Dynamic crossmodal links revealed by steady-state responses in auditory-visual divided attention.

Abstract

Frequency tagging has been often used to study intramodal attention but not intermodal attention. We used EEG and simultaneous frequency tagging of auditory and visual sources to study intermodal focused and divided attention in detection and discrimination performance. Divided-attention costs were smaller, but still significant, in detection than in discrimination. The auditory steady-state response (SSR) showed no effects of attention at frontocentral locations, but did so at occipital locations where it was evident only when attention was divided between audition and vision. Similarly, the visual SSR at occipital locations was substantially enhanced when attention was divided across modalities. Both effects were equally present in detection and discrimination. We suggest that both effects reflect a common cause: An attention-dependent influence of auditory information processing on early cortical stages of visual information processing, mediated by enhanced effective connectivity between the two modalities under conditions of divided attention.

Full text

Page 1

Dynamic crossmodal links revealed by steady-state responses in auditory–visual
divided attention
Ritske de Jong⁎, Paolo Toffanin, Marten Harbers
Experimental Psychology, University of Groningen, The Netherlands
a b s t r a c ta r t i c l ei n f o
Article history:
Received 23 March 2009
Received in revised form 13 August 2009
Accepted 15 September 2009
Available online xxxx
Keywords:
ASSR
VSSR
Intermodal attention
Multisensory integration
Crossmodal links
Frequency tagging has been often used to study intramodal attention but not intermodal attention. We used
EEG and simultaneous frequency tagging of auditory and visual sources to study intermodal focused and
divided attention in detection and discrimination performance. Divided-attention costs were smaller, but
still significant, in detection than in discrimination. The auditory steady-state response (SSR) showed no
effects of attention at frontocentral locations, but did so at occipital locations where it was evident only when
attention was divided between audition and vision. Similarly, the visual SSR at occipital locations was
substantially enhanced when attention was divided across modalities. Both effects were equally present in
detection and discrimination. We suggest that both effects reflect a common cause: An attention-dependent
influence of auditory information processing on early cortical stages of visual information processing,
mediated by enhanced effective connectivity between the two modalities under conditions of divided
attention.
© 2009 Elsevier B.V. All rights reserved.
1. Introduction
Selective attention plays an important role in managing the con-
tinuous stream of incoming sensory information. By focusing
attention on currently relevant information, the processing system
is biased towards processing this information relative to currently less
relevant and potentially conflicting or interfering information (Deco
and Zihl, 2006; Desimone and Duncan, 1995). Attention can be
focussed on a single source (e.g., object or location) but can also be
divided across multiple information sources. These different modes of
attention are referred to as focussed and divided attention.
Performance in divided-attention studies is usually interpreted in
terms of a tradeoff between the relative amount of attention allocated
to each task or source and the performance on each task. When a task
is resource limited, performance is expected to improve as a
monotonic function of the relative amount of attention allocated to
the task. Performance tradeoffs between two concurrent tasks can be
orderly represented in an “Attention Operating Characteristic” (AOC)
which plots the joint performance on two tasks as a function of the
attentional instructions (e.g. Bonnel and Hafter 1998, Sperling and
Melcher, 1978; Norman and Bobrow, 1975).
Divided-attention costs can vary dependingon both the features of
the stimulus and the type of response to be made (Duncan, 1980).
Bonnel et al. (1992) and Bonnel and Hafter (1998) showed that such
costs can also depend on the nature of the task — concurrent target
detection tasks did not yield statistically significant performance
tradeoffs whereas strong tradeoffs were found for concurrent target
discrimination tasks. Bonnel et al. (1992) and Bonnel and Hafter
(1998) interpreted their results in terms of a hierarchical processing
model (Hoffman, 1979; Duncan, 1981): Detection tasks depend on a
relatively low level capacity-free processing stage, whereas discrim-
ination additionally requires a higher level capacity-limited stage.
The formal distinction between detection and discrimination of
targets has been debated in the literature (Brawnand Snowden, 2000;
Kawahara et al., 2001). We follow the distinction made by Kawahara
and collaborators, which seems most compatible with the procedures
used by Bonnel et al. (1992) and Bonnel and Hafter (1998). They
proposed that detection is involved when an observer is asked to
distinguish between a uniform stimulus anda stimulus that contains a
non-uniformity or discontinuity in space or time. Only the presence or
absence of the non-uniformity is to be reported, not its identity. By
contrast, discrimination (also referred to as identification) requires
the observer to identify the nature of the discontinuity.
The notion that detection performance is not capacity-limited has
not gone unchallenged. Several studies found target-detection
performance to be enhanced when target location was cued in
advance, suggesting that detection can be affected by attentional
mechanisms (Muller and Humphreys, 1991; Hawkins et al., 1990;
Downing, 1988; see also Lappin and Uttal, 1976). Bonnel et al. (1992)
argued that such results do provide evidence that target detection can
be affected by spatialuncertaintybut do notundermine theirproposal
that detection performance is capacity free when, as in their own
studies, spatial uncertainty has been removed.
International Journal of Psychophysiology xxx (2009) xxx–xxx
⁎ Corresponding author. Grote Kruisstraat 2/1, 9712 TS, Groningen, The Netherlands.
Tel.: +31 50 363 6463.
E-mail address: ritske.de.jong@rug.nl (R. de Jong).
INTPSY-10074; No of Pages 13
0167-8760/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.ijpsycho.2009.09.013
Contents lists available at ScienceDirect
International Journal of Psychophysiology
journal homepage: www.elsevier.com/locate/ijpsycho
ARTICLE IN PRESS
Please cite this article as: de Jong, R., et al., Dynamic crossmodal links revealed by steady-state responses in auditory–visual divided
attention, International Journal of Psychophysiology (2009), doi:10.1016/j.ijpsycho.2009.09.013

Page 2

In the present study, we followed the procedure of Bonnel and
Hafter (1998), employing concurrent crossmodal auditory and visual
detection or discrimination tasks with fixed stimulus locations and
with several different attention allocation instructions. The tasks
involved either detecting or discriminating a small and brief upward
or downward change in tone volume in the auditory task and in
stimulus brightness in the visual task. This particular procedure was
chosen for several reasons. The data presented by Bonnel and Hafter
(1998; Table 4) do seem to show effects of instructed attention
allocation on performance even for concurrent detection. While these
effects may have been due, at least in part, to participants matching
performance to instructed relative attention (Navon, 1985), they
might also reflect subtle but theoretically important capacity limita-
tions in detection performance. Related to this point, participants may
have become aware that different attentional policies impacted
discrimination performance more than detection performance, and
this may have led them to adopt less differentiating attention
allocation policies in detection conditions. To address the possibility
of different attentional strategies in crossmodal detection and
discrimination, we recorded electroencephalographic (EEG) activity
and used event-related potentials (ERPs) and, in particular, steady-
state responses (SSRs) to gain information regarding the attentional
mechanisms operative in the various experimental conditions.
The apparent absence of divided-attention costs in detection was
reported by Bonnel et al. (1992) and Bonnel and Hafter (1998) for
both intramodal and intermodal divided attention. In the present
study, and as outlined more extensively below, we chose to focus on
intermodal rather than intramodal attention in order to benefit from
and contribute to the recent surge of research interest and findings
regarding crossmodal effects in intermodal perception and attention
(Driver and Noesselt, 2008; Macaluso and Driver, 2005; Stein and
Stanford, 2008).
1.1. Using steady-state responses for studying auditory–visual
intermodal attention
The auditory steady-state response (ASSR) refers to oscillatory EEG
orMEGactivitywithconstantamplitudeandphaseatthefrequencyofa
periodic eliciting stimulus. The ASSR can be elicited by stimuli such as
sequences of clicks, Gaussian tone pulses, or amplitude-modulated
tones, and its amplitude is typically found to be maximal at stimulation
frequencies around 40Hz (Ross et al., 2005a). The ASSR is generally
thought to result from periodic superposition of middle latency evoked
responses (Galambos et al., 1981; Bohorquez and Ozdamar, 2008), but
the alternative hypothesis that the ASSR reflects a separate neural
oscillation that is driven by the periodic stimulus has also received
empiricalsupport(Rossetal.,2005a).Convergingevidencefromseveral
neuroimaging studies and patient studies has shown the primary
auditorycortex(medialportionofHeschl'sgyrus)tobethemainsource
of the 40-Hz ASSR (Gutschalk et al., 1999; Ross et al., 2002, Ross et al.,
2005b; Simpson et al., 2005).
While the functional meaning of the ASSR remains unclear, an
interesting hypothesis is that its amplitude may index sensory gain
control in primary auditory cortex (Giard et al., 2000). Some studies
have found effects of attention on the 40-Hz ASSR amplitude, for
intramodal auditory selective attention (Bidet-Caulet et al., 2007;
Skosnik et al., 2007; Gander et al., 2007) as well as for auditory–visual
intermodal selective attention (Ross et al., 2004a,b; Saupe et al.,
2009), but such effects have not always been found (Linden et al.,
1987). A key feature of studies that obtained positive effects of
attention on ASSR amplitude may be that they used tasks that
required focused attention to the stimulus rhythm, or modulation
frequency, eliciting the ASSR. Whether such effects can also be found
when the stimulus rhythm is incidental to the task to be performed,
and thus merely serves as a frequency tag for quantification of the
ASSR, is uncertain — one of the aims of the present study is to assess
this possibility for the case of auditory–visual intermodal attention.
ThevisualcounterparttotheASSRisthevisualsteady-stateresponse
(VSSR; Regan, 1989), which is found with amplitude-modulated or
flickering visual stimuli across a wide range of frequencies (Herrmann,
2001). In a recent elegant study, two main cortical sources of a 12-Hz
VSSR were established: a main source in the primary visual cortex (V1)
and a secondary source in the motion sensitive areas (MT/V5) (Russo
etal.,2007).ThissuggeststhepossibilitythatVSSRamplitudemayindex
sensory gain control in primary visual cortex. Positive effects of
intramodal selective attention on VSSR amplitude have been reported
in manystudies thatemployed different paradigms suchasvisuospatial
attention or selective color-based processing (e.g., Morgan et al., 1996;
Muller et al. 1998a; Hillyard et al., 1998; Malinowski et al. 2007; Ding
et al., 2006; Pei et al., 2002). In a visuospatial divided-attention task,
VSSR amplitude for frequencies ranging from 8 to 23Hz exhibited a
graded dependency on attention, with substantially larger amplitudes
for attended as compared to unattended locations and intermediate
amplitudes when attention wasdividedacrosslocations(Toffanin etal.,
2009). In contrast to the manystudies that used the VSSR to investigate
neuralmechanismsofintramodalvisual attention,weareawareofonly
one study that used this measure to investigate effects of intermodal
attention on visual information processing — Talsma et al. (2006) used
theVSSRtodemonstratethatattentionalcapacitybetweenmodalitiesis
larger than that within one and the same modality.
1.2. The present study
ASSR amplitude and VSSR amplitude may provide useful indices of
sensory gain control in early cortical auditory processing and visual
processing, respectively. In the present study, we used these measures
to investigate effects of intermodal auditory–visual selective or divided
attention on early modality-specific cortical processing in detection or
discrimination. The almost exclusive emphasis on steady-state
responses may require some explanation, as we seem to have ignored
the much more extensive and arguably better developed literature on
the use of event-related potentials (ERPs) in the investigation of
intramodal and intermodal attention (for review, see Driver and
Noesselt, 2008; Macaluso and Driver, 2005). The reasons for this are
mostly practical — in the present study we were unable to reliably
measure short-latency and mid-latency ERP components (P50, C1, N1,
P1, P2) that reflect sensory processing. Most likely, this absence of early
ERP components was caused by the low saliency of the targets in both
detection and discrimination.
The auditory stimulus consisted of a 40-Hz amplitude-modulated
tone and the visual stimulus of a 24-Hz luminance-modulated colored
square, both presented simultaneously for 2.5s at approximately the
same central location. Participants judged visual and auditory stimuli
to share a common spatial source — thus, crossmodal selective
attentioncould not be achievedby means of spatial selective attention
to one or the other source of stimulation (Eimer and Schroger, 1998).
Auditory and visual targets consisted of a brief (125ms) decrease or
increase of the mean loudness of the tone and of the main brightness
of the colored square, respectively. Participants had to detect the
presence of targets in the detection task, and had to identify the
direction of change (decrease or increase) in the discrimination task.
In different attention conditions, they were instructed to selectively
attend to oneand to ignorethe othermodality (100% auditory or 100%
visual), or to divide their attention across modalities (20–80%, 50–
50%, or 80–20% auditory–visual).
As relevant prior data regarding intermodal attention effects on
ASSR and VSSR amplitudes in the literature are either inconsistent or
largely absent (but see Talsma et al., 2006), the present study must be
deemed exploratory.However,some possible outcomes will be briefly
discussed. One possibility is that intermodal attention effects on ASSR
and VSSR amplitude are mediated by largely the same neural
2
R. de Jong et al. / International Journal of Psychophysiology xxx (2009) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: de Jong, R., et al., Dynamic crossmodal links revealed by steady-state responses in auditory–visual divided
attention, International Journal of Psychophysiology (2009), doi:10.1016/j.ijpsycho.2009.09.013

