Activation of frontoparietal cortices during memorized triple

advertisement
European Journal of Neuroscience, Vol. 13, pp. 1177±1189, 2001
ã Federation of European Neuroscience Societies
Activation of frontoparietal cortices during memorized
triple-step sequences of saccadic eye movements:
an fMRI study
W. Heide,1 F. Binkofski,2 R. J. Seitz,2 S. Posse,3 M. F. Nitschke,1 H.-J. Freund2 and D. KoÈmpf1
1
Department of Neurology, Medical University at LuÈbeck, D-23538 LuÈbeck, Germany
Department of Neurology, Heinrich-Heine-University, DuÈsseldorf, Germany
3
Institute of Medicine, fMRI Unit, Research Center JuÈlich, Germany
2
Keywords: double-step saccades, efference copy, human frontal eye ®elds, parietal eye ®elds, visuospatial attention
Abstract
To determine the cortical areas controlling memory-guided sequences of saccadic eye movements, we performed functional
magnetic resonance imaging (fMRI) in six healthy adults. Subjects had to perform a memorized sequence of three saccades in
darkness, after a triple-step stimulus of successively ¯ashed laser targets. To assess the differential contribution of saccadic
subfunctions, we applied several control conditions, such as central ®xation with or without triple-step visual stimulation, selfpaced saccades in darkness, visually guided saccades and single memory-guided saccades. Triple-step saccades strongly
activated the regions of the frontal eye ®elds, the adjacent ventral premotor cortex, the supplementary eye ®elds, the anterior
cingulate cortex and several posterior parietal foci in the superior parietal lobule, the precuneus, and the middle and posterior
portion of the intraparietal sulcus, the probable location of the human parietal eye ®eld. Comparison with the control conditions
showed that the right intraparietal sulcus and parts of the frontal and supplementary eye ®elds are more involved in the execution
of triple-step saccades than in the other saccade tasks. In accordance with evidence from clinical lesion studies, we propose that
the supplementary eye ®eld essentially controls the triggering of memorized saccadic sequences, whereas activation near the
middle portion of the right intraparietal sulcus appears to re¯ect the necessary spatial computations, including the use of
extraretinal information (efference copy) about a saccadic eye displacement for updating the spatial representation of the second
or third target of the triple-step sequence.
Introduction
Single-neuron studies in nonhuman primates have identi®ed four
cortical areas with presaccadic neuronal activity, re¯ecting their
speci®c involvement in the preparation and initiation of saccadic eye
movements: the frontal eye ®eld (FEF; Bruce & Goldberg, 1985), the
supplementary eye ®eld (SEF; Schlag & Schlag-Rey, 1987), the
parietal eye ®elds, particularly the lateral intraparietal area (LIP;
Barash et al., 1991) on the lateral bank of the intraparietal sulcus (IPS),
and ± for delayed saccades and memory-guided saccades ± the
dorsolateral prefrontal cortex (PFC; Boch & Goldberg, 1989;
Funahashi et al., 1991). Other areas are also involved in the
preparation of saccades, such as the anterior cingulate cortex (ACr).
The probable location of these areas in human cerebral cortex has been
inferred from lesion studies and functional brain imaging (Fox et al.,
1985; Petit et al., 1993, 1996, 1997; Anderson et al., 1994; O'Sullivan
et al., 1995; MuÈri et al., 1996; Sweeney et al., 1996; Luna et al., 1998).
According to human lesions studies, each of these areas is critical
for the control of certain saccadic subfunctions (Pierrot-Deseilligny
et al., 1995; Heide & KoÈmpf, 1998): the FEF for the execution of
internally generated intentional saccades, the SEF for the triggering
of memorized saccadic sequences, the PFC for spatial working
memory, and the posterior parietal cortex (PPC) for visually triggered
Correspondence: Dr Wolfgang Heide, as above.
E-mail: heide_w@neuro.mu-luebeck.de
Received 25 April 2000, revised 5 January 2001, accepted 9 January 2001
re¯exive saccades and for saccade-related spatial transformations.
The latter were investigated using a ¯ashed version of the double-step
task that allows the dissociation of a target's retinal vector from its
saccadic motor vector (Hallett & Lightstone, 1976). In this task two
saccades have to be executed to the memorized locations of two
peripheral targets that had been ¯ashed previously in rapid succession. As the second target has disappeared before the ®rst saccade,
spatial accuracy of the second saccade requires updating of the
second target's retinal coordinates by subtracting extraretinal information (e.g. efference copy) about the motor vector of the ®rst saccade.
Normal human subjects perform this task quite accurately, except for
systematic perceptual and saccadic mislocalization when the second
target is presented as a very short ¯ash (2 ms) immediately before or
during the ®rst saccade (Honda, 1989; Dassonville et al., 1992;
Schlag & Schlag-Rey, 1995). Single-unit studies in monkeys have
detected visually responsive neurons in the superior colliculus (Mays
& Sparks, 1980), the FEF (Goldberg & Bruce, 1990) and the posterior
parietal area LIP (Goldberg et al., 1990; Barash et al., 1991) that
generate a spatially accurate signal preceding the second saccade of
the double-step task. This implies the availability of extraretinal
information about eye position or previous eye displacement in these
neuronal structures. Particularly in the FEF (Umeno & Goldberg,
1997) and in area LIP the efference copy signal is present even prior
to the ®rst saccade (by about 80 ms) and used to remap the neurons'
receptive ®elds according to the anticipated eye displacement of the
1178 W. Heide et al.
impending saccade (Duhamel et al., 1992a), thus updating the spatial
representation of the second target for the maintenance of spatial
constancy (Heide & KoÈmpf, 1997). In a computational model,
Dominey & Arbib (1992) have proposed that this dynamic remapping
of spatial target representations and the updating of retinal error caused
by changes in eye position might take place in posterior parietal area
LIP and is transferred from there to the FEF and the superior colliculus.
More recently these authors (Dominey et al., 1997) postulated that the
de®nite neuronal compensation for an eye movement (e.g. the ®rst
saccade of the double-step task) occurring between target presentation
and the corresponding saccade is accomplished by subtraction of a
damped signal of the change in eye position and that it should take
place on a subcortical level, downstream from the FEF, in order to
account for the ®ndings of collision experiments where after electrical
stimulation of the FEF during a visually guided saccade, the ®xedvector code of the stimulated saccade appears to be combined with the
estimated eye position at the time of the ®ctive target presentation, thus
compensating for eye displacement during the afferent and efferent
delay periods. Nevertheless, the critical role of the PPC for these
spatial transformations was con®rmed by clinical studies of doublestep saccades, where patients with posterior parietal lesions were
speci®cally impaired, in terms of dysmetric second saccades, whenever the ®rst saccade had been directed into the contralesional
hemi®eld (Duhamel et al., 1992b; Heide et al., 1995).
The objective of this study was to delineate cortical activation
patterns that might re¯ect these different subfunctions involved in the
performance of memorized saccadic sequences in humans. In contrast
to a similar study with positron emission tomography (Petit et al.,
1996), we used functional magnetic resonance imaging (fMRI) with its
higher spatial resolution, and we applied a different paradigm.
Analogous with the double-step task, our subjects performed
nonpredictive triple-step saccades. This should continuously activate
cortical areas involved in spatial updating of target representations and
in the processing of efference copy, whereas during the repetitive
performance of a prelearned saccadic sequence, as applied by Petit
et al. (1996), these computations might partly be compensated by
learning and prediction, and signal intensity in parietal areas might
decrease as an effect of task repetition (Dejardin et al., 1998). In order
to isolate the cortical activity speci®c to the triple-step task from its
basic visual and motor components, we used other paradigms for
control, either with an identical visual stimulus (during central
®xation) or with a similar motor output, in terms of single visually
or memory-guided saccades towards identical target locations. In
contrast to triple-step saccades, these tasks require neither in-advanceprogramming of saccadic sequences nor spatial updating of target
representations, but each saccade may be performed according to the
target's retinal coordinates. So the direct comparison of triple-step and
visually guided saccades should unveil cortical fMRI correlates of
these speci®c cognitive demands of the triple-step task and should test
the hypothesis that certain foci in the PPC are more involved or more
active in this respect than other cortical areas (Dominey & Arbib,
1992; Heide et al., 1995; Colby & Goldberg, 1999). To control for the
contribution of spatial working memory, we compared triple-step with
single memory-guided saccades. Preliminary results have been
published in abstract form (Heide et al., 1997).
Methods
Subjects and scanning procedures
We examined six naive, healthy and nonstrabismic right-handed
subjects, aged 27±41 years, with normal visual acuity and visual
®elds. All subjects gave their informed consent in written form, and
the study was approved by the local ethics committee of the HeinrichHeine-University (DuÈsseldorf, Germany). Functional MR images of
cerebral blood oxygen level-dependent signal changes (BOLD) were
performed using a 1.5 Tesla Siemens `Vision' MRI system
(SIEMENS Magnetom, Erlangen, Germany), equipped with echoplanar imaging (EPI) capabilities and standard radio frequency head
coil for signal transmission and reception. Using a midsagittal scout
image, 16 axial slice positions (0.1 mm interslice gap) were
orientated in the anterior±posterior commissure plane covering the
dorsal part of the brain above the temporal and occipital poles, thus
not including primary visual cortex. The following sequences were
used: gradient echoplanar imaging, TR = 3 s, TE = 66 ms, ®eld of
view = 200 3 200 mm2, a = 90°, matrix size = 64 3 64, in-plane
resolution = 3.125 3 3.125 3 4 mm3. In addition, high-resolution
anatomical images of the entire brain were obtained by using a
strongly T1-weighted gradient echo pulse sequence (fast low-angle
shot) with the following parameters: TR = 40 ms, TE = 5 ms,
a = 40°, 1 excitation per phase encoding step, ®eld of view =
25 cm, matrix size = 256 3 256, 128 sagittal slices with 1.25 mm
single slice thickness.
For each experiment, the protocol consisted of six cycles,
alternating between activation and control conditions. Each condition
lasted for 24 s, thus each cycle for 48 s, and one complete experiment
for 288 s. The respective fMRI data were stored in 96 volume image
®les, each containing 3 s of recording.
