alpha synchronization and anxiety_knyazev
TRANSCRIPT
www.elsevier.com/locate/ijpsycho
International Journal of Psychoph
Alpha synchronization and anxiety: Implications for inhibition vs.
alertness hypotheses
Gennady G. Knyazev*, Alexander N. Savostyanov, Evgenij A. Levin
State Research Institute of Physiology, Siberian Branch of the Russian Academy of Medical Sciences, Timakova str., 4, Novosibirsk 630117, Russia
Received 6 December 2004; received in revised form 8 February 2005; accepted 18 March 2005
Available online 14 June 2005
Abstract
Although there is much evidence that alpha oscillations are linked with processes of perception, attention and semantic memory, their
functional significance remains uncertain. Synchronization in the alpha frequency range is taken to be a marker of cognitive inactivity, active
inhibition of sensory information, or a means of inhibition of non-task relevant cortical areas. Here we propose an alternative interpretation
which posits that higher alpha power during reference interval signifies higher readiness of alpha system to information processing.
Predictions derived from the inhibition and alertness hypotheses were tested during presentation of acoustic stimuli (tone 1000 Hz) and
neutral words to 30 males (18–25 years) with different levels of trait anxiety. On the whole, predictions derived from the inhibition theory
were not confirmed and findings more corresponded to the alertness hypothesis. High-anxiety subjects showed higher alpha power during
reference interval simultaneously with higher magnitude of event-related desynchronization and higher amplitude of phase-locked alpha
responses. These findings are discussed in terms of functional significance of alpha band synchronization and desynchronization.
D 2005 Elsevier B.V. All rights reserved.
Keywords: EEG; Alpha oscillations; Anxiety; Event-related desynchronization; Averaged evoked potential
1. Introduction
Some data indicate that alpha oscillations are enhanced in
anxious individuals particularly in anxiogenic environment
(Bell et al., 1998; Herrmann and Winterer, 1996; Knyazev et
al., 2002, 2003, 2004a,b; Knyazev and Slobodskaya, 2003).
This enhancement has been interpreted as a sign of higher
readiness of alpha system for information processing
(Knyazev and Slobodskaya, 2003). Prima facie, this
interpretation seems dubious since enhanced alpha oscil-
lations have long been considered as an attribute of
relaxation. Indeed, starting from Berger’s (1929) pioneering
works, many studies have noted a task-related decrease in
alpha power. This finding was so pervasive that alpha power
has come to be considered as a reverse measure of
activation. More recently this idea has been reconceptual-
0167-8760/$ - see front matter D 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.ijpsycho.2005.03.025
* Corresponding author. Tel.: +7 383 2 33 48 65; fax: +7 383 2 32 42 54.
E-mail address: [email protected] (G.G. Knyazev).
ized to propose alpha as a mechanism for increasing signal
to noise ratios within the cortex by means of inhibition of
unnecessary or conflicting processes to the task in hand
(Klimesch et al., 1999, 2000)—the greater the task
demands, the more inhibition needed, the greater the
synchronization. Klimesch’s proposals are compatible with
the notion of ‘‘surround inhibition’’ wherein active cortical
areas, indexed by alpha desynchronization are surrounded
by a ‘‘doughnut’’ of alpha synchronization or inhibition
(Suffczynski et al., 2001; Pfurtscheller, 2003) in keeping
with Crick’s (1984) spotlight of attention hypothesis.
The idea of inhibitory function for alpha synchronization
is appealing but it raises some doubts. First, it is not clear
how the same mechanism might be linked with perceptual
activation, as in the case of phase-locked evoked alpha
oscillations (Basar, 1998, 1999), and perceptual inhibition
(as proposed for event-related alpha synchronization, ERS).
Next, if ERS served a function of selective attention (e.g.
inhibition of non-task-relevant perception), one would
expect that relatively small cortical area within a task-
ysiology 59 (2006) 151 – 158
G.G. Knyazev et al. / International Journal of Psychophysiology 59 (2006) 151–158152
relevant zone would show ERD whereas larger cortical
areas, which are not related to the task processing, would
show ERS. That would correspond to the Crick’s spotlight
of attention hypothesis. The more focused is attention (e.g.
in the beginning of an experiment) the more widespread
ERS and more localized event-related desynchronization
(ERD) should be observed. Actually the opposite applies.
Alpha ERD is a much more common and widespread
phenomenon than alpha ERS. It is usually more pronounced
and widespread during first presentations of a signal or a
task. It is stronger during more complex tasks compared to
the relatively simple ones (Neubauer et al., 1999). All these
observations are difficult to reconcile with the idea of lateral
inhibition as a function of ERS.