Page 3

structures and mechanisms as those underlying intramodal attention
effects (Talsma and Kok, 2001). In that case, graded effects of
intermodal attention, similar to the graded effects of visuospatial
attention on VSSR amplitude reported by Toffanin et al. (2009)),
might be expected. Moreover, such effects might depend on whether
the task involves detection or discrimination.
On the other hand, recent empirical and theoretical developments
regarding crossmodal influences on modality-specific brain regions may
suggest a very different outcome (for review, see Driver and Noesselt,
2008;Steinand Stanford, 2008), especiallywithregard tothedistinction
between divided-attention and focused-attention conditions. Many
studies have found multisensory influences even in relatively low-
order, early regions of sensory cortex (for review, Ghazanfar and
Schroeder, 2006). Such influences may underlie multisensory effects on
perceptual judgments, and are subject to constraints of spatial and
temporal proximity or overlap between incoming signals from different
sensory modalities (Macaluso and Driver, 2005). Possible mechanisms
underlyingcrossmodalinfluencesincluderapidfeedforwardintegration,
thalamic influences, direct connections between primary cortices, and
feedback from multisensory regions to sensory-specific brain areas
(DriverandNoesselt,2008).Apartfromspatialandtemporalconstraints,
whicharelikelytobefulfilledinthepresentstudy,crossmodalinfluences
on modality-specific brain regions may also depend on attentional
factors.Talsmaetal.(2007)foundcrossmodalmodulationoftheauditory
P50 only when participants attended to both audition and vision, and
stated “We thus conclude that when attention is directed to both
modalities simultaneously, auditory and visual stimuli are integrated
very early in the sensory flow of processing (∼50ms poststimulus).
Attention appears to play a crucial role in initiating such an early
integration of auditory and visual stimuli. When only one modality is
attended, the integration processes appear to be delayed.” (p. 689). This
idea suggests the interesting possibility of crossmodal effects on ASSR
and VSSR amplitudes, where these effects may be particularly evident in
the divided-attention conditions. More specifically, crossmodal interac-
tions may manifest in two possibly related ways in divided-attention
conditions. First, in these conditions 40-Hz activity, which specifically
tagsauditoryinput,mayshowupalsoatposterioroccipitalsites,whereas
24-Hz activity may show up at frontocentral, auditory sites. Second,
crossmodalinteractions mayserveto enhancesensorygaininlow-order
auditoryandvisualcortices,thussuggestingthepossibilitythatASSRand
VSSR amplitudes might actually be enhanced in divided-attention as
compared to focused-attention conditions. Again, such effects may
depend on whether the task involves detection or discrimination.
2. Experiment
2.1. Method
2.1.1. Participants
Ten healthy right-handed students (3 males) between the age of 19
and 23 (mean=21.1; sd=1.5) participated as paid volunteers. All
volunteers had normal or corrected-to-normal vision. The study
adhered to the Declaration of Helsinki and was approved by the local
ethics committee. Informed consent was obtained from all participants.
Due totechnical malfunctioning, EEG data for one participantwere lost.
2.1.2. Stimuli and apparatus
Visual stimuli were presented on a 17inch CRT screen graphics
(800×600 pixels with a 144-Hz refresh rate). Auditory stimuli were
presented through a single loudspeaker placed at the top center of the
computer monitor used for visual presentation; participants judged
visual and auditory stimuli to share a common spatial source. The
auditory stimulus was presented at 65 dB, measured at the location of
the participants' ears. Stimuli were generated with the Matlab
Psychtoolbox (Brainard, 1997; Pelli, 1997). Participants accommodated
their chin ona chinrest ata distance of 65cmfrom thedisplaymonitor.
The computer keyboard was used for the participants' responses.
The visual stimulus wasa 4×4cm(subtending3.52º of visual angle)
oscillating yellow square presented at the center of the screen. The
luminanceoscillation wascreatedbychangingtheredandgreen values
oftheRGBvaluesofthesquareaccordingtoa24Hzsinewavewhilethe
blue value of the RGB value was kept at zero. The mathematical
description of the red and green values at time t is:
colðtÞ = b⋅ð1 + m⋅sinð2π⋅f⋅tÞÞ;
where b is the base value of the colors (125), m the modulation value
(0.6), and f the modulation frequency (24Hz). The visual target was
an increase or decrease in the base value (b) for a period of 125ms
(three cycles of a 24Hz sine wave). The auditory stimulus was a
500Hz sine wave with 100% amplitude modulation at 40Hz. The
auditory target was an increase or decrease in the maximum volume
of the tone for a period of 125ms (exactly 5 cycles of a 40Hz sine
wave). To avoid any clearly perceptible discontinuities at target onset
and offset, auditory target onset always coincided with the modulated
amplitude being at or near its minimum value.
For both the visual and the auditory stimulus modulation onset
was at t=0. This was done to insure that the modulation was time/
phase locked on every trial.
2.1.3. Experimental design
All the participants performed both the detection and the dis-
crimination tasks which were administered in separate sessions with
a week in between sessions. Task order was counterbalanced across
participants.
The event structure of a trial is illustrated in Fig. 1 (Panel A). A trial
startedwitha presentationofa grey squareinthecenterof thescreenfor
2000ms — darkening of the square during the last 1000ms indicated the
upcoming presentation of the stimuli. After this warning period, the
amplitude-modulated auditory and visual stimuli were presented for
2500ms. During this 2500ms period visual and auditory targets were
presented. If a trial contained targets in both modalities, the targets were
presentedatthesametime.In80%ofthetrials,thetargetswerepresented
between 1000 and 2000ms. In the other 20%, the targets were presented
between 300 and 1000ms. The latter trials served to force participants to
remain focused during the entire interval and were excluded from
analysis, ensuring a sufficiently long pre-target interval on the remaining
analyzed trials for the SSR to develop and stabilize (Ross et al., 2005b).
In the detection task, participants were instructed to indicate
whether a visual and/or an auditory target was present. Target prob-
ability for each modality was 50%, with a random half of targets being a
decrease in volume or luminance and the other half an increase. Target
presence and target type was randomly determined on each trial and
independently for the two modalities.
In the discrimination task, auditory and visual targets were
presented on each trial. Participants had to indicate if the target was
a decrease oran increasein the volumeof the toneorluminanceof the
square. Target type was randomly determined on each trial and
independently for the two modalities.
Both tasks were administered in five different attentional condi-
tions. Two conditions were focused-attention conditions where par-
ticipants were instructed to focus their attention fully on one of the
modalities (100% auditory or 100% visual). In the other three attention
conditions they were instructed to divide attention between the two
modalities in one of three different ways: 80% attention visual and 20%
attentionauditory, 50% visualand 50% auditory, and 20% visualand 80%
auditory — in the 50–50% condition, they were told to divide attention
equally between the modalities, whereas in the 80–20% they were told
to focus most of their attention on one modality but not to completely
ignoretheothermodality.Eachoftheseconditionswasadministeredin
two consecutive blocks of 60 trials. The order of the attention
3
R. de Jong et al. / International Journal of Psychophysiology xxx (2009) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: de Jong, R., et al., Dynamic crossmodal links revealed by steady-state responses in auditory–visual divided
attention, International Journal of Psychophysiology (2009), doi:10.1016/j.ijpsycho.2009.09.013

