task-specific sensory and motor preparatory activation revealed by contingent magnetic variation
TRANSCRIPT
www.elsevier.com/locate/cogbrainres
Cognitive Brain Research 21 (2004) 59–68
Research report
Task-specific sensory and motor preparatory activation revealed by
contingent magnetic variation
Carlos M. Gomeza,*, Alberto Fernandezb, Fernando Maestub, Carlos Amob, J.J. Gonzalez-Rosaa,Encarnacion Vaqueroa, Tomas Ortizb
aDpto. Psicologıa Experimental, Facultad de Psicologıa, Avda. San Francisco Javier, s/n. Sevilla 41005, SpainbCentro de Magnetoencefalografıa Dr. Perez-Modrego, Universidad Complutense de Madrid, Madrid, Spain
Accepted 24 May 2004
Available online 4 July 2004
Abstract
The present report studied the magnetic counterpart (CMV) of the auditory contingent negative variation (CNV). The ear where the target
auditory stimulus would be presented was cued with a visual central arrow at a validity of 84%. The subject’s behavioral response and the
magnetoencephalographic (MEG) and electroencephalographic (EEG) signals were recorded. The central cue diminished reaction times
(RTs) to the auditory target in the valid conditions with respect to the invalid conditions, indicating that the attentional manipulation was
effective. The averaged magnetic field power during the preparatory period was significantly higher than baseline, suggesting the
simultaneous presence of a magnetic counterpart of the electric CNV—the CMV. The field maps of the CMV grand averages showed two
different and well-established periods: an early one with a magnetic field distribution that suggests a central source, and a late one with a field
topography comparable to a low-intensity auditory-evoked field (M1). Single-dipole analysis of the preparatory phase in the subject’s
magnetic resonance images (MRI) demonstrated the presence of dipolar activity in the posterior cingulate (PCC) and posterior parietal
cortices (PPC), superior temporal gyrus (STG) and motor cortices (MC). The lateralization of this activity depended on the orientation of the
central cue. These results suggest that the action and perceptual-related areas needed to process the expected subsequent imperative task are
recruited during the preparatory periods, influencing the behavioral RTs.
D 2004 Elsevier B.V. All rights reserved.
Theme: Neural basis of behavior
Topic: Cognition
Keywords: Contingent magnetic variation; Posner’s paradigm; Perceptual attention; Motor intention
1. Introduction location for the imperative stimulus produces a faster re-
In a situation of expectancy induced by a warning stimulus
(S1) providing pertinent information regarding the arrival of
a second stimulus (S2, the so-called imperative stimulus), the
preparation for S2 induced by the warning generates a change
in the cortical activity, which constitutes the contingent
negative variation (CNV) [29,32]. Sometimes S1 may indi-
cate certain characteristics about S2, such as the probable
locations of its appearance. This is the case with Posner’s
paradigm using central cues [27]. The cueing of a probable
0926-6410/$ - see front matter D 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.cogbrainres.2004.05.005
* Corresponding author. Tel.: +34-954-557-800; fax: +34-954-551-784.
E-mail address: [email protected] (C.M. Gomez).
sponse when the cue is valid than with an invalid or neutral
cue. This behavioral outcome implies that some neural
activation must occur during the preparatory period to allow
a faster reaction time (RT) when comparing valid with invalid
or neutral conditions. It should be noted that the directionality
of the cue in the Posner-type paradigm enables a clear
separation of the warning-induced arousal component from
the attentional effects that are specific to the cued location.
It has been proposed that in the classical CNV paradigm,
S1 acts as a warning stimulus that activates areas needed for
the subsequent processing of S2 [3]. How task-specific the
cortical activation during the preparatory period is, and how it
influences the subsequent imperative stimulus, is a subject of
much debate, and is currently under investigation [12–15].
C.M. Gomez et al. / Cognitive Brain Research 21 (2004) 59–6860
The CNV comprises at least two different phases: an
early one related to the S1-stimulus orientation, and a late
one related to preparation for the motor response [22].
Some reports suggest a participation of sites responsible
for sensory processing in the genesis of the CNV [3,13].
The sources for each of these CNV phases have been
investigated with different techniques, such as voltage and
current source maps, dipole localization, low-resolution
tomographic EEG analysis (LORETA); dipole localization
and magnetic field tomography applied to magnetoence-
phalography (MEG), and functional magnetic resonance
imaging (fMRI). A brief account of the results follows.
The early phase of CNV has a bilateral frontal distri-
bution, and is related to the orientation induced by the
warning stimulus [33]. Observations derived from event-
related potentials (ERPs) and lesional studies suggest a
role of the supplementary motor area (SMA) and the
anterior cingular cortex (ACC) as possible sources (gen-
erators) for this early component [5,13,34]. The use of
dipole localization techniques in a variety of tasks has
enabled CNV to be detected in the prefrontal area, but also
in the ACC, in both EEG [11,28] and MEG [1,2,17]
recordings. However, none of the dipole localization stud-
ies specifically correlated the activity of the frontal region
to the early phase of CNV. Recently, using the LORETA
technique, neural activity in the SMA and ACC has been
proposed for the early phase of CNV [15]. fMRI data
further corroborate this finding [21].
A contribution from motor—but also posterior—cortices
to the genesis of the late CNV phase, particularly in relation
to motor preparation, has been reported elsewhere [15].
