вход по аккаунту



код для вставкиСкачать
Human Brain Mapping 5:287–292(1997)
Combining Steady-State Visual Evoked Potentials
and fMRI to Localize Brain Activity
During Selective Attention
Steven A. Hillyard,1* Hermann Hinrichs,2 Claus Tempelmann,2
Stephen T. Morgan,1 Jonathan C. Hansen,1 Henning Scheich,3
and Hans-Jochen Heinze2
of Neurosciences, University of California, San Diego, La Jolla, California
of Clinical Neurophysiology, Otto-V-Guericke University, Magdeburg, Germany
3Federal Institute of Neurobiology, Magdeburg, Germany
Abstract: Brain activity was studied with fMRI and steady-state visual evoked potentials (SSVEPs) in
separate sessions as subjects attended to letter sequences in either the right or left visual field. The two
letter sequences were superimposed on small square backgrounds that flickered at 8.6 and 12.0 Hz,
respectively, and elicited SSVEPs at the flicker frequencies. The amplitude of the frequency-coded SSVEP
elicited by either of the flickering backgrounds was enlarged when attention was focused upon the letter
sequence at the same location. Source analysis of the SSVEP waveforms localized the attentional
modulation to ventral and lateral extrastriate visual cortex, which corresponded to zones of activation
observed with fMRI. Hum. Brain Mapping 5:287–292, 1997. r 1997 Wiley-Liss, Inc.
Key words: event-related potentials; neuroimaging; SSVEP; spatial attention; flicker; visual cortex
The steady-state visual evoked potential (SSVEP) is
a continuous, oscillatory pattern of brain wave activity
elicited in response to a regularly repeating visual
stimulus such as a flickering light source. The SSVEP
may be recorded from the scalp as a near-sinusoidal
Contract grant sponsor: NIMH; Contract grant number: MH-25594;
Contract grant sponsor: NIH; Contract grant number: NS 17778;
Contract grant sponsor: ONR; Contract grant number: N00014-93-10942; Contract grant sponsor: DFG; Contract grant numbers: 1531/
3-1 and INK 15/A1 TP A1; Contract grant sponsor: BMPF Verbundprojeckt; Contract grant number: 07 NBL 04/01 ZZ 9505 TP 11.
*Correspondence to: Dr. Steven A. Hillyard, Department of Neurosciences, 0608, University of California, San Diego, La Jolla, CA 92093.
Received for publication 28 February 1997; accepted 13 March 1997
r 1997 Wiley-Liss, Inc.
waveform having the same fundamental frequency as
the driving stimulus [Regan, 1989; Silberstein, 1995].
Few studies have attempted to identify the neural
generators responsible for the surface-recorded SSVEP,
although dipolar sources in visual cortex have been
inferred on the basis of both electrophysiological [Van
Dijk and Spekreijse 1990] and neuromagnetic [Mueller
et al., in press] recordings. Both the amplitude and
scalp topography of the SSVEP are reportedly affected
by cognitive manipulations including selective attention and memory search [Wilson and O’Donnell, 1986;
Silberstein et al., 1990, 1995; Morgan et al., 1996].
The SSVEP is readily quantifiable in the frequency
domain and can be used to study selective attention to
stimuli that are continuously visible (flickering) rather
than flashed at irregular intervals. Morgan et al. [1996]
found that the SSVEP was strongly modulated by
Hillyard et al. r
spatial selective attention in a task in which subjects
were cued to attend to a letter/number sequence in
one visual field and to ignore a similar, concurrent
sequence in the opposite field. The letter/number
sequences in the two fields were superimposed upon
small background squares flickering at 8.6 Hz in one
field and at 12 Hz in the other. It was found that the
SSVEP amplitude was substantially enlarged to the
background flicker at the attended relative to the
unattended location. This result suggests that the
SSVEP amplitude modulations reflect the enhancement of evoked neural activity in the visual pathways
to stimuli that fall within the spotlight of spatial
The present study combined fMRI with SSVEP
recordings obtained in the same subjects while performing the same spatial attention task used by Morgan et
al. [1996]. The aim was to use fMRI to help localize the
cortical areas where steady-state neural activity is
modulated by attention and thus to gain information
about the neural generators of the attention-related
changes in the SSVEP.
Five normal young adults (three female, two male)
served as paid volunteer subjects. Their ages ranged
from 19 to 34, and all had normal vision.
intertrial interval. The subject’s task was to press a
button to targets in the attended field sequence; the
sequence in the opposite field was to be ignored. In
separate blocks of ‘‘passive’’ trials, the subject was
instructed to maintain fixation but to ignore all the
stimuli, which were identical with those presented
during the ‘‘attend’’ blocks.