Page 4

conditions was determined by a reduced order five Latin square in such
a way that participant number six received the same condition order as
participant number one. The order of the conditions was the same in
the detection and the discrimination tasks for each participant.
Participants were prompted to respond after the 2.5 s period of
stimulus presentation had expired. As Bonnel and Hafter (1998) found
no effects of response order in a similar paradigm, participants were
prompted to respond to the dominant modality first, by pushing one of
twobuttonslabeledas‘no’and‘yes’inthedetectiontaskandas‘up’and
‘down’ in the discrimination task. In the two focused attention con-
ditions the name of the attended modality served as the single prompt
to respond. In the divided attention conditions participants were
consecutively prompted to respond for each modality. The response for
the dominant modality was made first in the 80/20 and the 20/80
conditions, and in the 50/50 condition response order was alternated
between participants with, for each participant, opposite orders being
used in the detection and the discrimination tasks. Responses for the
auditory modality were made with the left hand and for the visual
modality with the right hand.
2.1.4. Training
Each session started with a short explanation of the task to be
performed, followed by a training that involved 100 trials of practice
in each of the two focused attention conditions. Only stimuli and
targets in the relevant modality were presented in these practice runs.
During each practice run, a staircase tracking algorithm adjusted
the target intensity (i.e. the volume difference between the baseline
tone and the target tone or the luminance difference between the
baseline square and the target square) to achieve an 80% accuracy
level (d′=1.7). The resulting target intensities were used in the ex-
perimental part of the session. Note that this procedure aimed to
equate performance accuracies in the focused attention conditions
across all combinations of task and modality.
2.1.5. EEG recording and electrode placement
EEGwasrecorded usinga common referenceamplifier: All channels
were amplified against the average of all connected inputs. Electrical
activity measured at the electrodes was amplified 20.000 times and
digitally filteredusinga low-pass FIR filter witha cutoff value of 140Hz.
Fig. 1. Panel A. Schematic of trial procedure. A grey square was displayed for 1s in the center of the screen and darkened for another second indicating the upcoming presentation of
the 2.5s amplitude-modulated auditory and visual stimuli. Targets could be presented in the visual and/or auditory modality. To ensure that participants deployed attention on the
modulated stimuli for the whole trial duration, targets were presented between 1 and 2s in 80% of the trials, or between 0.3 and 1s in the remaining 20%. Only trials with targets
between 1 and 2s were analyzed. When targets were presented in both modalities they appeared simultaneously. Panels B and C depict the procedure for computing SSR amplitude.
In this example VSSR amplitude is computed for a specific channel (PO3) in one of the conditions for an individual participant. Panel B: The visual steady-state response (VSSR)
obtained by averaging across all available trials. Onset of the frequency tags is at t=0, and an oscillatory EEG response can be seen to develop quickly and last until offset of the tags
(t=2.5). The large component at t=−2 is related to onset of the visual warning stimulus. Panel C: Visual steady-state response (VSSR) after narrow band-pass filtering (f=24Hz,
σ(f)=0.2Hz). The amplitude envelope was obtained through Hilbert transform. Mean VSSR amplitude was computed across the interval 1–1.5s after stimulus onset.
4
R. de Jong et al. / International Journal of Psychophysiology xxx (2009) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: de Jong, R., et al., Dynamic crossmodal links revealed by steady-state responses in auditory–visual divided
attention, International Journal of Psychophysiology (2009), doi:10.1016/j.ijpsycho.2009.09.013

Page 5

The EEG signals were recorded using Brain Vision Recorder (Brain
Products GmbH, Munich, Germany). EEG signal was digitized at 500Hz
with a resolution of 0.02µV.
A 70-electrode cap (Electro-cap International Inc, Eaton, Ohio,
USA) was used for placement of the tin electrodes. All scalp positions
in the International 10–20 System were used, with additional sites
located midway between the 10–20 locations (Sharbrough et al.,
1991) and six electrode positions 10% inferior to the standard parieto-
occipital electrodes (FT9, PO9, O9, FT10, PO10, O10). Two additional
electrodes were placed on the left and right mastoid for offline re-
referencing of the EEG signal, and the ground electrode was placed on
the sternum. Eye movements were monitored via bipolar recordings
of the electrooculogram (EOG). For the horizontal EOG the electrodes
were placed at the left and right external canthi and for the vertical
EOG the electrodes were placed above and beneath the middle of the
left eye. All inter-electrode impedances were kept below 5kΩ for the
EEG electrodes and 10kΩ for the EOG electrodes.
2.1.6. Data analysis
Only trials where the target was presented after the first second of
the stimulation interval were used for analysis (80% of the trials). This
ensures that the SSR was well-developed and had stabilized by the
time the target was presented.
2.1.7. Behavioral data
Dprime (d′) values were calculated from hit and false alarm rates
for each subject and each experimental condition, using standard
procedures (Macmillan and Creelman, 2005):
d′= zðHÞ−zðFÞ;
where z(H) and z(F) are the z-transformations of the hit rate and the
false alarm rate respectively. Dprime values were subsequently
subjected to repeated-measures analysis of variance (ANOVA). For
all ANOVA's of behavioral and EEG data, Greenhouse–Geisser epsilon
correction was used to adjust p-values, whenever applicable.
2.1.8. Electrophysiological data
All signals were re-referenced to the average of the mastoids. A
0.1Hz high-pass filter (12dB/oct slope) was applied to correct for
slow wave fluctuations in signal amplitude. Ocular correction was
performed using the method developed by Gratton et al. (1983)). All
data was visually inspected for artifacts after segmentation.
2.1.9. ERPs
FortheERPanalysis,segmentsof1000msweremadestarting200ms
prior to target onset on trials where responses were correct. A baseline
correction was performed using the 200ms interval preceding target
onset.Segmentswereaveragedandthensmoothedwitha9Hzlow-pass
filter (48dB/oct slope). For the detection task, ERPs were calculated for
each attention condition, separately for trials containing only a visual
target, only an auditory target, or both targets. For the discrimination
task, where every trial contained both a visual and an auditory target,
ERPs were calculated for each attention condition. Because early
components, including the N2, could not be reliably identified in the
ERPs (see Fig. 3), presumably due to the use of the frequency tagging
technique,1analysis was restricted to peak amplitude and latency values
of the P300 component at Cz. These values were determined by simple
peak picking in the 350–750ms interval following target onset.
2.1.10. SSRs
The procedure used to estimate SSR amplitude is illustrated in
Fig. 1 (Panels B and C). We used segments of 8000ms that began
3000ms before the start of the 2500ms stimulation interval during
which the amplitude-modulated auditory and visual signals were
presented. Segments were averaged over trials, separately for each
combination of task and attention condition, resulting in a steady-
state potentials (SSRs; Fig. 1, panel B). Instantaneous SSR amplitudes
(IAs) were then computed in two steps. First, SSRs were filtered, using
FFT and a narrow Gaussian band-pass filter defined by a center
frequency f (the driving frequencies of the visual and auditory stimuli,
24 or 40Hz, respectively) and a standard deviation σ(f) (0.2Hz).
Instantaneous amplitudes of the band-pass filtered SSRs were
computed by means of (the absolute value of) the Hilbert transform
(Fig. 1, panel C). This procedure is formally equivalent to computing
instantaneous amplitude by means of a convolution with a complex
Morlet wavelet2(for details, see Gladwin et al., 2006; Tallon-Baudry
and Bertrand, 1999). SSR amplitude was computed as the average
amplitude across the 1000–1500ms interval following onset of the
2500ms stimulation interval.
Average scalp distributions of ASSR and VSSR amplitudes were
computed for each combination of task and attention condition. For
each subject, normalized scalp distributions were computed by
dividing the amplitude at each electrode by the maximum amplitude
across all electrodes (McCarthy and Wood, 1985). Normalized
amplitudes were then averaged across participants. Values for
individual electrodes of the average scalp distributions vary between
0 and 1, with high values indicating consistently high relative SSR
amplitudes across participants for that electrode and low values
indicating consistently low relative SSR amplitudes.
TheprocedureusedtoestimateSSRamplitudeasoutlinedabove,i.e.,
time-domain averaging across trials followed by frequency-domain
estimation of SSR amplitude, renders the nature of possible attentional
effects on SSR amplitude ambiguous. Such effects could reflect veridical
and consistent effects on SSR amplitude at the level of single-trials (i.e.,
veridical effects on sensory gain control) but could also reflect
attentionaleffects on intertrialcoherenceofthephaserelation between
the driving oscillatory stimulation and the resulting EEG activity (Ding
et al., 2006; Kim et al., 2007). To address this possibility, we computed
SSR amplitude for single-trials, using the procedure detailed above, and
then averaged these values across trials. By definition, average SSR
amplitude estimated in this way cannot be sensitive to degree of
intertrial phase coherence. Therefore, if attention primarily affects such
phase coherence, attentional effects on SSR amplitude should no longer
be present. In fact, attentional effects computed this way were virtually
identical to those obtained with the former method of estimating SSR
amplitude. We therefore conclude that the attentional effects on SSR
amplitude, computed with the former method and to be reported
below, almost certainly reflect genuine effects of attention on sensory
gain control.
3. Results
3.1. Behavioral results
Averaged d′-values for each combination of modality (auditory,
visual), task (detection, discrimination), and attention condition
(100%, 80%, 50%, and 20% — note that no responses were given in
the 0% condition) are presented in Fig. 2. Note that our procedure for
1Early ERP components also failed to become evident at any electrode position
when the cutoff of the low-pass filter was set at 20Hz.
2For the complex Morlet wavelet, frequency resolution (σ(f)) and time resolution
(σ(t)) are reciprocally related as follows: σ(f). σ(t)=1/(2π). Small values for σ(f)
provide a desirably narrow focus on the driving or tagging frequency, but also result in
a loss of temporal resolution in computing instantaneous amplitude. Because
stimulation extended across a 2500ms interval, during which SSR amplitude may be
assumed to remain fairly constant (see Fig. 1), some loss of temporal resolution could
be tolerated but up to a limit. Based on computer simulations, we determined that a
value of 0.2Hz for σ(f) – corresponding to σ(t)≈0.8s – provided a near-optimal
compromise between the opposing requirements of frequency specificity and
temporal specificity in the present study.
5
R. de Jong et al. / International Journal of Psychophysiology xxx (2009) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: de Jong, R., et al., Dynamic crossmodal links revealed by steady-state responses in auditory–visual divided
attention, International Journal of Psychophysiology (2009), doi:10.1016/j.ijpsycho.2009.09.013