According to those authors, the late CNV component is
usually contralateral to the hand used for motor responses.
Late CNV has been compared to the Bereitschaftspotential
(BP) [7,20]. Similarly to the BP2 component, the late CNV
phase could also represent a motor preparation process
(reviewed in Rockstroh et al. [29]). In fact, during self-
paced movements, the MEG recordings showed dipolar
sources in the primary motor cortex contralateral to the
movement [8]. However, it must be remarked that the late
phase of CNV also includes the activation of posterior
sensory-related sites [4,15].
On the other hand, the contribution of parietal lobe and
sensory cortices to the CNV has been scarcely reported
[6,10,15,18]. Elbert et al. [9] described a magnetic counter-
part of CNV that could be at least partially generated by the
temporal cortex. Using directional central cues, two studies
have shown slow waves that might represent the preactiva-
tion of sensory cortices, suggesting that the CNV could have
a sensory component [8,16]. Electrophysiological record-
ings of single neurons in animals [23], and fMRI studies in
humans [19], support the activation of frontal—as well as
striate and extrastriate—cortices, during preparatory periods
while visual stimulation is delivered. Based on these obser-
vations, the neural preparatory activity in primary motor and
posterior sites [5,15] could anticipate the activation of these
same areas that are needed for the current processing of the
imperative stimulus. However, the possibility that task-
related specific sensory and motor cortex is activated during
cued tasks has never been addressed using neural localiza-
tion techniques.
From the evidence in the literature we identified two
goals to be studied. First, to obtain a reliable magnetic
counterpart of the CNV component (CMV) in the context
of a modified Posner’s paradigm where a central visual cue
indicates the ear to be stimulated by a target tone. This
experimental design has been chosen to avoid confusion
between long-latency visual ERPs and the sensory CNV–
CMV component. The differences between CMV and noise
level would be measured and statistically evaluated. Sec-
ond, the hypothesis that task-relevant sensory and motor
cortices recruited during the imperative phase of the task
were previously activated during the preparatory period
would be tested by single-dipole localization (MEG) of the
CMV generators in the subject’s MRI. It was anticipated
that the information regarding CMV would shed some
light on the neurocognitive activity during the preparatory
period. How it relates to motor intention and perceptual
attention, as well as their behavioral consequences, will be
discussed. In particular, the possibility that the preparatory
pattern of activation could explain the cost–benefit pattern
of RTs in Posner-type paradigms will be addressed.
2. Materials and methods
2.1. Participants
Six men and three women, all right-handed (with 10/10
scores in the Edinburgh handedness questionnaire [26]) and
aged between 17 and 36 years, participated in the study.
Prior to the experiments, each subject underwent a training
period in which a detailed demonstration of the procedure
and apparatus was carried out. The experiments were con-
ducted with the informed consent of each subject.
2.2. Behavioral paradigm
The visual stimuli (arrows) were projected through an
LCD video-projector (SONY VPL-X600E), situated out-
side a magnetic-shielded room on to a series of in-room
mirrors, the last of which was suspended 100 cm above the
subject’s face. The arrows subtended 1.8j and 3j of
horizontal and vertical visual angle, respectively. The
auditory stimuli were delivered to the subject’s ears
through non-magnetic plastic tubes. The tones reached
the subject’s ears 19 ms after the electrical pulse had been
sent to the speakers. Subjects were asked to fix their eyes
on the center of the screen, and instructed to pay attention
to the ear signaled by a central arrow, and to press the
thumb-button of the non-magnetic response device (4D-
NeuroimagingR) as soon as an auditory target appeared at
C.M. Gomez et al. / Cognitive Brain Research 21 (2004) 59–68 61
either ear. The arrow stimulus was considered the warning
stimulus, and the auditory stimulus the imperative one.
The stimulus presentation was computer-controlled (Stim
system, NeuroscanR). The event sequence within a trial
was as follows: the central arrow pointer was on for 300
ms, followed by a temporal delay to the target of 700 ms;
thus the total preparatory period was 1000 ms. The trial
was ended with a monaural auditory stimulus that the
subject had to consider the imperative stimulus. The
auditory stimulus (1000 Hz, 90 dB) lasted for 100 ms
(20 ms rise and fall time) and was randomly presented to
the left and right ear with an equal probability (0.5). The
intertrial interval was 1.7 s.
Each subject was confronted with a total of 400 trials
divided into eight blocks. The warning stimulus had a
directional information: in half of the trials it pointed to the
right, and in the other half to the left. In 84% of the trials,
the arrow had a valid informative value about the target ear
(valid trials), and in 16% of the trials, the arrow pointed to
the ear opposite to where the auditory stimulus would
appear (invalid trials). Thus, the experiment presented 4
conditions: left valid (LV: 168 trials per subject), right
valid (RV: 168 trials), left invalid (LI: 32 trials), and right
invalid (RI: 32 trials). It should be noted that left/right in
the experimental conditions refers to the lateralization of
the auditory stimulus and not the directionality of the
warning/arrow stimulus. Thus, the LI condition refers to
preparation of the right side, although the actual target
appears in the left ear. The situation for RI is equivalent,
where a left target is expected, but a right target appears.
The subjects had to respond to the monaural auditory
stimulus with the thumb of the compatible hand. They
were informed that the visual cue had an informative value
indicating with high probability the location of the audi-
tory stimulus. RTs and proportion of correct responses
were computed.
