International Journal of Psychophysiology 65 (2007) 252 260

www.elsevier.com/locate/ijpsycho

Changes of EEG spectra and coherence following performance in a

cognitive task in Alzheimer's disease
Zoltn Hidasi a , Balzs Czigler b , Pl Salacz a , va Csibri a , Mrk Molnr b,
a

Department of Psychiatry and Psychotherapy, General Medical Faculty, Semmelweis University, Budapest, Hungary
b
Institute of Psychology, Hungarian Academy of Sciences, Budapest, Szondi u. 83-85. Hungary
Received 27 October 2006; received in revised form 2 March 2007; accepted 3 May 2007
Available online 13 May 2007

Abstract
Electroencephalographic measures combined with cognitive tasks are widely used for the assessment of cognitive and pathophysiological
changes in Alzheimer's disease (AD). Instead of the analysis of EEG data obtained during the performance of the task, in this study data recorded
in the immediate after-task period were analyzed. It was expected that this period would correspond to the electrophysiological consequences of
the cognitive effort. Data of 14 patients with AD (MMS score: 1624) were compared to that of 10 healthy control subjects. Reverse counting of a
fix duration was used as a cognitive task. Changes of relative frequency spectra, and those of inter-and intrahemispheric coherence were analyzed.
Relative theta power was significantly higher in AD patients compared to the controls both before and after the task. The performance of the task
resulted in an increase of the relative alpha2 band in the AD group, whereas it slightly decreased in the control group.
The most prominent coherence differences between AD and controls were found in the alpha1 band, especially for long-range coherence
values. Coherence in this frequency band increased in the control group following the task, not seen in the AD group. We conclude that EEG
parameters calculated from epochs following the completion of a cognitive task clearly differentiates patients with AD from normal controls. The
electrophysiological changes found in AD may correspond to the decrease of functional connectivity of cortical areas and to the malfunctioning of
the networks engaged in the cognitive task investigated.
2007 Elsevier B.V. All rights reserved.
Keywords: Alzheimer's disease; EEG frequency spectra; EEG coherence; Cognitive task; Mental arithmetic

1. Introduction
Detection of functional brain abnormalities, including the
impairment of cognitive processes in the early stages of
Alzheimer's disease (AD) and mild cognitive impairment
(MCI) is an important aspect for the diagnosis and effective
treatment of the disease. The EEG appears to be a particularly
sensitive method to detect and track functional changes in the
brain including those caused by deterioration of neuronal
connectivity and remains one of the best techniques in the
diagnosis (Claus et al., 1999), staging (Kowalski et al., 2001)
and prediction of the progress of AD Jonkman (1997).
According to the rather widely held view the development of
Corresponding author. Tel.: +36 1 3542 290; fax: +36 1 3542 416.
E-mail address: molnar@cogpsyphy.hu (M. Molnr).
0167-8760/$ - see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.ijpsycho.2007.05.002

AD is associated with the increase of theta and delta power and

the decrease of faster (alpha and beta) activities (Bennys et al.,
2001). Significant changes of mean frequency were found in
several studies, in correlation with psychometric scores
(Primavera et al., 1990; Gueguen et al., 1991; Schreiter-Gasser
et al., 1994; Claus et al., 1998; Briel et al., 1999 etc).
The question of staging of AD is an important issue.
Characteristic EEG spectral profile changes were found in
different stages of Global Deterioration Scale (GDS) (Rodriguez et al., 1999). EEG markers in the alpha and delta frequency
range were suggested for the prediction of the course of AD
(Nobili et al., 1999). Topographical spectral power changes,
especially occipital peak frequency, was related to MMSE
scores, and showed non-overlapping features in MCI, AD and
normal controls (Kwak, 2006). Cortical sources of EEG
rhythms changed across normal aging population, MCI, and

Z. Hidasi et al. / International Journal of Psychophysiology 65 (2007) 252260

mild AD subjects as a function of the global cognitive level and

were correlated with MMSE score in all subjects with
supposedly predictive value (Babiloni et al., 2006a) and was
found to correlate also with lobar brain volume (Babiloni et al.,
2006b). Alpha power of the EEG proved to be closely linked to
cognitive function and regional cerebral blood flow (rCBF),
while power in the beta and theta bands were suggested to be
related to subcortical changes (Claus et al., 2000).
Hypoperfusion verified by perfusion MRI was related to an
increase of EEG theta power and decrease in alpha power in AD
patients compared to control subjects (Mattia et al., 2003). The
amplitude of alpha1 and alpha2 sources using low-resolution brain
electromagnetic tomography in several brain regions was lower
both in AD and MCI, associated with carrying the apolipoprotein E
epsilon4 allele, which phenomenon might be used for the early
prediction of AD conversion in MCI (Babiloni et al., 2006c).
As multimodal information processing at the level of corticocortical projections are affected in AD, a hypothetical model of
disconnection syndrome for AD symptomatology was suggested (Morrison et al., 1996; Delbeuck et al., 2003). Methods
sensitive to electrophysiological characteristics of interneuronal
connectivity, such as coherence proved to be useful tools in the
investigation of AD, showing abnormal features in various
frequency bands and conditions (Stevens et al., 2001; Adler
et al., 2003; Pogarell et al., 2005.). In AD patients lower
interhemispheric and intrahemispheric EEG coherence was
found (Knott et al., 2000). Decreased temporo-parietal
coherence appeared to be a discriminant variable between AD
patients and controls (Jelic et al., 1996). A decrease in alpha coherence and an increase in delta coherence were found to significantly
correlate with the degree of dementia (Brunovsky et al., 2003).
Lower sychronization likelihood (SL, a method sensitive for both
linear and nonlinear synchronization) in the beta band was found to
correlate with MMS scores in AD (Stam et al., 2003) and the mean
level of sychronization and spontaneous fluctuations in synchronization were found to be lower in AD compared to normal controls in
the alpha and beta bands (Stam et al., 2005).
Magnetoencephalography (MEG) is also successfully used to
detect electrophysiological changes in AD. Increased power
scores in the low frequency bands, and reduced power scores in
the high frequency bands were found in AD, and also the decrease
of mean frequency of MEG in AD and MCI (Fernandez et al.,
2006a,b). Source distribution of MEG alpha activity is not
significantly changed in MCI, unlike that found in AD (Osipova
et al., 2006). Decreased synchronization likelihood and coherence
of MEG signals in various frequency bands were analyzed and
evaluated with respect to possible compensatory mechanisms in
AD (Stam et al., 2006). Increased SL was found in the theta and
beta bands and short-range coherence showed a similar pattern of
increase in AD in alpha2, beta and gamma bands. However, SL
decreased in the alpha1 band for long-range intrahemispheric
connections. Reduced MEG alpha rhythm (914 Hz), pre-alpha
rhythm (79 Hz), increased slow band (37 Hz) reactivity and
decrease of alpha band coherence, more pronounced in the long
distance derivations were found in AD (Franciotti et al., 2006).
According to Gnther et al (1993) the analysis of the
spontaneous EEG recorded at rest without perceptivecognitive