Page 6

adjustingtargetintensityduringtraininginordertoachievecomparable
levels of performance in the focused-attention (100%) condition for all
combinations of modality and task, seems to have been quite effective.
Divided-attention costs on performance were much stronger for
discrimination than for detection, but some costs seem to be present
also for detection. A 2×2×4 (task ×modality×attention) repeated-
measures analysis of variance (ANOVA) yielded significant main effects
of task, F(1,8)=6.4, p<.05, and attention, F(3,24)=74.9, p<.001, and
a significant interaction of task and attention,F(3,24)=14.0,p<.001. A
separate analysis on the results for detection yielded a significant main
effect of attention, F(3,24)=18.8, p<.001.
3.2. Electrophysiological results
3.2.1. ERPs
ERPs at electrode Cz and time-locked to target onset are shown in
Fig. 3, for all attention conditions in the detection and discrimination
tasks. Note that these ERPs were computed from correct trials only.
Separate panels show the ERPs associated with the three possible
target events in the detection task: visual-only, auditory-only, and
visual+auditory. The waveforms are dominated by the P300 that
started after about 300ms, whereas most of the earlier ERP
components are hardly visible — as suggested earlier, this latter
phenomenon is likely to result from the usage of frequency tagging in
both modalities. Analysis of ERP waveforms was therefore restricted
to peak P300 amplitude in the 400–800ms interval.
The first analysis focused on P300 amplitudes for visual-only and
auditory-only targets in the detection task. A repeated-measures
ANOVA with target type (visual, auditory) and attention condition
(recodedtoindicateattentiontothetargettyperangingfrom0to100%)
as within-subject factors yielded main effects of target type, F(1,8)=
11.6, p<.01, and attention condition, F(4,32)=18.3, p<.001, and a
significant interaction, F(4,32)=3.1, p<.05. Pairwise comparisons
between the attention conditions showed no significant differences
between the 100%, 80%, and 50% conditions, significant differences
between each of these three conditions and both the 0% and 20%
conditions, and a significant difference between the 0% and 20%
conditions. Thus, visual targets elicited larger P300 components than
auditorytargets,and,with theexceptionofthe20%condition,theeffect
of attention on P300 amplitude was approximately all-or-none.
Inspection of Fig. 3 suggests that P300 amplitudes for bimodal
targets in the detection task may roughly correspond to the sum of
amplitudes for visual-only and auditory-only targets in that task. An
analysis designed to test for possible differences between amplitudes
for bimodal targets and the sum of unimodal targets yielded no
significant differences. This result is consistent with the notion that
the P300 response to bimodal targets in detection is composed of the
linear sum of independent P300 responses to each of the constituent
unimodal target events (Teder-Salejarvi et al., 2002; Talsma and
Woldorff, 2005).
P300 amplitudes for bimodal targets in detection and discrimina-
tion were analyzed in a repeated-measures ANOVA with task and
attention condition as within-subject factors. Apart from the main
effects of attention condition, this analysis yielded as a new result a
borderline significant main effect of task, F(1,8)=4.5, p<.07,
reflecting somewhat smaller P300 amplitudes when the task required
discrimination as compared to detection.
3.2.2. VSSR
Averaged topographical maps of the 24Hz VSSR amplitude are
shown in Fig. 4, for all combinations of task and attention condition.
These maps show a pronounced and consistent maximum at the most
posterior electrodes. Also evident, for detection and discrimination,
are the increased values at these electrodes for divided-attention as
compared to the focused-attention conditions. Statistical analysis of
VSSR amplitude was based on the average amplitudes computed
across the four electrodes where this amplitude, averaged across all
task by attention condition combinations, was maximal. This subset of
electrodes was determined separately for each participant, but
showed a high degree of consistency across participants, containing
only electrodes at posterior occipitoparietal locations. Estimates of the
signal-to-noise ratio (SNR) of VSSR amplitudes that were used in
statistical analysis were computed in the following way. We used
exactly the same procedure as that used for computing the 24 Hz
response, but now centered the Gaussian filter at 22Hz or 26Hz.
Dividing the 24Hz amplitude estimate, averaged across all task x
attention condition combinations, by the average of the 22 and 26Hz
estimates, similarly averaged across experimental conditions, pro-
vides an estimate of the frequency specificity, or SNR, of the 24Hz
VSSR amplitude. Values of this estimate across participants ranged
from 11.5 to 48.3, with an average of 28.4, indicating excellent
frequency specificity for VSSR amplitude estimation.
To accommodate the substantial individual differences in VSSR
amplitude, values were normalized for each participant by dividing by
the average value across all ten (task x attention condition) con-
ditions. Fig. 5 shows normalized mean VSSR amplitude as a function of
task and attention condition. For both detection and discrimination,
VSSR amplitude increasedmarkedly fromfocused-attentiontodivided-
attention conditions, culminating at the 50–50% condition. Amplitudes
also appear to be slightly larger in the 100%-visual as compared to the
100%-auditory condition. A repeated-measures ANOVA with task and
attention condition as within-subject factors yielded only a main effect
of attention condition, F(4,32)=8.5, p<.001. Polynomial contrast
analysis showed this effect to be explained completely by the quadratic
contrast, F(1,8)=24.2, p<.001 — the linear contrast did not approach
significance, F(1,8)=1.5, p>.26.
To assess the relation between VSSR amplitude and performance
accuracy, VSSR amplitude was computed separately for correct trials
and incorrect trials, and separately so for visual targets and auditory
targetsandfordetectionanddiscrimination.Inordertohaveasufficient
number of trials for stable estimates of VSSR amplitude, errors were
pooled across attention conditions — corresponding ‘pooled’ values for
correcttrialswereobtainedbyweightedaveragingoftheamplitudesfor
individual attention conditions, with the weights corresponding to the
relative frequency of errors in each condition. For each participant, the
resulting values were normalized by dividing by the average value of
Fig. 2. Performance accuracies (d′) as a function of attention condition for each
combination of target modality (auditory, visual) and task (detection, discrimination).
Attention condition refers to the proportion instructed attention to be allocated to the
target modality (100% thus refers to the focused-attention conditions).
6
R. de Jong et al. / International Journal of Psychophysiology xxx (2009) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: de Jong, R., et al., Dynamic crossmodal links revealed by steady-state responses in auditory–visual divided
attention, International Journal of Psychophysiology (2009), doi:10.1016/j.ijpsycho.2009.09.013

Page 7

Fig. 4. Topographic maps of the visual steady-state response (24Hz). The upper panel depicts separate maps for each attention condition in the detection task; the lower panel does
so for the discrimination task. Higher values represent higher average normalized amplitudes across participants at that location.
Fig. 3. ERP waveforms at Cz. Target onset is at t=0. The upper and lower left panels display the waveforms in each attention condition separately for each of the three possible target
types in the detection task (auditory-only, visual-only, auditory+visual). The lower right panel shows the waveforms for the discrimination task.
7
R. de Jong et al. / International Journal of Psychophysiology xxx (2009) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: de Jong, R., et al., Dynamic crossmodal links revealed by steady-state responses in auditory–visual divided
attention, International Journal of Psychophysiology (2009), doi:10.1016/j.ijpsycho.2009.09.013

Page 8

‘correct’ amplitudes across the four combinations of task and target
modality. The results of this analysis, averaged across participants, are
shownintheleftpanelofFig.6.Amplitudesseemsomewhatsmallerfor
detection as compared to discrimination. Amplitudes for auditory
targets appear to be somewhat larger for errors than for correct
responses in both detection and discrimination, whereas an opposite
but more pronounced effect seems to be present for visual targets. A
repeated-measures ANOVA with correctness, task, and target modality
aswithin-subjectfactorsyieldedasignificantmaineffectofcorrectness,
F(1,8)=9.0, p<.02, and a significant interaction of correctness and
target modality, F(1,8)=10.9, p<.015. Separate analyses for visual and
auditorytargetsyieldedasignificantmaineffectofcorrectnessforvisual
targets,F(1,8)=22.2,p<.002, but not for auditory targets,F(1,8)=4.1,
p>.10.
3.2.3. ASSR
Averaged topographical maps of the 40Hz ASSR amplitude are
shown in Fig. 7, for all combinations of task and attention condition.
Replicating previous results (e.g., Saupe et al., 2009), these mapsshow
ASSR amplitude to be largest across a fairly wide frontocentral region.
As for the VSSR, statistical analysis of ASSR amplitude was based on
the average value computed across the four electrodes where this
amplitude was maximal — all 36 electrodes (9 participants ⁎ 4
electrodes) used in this computation had a frontocentral location.
Following the same procedure as described above for VSSR (24Hz)
amplitude, but now with 38 and 42Hz as the bracketing frequencies,
yielded forthe40-Hz ASSRamplitudean SNR range of2.6 to10.0, with
an average of 5.5. These values indicate acceptable and useful
frequency specificity for ASSR amplitude.
ASSR amplitudes were normalized, using the same procedure as
that described earlier for VSSR amplitude. A repeated-measures
ANOVA with task and attention condition as within-subject factors
yielded no significant effects (all p's>0.17).
A possible relation between ASSR amplitude and accuracy for
auditory and visual targets was analyzed in the same way as
previously outlined for VSSR amplitude. The results of this analysis,
averaged across participants, are shown in the right panel of Fig. 6.
Amplitudes were somewhat smaller for detection as compared to
discrimination. More importantly, amplitudes for auditory targets
appeared to be smaller for errors than for correct responses in both
detection and discrimination, whereas the opposite effect seems to be
present for visual targets. A repeated-measures ANOVA with within-
subject factors correctness, task, and target modality yielded as the
only significant effect an interaction of correctness and target
modality, F(1,8)=10.1, p<.02. Separate analyses at each level of
target modality failed to show significant effects of correctness
(p's>.13). Thus, these results suggest an opposite but weak relation
between ASSR amplitude and performance accuracy for auditory
targets and visual targets. Note that these latter effects are almost the
mirror image of those found for VSSR amplitude depicted in the left
panel of Fig. 6.
Fig. 5. Average normalized amplitudes of the visual steady-state response, for all
attention conditions and separately for detection and discrimination.
Fig. 6. Normalized amplitudes of the visual steady-state potential (VSSR; left panel) and auditory steady-state potential (ASSR; right panel), as a function of correctness of the target
response, target type (auditory, visual), and task (detection, discrimination).
8
R. de Jong et al. / International Journal of Psychophysiology xxx (2009) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: de Jong, R., et al., Dynamic crossmodal links revealed by steady-state responses in auditory–visual divided
attention, International Journal of Psychophysiology (2009), doi:10.1016/j.ijpsycho.2009.09.013

