplitude transition
function described in previous double-step studies [26,
271, until finally, with the longest latencies, the saccade
to the 1st target was often aborted, as mentioned in
the previous section. When patients executed saccades
to the 1st target, their amplitudes were within normal
limits, except for hypometria of contralesional saccades
in parietal patients, analogous to single-step saccades
(significant for right PPC lesions with a mean gain of
7575 in condition 3 and 84% in condition 4; tendency
for left PPC lesions with a mean gain of 79% in condition 2).
Heide et al: Double-Step Saccades and Cortex
Error of mean eye position
i[2 r
5) single-step
8 peg
I 1
r .
4) L R
-1 -
-1 -
-2 L
-3 c
-2 -3 -
-5 k
FEP pi
lrcontrol aleft PFC *left SMA eleft PPC fright PFC +right FEF Wright PPC
Fig 4. Difference between mean horizontal eye position ( i n caje
of retino-spatial dissonance)and target position (the latter is indicated by the horizontal0 line) after the 1st sacrude (= P l ) .
after the 2nd .taccade ( = P2). and after c0rrectii.e saccades toward the 2nd target ( = FEP), plotted separately for the four
different types of double-step stimuli. according t o Figure 3, and
for rightuurd ( = R) and Leftuaurd ( = L) i'isziaL4 guided
single-step saccades (condition 5 ) . Upward deJection means
rightward deviation. dmc'nuiard'dejection leftward deviation
from target position. FEP = final eye position.
In all groups of subjects, spatial accuracy of the 2nd
saccade in case of retino-spatial dissonance was lower
than that of the 1st saccade, reflected in a larger deviation of mean eye position (P2 in Fig 4, compared with
P1) from target position, which was normalized to zero
in Figure 4.In the control group, this difference was
most marked in conditions 2 and 4 , where even the
mean FEP after corrective saccades deviated more
from target position than P1. Further, standard deviations of P2 and FEP (ranging between 1.3" and 3.0",
not shown in Fig 4 ) were much larger than those of
P1 (0.1-0.2"). After parietal lesions, mean P2 and FEP
deviated significantly more from target position (toward the fixation point, thus being hypometric) than
in the control group, when either the 1st or the 2nd
target of the double-step stimulus was located in the
contralateral hemifield (stimulus conditions 2, 3, and 4
for right PPC lesions and conditions 2 and 4 for left
PPC lesions), although P2 was improved to some ex-
744 Annals of Neurology
Vol 38 No 5
I -7
I -8
FEP 10" R
10" L
tent by corrective saccades. In frontal patients, mean
P2 and FEP were not significantly different from normals, except for right PFC cases, where the larger deviation from target position in condition 2 is due to the
increased percentage of erroneous 2nd saccades, which
was discussed in connection with Figure 3, occurring
independent of retino-spatial dissonance.
Whereas the deviation of mean eye position from
target position in Figure 4 has illustrated hypometria
of double-step saccades in parietal patients, Figure 5
shows the extent of dysmetria reflected in the mean
absolute error of FEPs in the single trials. As a consequence of aborted 2nd saccades (see Fig 3), right PFC
patients showed a high error in response to crossed
double-step stimuli (conditions 2 and 4 ) with and without retino-spatial dissonance, further parietal and right
FEF patients were impaired in condition 2 or 4, because they often neglected contralateral 2nd targets.