2.3. Data collection and analysis
2.3.1. EEG recordings
Simultaneously to MEG recording, EOG, ECG, and
EEG were also collected. Eye movements (EOG) were
recorded from four electrodes attached to the left and right
outer canthus and above and below the left eye. The ECG
was monitored with electrodes attached to the right collar-
bone and the lowest left rib. EEG data were collected from
the Fz, Cz, C3, C4, Pz, P3, P4, Oz, O1, and O2 electrodes
of the International 10–20 system. All the electrodes were
referred to an electrode on the forehead midline. Imped-
ance was maintained below 5000 V. Data were amplified
using a band-pass of 0.01–100 Hz (1/2 amplitude low-
and high-frequency cut-offs); the amplification gain was
30,000 (Synamps, NeuroscanR). Recordings were notch-
filtered at 50 Hz.
Recordings were averaged off-line using an artifact-
rejection protocol based on voltage amplitude. All the
epochs for which the EEG exceeded 50 mV were dis-
carded with an automatic procedure. ERPs were obtained
by averaging the EEG, using the auditory stimulus onset as
trigger.
2.3.2. MEG recordings
The magnetoencephalographic (MEG) recordings were
made using a 148-channel whole-head magnetometer
(MAGNESR 2500 WH, 4D Neuroimaging. San Diego,
USA). The MEG was recorded with a 678.17 Hz sampling
rate, using an on-line band-pass filter of 0.1–200 Hz.
The data were submitted to a data analysis protocol
involving linear filtering between 0.01 and 20 Hz, rejection
of artifact-contaminated epochs, and estimation of the
intracranial sources most likely responsible for the observed
surface distribution of magnetic flux at intervals of 4 ms.
The single trial event-related fields (ERFs) elicited by
valid and invalid conditions were then averaged together
after manually removing those during which an eye move-
ment or blink had occurred (as indicated by a peak-to-peak
amplitude in excess of 50 AV in the electro-oculogram
channel). The ECG allowed to monitor cardiac artifacts in
the MEG signal. Once detected, ECG artifacts were rejected
and subsequently eliminated using the ‘‘artifact detector’’
tool, which is part of the 4D Neuroimaging software.
2.3.3. Dipole fitting
Although different investigators have proposed a variety
of source modeling approaches, we relied on the single,
equivalent-current dipole (ECD) source model that is part of
the 4D Neuroimaging software. The intracranial generators
(i.e., activity sources) of the magnetic signals at successive
intervals of 4 ms during the course of the ERF waveform
were modeled using a finite version of the non-linear
Levenberg–Marquardt algorithm [31]. The algorithm used
in this study searched for the activity source that was most
likely to have produced the observed magnetic field distri-
bution at a given time. The location of activity sources was
computed with reference to a Cartesian coordinate system
defined by a set of three anatomical landmarks (fiduciary
points): the right and left external meatus and the nasion.
The position of the magnetometers relative to the subject’s
head was precisely determined using five coils, three of
which were attached at the fiduciary points and two on the
forehead. The coils were activated briefly at the beginning
and again at the end of the recording session, and their
precise location in three-dimensional space was determined
using a localization algorithm built into the system. During
the recording session, a fiber-optic motion detector was used
to ensure that the subject’s head did not change position
relative to the sensor. Head positions relative to the sensor
were thus controlled and equal for all subjects across
measurements.
T1-weighted magnetic resonance images were obtained
from the nine subjects. In order to identify the anatomical
regions where the activity sources were localized, activity
Table 1
2� 2 (Condition�Hand) reaction times
Condition Hand Mean S.D.
Valid Left 424.151 21.947
Right 431.184 22.308
Invalid Left 522.909 14.047
Right 484.744 21.063
Units in ms.
C.M. Gomez et al. / Cognitive Brain Research 21 (2004) 59–6862
source coordinates were overlaid on to T1-weighted MRIs
using the STAR component of the 4-D Neuroimaging
software. Precise co-registration of the MEG coordinate
system on to the MRI was achieved by aligning the MEG
fiduciary points with high contrast cod liver oil capsules (3
mm in diameter) which were affixed to the subject’s
nasion and inserted in the external meatus prior to the
MRI scan.
Activity-source solutions were considered satisfactory
only upon meeting the following criteria: correlation and
goodness of fit z 0.90 between the observed and the best
predicted magnetic field distribution, and a 95% confidence
volume < 15 cm3.
2.3.4. Data analysis
Differences in reaction time were evaluated by means of a
2� 2 (Condition�Hand) repeated-measures analysis of var-
iance (ANOVA). In cases of statistical significance, a pair-
wise mean comparison was carried out using Tukey’s post
hoc analysis. Differences in magnetic field power between
baseline and preparatory periods were subjected to T-Student
paired-samples means comparison. All of these data analyses
were made using the SPSSR 8.0 statistical package.
Fig. 1. This figure displays the MEG and EEG grand averages for Valid Right cond
the CNV task during the preparatory period: visual field/potential, early CMV, late
148 MEG channels. Bottom row, Cz electrode position. fT, femtoteslas. nv, nano
3. Results
3.1. Behavioral analysis
RT values evidence a significant main effect of the factor
Condition (F1,8 = 25.406; p < 0.01), with an increased reac-
tion time in Invalid conditions (Table 1). There was also a
significant Condition�Hand interaction (F1,8 = 7.318; p <
0.05). The results demonstrate that in this case the increased
reaction time was more evident for the left hand ( q = 4.55;
p < 0.01). Finally, a significant main effect of the factor
Hand also appeared (F1,8 = 5.575; p < 0.05), showing a
faster RT for the right hand in all subjects.