253

demand is not sensitive enough to produce significant group

differences in spectral EEG parameters in AD patients. In this
study manual-motor and music perception tasks elicited a
decrease of EEG power in the delta frequency range compared
to resting activity, not seen in age matched and younger control
groups. In some recent studies the usefulness of resting EEGanalysis is emphasized in AD and MCI patients (Babiloni et al.,
2006a,c; Kwak, 2006), considering staging and prognostic
aspects as well. The use of mental tasks improves the classification accuracy of subjects with normal aging, MCI and AD,
with the hypothesis that there are pronounced differences
between the cognitive tasks used for this purpose (Sneddon
et al., 2005). In a memory activation task EEG power differences were revealed in the lower alpha band between MCI
patients and controls, while in the resting conditions no such
differences were observed (van der Hiele et al., 2007). Changes
in event related desynchronization/synchronization (ERD/ERS)
of alpha and beta rhythms with abnormal ERD/ERS in the
frontolateral, centromedial and ipsilateral rolandic areas were
found in AD patients (Babiloni et al., 2000). The method of
ERD/ERS, using auditory verbal memory task was able to
provide information about compensatory mechanisms in MCI
and failure of lexical semantic processing in AD (Karrasch
et al., 2006). Differences in theta ERS was found useful to
differentiate patients in a progressive MCI group from clinically
stable MCI patients using working memory task (Missonnier
et al., 2006). Application of a complex perceptivecognitive
haptic task revealed significant differences in theta power
between AD and in patients with MCI, and in healthy controls
(Grunwald et al., 2002).
In all of the aforementioned studies using cognitive tasks
baseline conditions of various EEG parameters were compared
to those recorded during cognitive activity. This is only possible
in case of well cooperating subjects, otherwise numerous
artifacts render data analysis nearly impossible. Difficulty to
perform in a given task may also limit the number of subjects
participating in such a study. These difficulties may be the
reason why no simple electrophysiological method is known for
the assessment of cognitive functions of neuropsychiatric
patients for everyday practice. A simple but also rather challenging task is reverse counting, a quite demanding cognitive effort
(Warrington, 1982; McCloskey et al., 1985; Caramazza and
McCloskey, 1987; Roselli and Ardila, 1989; Gnther et al.,
1993; Harmony et al., 1999), which can still be executed by
patients with various degrees of cognitive decline.
The analysis of electrophysiological data obtained after the
performance of a cognitive task may provide information about
the dysfunction of neuronal systems involved in the type of
cognitive activity investigated. To the best of our knowledge, no
data are known in the literature concerned with the analysis of
immediate post-task EEG activity following cognitive effort of
short duration. In a study involving a particularly long session,
enhanced theta and beta2 band power was found after sustained
(70 min) mental effort (Smit et al., 2004).
The basic rationale of the study was the assumption that the
short term after-effects of mental effort as manifested in the
EEG would be substantially different in AD patients compared

254

Z. Hidasi et al. / International Journal of Psychophysiology 65 (2007) 252260

to those observed in control subjects and as such, can quantitatively be presented. Thus, it was assumed that 1) the state of
the subject as reflected by the electrophysiological activity
recorded immediately after the performance of a cognitive task
reflects the consequences of the effort required by the task and
2) that the fixed duration and level of difficulty of the task
represented an approximately equal load for all subjects within
the groups making the interindividual comparison of the effect
of the task possible.
Our primary hypothesis was that differences observed in the
electrophysiological parameters (spectral parameters and
changes of coherence in the present study) between the before-task, and after-task conditions will be markedly different in
AD patients compared to those seen in controls. We expected to
find less prominent changes in the AD patients than those seen
in normal controls since the available cognitive capacity is
presumably decreased in AD. In addition, the neural circuitry
involved in the cooperative activity of interactive neuronal
systems during cognitive processes is presumably impaired in
AD depending upon the stage of the disease.
2. Materials and methods
2.1. Patients and subjects
The study was approved by the Medical Ethical Committee
of the Semmelweis University, Budapest. Participants signed an
informed consent after being given complete explanation about
the protocol and the purpose of the study.
The group of AD patients consisted of 8 women and 6 men.
Their mean age was 67.4 years (range 5879), and their average
MMS score was 20.21 (range: 1624). The group of control
participants consisted of 6 women and 4 men. Their mean age
was 67.2 years (range: 5578), average MMS score was 29.8
(range: 2930).
Inclusion criteria for the AD group were based on those
defined by NINCDS-ADRDA, score on mini mental state
(MMS) between 15 and 24 (inclusive), age 50 years or older,
and stable medication history for one month prior to the beginning of the study. Exclusion criteria were the following: possible causes of dementia other than AD, evidence of mass lesions
(except atrophy) on computer tomography (CT) or magnetic
resonance imaging (MRI), signs of probable vascular dementia,
marked depression (Hamilton Depression Rating Scale N 15).
No subject received psychotropic medication, for at least three
weeks before the beginning of the study. Inclusion criteria for
the healthy control group were: ageN 50 years, absence of any
neurological and psychiatric diseases and no treatment with
drugs of psychotropic effect.