Specifically parietal patients, however, showed significant dysmetria of 2nd saccades only in case of retinospatial dissonance whenever the 1st saccade had been
directed into the contralateral hemifield (conditions 3
and 4 for right and 2 for left PPC lesions). This parietal
deficit of spatial accuracy did not correlate with the
neglect score, patients' age, duration, or volume of lesions. In right PPC patients, it correlated with impairment in copying Rey's figure ( r = 0.72; p < 0.05),
thus implying a more general deficit of visuo-spatial
November 1995
Mean error of FEP
/Retino-spatial dissonance
1) R - R
2) R - L
3) L - L
4) L - R
3) L - L
4) L - R
[No retino-spatial dissonance
6 t
1) R - R
2) R - L
stimulus condition
(Ocontrol Oleft PPC mright PFC Wright FEF l r i g h t PPC
F i g 5 . Mean ab.rolute error offinal t
y positions (FEW after
double-step triah. plotted separateJyfor the four different stirnub s conditions of Figure 3. Standard deviations of the conrrol
group are indicated by tiertical lines. Residts of' left supplementa y motor area and ldt prefiontal cortex patients are not shoumn
in this diagram. because they usere not signif candy different
from the control group.
Latencies of Double-Step Saccades
Mean latencies of contralesional 1st saccades were significantly ( p < 0.01) prolonged in the parietal groups,
ranging between 317 and 357 msec (versus 267 msec
in the control group) and correlating with the prolonged latencies of contralesional visually guided single-step saccades ( r = 0.75; p < 0.05). In the frontal
groups, most latencies were within normal limits, except for a delay of 1st saccades in right FEF patients
(310 msec; p < 0.05) only in condition 3, where both
targets were presented in the contralesional hemifield.
The latency of the 2nd saccade was assessed in terms
of the intersaccadic interval. It showed considerable
variability and was prolonged specifically in left SMA
patients (527 msec versus 310 & 215 msec in normal
subjects; p < 0.05), concerning only contraversive 2nd
saccades that had to cross the vertical meridian (condition 4 ) , no matter if retino-spatial dissonance had occurred or not.
Nornzal and Pathologiizf Performance
of Double-Step Saccades
Introduced by Westheimer [28}, visual double-step
stimuli have frequently been used to investigate the
spatio-temporal programming of saccades. It has been
shown that the two saccades are programmed in parallel and can interfere with each other [26, 27). For example, the amplitude of the 1st saccade is the more
influenced by the location of the 2nd target, the later
it starts after its appearance (amplitude transition function). This could be confirmed in our normal subjects.
Spatial programming of the 2nd saccade is particularly
interesting in Hallett's and Lightstone's flashed version
of the double-step stimulus [ 111, because the arising
retino-spatial dissonance permits the investigation of
the specific contribution of extraretinal signals that
monitor eye position or eye displacement associated
with a previous saccade. We applied this stimulus in
the present study and could confirm that normal subjects are able to perform this task rather well, but with
a lower spatial accuracy of the 2nd saccade, compared
with the 1st saccade or to visually guided single-step
saccades. For explanation, the following reasons might
be discussed: First, saccades without visual feedback
are less accurate than visually guided saccades. Second,
the extraretinal signal on eye position or eye displacement has been shown to be quite variable, sluggish,
and damped around the time of a saccade, thus leading
to perceptual and saccadic mislocalization of targets
flashed during that time [29-317.
In our patients, we identified the following deficits
of double-step saccades that we will discuss later on
(Table 2): (1)Parietal lesions cause a retinotopic deficit
of contralateral 1st saccades in terms of increased latencies and hypometria, analogous to visually triggered
single-step saccades. (2) When the two targets are presented in opposite visual hemifields, parietal and right
FEF lesions cause neglect of the contralateral target.
(3) The most important and specific parietal deficit was
failure or dysmetria of 2nd saccades only in case of
retino-spatial dissonance when the 1st saccade had
been directed into the hemifield contralateral to the
side of the lesion. ( 4 ) Lesions of frontal areas (FEF,
PFC, SMA) impair the temporal order and the triggering of double-step saccadic sequences, irrespective
of the occurrence of retino-spatial dissonance.