Owing to extensive studies by Basar (1998, 1999) as well
as other authors (Klimesch, 1999), a considerable body of
knowledge has been accumulated indicating that depending
on background activity, different reactions of EEG bands
could be observed. According to the concept proposed by
Basar (1998), the ongoing EEG determines (controls)
evoked activity. This signifies that through the maintenance
of higher or lower power of specific oscillations, brain could
be prepared or predisposed to specific pattern of responses.
Klimesch (1999) states that the reactivity in band power can
be predicted from the amount of absolute power as
measured during a resting state. Considering the reactivity
of alpha band he notes that large alpha power is associated
with a large amount of desynchronization during task
performance. He concludes that the most reactive individ-
uals would show in resting condition significantly more
power in the alpha band.
It is not intuitively clear why a state of expectation of a
perceptual event during the reference interval should be
associated with inhibition of those same cortical areas,
which later will be engaged in perception of this event. It is
well known that inhibition is associated with diminished
responsiveness. For example an inhibited (that is, hyper-
polarized) neuron is not responsive to external stimuli.
Contrary to that, above discussed evidences imply that alpha
enhancement during inter-stimulus interval is associated
with enhanced responsivity. Therefore we suggest that in
this case alpha synchronization should not be considered as
‘‘inhibition’’ of correspondent networks. On the contrary, it
reflects active preparation of the alpha system to a
demanding task.
Our interpretation is as follows. The EEG consists of the
activity of an ensemble of generators producing rhythmic
activity in several frequency ranges. These oscillators are
active usually in a random way. By application of sensory
stimulation these generators may be coupled and act
together in a coherent way. This synchronization and
enhancement of EEG activity gives rise to ‘‘evoked’’ or
‘‘induced rhythms’’ (Basar et al., 2000). Synchronization is
the typical arousal reaction for delta, theta, high beta, and
gamma rhythms. Desynchronization is mainly peculiar to
alpha and lower beta but these rhythms also show
synchronization when evoking events are relatively simple
for processing. If a task demands additional cortical
structures and diverse processes to be involved, alpha
oscillators disintegrate to smaller groups participating in
different processes and this reveals itself in event-related
desynchronization. The extent of alpha desynchronization
should correlate with perceived complexity of a task or
importance of a signal and reflects allocation of resources
needed for its management. But in order to effectively
process a demanding signal, the alpha system has to be
prepared. That means that alpha oscillators should be
disengaged from their current random activity and gathered
into a united system ready to action. This reveals itself in
anticipatory or preparatory alpha synchronization which is
the best background for ERD. Research indicates that this is
also the best condition for good performance. For example,
a large upper alpha power in a reference interval preceding a
task is related to both large suppression of upper alpha
power during the task and good performance (Klimesch,
1999). Moreover, artificial enhancing of alpha power by
means of repetitive transcranial magnetic stimulation at
individual upper alpha frequency can enhance task perform-
ance and, concomitantly, the extent of task-related alpha
desynchronization (Klimesch et al., 2003).
In the present study, we intended to check whether
assumptions derived from the inhibition theory may
explain changes of alpha power and reactivity during
repetitive presentation of neutral words to subjects with
different levels of trait anxiety. According to this theory,
alpha synchronization is associated with inhibition of non-
relevant cortical areas. The more attention to the stimuli,
the more inhibition is needed. Therefore, one would expect
that perception of acoustic stimuli should be associated
with alpha ERD within central cortical regions, and alpha
ERS within cortical areas not associated with acoustic
perception (e.g., posterior regions). ERS should be more
pronounced and widespread in the beginning of stimuli
presentation and should diminish after repetitive presenta-
tion of the same stimuli due to extinction of attention. It
also should be more evident in subjects with higher trait
anxiety, since it is well established that these subjects are
predisposed to enhanced attention to novel stimuli in
unfamiliar environment (Gray, 1987).
Alternative interpretation posits that higher background
alpha power signifies higher readiness of alpha system and
should be associated with higher reactivity (that is, ERD). In
the between-subject domain this implies that anxious
individuals, tending to show more power of alpha oscil-
lations in resting conditions (Knyazev et al., 2002, 2003,
2004a,b; Knyazev and Slobodskaya, 2003), would be more
predisposed to react to external stimuli by alpha desynch-
ronization. In the within-subject domain, we expected that
ERD would be most pronounced within cortical area with
highest alpha power (i.e. posterior region). Note that this is
in sharp contrast with expectation derived from the
inhibition theory, which predicts ERS in this region. Further,
G.G. Knyazev et al. / International Journal of Psychophysiology 59 (2006) 151–158 153
we expected that beginning of words presentation would
produce increase of alpha power during the reference
interval in anxious individuals but not in non-anxious ones.