head. The impedance was kept below 5 k. Of the continuously

recorded EEG 2560 ms (512 data points) long EEG-epochs
were made and subjected to automatic artifact screening, during
which epochs exceeding +/ 70 V were rejected. Visual artifact screening was also performed to ensure the epochs were
free of other types of artifacts (muscle activity, etc.) as well.
Data acquisition started with the recording of spontaneous
resting EEG for two minutes with eyes closed (before-task
period). Following this, a 20-second recording was performed
with eyes open for the stabilization of vigilance level. This was
followed by performing the cognitive task (reverse counting
[RC], see below) lasting for 45 s. Another 2-minute EEG
recording epoch followed which was the EEG section evaluated
for immediate post task changes (after-task period). The
session ended by an epoch lasting for 20 s with eyes open and
then hyperventilation for 3 min.
2.3. Reverse counting task
The subjects had to count backwards from 100 by 7 for 45 s.
A detailed description of the method used is given by Rajna
et al. (2003) and will be only briefly described here. The task
was outlined to the participants before the beginning of the EEG
recording. During the EEG recording the subjects were
instructed verbally when to start counting. The subjects performed the task with eyes closed, and listed aloud the results of
each step of the counting. Perfect performance was not the goal
during reverse counting, still five consecutive mistakes or
breaks longer than 5 s would have excluded that subject from
the study. This way a continuous mental effort was achieved and
could be monitored, in harmony with the description of the
original article (Rajna et al., 2003). All subjects both in the AD
and control group were able to perform the task. Errors committed during the procedure were documented but were not
analyzed afterwards.

2.2. EEG recording

The EEG was recorded by 19 AgAgCl electrodes placed
according to the international 1020 system using BRAINLIFT
21-5 (Medicor, Micromed, Hungary) amplifiers (sampling rate:
200 Hz, bandpass: 0.145 Hz). Linked ear electrodes were used
as reference. The ground electrode was positioned on the fore-

Fig. 1. Changes of the relative theta power in the two groups before and after the
completion of the counting task. Group differences were significant in both
conditions. (AD: AD patients, C: controls).

Z. Hidasi et al. / International Journal of Psychophysiology 65 (2007) 252260

255

Table 1
Significant group differences of coherence in the different conditions and
frequency bands
Type of
coherence

alpha1
Before

Interhemispheric
Intrahemispheric
short-range

Intrahemispheric
long-range

Fig. 2. Changes of the relative beta1 band in the two groups before and after the
completion of the reverse counting task. (AD: AD patients, C: controls).

2.4. Data analysis

The analyses were performed in the following frequency
bands (delta: 0.54 Hz, theta: 48 Hz, alpha1: 811 Hz,
alpha2:1114 Hz, beta1:1425 Hz, beta2: 2535 Hz). Relative
frequency spectra were calculated by means of Fast Fourier
Transform for the above bands by the Neuroscan 4.3 software.
Coherence was computed with the same software for interhemispheric (F7F8, F3F4, T3T4, T5T6, C3C4, P3P4, O1O2)
and intrahemispheric short-range (F7T3, F8T4, F3C3, F4C4,
T3T5, T4T6, C3P3, C4P4, P3O1, P4O2, FZCZ, CZPZ)
and long-range (F3T3, F4T4, F3T5, F4T6, F3P3, F4P4,
F3O1, F4O2, T3P3, T4P4, T3O1, T4O2) electrode pairs.

T3T4
T4T6
C4P4
P3O1
P4O2
F3T3
F4T4
F3T5
F3P3
F4P4
F4O2
T4P4
T3O1

alpha2
After

Before

beta2
After

Before

After

4.47

6.19
5.26

6.20

9.01

5.24
8.69
8.80
4.81

4.85
12.82
8.16
5.65
7.58

5.60

4.34

4.47
4.73
8.88
5.10

5.42

5.03
9.25
5.47

The numbers shown are F-values. The critical F-values for df(1,22) are 4.3 and 7.95
for p b 0.05 and p b 0.01, respectively. The F-values where coherence was lower in
AD patients compared to controls are shown in italics. For all other F-values,
coherence was higher in AD patients than in controls.

given frequency bands. Power spectra and alpha peak frequencies were measured separately at the O1 and O2 electrodes.
Group statistics were computed based on these averages.
Two-way Multivariate Analyses of Variance (MANOVAs) were
calculated in Group (AD vs. Control) Condition (Before vs.
After task) design for the relative power and coherence values
on each and every frequency band in both conditions. Following
MANOVAs ANOVAs were also calculated in the same twoway and one-way between groups (AD vs. Control) designs to
reveal between group differences for each variable.
3. Results

2.5. Statistical analysis

3.1. Frequency spectrum analysis
Individual means for the relative spectra were calculated by
averaging the values measured for each EEG channel for the

Significant differences were found in the theta band between

the two groups. The relative theta was higher in the AD group
compared to the controls both before and after the calculation
Table 2
Long-range alpha1 coherence before and after the reverse counting task
alpha1
F3T3
F4T4
F3T5
F4T6
F3P3
F4P4
F3O1
F4O2
T3P3
T4P4
T3O1
T4O2

Fig. 3. Changes of the relative frequency spectra of the delta, alpha2, beta1
bands before and after task in AD patients (AD) and in the controls (C).

AD

Control

Before

After

Difference

Difference

Before

After

0.093
0.161
0.201
0.732
0.326
0.109
0.443
0.137
0.652
0.664
0.637
0.473

0.08
0.155
0.194
0.705
0.302
0.113
0.424
0.138
0.645
0.679
0.662
0.496

0.013
0.006
0.007
0.027
0.024
0.004
0.019
0.001
0.007
0.015
0.025
0.023

0.034
0.042
0.079
0.017
0.066
0.048
0.06
0.055
0.03
0.055
0.025
0.085

0.112
0.257
0.225
0.706
0.337
0.149
0.44
0.197
0.659
0.62
0.577
0.421

0.146
0.299
0.304
0.723
0.403
0.197
0.5
0.252
0.689
0.675
0.602
0.506

Asterisks denote the sites of significant Condition Group interactions.

Differences denote before-task minus after-task values.