Contralateral Defcits and NegLect: PPC and FEF
Delayed initiation and hypometria of contralateral 1st
saccades following posterior parietal lesions correspond to analogous deficits of single-step saccades toward visual targets in the contralesional hemifield [ 12,
25, 323, thus confirming the important role of the PPC
for visually triggered saccades 181. We consider this
saccadic disorder as a contralateral retinotopic deficit,
Heide et al: Double-Step Saccades and Cortex
Table 2. Summayi of Signqicant Dejcits in the Double-Step Task
Lesion Site
Deficit of the 1st Saccade
Deficit of the 2nd Saccade
PPC (R > L)
Contralateral retinotopic
Accuracy: hypometric
Initiation: delayed
Contralateral hemineglect
Initiation: failure with C-I targets
Spatial, only in case of RSD (deficit of corollary discharge)
Accuracy: dysmetricifailure after contralateral 1st saccades
Contralateral hemineglect
Initiation: failure with I-C targets
Motorispatial working memory?
Failure of crossed saccades (C-1, I-C) (independent of
PPC > FEF (R > L)
Wrong temporal order
Initiation: delayed for I-C targets (independent of RSD)
Wrong temporal order
C-I, I-C ( = crossed stimuli), 1-1 and C-C ( = uncrossed stimuli) are the four target combinations, with C = contralcsional, 1
RSD = retino-spatial dissonance. R > L means that deficits were more pronounced with right hemispheric lesions.
PPC = posterior parietal cortex; FEF = frontal eye field; PFC = prefrontal cortex; SMA = supplementary motor area.
for the following two reasons: First, it is not a directional deficit, because contraversive 2nd saccades with
two ipsilateral double-step targets were performed normally; second, in a previous study [ 3 3 ] , the deficit did
not correspond to head- or hod y-centered coordinates,
because it was independent of the orbital position from
where the saccade was started.
The neglect of the contralateral target with crossed
double-step stimuli (mirror-symmetric target locations
in both hemifields) in parietal and FEF patients is obviously a consequence of contraiateral visual hemineglect, as it correlated with its severity. This deficit
occurred both with and without retino-spatial dissonance, but only when there was an ipsilateral target
to attract attention, like in the cuing task of Posner
and co-workers 134’1. These findings confirm the close
relationship between the orienting of attention and the
initiation of saccadic eye movements that specifically
characterizes neurons in the PPC coding “attentional
vectors” 19, 32).
Spatial Ddcit of Saccade Mttrics: PPC
The dissociation of a visual. target’s retinal and spatial
coordinates during the double-step task made it possible to identify a deficit of saccade metrics that was
specific for parietal lesions, being more marked in right
than left hemispheric lesions; following a 1st saccade
into the contralesional hemifield, the 2nd saccade was
often dysmetric or missing, but only in case of retinospatial dissonance when the extraretinal signal is essential for spatial accuracy. In contrast to the contralateral
retinotopic deficit mentioned in the previous section,
this occurred even when the 2nd target was located in
the healthy ipsilateral hemifield or when the 2nd saccade had to be performed in ipsiversive direction, indicating that it is neither a retinotopic nor a directional
deficit nor one of spatiotopic coding. Rather it is a
failure of efference copy (corollary discharge), because
PPC patients are impaired in registering extraretinal
information about the motor vector (amplitude and direction) of a contralateral saccade and using this for
updating the spatial representation of the next target.
These results support the evidence derived from
monkey studies that neurons in the PPC that have access to saccade-related extraretinal signals are essential
for the maintenance of spatial stability across saccades
and for the necessary transformations of spatial coordinates {7, 9). However, our findings do not support the
view that the PPC contains a spatially invariant target
representation in stable head- or body-centered coordinates, requiring continuous reliance on eye- and headposition signals {7, 35, 361. According to this model,
parietal lesions should cause spatiotopic deficits, restricted to saccade targets in the contralateral hemispace. This was not the case in our parietal patients;
rather they also failed to acquire ipsilateral targets that
should be accessible in case of a stable spatial map. In
contrast, our results are in perfect agreement with the
theory that target position is represented in retinocentric coordinates, with continuous spatial updating according to the current gaze location. The critical neurons, found in the lateral intraparietal area (LIP) of
monkeys, anticipate the retinal consequences of an intended saccade by shifting their receptive fields appropriately, thus updating and remapping the retinal representation of the visual world for a continuous
maintenance of spatial accuracy [lo]. As they do this
up to 80 msec prior to the saccade, they must use an
efference copy signal providing information about the
intended eye displacement. Lesions of these neurons
should cause deficits of taking into account saccadeassociated eye displacement, like observed in our patients.