Page 9

Closeinspection ofthetopographical mapsinFig.7 reveals a slightly
enhanced presence of 40Hz activity at posterior occipital electrodes for
the 50–50% divided-attention condition in both detection and discrim-
ination. Because such an effect would be of considerable theoretical
significance butthe observationis post hoc,we analyzed itin detail.We
averaged ASSR amplitude across the three most posterior electrodes
(O9, Iz, O10), as the effect seems most prominent at those electrodes —
very similar results were obtained when we used the same set of four
electrodesused for computingVSSR amplitude. Becausetheamplitudes
involved arevery small(lessthan0.1 µV) andtoaccommodatethevery
substantial individual differences, we also averaged amplitudes across
detection and discrimination to enhance robustness and then log-
transformed them (using the natural logarithm). Because the posterior
electrodesarepronetoEMGartifacts,andthe50–50%divided-attention
condition might be considered the most challenging one and lead
participants to tense up, it is possible that the effect represents an EMG
artifact. Inaddition,theeffectmightreflectenhanced gammaactivityin
background EEG at posterior sites under more challenging conditions
(Fell et al., 2003). To explicitly check for these alternative explanations,
weusedexactlythesameprocedureused tocompute40Hzamplitudes
to compute amplitudes at two neighboring frequencies, 38 and 42Hz,
and then averaged across these two frequencies. As both EMG activity
and gamma-band EEG activity have broad spectra, we reasoned that
such activities should be equally evident at those neighboring
frequencies, thus providing an appropriate check for the possibility
that such activities might underlie the 40Hz effect at posterior sites.
Average (log-transformed) ASSR (40Hz) amplitudes and corre-
sponding mean 38–42Hz amplitudes are presented in Fig. 8, as a
function of attention condition. The ASSR amplitude is slightly larger
thanthe38/42Hzcontrol valueatthe focused-attention conditions, but
this difference is substantially increased at the divided-attention
conditions and is maximal at the 50–50% condition. The relatively
small attention effects on the 38/42Hz control amplitude that can be
seen might reflect a modest contribution of EMG or gamma-band
activity to the attentional effects on ASSR amplitude, but might also
reflect leakage of 40Hz activity to neighboring frequencies due to
nonlinear neural interactions. Repeated-measures ANOVA with com-
ponent (40Hz, 38/42Hz) and attention condition as within-subject
factors yielded main effects of component, F(1,8)=14.0, p<.01, and
attention condition, F(4,32)=7.3, p<.005, and a significant interaction
of these factors, F(4,32)=5.4, p<.01. Polynomial contrast analysis
revealed that the latter effects involving attention condition were
almost completely explained by the quadratic contrast (F(1,8)=13.8,
p<.01, for the main effect of attention condition, and F(1,8)=11.6,
p<.01, for the interaction). Analysis of simple effects showed that the
effect of attention was significant for the 40 Hz component (p<.001),
but not for the 38/42Hz component (p>.15). These results seem to
provide compelling evidence that the enhanced presence of 40Hz
activity at posterior sites in divided-attention conditions represents
genuine neural activity at occipital sites that is specifically associated
with the auditory input stream.
Inspired by this latter effect, we also undertook several analyses
aimed towards identifying possible attention-dependent VSSR
(24Hz) activity at frontocentral electrodes, which could not be
explained in terms of volume conduction from the primary occipital
sources of the VSSR. None of these produced even a hint of a possible
presence of such effects.
Fig. 7. Topographic maps of the auditory steady-state response (40Hz). The upper panel depicts separate maps for each attention condition in the detection task; the lower panel
does so for the discrimination task. Higher values represent higher average normalized amplitudes across participants at that location. Note the emergence of the 40Hz response at
posterior locations during the 50–50% divided-attention condition.
Fig. 8. Mean log-transformed amplitude of the 40Hz response at posterior occipital
locations, as a function of attention condition. The amplitude of the 38/42Hz response
serves as an important control to assess possible influences of EMG artifacts and
gamma-band EEG activity. See text for further explanation.
9
R. de Jong et al. / International Journal of Psychophysiology xxx (2009) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: de Jong, R., et al., Dynamic crossmodal links revealed by steady-state responses in auditory–visual divided
attention, International Journal of Psychophysiology (2009), doi:10.1016/j.ijpsycho.2009.09.013

Page 10

4. Discussion
In the present study we explored the effects of bimodal divided
attention on performance during auditory–visual detection and
discrimination, using a paradigm similar to that developed by Bonnel
and Hafter (1998). The behavioral results showed clear dual-task
costs when attention had to be divided between the auditory and
visual modalities for discrimination performance and also, though the
costs were substantially smaller, for detection performance. These
results replicate earlier results by Bonnel et al. (1998) who, as we
argued before, also seem to have obtained divided-attention costs in
detection performance when, as in the present study, target intensity
was adjusted in such a way that overall performance levels were
approximately equated for the focused-attention conditions in the
detection and discrimination tasks. Thus, the present evidence fails to
support the notion that auditory–visual bimodal detection and
discrimination are mediated by qualitatively different processing
stages, a pre-attentive capacity free stage in detection and an
attention-dependent capacity limited stage in discrimination (Bonnel
et al., 1992; Bonnel and Hafter, 1998). Instead, the evidence seems
more consistent with the notion that bimodal detection is in fact
subject to capacity limitations but that these limitations are generally
less severe than those delimiting dual-task performance with
auditory–visual discrimination tasks.
Results from a study by Kawahara et al. (2001), who used an
attentional-blink paradigm,further substantiate this view. They found
an attentional blink (i.e., impaired detection of the second target
when it followed the first target within 100 to 500ms — see Raymond
et al., 1992, for an introduction to the paradigm) when the second
target had to be identified but not when it had to be detected — an
attentional blink in detection performance was obtained only when
the stimuli were highly degraded. These results also suggest that
detection is subject to capacity limitations but that such restrictions
become manifest in performance only when detection is sufficiently
difficult in terms of signal-to-noise ratios.
The fact that divided-attention costs were much smaller for
detection than for discrimination, even when performance in the two
tasks was equated in the focused-attention conditions, might be
explained in terms of a steeper performance-resource tradeoff for
discrimination than for detection (Norman and Bobrow, 1975). In the
present study we sought to address an alternative, but compatible,
account — that people may tend to adopt less differentiating attention
allocation policies in detection conditions, perhaps because they
become aware that different allocation policies may impact perfor-
mance in concurrent detection less than that in concurrent discrimina-
tion. We will discuss the results for the various electrophysiological
measuresindetailbelow,butthefactthatattentionaleffectsonallthese
measures were remarkably similar for detection and discrimination,
fails to provide any support for this alternative account.
4.1. Event-related potentials
Because the various early-onset ERP components were either
absent in the present study or too small to allow for useful analysis,
presumably due to the low salience of the targets, only the P300
component could be considered. In the detection task, mean P300
amplitude, which was computed across correctly detected targets
only, was larger for visual-only targets than for auditory-only targets.
More important, P300 amplitude on single-target trials was found to
scale systematically with instructed attention allocation — as more
attention was allocated to one modality, P300 amplitude increased for
targets in that modality while it decreased for targets in the other
modality. These results are in agreement with findings by previous
studies that also showed an increase of P300 amplitude as a function
of attention allocation (e.g. Mangun and Hillyard, 1990; Hoffman et
al., 1985, Wickens et al., 1983; Isreal et al., 1980). Such results have
been interpreted in terms of P300 amplitude reflecting the resource
distribution of a limited capacity processing system (Hoffman et al.,
1985). However, others have argued that P300 amplitude is closely
tied to the behavioral outcome, and might reflect the effects of
attentional selection rather than the attentional modulation and
selection of relevant information itself, with the latter causal
processes being reflected by earlier ERP components such as the P1
and N1 (Mangun and Hillyard, 1990; Toffanin et al., 2009). Consistent
with this latter notion, the effects of relative attention allocation on
P300 amplitude were distinctly nonlinear, with relatively large and
similar amplitudes in the 100%, 80%, and 50% conditions, a smaller
amplitude in the 20% condition, and an essentially absent P300 in the
0% condition. Thus, it is questionable whether the attentional effects
on P300 amplitude can provide any more direct information on the
allocation of attention across the two modalities than what is
provided by effects on overt behavior. Interestingly, P300 amplitude
for bimodal targets closely equaled the summed amplitudes for
unimodal targets — this might suggest that at the level of processing
indexedby theP300, there is little or no overheadcost associated with
the requirement to divide attention in bimodal detection (Bonnel et
al., 1992; Bonnel and Hafter, 1998).
In the discrimination task, targets in the different modalities were
always presented simultaneously. Consistent with the larger ampli-
tudes for visual-only targets than for auditory-only targets in the
detection condition, P300 amplitude in the discrimination condition
steadily increased as more attention was allocated to the visual
modality. When compared to the P300 amplitudes for bimodal targets
in the detection condition, amplitudes in the discrimination condition
were somewhat smaller. As bimodal targets were far less common in
the detection condition, this latter effect may reflect the well-known
effect of probability on P300 amplitude (Johnson, 1984, 1986).
4.2. Frequency tagging: ASSR and VSSR
Several aspects of the results from frequency tagging were
unexpected and, we suggest, quite remarkable. First, ASSR amplitude
(40Hz) at frontocentral electrodes, where the amplitude exhibited a
clear maximum,showed no significant effects of intermodalattention.
The lack of a difference between the focused-auditory and focused-
visual conditions replicates earlier findings (Linden et al., 1987) and
reinforces the notion that such a difference may be found only when
the target event is defined in terms of a change in the modulation
frequency itself (Ross et al., 2004a, 2004b; Saupe et al., 2009; Skosnik
et al., 2007). ASSR amplitude fluctuates from trial to trial. Such
fluctuations were found to predict performance accuracy, which high
ASSR amplitude being associated with accurate performance with
auditory targets but with inaccurate performance with visual targets.
Though these results are only correlational, it is tempting to infer a
causal relationship where ASSR amplitude indexes relative prepared-
ness to process auditory as compared to visual information. The
contrast with our failure to find effects of intermodal attention
instructions on ASSR amplitude may be more apparent than real. The
correct–incorrect distinction may simply provide a more powerful
contrast than that provided by the various attention conditions – after
all, different attention instructions had significant but relatively mild
effects on performance accuracy. Furthermore, even the correct–
incorrect contrast failed to produce significant effects on accuracy
when visual targets or auditory targets were considered separately.
The combined results may point to a subtle, but theoretically
significant, relation between (EEG-based) ASSR amplitude and
relative processing bias towards auditory versus visual sources of
information, which requires considerable power to detect. This
perspective makes it understandable why and how such power is
enhanced when the target is defined in terms of a change in the
tagging frequency itself Ross et al. (2004a,b)). Our only reservation
with this solution to boost power is that it would seem to do away
10
R. de Jong et al. / International Journal of Psychophysiology xxx (2009) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: de Jong, R., et al., Dynamic crossmodal links revealed by steady-state responses in auditory–visual divided
attention, International Journal of Psychophysiology (2009), doi:10.1016/j.ijpsycho.2009.09.013

