INTPSY-10656; No of Pages 8

International Journal of Psychophysiology xxx (2013) xxx–xxx

Contents lists available at SciVerse ScienceDirect

International Journal of Psychophysiology

journal homepage: www.elsevier.com/locate/ijpsycho

EEG beta band activity is related to attention and attentional deficits in the visual
performance of elderly subjects
Mateusz Gola a,⁎, Mikołaj Magnuski a, Izabela Szumska b,c, Andrzej Wróbel d
a
Department of Psychology, University of Social Sciences and Humanities, Poland
b
Department of Cognitive Psychology, University of Finance and Management, Warsaw, Poland
c
Laboratory of Psychophysics, Brain Mind Institute, Ecole Polytechnique Fédérale de Lausanne (EPFL), Switzerland
d
Department of Neurophysiology, Nencki Institute of Experimental Biology, Poland

a r t i c l e i n f o a b s t r a c t

Article history: We have previously shown that beta-band EEG activity is related to attentional modulation in the visual
Received 5 February 2013 system of cats and humans. In a separate experiment we also observed that some elderly subjects expressed
Received in revised form 8 May 2013 beta-band power decreases during a simple visual attention task, an effect which was accompanied by low
Accepted 12 May 2013
behavioral accuracy in this subgroup. Here, we conducted a detailed examination of beta power deficits in
Available online xxxx
elderly subjects in comparison to young controls. In order to do so, we equalized the subjective level of
Keywords:
task difficulty by adjusting visual stimuli presentation duration in such a way that elderly and young subjects
Attention achieved similar behavioral results. We found that: (1) beta-band power of EEG signals recorded over occip-
Beta band ital regions in elderly and young groups is related to visual attention, as judged from increases in beta power
Aging preceding correct responses and lack of beta activity change before erroneous responses; (2) despite forming
Human EEG a homogeneous group when screened for dementia (MMSE), age, education level, visual correction, and
speed–accuracy trade-off strategy, elderly subjects could be assigned into one of the two subgroups: high
performers, who did not differ from young performers in terms of beta-band power increases, and low
performers, whose beta power decreased during the most difficult attentional conditions (shortest — 3 s
and longest — 11 s cue-target delays). These findings posit that the beta-band activity decrease recorded in
low performing elderly subjects reflects difficulty in activation and deficits in sustaining attentional processes.
© 2013 Elsevier B.V. All rights reserved.

1. Introduction It is often neglected that all of the observed beta activity changes in
sensorimotor behavior are accompanied by various types of attention
Over a hundred years of electrical signal recordings from the (endogenous or exogenous), and it is hard to disentangle attention-
brain have shown that activity in different EEG frequency bands related processes from sensorimotor control (see Engel and Fries,
can be related to specific psychophysiological states. One of the less 2010). A slowly growing body of evidence shows that beta activity
frequently described bands is activity in the beta-band (12–30 Hz), may play an important role not only in sensorimotor but also in cogni-
studied mostly in relation to sensorimotor behavior. It has been tive functions, particularly as a result of their relation to attentional
well established that beta band power decreases during the prepara- processes.
tion and execution of voluntary movements (Jasper and Penfield,
1949; Pfurtscheller and Berghold, 1989; Pfurtscheller and Lopes da 1.1. Beta activities and attentional processes
Silva, 1999). Recent studies have shown that bursts of beta activity
are associated with strengthened sensory feedback during static The initial report on function related beta band power change was
motor control and reduced beta activity during movement change published by Mundy-Castle (1951), who observed significantly more
(Lalo et al., 2007). Beta activity also increases when movement activity at 12–30 Hz frequencies in subjects with good visual imagi-
has to be withstood or voluntarily suppressed (Kühn et al., 2004; nation skills when compared to less skilled subjects. Further on, two
Androulidakis et al., 2007; Zhang et al., 2008) and, in parallel, an in- similar studies by Townsend and Johnson (1979) and Belyavin and
crease in beta activity induced by transcranial alternating-current Wright (1987) found a positive correlation between accuracy levels
stimulation over the motor cortex slows subject's movements in a visual vigilance task and beta band power increase in occipito-
(Pogosyan et al., 2009). parietal brain areas.
Recent studies are in line with these early observations. Our
⁎ Corresponding author. Tel.: +48 729030030. experiments on cats (Bekisz and Wróbel, 1993, 1999, 2003; Wróbel
E-mail address: mateusz.gola@swps.edu.pl (M. Gola). et al., 1994a, 1994b, 2007; reviewed in Wróbel, 2000, 2014) provided

0167-8760/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.ijpsycho.2013.05.007