The percentages of correct responses were 96.74% (LV),
96.35% (RV), 91.65% (LI), and 89.9% (RI).
3.2. Analysis of ERFs (CNV and CMV) and topographic
patterns (CMV)
A grand average was performed on the MEG and EEG
signals. This procedure is considered a standard in ERP
analyses, but is not as common for ERFs. Essentially, the
grand average calculation is possible when the sensors
position is stable across subjects. Since EEG measurements
rely on the well-known International 10/20 system, the
position of the sensors (electrodes) is ensured. As mentioned
above, the position of the MEG sensors was controlled and
kept stable, not only during every individual recording but
also across subjects, in order to ensure the reliability of the
MEG grand average.
Fig. 1 displays the grand average for the LV condition.
This grand average reflects the typical sequence observed
ition with the temporal sequence found and the main components elicited by
CMV, and auditory field/potential. Top row, collapsed representation of the
volts.
Table 2
Means and standard deviations of preparatory (CMV) and baseline field
powers in Left Valid (LV), Right Valid (RV), Left Invalid (LI), and Right
Invalid (RI) conditions
Condition Mean S.D.
LV_Baseline 11.22 1.81
LV_CMV 22.04 4.47
RV_Baseline 10.97 2.85
RV_CMV 21.81 4.96
LI_Baseline 17.81 2.41
LI_CMV 35.83 4.08
RI_Baseline 20.05 2.89
RI_CMV 37.53 4.68
Units in femtoteslas.
C.M. Gomez et al. / Cognitive Brain Research 21 (2004) 59–68 63
in the four conditions, and highlights the temporal concor-
dance between the CNV and its magnetic counterpart. The
so-considered CMV was encompassed by the visual and
auditory fields elicited by the warning and imperative
stimuli, respectively. In order to ensure the reliability of
the CMV as a real signal, the averaged field power of all
the MEG channels was calculated for the time period
between the visual and auditory fields (the preparatory
period) and compared with the averaged field power
calculated for an equal-duration period at the baseline.
The differences between preparatory period and baseline
were significant in all conditions: LV (T =� 6.88;
p < 0.001, DF = 8), RV (T=� 8.22; p < 0.001, DF = 8), LI
(T = � 13.70; p < 0.001, DF = 8), and RI (T = � 6;
p < 0.001; DF = 8), with a much higher field power during
the CMV period than at baseline. Table 2 shows the values
(in femtoteslas) of averaged field powers for CMV and
baseline. The high reliability of CMV in individual sub-
jects can be observed in Table 3, where every single
subject had a higher ERF field during the CMV with
respect to baseline.
Once the reliability of the CMV signal was established,
the second aim of this analysis was to define the temporal
stability of the CMV, and the possible existence of different
components. The field distributions in the time period
between the visual and auditory fields (the so-considered
CMV or preparatory period) were analyzed millisecond by
Table 3
Means and standard deviations of preparatory (CMV) and baseline (BL) field pow
(RI) conditions for individual subjects to show the intersubject reliability of CMV
Subjects BL-VL CMV-VL BL-VR MV-V
1 11.90 21.50 10.00 19.50
2 10.90 31,70 10.70 29.90
3 8.90 20.00 9.40 19.30
4 9.60 19.10 7.60 16.20
5 12.60 23.40 14.10 28.50
6 8.90 18.40 11.30 22.10
7 11.60 26.40 14.70 25.00
8 12.40 19.50 14.10 19.20
9 14.20 18.40 6.90 16.60
Units in femtoteslas.
millisecond for each condition; two well-defined compo-
nents were found by visual inspection of the ERF topo-
graphic maps: an early component with a mean duration-
interval between � 569.7 to � 464.7 ms pre-tone (range
� 583 to � 439) and 431 to 536 ms post-cue, and a second,
late component, with a longer duration—between � 431.6
to � 142 ms (range � 436 to � 136 ms) pre-tone and 569
to 858 ms post-cue. Within these time intervals the field
distributions were consistent, and may, therefore, reflect the
stability of the intracranial activity sources. These time
windows appear in Fig. 1 for both CMV and CNV. The
lack of a clear early electric CNV is due to the fact that the
reference used on the forehead is quite close to the brain
generators of the early CNV.
Figs. 2 and 3 display a portrait of the early and late CMV
components, and the field distribution of the auditory M1
field. The early component presents a medial–central dis-
tribution, while the late component has a completely differ-
ent field pattern, very similar (at least in RV, LI, and RI) to
the auditory M1 component, but with a smaller amplitude.
As far as the final goal of this study was to determine the
patterns of cortical activation during the preparatory period,
the intracranial sources of these fields were calculated.
3.3. Dipole localization
Only four areas were consistently activated during the
preparatory period, for all conditions: the inner part of the
superior temporal gyrus (STG), the anterior (primary motor)
region of the Rolandic sulcus (MC), the posterior cingulate
cortex (PCC), and the posterior region of the parietal cortex
(PPC). Fig. 4 shows the dipole localization in these four
brain areas for the left valid condition in a single subject.