Page 11

with the very elegance and promise of frequency tagging as a tech-
nique that can provide a continuous window on perceptual
information processing in the brain without influencing or interfering
with such processing.
Clear attention effects on ASSR amplitude were found at posterior
occipital sites, where the 40Hz response was markedly stronger in
divided-attention as compared to focused-attention conditions. It is
important to note that these effects were very small in absolute size,
approaching, but not exceeding, the resolution limits of our EEG
recording system. These effects are unlikely to reflect either EMG
artifacts or gamma-band EEG activity, as they were not found, or were
much smaller, at neighboring frequencies (38/42Hz). Neither can
they be due to passive volume conduction from primary ASSR sources
in auditory cortex, as no significant attention effects were found there.
By exclusionthen, as the 40Hz frequency specifically tags theauditory
input stream, the enhanced 40Hz response at occipital sites in
divided-attention conditions must be taken to reflect an enhanced
influence of auditory input or auditory processing on neural activity in
occipitalcortex.To ourknowledge,thisis thefirst demonstrationofan
unambiguous attention-dependent influence of auditory information
on occipital activity using EEG (Talsma et al., 2007, may have
demonstrated such an influence of visual information on auditory
regions). The nature of such an enhanced influence in divided-
attention conditions remains to be clarified. The fact that ASSR
amplitude at anterior-central sites showed no attention effects would
seem to rule out the possibility that the effects at posterior sites are
due to a stronger neural drive (as opposed to passive volume
conduction) from auditory regions, presumably primary auditory
cortex, in divided-attention conditions. Rather, these effects might
reflect stronger effective connectivity from auditory to visual cortical
areas in divided-attention conditions. That issue will be discussed in
more detail later.
VSSR amplitude (24 Hz) exhibited a clear maximum at posterior
occipital sites, with a highly favorable signal-to-noise ratio. Attention
effects on posterior VSSR amplitude resembled those found for ASSR
amplitude at posterior sites, with higher amplitudes obtained in
divided-attention conditions as compared to focused-attention con-
ditions and with both effects being fully accounted for by the
quadratic contrast in polynomial contrast analyses. We stress that
almost identical attention effects were obtained when VSSR ampli-
tudewascomputed in away thatrendereditcompletelyinsensitiveto
possible attention-related differences in intertrial phase coherence
(see Method). Thus, these effects on VSSR amplitude almost certainly
reflect genuine effects on sensory gain control (Kim et al., 2007).
VSSR amplitude in the focused-visual condition did not differ from
that in the focused-auditory condition. This lack of an effect of
intermodal selective attention contrasts strongly with the many
positive findings of clear intramodal visuospatial attention effects on
VSSR amplitude, also in studies that used tagging frequencies similar
to the 24Hz used in the present study (Muller and Hillyard, 2000;
Muller et al. 1998a,b; Morgan et al., 1996; Toffanin et al., 2009). The
latter findings have generally been interpreted in terms of modulation
of sensory gain control in primary or secondary visual cortex, and this
interpretation is corroborated by evidence from numerous fMRI
studies of visuospatial attention (Kastner et al., 1998; Russo et al.,
2003; Tootell et al., 1998). Several recent fMRI studies of auditory–
visual intermodal attention have also found evidence for enhanced
activity in modality-specific cortical areas associated with the
attended modality in conjunction with diminished activity in cortical
areas associated with the unattended modality (for a recent example
and overview, see Johnson and Zatorre, 2005, 2006). Importantly,
these cortical areasare typically found nottoincludeprimaryauditory
and visual cortex (e.g., Johnson and Zatorre, 2005, 2006, Shomstein
and Yantis, 2004). As occipital VSSR has been associated with sources
in primary visual cortex (Russo et al., 2007), the fact that evidence
from fMRI studies indicates that effects of selective intermodal
attention are largely restricted to secondary or higher visual cortical
areas, may thus be consistent with the lack of selective intermodal
attention on VSSR amplitude in the present study.3
As noted before, the enhanced VSSR amplitude in divided-attention
conditionsresemblestheresultsforASSRamplitude(40 Hz)atoccipital
sites, and we suggest that both effects may share a common cause:
Enhanced effective connectivity from auditory to visual cortex in
divided-attention conditions so that activity in auditory cortex can
affect, directly or indirectly, activity in visual cortex during crossmodal
auditory–visual attention. The central idea here is that enhanced 40 Hz
activity at occipital sites serves the critical function of unambiguously
indexing such enhanced effective connectivity, but that 40 Hz activity
itself constitutes only a small portion of total neural activity in auditory
cortexthat,onceeffectiveconnectivityhasbeenestablished,can exerta
more general influence on neural activity levels and patterns in visual
cortex. If VSSR amplitude is assumed to reflect sensory gain control in
primary visual cortex (Russo et al., 2007), the present findings would
seem to provide strong evidence for an influence of auditory
information processing on early cortical stages of visual information
processing. Possibly related small but reliable changes in occipital
activity due to auditory-spatial attention were recently reported by Wu
et al. (2007). More speculatively, such an influence may have been
responsible for results by van der Burg et al. (2008) who demonstrated
that a nonspecific auditory beep could critically facilitate the detection
of low-salience visual targets. Finally, such an influence on sensory gain
in visual cortex may have been responsible for the finding of Talsma
etal.(2006)thatattentionalcapacitybetweenmodalities(auditoryand
visual) is larger than that within the visual modality alone.
An alternative interpretation of the enhanced VSSR amplitude in
divided-attention conditions, is that VSSR amplitude depends on
general arousal level and that this level was higher in the presumably
more challenging divided-attention conditions. One problem with
this interpretation is that it fails to explain why, as discussed earlier,
VSSR amplitude is not enhanced in divided-attention conditions in
intramodal studies. Also, without a specification of the neural systems
and mechanisms underlying possible effects of general arousal on
VSSR amplitude, this interpretation may be difficult to distinguish
from interpretations that account for crossmodal effects in terms of
neural interactions at midbrain structures (e.g. thalamus) that
influence cortical processing (Driver and Noesselt, 2008). Such a
distinction could perhaps be made in future studies by manipulating
the relative positionsof the auditory and visual sources, as crossmodal
interactions are typically dependent on close spatial vicinity of
sources (Driver and Noesselt, 2008), whereas the difficulty of
divided-attention conditions, and thus general arousal level, should
instead be greater when sources are spatially separate. For now, we
prefer the account in terms of enhanced effective connectivity over
that in terms of general arousal as the former but not the latter can
also readily account for the presence of the ASSR at occipital sites in
divided attention conditions.
In the present study, we failed to obtain any evidence for a reverse
attention-dependent influence from visual information processing on
auditory information processing. This might point to a truly
asymmetric interaction between vision and audition under the
conditions of the present experiment, or it could be due to limitations
of the measures and techniques we used to find evidence for such a
reverse influence. Seemingly consistent with the former possibility,
Giard and Peronnet (1999), in what could arguably be considered an
auditory–visual divided-attention task, found influences of auditory
information on occipital activity to be both earlier (∼50 ms) and
stronger than the reverse influences.
3Some previous studies using the frequency tagging technique have found the
second harmonic response to be particularly sensitive to attentional manipulations
(e.g., Pastor et al., 2007). For this reason we also checked for possible effects of
selective intermodal attention at 48 Hz — no such effects were found.
11
R. de Jong et al. / International Journal of Psychophysiology xxx (2009) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: de Jong, R., et al., Dynamic crossmodal links revealed by steady-state responses in auditory–visual divided
attention, International Journal of Psychophysiology (2009), doi:10.1016/j.ijpsycho.2009.09.013