Please cite this article as: Gola, M., et al., EEG beta band activity is related to attention and attentional deficits in the visual performance of elderly
subjects, International Journal of Psychophysiology (2013), http://dx.doi.org/10.1016/j.ijpsycho.2013.05.007
2 M. Gola et al. / International Journal of Psychophysiology xxx (2013) xxx–xxx

evidence which supports the hypothesis that beta band activity may by any factor which was controlled for in our experiment (screening
serve as a carrier for attentional activation in many thalamic and cor- test for dementia, age, and education level), suggesting that it could
tical centers of the visual system. This hypothesis was based on the be interpreted in terms of a speed–accuracy trade-off.
finding that beta band power increased during the anticipatory time The procedure used in the experiment described above did not
period of a visual spatial differentiation task only in trials resulting allow us to examine if the distinguished groups of older adults
in correct responses. Similar data were described by Buschman and chose different strategies to cope with the task. Our experimental
Miller (2007), who noticed that visual attention is related to beta task was also visibly much easier for younger subjects, which does
band coherence between the frontal and occipito-parietal cortexes not favor intergroup comparisons (Gola et al., 2012). Another limita-
of monkeys. In a recent study, Buschman et al. (2012) showed an tion of the procedure prevented us from assessing whether erroneous
increased beta band coupling between dorsolateral prefrontal cortex response trials followed low beta power in occipital regions as shown
neurons when one of two available rules was cued for a given trial in our previous animal studies (Wróbel, 2000; Wróbel, 2014).
(monkeys had to react either to bar color or bar orientation; for When planning the current experiment, we aimed to remove such
review see: Engel, 2012). procedural inconveniences. In this study, we designed a new experi-
An increase of beta band EEG activity during the stimulus ex- mental procedure which meets three main objectives: (1) a high
pectancy period was also found in humans (Basile et al., 2007). level of task difficulty to obtain a sufficient number of erroneous trials
Hanslmayer et al. (2007) demonstrated that brief visual stimuli for comparisons between attentive and inattentive states, (2) a simi-
were perceived by subjects only when the preceding EEG beta band lar level of difficulty for both younger and older subjects, and (3) pro-
activity was high. MacLean et al. (2012) showed that high resting viding data to solve the speed–accuracy trade-off ambiguity. To
state beta band power is related to high accuracy in the attentional address all of these issues, we developed an adaptive procedure in-
blink task, while increased alpha band power is related to lower per- spired by the signal detection theory (Green and Swets, 1966). Trial
formance. Our study, in which we adopted a paradigm formerly used to trial adjustments allow us to tune the overall and individual diffi-
to study the anticipatory attention of cats, confirmed that increased culty level in order to obtain a sufficient number of correct and incor-
activity in the beta frequency range might also serve as a carrier for rect trials, while maintaining a similar level of performance between
attentional arousal in human subjects (Kamiński et al., 2012). In this the age groups. We believe that this approach can also provide a more
study we showed that increased alertness, manifested by faster re- specific relation of beta-band activity to attentional engagement and
sponses to target visual stimuli, was preceded by higher EEG activa- attentional deterioration in aging.
tion in the beta band over parieto-occipital regions and a decrease
of alpha band oscillations over occipital areas (Kamiński et al., 2012). 2. Subjects and methods
In parallel to the beta band research described above, many con-
tributions indicate a relation between alpha-band activity and atten- 2.1. Subjects
tive behavior (Ray and Cole, 1985; Busch and Herrmann, 2003;
Leiberg et al., 2006; Jokisch and Jensen, 2007; Busch et al., 2009; Two groups consisting of 35 university students (18–30 y.o.) and
Grimault et al., 2009; Haenschel et al., 2009; Mathewson et al., 40 preselected older subjects (65–85 y.o.) were recruited for this
2009, 2010; vanRullen et al., 2011). Prominent reviews (Klimesch study. Each subject was examined using the Mini Mental State Exam-
et al., 2007; Palva and Palva, 2007, 2011) have provided models for ination scale (MMSE; Folstein and Robins, 1983) and obtained a min-
the interaction of alpha oscillatory activity phase and amplitude and imum of 28 points. The Snellen visual correction test (Snellen, 1862)
its role in active inhibition within the cortical network. While the was also administered, and its inclusion criterion was set for sight
importance of alpha band oscillation during attentional behavior is deficits of less than 1 diopter (uncorrected or with use of corrective
widely agreed upon, the role of beta band activity in attentional acti- lenses). The last inclusion criterion was a negative result on the
vation requires further study. screening interview for psychiatric diseases, history of brain injury,
Deriving from our past studies (Bekisz and Wróbel, 1993, 1999, and current medication potentially impairing cognitive abilities.
2003; Wróbel et al., 1994a, 1994b, 2007; Gola et al., 2012; Kamiński After these examinations, we selected 35 subjects in two age
et al., 2012), we posit that beta-band activity serves as an activation groups: younger (n = 17; 8 males; age range from 19 to 28 y.o.,
mechanism which enhances the gain of feedback loops at subsequent M = 22.3; SD = 2.78) and older (n = 18; 9 males; age ranged
stages of visual information processing (reviewed in Wróbel, 2014). from 65 to 85 y.o., M = 74.8; SD = 7.54). Subjects received 120
PLN (approximately 30 EUR) as monetary compensation for partici-
1.2. Does beta band activity reveal attentional deficits in aging? pating in the study. This study was approved by the Ethics Committee
of the University of Social Sciences and Humanities. All subjects
Earlier research on vigilance task performance in elderly subjects signed consent forms specific to this study.
provided ambiguous results. Some studies (Moss et al., 1988; Lai
et al., 1995) pointed to slightly decreased performance of elderly sub- 2.2. Experimental procedure
jects as compared to young, while others (reviewed in Allbert and
Moss, 1996; Gola et al., 2012) found a significant deterioration in Subjects were asked to determine whether or not the target
elderly performance on attention-dependent tasks. Assuming that (a horizontal, diagonal, or vertical bar) was present in a matrix of
beta activity is involved in attentional processes, we recently conducted 16 bars. There were four possible outcomes: (1) Hit (when target
an experiment investigating beta power deficits in elderly subjects was present and the subject saw it), (2) Rejection (when there
performing a delayed attentional task with spatial differentiation be- was no target and the subject confirmed its absence), (3) Omission
tween target visual stimuli (Gola et al., 2012). (when target was present but the subject claimed it was not in the
As predicted, we found that a lower beta (17–20 Hz) power over oc- matrix), and (4) False Alarm (when there was no target, but the
cipital regions in older adults correlated with their decreased behavioral subject answered that there was). The delay between the presen-
performance. Additionally, we observed that in comparison to young tation of the cue and the appearance of the target matrix varied
subjects (18–30 y.o.), elderly subjects (65–80 y.o.) performed either between 3 and 11 s (in discrete values of 3, 5, 7, 9, and 11 s; see
with increased reaction times (RTs) and no accuracy decrease (elderly Fig. 1). The cue was displayed during the entire cue-trial delay period
high performance group — EHP) or with similar RTs but lower accuracy and disappeared exactly at the moment of target matrix presentation.
(elderly low performance group — ELP). The differences in behavioral This allowed us to minimize memory process involvement in the task.
results observed between the two subgroups could not be explained Subjects had no idea if the matrix would appear to the right or left of