256

Z. Hidasi et al. / International Journal of Psychophysiology 65 (2007) 252260

long-range electrode-pairs: in the controls coherence increased

in all considered electrode-pairs as a result of the calculation
task which was significant at the F4T4, F3T5, F3P3, F4
P4, F3O1, F4O2, T4O2 electrode-pairs. Contrary to this, no
significant increase was found in the AD group following task
execution. Furthemore, coherence showed a small (nonsignificant) decrease after the task (Table 2., Fig. 4.). The
Group Condition interaction in the MANOVA performed on
the 12 long-range electrode pairs proved to be significant (F
(1,22) = 8.11, p b 0.01).
4. Discussion

Fig. 4. Changes of the long-range coherence values (shown as averages) in the

alpha1 band in the two groups before and after the reverse counting task shown
in Table 2. In the Control group (C) the coherence between long-range
intrahemispheric electrode-pairs was significantly higher in the after task
period but it was slightly lower in the AD group (AD).

task (F(1,22) = 10.49; p b 0.004 and F(1,22) = 6.99; p b 0.02,

respectively) (Fig. 1.).
Compared to the control group, a marginally significant
decrease of the beta1 band was found in the patients before the
task (F (1,22) = 3.59; p b 0.07). The same tendency was found
following performance of the task, but in this case the difference
was not significant. (Fig. 2.)
A significant Group Condition interaction was found in the
alpha2 band (F(1,22) = 4.2594, p b 0.05). Following the task, the
amount of alpha2 band increased in the AD patients whereas it
showed a small, non-significant decrease in the controls. The
same tendency could be seen in the beta1 band (F(1, 22) = 3.3559,
p b 0.08). In the delta band a marginally significant interaction
was observed (F(1, 22) = 2.8615, p b 0.10). After the completion
of the task, the delta power increased in the controls, but decreased
in the AD group. No such interaction was found in the theta,
beta2, and alpha1 bands (Fig. 3.).
Although alpha power measured on the O1 and O2
electrodes was not different in the two groups, a significantly
lower alpha frequency peak was found in AD than that seen in
the controls (F (1, 22) = 5.3127, p b 0.05). However, no significant Group Condition interaction of the alpha peak frequency
as a result of task performance was observed.
3.2. Coherence
The electrode-pairs where significant group-differences were
found in the apha1, alpha2 and beta2 bands in both conditions
are shown in Table 1.
Compared to that seen in the controls coherence was lower in
AD in the alpha1 band but only in the after task condition.
Coherence was higher in the alpha2 and beta2 bands in AD in
both conditions.
Significant Group Condition interactions were found in the
alpha1 frequency band for the majority of intrahemispheric

In the present study spectral and coherence changes of the

EEG were analyzed in a group of AD patients and a group of
healthy controls before and after the performance in a cognitive
task. Our primary hypothesis was that the two groups will differ
as before-task and after-task conditions are compared, in which
conditions available resource capacities engaged in the task are
different in the controls and in AD patients. Because the number
of participants was relatively low, only preliminary conclusions
can be drawn. Other limitations of the present study rendering it
to be a preliminary one include the following aspects: 1)
application of more than one task with varying difficulty would
likely yield additional data helping the staging of the patients, 2)
level of educational background of the participants in future
studies would similarly be desirable to be able to select
appropriate participants, to design cognitive tasks and to assess
the effect of mental effort.
Since in the present study EEG epochs recorded before and
after and not during the execution of a cognitive task were
analyzed, the neurophysiological processes underlying the observed changes are not directly comparable to data of the literature
where these were collected during cognitive performance.
4.1. Frequency spectrum analysis
Irrespective of recording conditions (before and after task)
significantly higher relative theta and marginally significant
lower beta1 were found in the AD group. Whereas the relative
alpha2 band increased in the AD group after the performance of
the task, it slightly decreased in the control group. The same
tendency was found for the beta1 and the opposite for the delta
frequency band. In general, there was an increase in fast (alpha2
and beta1), and a decrease in slow (delta) frequencies in the AD
group following the completion of the cognitive task.
The spectral differences seen in the theta, (posterior) alpha
and beta bands between the controls and AD patients, which
were unaffected by the cognitive effort accomplished in the
present study are in agreement with numerous previous findings
indicating general slowing of the EEG in AD. Since episodes
that follow immediately after the execution of a cognitive task
had not been analyzed before, only tentative explanations can
be offered for our findings. The increase of the slow and
decrease of fast frequency activity occurring during the task
is a well known phenomenon in task-related EEG studies in
healthy subjects. Concerning complex memory-related tasks,

Z. Hidasi et al. / International Journal of Psychophysiology 65 (2007) 252260

theta power was found to increase as a result of theta-synchronization, while alpha band power to decrease as a result of alphadesynchronization with increased task difficulty (Klimesch,
1999a; Smith et al., 1999). Increase of delta power and generalized decrease of alpha power, associated with the decrease of
beta power was observed in a study evaluating the effects of
different components of mental calculation on the EEG in healthy
volunteers (Fernandez et al., 1995). Our results in the control
group, although not reaching the level of significance show the
same type of changes, that can be due to the fact that we analysed
the episode following and not during the task.
The task applied in our study was a combination of a mental
calculation and a working memory task. Mental calculation
itself is thought to represent a complex process, involving the
interaction of different cognitive mechanisms (Caramazza and
McCloskey, 1987). Mental arithmetic is also thought to include
activation of working memory (Hitch, 1978). According to our
data the effect of the task was the opposite in AD patients than
those seen in the controls in the delta, alpha2 and beta1 frequencies, possibly showing different responses of neural networks involved in task performance. Spectral analysis revealed
a connection between different kinds of mental calculation
processes and topographical changes of EEG activity in healthy
subjects (Harmony et al., 1999). According to this study an
increase of the 3.9 Hz activity within Broca's and left parietotemporal cortices corresponds to processes related to internal
speech, storage and rehearsal of working memory. The
increased activity of the 5.64 Hz frequency within the right
dorsolateral prefrontal cortex was suggested to be related to
sustained attention, while the decrease at the 12.46 Hz within
the left parietal cortex was interpreted as the sign of retrieval of
arithmetic facts from long term memory. In a study using
functional MRI in evaluating the effect of mental calculation
found induced activity in a cerebral network including the
supplementary motor area, the posterior parietal cortices
bilaterally, the Broca's area, the dorso-lateral prefrontal cortices
bilaterally, the cerebellum bilaterally and premotor cortices
bilaterally were found in healthy subjects. In AD patients the
inferior parietal and lateral prefrontal activations were significantly reduced compared to normal controls (Rmy et al.,
2004). Activation of the right ventral prefrontal cortex caused
by the emotional stress related to the performance of mental
arithmetic task (serial subtraction) was observed using perfusion MRI in healthy subjects (Wang et al., 2005).
Decreases of EEG power in the delta frequency range also
observed in our patient group as a result of performing mental
arithmetic during manual-motor tasks and music perception in
AD patients were described (Gnther et al., 1993). The
significant differences between the two groups of subjects
regarding before-and after-task conditions according to our
results was found in the alpha2 band, where the increase of the
relative power in the AD group might be a correlate of a
rebound effect, which was more pronounced than that seen in
the control group. The higher amount of relative alpha2 band
seen in AD patients following the task may correspond to a
long-lasting effect of the increased effort invested in the
execution of the task. This effect may represent some kind of