This expectation is based upon the assumption that
presentation of neutral words would not be a demanding
task for subjects with low anxiety whereas for high-anxiety
subjects any unexpected stimulus represents a potential
threat thus provoking additional adjustment of point of
regulation for alpha system.
There are observations that a larger amplitude evoked
potential (EP) occurs when a stimulus falls against a low-
amplitude background and this has been interpreted as an
evidence of more effective information processing (that is,
primary sensory perception) (Basar, 1998). This looks like a
contradiction. We have earlier discussed evidences that
enhanced background alpha power is associated with better
task performance (Klimesch et al., 2003) and individuals
with higher background alpha power are better performers
on some cognitive tasks (Klimesch, 1999). Might it be that
primary sensory perception is poorer in these states and in
these individuals? Theoretically, since evoked response is
actually re-organization and phase-locking of ongoing
activity (Basar, 1998, 1999), high background alpha should
result in higher evoked alpha amplitude. When considering
a meaning of observations about dependence of EP
amplitude on preceding activity, the spontaneous variations
of EEG amplitude must be taken into account. It seems that
in each state the oscillatory activity is dynamically main-
tained around some level or point of regulation so that the
mean power of oscillations remains relatively stable. When
a moderate stimulus falls against a low-amplitude back-
ground, there is high probability that a splash of activity
would appear during subsequent period. In this study, we
sought to test whether the phase-locked alpha response (that
is, alpha band-pass filtered auditory EP, AEP) is lower in
subjects with higher background alpha power which should
be expected basing on the above-described within-subject
observations (Basar, 1998). If, on the other hand, EP
represents phase-locked on-going activity, the opposite
might be expected.
2. Materials and methods
2.1. Subjects
Subjects were 30 right-handed non-psychology stu-
dents, males, Caucasian, aged 18 to 25 years (mean 21.2;
S.D. 3.5). All participants gave consent to completing the
self-report questionnaires and the psychophysiological
protocols.
2.2. Instruments and procedures
Russian versions of the Spielberger State Trait Anxiety
Inventory (STAI, Spielberger et al., 1970; Hanin, 1989) and
Taylor Manifest Anxiety Scale (MAS, Taylor, 1953) were
used as psychometric measures of anxiety. Test Annet
(1970) was used to evaluate handedness. These were filled
out just before the experimental procedure.
Physiological measures were obtained in afternoon. Each
participant was seated comfortably in a reclined armchair
with eyes closed within a sound insulated dimly lit chamber.
The participants were asked to minimize their movements
during the recording. Prior to recordings, the individual
thresholds of acoustical sensitivity were measured for each
subject. After 1 min of baseline recording, subjects were
presented with 30 acoustic stimuli (tone 1000 Hz, 500 ms,
50 dB above individual threshold, with inter-stimulus
interval of 1.28 s.) for AEP registration. We wittingly used
a non-verbal acoustic stimulus for AEP registration since it
supposedly would not produce specific reactions and would
allow to measure ‘‘pure’’ background sensory perception.
Then three previously tape-recorded words (conclusion,
finger, porch) were 20 times presented in triplets with the
inter-stimulus interval of 10 s. These words were randomly
selected from a list of words which in previous studies were
rated by a representative sample of experts as emotionally
neutral. The rationale for using neutral words was that they
supposedly would not evoke substantial reactions in non-
anxious subjects but in anxious subjects even these stimuli
would produce some reactions at least in the beginning of
their presentation.
2.3. Psychophysiological recording
EEGs were recorded using a 32-channel PC based
system via silver–silver chloride electrodes. A mid-fore-
head electrode was the ground. The electrode resistance was
maintained below 5 kV. The signals were amplified with a
multichannel biosignal amplifier with bandpass 0.05–70
Hz, �6 dB/octave and continuously digitised at 300 Hz.
The electrodes were placed to 30 head sites according to the
International 10–20 system and referred to linked-ear
electrode. The horizontal and vertical EOG was registered
simultaneously. All recordings were visually inspected off-
line and data contaminated with muscle or blink artifact was
discarded.
2.4. Psychophysiological data reduction
To reduce the number of variables for ERD analysis, the
cortical sites were grouped into three zones and all EEG
data were averaged within these zones. Fp1, Fp2, F3, F4, F7
and F8 represented frontal, T1, T2, T3, T4, C3 and C4—
central, and O1, O2, P7, P8, P3, and P4—posterior zones.
AEPs were evaluated at T3 and Cz, which are frequently
used in studies of acoustic perception and were considered
as stimulus–adequate sites and at O1, which was considered
as stimulus-inadequate site (Basar, 1998).