Another area—the mesial structures of the temporal lobe—
was less-frequently activated. However, there was a high
level of variability across subjects, in terms of the temporal
patterns of activation in each cortical region. Fig. 5 shows
the sequential pattern of activation for each subject. Every
subject presents a unique sequence of activation where each
area can be activated just once or up to three times (i.e., the
STG) during the preparatory period. All the subjects pre-
ers in Left Valid (LV), Right Valid (RV), Left Invalid (LI), and Right Invalid
R BL-IL CMV-IL BL-IR CMV-IR
16.80 33.10 20.20 39.30
17.40 39.30 20.60 41.60
19.60 30.00 19.00 35.30
14.10 32.00 16.70 29.00
18.60 41.50 22.50 39.90
19.00 39.40 23.60 33.70
18.20 38.60 14.60 44.70
14.70 32.30 20.90 38.90
21.90 36.30 22.40 35.40
Fig. 2. Isocontour maps for Valid Left and Valid Right conditions displayed during the early and late CMV phases (left and central columns, respectively) and
the auditory evoked field M1 (right column). The represented values correspond to the mean value of the ERF grand average for the periods 431 to 536 ms
post-cue (early CMV), 569 to 858 ms post-cue (late CMV), and 80–120 ms post-imperative stimulus (auditory M1 field). The step between isocontour lines is
100 fT. ingoing and outgoing magnetic fields are indicated in the figure.
Fig. 3. Isocontour maps for Invalid Left and Invalid Right conditions displayed during the early and late CMV phases (left and central columns, respectively)
and the auditory evoked field M1 (right column). The represented values correspond to the mean value of the ERF grand average for the periods 431 to 536 ms
post-cue (early CMV), 569 to 858 ms post-cue (late CMV), and 80–120 ms post-imperative stimulus (auditory M1 field). The step between isocontour lines is
100 fT. Ingoing and outgoing magnetic fields are indicated in the figure.
C.M. Gomez et al. / Cognitive Brain Research 21 (2004) 59–6864
Fig. 4. Sagittal and coronal views of the stimulated dipolar source locations, superimposed on MRI, are shown for one subject in Left Valid condition. MC,
motor cortex; PPC, posterior parietal cortex; PCC, posterior cingular cortex; STG, superior temporal gyrus. Note the activation of the right MC and STG.
C.M. Gomez et al. / Cognitive Brain Research 21 (2004) 59–68 65
sented an early activation in the occipital cortex, cor-
responding to the visual areas activation induced by the
cue, and then the intermingled preparatory activation of the
previously described brain areas. Only a weak general
tendency indicates that motor—and posterior cingulate and
parietal—regions are more-frequently active during the
early phase of CMV, while STG was more-frequently active
during the late phase of CMV.
Fig. 5. Time of dipole fitting. The codes for brain areas are indicated below, and
cortex (MC), medial temporal (MT), posterior cingular cortex (PCC), and PPC (
contralateral activity with respect to the cued location is indicated. The cue appea
activity of different cortices, where a clear pattern of sequential activation cannot
Table 4 shows the percentage of subjects in whom these
areas were activated in every condition. As can be observed,
there was a tendency to activate the cortex contralateral to the
cued signal, for both temporal and motor cortices. The STG
contralateral to the arrow direction was always (LV) or
almost always (RV, LI, RI) active during the preparatory
period. In the case of an exclusively unilateral activation
(subjects 5 and 6 in RV, LI, and RI conditions), the
correspond to occipital cortex (OC), superior temporal gyrus (STG), motor
posterior parietal cortex). For all these areas, except OC, the ipsilateral or
rs at time 0. Note the early activation of the OC, and then the intermingled
be observed.
Table 4
Percentages of subjects who presented activation for every condition in
superior temporal gyrus (STG), motor cortex (MC), posterior cingulate
cortex (PCC), and posterior parietal cortex (PPC) areas
Condition STG MC PCC PPC
LV Left: 66.6% Left: 0% Left: 0% Left: 0%
Right: 100% Right: 66.6% Right: 33.3% Right: 44.4%
RV Left: 88.8% Left: 66.6% Left: 11.1% Left: 33.3%
Right: 66.6% Right: 22.2% Right: 22.2% Right: 0%
LI Left: 77.7% Left: 66.6% Left: 11.1% Left: 0%
Right: 55.5% Right: 22.2% Right: 11.1% Right: 22.2%
RI Left: 44.4% Left: 22.2% Left: 11.1% Left: 0%
Right: 77.7% Right: 77.7% Right: 11.1% Right: 22.2%
C.M. Gomez et al. / Cognitive Brai66
contralateral (to the arrow direction) STG was the only one
active. In the case of bilateral activation, the contralateral (to
the arrow direction) STG was always the first to be activated.
The patterns of motor cortex activation seem to be more
evident. Bilateral activation is less frequent (22.2% in RV,
LI, and RI conditions). In this case, the cortical lateralization
clearly depends on the arrow (central cue), and only subjects
5 and 6 showed an absence of motor preparatory activity.
For all conditions, the cortex most-frequently active (be-
tween 66% and 77% of the subjects) is the one contralateral
to the direction of the arrow.
4. Discussion
4.1. Behavioral analysis
The behavioral results presented in this study are depen-
dent upon the validity or invalidity of the visual central cue.