Page 12

It is important to point out that the effects of intermodal attention
on VSSRamplitudewereverysimilar fordetectionanddiscrimination.
Apart from serving as a useful indication of the robustness of the
present results, this similarity also serves to indicate that attentional
strategies and mechanisms are likely to have been largely the same in
detection and discrimination. Attention-dependent crossmodal links
of the type demonstrated here might have important functions in
supporting flexible multisensory integration. The tasks we used in the
present study, with target events in one modality being independent
of and unrelated to those in the other modality, provided no basis for
useful multisensory integration, begging the question of whether and
how such links may have influenced task performance. We have
interpreted VSSR amplitude to reflect sensory gain at low levels of
cortical visual processing, but the positive functional connotations of
the concept of sensory gain may be reason to reconsider that
interpretation. As noted earlier, recent evidence tends to support
the hypothesis that the ASSR reflects a separate neural oscillation that
is driven by the periodic stimulus (Ross et al., 2005a). If we apply the
same hypothesis to the VSSR, it becomes possible to see how input
from auditory processing regions, even when it carries no useful
information – and thus should be considered noise – may raise the
level of background activity in early visual cortex and thereby
enhance VSSR amplitude as background activity is entrained by, or
feeds, the neural oscillation. However, a higher level of background
activity, or noise, would not necessarily facilitate the processing of
discrete visual events, and would perhaps be more likely to actually
hamper such processing. Thus, attention-dependent crossmodal links,
by enabling crosstalk, may have contributed to divided-attention
costs in performance in the present study (Navon and Miller, 2002).
The more important and general point is that tasks, such as the ones
used here, that provide little or no possibilities for useful multisensory
integration, may nevertheless be powerful tools to establish potential
neural mechanisms underlying such integration. To assess the
functional significance of such mechanisms requires the complemen-
tary use of tasks or paradigms that allow for useful types and degrees
of multisensory integration.
4.3. Multisensory integration: ERPs and SSRs
The SSRis easily quantified inthe frequencydomain and knowledge
aboutthe neuralgenerators of the ASSRand VSSR greatly enhance their
utilityin assessingneural activity in specific cortical regions. Twomajor
sources of theVSSRhave been identified in primary visual cortex and in
the motion-sensitive area MT/V5 — in addition,activity in mid-occipital
and ventral occipital region appear to make minor contributions to the
VSSR (Russo et al., 2007). Together with the posterior occipital
distribution of theVSSR in the present study, this has led us to interpret
VSSRamplitudetoindex neuralactivityin earlyregionsof visualcortex.
While several studies have identified primary auditory cortex as the
main source of the ASSR, a recent study analyzed sources by means of
PET and found six sources of the ASSR, including temporal, frontal, and
parietal regions as well as regions in the brain stem and cerebellum
(Reyesetal.,2005).Giventheevidence,itseemsreasonabletoassociate
ASSR amplitude primarily with neural activity in early regions of au-
ditory cortex.
The regional specificity of ASSR and VSSR in conjunction with the
fact that their temporal resolution is inherently poor, imply that these
measures do not allow for a chronometric analysis of brain activity
across different cortical areas. Several possible mechanisms underly-
ing crossmodal influences have been proposed, including rapid feed-
forward integration, thalamic influences, direct connections between
primary sensory cortices, and feedback from multisensory regions to
sensory-specific brain areas (Driver and Noesselt, 2008). We believe
that the present results, while providing clear evidence for attention-
dependent crossmodal influences, are mute regarding the issue of
underlying mechanisms, because of this inherent limitation of the SSR
measure. In contrast, the high temporal resolution of ERPs combined
with the high spatial resolution of fMRI, can provide a picture of the
sequential activation of cortical areas during perceptual processing.
Some interesting results that demonstrate the potential of this
technique and speak to underlying neural mechanisms are already
available, for both intramodal attention (e.g., Russo et al., 2003) and
intermodal attention (e.g., McDonald et al., 2003; see Driver and
Noesselt, 2008, and Macaluso and Driver, 2005, for a more extensive
overview). The present results indicate that bimodal frequency
tagging may amply compensate for the loss of temporal resolution
by providing great sensitivity and power to detect tonic and subtle
changes in regionally specific background neural activity related to
crossmodal influences. Frequency tagging and evoked potentials thus
seem to represent somewhat complementary techniques, which
probably can be used profitably as such in EEG/MEG research on
intermodal attention and crossmodal interactions.
References
Bidet-Caulet, A., Fischer, C., Besle, J., Aguera, P.-E., Giard, M.H., Bertrand, O., 2007. Effects
of selective attention on the electrophysiological representation of concurrent
sounds in the human auditory cortex. The Journal of Neuroscience 27, 9252–9261.
Bonnel, A.M., Hafter, E.R., 1998. Divided attention between simultaneous auditory and
visual signals. Perception and Psychophysics 60, 179–190.
Bonnel, A.M., Stein, J.F., Bertucci, P., 1992. Does attention modulate the perception of
luminance changes? Quarterly Journal of Experimental Psychology A 44, 601–626.
Bohorquez, J., Ozdamar, O., 2008. Generation of the 40-Hz auditory steady-state
response (ASSR) explained using convolution. Clinical Neurophysiology 119,
2598–2607.
Brainard, D.H., 1997. The psychophysics toolbox. Spatial Vision 10, 433–436.
Brawn, P.T., Snowden, R.J., 2000. Attention to overlapping objects: detection and
discrimination of luminance changes. Journal of Experimental Psychology: Human
Perception and Performance 26, 342–358.
Deco, G., Zihl, J., 2006. The neurodynamics of visual search. Visual Cognition 14,
1006–1024.
Desimone, R., Duncan, J., 1995. Neural mechanisms of selective visual attention. Annual
Review Neuroscience 18, 193–222.
Ding, J., Sperling, G., Srinivasan,R., 2006. Attentional modulation of SSVEP power depends
on the network tagged by the flicker frequency. Cerebral Cortex 16, 1016–1029.
Downing, C.J., 1988. Expectancy and visual-spatial attention: effects on perceptual
quality. Journal of Experimental Psychology Human Perception and Performance 14,
188–202.
Driver, J., Noesselt, T., 2008. Multisensory interplay reveals crossmodal influences on
‘sensory-specific’ brain regions, neural responses, and judgments. Neuron 57,
11–23.
Duncan, J., 1980. The locus of interference in the perception of simultaneous stimuli.
Psychological Review 87, 272–300.
Duncan, J., 1981. Directing attention in the visual field. Perception and Psychophysics 30,
90–93.
Eimer, M., Schroger, E., 1998. ERP effects of intermodal attention and cross-modal links
in spatial attention. Psychophysiology 35, 313–327.
Fell, J., Fernndez, G., Klaver, P., Elger, C.E., Fries, P., 2003. Is synchronized neuronal
gamma activity relevant for selective attention? Brain Research: Brain Research
Review 42, 265–272.
Galambos, R., Makeig, S., Talmacho, P.J., 1981. A 40-Hz auditory potential recorded from
the human scalp. Proceedings of the National Academy of Sciences U.S.A. 78,
2643–2647.
Gander, P.E., Bosnyak, R., Wolek, R., Roberts, L.E., 2007. Modulation of the 40-Hz
auditory steady-state response by attention during acoustic training. International
congress series 1300, 37–40.
Ghazanfar, A.A., Schroeder, C.E., 2006. Is neocortex essentially multisensory? Trends in
Cognitive Sciences 10, 278–285.
Giard, M.H., Fort, A., Mouchetant-Rostaing, Y., Pernier, J., 2000. Neurophysiological
mechanisms of auditory selective attention in humans. Frontiers in Bioscience 5,
D84–D94.
Giard, M.H., Peronnet, F., 1999. Auditory–visual integration during multimodal object
recognition in humans: a behavioral and electrophysiological study. Journal of
Cognitive Neuroscience 11, 473–490.
Gladwin, T.E., Lindsen, J.P., de Jong, R., 2006. Pre-stimulus EEG effects related to
response speed, task switching and upcoming response hand. Biological Psychol-
ogy 72, 15–34.
Gratton, G., Coles, M.G., Donchin, E., 1983. A new method for off-line removal of ocular
artifact. Electroencephalography and Clinical Neurophysiology 55, 468–484.
Gutschalk, A., Mase, R., Roth, R., Ille, N., Rupp, A., Hahnel, S., Picton, T.W., Scherg, M.,
1999. Deconvolution of 40Hz steady-state fields reveals two overlapping source
activities of the human auditory cortex. Clinical Neurophysiology 110, 856–868.
Hawkins, H.L., Hillyard, S.A., Luck, S.J., Mouloua, M., Downing, C.J., Woodward, D.P.,
1990. Visual attention modulates signal detectability. Journal of Experimental
Psychology: Human Perception and Performance 16, 802–811.
12
R. de Jong et al. / International Journal of Psychophysiology xxx (2009) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: de Jong, R., et al., Dynamic crossmodal links revealed by steady-state responses in auditory–visual divided
attention, International Journal of Psychophysiology (2009), doi:10.1016/j.ijpsycho.2009.09.013