The same type of disorder (even more severe) has
746 Annals of Neurology Vol 38 No 5 November 1995
already been reported in a patient with a large right
fronto-parietal lesion 1331. Our study has shown that
it has to be attributed predominantly to the parietal
lesion. Even though a spatially accurate neuronal signal
during the performance of double-step saccades has
been found not only in parietal, but also in FEF and
collicular neurons of monkeys [4, 5,9}, the deficit did
not turn out in patients with FEF lesions or might have
already recovered. There seems to be a functional specialization, i.e., whereas FEF neurons code the location
of a saccade target rather in oculomotor coordinates,
the important phenomenon of remapping its retinal
coordinates prior to a saccade was found only in PPC
Significance for Visuo-Spatial Perception
The parietal deficit of spatial stability across saccades
has implications for visuo-spatial orientation in general, reflected in these patients’ impaired performance
in copying Rey’s figure, which indicates an inaccurate
or unstable representation of spatial location. Thus, our
data confirm the key role of the PPC for the analysis
of visual space, based on clinical [ 3 7 , 381 and neurophysiological evidence [39, 401, with a dominance of
the right hemisphere that also happened in our cases.
An accurate representation of visual space is a prerequisite for the visual guidance of actions in general. That
the saccadic deficit in our patients did not correlate
with other parietal disorders such as visual hemineglect
or optic ataxia could be attributed to a functional specialization of areas within the PPC, which is evident at
least from primate studies [ 8 ] . Lesions in our patients
seem to have damaged predominantly the homologue
of area LIP, which is particularly involved in the control
of saccades and saccade-related spatial constancy.
Frontal Lobe Control of Saccadic Sequences
In our study, frontal lesions did not significantly affect
spatial accuracy of saccades, but rather the initiation of
the 2nd saccade in the double-step task, independent
of retino-spatial dissonance. In SMA patients, it was
delayed, in terms of a prolonged intersaccadic interval,
when it had to be directed from the ipsilateral into
the contralateral hemifield. This delay might reflect a
specific deficit of timing voluntary saccadic sequences,
in correspondence with the role of the SMA for the
execution of motor sequences, particularly if they are
internally triggered or memorized, which has been
shown for skeletal movements [41, 4 2 ) and for sequences of memory-guided saccades 14 31. Patients
with prefrontal lesions often aborted contra- and ipsiversive 2nd saccades, when they had to cross the vertical meridian. This reflects a motor impairment of
crossed saccadic sequences and can be explained with
respect to the role of the PFC for a spatial working
memory E13, 441, as the memory trace of the 2nd
target (located in the opposite visual hemifield) could
have been lost during the processing time of the 1st
saccade. Further, frontal patients generally tended to
perform the double-step task in the wrong temporal
order, in accordance with a theoretical model [45] that
assigns computation and storage of temporal order information to frontal lobe (FEF or PFC) neurons involved in visual fixation.
In summary, the frontal and parietal cortices supplement each other in the control of double-step saccadic
sequences; i.e., frontal areas are more involved in temporal aspects and the posterior parietal cortex in the
spatial programming. For diagnostic purposes, the double-step paradigm should be helpful for the assessment
of a specific parietal deficit of spatial constancy.
We rhank H.-M. Teicherr and H.-J. Berkenrien for help in statistical
problems, H. Briickmann and H. D. Weiss for giving us the CT
and MR scans, and, in addition, K. Miiller (AMTECH company,
Weinheim) and T. Piskol for rheir assistance in technical and software
problems and TWO reviewers for many helpful comments to improve
the manuscript.