Visual stimuli and behavioural tasks
For visual stimulation, a red laser spot (diameter 0.5°, luminance
23.1 cd/m2, background luminance < 0.02 cd/m2) was projected to a
screen that was shown to the subjects by a mirror, positioned within
the head coil 15 cm above the eyes. The target task was a triple-step
stimulus (Fig. 1A), which is an extension of the double-step stimulus,
designed to achieve higher levels of cortical activation, and has also
been applied in monkey experiments (Tian et al., 2000). It started
with the presentation of a central ®xation point. After 1.5 s, three
peripheral targets of different horizontal eccentricities (5°, 7°, 8° or
10°) were ¯ashed successively for 265, 235 and 200 ms, respectively,
in a pseudorandomized order. Target locations in the right and left
visual hemi®elds were alternated and equally distributed within the
whole sample of trials, thus excluding a directional or hemi®eld bias
that might contaminate hemispheric lateralization. Subjects were
instructed to maintain central ®xation during target presentation and
to memorize the locations of the three targets. After a delay of
500 ms following the extinction of the third target, the ®xation point
disappeared, and three saccades had to be performed in darkness
towards these remembered locations in the correct order. After 3.3 s,
the central ®xation point reappeared to start the next trial. Thus,
altogether, each triple-step trial lasted for 6 s, and one activation
condition (24 s) consisted of four different triple-step sequences.
Three tasks served as control conditions without saccades: rest in
darkness, ®xation of a central laser spot and central ®xation during
triple-step stimulation, requiring the suppression of visually triggered
re¯exive saccades to the targets. Three further control conditions
implied the execution of saccades (Fig. 1), namely visually guided
saccades, single memory-guided saccades, and self-paced saccades in
darkness. The ®rst two of these tasks were identical to the triple-step
task in terms of motor output: they led to the same number and
sequence of saccades. In the visually guided task (Fig. 1B), target
steps occurred every 1.5 s, with pseudorandom amplitudes. In the
single memory-guided task (Fig. 1C), a single target was ¯ashed for
200 ms. After a delay of 500 ms following the extinction of the
ã 2001 Federation of European Neuroscience Societies, European Journal of Neuroscience, 13, 1177±1189
Functional MRI of triple-step saccades
1179
Data analysis
FIG. 1. Oculographic records during the saccade tasks in one of the
subjects. In each example, the upper trace plots horizontal eye position
(right eye, recorded with infrared re¯ection oculography outside the MRI
scanner), the lower trace plots horizontal target position. Upward de¯ection
means rightward deviation, ranging between 0° and 10° of visual
eccentricity, downward de¯ection means leftward deviation, respectively.
On the abscissa, time is plotted in milliseconds. (A) One sequence of triplestep saccades, with target positions of 10° right, 8° left and 5° right. After
6 s the central ®xation point reappears to start the next sequence. (B)
Analogous sequences of visually guided saccades and (C) of single
memory-guided saccades, with target positions identical to those in A.
target, the ®xation point disappeared and a saccade had to be
performed to the memorized target location in darkness. Two seconds
after the start of the trial, the target reappeared and served as a
®xation point for the next trial. Self-paced saccades were alternating
horizontal saccades of about 25° amplitude and a frequency of about
0.5±1.0 Hz, executed in darkness from the imagined left edge of the
screen to its right edge and backwards.
In all these tasks, each activation or control condition lasted for
24 s and was composed of four blocks of trials each lasting 6 s.
Before each measurement, at least one cycle of the respective task
was presented for practice. Performance was controlled by infrared
re¯ection oculography (Fig. 1) outside the scanner prior to the
experiments. In addition, a qualitative assessment of the subjects' eye
movements was obtained by electro-oculographic recordings during
fMRI measurements. For further control of subjects' ®xation
behaviour inside the MRI scanner, we performed an additional
fMRI experiment where the active condition consisted of central
®xation and the control condition of rest in darkness. As there was no
signi®cant activation of any of the cortical saccade areas, not even of
the frontal eye ®elds, we took this as a con®rmation that ®xation must
have been fairly stable.
Image analysis was performed on an SPARC II work-station (Sun
Microsystems, Palo Alto, CA, USA) using MATLAB (Mathworks
Inc., Natiek, MA, USA) and statistical parametric mapping (SPM96b,
Wellcome Department of Cognitive Neurology, London, UK; Friston
et al., 1994a, b, 1995a, b, 1997). First, the 96 volume images of each
condition were automatically realigned to each tenth image to correct
for head movement between scans (Friston et al., 1995a, 1997). Then
the images were coregistered and transformed into a standard
stereotactic space using the intercommissural line as the reference
plane for transformation (Talairach & Tournoux, 1988). For the
normalization procedure pixels were slightly smoothed to achieve
isotropic voxels representing 4 3 4 mm2 in the x- and y-dimensions,
with an interplanar distance of 4.4 mm. The effects of global volume
activity and time were removed as confounds, using linear regression
and sine/cosine functions (up to a maximum of 2.5 cycles per 75
scans). Removing the latter confounds corresponds to high-pass
®ltering of the time series to remove low-frequency artefacts, which
can arise due to aliased cardiac and other cyclical components.
In order to present the overall pattern of activation across subjects,
the stereotactically transformed data sets from each subject were
smoothed slightly by a Gaussian ®lter (root mean square radius of
4 mm) to compensate for individual variation in brain anatomy
(Steinmetz et al., 1990). The alternating periods of `baseline' and
`activation' were modelled using a simple delayed box-car reference
vector accounting for the delayed cerebral blood ¯ow after stimulus
presentation. Signi®cantly activated pixels were searched for by using
the `General Linear Model' approach for time-series data suggested
by Friston and colleagues (Friston et al., 1994a, b, 1995a, b, 1997;
Friston, 1995; Poline et al., 1995; Worsley & Friston, 1995).
Therefore we de®ned a design matrix comprising contrasts that tested
for signi®cant activations during each task separately (tests for simple
main effects). With these data group-activation maps were calculated
by pooling the data for each condition across all six subjects. Only
pixels passing a height threshold of Z = 2.33 (P < 0.01; degrees of
freedom corrected for correlation between adjacent time points) and a
cluster of at least 10 voxels (P < 0.05 for spatial extent, corrected for
multiple comparisons) were considered as signi®cant. Only for
intertask comparisons, e.g. triple-step saccades vs. visually guided
saccades, where the change in MR signal intensity was expected to be
much lower than in saccade tasks controlled vs. ®xation, the height
threshold was set to Z = 1.64 (P < 0.05) for each voxel. These
thresholds appeared reasonable on the basis of prior fMRI studies in
this and other labs (Petit et al., 1997; Binkofski et al., 1998). They
were de®ned a priori and used in all experiments for the analysis of
both individual and group data. The activated voxels surviving this
procedure were superimposed on `SPM-brain-projections' and on
high-resolution, MR-anatomical scans. With the aid of published
Talairach-coordinates (Roland & Zilles, 1996) and prominent sulcal
landmarks of each individual brain (precentral, superior frontal,
central, postcentral, intraparietal, cingulate and parieto-occipital
sulcus), clusters of activated voxels were assigned according to
their centre of mass activity to the following regions of interest in
each hemisphere: posterior parietal cortex, primary sensorimotor
cortex, posterior and anterior cingulate cortex, lateral and medial
premotor cortex (including the FEF and the SEF). For these clusters
we determined the numbers of signi®cantly activated voxels, as well
as mean signal intensity changes relative to baseline (Z-values) of the
voxels showing peak activation in each region of interest. If there
were several foci of peak activation within one region of interest, we
will mention only those voxels that are separated by a distance of at
ã 2001 Federation of European Neuroscience Societies, European Journal of Neuroscience, 13, 1177±1189
1180 W. Heide et al.
TABLE 1. Behavioural data: latencies and accuracies of saccades
TRS
VIS
MEM
Latency (ms) (for TRS: intersaccadic interval)
1st saccade
376 6 196
167 6 33
2nd saccade
439 6 202
161 6 24
3rd saccade
423 6 96
148 6 28
314 6 109
305 6 115
325 6 125
Amplitude gain
1st saccade
2nd saccade
3rd saccade
0.89 6 0.05
0.90 6 0.04
0.88 6 0.03
0.73 6 0.12
0.68 6 0.13
0.71 6 0.07
0.3° 6 0.1
0.2° 6 0.1
0.3° 6 0.2
1.2° 6 1.6
1.6° 6 1.6
1.9° 6 0.3
0.96 6 0.13
0.91 6 0.12
0.90 6 0.11
Error of ®nal eye position (°)
1st saccade
1.7° 6 1.6
2nd saccade
3.9° 6 3.0
3rd saccade
2.6° 6 0.9
Data are presented as means 6 SD. TRS, triple-step saccades; VIS, visually
guided saccades; MEM, single memory-guided saccades.
least 8 mm and that reach a Z-value of at least 4.0 (P < 0.0001) for
tasks controlled against ®xation, and of at least 3.09 (P < 0.001) for
intertask comparisons. We also determined Brodmann's areas (BA)
that were most probably involved.
Results
Behavioural data
Table 1 summarizes the behavioural data of saccadic performance
recorded with infrared re¯ection oculography outside the scanner.
Latencies of visually guided saccades were around 160 ms, as
subjects performed a considerable amount of re¯exive saccades with
latencies below 150 ms (37% on average). In contrast, latencies of
memory-guided saccades were much longer, often exceeding 300 ms,
and intersaccadic intervals of triple-step saccades were above
400 ms. Concerning saccadic accuracy, visually guided primary
saccades usually undershot the target by about 11% (their gain being
0.89 on average), which can be attributed to their relatively short
latencies. Target position was reached by means of a corrective
saccade so that the error of ®nal eye position was almost 0° (below
0.3°, which re¯ects just the inaccuracy of the recording procedure). In
contrast, due to their generation without visual feedback, the accuracy
of triple-step saccades and single memory-guided saccades was much
lower in terms of a higher standard deviation of their gain (ranging
around 6 0.12) and a signi®cantly higher error of ®nal eye position
(around 1.6°). The latter was maximal after the second and third
saccades of the triple-step sequence (on average 3.9° and 2.6°,
respectively), probably re¯ecting the fact that in contrast to singlememory-guided saccades, their spatial programming could not rely
on retinotopic information, but had to use extraretinal information
(efference copy) to update the spatial representation of the second and
third targets. In addition, primary saccades in the single memoryguided task were more hypometric than in the visually guided task,
whereas primary saccades in the triple-step task were quite often
hypermetric. These ®ndings are in accordance with previous studies
using the double-step task (Duhamel et al., 1992a; Heide et al., 1995).