Please cite this article as: Gola, M., et al., EEG beta band activity is related to attention and attentional deficits in the visual performance of elderly
subjects, International Journal of Psychophysiology (2013), http://dx.doi.org/10.1016/j.ijpsycho.2013.05.007
 M. Gola et al. / International Journal of Psychophysiology xxx (2013) xxx–xxx 3

the fixation cross, which caused increased activation in the visual SPSS 19 software. As the first step, we compared the overall results
cortex of both hemispheres. of both age groups to examine if the experimental procedure reached
Task difficulty was adjusted by lengthening (+10% after each error) our expectations and resulted in a comparable level of accuracy
or shortening (−5% after each correct response) the duration of the tar- irrespective of age. The second step followed the method used in
get matrix presentation. The target matrix display starting time was set our previous paper (Gola et al., 2012), where we classified our data
at 800 ms. In the instructions, subjects were asked not to guess if not by K-mean analysis (Hartigan and Wong, 1979) to verify whether
sure of the response, and instead to wait for the next trial. Being unsure older subjects could be split into high (EHP) and low (ELP) perfor-
may be treated as an inattentive state or at least a state of lower concen- mance level groups. Individual accuracy and mean time of target ma-
tration, and seems to provide a clearer indication of an inattentive state trix presentation were used as criterions for this analysis. A mean
than omissions and false alarms. As omissions may be caused by inat- time parameter reflected the overall level of difficulty of the proce-
tention, false alarms are more likely to be caused by oversensitivity dure, resulting from changes in time of target matrix presentation
and insufficient response inhibition. We decided to compare the correct depending on response accuracy (see Subjects and methods).
trials (hits and rejections) with no response trials (see Analysis), as in In the third step, we compared accuracy and RTs between all three
this case the subjective state of subjects seemed to be clear. Refraining groups (young, EHP, and ELP) to investigate the possibility of differ-
from response was considered an error, and duration of matrix presen- ences in speed–accuracy trade-off strategies. We also computed a d′
tation in the next trial was increased by 10%. Response time was limited index (d′ = Z(hit rate) − Z(false alarm rate)), which could help to
to 2.5 s, after which the next trial would start. This type of procedure is a verify whether the subgroups of older subjects (EHP and ELP) differed
result of pretests performed on a group of 120 subjects, allowing us to in sensitivity of responses (i.e. if one group had significantly higher
align the overall level of difficulty between young and elderly subjects. rate of false alarms in comparison to hits). We also examined if the
There were 90 trials combined into 3 blocks of 30 trials, with self-paced two subgroups differed according to control variables (age, education
breaks between blocks. level, visual correction, MMSE scores). ANOVA was used for those
analyses and all multiple comparisons were Bonferroni corrected.
2.3. EEG recordings

EEG signals were recorded with cap attached electrodes (actiCap, 3.2. EEG signal analysis
Inc. Brain Products) using 64 channels in standard 10–20 system
positions, referenced to the FCz electrode. To control for artifacts For EEG signal analysis, EEGLAB 10.2.2.4b (Delorme and Makeig,
associated with eye movements and blinking, we used two bipolar 2004), Matlab, and SPSS 19 software were used. At the preprocessing
EOG channels — one pair of electrodes placed above and below the stage, signals were referenced to a global average excluding two EOG
right eye (the vertical line defined at the pupil), and a second at the channels and then filtered with a 0.1 Hz high pass FIR filter. Extracted
left and right temples (in horizontal line of both pupils). Recorded epochs started 3000 ms before the target matrix (it was the shortest
signal was amplified by a BrainVision QuickAmp (Brain Products cue-target delay) and ended 1 ms before the target matrix appear-
Inc.) 72-channel amplifier, with a sampling rate of 1000 Hz and ance (to avoid the evoked response for stimuli). All trials were
low-pass filtered with a frequency of 140 Hz. Impedance was kept visually inspected for artifacts. Epochs containing high amplitude
below 5 kΩ for all EEG electrodes. muscle artifacts (>± 50 μV) and visible multichannel distortion
were removed (21% of all epochs). Eye movement and eye blink arti-
3. Analysis facts were removed from the signal with independent component
analysis (ICA; Makeig et al., 1996) in the following step of preprocess-
3.1. Behavioral data analysis ing. All channels whose impedance increased and reached a value
over 10 kΩ during the study were excluded and replaced by interpo-
T-tests and ANOVA were used for the behavioral data analysis lated signals from the neighboring channels (average of 5 channels
(comparisons of accuracy and RTs). Analyses were performed using per subject) for all epochs.