257

cognitive capacity characterizing in the AD group. Mobilization

of this reserve capacity might correspond to the preserved
cognitive abilities in AD of mild and moderate degree,
represented by the patients in the present study. An increased
power in alpha frequencies was interpreted as sign of enhanced
effort of AD patients during working memory task (Hogan
et al., 2003). The upper alpha band, especially its desychronization is generally thought to be related to search and retrieval
processes in semantic long-term memory, induced by thalamocortical feedback loops (Klimesch, 1999b).
4.2. Coherence
Significant differences in coherence values between the two
groups were found in the alpha1, alpha2 and beta2 bands.
Although the use of linked-ear reference (as opposed to common
average reference) may represent a bias from the technical point
of view, the identical recording conditions used in the two groups
of subjects allow a number of conclusions to be drawn. Since
differences in the alpha2 band were found only at one electrode
pair (P4O2) before and at another pair (T4P4) after the task, the
clinical relevance of these data is questionable, and will not be
discussed further on.
EEG coherence, being the covariance of spectral activity at
two electrode sites, is a measure for the synchrony of neuronal
activity and thus can be used as an indicator of effective cortical
connectivity. In AD patients decreased coherence was found for
the faster frequency bands, especially for alpha frequencies, and
in some cases also for the beta band (Locatelli et al., 1998;
Leocani and Comi, 1999; Knott et al., 2000; Stam et al., 2003).
In our study the most prominent coherence differences
between AD and controls were found in the alpha1 band.
Significant interactions were found in the alpha1 band for most
of the long-range electrode-pairs. According to the twocompartmental model of EEG-coherence (Thatcher et al.,
1986) long-range EEG coherence is based upon long axonal
cortical connectivity, and thus depends on interactions of distant
cortical areas. In AD long-range coherence seems to be more
affected than short-range coherence, while the opposite was
found in vascular dementia (Leuchter et al., 1992; Dunkin et al.,
1994). According to the same studies both state and trait factors
are involved in the long-range coherence changes in AD. Longrange alpha and beta coherence decrease was found in AD at
rest in fronto-central and occipito-parietal derivations (Calderon-Gonzalez et al., 2004). Significantly reduced alpha band
long-range intrahemispheric coherence in AD was attributed to
the impairment of fronto-parietal connections, such as the
superior longitudinal fascicle (Locatelli et al., 1998). According
to the above authors the decrease of coherence of the faster (alpha
and beta) bands begins in the earliest stages of AD, and is due to
the impairment of cortico-cortical networks. The increase of slow
band (delta and theta) coherence becomes evident later, and is
related to cortical deafferentation from subcortical structures.
According to Hogan et al. (2003), the increase of upper alpha
power (911 Hz) was observed with increasing memory
demands in normal controls, but this phenomenon was not
observed in AD patients. In the same study reduced alpha

258

Z. Hidasi et al. / International Journal of Psychophysiology 65 (2007) 252260

coherence was found in the same frequencies, compared to

normal controls (Hogan et al., 2003). The decrease of anterior
short-range alpha coherence and a simultaneous increase of
fronto-parietal long-range theta coherence was found in healthy
subjects during visuospatial working memory task (Sauseng
et al., 2005). These changes were attributed to the involvement
of prefrontal areas and the activation of fronto-parietal network
in executive functions. Taking into account the studies mentioned
above, it is not surprising that the changes of long-range alpha1
coherence were different in the two groups in our study, since the
effect of a rather complex cognitive task was evaluated.
The alpha band coherence seems to be affected in early
stages of AD. In harmony with earlier findings (Thatcher et al.,
1986; Calderon-Gonzalez et al., 2004; Locatelli et al., 1998;
Hogan et al., 2003; Stam et al., 2006; Franciotti et al., 2006), we
also found significant alpha1 band coherence changes in longrange derivations which were not specific for distinct brain
regions and thus might indicate diffuse impairment of longdistance cortical connectivity. These coherence changes may
have become apparent as a result of the demand represented by
the cognitive task, involving multiple neuronal systems. Our
findings are consistent with the neocortical disconnection
hypothesis of AD, related to the loss of structural and functional
integrity of long cortico-cortical tracts. We found a significant
increase in alpha1 coherence in the control group following the
execution of the cognitive task, which was not seen in the AD
group. In earlier studies where task-dependent EEG-changes
were investigated, the importance of alpha coherence both in
AD and healthy subjects were also confirmed, although the
interpretation of these results were rather controversial (Stevens
et al., 2001; Hogan et al., 2003; Sauseng et al., 2005). The
increase of alpha1 coherence in the control subjects may reveal
the level of functional connectivity of cortical areas, recruited
during the participation in the cognitive task. The unchanged
alpha1 coherence observed in the AD group may correspond to
the decrease of this connectivity.
Differences were found between the two groups in beta2
coherence both before and after task performance. These differences occurred in the long-range comparisons where higher
beta2 coherence values were observed in the AD group than in
the control group, irrespective of task performance. Increased
beta coherence has not been reported earlier in AD, and in our
case it was only present in the higher beta band. Furthermore,
since these beta2 coherence differences were present both
before-and after-task performance, they seem not to be affected
by cognitive effort.
In conclusion our findings seem to support our basic hypothesis that electrophysiological parameters calculated from
epochs following the completion of a cognitive task are able
to differentiate patients with AD from normal controls. With
respect to the spectral EEG parameters the significant differences between the groups appeared in the (relative) alpha2
band which increased in the AD group after the performance of
the task, while it slightly decreased in the control group.
Significant group differences in coherence measures following
the cognitive task were observed in the alpha1 band, exclusively
for the long-range coherence values for the majority of long-