In the present experiment, RTs are shorter when the auditory
imperative stimulus is preceded by a valid visual cue than
when the auditory stimulus has been preceded by a cue in
the opposite (invalid) direction. A similar effect has been
reported in a great number of studies, basically using the
same modality for cue and target (for a review see Posner
and Cohen [27]).
The behavioral results obtained in present and other
experiments [27] suggest that there must be some kind of
neural activity during the preparatory periods which enables
a better processing of the expected stimulus, producing a
faster response in the valid condition than in the invalid one.
In fact, visual ERP components such as P1 and N1 have
been shown to be modulated by the presence of a peripheral
cue, suggesting an interaction between the preparatory
peripheral stimulus and the sensory gating [24].
However, the possible sensory and motor modulation
during the preparatory periods within a central-cue Posner-
type paradigm have never before been addressed using
neural source localization techniques. In this kind of
paradigm, the brain activity linked to the arousal effects
should not be affected by cue directionality—in contrast,
attentional effects should be driven by the directionality of
the cue.
4.2. Magnetic and electrical recordings
During the preparatory period, an evident CNV potential
appeared in the grand average of the nine subjects. The early
CNV was not so clearly developed as the late CNV, possibly
due to the forehead reference used, which would diminish
the frontal contribution to the early CNV [15]. The electrical
CNV was simultaneously accompanied by a statistically
significant change in the ERF values (comparing the pre-
paratory period with baseline). This change in the ERF
could be interpreted as the magnetic counterpart of the
CNV—the CMV—given the similar time dynamics of the
two (Fig. 1), but also because of its relative topographic
stability, as seen in Figs. 2 and 3. Moreover, the increase of
ERF during CMV periods with respect to baseline showed a
high individual reliability. Previous experiments have
evidenced the feasibility of CMV recordings in paradigms
similar to those used to obtain the CNV, and—as far as we
are aware—always in Go/NoGo conditions [1,2,6,9,10,
17,18]. All those studies showed variable, subject-depen-
dent topographies, which could reflect the different geom-
etries of neural generators. However, the intersubject
variability probably also reflects magnetic noise due to
the relatively low number of trials that can be presented in
S1/S2-type experiments, given the long duration of single
trials. We have applied an ERF grand average procedure
similar to that generally used in ERP studies, allowing
highlighting of the topographical and dynamic intersubject
commonalities of a given ERP (or ERF) component in
time and space.
The topography of a given component represents a qual-
itative approach to the neural source localization problem. In
the CMV grand average obtained, two well-defined periods
were found: an early one where the topography suggests the
presence of a central-midline generator for the four condi-
tions considered, and a late phase, which could be explained
by generators in the superior temporal cortex. In fact, a
qualitative comparison between the late CMV and the audi-
tory M1 component reveals a very similar topography in the
preparatory and perceptual periods. The late CMV topogra-
phy is compatible with that described for the auditory M1
component, since it was its origin in the STG [25]. Thus, from
the topographic analysis, a neural source located in the
auditory cortex might be proposed for the late CMV. It is
more difficult to propose a single neural source localization
based on magnetic field topography analysis for the early
CMV,. However, magnetic field distributions with incoming
and outgoing flux around the midline has been obtained by
Hultin et al. [17] in the CMV, and a source in the Rolandic
sulcus for such a distribution was proposed. In any case, Fig.
5 shows that in the early CMV there is intermingled activity
of several brain areas, including the STG and PPC, but also
midline activity situated in the cingulate gyrus that could be
contributing to the ERF distribution.
The topographically predicted neural sources were fully
confirmed by the dipole localization analysis: (i) primary
n Research 21 (2004) 59–68
C.M. Gomez et al. / Cognitive Brain Research 21 (2004) 59–68 67
motor and auditory areas are more-frequently active during
the preparatory (CMV) period than either of the other
activated areas (the PPC and the PCC), and (ii) the
lateralization of this activation depends, essentially, on
the orientation of the central cue for motor and auditory
cortices. As has been seen in Table 4, there was a
tendency towards a bilateral pattern of auditory cortex
activation, but the cortex contralateral to the arrow direc-
tion was more-frequently active. An almost identical
pattern appears for the primary motor cortex, where the
MC contralateral to the cued ear is more-often active than
the ipsilateral one.
Our findings are in keeping with at least four previous
studies. Fenwick et al. [10], using a Go/NoGo paradigm,
showed that the CMV consisted of multiple generators, not
only frontal but temporal and parietal. Ioannides et al.
[18], using magnetic field tomography, obtained a similar
pattern of activation in motor and auditory cortices during
the preparatory periods (but also of SMA) in two subjects.
The topographic analyses performed by Elbert et al. [9] in
their CMV experiment suggested that during the prepara-
tory periods, activations occur in the auditory, motor, and
association areas, in at least some 50% of the subjects
analyzed. Finally, Dammers and Ioannides [6], using
magnetic field tomography, have obtained activation in
the same brain areas as in the present report during the
preparatory period of a Go/NoGo paradigm. In addition,
they obtained activation in inferior prefrontal cortex and
the SMA. The involvement of frontal areas is well
documented in the CNV literature [15]. The lack of
evident dipolar frontal activation in the present report
could be due to the use of a single-dipole model, which
would preclude low-intensity sources being obtained when
high-intensity sources are simultaneously present. Regard-
ing the activation in PCC and PPC, it must be remarked
that a similar pattern of activation has also been obtained
in CNV–CMV experiments [6,15], suggesting that these
brain areas are involved in a posterior preparatory atten-
tional network.