Fig. 1. Schema of experimental procedure used in the study. After random fixation cross presentation (1.2–2 s), cue appears over the fixation cross and lasts for 3, 5, 7, 9 or 11 s until
target matrix presentation. Subjects were asked to say if the bar which was presented as a cue (diagonal, horizontal or vertical) is present within the target matrix. Subjects
answered yes or no. Duration of target matrix is variable and depends on accuracy of previous response, which adapts the level of task difficulty to individual subjects' abilities.
Subjects were asked not to respond if they were not sure if the target was present or not within the matrix.

Please cite this article as: Gola, M., et al., EEG beta band activity is related to attention and attentional deficits in the visual performance of elderly
subjects, International Journal of Psychophysiology (2013), http://dx.doi.org/10.1016/j.ijpsycho.2013.05.007
4 M. Gola et al. / International Journal of Psychophysiology xxx (2013) xxx–xxx

In the next step, we performed an ICA with an Infomax algorithm level was over 2 SD's below the mean of the whole group. It should be
separately for each subject. ICA decomposes EEG signal into a number noted that one subject from the young group seems to be an outlier
of components equal to the number of sensors. Derived components (yellow dot on the bottom of Fig. 2), but his results do not match our
have distinct but not necessarily orthogonal scalp projections. The al- exclusion criterions (−1.36 SD of the mean for all subjects; −1.79 SD
gorithm by itself does not model the locations of component genera- of the mean result of his age group).
tors in the head, but allows for it to be done at further stages (Makeig The older subject subgroups (ELP and EHP) did not differ when
et al., 1996). We have chosen this analysis for four reasons: (1) this compared according to our control variables: age, visual correction,
type of analysis allows us to separate independent components of sig- MMSE scores, and sensitivity index (d′; see Table 1). In contrast,
nals with occipital scalp projection. According to Makeig and Onton's dependent variable comparisons (see Subjects and methods) revealed
work (Onton et al., 2006; Makeig and Onton, 2011), this method al- that the elderly low performing group obtained a significantly lower ac-
lows specific occipito-parietal signals to be selected reliably; (2) inde- curacy level (sum of Hits and Rejections) and a higher rate of no re-
pendent components may be localized in the brain, giving a hint for sponse trials than high performers (Fig. 3B). However, the subgroups
localizing a putative source of the signal; (3) ICA produces individual did not differ in mean RTs (Fig. 3A). Neither of the older subjects groups
components for each subject. This is important, as some authors claim differed in RTs from younger subjects (F(2,31) = 1.14; p > .05;
that cognitive beta-band activity is characterized by high inter- Fig. 3A). Accuracy levels for the ELP group were significantly lower
individual variability of source localization (compare Basile et al., than for younger subjects (F(2,31) = 10.99; p b .001; post hoc test:
2007); and (4) by using ICA, it is possible to subtract eye-related arti- p = .0014; Fig. 3B), but rate of the ‘no response’ trials did not differ sig-
fact components from the signal and save many epochs for further nificantly (p = .074; Fig. 3B). The EHP group did not differ from youn-
analysis. ger subjects on rates of different response types (Fig. 3B).
We obtained 1960 independent components for all subjects As task difficulty was adjusted by lengthening or shortening of tar-
(averaged 58 per subject) and localized their sources with the get matrix presentation time, we also examined whether the target
4-shell BESA brain model and dipole fitting DIPFIT 2 toolbox for duration affected RTs. Pearson's correlations of these two variables
EEGLab (Kavanagh et al., 1978; Oostendorp and van Oosterom, were insignificant when calculated for each subject separately, and
1989; Oostenveld and Praamstra, 2001). For the subsequent analysis, R = − .01 (p = .53) at the group level. The lack of significant corre-
we chose all components with a residual variance under 15%. Then, lation suggests that RT can be successfully used as a reliable measure
as in the previous study (Gola et al., 2012), the last 1000 ms of each for reasoning about the type of speed–accuracy trade-off strategy.
epoch was extracted (1 s preceding the target matrix appearance) Finally, we checked accuracy level dependency on the duration
and Fast Fourier Transform (FFT) was performed on the signal of the of cue-target delay. These comparisons were conducted separately
selected components. FFT results were compared between groups and for each group. No significant differences were observed in the
between correct and incorrect trials. Additionally, we compared FFT group of younger and EHP subjects, and two significant differences
spectra for trials grouped according to the cue-target delays (3, 5, 7, 9 were noted in the ELP group: increase of accuracy level between 3
and 11 s) for the ELP group. We decided to compare these EEG condi- and 9 second delay trials (p b .05), and decrease of accuracy level
tions because behavioral data analysis revealed that the shortest and between 9 and 11 second delay trials (p b .05; Fig. 4A).
longest (3 and 11 s) delay trials were significantly more difficult than
medium (9 s) delay trials. All FFT comparisons were performed within
the 1 Hz window for frequencies in the 2.5 to 47.5 Hz range with 4.2. EEG results
ANOVA and Bonferroni multiple comparisons correction. The signifi-
cance level for each analysis was set to p b 0.01. Out of the 1960 independent components extracted from EEG sig-
nals with ICA analysis we selected 449 with dipoles characterized by a
4. Results residual variance under 15%, meaning that their localizations were
plausible and dipole projections fit the component scalp projections.
4.1. Behavioral results K-mean analysis was performed on the selected components using