ERD represents the percentage of a change in band
power during a test interval with respect to a reference
G.G. Knyazev et al. / International Journal of Psychophysiology 59 (2006) 151–158154
within a defined frequency band. When calculating ERD, in
a first step the EEG data are usually digitally bandpass
filtered. In the present study we meant to analyze relatively
long periods, therefore Fast Fourier transformation of
respective intervals was used instead of digital filtering.
An interval of 3000 to 1000 ms before the onset of a word
presentation was used as reference. Test interval was the
time periods of 1000 ms following onset of the word
presentation. Alpha power density was calculated for
respective artifact-free EEG chunks for presentation of first
30 words. We used alpha peak frequency, averaged over all
leads and all baseline epochs, as the anchor to adjust
frequency bands individually for each subject (Klimesch,
1999). In three subjects with no alpha peak we used the
gravity frequency ( f(i)), which was calculated as the
weighted sum of spectral estimates, divided by alpha power:
f(i)= (~(a( f) * f)) / (~(a( f)). Power spectral estimates at
frequency f are denoted a( f). The index of summation is
in the range 7–13 Hz. In the present study only alpha2 and
alpha3 sub-bands (in Klimesch’s notification) were consid-
ered. They were defined in relation to f(i) determined either
as gravity or peak frequency: alpha2—f(i)*0.8 to f(i) and
alpha3—f(i) to f(i)*1.2 (Doppelmayr et al., 1998). To avoid
non-normal distribution, alpha power density was log-
transformed. Consequent analyses were performed with
both log-transformed and non-transformed power values.
Since results did not differ substantially, only results for
non-transformed variables are reported for the sake of
clarity. All power density measures, taken before and after a
word presentation, were averaged within first, second and
third decades of words. If respective epochs for a word
presentation were contaminated by artifacts, they were
omitted and average value within this decade was calculated
for remaining words. Reference interval before presentation
of the first word was used as baseline and therefore was not
included in calculation of average inter-stimuli power
density during presentation of first ten words. ERD was
calculated by dividing the difference between alpha power
density before and after a word presentation by the pre-
stimulus alpha power density and multiplying it by 100
(Pfurtscheller and Aranibar, 1977). Note that positive ERD
values indicate desynchronization, whereas negative values
indicate synchronization. Averaged AEPs (30) were
obtained for all subjects and off-line filtered in individually
adjusted alpha diapason, f(i)�2 to f(i)+2 by means of
sharp FFT method. The maximal peak-to-peak amplitude of
alpha oscillations was evaluated within 500 ms after
stimulus (Basar, 1998).
3. Results
There are no normative data on trait anxiety (TA) in
Russian population. In this sample mean (S.D.) TA (STAI)
was 41.1 (9.5). In a reasonably large sample (N =307) of
students and their relatives we (Knyazev et al., 2004a,b)
have recently collected data on a number of psychometric
measures including TA. In that sample mean (S.D.) TA was
41.8 (9.9). Therefore, the mean level of TA in the present
study sample could be considered as average. The two
measures of trait anxiety (TA and MAS) correlated at 0.92
signifying that they actually capture the same construct.
State anxiety on the other hand showed only moderate
correlations with TA (r =0.56, P <0.001) and MAS
(r =0.56, P <0.001).
First, correlations between anxiety measures and both
pre-stimulus alpha power density and ERD magnitude
averaged across all cortical sites and all stages of words
presentation were calculated. Since the direction of an effect
was specified in advance, a one-tailed significance level test
was applied. For pre-stimulus alpha power density, all
correlations were positive and significant. They were higher
for alpha3 (r =0.34, P=0.032, r =0.47, P=0.005, and
r =0.52, P=0.002, for TA, MAS, and SA, respectively)
than for alpha2 (r =0.39, P=0.017, r=0.43, P=0.009, and
r =0.34, P=0.035, for TA, MAS, and SA, respectively). For
ERD magnitude, the correlations were also positive. The
correlation of alpha3 ERD with SAwas significant (r =0.41,
P=0.011) and with MAS marginal (r=0.30, P=0.053).
Correlations of alpha2 ERD did not reach significance level.
Next, correlations between ERD magnitude and respective
pre-stimulus alpha power density were calculated. All of
them were positive ranging from 0.16 to 0.46. 8 (out of 18)
correlations were significant and 6 others marginal
(P <0.1).
Greenhouse–Geisser corrected for sphericity assumption
violation repeated measures ANOVA was used to test the
effects of MAS, repeated presentation of words (RPW, viz.
first ten words, second ten words, and third ten words),
cortical zone (frontal vs. central vs. posterior), and alpha
sub-band (alpha2 vs. alpha3) on ERD magnitude. To
preserve statistical power, MAS was entered as a covariate.