Six different types of trials were presented as defined by the orthogonal factors of attention (attendleft, attend-right, and passive) and background flicker
location (12 Hz left/8.6 Hz right or 12 Hz right/8.6 Hz
left). Each subject received 30–40 trials of each type in
counterbalanced order during the SSVEP recording
SSVEP recording and analysis
Brain electrical activity was recorded from 30 scalp
electrodes mounted in an elastic cap all referred to
right mastoid. The approximate positions of the posterior electrodes are indicated in Figure 1A and 1B. Eye
movements were monitored via horizontal and vertical electrooculogram recordings. Recording bandpass
was 0.01–64 Hz.
Separate time-domain averages of the SSVEPs were
obtained for each combination of flicker frequency (8.6
or 12 Hz), stimulus location (left or right visual field),
and direction of attention (left, right, passive). Fourier
transforms of each averaged SSVEP waveform were
also calculated, with response amplitude quantified as
Experimental design
Each subject participated in two separate sessions in
which the same visual attention task was performed
while either SSVEP recording or fMRI scanning was
carried out. During testing the subject maintained
central fixation while randomized sequences of alphanumeric characters were presented concurrently in the
left and right visual fields, 5.7 degrees lateral to
fixation. Each sequence consisted of equiprobable
(P 5 .08) black characters (the letters A through K and
the target number 5, presented in random order at a
rate of 5 characters/sec) superimposed on a filled
white square that was flickered on and off at a rate of
12 Hz in one visual field and 8.6 Hz in the other (see
Fig. 1A).
On each trial a left or right pointing central arrow
indicated the visual field to be attended. The concurrent letter/number sequences in the left and right
fields began 2 seconds after the arrow, together with
the repetitive flickering of the background squares;
these stimuli continued for 10 sec, followed by an
Figure 1.
A: Visual display as seen by the subject for condition with 12-Hz
flicker at the left stimulus location and 8.6 Hz at the right location.
SSVEP waveforms shown are the time domain averages for subject
05 time-locked to the 8.6-Hz flicker, with averages for attend-right
and passive conditions superimposed. Head map shows the voltage
topography of the 8.6-Hz attention effect (i.e., attend-left minus
passive amplitude). Waveform calibration 5 1.0 µV. B: Grand
average topography of SSVEP attention effects over all five subjects.
Amplitude topographies were averaged over the 8.6-Hz and 12-Hz
SSVEPs for each stimulus position and then subtracted for attendleft minus passive and attend-right minus passive conditions. C:
Grand average occipitotemporal fMRI activations and superimposed best-fit dipoles (circles with pointers) for grand average
SSVEP attention effects. A-P Talaraich coordinate of section is 278
mm. The raw fMRI signal strength values for each attention
condition were averaged across all replications of that condition
and then across all subjects. Hotter colors indicate stronger signals
in the indicated attend minus passive subtractions. The dipoles
shown were calculated using BESA on the basis of the SSVEP
waveform distributions averaged over all subjects and over both
stimulus frequencies for the indicated attend minus passive differences. Dipole orientation is indicated by the pointer direction.
Figure 1.
Hillyard et al. r
the square root of the sum of the squares of the sine
and cosine coefficients.
To estimate the locations of the SSVEP generators
modulated by attention, dipole modeling was carried
out using BESA (Scherg 1990). Multiple dipoles were
fit concurrently to the scalp distribution of the sine and
cosine coefficients of the grand average SSVEP difference waveforms (attend-left minus passive and attendright minus passive), separately for each frequency.
The positions of the best-fit dipoles were related to
brain anatomy by co-registering the spherical BESA
coordinate system with anatomical magnetic resonance imaging scans and with the Talaraich and
Tourneaux [1988] atlas.
fMRI analysis
In a separate session, subjects attended to the same
stimulus sequences (viewed with a mirror) while fMRI
scanning was performed using a Brucker Biospec
3.0T/60 cm system with a standard birdcage headcoil.
Functional images were collected using a gradient
echo sequence with TR/TE/flip angle values of 225
msec/40 msec/10 deg. Experimental sessions consisted of alternating 36-sec periods of attend-left/
passive conditions or attend-right/passive conditions,
with each condition repeated 10 times per session.
During each 720-sec session, five adjacent coronal
slices (thickness 8 mm, in-plane resolution 4 3 4 mm)
through the occipital and posterior temporal and
parietal lobes were scanned at intervals of 9 sec (i.e.,
four repetitions of each slice for each attention condition). Significance of attention effects was assessed by
determining the linear correlation between signal
strength at each pixel and the alternating block design
of the experiment [Bandettini et al., 1992]. The resulting P values were subjected to two-dimensional 3 3 3
median filtering to reduce the likelihood of single,
isolated voxels reaching significance by chance. Significant activations for each subject were superimposed
on anatomical fMRI scans, and their positions were
determined in coordinates of the Talaraich and Tourneaux atlas.