Page 13

Herrmann, C.S., 2001. Human EEG responses to 1–100 Hz flicker: Resonance
phenomena in visual cortex and their potential correlation to cognitive phenom-
ena. Experimental Brain Research 137, 346–353.
Hillyard, S.A., Teder-Salejarvi, W.A., Munte, T.F., 1998. Temporal dynamics of early
perceptual processing. Current Opinion in Neurobiology 8, 202–210.
Hoffman, J.E., 1979. A two-stage model of visual search.Perception and Psychophysics 25,
319–327.
Hoffman, J.E., Houck, M.R., MacMillan, F.W., Simons, R.F., Oatman, L.C., 1985. Event-
related potentials elicited by automatic targets: a dual-task analysis. Journal of
Experimental Psychology: Human Perception and Performance 11, 50–61.
Isreal, J.B., Chesney, G.L., Wickens, C.D., Donchin, E., 1980. P300 and tracking difficulty:
evidence for multiple resources in dual-task performance. Psychophysiology 17,
259–273.
Johnson, J.A., Zatorre, R.J., 2005. Attention to simultaneous unrelated auditory and
visual events: behavioral and neural correlates. Cerebral Cortex 15, 1609–1620.
Johnson, J.A., Zatorre, R.J., 2006. Neural substrates for dividing and focusing attention
between simultaneous auditory and visual events. Neuroimage 31, 1673–1681.
Johnson, R., 1984. P300: A model of the variables controlling its amplitude. Annals New
York Academy of Sciences 425, 223–229.
Johnson, R., 1986. A triarchic model of p300 amplitude. Psychophysiology 23, 367–384.
Kastner, S., Weerd, P.D., Desimone, R., Ungerleider, L.G., 1998. Mechanisms of directed
attention inthehumanextrastriate cortexas revealedbyfunctionalMRI.Science 282,
108–111.
Kawahara, J., Lollo, V.D., Enns, J.T., 2001. Attentional requirements is visual detection
and identification: evidence from the attentional blink. Journal of Experimental
Psychology: Human Perception and Performance 27, 969–984.
Kim, Y.J., Grabowecky, M., Paller, K.A., Muthu, K., Suzuki, S., 2007. Attention induces
synchronization-based response gain in steady-state visual evoked potentials.
Nature Neuroscience 10, 117125.
Lappin, J.S., Uttal, W.R., 1976. Does prior knowledge facilitate the detection of visual
targets in random noise? Perception and Psychophysics 20, 367–374.
Linden, R.D., Picton, T.W., Hamel, G., Campbell, K.B., 1987. Human auditory steady-state
evoked potentials during selective attention. Electroencephalography and Clinical
Neurophysiology 66, 145–159.
Macaluso, E., Driver, J., 2005. Multisensory spatial interactions: a window onto
functional integration in the human brain. Trends in Neurosciences 28, 264–271.
Macmillan, N.A., Creelman, C.D., 2005. Detection Theory: A User's Guide, 2nd ed.
Lawrence Erlbaum Associates, Mahwah, N.J.
Malinowski, P., Fuchs, S., Muller, M.M., 2007. Sustained division of spatial attention to
multiple locations within one hemifield. Neuroscience Letters 414, 65–70.
Mangun, G.R., Hillyard, S.A., 1990. Allocation of visual attention to spatial locations:
tradeoff functions for event-related brain potentials and detection performance.
Perception and Psychophysics 47, 532–550.
McCarthy, G., Wood, C.C., 1985. Scalp distributions of event-related potentials: an
ambiguity associated with analysis of variance models. Electroencephalography
Clinical Neurophysiology 62, 203–208.
McDonald, J.J., Teder-Salejarvi, W.A., Russo, F.D., Hillyard, S.A., 2003. Neural substrates
of perceptual enhancement by cross-modal spatial attention. Journal of Cognitive
Neuroscience 15, 10–19.
Morgan, S.T., Hansen, J.C., Hillyard, S.A., 1996. Selective attention to stimulus location
modulates the steady-state visual evoked potential. Proceedings of the National
Academy of Sciences U.S.A. 93, 4770–4774.
Muller, H.J., Humphreys, G.W., 1991. Luminance-increment detection: capacity-limited
or not? Journal of Experimental Psychology: Human Perception and Performance 17,
107–124.
Muller, M.M., Hillyard, S., 2000. Concurrent recording of steady-state and transient
event-related potentials as indices of visual–spatial selective attention. Clinical
Neurophysiology 111, 1544–1552.
Muller, M.M., Picton, T.W., Valdes-Sosa, P., Riera, J., Teder-Salejarvi, W.A., Hillyard, S.A.,
1998a.Effectsofspatialselectiveattentiononthesteady-statevisualevokedpotential
in the 20–28Hz range. Brain Research: Cognitive Brain Research 6, 249–261.
Muller, M.M., Teder-Salejarvi, W., Hillyard, S.A., 1998b. The time course of cortical
facilitation during cued shifts of spatial attention. Nature Neuroscience 1, 631–634.
Navon, D., 1985. Attention and performance XI. In: Posner, I., Marin, O.S.M. (Eds.),
Erlbaum, Hillsdale, NJ, pp. 133–146.
Navon, D., Miller, J., 2002. Queuing or sharing? a critical evaluation of the single-
bottleneck notion. Cognitive Psychology 44, 193–251.
Norman, D., Bobrow, D., 1975. On data-limited and resource-limited processing. Journal
of Cognitive Psychology 7, 44–60.
Pastor, M.A., Valencia, M., Artieda, J., Alegre, M., Masdeu, J.C., 2007. Topography of
cortical activation differs for fundamental and harmonic frequencies of the steady-
state visual-evoked responses. An EEG and PET H2O15study. Cerebral Cortex 17,
1899–1905.
Pei, F., Pettet, M.W., Norcia, A.M., 2002. Neural correlates of object-based attention.
Journal of Vision 2, 588–596.
Pelli, D.G., 1997. The videotoolbox software for visual psychophysics: transforming
numbers into movies. Spatial Vision 10, 437–442.
Raymond,J.E.,Shapiro,K.L.,Arnell,K.M.,1992.Temporarysuppressionofvisualprocessing
in an RSVP task: an attentional blink? Journal of Experimental Psychology Human
Perception and Performance 18, 849–860.
Regan, D., 1989. Human Brain Electrophysiology: Evoked Potentials and Evoked
Magnetic Fields in Science and Medicine. Elsevier, Amsterdam.
Reyes, S.A., Lockwood, A.H., Salvi, R.J., Coad, M.L., Wack, D.S., Burkard, R.F., 2005. Mapping
the 40-Hz auditory steady-state response using current density reconstructions.
Hearing Research 204, 1–15.
Ross, B., Herdman, A.T., Pantev, C., 2005a. Right hemispheric laterality of human 40Hz
auditory steady-state responses. Cerebral Cortex 15, 2029–2039.
Ross, B., Herdman, A.T., Pantev, C., 2005b. Stimulus induced desynchronization of
human auditory 40-Hz steady-state responses. Journal of Neurophysiology 94,
4082–4093.
Ross, B., Herdman, A.T., Wollbrink, A., Pantev, C., 2004a. Auditory cortex responses to
the transition from monophonic to pseudo-stereo sound. Neurology and Clinical
Neurophysiology 2004, 18.
Ross, B., Picton, T.W., Herdman, A.T., Pantev, C., 2004b. The effect of attention on the
auditory steady-state response. Neurology and Clinical Neurophysiology 2004, 22.
Ross, B., Picton, T.W., Pantev, C., 2002. Temporal integration in the human auditory
cortex as represented by the development of the steady-state magnetic field.
Hearing Research 165, 68–84.
Russo, F.D., Martinez, A., Hillyard, S.A., 2003. Source analysis of event-related cortical
activity during visuo-spatial attention. Cerebral Cortex 13, 486–499.
Russo, F.D., Pitzalis, S., Aprile, T., Spitoni, G., Patria, F., Stella, A., Spinelli, D., Hillyard, S.A.,
2007. Spatiotemporal analysis of the cortical sources of the steady-state visual
evoked potential. Human Brain Mapping 28, 323–334.
Saupe, K., Widmann, A., Bendixen, A., Muller, M.M., Schroger, E., 2009. Effects of
intermodal attention on the auditory steady-state response and the event-related
potential. Psychophysiology 46 (2), 321–327 2009.
Sharbrough, F., Chatrian, G., Lesser, R., Luders, H., Nuwer, M., Picton, T., 1991. American
Electroencephalographic Society guidelines for standard electrode position
nomenclature. Journal of Clinical Neurophysiology 8, 200–202.
Shomstein, S., Yantis, S., 2004. Control of attention shifts between vision and audition in
human cortex. Journal of Neuroscience 24, 10702–10706.
Simpson, M.I.G., Hadjipapas, A., Barnes, G.R., Furlong, P.L., Witton, C., 2005. Imaging the
dynamics of the auditory steady-state evoked response. Neuroscience Letters 385,
195–197.
Skosnik, P.D., Krishnan, G.P., O'Donnell, B.F., 2007. The effect of selective attention on the
gamma-band auditory steady-state response. Neuroscience Letters 420, 223–228.
Sperling, G., Melcher, M.J., 1978. The attention operating characteristic: examples form
visual search. Science 202, 315–318.
Stein, B.E., Stanford, T.R., 2008. Multisensory integration: current issues from the
perspective of the single neuron. Nature Review Neuroscience 9, 255–266.
Tallon-Baudry, Bertrand, 1999. Oscillatory gamma activity in humans and its role in
object representation. Trends in Cognitive Sciences 3, 151–162.
Talsma, D., Doty, T.J., Strowd, R.E., Woldorff, M.G., 2006. Attentional capacity for
processing concurrent stimuli is larger across sensory modalities than within a
modality. Psychophysiology 43, 541–549.
Talsma, D., Doty, T.J., Woldorff, M.G., 2007. Selective attention and audiovisual
integration: is attending to both modalities a prerequisite for early integration?
Cerebral Cortex 17, 679–690.
Talsma, D., Kok, A., 2001. Nonspatial intermodal selective attention is mediated by
sensory brain areas: evidence from event-related potentials. Psychophysiology 38,
736–751.
Talsma, D., Woldorff, M.G., 2005. Selective attention and multisensory integration:
multiple phases of effects on the evoked brain activity. Journal of Cognitive
Neuroscience 17, 1098–1114.
Teder-Salejarvi, W.A., McDonald, J.J., Russo, F.D., Hillyard, S.A., 2002. An analysis of
audio-visual crossmodal integration by means of event-related potential (ERP)
recordings. Brain Research: Cognitive Brain Research 14, 106–114.
Toffanin, P., de Jong, R., Johnson, A., Martens, S., 2009. Using frequency tagging to
quantify attentional deployment in a visual divided attention task. International
Journal of Psychophysiology 72, 289–298.
Tootell, R.B., Hadjikhani, N.K., Vandu_el, W., Liu, A.K., Mendola, J.D., Sereno, M.I., et al.,
1998. Functional analysis of primary visual cortex (V1) in humans. Proceedings of
the National Academy of Sciences U.S.A. 95, 811–817.
Van der Burg, E., Olivers, C.N.L., Bronkhorst, A.W., Theeuwes, J., 2008. Pip and pop: non-
spatial auditory signals improve spatial visual search. Journal of Experimental
Psychology: Human Perception and Performance 34, 1053–1065.
Wickens, C., Kramer, A., Vanasse, L., Donchin, E., 1983. Performance of concurrent tasks:
a psychophysiological analysis of the reciprocity of information-processing
resources. Science 221, 1080–1082.
Wu, C.T., Weissman, D.H., Roberts, K.C., Woldorff, M.G., 2007. The neural circuitry of the
executive control of auditory spatial attention. Brain Research 1134, 187–198.
13
R. de Jong et al. / International Journal of Psychophysiology xxx (2009) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: de Jong, R., et al., Dynamic crossmodal links revealed by steady-state responses in auditory–visual divided
attention, International Journal of Psychophysiology (2009), doi:10.1016/j.ijpsycho.2009.09.013