range electrode-pairs. At most of these electrode pairs the

alpha1 coherence significantly increased in the control group
following the calculation task, but remained unchanged in the
AD group. Thus cognitive activity seems to affect both spectral
and coherence characteristics of the EEG which are conspicuous even after cognitive performance. The increase of the
relative alpha2 band power in AD patients may correspond to
compensatory mechanisms of the cognitive networks involved in
the task. From this point of view, besides helping in the detection of
functional impairment of different neural networks, these changes
in EEG spectra and coherence characteristics may yield valuable
information about reserve capacity of cognitive functioning in AD
patients with important diagnostic and therapeutic consequences.
The findings in spectral power and coherence values seem
not to be closely related to each other. Differences in coherence
between the controls and AD patients were found to be more
conspicuous than those seen for the spectral measures. This is
not surprising, since spectral and coherence attributes reflect
different aspects of the EEG, and they are mathematically
independent. As the underlying neurophysiological mechanisms of the findings related to the differences between the
controls and AD patients are not clear, the two types of analyses
may reflect different, perhaps complementary pathophysiological aspects. However, since coherence provides quantitative
information with respect to the functional connectivity between
different brain regions in a frequency-specific way, it appears to
be particularly informative from the point of view of the present
study since the pathology of Alzheimer's disease concerns
specifically this aspect. Further evaluation of EEG data
recorded following task performance in a larger pool of patients
is needed, both for the better understanding of neurophysiological processes underlying the mechanisms of the disease and
for ultimate diagnostic purposes in everyday clinical practice.
Acknowledgements
We thank Professor Peter Rajna (Department of Psychiatry
and Psychotherapy, General Medical Faculty, Semmelweis
University, Budapest, Hungary) for his valuable help in study
design and data collection.
Grant support: OTKA Hungarian Research Fund T048338.
NKFP Hungarian National Research and Developmental
Fund 1B/020/04.
References
Adler, G., Brassen, S., Jajcevic, A., 2003. EEG Coherence in Alzheimer's
dementia. J. Neural Transm. 110, 10511058.
Babiloni, C., Babiloni, F., Carducci, F., Cincotti, F., Del Percio, C., De Pino, G.,
Maestrini, S., Priori, A., Tisei, P., Zanetti, O., Rossini, P.M., 2000.
Movement-related electroencephalographic reactivity in Alzheimer disease.
NeuroImage 12 (2), 139146.
Babiloni, C., Binetti, G., Cassetta, E., Dal Forno, G., Del Percio, C., Ferreri, F.,
Ferri, R., Frisoni, G., Hirata, K., Lanuzza, B., Miniussi, C., Moretti, D.V.,
Nobili, F., Rodriguez, G., Romani, G.L., Salinari, S., Rossini, P.M., 2006a.
Sources of cortical rhythms change as a function of cognitive impairment in
pathological aging: a multicenter study. Clin. Neurophysiol. 117, 252268.
Babiloni, C., Frisoni, G., Steriade, M., Bresciani, L., Binetti, G., Del Percio, C.,
Geroldi, C., Miniussi, C., Nobili, F., Rodriguez, G., Zappasodi, F., Carfagna, T.,

Z. Hidasi et al. / International Journal of Psychophysiology 65 (2007) 252260

Rossini, P.M., 2006b. Frontal white matter volume and delta EEG sources
negatively correlate in awake subjects with mild cognitive impairment and
Alzheimer's disease. Clin. Neurophysiol. 117, 11131129.
Babiloni, C., Benussi, L., Binetti, G., Cassetta, E., Dal Forno, G., Del Percio, C.,
Ferreri, F., Ferri, R., Frisoni, G., Ghidoni, R., Miniussi, C., Rodriguez, G.,
Romani, G.L., Squitti, R., Ventriglia, M.C., Rossini, P.M., 2006c.
Apolipoprotein E and alpha brain rhythms in mild cognitive impairment: a
multicentric electroencephalogram study. Ann. Neurol. 59, 324324.
Bennys, K., Rondouin, G., Vergnes, C., Touchon, J., 2001. Diagnostic value of
quantitative EEG in Alzheimer's disease. Clin. Neurophysiol. 31 (3), 153160.
Briel, R.C., McKeith, I.G., Barker, W.A., Hewitt, Y., Perry, R.H., Ince, P.G.,
Fairbairn, A.F., 1999. EEG findings in dementia with Lewy bodies and
Alzheimer's disease. J. Neurol. Neurosurg. Psychiatry 66, 401403.
Brunovsky, M., Matousek, M., Edman, A., Cervena, K., Krajca, V., 2003.
Objective assessment of the degree of dementia by means of EEG.
Neuropsychobiology 48 (1), 1926.
Calderon-Gonzalez, P.L., Parra-Rodrigez, M.A., Llibre-Rodrigez, J.J., Gutierrez, J.V.,
2004. Spectral analysis of EEG coherence in Alzheimer's disease. Rev. Neurol.
38 (5), 422427.
Caramazza, A., McCloskey, M., 1987. Dissociation of calcuation process. In:
Deloche, G., Seron, X. (Eds.), Mathematical Disabilities. A cognitive
neuropsychological perspective. Lawrence Erlbaum, Hillsdale NJ, pp. 221234.
Claus, J.J., Kwa, V.I., Teunisse, S., Walstra, G.J.M., van Gool, W.A., Koelman, J.H.,
Bour, L.J., Ongerboer de Visser, B.W., 1998. Slowing on quantitative spectral
EEG is a marker for rate of subsequent cognitive and functional decline in early
Alzheimer disease. Alzheimers Dis. Assoc. Disord. 12, 167174.
Claus, J.J., Striers, R.L., Jonkman, E.J., Ongerboer De Visser, B.W., Jonker, C.,
Walstra, G.J., Scheltens, P., van Gool, W.A., 1999. The diagnostic value of
electroencephalography in mild senile Alzheimer'disease. Clin. Neurophysiol. 110 (5), 825832.
Claus, J.J., Ongerboer De Visser, B.W., Bour, L.J., Walstra, G.J., Hijdra, A.,
Verbeeten Jr., B., Van Royen, E.A., Kwa, V.I., van Gool, W.A., 2000.
Determinants of quantitative spectral electroencephalography in early
Alzheimer's disease: cognitive function, regional cerebral blood flow, and
computed tomography. Dement. Geriatr. Cogn. Disord. 11 (2), 8189.
Delbeuck, X., Van der Linder, M., Colette, F., 2003. Alzheimer's disease as a
disconnection syndrome? Neuropsychol. Rev. 13, 7992.
Dunkin, J.J., Leuchter, A.F., Newton, T.F., Cook, I.A., 1994. Reduced EEG
coherence in dementia: State or trait marker? Biol. Psychiatry 35, 970879.
Fernandez, T., Harmony, T., Rodrigez, M., Bernal, J., Silva, J., Reyes, A.,
Marosi, E., 1995. EEG activation patterns during the performance of tsks
involving different components of mental calculation. Electroencephalogr.
Clin. Neurophysiol. 94, 175182.
Fernandez, A., Hornero, R., Mayo, A., Poza, J., Maestu, F., Ortiz, T., 2006a.
Quantitative magnetoencephalography of spontaneous brain activity in
Alzheimer disease an exhaustive frequency analysis. Alzheimers Dis. Assoc.
Disord. 20 (3), 153159.
Fernandez, A., Hornero, R., Mayo, A., Poza, J., Gil-Gregorio, P., Ortiz, T.,
2006b. MEG spectral profile in Alzheimer's disease and mild cognitive
impairment, vol. 117, pp. 306314.
Franciotti, R., Iacono, D., Della Penna, S., Pizzella, V., Torquati, K., Onofrj, M.,
Romani, G.L., 2006. Cortical rhythms reactivity in AD, LBD and normal
subjects: a quantitative MEG study. Neurobiol. Aging 27, 11001109.
Grunwald, M., Busse, F., Hensel, A., Riedel-Heller, S., Kruggel, F., Arendt, T.,
Wolf, H., Gertz, H.J., 2002. Theta-power differences in patients with mild
cognitive impairment under rest condition and during haptic tasks. Alzheimers
Dis. Assoc. Disord. 16 (1), 4048.
Gueguen, B., Derouesne, C., Bourdel, M.C., Guillou, S., Landre, E., Gaches, J.,
Hossard, H., Ancri, D., Mann, M., 1991. Quantified EEG in the diagnosis of
Alzheimer's type dementia. Clin. Neurophysiol. 21, 357371.
Gnther, W., Giunta, R., Klages, U., Haag, C., Steinberg, R., Satzger, W., Jonitz,
L., Engel, R., 1993. Findings of electroencephalographic brain mapping in
mild to moderate dementia of the Alzheimer type during resting, motor, and
music-perception conditions. Psychiatry Res. 50, 163176.
Harmony, T., Fernandez, T., Silva, J., Bosch, J., Valds, P., Fernndez-Bouzas,
A., Galn, L., Aubert, E., Rodriguez, D., 1999. Do specific EEG frequencies
indicate different processes during mental calculation? Neurosci. Lett. 266,
2528.