The pattern of sensory and motor activation obtained
during the preparatory period in the present experiment,
but also in those of Ioannides [18], Elbert [9], and
Dammers and Ioannides [6], indicates that subjects do
anticipate the brain areas which will be required for
processing the expected auditory stimulus and the subse-
quent requested action as cued by the visual stimulus. The
congruence or incongruence of the anticipated motor or
sensory regions with the areas that are finally requested by
the imperative auditory stimulus may well explain the
behavioral effects elicited by the task, where the presence
of a valid cue stimulus produces a faster response than that
to an invalid cue. Some previous reports using fMRI [19]
and low-resolution electromagnetic tomography (LOR-
ETA) [15] suggest that during the preparatory periods
there is a sensory and motor network anticipating the
stimulus and the requested actions.
In this sense, and according to Brunia (for a review of
this perspective, see Brunia [3,4]), our evidence could well
be explained within a global attentional and predictive
system where the perceptual domain is considered prepara-
tory sensory attention and the motor domain is considered
the preparation to respond. The anatomical bases of this
attentional–anticipatory system are well defined for the
motor domain: they include not only the frequently de-
scribed prefrontal, SMA, and primary motor cortices, but
posterior parietal cortex, cingular cortex, and pulvinar
thalamic nuclei too. The neural substrate of the perceptual
domain is not so well-described, but, of course, the partic-
ipation of primary sensory areas has been hypothesized. Our
results not only evidence an important level of correspon-
dence with Brunia’s predicted neural network but also
confirm that the patterns of temporal activation in those
areas would mirror the sensory and motor overlapping that
S1/S2 tasks demand in attentional processing. The presence
of task-specific cortical networks has also been detected in
the differential topographies of the negative slow waves
registered during the retrieval of different items during a
memory task [30].
Therefore, the anticipatory activity could be considered a
purposeful adaptive mechanism that would enable a faster
processing of the expected stimulus when cues are valid;
however, when the predictions are not accurate, the neces-
sary functional sensorimotor reorganization would produce
a cost associated to invalid cues [27]. How relevant this
conclusion is for the case when the central cue and the target
appear in the visual modality, as in the standard Posner’s
paradigm, remains to be tested.
Acknowledgements
The present work was supported by grant number
BSO2001-2383 of the Spanish DGYCIT and by a grant
from the Junta of Andalucia.
References
[1] L.F. Basile, R.L. Rogers, W.T. Bourbon, A.C. Papanicolau, Slow
magnetic flux from human frontal cortex, EEG and Clin. Neurophy-
siol. 90 (1994) 157–165.
[2] L.F. Basile, D.G. Brunder, I.M. Tarkka, A.C. Papanicolau, Magnetic
fields from human prefrontal cortex differ during two recognition
tasks, Int. J. Psychophysiol. 27 (1997) 29–41.
[3] C.H.M. Brunia, Neural aspects of anticipatory behavior, Acta Psy-
chol. 101 (1999) 213–242.
[4] C.H.M. Brunia, G.J.M. van Boxtel, Wait and see, Int. J. Psychophy-
siol. 43 (2001) 59–75.
[5] R.Q. Cui, A. Egkher, D. Huter, W. Lang, G. Lindinger, L. Deecke,
High resolution spatio-temporal analysis of the contingent negative
variation in simple or motor complex motor tasks and a non-motor
task, Clin. Neurophysiol. 111 (2000) 1847–1859.
[6] J. Dammers, A.A. Ioannides, Neuromagnetic localization of CMV
generators using incomplete and full-head biomagnetometer, Neuro-
Image 11 (2000) 167–178.
C.M. Gomez et al. / Cognitive Brain Research 21 (2004) 59–6868
[7] L. Deecke, The Bereitschaftspotential as an electrophysiological tool
for studying the cortical organization of human voluntary action,
Suppl. Clin. Neurophysiol. 53 (2000) 199–206.
[8] L. Deecke, J. Boschert, P. Brickett, H. Weinberg, Magnetoencepha-
lographic evidence for possible supplementary motor area participa-
tion in human voluntary movement, in: H. Weinberg, G. Stroink, T.
Katila (Eds.), Biomagnetism: Applications and Theory, Pergamon,
New York, 1985, pp. 369–372.
[9] T. Elbert, B. Rockstroh, S. Hampson, C. Pantev, M. Hoke, The mag-
netic counterpart of the contingent negative variation, EEG and Clin.
Neurophysiol. 92 (1994) 262–272.
[10] P.B.C. Fenwick, A.A. Ioannides, G.W. Fenton, J. Lumsden, P.
Grummich, H. Kober, A. Daun, J. Vieth, Estimates of brain activity
using magnetic field topography in a GO/NOGO avoidance para-
digm, Brain Topogr. 5 (1993) 275–282.
[11] A.M. Ferrandez, V. Pouthas, Does cerebral activity change in middle-
aged adults in a visual discrimination task? Neurobiol. Aging 22
(2001) 645–657.
[12] C. Frith, A framework for studying the neural basis for attention,
Neuropsychologia 39 (2001) 1367–1371.