Functional MRI data
Distributed frontoparietal networks of areas were activated by each of
the saccade tasks. We will mention only those foci of activation
whose reliability and statistical signi®cance was con®rmed by group
analysis using SPM. Although we used common statistical limits,
small foci of weaker activation might have been missed. The
procedure of normalization, smoothing, and coregistration across
subjects degrades anatomical precision. Nevertheless most foci of
activation in the population data were circumscript, distinct and
reliable across different tasks, and their location corresponded exactly
to recent high-resolution fMRI studies using individual and group
analysis (Luna et al., 1998; Berman et al., 1999). Due to the ®eld of
view limits of our MRI sequences, our study was restricted to the
dorsal parts of the cerebral hemispheres.
Triple-step saccades vs. ®xation
The performance of triple-step saccades, compared with pure central
®xation (without peripheral visual stimulation) as control condition,
yielded signi®cant activations in the regions of the key areas (cortical
`eye ®elds') known to be involved in the generation of saccades
(Figs 2A and 3; Table 2, upper part), i.e. the FEF of both
hemispheres, the left SEF and several distinct regions of the posterior
parietal cortex (PPC). Further foci of activation were found bilaterally
in the lateral premotor cortex (PMC), ventral and slightly anterior to
the FEF (more than 12 mm apart), along the inferior portion of the
precentral sulcus, at the border between BA6, 44, and 9 (for Talairach
coordinates see Table 2). It is remarkable that there was no signi®cant
activation of the PFC in any of these tasks. Lateral premotor
activation in the region of the FEF showed two distinct peaks in each
hemisphere located more than 10 mm apart. In accordance with
previous reports (Paus, 1996; Petit et al., 1996; Luna et al., 1998;
Berman et al., 1999) one of these peaks (called `sFEF' in Table 2)
was centred around the superior portion of the precentral sulcus and
its posterior bank, extending dorsally towards the posterior end of the
superior frontal sulcus (as shown in Fig. 2A). The second peak was
located more ventrally and laterally, centred around the middle
portion of the precentral sulcus (called `iFEF' in Table 2). Both FEF
foci were located in BA6. Anatomical analysis in each individual
subject demonstrated that they were more or less restricted to the
precentral sulcus and the lip of the precentral gyrus.
Activation of medial premotor cortex centred on the left anterior
BA6 on the dorsomedial wall of the hemisphere, where the human
homologue of the SEF (as de®ned in the monkey) has been allocated
in most recent fMRI studies of high spatial resolution (Luna et al.,
1998; Berman et al., 1999; Grosbras et al., 1999; Petit & Haxby,
1999), anterior to the paracentral sulcus and superior to the cingulate
sulcus. In most of their subjects these authors located the peak of SEF
activation posteriorly to the anterior commissure line, thus with
negative y-coordinates. In contrast, the medial premotor activation in
our experiment peaked at a positive y-coordinate of +4, but extended
far more posteriorly, beyond the anterior commissure line, and thus
within the region of the SEF. The discrepancy of y-coordinates could
be due to intersubject variability. Alternatively, the peak of activity at
y = +4 could re¯ect a distinct area anterior to the SEF, which might
be named `preSEF', analogous to the `preSMA' (Hikosaka et al.,
1996; Petit et al., 1996; Grosbras et al., 1998), which has been
reported to be located anterior to the classical supplementary motor
area (SMA) called `SMA-proper'. In this experiment we observed no
distinct peak of activation in the region of the corresponding `SEFproper' (with negative y-coordinates), but only when triple-step
saccades were compared with central ®xation during triple-step visual
stimulation as control condition (Table 2, lower part). There the
population analysis revealed two signi®cant peaks of activity in each
hemisphere. Their distance of 14 mm towards each other is suf®cient
to assume two distinct foci of activation, possibly corresponding to
the preSEF and the SEF-proper. So we have attributed the medial
premotor activity in all our experiments to either the preSEF or the
ã 2001 Federation of European Neuroscience Societies, European Journal of Neuroscience, 13, 1177±1189
Functional MRI of triple-step saccades
1181
FIG. 2. Activation maps (group statistics, SPM 96) during the performance of triple-step and visually guided saccades. (A and B) Triple-step saccades,
relative to central ®xation. The sections are centred (A) in the region of the left superior FEF (sFEF; Talairach coordinates of the voxel with peak activation:
x, y, z = ±32, ±8, 52) and (B) in the right PPC (20, ±72, 52). (C) Visually guided saccades, relative to central ®xation. Sections are centred in the region of
the left SEF and preSEF (±4, 4, 48). (D) Triple-step saccades, relative to visually guided saccades. Sections are centred in the right IPL (44, ±48, 36). FEF,
frontal eye ®eld; sFEF/iFEF, superior and inferior portion of the FEF; SEF, supplementary eye ®eld; preSEF, presupplementary eye ®eld; PMC, premotor
cortex; ACc, caudal part of the anterior cingulate cortex; mIPS, middle portion of the intraparietal sulcus (IPS); IPL, inferior parietal lobule; SPL, superior
parietal lobule; MP, medial posterior parietal region, extending into the precuneus (Prec.); ®xation 0, central ®xation without peripheral visual stimulation;
®xation N, central ®xation during triple-step visual stimulation.
SEF itself, but we admit that this distinction remains hypothetical and
cannot be proven de®nitely by our data, as SEF activation exhibits
interindividual variability particularly in anterior±posterior direction.
In the posterior parietal cortex (PPC), the triple-step task activated
several regions, with right-hemispheric dominance, most of them
centred along the IPS in each individual subject. When the control
condition was central ®xation without peripheral visual stimulation
(Table 2, upper part), these regions comprised the lateral bank of the
middle portion of the IPS, extending into the inferior parietal lobule
(IPL), further towards the medial bank of the posterior portion of the
IPS, extending into the superior parietal lobule (SPL), and the
posterior-most portion of the medial PPC (MP/Prec.), including the
dorsal precuneus, adjacent to the parieto-occipital sulcus (Fig. 2B).
Within these three parietal regions of interest, there were ®ve distinct
clusters of activation in the left hemisphere, and six distinct clusters
in the right hemisphere (Fig. 3; Table 2): two of the latter were
located in the SPL, one of them in the posteromesial PPC (MP/Prec.),
and three of them along the IPS, one in its posterior portion and two
ã 2001 Federation of European Neuroscience Societies, European Journal of Neuroscience, 13, 1177±1189
1182 W. Heide et al.
TABLE 2. Triple-step saccades vs. ®xation: clusters and foci of signi®cant activation (P < 0.01)
Left hemisphere
Region of interest
n
x
Right hemisphere
y
z
Z
Location
BA
Triple-step saccades vs. central ®xation without peripheral visual stimulation
Lateral premotor
139
±32
±8
52
6.26
sFEF
±48
±4
40
7.39
iFEF
±48
4
28
5.62
PMC
Medial premotor
71
±8
4
48
7.51
(pre)SEF
PPC
123
±40
±44
40
3.88
mIPL
±32
±56
56
6.99
IPS/SPL
±20
±72
56
6.58
SPL/MP
±16
±76
44
4.35
MP/Prec
±16
±60
68
3.62
MP/Prec
6
6
6/44
6
7 > 40
7 > 40
7
7
7
Triple-step saccades vs. central ®xation during triple-step visual stimulation
Lateral premotor
144
±32
±8
56
7.14
sFEF
±44
±4
44
7.58
iFEF
±48
4
28
6.01
PMC
Medial premotor
52
±4
4
48
5.76
preSEF
±8
±4
60
4.96
SEF
PPC
94
±32
±56
56
6.24
IPS/SPL
±20
±72
56
6.85
SPL
±12
±64
64
4.62
MP/Prec
6
6
6/44
6
6
7 > 40
7
7
n
x
y
z
Z
Location
BA
99
32
44
52
±4
±8
0
56
48
36
7.14
7.11
6.62
sFEF
iFEF
PMC
193
44
38
36
16
20
12
±48
±56
±76
±68
±60
±80
48
52
44
56
68
48
4.78
4.03
5.28
6.53
5.85
6.39
mIPS/IPL
mIPS
pIPS
SPL
SPL
MP/Prec.
40
7/40
7/19
7
7
63
28
48
48
8
4
20
20
12
0
32
40
±4
±4
4
0
8
±60
±72
±80
±88
±76
±80
56
44
28
52
44
68
52
48
28
40
24
6.27
4.41
4.60
4.62
4.21
5.12
6.35
6.20
5.38
4.81
4.10
sFEF
iFEF
PMC
SEF
pSEF/ACc
SPL
SPL
MP/Prec.
Prec./Cu.
pIPS/OC
IPL/OC
6
6
6/44
6
6/32
7
7
7
7/19
7/19
39/19
210
6
6
6/9
n, numbers of voxels signi®cantly activated in each region of interest; x, y, z, Talairach coordinates (in mm) of the voxel showing peak activation in each cluster.