Target detection performance
Targets were detected with an average hit rate of
89.4% and a mean reaction time of 295 msec; falsealarm rates averaged 2.1%. There were no visual field
differences in performance.
SSVEP modulation by attention
The SSVEP amplitudes over the posterior scalp were
significantly modulated by attention, being considerably larger in response to the flickering backgrounds
when their locations were attended than during the
passive control conditions (Fig. 1A). Averaged over
both the 8.6- and 12-Hz frequencies and the left and
right stimulus positions, the mean SSVEP amplitude
increased in attend versus passive conditions by 71%
(t 5 4.17, P , .0001) over posterior scalp sites (average
of sites 01/02, TO1/2, IN3/4). This amplitude enhancement tended to be more pronounced over the hemisphere contralateral to the eliciting stimulus (74%,
t 5 3.57, P , .002) as compared to the ipsilateral hemisphere (68%, t 5 2.49, P , .03).
The scalp distribution of the SSVEP attention effect
averaged over all subjects (and both frequencies) is
shown in Figure 1B. When the left stimuli were
attended the SSVEP enhancement was strongly contralateral, while attention to the right produced a more
bilaterally distributed amplitude increase.
fMRI activation during attention
Significant activations for both attend-left and attendright versus passive comparisons, were clustered in
two extrastriate visual cortical areas: (1) in the fusiform/
inferior occipital gyri, Brodmann’s areas 18 and 19,
with Talaraich coordinates ranging between 30 to 40
(M-L), 270 to 280 (A-P), and 25 to 210 (S-I); and (2) in
the medial occipital/inferior-medial temporal gyri,
Brodmann’s areas 19 and 37, with Talaraich coordinates ranging between 35 to 45 (M-L), 268 to 275
(A-P), and 15 to 115 (S-I). Both of these zones were
activated concurrently in three of the five subjects; the
other two subjects showed activation in only one or the
other of the areas.
These extrastriate activations were generally largest
in the hemisphere contralateral to the attended visual
field, although two subjects also showed substantial
ipsilateral activations, particularly during the attendright condition. No significant activations were found
in the primary visual cortex (area 17) in any of the
attend-passive comparisons.
To give an overview of these attention effects, the
fMRI activation patterns averaged over the five subjects are shown in Figure 1C, superimposed on a
coronal slice through the occipital/temporal cortex.
Both the inferior and lateral zones of activation can be
seen, along with secondary foci in the parietal cortex
(significant in two subjects). As with the SSVEPs, a
SSVEP and fMRI During Visual Attention r
more bilateral activation was seen for the attend-right
Also shown in Figure 1C are the posterior dipole
pairs that were fit using BESA to the grand average
scalp distributions of the SSVEP attention effects.
Separate dipole fits were carried out for the attend
minus passive SSVEPs to the 8.6-Hz and 12-Hz stimuli
in each visual field. For both frequencies a similar pair
of occipital dipoles was obtained, and their coordinates and orientations were averaged together to
obtain the dipoles shown in Figure 1C. These occipital
dipoles were positioned at the same A-P levels as the
maximal fMRI activations, but were situated more
medially by 10–18 mm. Such a medial dipole displacement would be expected because the localization
algorithm attempted to account for a broad arc of more
lateral generators with a single dipole. Additional
dipoles in parietal cortex (not shown in Fig. 1C) were
required to bring the residual variance of the dipole fits
down below 5%, but they only accounted for a small
proportion of the distributional variance.
The present study confirms the finding of Morgan et
al. [1996] that the SSVEP to a flickering visual stimulus
is increased in amplitude when attention is focused
upon its location. Since the 8.6-Hz and 12-Hz flickering
backgrounds were not relevant to the target detection
task, the SSVEP modulation supports the hypothesis
that evoked neuronal responses to all stimuli falling
within the spotlight of attention are facilitated, regardless of whether or not they are task-relevant. The
SSVEP to such irrelevant background flickers can thus
serve as a ‘‘probe’’ to monitor spatial attention to any
type of superimposed stimulus or array of stimuli.
The fMRI data obtained in the same attention task
provide converging evidence about the visual cortical
areas in which this attentional facilitation is carried
out. Two specific zones of extrastriate visual cortex
were found to be activated during the focusedattention conditions, one in the fusiform/inferior occipital area and the other in more lateral occipito-temporal
cortex. Dipole modeling of the grand average SSVEP
scalp distributions indicated that the attentional modulation of the SSVEP could be accounted for by dipolar
sources in the same general extrastriate regions. While
a more definitive picture of the SSVEP generators will
require comparisons of SSVEP source localizations and
patterns of fMRI activation in individual subjects, the
present preliminary analyses are consistent with the
proposal that these task-related modulations of SSVEP/
fMRI both reflect the same patterns of facilitated
neuronal processing of attended visual inputs in ventral-lateral extrastriate cortex.