Later, to elucidate the effects of MAS, this analysis was also
repeated with entering MAS as a binary factor (above vs.
below median). The effects of zone (F =4.01, df =1.58,
P = 0.034), zone�band�MAS ( F = 6.63 df = 1.68,
P=0.005), and RPW�band�MAS (F =3.5, df =1.98,
P=0.037) emerged as significant. The main effect of
MAS was marginal (F =3.49 df =1, P=0.072). Main effect
of zone indicated that ERD magnitude decreased from the
posterior to the anterior cortical sites (estimated marginal
means=�5.68, 0.77, and 4.99 for frontal, central and
posterior regions, respectively). The zone�band�MAS
interaction is presented at Fig. 1 and respective marginal
means and standard errors are presented in Table 1. In low-
anxiety subjects within posterior cortical zone, the mean
ERD magnitude for both alpha sub-bands approached zero.
At frontal and especially at central cortical sites, mean
alpha2 ERD values were substantially lower than respective
mean alpha3 ERD values. It should be noted that in low-
anxiety subjects, the only positive mean ERD values were
those for alpha3 within central (estimated marginal
Alpha2
-25-20
-15-10-5
05
10
1520
ER
D (
%)
Alpha3
-20
-15-10
-5
0
510
15
1st ten words 2nd ten words 3rd ten words
ER
D (%
)
Low MAS
High MAS
Fig. 2. Effect of repeated presentation of words on alpha2 (top) and alpha3
(bottom) ERD in subjects with low and high Manifest Anxiety (MAS).
Alpha2
-30-25-20
-15-10-5
05
1015
ER
D (
%)
ER
D (
%)
Alpha3
-15
-10
-5
0
5
10
15
20
Frontal Central Posterior
Low MAS
High MAS
Fig. 1. Magnitude of alpha2 (top) and alpha3 (bottom) ERD at different
cortical zones in subjects with low and high Manifest Anxiety (MAS).
G.G. Knyazev et al. / International Journal of Psychophysiology 59 (2006) 151–158 155
mean=0.98) and posterior (estimated marginal mean=0.25)
regions. All other values were negative implying ERS rather
than ERD. In high-anxiety subjects, on the other hand, only
mean alpha3 ERD within frontal zone was negative
(estimated marginal mean=�1.81). All other values were
positive clearly indicating prevalence of desynchronization
in these subjects. Comparison of the two groups of subjects
shows that whereas in subjects with low anxiety prevalence
of specific (alpha3 ERD) over unspecific (alpha2 ERD)
activation is most marked at specific for acoustic perception
cortical zone (that is central zone) and is virtually absent at
nonspecific cortical zone (posterior zone), in high-anxiety
subjects the opposite applies. Here prevalence of alpha3
over alpha2 ERD is most marked within posterior region, it
Table 1
Estimated marginal means and standard errors of alpha2 and alpha3 ERD
(%) at different cortical areas in subjects with low and high manifest anxiety
(MAS)
MAS Zone Band Mean ERD Standard error
Below median Frontal Alpha2 �15.8 7.5
Alpha3 �6.2 6.9
Central Alpha2 �11.4 7.8
Alpha3 1.0 6.3
Posterior Alpha2 �0.8 8.1
Alpha3 0.2 6.7
Above median Frontal Alpha2 0.9 7.5
Alpha3 �1.8 6.9
Central Alpha2 3.9 7.8
Alpha3 5.7 6.3
Posterior Alpha2 4.1 8.1
Alpha3 11.3 6.7
is small at central sites, and reverses to opposite within
frontal area.
The MAS�RPW�band interaction is depicted at Fig.
2 and respective marginal means and standard errors are
presented in Table 2. In low-anxiety subjects, only alpha3
during presentation of first ten words shows signs of ERD
(positive values). Alpha2 at the same time clearly
synchronizes. Repetitive presentation of the same words
diminishes magnitude of alpha2 synchronization and
alpha3 desynchronization. On the other hand, in high-
anxiety subjects, only alpha2 ERD during presentation of
second ten words reaches negative values (estimated
marginal mean=�1.67). Contrary to low-anxiety subjects,
in these subjects, in the beginning of words presentation,
alpha2 desynchronization prevails. It drops toward the
middle of the session and again rises during presentation
Table 2
Estimated marginal means and standard errors of alpha2 and alpha3 ERD
(%) during repeated presentation of words (RPW) in subjects with low and
high manifest anxiety (MAS)
MAS RPW Band Mean ERD Standard error
Below median First ten words Alpha2 �12.8 7.9
Alpha3 6.4 6.2
Second ten words Alpha2 �9.8 8.2
Alpha3 �7.7 7.6
Third ten words Alpha2 �5.5 8.2
Alpha3 �3.6 7.4
Above median First ten words Alpha2 7.3 7.9
Alpha3 6.5 6.2
Second ten words Alpha2 �1.7 8.2
Alpha3 4.4 7.6
Third ten words Alpha2 3.2 8.2
Alpha3 4.3 7.4
Alpha2
0
0.5
1
1.5
2
2.5
3
3.5
Bsl 1tw 2tw 3tw
Po
wer
den
sity
Low MAS
High MAS
Fig. 3. Effect of repeated presentation of words on alpha2 power density
during reference interval in subjects with low and high Manifest Anxiety
(MAS). Bsl—Baseline; 1tw—First ten words; 2tw—Second ten words;
3tw—Third ten words.