Z, its Z-value; BA, Brodmann areas involved by each focus of activation. ACc, rostral part of the anterior cingulate gyrus; Cu, cuneus; sFEF/iFEF, superior and
inferior portion of the frontal eye ®eld; IPL, inferior parietal lobule; IPS, intraparietal sulcus; mIPS, middle portion of the IPS; pIPS, posterior IPS; MP/Prec.
medial posterior parietal region and adjacent dorsal precuneus; OC, lateral occipital cortex; PMC, lateral premotor cortex; Prec. precuneus; preSEF,
presupplementary eye ®eld; pSEF/ACc, border between the ACc and the preSEF; SPL, superior parietal lobule.
in its middle portion, extending towards its lateral bank into the
adjacent IPL (BA40). The latter foci in the right IPS/IPL were
speci®cally activated during triple-step saccades, but not during
visually guided, single memory-guided or self-paced saccades
(Fig. 3).
When the control condition was central ®xation during triple-step
visual stimulation (Table 2, lower part), triple-step saccades activated
almost the same network of frontoparietal areas. As mentioned
earlier, three additional foci were activated in the medial premotor
region on the medial wall of the hemispheres, namely the region of
left and right SEF-proper and of the right preSEF bordering the
caudal portion of the anterior cingulate cortex (ACc). However, in
contrast to central ®xation without triple-step stimulation as baseline
condition, there was no signi®cant activation in the middle portion of
the right IPS/IPL (Fig. 2B; Table 2). So this focus appears to be
related to common features of the activation and control conditions,
such as the triple-step stimulus itself or the required suppression of
re¯exive visually triggered saccades towards the ¯ashed targets.
Overall there was a prominent hemispheric asymmetry in this
experiment, showing left hemispheric dominance for the lateral
premotor areas (144 vs. 63 activated voxels) and right hemispheric
dominance for the PPC (210 vs. 94 voxels; Table 2).
Other saccade tasks
When visually guided saccades were used as active condition and
compared with central ®xation as control condition, a similar
frontoparietal network was activated, though to a lesser extent than
during triple-step saccades, re¯ected in lower numbers of voxels and
lower Z-values. Foci of signi®cant activation (Fig. 2C; Table 3) were
found in the superior and inferior portions of the FEF of both
hemispheres, in the right lateral PMC, ventral to the FEF, in the
region of the left preSEF, and in the PPC of both hemispheres,
centred in the SPL and on the medial bank of the IPS. Another focus
of activation located in the right precuneus did not reach statistical
signi®cance (Z > 4.0) in this condition, but was activated much
stronger in the memory-guided saccade tasks. In contrast to the triplestep task, there were no distinct peaks of activation in the regions of
the SEF-proper, the right preSEF, and in the middle portion of the IPS
or IPL (Fig. 3).
When successive single memory-guided saccades were performed
in the active condition and compared with central ®xation during
peripheral visual stimulation (¯ashing of the saccade targets) as
control condition (Table 4), the population data yielded foci of
signi®cant activation in the inferior portion of the FEF (iFEF), in the
right PMC, in the SPL bilaterally and in the right precuneus (MP/
Prec.), but not in the medial premotor region (SEF or ACc).
Self-paced saccades, executed in darkness, with rest in darkness as
control, activated the regions of the FEF and the ventro-anterior SEF
bilaterally, and also the left and right ventrolateral PMC, the right
ACc, and the left PPC (Fig. 3). Although the saccades were equally
distributed across both horizontal directions and both hemi®elds, FEF
activation was stronger in the left [two peaks at (28, ±8, 48), Z = 5.55,
and at (±52, ±12, 44), Z = 7.88] than in the right hemisphere, where
only the inferior portion of the FEF was activated [peak at (44, ±12,
44), Z = 7.91]. The peak of FEF activation during self-paced
saccades was located 4±8 mm more posterior (within the precentral
gyrus) than during the execution of visually guided or triple-step
saccades. The ventrolateral PMC was activated bilaterally [peaks at
(60, ±4, 32), Z = 6.51, in the left and at (56, ±4, 32), Z = 4.44, in the
right hemispheres]. The premotor cortex on the medial wall was
ã 2001 Federation of European Neuroscience Societies, European Journal of Neuroscience, 13, 1177±1189
Functional MRI of triple-step saccades
1183
FIG. 3. Task-dependent variability of cortical
foci activated during the performance of
visually guided saccades (vs. central ®xation),
triple-step saccades (vs. central ®xation) and
self-paced saccades (vs. rest in darkness). The
respective voxels of peak activation (Z > 4.0)
are plotted in Talairach space for each cluster
of the population data. Compared with the
coordinates in Tables 1 and 4, z-coordinates
of the parietal foci were corrected for
incongruency between Montreal Neurological
Institute (MNI) templates and Talairach
coordinates (Stephan et al., 1997) by
subtracting 13 mm. For the sake of clarity, the
right FEF is not shown in the sagittal plot (top
left) and parietal foci not in the coronal plot
(top right). The symbols for the different tasks
are explained in the ®gure (lower right). For
abbreviations see Fig. 2; pIPS, posterior
portion of the IPS.
TABLE 3. Visually guided saccades vs. ®xation: clusters and foci of signi®cant activation (P < 0.01)
Left hemisphere
Region of interest
n
x
Lateral premotor
54
±28
±44
Medial premotor
PPC
28
36
±4
±28
±24
Right hemisphere
y
z
Z
Location
BA
n
x
±8
±8
52
44
4.36
5.59
sFEF
iFEF
6
6
77
32
48
52
4
±56
±68
48
60
60
4.95
5.44
4.85
(pre)-SEF
SPL
SPL
6
7
7
28
20
[8
y
z
Z
Location
BA
±8
±8
0
56
44
36
6.69
6.44
5.64
sFEF
iFEF
PMC
6
6
6/9
±60
±68
68
44
5.82
3.12
SPL
MP/Prec.
7
7]
For abbreviations see Table 2. Peak activation of the cluster in square brackets was not signi®cant in this study (Z = 3.12, thus remaining below the postulated limit
of 4.0).
activated bilaterally in the region of the preSEF, at its border to the
cingulate cortex [peaks at (4, 8, 44) and (±4, 8, 44)], and more
rostroventrally, in the right ACc (8, 12, 36). Posterior parietal
activation was signi®cant in the left hemisphere, showing three foci
along the IPS, namely on its lateral bank (IPL) at (±40, ±52, 48),
Z = 4.14; on its medial bank (SPL) at (±32, ±60, 56), Z = 6.29; and
around its posterior end at (±24, ±72, 48), Z = 4.8.
Triple-step saccades vs. other saccade tasks
In order to delineate patterns of activity that might be related to
speci®c aspects of the triple-step task, we used the other saccade tasks
as control conditions. This reduced not only the number of activated
regions, but also their spatial extent and their activation level
(Table 5). With triple-step saccades, controlled vs. visually guided
saccades (Fig. 2D), there was no activity in the PMC, SPL or
precuneus, but signi®cant activation was found in the region of the
FEF (its superior portion bilaterally and its right inferior portion), in
the right and left ACc, further in the rostral portion of the right ACr,
in the right ventral PFC (BA10) and in the right IPL (BA40). The
latter activation overlaps with the ventral border of the focus in the
right middle IPS/IPL, evoked during the performance of triple-step
saccades vs. pure central ®xation (Table 2). It appears to be speci®c
for the execution of triple-step saccades, in contrast to all other types
of saccades investigated in this study.
If single memory-guided saccades were used as control, the triplestep task activated the regions of the FEFs (its right superior portion
and its inferior portion bilaterally), the left SEF and preSEF, the right
ACc (BA24), and the right medial PPC and dorsal precuneus as the
only parietal region. A focus in the right dorsolateral PFC (BA9/46)
did not reach signi®cance.
Discussion
Saccade-related fMRI activations revealed the classical frontoparietal
saccade areas, as de®ned in experimental studies: FEF, SEF and
several areas in the PPC. Some of the tasks activated also the AC and,
unexpectedly, the lateral PMC, ventral to the FEF (Fig. 4). No
signi®cant activation was observed in the dorsolateral PFC. This is
probably due to the fact that delay periods of memory-guided
saccades did not exceed 0.5 s, which might be too short for activating
ã 2001 Federation of European Neuroscience Societies, European Journal of Neuroscience, 13, 1177±1189
1184 W. Heide et al.
TABLE 4. Single memory-guided saccades vs. ®xation: clusters and foci of signi®cant activation (P < 0.01)
Left hemisphere
Region of interest
n
x
Lateral premotor
43
±44
PPC
28
±32
Right hemisphere
y
z
Z
Location
BA
n
x
±8
48
6.10
iFEF
6
88
44
48
48
±56
60
4.84
SPL
7
71
24
12
y
z
Z
Location
BA
±8
0
8
48
48
24
5.43
5.6
4.7
iFEF
iFEF/PMC
PMC
6
6
6/44
±72
±80
52
48
5.54
6.08
SPL
MP/Prec.
7
7
For abbreviations see Table 2.
TABLE 5. Triple-step saccades vs. other tasks: clusters and foci of signi®cant activation (P < 0.05)
Left hemisphere
Region of interest
n
x
y
Triple-step saccades vs. visually guided saccades
Lateral premotor, prefrontal
12
±28
Medial premotor
15
Right hemisphere
±4
z
Z
Location
±8
52
3.43
sFEF
12
32
4.37
ACc
BA
n
x
y
z
Z
Location
BA
6
12
24
10
36
44
32
4
4
44
±4
±4
48
8
20
±48
52
44
0
40
28
36
3.95
3.52
3.21
3.85
3.42
3.11
sFEF
iFEF
PFC
ACc
ACr
IPL
6
6
10
24/32
32
40
±8
±8
24
12
56
44
24
32
4.14
4.01
2.91
4.12
sFEF
iFEF
PFC
ACc
6
6
9/46]
24
±80
48
5.02
MP/Prec
PPC
11
Triple-step saccades vs. single memory-guided saccades
Lateral premotor, prefrontal
38
±44
±12
Medial premotor
PPC
29
±8
±8
4
±8
44
4.73
iFEF
6
59
48
48
5.17
3.65
preSEF
SEF
6
[10
17
32
48
36
4
10
12
7
ACr/ACc, rostral/caudal anterior cingulate gyrus; PFC, prefrontal cortex. Peak activation of the cluster in square brackets was not signi®cant in this study
(Z = 2.91, thus remaining below the postulated limit of 4.0). For other abbreviations see Table 2.
the classical areas controlling spatial working memory (Funahashi
et al., 1993; Brandt et al., 1998).