259

Hitch, G.J., 1978. The role of short-term working memory in mental arithmetic.
Cogn. Psychol. 10, 302323.
Hogan, M.J., SwanWick, G.R.J., Kaiser, J., Rowan, M., Lawlor, B., 2003.
Memory-related EEG power and coherence reductions in mild Alzheimer's
disease. Int. J. Psychophysiol. 49, 147163.
Jelic, V., Shigeta, M., Julin, P., Almkvist, O., Winblad, B., Wahlund, L.O., 1996.
Quantitative electroencephalography power and coherence in Alzheimer's
disease and mild cognitive impairment. Dementia 7 (6), 314323.
Jonkman, E.J., 1997. The role of the electroencephalogram in the diagnosis of
dementia of the Alzheimer type: an attempt at technology assessment. Clin.
Neurophysiol. 27, 211219.
Karrasch, M., Laine, M., Rinne, J.O., Rapinoja, P., Sinerva, E., Krause, C.M., 2006.
Brain oscillatory responses to an auditory-verbal working memory task in mild
cognitive impairment and Alzheimer's disease. Int. J. Psychophysiol. 59,
168178.
Klimesch, W., 1999a. Event-related band power changes and memory performance.
In: Pfurtscheller, G., da Silva, L. (Eds.), Event-related desynchonization Handbook of Electroencephalography and Clinical Neurophysiology. Elsevier,
Amsterdam, pp. 161178.
Klimesch, W., 1999b. EEG alpha and theta oscillations reflect cognitive and
memory performance: a review and analysis. Brain Res. Rev. 29, 169195.
Knott, V., Mohr, E., Mahoney, C., Ilivitsky, V., 2000. Electroencephalographic
coherence in Alzheimer's disease: comparisons with a control group and
population norms. J. Geriatr. Psychiatry Neurol. 13, 18.
Kowalski, J.W., Gawel, M., Pfeffer, A., Barcikowska, M., 2001. The diagnostic
value of EEG in Alzheimer Disease: correlation with severity of mental
impariment. J. Clin. Neurophysiol. 18 (6), 570575.
Kwak, Y.T., 2006. Quantitative EEG findings in different stages of Alzheimer's
disease. J. Clin. Neurophysiol. 23, 456461.
Leocani, L., Comi, G., 1999. EEG coherence in pathological conditions. J. Clin.
Neurophysiol. 16 (6), 548555.
Leuchter, A.F., Newton, T.F., Cook, I.A., Walter, D.O., Rosenberg-Thompson,
S., Lachenbruch, P.A., 1992. Changes in brain functional connectivity in
Alzheimer-type and multi-infarct dementia. Brain 115, 15431561.
Locatelli, T., Cursi, M., Liberati, D., Franceschi, M., Comi, G., 1998. EEG coherence
in Alzheimer's disease. Electroencephalogr. Clin. Neurophysiol. 106 (3),
229237.
Mattia, D., Babiloni, F., Romigi, A., Cincotti, F., Bianchi, L., Sperli, F., Placidi, F.,
Bozzao, A., Giacomini, P., Floris, R., Grazia Marciani, M., 2003. Quantitative
EEG and dynamic susceptibility contrast MRI in Alzheimer's disease: a
correlative study. Clin. Neurophysiol. 114 (7), 12101216.
McCloskey, M., Caramazza, A., Basili, A., 1985. Cognitive mechanisms in
number processing and calculation: evidence from dyscalculia. Brain Cogn.
4, 171196.
Missonnier, P., Gold, G., Herrmann, F.R., Fazio-Costa, L., Michel, J.P., Deiber, M.P.,
Michon, A., Giannakopoulos, P., 2006. Decreased theta event-related
synchronization during working memory activation is associated with
progressive mild cognitive impairment. Dement. Geriatr. Cogn. Disord. 22,
250259.
Morrison, J.H., Rogers, J., Scherr, S., Levis, D.A., Campbell, M.J., Bloom, F.E.,
1996. The laminar and regional distribution of neocortical somatostatin and
neuritic plaques: implications for Alzheimers disease as a global neocortical
disconnection syndrome. In: Scheibel, A.B., Wechsler, A.F. (Eds.), The
biological substrates of Alzheimer's disease. Academic Press, Orlando.
Nobili, F., Copello, F., Vitali, P., Prastaro, T., Carozzo, S., Perego, G.,
Rodriguez, G., 1999. Timing of disease progression by quantitative EEG in
Alzheimer's patients. Clin. Neurophysiol. 16 (6), 566.
Osipova, D., Rantanen, K., Ahveninen, J., Ylikoski, R., Happola, O.,
Strandberg, T., Pekkonen, E., 2006. Source estimation of spontaneous
MEG oscillations in mild cognitive impairment. Neurosci. Lett. 405, 5761.
Pogarell, O., Teipel, S.J., Juckel, G., Gootjes, L., Moller, T., Burger, K.,
Leisinger, G., Moller, H.-J., Hegerl, U., Hampel, H., 2005. EEG coherence
reflects regional corpus callosum area in Alzheimer's disease. J. Neurol.
Neurosurg. Psychiatry 76, 109111.
Primavera, A., Novello, P., Finocchi, C., Canaveri, E., Corsello, L., 1990.
Correlation between mini-mental state examination and quantitative
electroencephalography in senile dementia of the Alzheimer type.
Neuropsychobiology 23, 7478.