The present results are consistent with those of
previous studies of transient ERPs in spatial attention
tasks [reviewed in Hillyard et al., 1995]. Attentionsensitive transient evoked components have been localized to the same ventral and ventral-lateral extrastriate
cortical areas as the present steady-state modulations
by means of source modeling [Mangun et al., 1993;
Gomez et al., 1994; Clark and Hillyard, 1996] and
positron emission tomography imaging [Heinze et al.,
1994]. Also consistent with the present results, none of
these experiments with transient stimuli observed any
attentional modulation of evoked activity in primary
visual cortex.
Taken together, the evidence from these studies of
steady-state and transient stimulation support a model
of spatial attention in which visual signals arising from
attended locations are facilitated at the level of extrastriate cortical areas but not in the primary cortex itself.
These ventral and lateral cortical regions may include
the human homologs of monkey visual areas V4 and
MT [Sereno et al., 1995] and perhaps adjacent areas as
well. Further studies are needed to delineate the exact
sites along the visual pathways at which attentional
control is exerted for different classes of stimuli and
This work was supported by grants from NIMH
(MH-25594), NIH (NS 17778), and ONR (N00014-93-10942) to S.A.H. and grants from the DFG (1531/3-1 and
INK 15/A1 TP A1) and BMBF Verbundprojeckt (07
NBL 04/01 ZZ 9505 TP 11) to H.J.H.
Bandettini PA, Wong EC, Hinks RS, Tikofsky RS, Hyde JS (1992):
Time course EPI of human brain function during task activation.
Magn Reson Med 25:390–397.
Clark VP, Hillyard SA (1996): Spatial selective attention affects early
extrastriate but not striate components of the visual evoked
potential. J Cognitive Neurosci 8:387–402.
Gomez CM, Clark VP, Luck SJ, Fan S, Hillyard SA (1994): Sources of
attention-sensitive visual event-related potentials. Brain Topogr
Heinze HJ, Luck SJ, Muente TF, Goes A, Mangun GR, Hillyard SA
(1994): Attention to adjacent and separate positions in space: an
electrophysiological analysis. Perception Psychophysics, 56:42–
Hillyard SA, Mangun GR, Woldorff MG, Luck SJ (1995): Neural
systems mediating selective attention. In: Gazzaniga MS (ed):
The Cognitive Neurosciences. Cambridge, MA: MIT Press, pp
Hillyard et al. r
Mangun GR, Hillyard SA, Luck SJ (1993): Electrocortical substrates
of visual selective attention. In: Meyer D, Kornblum S (eds):
Attention and Performance XIV. Cambridge, MA: MIT Press, pp
Morgan ST, Hansen JC, Hillyard SA (1996): Selective attention to
stimulus location modulates the steady-state visual evoked
potential. Proc Nat Acad Sci USA 93:4770–4774.
Mueller MM, Teder W, Hillyard SA (in press): Magnetoencephalographic recording of steady-state visual evoked cortical activity.
Brain Topography.
Regan D (1989): Human Brain Electrophysiology: Evoked Potentials
and Evoked Magnetic Fields in Science and Medicine. New York:
Elsevier Publishers.
Scherg M (1990): Fundamentals of dipole source potential analysis.
In: Grandori F, Hoke M, Romani GL (eds): Auditory Evoked
Magnetic Fields and Potentials. Advanced Audiology. Basel:
Karger, Vol 6, pp 40–69.
Sereno MI, Dale AM, Reppas JB, Kwong KK, Belliveau JW, Brady TJ,
Rosen BR, Tootell RBH (1995): Borders of multiple visual areas in
humans revealed by functional magnetic resonance imaging.
Science 889–893.
Silberstein RB (1995): Steady-state visually evoked potentials, brain
resonances, and cognitive processes. In: Nunez PL (ed): Neocortical Dynamics and Human EEG Rhythms. Oxford: Oxford University Press, pp 272–303.
Silberstein RB, Schier MA, Pipingas A, Ciorciari J, Wood SR,
Simpson DG (1990): Steady-state visually evoked potential topography associated with a visual vigilance task. Brain Topography
Silberstein RB, Ciorciari J, Pipingas A (1995): Steady-state visually
evoked potential topography during the Wisconsin card sorting
test. Electroencephalogr Clin Neurophysiol 96:24–35.
Talairach J, Tourneaux P (1988): Co-Planar Stereotaxic Atlas of the
Human Brain. New York: Thieme.
Van Dijk B, Spekreijse H (1990): Localization of electric and magnetic
sources of brain activity. In: Desmedt JE (ed): Visual Evoked
Potentials. Amsterdam: Elsevier Science Pub., pp 57–74.
Wilson GF, O’Donnell RD (1986): Steady-state evoked responses:
Correlations with human cognition. Psychophysiology 23:57–61.
Без категории
Размер файла
116 Кб
Пожаловаться на содержимое документа