Each of the saccade tasks activated similar networks of cortical
areas. Nevertheless, the activation patterns in the different tasks and
task-control comparisons (Table 6) reveal some clear task-related
dissociations that have implications for the cortical representation of
saccade-related subfunctions and will be outlined in the following
sections.
Posterior parietal cortex and the representation of spatial
transformations
In congruence with the fMRI study by Luna et al. (1998), we found
three main regions of saccade-related activation in the PPC: the
lateral bank of the middle portion of the IPS (IPL), the medial bank of
the posterior IPS (SPL), and the precuneus. Similar foci were
activated by covert shifts of visuospatial attention (Nobre et al., 1997;
Corbetta et al., 1998; Gitelman et al., 1999; Kastner et al., 1999;
Nobre et al., 2000). Accordingly, neurons of area LIP in macaque
monkeys respond not only in association with saccades to visual or
remembered targets, but also with the allocation of an attentional
vector to behaviourally relevant or salient objects in their visual
receptive ®elds (Goldberg et al., 1990; Colby et al., 1996; Colby &
Goldberg, 1999) and with the intention to perform a saccade to this
spatial location (Andersen, 1995; Andersen et al., 1997). When the
visual stimulus was spatially dissociated from the saccade goal in the
antisaccade task, the presaccadic activity of most LIP neurons
encoded the location of the visual cue, rather than the direction of the
upcoming saccade (Gottlieb & Goldberg, 1999). However, it is
improbable that parietal activations in our fMRI study re¯ect merely
attentional shifts, as there were distinct activation patterns in the
different saccade tasks. Visually guided saccades activated only the
SPL, triple-step saccades all three regions bilaterally, more in the
right hemisphere, which re¯ects right-hemispheric dominance for
visuospatial and attentional processes involved in this task. In
contrast, self-paced saccades activated only the left PPC.
Our ®ndings demonstrate the existence of multiple saccade areas
(`eye ®elds') in the human PPC, similar to macaques (Thier &
Andersen, 1998). It is dif®cult to decide which of the parietal foci
corresponds to macaque area LIP. According to the pioneering fMRI
study by MuÈri et al. (1996), it should be the focus in the middle
portion of the IPS. However, this area was not signi®cantly activated
during visually guided or memory-guided saccades, but only during
triple-step saccades. Therefore it cannot be related to the execution of
saccades in general, nor to the allocation of visuospatial attention, nor
to spatial working memory, but it must re¯ect some speci®c cognitive
demand of the triple-step task. It was the only parietal focus that
remained activated in the critical intertask-comparison (triple-step vs.
visually guided saccades). Its location in caudal BA40 at the lateral
bank of the IPS and its right-hemispheric dominance correspond to
the common overlap of circumscribed parietal lesions that led to an
impaired spatial performance of double-step saccades in stroke
patients (Heide et al., 1995). As outlined earlier, these patients have
de®cits in using efference copy signals to update the retinotopic
representation of saccade targets in order to compensate for eye
ã 2001 Federation of European Neuroscience Societies, European Journal of Neuroscience, 13, 1177±1189
Functional MRI of triple-step saccades
displacement associated with a preceding saccade in contralesional
direction. We therefore assume that activation of this area during
triple-step saccades re¯ects the processing of these spatial transformations, as postulated for LIP neurons in monkeys, according to
evidence from single neuron (Goldberg et al., 1990; Duhamel et al.,
1992a; Andersen et al., 1997) and simulation studies (Droulez &
Berthoz, 1991; Dominey & Arbib, 1992). The latter have proposed
network models where the respective neurons (e.g. of area LIP) are
components of a `sensor-related' map in which the locations of
possible saccade targets are stored (memorized) with respect to
current eye position (i.e. in retinotopic coordinates). Each time the
eye direction changes the map is updated as neurons shift their spatial
target representations in the direction of eye displacement. Our
human data support the assumption that this dynamic remapping of
FIG. 4. Schematic lateral and medial surfaces of a human brain, with
cytoarchitectonic Brodmann areas of the premotor and parietal cortices. The
scheme shows the network of cortical areas involved in the control of
sequences of memory-guided saccades. The numbers indicate Brodmann
areas. The SEF, ACr/ACc and precuneus are located on the medial wall of
the hemisphere. ACr, rostral part of the anterior cingulate gyrus; mIPS,
middle portion of the intraparietal sulcus (lateral bank) and the adjacent
inferior parietal lobule; MP, medial posterior parietal cortex and dorsal
precuneus; M1, primary motor cortex; CS, central sulcus; CMAr/CMAc,
rostral and caudal part of the cingulate motor area. For other abbreviations
see Fig. 2.
1185
receptive ®elds might take place in the human homologue of LIP
within the PPC. However, the mechanism of this remapping remains
unclear. Neuronal responses with evidence for remapping during
double- or triple-step saccades were also found in the FEF of
monkeys (Umeno & Goldberg, 1997; Tian et al., 2000). They might
receive this information from LIP, but on the other hand, there is
experimental evidence from colliding saccade studies (Dominey et al.,
1997) that the de®nite neuronal compensation for an intervening
saccadic eye displacement takes place downstream from the FEF,
probably in subcortical structures, such as the cerebellum, and
appears to be accomplished by subtraction of a damped signal of the
change in eye position, in order to compensate for sensorimotor
transmission delays. Consequently, these authors speculate that
cortical remapping mechanisms (e.g. in LIP) might be different,
possibly achieving compensation for an eye displacement by taking
into account allocentric cues that are provided by a consecutive
presentation of double- or triple-step visual targets without a temporal
gap. As this was the case in our studies, we cannot exclude this
interpretation, but we consider it improbable, as in cases of parietal
lesions the de®cit was not spatiotopic, i.e. it could not be described in
allocentric or egocentric coordinates. Rather it was dependent on the
direction of eye displacement (motor vector) during the ®rst saccade,
indicating that this information is not adequately provided by
efference copy signals to update the spatial representation of the
second saccade goal (Duhamel et al., 1992b; Heide et al., 1995).
The importance of the focus in the right mid-IPS for spatial
transformations during visuomotor tasks is supported by numerous
other functional imaging studies (Dieterich et al., 1998; Goebel et al.,
1998; Vallar et al., 1999; Harris et al., 2000). Surprisingly, this focus
was not activated in our study when the triple-step stimulus appeared
in the control condition, during central ®xation. Therefore its
activation might be in¯uenced by the suppression of visually
triggered saccades towards the triple-step targets (Law et al., 1977)
or by the triple-step stimulus itself. It could re¯ect the coding of
stimulus shape (Sereno & Maunsell, 1998) or the spatial representation of its trajectory. Such a relationship would be in line with
recent PET and fMRI studies where imagery of visually guided ®nger
movements or of graphomotor trajectories activated the same site in
the middle IPS (Seitz et al., 1997; Binkofski et al., 2000). According
to fMRI and lesion data (Binkofski et al., 1998), the anterior IPS is a
critical site for the control of grasping visual objects and therefore is
likely to contain the human homologue for macaque area AIP
(anterior intraparietal), located anterior to LIP. Accordingly, the
middle IPS may represent the LIP-homologue. As further support for
this assumption, a PET study by Kawashima et al. (1996) revealed a
reaching-related focus in the anterior IPS and a saccade-related focus
adjacent to it, more posterior in the IPS.
TABLE 6. Survey of cortical areas showing signi®cant activation during the different task-control combinations
Task
FEF
PMC
VIS ± FIX
SP ± REST
MEM ± FIX
TRS ± FIX
TRS ± VIS
TRS ± MEM
++
++
++
++
++R > L
++
+R
++
+R
++
PFC
+R*
SEF/preSEF
ACc
+L(preSEF)
++(preSEF)
+R
++L > R
+L
+R
++
+R
ACr
IPL/IPS
+L
+R
++
+R
SPL/IPS
MP/Prec.
++
+L
++
++R > L
+L
+R
++
+R
Upper part of table, tasks vs. central ®xation or rest; lower part, intertask comparisons. VIS, visually guided saccades; FIX, central ®xation; SP, self-paced
saccades; REST, rest in darkness; MEM, single memory-guided saccades; TRS, triple-step saccades; PMC, ventrolateral premotor cortex; PFC, dorsolateral
prefrontal cortex; ACr, rostral anterior cingulate gyrus; ACc, caudal anterior cingulate gyrus; MP/Prec. medial posterior parietal region and dorsal precuneus; + or
++, unilateral or bilateral activation; R, right hemispheric focus; L, left hemispheric focus. *Ventrorostral PFC. For other abbreviations see Fig. 2.
ã 2001 Federation of European Neuroscience Societies, European Journal of Neuroscience, 13, 1177±1189
1186 W. Heide et al.
Table 6 demonstrates coactivation of posterior parietal (posterior
IPS and SPL) and lateral premotor areas (PMC) that was consistently
observed in all saccade tasks employing visual stimulation. This is in
agreement with the parieto-premotor network, as proposed for covert
shift of visuospatial attention (Corbetta et al., 1998; Gitelman et al.,
1999; Nobre et al., 2000) and for visuomotor control, subserving the
gradual transformation of extrinsic visuospatial information about
target location and movement trajectory into limb-centred motor
commands (Rizzolatti et al., 1997). The activation in the dorsal
precuneus with its location at the anterior bank of the parietooccipital sulcus corresponds anatomically to area V6A [parietooccipital (PO)] of macaque monkeys. In this visual area neurons
encode post-saccadic eye position and the spatial location of visual
targets for arm and eye movements in nonretinotopic coordinates
(Galletti et al., 1997; Nakamura et al., 1999). Alternatively, the focus
in the precuneus could be the homologue of the medial parietal eye
®eld (MP; Thier & Andersen, 1998). In humans the right dorsal
precuneus is also activated during covert attentional shifts and during
visuomotor sequence learning (Sakai et al., 1998). So this activation
does not appear to be saccade-speci®c, but involved in supramodal
encoding of space for visuomotor control.