260

Z. Hidasi et al. / International Journal of Psychophysiology 65 (2007) 252260

Rajna, P., Csibri, ., Pal, I., Szelenberger, W., 2003. Task related difference EEG
spectrum a new diagnostic method for neuropsychiatric disorders. Med.
Hypotheses 61 (3), 390397.
Rmy, F., Mirrashed, F., Campbell, B., Richter, W., 2004. Mental calculation
impairment in Alzheimer's disease: a functional magnetic resonance
imaging study. Neurosci. Lett. 358, 2528.
Rodriguez, G., Copello, F., Vitali, P., Perego, G., Nobili, F., 1999. EEG spectral
profile to stage Alzheimer's disease. Clin. Neurophysiol. 110 (10), 18311837.
Roselli, M., Ardila, A., 1989. Calculation deficit in patients with right and left
hemisphere damage. Neuropsychologia 27, 607617.
Sauseng, P., Klimesch, W., Schabus, M., Doppelmayr, M., 2005. Fronto-parietal
EEG coherence in theta and upper alpha reflect executive functions of
working memory. Int. J. Psychophysiol. 57, 97103.
Schreiter-Gasser, U., Gasser, T., Ziegler, P., 1994. Quantitative EEG analysis in early
onset Alzhimer's disease: correlations with severity, clinical characteristics, visual
EEG and CT. Electroencephalogr. Clin. Neurophysiol. 90, 267272.
Smith, M.E., McEvoy, L.K., Gevins, A., 1999. Neurophysiological indices of
strategy development and skill acquisition. Cogn. Brain Res. 7, 389404.
Smit, A.S., Eling, P.A.T.M., Coenen, A.M.L., 2004. Mental effort affects vigilance
enduringly: after-effects inn EEG and behaviour. Int. J. Psychophysiol. 53,
239243.
Sneddon, R., Shankle, W.R., Hara, J., Rodriquez, A., Hoffmann, D., Saha, U.,
2005. EEG detection of early Akzheimer's disease using psychophysical
tests. Clin. EEG. Neurosci. 36, 141150.
Stam, C.J., van der Made, Y., Pijnenburg, Y.A., Scheltens, P., 2003. EEG
sychronization in mild cognitive impairment and Alzheimer's disease. Acta
Neurol. Scand. 108 (2), 9096.

Stam, C.J., Montez, T., Jones, B.F., Rombout, S.A.R.B., van der Made, Y.,
Pijnenburg, Y.A.L., Scheltens, P., 2005. Disturbed fluctuations of resting state
EEG synchronization in Alzheimer's disease. Clin. Neurophysiol. 116, 708715.
Stam, C.J., Jones, B.F., Manshanden, I., van Cappellen van Walsum, A.M.,
Montez, T., Verbunt, J.P.A., de Munck, J.C., van Dijk, B.W., Berendse, H.W.,
Scheltens, P., 2006. Magnetoencephalographic evaluation of resting-state
functional connectivity in Alzheimer's disease. NeuroImage 32, 13351344.
Stevens, A., Kircher, T., Nickola, M., Bartels, M., Rosellen, N., Wormstall, H.,
2001. Dynamic regulation of EEG power and coherence is lost early ang
globally in probable DAT. Eur. Arch. Psychiatry Clin. Neurosci. 251, 199204.
Thatcher, R.W., Krause, P.J., Hrybyk, M., 1986. Cortico-cortical associations
and EEG coherence: A two compartmental model. Electroencephalogr. Clin.
Neurophysiol. 64, 123143.
van der Hiele, K., Vein, A.A., Kramer, C.G.S., Reijntjes, R.H.A.M., van Buchem, M.A.,
Westendorp, R.G.J., Bollen, E.L.E.M., van Dijk, J.G., Middelkoop, H.A.M., 2007.
Memory activation enhances EEG abnormality in mild cognitive impairment.
Neurobiol. Aging 28, 8590.
Wang, J., Rao, H., Wetmore, G.S., Furlan, P.M., Korczykowski, M., Dinges, D.F.,
Detre, J.A., 2005. Perfusion functional MRI reveals cerebral blood flow
pattern under psychological stress. Proc. Natl. Acad. Sci. USA/PNAS 102 (49),
1780417809.
Warrington, E.K., 1982. The fractionation of arithmetical skills: a single case
study. Quart. J. Exp. Psychol. 34A, 3151.