1 Von Helmholtz H. Handbuch der physiologischen Optik, Vol
111. Hamburg: Voss, 1866 (English translation by Southall JPC.
Helmholtz treatise on physiological optics, vol 111. The perceptions of vision. New York: Dover, 1962)
2. Bridgeman B, van der Heijden AHC, Velichkovsky BM. A
rheory of visual stability across saccadic eye movements. Behav
Brain Sci 1394;17:247-292
3 . Von Holst E, Mittelstaedt H . Das Reafferenzprinzip (Wechselwirkungen zwischen Zentralnervensystem und Peripherie).
Naturwissenschafren 1950;3?:464-476
4. Mays LE, Sparks DL. Dissociation of visual and saccade-related
responses in superior colliculus neurons. J Neurophysiol 1980;
5. Goldberg ME, Bruce CJ. Primare frontal eye fields. 111. Mainrenance of a spatially accurate saccade signal. J Neurophysiol
6. Schlag J , SchlagRey M, Pigarev 1. Supplementary eye fieldinfluence of eye position o n neural signals of fixation. Exp Brain
Res 1990;OO:302-306
7 . Andersen RA, Essick GK, Siege1 RM. Encoding of spatial location by posterior parietal neurons. Science 1985;230:456-458
8. Andersen RA, Gnadt JW. Posterior parietal cortex. In: Wurtz
R H , Goldberg ME, eds. The ncurobiology of saccadic eye
movements. Review of oculomotor research, vol 3. Amsterdam:
Elsevier, 1989:315-335
9. Goldberg ME, Colby CL, Duhamel J-R. Represenration of visuomotor space in the parietal lobe of rhe monkey. Cold Spring
Harbor Symp 1390;55:729-739
10. Duhamel J-R, Colby CL, Goldberg ME. The updaring of the
representarion of visual space in parieral cortex by intended eye
movements. Science 1992;255:90-92
11. Halletr PE, Lightstone AD. Saccadic eye movements to flashed
targets. Vision Res 1976;16:107-114
12. Pierrot-Deseilligny C. Corrical control of saccades. Neuroophrhalmology 1991;11:63-75
13. Funahashi S , Bruce CJ, Goldman-Rakic PS. Neuronal activity
related to saccadic eye movemenrs in the monkey’s dorsolateral
prefrontal cortex. J Neurophysiol 1931;65:1464-1483
Heide et al: Double-Step Saccades and Cortex
14. Heide W, Zimmermann E, Kompf D. Double-step saccades
in patients with frontal or parietal lesions. Soc Neurosci Abstr
1993;19:427 (Abstract)
15. Lczak MD. Neuropsychological assessment. 2nd ed. New York:
Oxford University Press, 1083
16. Halligan PW, Cockburn J. Wilson BA. The behavioural assessment of visual neglect. Nruropsychol Rehabil 1991; 1:5-32
17. Damasio H, Damasio AR. Lesion analysis in neuropsychology.
New York. Oxford University Press, 1989
18. Duvernoy HM. The human brain. Surface, three-dimensional
sectional anatomy and MRI. New York: Springer, 1991:31133 1
19. Lundin P, Pedersen F. Volun-ie of pituitary macroadeneomas:
assessment b y MRI. J Comput Assist Tomogr 1992;16:5 19-528
20. Filipek PA, Kennedy D N . Caviness VS, et al. Magnetic resonance imaging-based brain morphometry: development and application to normal subjects. Ann Neurol 1989;25:61-67
2 1 Godoy J, Liiders H, Dinner DS, et al. Versive eye movements
elicited by cortical stimulation of the human brain. Neurology
22. Petit L, Orssaud C , Tzourio N . et al. PET study of voluntary
saccadic eye movements in humans: basal ganglia-thalamocortical system and cingulate cortex involvement. J Neurophysiol 1993;69: 1009-10 I7
23. Fried I, Katz A, McCarthy G, et al. Functional organization of
human supplementary motor cortex studied by electrical stimulation. J Neurosci 1991;11:3656-3666
24. Katz B, Miiller K, Helmle H. Binocular eye movement recording with C C D arrays. Neuroopthalniology 1987;7:81-9 1
25. Heide W, Ktimpf D. Saccades after frontal and parietal lesions.
In: Fuchs AF, Brandt T. Biittnrr U, Zee DS. eds. Contemporary
ocular motor and vestibular research: a tribute to David A.