Frontal eye ®elds
In the lateral premotor region, our data showed three distinct clusters
of activation around the precentral sulcus. The most ventral focus is
located clearly outside the FEF in the ventrolateral PMC. The two
dorsal clusters coincide with the assumed location of the human FEF
(Paus, 1996), centred around the precentral sulcus and the adjacent
portion of the precentral gyrus, within BA6, with z-coordinates
ranging between 44 and 51 mm. The existence of two distinct foci of
FEF activation (sFEF and iFEF) has recently been reported by PET
(Petit et al., 1996) and fMRI studies (Petit et al., 1997; Luna et al.,
1998), which corresponds exactly to our data, including the Talairach
coordinates. Depending on the saccadic paradigm, we found
variabilities in the location of peak FEF activation (Fig. 3): internally
triggered self-paced saccades led to a slightly more posterior (y =
±12) and inferior activation than visually guided or triple-step
saccades (y = ±4 or ±8), as in previous PET studies on self-paced
saccades (Petit et al., 1993, 1996; Law et al., 1998). This effect might
be attributed to the larger amplitude of self-paced saccades, as largeamplitude saccades led to a more ventrolateral activation in all
previous studies (Paus, 1996).
Functionally, FEF activation in our study re¯ects the execution of
saccades in general, as it was not activated during visual ®xation, but
in each of the saccade tasks. The intertask comparisons show that
FEF activation was stronger in the triple-step and self-paced tasks
than during externally triggered saccades (visually guided or
memory-guided). This con®rms the clinical and experimental
evidence (Pierrot-Deseilligny et al., 1995; Hanes & Schall, 1996;
Heide & KoÈmpf, 1998) that the FEF predominantly controls
internally generated intentional saccades. Additionally it appears to
be involved in generating sequences of memory-guided saccades, at
least the right sFEF, that was not activated during self-paced
saccades, but during triple-step saccades, even in the intertask
comparisons (Table 5). A similar activation was reported by Petit
et al. (1996) in their PET study on prelearned saccadic sequences. It
could re¯ect the triggering of such sequences as well as the spatial
computations needed for their spatial accuracy, as the saccade-related
efference copy signal has been shown to be represented in FEF
neurons of macaques (Goldberg & Bruce, 1990; Umeno & Goldberg,
1997; Tian et al., 2000), not only in LIP neurons. However, according
to the results of lesion studies in rhesus monkeys (Schiller & Sandell,
1983) and human stroke patients (Heide et al., 1995), FEF lesions did
not impair the ability to compensate for a presaccadic eye displacement caused by either electrical stimulation of the superior colliculus
(in monkeys) or during the double-step task, leaving the spatial
accuracy of double-step saccades rather intact. Thus in contrast to the
PPC, the FEF is not essential for the spatial updating of saccade goals
by the use of extraretinal information (efference copy).
Ventrolateral premotor cortex
Outside the FEF, along the inferior portion of the precentral sulcus,
the ventrolateral PMC was activated during all tasks of this study,
with right hemispheric dominance. This has never been reported in
relation with saccades. PMC activation disappeared in the intertask
comparisons, like parietal foci in the SPL. Thus it might re¯ect
general saccade-related or mere attentional processes, as these foci
were also activated during covert attentional movements (Gitelman
et al., 1999; Nobre et al., 2000). Within the ventrolateral PMC of
monkeys, an oculomotor area has recently been identi®ed by
intracortical microstimulation (Fuji et al., 1998), evoking goaldirected saccades, in contrast to the FEF. We might have found its
human homologue. Together with the surrounding PMC, it seems to
coordinate saccadic commands with attentional, visuospatial and
skeletomotor information for goal-directed motor tasks, e.g. the
manipulation of complex objects (Binkofski et al., 1999). Thus the
PMC works on a higher level of sensorimotor integration than the
FEF, as part of a parieto-premotor network for visuomotor control,
which transforms extrinsic visuospatial information about target
location and movement trajectory into limb-centred motor commands
(Rizzolatti et al., 1997; Vallar et al., 1999). For skeletomotor
activities the dorsolateral PMC was shown to be involved in the
processing of cue-related motor behaviour (Halsband & Freund,
1990; Halsband et al., 1993).
Supplementary eye ®elds and cingulate cortex
Schlag & Schlag-Rey (1987) were the ®rst to describe the SEF and its
neuronal properties in nonhuman primates. In our fMRI study most
saccade tasks activated a premotor region at the dorsomedial wall of
the frontal lobe, close to the paracentral sulcus and superior to the
cingulate sulcus, which corresponds to the assumed location of the
human SEF (Petit et al., 1996; Luna et al., 1998; Grosbras et al.,
1999). Visually guided saccades activated the anterior portion of the
left SEF, that might be classi®ed as `preSEF'. It should be noted,
however, that up to now there are no data supporting the existence of
a separate preSEF in nonhuman primates. The human preSEF was
proposed to be located rostral to the anterior commissure line (ycoordinate > 0), the SEF-proper directly caudal to it (y < 0; Grosbras
et al., 1998). Only the triple-step task evoked a distinct peak of
activity in the region of the SEF-proper. As there was no SEF
activation during single memory-guided saccades, but only during
memory-guided sequences, this activity appears to be related to
visuomotor sequence learning. In line with other imaging studies on
this subject (Hikosaka et al., 1996; Kawashima et al., 1998; Sakai
et al., 1998) we found evidence for a functional dissociation. The
preSEF, like the preSMA in other studies, was activated whenever
new sequences were presented or had to be planned or memorized,
whereas the SEF-proper was activated only during the execution of
memorized sequences in the triple-step task. This interpretation is
consistent with single-neuron data in monkeys (Nakamura et al.,
1998) and with clinical studies (Gaymard et al., 1990; Heide et al.,
1995), where patients with focal lesions of the left SMA/SEF were
impaired in triggering memorized sequences of saccades.
ã 2001 Federation of European Neuroscience Societies, European Journal of Neuroscience, 13, 1177±1189
Functional MRI of triple-step saccades
The preSEF is also involved in triggering self-paced sequences of
saccades in darkness (Petit et al., 1993, 1996), together with the
adjacent caudal portion of the ACc. In our study, self-paced saccades
and triple-step saccades activated just the border region between the
preSEF and the ACc, centring around the cingulate sulcus, with right
hemispheric dominance. It could re¯ect the self-initiation of saccades.
Thus the right ACc appears to play an important role in the control of
intentional saccades and might correspond to a `cingulate eye ®eld'.
Lesions of this area impair the initiation and accuracy of internally
generated intentional saccades (Gaymard et al., 1998). Another area in
the right ACr, located anterior and ventral to the ACc, was activated
only when triple-step saccades were controlled vs. visually guided
saccades. This might re¯ect sustained attention and on-line monitoring
of performance, as the ACr is part of the cortical network for
visuospatial attention (Nobre et al., 1997, 2000; Carter et al., 1998).
In summary, the present fMRI study is the ®rst to investigate cortical
activation patterns during the performance of memorized triple-step
saccades. In contrast to previous PET or fMRI studies on saccades, the
combined use of several control conditions, either with identical visual
stimulation or with identical motor output (visually guided or single
memory-guided saccades), permitted some important conclusions on
saccadic and cognitive subfunctions of the triple-step task. Each of
these functions is represented in overlapping frontoparietal networks
of multiple cortical areas, the exact homology of which still requires
further research. For the ®rst time, an area in the right middle IPS was
identi®ed that was congruent with the common lesioned area in
parietal patients with dysmetric double-step saccades (Heide et al.,
1995) and appears to be essentially involved in spatial transformations
updating (remapping) spatial representations of saccade targets by the
use of efference copy signals in order to compensate for a previous eye
displacement, to achieve saccadic accuracy, and to maintain spatial
constancy. This area is a putative homologue of monkey area LIP. Also
frontal areas contribute to the processing of such saccadic sequences.
FEF activation appeared to re¯ect the execution of saccades in general,
ACc activation the execution of self-paced saccades, and SEF
activation the execution (possibly the timing and triggering) of
saccadic sequences. In contrast to most previous studies we found
some evidence for a functional dissociation of activity along the
rostrocaudal axis that may be divided in a SEF and a preSEF. Further,
we were the ®rst to identify saccade-related activity in the right
ventrolateral PMC that might integrate saccadic commands into the
parieto-premotor network for visuomotor control.
Abbreviations
ACc, caudal anterior cingulate gyrus; ACr, rostral anterior cingulate gyrus;
AIP, anterior intraparietal; BA, Brodmann area; BOLD, blood oxygen leveldependent signal changes; FEF, frontal eye ®eld: sFEF, superior portion; iFEF,
inferior portion; fMRI, functional magnetic resonance imaging; IPL, inferior
parietal lobule; IPS, intraparietal sulcus: mIPS, middle portion of the IPS;
pIPS, posterior portion; LIP, lateral intraparietal area of the macaque; MP/
Prec., medial posterior parietal cortex and dorsal precuneus; PET, positron
emission tomography; PFC, prefrontal cortex; PMC, premotor cortex; PPC,
posterior parietal cortex; SEF, supplementary eye ®eld; SMA, supplementary
motor area; SPL, superior parietal lobule; SPM, statistical parametric
mapping.
References
Andersen, R.A. (1995) Encoding of intention and spatial location in the
posterior parietal cortex. Cereb. Cortex, 5, 457±469.
Andersen, R.A., Snyder, L.H., Bradley, D.C. & Xing, J. (1997) Multimodal
representation of space in the posterior parietal cortex and its use in
planning movements. Annu. Rev. Neurosci., 20, 303±330.
1187
Anderson, T.J., Jenkins, I.H., Brooks, D.L., Hawken, M.B., Frackowiak, R.S.J.
& Kennard, C. (1994) Cortical control of saccades and ®xation in man. A
PET study. Brain, 117, 1073±1084.
Barash, B., Bracewell, R.M., Fogassi, L., Gnadt, J.W. & Andersen, R.A.
(1991) Saccade-related activity in the lateral intraparietal area. J.