Intergroup (young vs. all older subjects) comparisons of accuracy

and RTs revealed that our experimental procedure met the required de-
mands. The groups did not differ on any response type (Hits, Rejections,
Omissions, False Alarms, No response; p > .05) or overall accuracy
(sum of correct responses: Hits and Rejections; p > .05). Mean RTs
were only slightly shorter for young subjects (1487 ms vs. 1538 ms;
p = .05); however, variance within the younger group was signifi-
cantly higher compared to older subjects (p b .05). No differences in
the d′ index between age groups were observed (younger: M = −.27;
SD = 1.08; older: M = .25; SD = 1.18; t(33) = 1.36; p > .05). The
lack of intergroup differences (young vs. all older subjects) confirmed
that the procedure met the main objectives and equalized the difficulty
level of the task for both younger and older subjects.
Splitting the older participant group into two subgroups of high and
low performers was performed using the K-means algorithm with two
criterions: individual mean accuracy (sum of Hits and Rejections) and
mean time of target matrix presentation (this parameter reflected the
individual level of difficulty set by adjusting the duration of target
presentation, and is negatively correlated with accuracy: r = −.82;
p b .001). This analysis clustered the results (accuracy and mean time
of the target matrix presentation) of all subjects into two significantly Fig. 2. Results of the K-means analysis distinguishing the subgroup of high performing
different groups (F(1.13) = 40.74; p b .001; Fig. 2). One subject in the (light blue color) and low performing (dark blue) older subjects. Orange dots indicate
older group was excluded from further analysis as his performance younger subjects. Correct responses are the sum of Hits and Rejections.

Please cite this article as: Gola, M., et al., EEG beta band activity is related to attention and attentional deficits in the visual performance of elderly
subjects, International Journal of Psychophysiology (2013), http://dx.doi.org/10.1016/j.ijpsycho.2013.05.007
 M. Gola et al. / International Journal of Psychophysiology xxx (2013) xxx–xxx 5

Table 1
Comparisons of the high and low performing elderly subgroups.

High performers Low performers T-tests

n = 8; M (SD) n = 9; M (SD) t(15) =

Age 75.83 (9.02) 73.86 (6.62) .55; p > .05

MMSE scores 28.83 (.75) 28.43 (1.18) .12; p > .05
Visual correction .63 (.38) .75 (.35) .19; p > .05
d′ −.50 (.80) .74 (1.62) 1.85; p > .05

10 clusters defined by two parameters: spatial (in 3D) localization of

dipoles, and similarity of the dipole projection on the scalp.
The first criterion received a weight which was five times higher
than the second. Components outlying more than 3 SD's from the center
of each cluster were excluded (21 outliers). A cluster containing 38 in-
dependent components derived from the 32 subjects with occipital
scalp projection was chosen for further analysis (Fig. 5). For two sub-
jects (one younger and one from the ELP group), no component meet-
ing the criteria (localization of dipoles with occipital projections
localized in occipito-parietal area) was found in the selected cluster.
Comparison of FFT performed on the selected components of EEG
signal during the 1000 ms preceding onset of the target matrix
for correct (Hits and Rejections) and incorrect (No response; see
Subjects and methods) trials did not reveal significant (p > .01)
differences at any frequencies, although a slight beta-band (17 Hz)

Fig. 4. Accuracy level (sum of Hits and Rejections) presented separately for each of
cue-target delay conditions. Groups were marked by colors: young (yellow), elderly
with high performance (light blue; EHP), and elderly with low performance (dark blue;
ELP). Error bars indicate SEM, and two arrows point out a significant (p b 0.05) difference
of accuracy level between the 3, 9, and 11 second delay conditions in the elderly with low
performance group. B: EEG signal power spectrum comparison of 3 and 9 second
cue-target delay conditions for ELP group and C: parallel comparison of 9 and 11 second
conditions. Black bars indicate a p b 0.01 level of significance.

peak in FFT for correct trials was visible. We supposed that this lack
of difference was due to the common analysis performed for both
elderly groups; elderly low performing subjects (ELP) could display
attentional deficits that would increase the variability of the results.
This hypothesis was supported by two observations: (1) ELP was
the only group manifesting significantly lower accuracy with respect
to other groups (Fig. 3B), and (2) our previous study (Gola et al.,
2012) showed that this group of subjects was characterized mainly
by decreases in beta activity. Taking those arguments into account,
we compared FFT for the correct (Hits and Rejections) and incorrect
(No response) trials once again including only EHP and younger
groups (Fig. 3B). The obtained results uncovered a significant (p b .01)
increase of beta band power activity (17–19 Hz) for trials ending with
correct responses (Fig. 3C), as well as significant decreases in theta
(6 Hz) and alpha (11 Hz) bands (Fig. 3C).
Omissions and false alarms were not analyzed due to an insuffi-
cient number of such trials, and they were not included in the class
Fig. 3. A: Mean reaction times (error bars indicates SD). B: Comparisons of all behavioral of incorrect trials (which encompassed only ‘no response’ trials)
responses types between the groups of young, elderly high, and elderly low performing because omissions and false alarms are different types of errors
subjects. C: EEG signal power spectra calculated for trials ending with correct responses
(green line; assumed as representing an attentive state) and trials with no response
(see Subjects and methods).
(black; inattentive state) for all younger and older high performing subjects. Black bars Although we did not find differences between FFT power spectra
indicate a p b .01 level of significance. for the correct and incorrect trials in the ELP group data, we expected