G.G. Knyazev et al. / International Journal of Psychophysiology 59 (2006) 151–158156
of third ten words. Alpha3 ERD scarcely changes during
all stages of words presentation remaining within the range
of positive values.
Next, repeated measures ANOVAs were conducted
separately for each alpha sub-band with pre-stimulus alpha
power density as a dependent variable. In this case the RPW
factor consisted of four levels, since baseline period (2 s
before presentation of the first word) was added. For alpha3,
apart from the significant main effects of MAS (F =7.77,
df =1, P=0.009) indicating higher alpha3 power density in
anxious subjects, only zone�MAS interaction emerged as
significant (F =7.77, df =1, P=0.009). This interaction
showed that alpha3 power density, being higher in anxious
subjects at all cortical sites, was particularly high within
posterior region. For alpha2 power density, the main effect
of MAS was also significant (F =6.34, df=1, P=0.018).
There was also a significant interaction RPW�MAS
(F =3.92, df =1.31, P=0.045), which is depicted at Fig. 3.
In low-anxiety subjects, the alpha2 power density did not
change during the whole session. In high-anxiety subjects, it
substantially increased in the beginning of words presenta-
tion and continued to increase till the middle of the session,
only then starting to decrease.
Finally, we analyzed time-locked evoked alpha activity
in the form of averaged AEP following presentation of
acoustic stimuli (tone 1000 Hz). In these analyses alpha2
and alpha3 sub-bands were not distinguished and were
treated as alpha band. First, we tested whether the amplitude
of averaged and alpha band-pass filtered AEP is lower in
anxious subjects and subjects with higher baseline alpha
power which would be expected basing on published
within-subject observations (Basar, 1998). Actually AEP
amplitude was positively correlated with the anxiety
measure (r =0.33, P =0.077, r =0.46, P =0.013, and
r =0.53, P=0.003, for O1, Cz, and T3, respectively). It
was also strongly positively associated with mean alpha
power during respective session (r =0.76, P <0.001,
r =0.87, P <0.001, and r=0.69, P <0.001, for O1, Cz, and
T3, respectively).
4. Discussion
On the whole, predictions derived from the inhibition
theory were not confirmed in the present study. First, there
were no signs of alpha synchronization within non-relevant
(i.e. posterior) cortical areas. Actually these areas showed
the most marked alpha desynchronization. Correspondingly
there were no evidences that alpha synchronization within
non-relevant regions was most pronounced in the beginning
of words presentation and was more marked in anxious
subjects, which would be expected on the assumption of
inhibition theory. Except for alpha2 in low-anxiety subjects,
alpha desynchronization was most pronounced in the
beginning of words presentation at all cortical sites, and
except for alpha3 in the beginning of words presentation, it
was more marked in high-anxiety subjects. Particularly
alpha3 ERD in anxious subjects was considerably higher
within posterior region. Pre-stimulus alpha3 power density,
being higher in anxious subjects at all cortical sites, was also
particularly high within posterior region. It is possible to
speculate that for high-anxiety subjects, being fixed in
unfamiliar environment with eyes closed is especially linked
with anxious visual imagery and preparedness for process-
ing of visual information. Therefore presentation of even
acoustic stimuli provokes marked desynchronization within
this area.
All these findings more correspond to predictions
derived from the alternative interpretation, which posits
that higher alpha power signifies higher readiness of alpha
system and should be associated with higher alpha
reactivity. We have shown recently (Knyazev et al.,
2004a,b) that in high-anxiety subjects, alpha2 sub-band
seems to be the most reactive, whereas low-anxiety
subjects tend to adjust to environmental challenges by
alpha3 power modification. This has been interpreted so
that low-anxiety subjects tend to react to unexpected
events by increase of specific attention in attempt to
understand the meaning of a happening. That would imply
use of semantic memory and hence alpha3 activation.