Neurophysiol., 66, 1095±1124.
Berman, R.A., Colby, C.L., Genovese, C.R., Voyvodic, J.T., Luna, B.,
Thulborn, K.R. & Sweeney, J.A. (1999) Cortical networks subserving
pursuit and saccadic eye movements in humans: An fMRI study. Hum.
Brain Mapp., 8, 209±225.
Binkofski, F., Amunts, K., Stephan, K.M., Posse, S., Schormann, T., Freund,
H.-J., Zilles, K. & Seitz, R.J. (2000) Broca's region subserves imagery of
motion: a combined cytoarchitectonic and fMRI study. Hum. Brain
Mapping, 11, 273±285.
Binkofski, F., Buccino, G., Posse, S., Seitz, R.J., Rizzolatti, G. & Freund, H.-J.
(1999) A fronto-parietal circuit for object manipulation in man: evidence
from an fMRI study. Eur. J. Neurosci., 11, 3276±3286.
Binkofski, F., Dohle, C., Posse, S., Stephan, K.M., Hefter, H., Seitz, R.J. &
Freund, H.-J. (1998) Human anterior intraparietal area subserves
prehension. A combined lesion and functional MRI activation study.
Neurology, 50, 1253±1259.
Boch, R.A. & Goldberg, M.E. (1989) Participation of prefrontal neurons in the
preparation of visually guided eye movements in the rhesus monkey. J.
Neurophysiol., 61, 1064±1084.
Brandt, S.A., Ploner, C.J., Meyer, B.-U., Leistner, S. & Villringer, A. (1998)
Effects of repetitive transcranial magnetic stimulation over dorsolateral
prefrontal and posterior parietal cortex on memory-guided saccades. Exp.
Brain Res., 118, 197±204.
Bruce, C.J. & Goldberg, M.E. (1985) Primate frontal eye ®elds. I. Single
neurons discharging before saccades. J. Neurophysiol., 53, 603±635.
Carter, C.S., Braver, T.S., Barch, D.M., Botvinick, M.M., Noll, D. & Cohen,
J.D. (1998) Anterior cingulate cortex, error detection, and the online
monitoring of performance. Science, 280, 747±749.
Colby, C.L., Duhamel, J.-R. & Goldberg, M.E. (1996) Visual, presaccadic,
and cognitive activation of single neurons in monkey lateral intraparietal
area. J. Neurophysiol., 76, 2841±2852.
Colby, C.L. & Goldberg, M.E. (1999) Space and attention in parietal cortex.
Annu. Rev. Neurosci., 22, 319±349.
Corbetta, M., Akbudak, E., Conturo, T.E., Snyder, A.Z., Ollinger, J.M., Drury,
H.A., Linenweber, M.R., Petersen, S.E., Raichle, M.E., van Essen, D.C. &
Shulman, G.L. (1998) A common network of functional areas for attention
and eye movements. Neuron, 21, 761±773.
Dassonville, P., Schlag, J. & Schlag-Rey, M. (1992) Oculomotor localization
relies on a damped representation of saccadic eye displacement in human
and nonhuman primates. Visual Neurosci., 9, 261±269.
Dejardin, S., Dubois, S., Bodart, J.M., Schiltz, C., Delinte, A., Michel, C.,
Roucoux, A. & Crommelinck, M. (1998) PET study of human voluntary
saccadic eye movements in darkness: effect of task repetition on the
activation pattern. Eur. J. Neurosci., 10, 2328±2336.
Dieterich, M., Bucher, S.F., Seelos, K.C. & Brandt, Th (1998) Horizontal or
vertical optokinetic stimulation activates visual motion-sensitive, ocular
motor and vestibular cortex areas with right hemispheric dominance. An
fMRI study. Brain, 121, 1479±1495.
Dominey, P.F. & Arbib, M.A. (1992) A cortico-subcortical model for
generation of spatially accurate sequential saccades. Cerebral Cortex, 2,
153±175.
Dominey, P.F., Schlag, J., Schlag-Rey, M. & Arbib, M.A. (1997) Colliding
saccades evoked by frontal eye ®eld stimulation: artifact or evidence for an
oculomotor compensatory mechanism underlying double-step saccades?
Biol. Cybern., 76, 41±52.
Droulez, J. & Berthoz, A. (1991) A neural network model of sensoritopic maps
with predictive short-term memory properties. Proc. Natl. Acad. Sci. USA,
88, 9653±9657.
Duhamel, J.-R., Colby, C.L. & Goldberg, M.E. (1992a) The updating of the
representation of visual space in parietal cortex by intended eye movements.
Science, 255, 90±92.
Duhamel, J.-R., Goldberg, M.E., Fitzgibbon, E.J., Sirigu, A. & Grafman, J.
(1992b) Saccadic dysmetria in a patient with a right frontoparietal lesion.
Brain, 115, 1387±1402.
Fox, P.T., Fox, J.M., Raichle, M.E. & Burde, R.M. (1985) The role of cerebral
cortex in the generation of voluntary saccades: a positron emission
tomographic study. J. Neurophysiol., 54, 348±369.
Friston, K.J. (1995) Commentary and opinion: II. Statistical parametric
mapping: ontology and current issues. J. Cereb. Blood Flow Metab., 15,
361±370.
ã 2001 Federation of European Neuroscience Societies, European Journal of Neuroscience, 13, 1177±1189
1188 W. Heide et al.
Friston, K.J., Ashburner, J., Poline, J.B., Frith, C.D., Heather, J.D. &
Frackowiak, R.S.J. (1997) Spatial realignment and normalization of images.
Hum. Brain Mapp., 2, 165±189.
Friston, K.J., Holmes, A.P., Poline, J.-B., Grasby, P.J., Williams, S.C.R.,
Frackowiak, R.S.J. & Turner, R. (1995a) Analysis of fMRI time-series
revisited. Neuroimage, 2, 5±53.
Friston, K.J., Holmes, A.P., Worsley, K.J., Poline, J.-B., Frith, C.D. &
Frackowiak, R.S.J. (1995b) Statistical parametric maps in functional
imaging: a general linear approach. Hum. Brain Mapp., 2, 189±210.
Friston, K.J., Jezzard, P. & Turner, R. (1994a) Analysis of functional MRI
time-series. Hum. Brain Mapp., 1, 153±171.
Friston, K.J., Worsley, K.J., Frackowiak, R.S.J., Maziotta, J.C. & Evans, A.C.
(1994b) Assessing the signi®cance of focal activation using their spatial
extent. Hum. Brain Mapp., 1, 210±220.
Fuji, N., Mushiake, H. & Tanji, J. (1998) An oculomotor representation area
within the ventral premotor cortex. Proc. Natl. Acad. Sci. USA, 95, 12034±
12037.
Funahashi, S., Bruce, C.J. & Goldman-Rakic, P.S. (1991) Neuronal activity
related to saccadic eye movements in the monkey's dorsolateral prefrontal
cortex. J. Neurophysiol., 65, 1464±1484.
Funahashi, S., Bruce, C.J. & Goldman-Rakic, P.S. (1993) Dorsolateral
prefrontal lesions and oculomotor delayed-response performance: evidence
for `mnemonic' scotomas. J. Neurosci., 13, 1479±1497.
Galletti, C., Fattori, P., Kutz, D.F. & Battaglini, P.P. (1997) Arm movementrelated neurons in the visual area V6A of the macaque superior parietal
lobe. Eur. J. Neurosci., 9, 110±113.
Gaymard, B., Rivaud, S., Cassarini, J.F., Dubard, T., Rancurel, G., Agid, Y. &
Pierrot-Deseilligny, C. (1998) Effects of anterior cingulate cortex lesions on
ocular saccades in humans. Exp. Brain Res., 120, 173±183.
Gaymard, B., Rivaud, S. & Pierrot-Deseilligny, C. (1990) Impairment of
sequences of memory-guided saccades after supplementary motor area
lesions. Ann. Neurol., 28, 622±626.
Gitelman, D.R., Nobre, A.C., Parrish, T.B., LaBar, K.S., Kim, Y.-H., Meyer,
J.R. & Mesulam, M.-M. (1999) A large-scale distributed network for covert
spatial attention. Brain, 122, 1093±1106.
Goebel, R., Linden, D.E.J., Lanfermann, H., Zanella, F.E. & Singer, W. (1998)
Functional imaging of mirror and inverse reading separate coactivated
networks for oculomotion and spatial transformations. Neuroreport, 9, 713±
719.
Goldberg, M.E. & Bruce, C.J. (1990) Primate frontal eye ®elds. III.
Maintenance of a spatially accurate saccade signal. J. Neurophysiol., 64,
489±508.
Goldberg, M.E., Colby, C.L. & Duhamel, J.-R. (1990) Representation of
visuomotor space in the parietal lobe of the monkey. Cold Spring Harbor
Symp., 55, 729±739.
Gottlieb, J. & Goldberg, M.E. (1999) Activity of neurons in the lateral
intraparietal area of the monkey during an antisaccade task. Nature
Neurosci., 2, 906±912.
Grosbras, M.-H., Lobel, E., leBihan, D., Berthoz, A. & Leonards, U. (1998)
Evidence for a pre-SEF in humans: a fMRI study. Neuroimage, 7, S988.
Grosbras, M.-H., Lobel, E., van de Moortele, P.-F., leBihan, D. & Berthoz, A.
(1999) An anatomical landmark for the supplementary eye ®eld in human
revealed with functional magnetic resonance imaging. Cereb. Cortex, 9,
705±711.
Hallett, P.E. & Lightstone, A.D. (1976) Saccadic eye movements to ¯ashed
targets. Vision Res., 16, 107±114.
Halsband, U. & Freund, H.-J. (1990) Premotor cortex and conditional motor
learning in man. Brain, 113, 207±222.
Halsband, U., Ito, N., Tanji, J. & Freund, H.-J. (1993) The role of premotor
cortex and the supplementary motor area in the temporal control of
movement in man. Brain, 116, 243±266.
Hanes, D.P. & Schall, J.D. (1996) Neural control of voluntary movement
initiation. Science, 274, 427±430.