[13] C.M. Gomez, A. Delinte, E. Vaquero, M.J. Cardoso, M. Vazquez, M.
Crommelinck, A. Roucoux, Current source density analysis of
CNV during temporal gap paradigm, Brain Topogr. 13 (2001)
149–159.
[14] C.M. Gomez, A. Delinte, E. Vaquero, M. Vazquez, M. Atienza, S.
Millan, M. Crommelinck, A. Roucoux, Sensory, motor, attentional
modulation in the manual gap effect. A high-density ERP study,
Exp. Brain Res. 142 (2002) 385–394.
[15] C.M. Gomez, J. Marco, C. Grau, Visuo-motor cortical network of the
contingent negative variation, NeuroImage 20 (2003) 216–226.
[16] M.R. Harter, L.M. Steven, N.J. Price, M.E. Lalonde, A.L. Keyes,
Neural processes involved in directing attention, J. Cogn. Neurosci.
1 (3) (1989) 223–237.
[17] L. Hultin, P. Rossini, G.L. Romani, P. Hogstedt, F. Tecchio, V. Piz-
zella, Neuromagnetic localization of the late component of the cog-
nitive negative variation, EEG and Clin. Neurophysiol. 98 (1996)
435–448.
[18] A.A. Ioannides, P.B. Fenwick, J. Lumsden, M.J. Liu, P.D. Bamidis,
K.C. Squires, D. Lawson, G.W. Fenton, Activation sequence of
discrete brain areas during cognitive processes: results from mag-
netic field tomography, EEG and Clin. Neurophysiol. 91 (1994)
399–402.
[19] S. Kastner, M.A. Pinsk, P. De Weerd, R. Desimone, L.G. Ungerleider,
Increased activity in human visual cortex during directed attention in
the absence of visual stimulation, Neuron 22 (1999) 751–761.
[20] H.H. Kornhuber, L. Deecke, Hirnpotentialanderungenbei Willkurbe-
wegungen und passiven Bewegungen des Menschen: Bereitschaftspo-
tential und reafferente Potentiale, Pflugers Arch. 284 (1965) 1–17.
[21] K.M. Lee, K.H. Chang, J.K. Roh, Subregions within the supplemen-
tary motor area activated at different stages of movement preparation
and execution, NeuroImage 9 (1999) 117–123.
[22] N.E. Loveless, A.J. Sanford, Slow potentials correlates of preparatory
set, Biol. Psychol. 1 (1974) 303–314.
[23] S.J. Luck, L. Chelazzi, S.A. Hillyard, R. Desimone, Neural mecha-
nisms of spatial selective attention in areas V1, V2, and V4 of ma-
caque visual cortex, J. Neurophysiol. 77 (1997) 24–42.
[24] G.R. Mangun, S.A. Hillyard, S.J. Luck, Electrocortical substrates of
visual selective attention, in: D. Meyer, S. Konblum (Eds.), Atten-
tion and Performance, vol. 14, MIT Press, Cambridge, MA, 1993,
pp. 219–243.
[25] N. Nakasato, S. Fujita, K. Seki, Functional localization of bilateral
auditory cortices using an MRI-linked whole head magnetoencepha-
lography (MEG) system, EEG and Clin. Neurophysiol. 94 (1995)
183.
[26] R.L. Oldfield, The assessment and analysis of handedness: the Edin-
burgh Inventory, Neuropsychology 9 (1971) 97–113.
[27] M.I. Posner, Y. Cohen, Components of visual orienting, in: H.
Bouma, D.G. Bouwhuis (Ed.), Attention and Performance, Erl-
baum, Hillsdale, NJ, 1984, pp. 531–556.
[28] V. Pouthas, L. Garnero, A.M. Ferrandez, B. Renault, ERPs and PET
analysis of time perception: spatial and temporal brain mapping dur-
ing visual discrimination tasks, Hum. Brain Mapp. 10 (2000) 49–60.
[29] B. Rockstroh, T. Elbert, N. Birbaumer, W. Lutzenberger, Slow Brain
Potentials and Behavior, Urban and Schwarzenberg, Baltimore–
Munich, 1982.
[30] F. Rosler, M. Heil, E. Henninghausen, Distinct cortical activation
patterns during long-term retrieval of verbal, spatial and color infor-
mation, J. Cogn. Neurosci. 7 (1995) 51–65.
[31] J. Sarvas, Basic mathematical and electromagnetic concept of the
biomagnetic problem, Phys. Med. Biol. 32 (1987) 11–22.
[32] W.G. Walter, R. Cooper, W.J. Aldridge, W.C. McCallum, Contingent
negative variation: an electrophysiological sign of sensoriomotor as-
sociation and expectancy in the human brain, Nature 203 (1964)
380–384.
[33] T.C. Weerts, P.I. Lang, The effects of eye fixation and stimulus and
response location in the contingent negative variation (CNV), Biol.
Psychol. 1 (1973) 1–19.
[34] R. Zappoli, A. Versari, F. Zappoli, R. Chiaramonti, G.D. Zappoli-
Thyrion, M. Grazia-Arneodo, V. Zerauschek, The effects on auditory
neurocognitive evoked responses and cognitive negative variation ac-
tivity of frontal cortex lesions or ablations in man: three new case
studies, Int. J. Psychophysiol. 38 (2000) 109–144.