Those high on anxiety react by increase of unspecific
attention (alpha2 activation), which is an evidence of their
higher general vigilance. This difference is most pro-
nounced in the beginning of words presentation (see Fig.
2). Whereas in low-anxiety subjects, alpha3 ERD clearly
prevails, in high-anxiety subjects, alpha2 ERD is higher at
this time. Changes of alpha2 ERD in these subjects are
accompanied by respective changes of pre-stimulus alpha2
power, which increases from baseline to the middle of the
session, only then starting to decrease. This corresponds to
the notion that on-going EEG might act as a means for
regulation reactivity (Basar, 1998). Comparison of the
ERD distribution across cortical sites in the two groups of
subjects also shows that in subjects with low anxiety,
prevalence of specific (alpha3 ERD) over unspecific
(alpha2 ERD) activation is most marked within specific
(i.e. central) zone whereas in high-anxiety subjects it is
G.G. Knyazev et al. / International Journal of Psychophysiology 59 (2006) 151–158 157
most marked within unspecific (i.e. posterior) region (see
Fig. 1). That implies that low-anxiety subjects process the
stimuli in a more economic way, which suffices in this
simple and undemanding situation. High-anxiety subjects
however invest greater resources, perhaps searching for
information, which is not directly related to perceived
acoustic stimuli. The stimuli used in the present study were
not sufficient to necessitate an increase of alpha system
preparedness (hence preparatory alpha synchronization) in
low-anxiety subjects. But in high-anxiety subjects, the
beginning of words presentation provoked an increase of
alpha2 power.
Our data show that subjects with higher background
alpha power tend to show higher AEP amplitude both at
specific and non-specific cortical sites (although correla-
tion of anxiety with AEP amplitude at O1 only
approached significance). As we have discussed in
Introduction, there are no reasons to expect that the
phase-locked alpha amplitude should be negatively related
to the amplitude of pre-stimulus alpha oscillations since it
represents phase-locking of these same oscillations. Since
higher amplitude of the phase-locked alpha response to
sensory stimuli is generally considered as an evidence of
higher perceptual sensitivity (Basar, 1998, 1999), the
present study findings indicate that subjects with higher
alpha power (particularly high-anxiety subjects) are more
perceptually sensitive.
On the whole, this study findings conform to the idea
that enhanced alpha oscillations should not be considered
as a sign of inhibition. On the contrary, they reflect a state
of enhanced preparedness of corresponding networks to
information processing. Particularly in high-anxiety sub-
jects, higher alpha power coincides with higher perceptual
sensitivity (as reflected in higher amplitude of phase-
locked response) and predisposition to higher alpha
desynchronization in response to even neutral stimuli. This
is in keeping with well-known evidences of a greater effort
investment (e.g. Brocke et al., 1996) and an increased
receptivity or cortical excitability of the nervous system to
afferent stimuli (Lacey and Lacey, 1974) in these subjects.
There is no reason however to allege that alpha enhance-
ment is specifically linked with anxious endophenotype
and anxious states. It might be expected that any state with
enhanced attention and concentration should be associated
with alpha enhancement and individuals who are predis-
posed to these states should tend to show higher alpha
power in the reference interval and higher magnitude of
alpha desynchronization.
Acknowledgements
This study was supported by a grant of the Russian
Foundation for Basic Research # 05-06-80033-a. We are
grateful to D.A. Savostyanova, L.G. Mitrofanova, and N.V.
Dmitrienko for assistance with data collection.
References
Annet, M.A., 1970. A classification of hand preference by association
analysis. Br. J. Psychol. 61, 303–321.
Basar, E., 1998. Brain function and oscillations: I. Brain oscillations.
Principles and Approaches. Springer, Berlin, Heidelberg.
Basar, E., 1999. Brain function and oscillations: II. Integrative brain
function. Neurophysiology and Cognitive Processes. Springer, Berlin,
Heidelberg.
Basar, E., Basar-Erogly, C., Karakas, S., Schurmann, M., 2000. Brain
oscillations in perception and memory. Int. J. Psychophysiol. 35,
95–124.
Bell, I.R., Schwartz, G.E., Hardin, E.E., Baldwin, C.M., Kline, J.P., 1998.
Differential resting quantitative electroencephalographic alpha patterns
in women with environmental chemical intolerance, depressives, and
normals. Biol. Psychiatry 43 (5), 376–388.
Berger, H., 1929. Uber das elecktroenzephalogramm des menschen I. Arch.
Psychiatr. Nervenkr. 87, 527–570.
Brocke, B., Tasche, K.G., Beauducel, A., 1996. Biopsychological founda-
tions of extraversion: differential effort reactivity and the differential
P300 effect. Pers. Individ. Differ. 21, 727–738.