Robinson. Stuttgart: Georg Thiemc Verlag, 1994:225-227
26. Becker W, Jurgens R. An analysis of the saccadic system by
means of double step stimuli. Vision Res 1979;19:967-983
27. Gellman RS, Carl JR. Early responses to double-step targets are
independent of step amplitude. Exp Brain Res 1991;87:433437
28. Westheimer G . Eye movement responses to a horizontally moving stimulus. AMA Arch Ophthalmol 1954;52:932-94 1
29. Honda H. Eye movements to a visual stimulus flashed before,
during or after a saccade. In: Jeannerod M, ed. Attention and
performance, vol 13. Hillsdale: Erlbaum, 1990:567-582
30. Dassonville P, Schlag J, Schlag,-Rey M. Oculomoror localization
748 Annals of Neurology Vol 38 No 5
November 1995
relies on a damped representation of saccadic eye displacement
in human and nonhuman primates. Visual Neurosci 1992;9:
Gellman RS, Fletcher WA. Eye position signals in human saccadic processing. Exp Brain Res 1992;89:425-434
Lynch JC, McLaren JW. Deficits of visual attention and saccadic
eye movements after lesions of parieto-occipital cortex in monkeys. J Neurophysiol 1989;61:74-30
Duhamel J-R, Goldberg ME, Fitzgibbon EJ, et al. Saccadic dysmetria in a patient with a right frontoparieral lesion. Brain 1992;
Posner MI, Walker JA. Friedrich FJ, Rafal RD. Effects of parietal injury o n covert orienting of atrention. J Neurosci 1984;4:
Galletti C , Battaglini PP, Fattori P. Parietal neurons encoding
spatial locations in craniotopic coordinates. Exp Brain Res 1993;
Robinson DA. Models of the saccadic eye movement control
system. Kybernetik 1973;14:71-83
Critchley M. T h e parietal lobes. London: Edward Arnold, I953
DeRenzi E. Disorders of space exploration and cognition. New
York: Wiley, 1982
Mountcastle VB, Lynch JC, Georgopoulos A, et al. Posterior
parietal association cortex of the monkey: command functions
for operations within extrapersonal space. J Neurophysiol 1975;
3 8 ~ 8 1-908
Stein JF. T h e representation of egocentric space in the posterior
parietal cortex. Behav Brain Sci 1992.15:691-700
Deecke L, Kornhuber H H , Lang W, er al. Timing function of
the frontal cortex in sequential motor and learning tasks. H u m
Neurobiol 1985;4:143-154
Mushiake H, Inase M, Tanji J. Selective coding of motor sequence in the supplementary motor area of monkey cerebral
cortex. Exp Brain Res 1990;82:208-2 10
Gaymard B, Pierrot-Deseilligny C, h v a u d S. Impairment of
sequences of memory-guided saccades after supplementary motor area lesions. Ann Neurol 1990;28:622-626
Funahashi S, Bruce CJ. Goldman-Rakic PS. Dorsolareral prefrontal lesions and oculomotor delayed-response performance:
evidence of “mnemonic” scotomas. J Neurosci 1993;13: 14791497
Grossberg S, Kuperstein M. Neural dynamics of adaptive sensory-motor control. Expanded ed. New York: Pergamon Press,
Без категории
Размер файла
992 Кб
implications, orientation, step, cortical, saccade, double, spatial, control
Пожаловаться на содержимое документа