Please cite this article as: Gola, M., et al., EEG beta band activity is related to attention and attentional deficits in the visual performance of elderly
subjects, International Journal of Psychophysiology (2013), http://dx.doi.org/10.1016/j.ijpsycho.2013.05.007
6 M. Gola et al. / International Journal of Psychophysiology xxx (2013) xxx–xxx

Fig. 5. Localization of selected EEG independent components dipoles cluster (left) and its scalp projections (right).

that the comparison between the cue-target delay conditions would for upstream visual information (Lindström and Wróbel, 1990; see
reveal differences. As the behavioral data analysis revealed significant Wróbel, 2014 for review). Consequently, we proposed that during
changes in accuracy for the different cue-target delay conditions mea- high attentional demand tasks beta-band activity caused subthreshold
sured for the ELP group (Fig. 4A), we decided to compare activities depolarization in structures modulated by feedback connections and
recorded in this group during the shortest and longest trials (3 and increased their sensitivity for visual input (Wróbel, 2000, 2014). Such
11 second cue-target delays) with medium trials (9 second delay). a modulatory mechanism exerted by feedback pathways within all
The shortest and longest trial conditions appeared to be significantly the sensory systems may underlie many phenomena described as
more difficult (accuracy was lower) than the medium trial (Fig. 4A). attention or vigilance (Bekisz and Wróbel, 1993; Briggs and Usrey,
Comparing the FFT spectra for the 3 vs. 9 s delay trials, we ob- 2009; see Wang, 2010 for review).
served a significant (p b .01) decrease in the 13–14 Hz, 22–24 Hz, The results obtained in this study showing that beta activity in
and 37–39 Hz frequencies accompanied by an increase of power in occipital regions is related to attentive visual performance perfectly
the 6–7 Hz frequency band (Fig. 4B). Similar results were obtained match previous animal models. This conclusion follows the observation
when comparing the FFT power spectrum from 11 vs. 9 s delay trials. that increased beta power differentiated trials ending with correct
In this comparison, the more difficult condition (11 s) was character- response from those without a response (subjects were instructed to
ized by significant decreases in 22–24 Hz and 36 Hz frequency win- respond only if confident of their observation; Fig. 3C). Quantitative
dows and an increase in 9 Hz (Fig. 4C). analysis of independent components revealed that the sources of
these beta-band activities originated in occipital EEG signals (Fig. 5).
5. Discussion The present results are in line with our proposed hypothesis.
We observed a significant increase of beta-band power in occipital
In this study we have shown that: (1) beta-band power of EEG sig- visual regions of young and elderly subjects with high performance
nal recorded over occipital regions is related to visual attention, levels (EHP; Fig. 3C) only for successful trials. Conversely, signals
as correct responses were preceded by increased beta activity while from elderly subjects with low performance levels (ELP) were charac-
trials without such activity ended with erroneous responses; (2) de- terized by decreased beta activity in the difficult (defined by the
spite a lack of differences on the screening test for dementia (MMSE), lowest accuracy level) delay conditions (shortest: 3 s, and longest:
age, visual correction, and speed–accuracy trade-off strategy, older sub- 11 s) in comparison to the easiest condition (medium: 9 s; Fig. 4C).
jects could be divided into two subgroups: high performers, who did
not differ from young subjects in terms of beta-band power changes, 5.2. Beta-band and age related deficits
and low performers, whose beta band activity was higher only during
the relatively easy condition (9 s cue-target delay) and decreased fol- In our previous study (Gola et al., 2012) we showed that (1) in
lowing difficult conditions (shortest: 3 s, and longest: 11 s; cue-target comparison to young subjects, elderly subjects may perform either
delay). These findings allow us to conclude that in the case of low with increased RTs and no accuracy difference (EHP) or with similar
performing elderly subjects, beta-band decrease reflects difficulty in RTs but lower accuracy (ELP); (2) accuracy of performance in elderly
activation of attentional processes (alertness deficits in short delay subjects is positively related to beta EEG activity in the 18–19 Hz
condition) and deficits in sustaining those processes (longest delay). range; and (3) the two groups of elderly adults expressed different
dynamics of beta activity during anticipation of the target stimulus.
5.1. Beta-band activity is related to attention In addition to replicating these results, the current experiment aimed
to determine whether the differences observed between our ELP and
We have previously shown that attentive visual behavior is accom- EHP groups were caused by their speed–accuracy trade-off strategy. If
panied by increased local field potential activity in the beta band they indeed resulted from a different strategy undertaken by the two
recorded from the visual centers of cats (lateral geniculate nucleus, elderly groups, beta-band decrease could be interpreted as reflecting
lateral-posterior nuclear complex, primary and medial suprasylvian vi- a temporary decline in attention. Alternatively, we might assume a
sual cortices; Bekisz and Wróbel, 1993; Wróbel et al., 2007). All of these more durable depletion of attentional resources in the ELP group.
structures are hierarchically connected via driving feed-forward and The present study confirmed our previous finding (Gola et al., 2012)
modulatory feedback pathways (Sherman, 2005). Some of these feed- that the elderly group was composed of low-performing (ELP) and
back pathways, which were checked for gain transfer, appeared to high-performing (EHP) subjects, and showed that the two subgroups
possess a frequency potentiation mechanism at the beta band, and did not differ in the speed–accuracy trade-off strategy since no sig-
therefore may depolarize their input structures and increase the gain nificant differences were found for RTs and sensitivity indexes (d′)