Crick, F., 1984. Function of the thalamic reticular complex: the searchlight
hypothesis. Proc. Natl. Acad. Sci. U. S. A. 81, 4586–4590.
Doppelmayr, M., Klimesch, W., Pachinger, T., Ripper, B., 1998. Individual
differences in brain dynamics: important implications for the calculation
of event-related band power. Biol. Cybern. 79, 49–57.
Gray, J.A., 1987. Perspectives on anxiety and impulsivity. J. Res. Pers. 21,
493–509.
Hanin, Y.L., 1989. Cross-cultural perspectives of the individual differences
diagnostic. Vopr. Psihol. 4, 118–125.
Herrmann, W.M., Winterer, G., 1996. Electroencephalography in psychia-
try—current status and outlook. Nervenarzt 67, 348–359.
Klimesch, W., 1999. EEG alpha and theta oscillations reflect cognitive
and memory performance: a review and analysis. Brain Res. Rev. 29,
169–195.
Klimesch, W., Doppelmayr, M., Schwaiger, J., Auinger, P., Winkler, T.,
1999. Paradoxical’ alpha synchronization in a memory task. Brain Res.
Cogn. Brain Res. 7, 493–501.
Klimesch, W., Doppelmayr, M., Rohm, D., Pollhuber, D., Stadler, W., 2000.
Simultaneous desynchronization and synchronization of different alpha
responses in the human electroencephalograph: a neglected paradox?
Neurosci. Lett. 284 (1–2), 97–100.
Klimesch, W., Sauseng, P., Gerloff, C., 2003. Enhancing cognitive
performance with repetitive transcranial magnetic stimulation at human
individual alpha frequency. Eur. J. Neurosci. 17 (5), 1129–1133.
Knyazev, G.G., Slobodskaya, H.R., 2003. Personality trait of behavioral
inhibition is associated with oscillatory systems reciprocal relationships.
Int. J. Psychophysiol. 48, 247–261.
Knyazev, G.G., Slobodskaya, H.R., Wilson, G.D., 2002. Psychophysio-
logical correlates of behavioural inhibition and activation. Pers. Individ.
Differ. 33, 647–660.
Knyazev, G.G., Slobodskaya, H.R., Safronova, M.V., Sorokin, O.V.,
Goodman, R., Wilson, G.D., 2003. Personality, psychopathology and
brain oscillations. Pers. Individ. Differ. 35 (6), 1331–1349.
Knyazev, G.G., Savostyanov, A.N., Levin, E.A., 2004a. Alpha
oscillations as a correlate of trait anxiety. Int. J. Psychophysiol. 53
(2), 147–160.
Knyazev, G.G., Slobodskaya, H.R., Wilson, G.D., 2004b. Comparison of
construct validity of the Gray-Wilson personality questionnaire and the
BIS/BAS scales. Pers. Individ. Differ. 37, 1565–1582.
Lacey, J.I., Lacey, B.C., 1974. Some autonomic–central nervous system
interrelationships. In: Black, P. (Ed.), Physiological Correlates of
Emotion. Academic Press, New York.
Neubauer, A.C., Sange, G., Pfurtscheller, G., 1999. Psychometric intelli-
gence and event-related desynchronization during performance of a
letter matching task. In: Pfurtscheller, G., Lopes da Silva, F.H. (Eds.),
Event-related Desynchronization and Related Oscillatory EEG-Phe-
G.G. Knyazev et al. / International Journal of Psychophysiology 59 (2006) 151–158158
nomena of the Awake Brain, Handbook of EEG and Clinical Neuro-
physiology, Revised Series, vol. 6. Elsevier, Amsterdam, pp. 219–231.
Pfurtscheller, G., 2003. Induced oscillations in the alpha band: functional
meaning. Epilepsia 44 (Suppl. 12), 2–8.
Pfurtscheller, G., Aranibar, A., 1977. Event-related cortical desynchroniza-
tion detected by power measurement of scalp EEG. Electroencephalogr.
Clin. Neurophysiol. 42, 817–826.
Spielberger, C.D., Gorsuch, R.L., Lushene, R.E., 1970. Manual for
the state-trait anxiety inventory. Consulting Psychologists Press, Palo
Alto, CA.
Suffczynski, P., Kalitzin, S., Pfurtscheller, G., Lopes da Silva, F.H., 2001.
Computational model of thalamo-cortical networks: dynamical control
of alpha rhythms in relation to focal attention. Int. J. Psychophysiol. 43,
25–40.
Taylor, J.A., 1953. A personality scale of manifest anxiety. J. Abnorm. Soc.
Psychol. 48, 285–290.