Harris, I.M., Egan, G.F., Sonkkila, C., Tochon-Danguy, H.J., Paxinos, G. &
Watson, J.D.G. (2000) Selective right parietal lobe activation during mental
rotation. A parametric PET study. Brain, 123, 65±73.
Heide, W., Binkofski, F., Posse, S., Seitz, R.J., KoÈmpf, D. & Freund, H.-J.
(1997) Functional anatomy of memory-guided saccade sequences in human
cerebral cortex. Soc. Neurosci. Abstr., 23, 2222.
Heide, W., Blankenburg, M., Zimmermann, E. & KoÈmpf, D. (1995) Cortical
control of double-step saccades ± implications for spatial orientation. Ann.
Neurol., 38, 739±748.
Heide, W. & KoÈmpf, D. (1997) Speci®c parietal lobe contribution to spatial
constancy across saccades. In Thier, P., Karnath, H.-O. (eds), Parietal Lobe
Contributions to Orientation in 3D Space. Springer Verlag, Heidelberg, pp.
149±172.
Heide, W. & KoÈmpf, D. (1998) Combined de®cits of saccades and visuospatial orientation after cortical lesions. Exp. Brain Res., 123, 164±171.
Hikosaka, O., Sakai, K., Miyauchi, S., Takino, R., Sasaki, Y. & PuÈtz, B.
(1996) Activation of human presupplementary motor area in learning of
sequential procedures: a functional MRI study. J. Neurophysiol., 76, 617±
621.
Honda, H. (1989) Perceptual localization of visual stimuli ¯ashed during
saccades. Percept. Psychophys., 45, 162±174.
Kastner, S., Pinsk, M.A., de Weerd, P., Desimone, R. & Ungerleider, L.G.
(1999) Increased activity in human visual cortex during directed attention in
the absence of visual stimulation. Neuron, 22, 751±761.
Kawashima, R., Naitoh, E., Matsumara, M., Ito, H., Ono, S., Sato, K., Goto,
H., Koyama, M., Inoue, K., Yoshioka, S. & Fukuda, H. (1996) Topographic
representation in human intraparietal sulcus of reaching and saccade.
Neuroreport, 7, 1253±1256.
Kawashima, R., Tanji, J., Okada, K., Sugiura, M., Sato, K., Kinomura, S.,
Inoue, K., Ogawa, A. & Fukuda, H. (1998) Oculomotor sequence learning: a
positron emission tomography study. Exp. Brain Res., 122, 1±8.
Law, I., Svarer, C., Holm, S. & Paulson, O.B. (1997) The activation pattern in
normal humans during suppression, imagination and performance of
saccadic eye movements. Acta Physiol. Scand., 161, 419±434.
Law, I., Svarer, C., Rostrup, E. & Paulson, O.B. (1998) Parieto-occipital
cortex activation during self-generated eye movements in the dark. Brain,
121, 2189±2200.
Luna, B., Thulborn, K.R., Strojwas, M.H., McCurtain, B.J., Berman, R.A.,
Genovese, C.R. & Sweeney, J.A. (1998) Dorsal regions subserving visually
guided saccades in humans: An fMRI study. Cereb. Cortex, 8, 40±47.
Mays, L.E. & Sparks, D.L. (1980) Dissociation of visual and saccade-related
responses in superior colliculus neurons. J. Neurophysiol., 43, 207±232.
MuÈri, R.M., Iba-Zizen, M.T., Derosier, C., Cabanis, E.A. & PierrotDeseilligny, C. (1996) Location of the human posterior eye ®eld with
functional magnetic resonance imaging. J. Neurol. Neurosurg. Psychiatry,
60, 445±448.
Nakamura, K., Chung, H.H., Graziano, M.S.A. & Gross, C.G. (1999) Dynamic
representation of eye position in the parieto-occipital sulcus. J.
Neurophysiol., 81, 2374±2385.
Nakamura, K., Sakai, K. & Hikosaka, O. (1998) Neuronal activity in medial
frontal cortex during learning of sequential procedures. J. Neurophysiol.,
80, 2671±2687.
Nobre, A.C., Gitelman, D.R., Dias, E.C. & Mesulam, M.M. (2000) Covert
visual spatial orienting and saccades: overlapping neural systems.
Neuroimage, 11, 210±216.
Nobre, A.C., Sebestyen, G.N., Gitelman, D.R., Mesulam, M.M., Frackowiak,
R.S.J. & Frith, C.D. (1997) Functional localization of the system for
visuospatial attention using positron emission tomography. Brain, 120, 515±
533.
O'Sullivan, E.P., Jenkins, I.H., Henderson, L., Kennard, C. & Brooks, D.J.
(1995) The functional anatomy of remembered saccades: a PET study.
Neuroreport, 6, 2141±2144.
Paus, T. (1996) Location and function of the human frontal eye-®eld: a
selective review. Neuropsychologia, 34, 475±483.
Petit, L., Clark, V.P., Ingeholm, J. & Haxby, J.V. (1997) Dissociation of
saccade-related and pursuit-related activation in human frontal eye ®elds as
revealed by fMRI. J. Neurophysiol., 77, 3386±3390.
Petit, L. & Haxby, J.V. (1999) Functional anatomy of pursuit eye movements
in humans as revealed by fMRI. J. Neurophysiol., 81, 463±471.
Petit, L., Orssaud, C., Tzourio, N., Crivello, F., Berthoz, A. & Mazoyer, B.
(1996) Functional anatomy of a prelearned sequence of horizontal saccades
in humans. J. Neurosci., 16, 3714±3726.
Petit, L., Tzourio, N., Orssaud, C., Salamon, G., Mazoyer, B. & Berthoz, A.
(1993) PET study of voluntary saccadic eye movements in humans: basal
ganglia-thalamocortical system and cingulate cortex involvement. J.
Neurophysiol., 69, 1009±1017.
Pierrot-Deseilligny, C., Rivaud, S., Gaymard, B., MuÈri, R. & Vermersch, A.-I.
(1995) Cortical control of saccades. Ann. Neurol., 37, 557±567.
Poline, J.B., Worsley, K.J., Holmes, A.P., Frackowiak, R.S.J. & Friston, K.J.
(1995) Estimating smoothness in statistical parametric maps: variability of
p-values. J. Comput. Assist. Tomogr., 19, 788±796.
Rizzolatti, G., Fogassi, L. & Gallese, V. (1997) Parietal cortex: from sight to
action. Curr. Opin. Neurobiol., 7, 562±567.
Roland, P.E. & Zilles, K. (1996) Functions and structures of the motor areas in
humans. Curr. Opin. Neurobiol., 6, 773±781.
Sakai, K., Hikosaka, O., Miyauchi, S., Takino, R., Sasaki, Y. & PuÈtz, B.
ã 2001 Federation of European Neuroscience Societies, European Journal of Neuroscience, 13, 1177±1189
Functional MRI of triple-step saccades
(1998) Transition of brain activation from frontal to parietal areas in
visuomotor sequence learning. J. Neurosci., 18, 1827±1840.
Schiller, P.H. & Sandell, J.H. (1983) Interactions between visually and
electrically elicited saccades before and after superior colliculus and frontal
eye ®eld ablations in the rheusus monkey. Exp. Brain Res., 49, 381±392.
Schlag, J. & Schlag-Rey, M. (1987) Evidence for a supplementary eye ®eld. J.
Neurophysiol., 57, 179±200.
Schlag, J. & Schlag-Rey, M. (1995) Illusory localization of stimuli ¯ashed in
the dark before saccades. Vision Res., 35, 2347±2357.
Seitz, R.J., Canavan, A.G.M., Herzog, H., Tellmann, L., Knorr, U., Huang, Y.
& HoÈmberg, V. (1997) Representation of graphomotor trajectories in the
human parietal cortex: evidence for controlled processing and automatic
performance. Eur. J. Neurosci., 9, 378±389.
Sereno, A.B. & Maunsell, J.H. (1998) Shape selectivity in primate lateral
intraparietal cortex. Nature, 395, 500±503.
Steinmetz, H., FuÈrst, G. & Freund, H.-J. (1990) Variation of perisylvian and
calcarine anatomical landmarks within stereotactic proportional
coordinates. Am. J. Neuroradiol., 11, 1123±1130.
Stephan, K.M., SchuÈller, M., Hoȯich, P., Knorr, U., Binkofski, F. & Seitz, R.J.
(1997) Normalization into `Talairach space': variability of reference
coordinates [abstract]. Neuroimage, 5, S416.
1189
Sweeney, J.A., Mintun, M.A., Kwee, S., Wiseman, M.B., Brown, D.L.,
Rosenberg, D.R. & Carl, J.R. (1996) Positron emission tomography study of
voluntary saccadic eye movements and spatial working memory. J.
Neurophysiol., 75, 454±468.
Talairach, J. & Tournoux, P. (1988) Co-Planar Stereotactic Atlas of the
Human Brain. Thieme-Verlag, Stuttgart.
Thier, P. & Andersen, R.A. (1998) Electrical microstimulation distinguishes
distinct saccade-related areas in the posterior parietal cortex. J.
Neurophysiol., 80, 1713±1735.
Tian, J., Schlag, J. & Schlag-Rey, M. (2000) Testing quasi-visual neurons in
the monkey's frontal eye ®eld with the triple-step paradigm. Exp. Brain
Res., 130, 433±440.
Umeno, M.M. & Goldberg, M.E. (1997) Spatial processing in the monkey
frontal eye ®eld. I. Predictive visual responses. J. Neurophysiol., 78, 1373±
1383.
Vallar, G., Lobel, E., Galati, G., Berthoz, A., Pizzamiglio, L. & Le Bihan, D.
(1999) A fronto-parietal system for computing the egocentric spatial frame
of reference in humans. Exp. Brain Res., 124, 281±286.
Worsley, K.J. & Friston, K.J. (1995) Analysis of fMRI time-series revisited
again. Neuroimage, 3, 173±181.
ã 2001 Federation of European Neuroscience Societies, European Journal of Neuroscience, 13, 1177±1189
Download