Please cite this article as: Gola, M., et al., EEG beta band activity is related to attention and attentional deficits in the visual performance of elderly
subjects, International Journal of Psychophysiology (2013), http://dx.doi.org/10.1016/j.ijpsycho.2013.05.007
 M. Gola et al. / International Journal of Psychophysiology xxx (2013) xxx–xxx 7

between EHP and ELP subjects. The d′ index indicates a sensitivity Acknowledgments
of separation between the signal and noise in the classical signal-
detection task, and thus can help to determine the strategy if RTs differ. We would like to thank Marta Ścisło for recruiting our older subjects.
However, the additional response choice introduced in the current Marta's visible state of pregnancy was the cause of considerable
paradigm — to withdraw response in case of uncertainty (which allowed enthusiasm among them. We also thank Zuza Kłyszejko, who helped
us to obtain a sufficient number of attentive and inattentive trials for us greatly during testing of parameters for the experimental procedure.
EEG comparisons) — broke signal-detection principles and caused diffi- This research project was supported by the Polish National Science
culties in interpretation of the d′ index. Nonetheless, the lack of differ- Center grant NCN 2011/01/N/NZ4/03695.
ences between RTs measured for EHP and ELP (Fig. 3A) indicated
those groups did not differ in the speed–accuracy trade-off. This result
suggests that low performing elderly subjects may have disturbed References
mechanisms underlying their attentional processes. We found that
Allbert, S., Moss, E., 1996. Neuropsychology of aging, In: Schneider, E.i., Rowe, J.W.
elderly subjects with low performance (and without symptoms of (Eds.), Handbook of the Biology of Aging, 4th Edition. Academic Press.
dementia) expressed reduced power at these frequencies instead of Androulidakis, A.G., Doyle, L.M.F., Yarrow, K., Litvak, V., Gilbertson, T.P., Brown, P., 2007.
smaller activation of beta-band activity during difficult attentive tasks Anticipatory changes in beta synchrony in the human corticospinal system and
associated improvements in task performance. The European Journal of Neuroscience
(3 and 11 s cue-target delay trials). This finding differentiates ELP sub-
25, 3758–3765.
jects from EHP and younger subjects, raising the question of whether Basile, L.F.H., Anghinah, R., Ribeiro, P., Ramos, R.T., Piedade, R., Ballester, G., Brunetti,
the neural processes evoked by the attentive situation in ELP subjects E.P., 2007. Interindividual variability in EEG correlates of attention and limits of
were different. One possible explanation could be that those subjects functional mapping. International Journal of Psychophysiology 65, 238–251.
Bekisz, M., Wróbel, A., 1993. 20 Hz rhythm of activity in visual system of perceiving cat.
operate permanently at the verge of ‘attentional depletion’, which is Acta Neurobiologiae Experimentalis 53, 175–182.
expressed by a decrease of beta activation in the shortest and longest Bekisz, M., Wróbel, A., 1999. Coupling of beta and gamma activity in cortico-thalamic
cue-delay conditions. In contrast, younger and EHP subjects can main- system of cats attending to visual stimuli. NeuroReport 10, 3589–3594.
Bekisz, M., Wróbel, A., 2003. Attention-dependent coupling between beta activities
tain attention more efficiently, which is reflected by beta-band increases recorded in the cat's thalamic and cortical representations of the central visual
in most of the trials. field. The European Journal of Neuroscience 17, 421–426.
While this paper is focused on beta band activity, it is worthwhile Belyavin, A., Wright, N.A., 1987. Changes in electrical activity of the brain with vigilance.
Electroencephalography and Clinical Neurophysiology 66 (2), 137–144.
to note that similarly to beta, activity in other frequencies also dif- Briggs, F., Usrey, W.M., 2009. Modulation of gamma-band activity across local cortical
fered in the ELP group in opposition to what we found in the younger circuits. Frontiers in Integrative Neuroscience 3. http://dx.doi.org/10.3389/
adult and EHP. For these two groups, we observed decreases in alpha neuro.07.015.2009.
Busch, N.A., Herrmann, C.S., 2003. Object-load and feature-load modulate EEG in a
and theta powers preceding correct trials (Fig. 3C). In contrast,
short-term memory task. NeuroReport 14, 1721–1724.
ELP subjects exhibited increases in these frequencies for the most Busch, N.A., Dubois, J., VanRullen, R., 2009. The phase of ongoing EEG oscillations
difficult conditions (Fig. 4B, C). Alpha band desynchronization is a predicts visual perception. The Journal of Neuroscience 29, 7869–7876.
Buschman, T.J., Miller, E.K., 2007. Top-down versus bottom-up control of attention in
well-established effect related to attentional behavior in perceptual
the prefrontal and posterior parietal cortices. Science 315, 1860–1862.
tasks (for reviews see: Klimesch et al., 2007; Palva and Palva, 2007, Buschman, T.J., Miller, E.K., 2010. Shifting the spotlight of attention: evidence for
2011). In our ELP group, the mechanism of alpha desynchronization discrete computations in cognition. Frontiers in Human Neuroscience 4, 194–203.
seemed to be deregulated in the most difficult conditions (when the Buschman, T.J., Denovellis, E.L., Diogo, C., Bullock, D., Miller, E.K., 2012. Synchronous
oscillatory neural ensembles for rules in the prefrontal cortex. Neuron 76, 838–846.
cue-target delay is too short or too long). These difficult conditions Delorme, A., Makeig, S., 2004. EEGLAB: an open source toolbox for analysis of single-
were also accompanied by a decrease of gamma band power in ELP trial EEG dynamics. Journal of Neuroscience Methods 134, 9–21.
subjects in contrast to the easiest condition (9 s cue-target delay; Engel, A.K., 2012. Rules got rhythm. Neuron 76, 673–676.
Engel, A.K., Fries, P., 2010. Beta-band oscillations — signaling the status quo? Current
Fig. 4B and C). Simultaneous beta and gamma band decreases in Opinion in Neurobiology 20, 156–165.
ELP subjects are in agreement with our hypothesis that the depolar- Folstein, M.F., Robins, L.N., 1983. The Mini-Mental State Examination. Archives of General
izations produced by beta activity might lower thresholds within Psychiatry 40, 812.
Gola, M., Kamiński, J., Brzezicka, A., Wróbel, A., 2012. Beta band oscillations as a corre-
the required ensemble of neurons (Lindström and Wróbel, 1990; late of alertness — changes in ageing. International Journal of Psychophysiology 85,
Wróbel, 2014) and evoke an activation state allowing for gamma syn- 62–67.
chronization in local networks (Wróbel, 2000; Briggs and Usrey, Green, D.W., Swets, J.A., 1966. Signal Detection Theory and Psychophysics. Wiley,
New York.
2009; Buschman and Miller, 2010; Lakatos et al., 2008; Steriade Grimault, S., Robitaille, N., Grova, C., Lina, J.M., Dubarry, A.S., Jolicoeur, P., 2009. Oscillatory
et al., 1996). If gamma activity depends on beta-related depolariza- activity in parietal and dorsolateral prefrontal cortex during retention in visual short-
tions, the decreases of gamma band expressed by low performing term memory: additive effects of spatial attention and memory load. Human Brain
Mapping 30, 3378–3392.
elderly subjects might be related to their beta-band activation deficits.
Gross, J., Schmitz, F., Schnitzler, I., Kessler, K., Shapiro, K., Hommel, B., Schnitzler, A.,
2004. Modulation of long-range neural synchrony reflects temporal limitations of
5.3. Conclusions visual attention in humans. Proceedings of the National Academy of Sciences
USA 101, 13050–13055.
Haenschel, C., Bittner, R.A., Waltz, J., Haertling, F., Wibral, M., Singer, W., Linden, D.E.,
In this study we present results supporting the hypothesis that Rodriguez, E., 2009. Cortical oscillatory activity is critical for working memory as
beta oscillation serves as a carrier of visual attention in humans revealed by deficits in early-onset schizophrenia. The Journal of Neuroscience 29,
(Gross et al., 2004; Basile et al., 2007; Hanslmayer et al., 2007; Gola 9481–9489.
Hanslmayer, S., Aslan, A., Staudigl, T., Klimesch, W., Herrmann, C.S., Bauml, K., 2007.
et al., 2012). This conclusion is based on the increase of beta band Prestimulus oscillations predict visual perception performance between and within
power recorded from occipital sites during a spatial discrimination subjects. NeuroImage 37, 1465–1473.
task in young and elderly high performing subjects. As these increases Hartigan, J.A., Wong, M.A., 1979. Algorithm AS 136: a K-means clustering algorithm.
Journal of the Royal Statistical Society: Series C (Applied Statistics) 28 (1), 100–108.
were observed only during successful trials, we attribute the beta Jasper, H., Penfield, W., 1949. Electrocorticograms in man: effect of voluntary
activity to an attention related mechanism. Moreover, the beta band movement upon the electrical activity of the precentral gyrus. Archiv fur
increases in these subjects were related to response accuracy. Elderly Psychiatrie und Zeitschrift Neurologie 183, 163–174.
Jokisch, D., Jensen, O., 2007. Modulation of gamma and alpha activity during a working
low performing subjects expressed decreased activity in the beta memory task engaging the dorsal or ventral stream. The Journal of Neuroscience
band during the most difficult trials (the shortest and the longest 27, 3244–3251.
cue-target delays). It seems that elderly low performing subjects Kamiński, J., Brzezicka, A., Gola, M., Wróbel, A., 2012. Beta band oscillations engagement in
human alertness process. International Journal of Psychophysiology 85, 125–128.
have particular difficulties in the activation and sustainment of atten-
Kavanagh, R., Darcey, T.M., Lehmann, D., Fender, D.H., 1978. Evaluation of methods
tional processes. Further investigations are needed for a better under- for three-dimensional localization of electric sources in the human brain. IEEE
standing of the possible causes of these deficits. Transactions on Biomedical Engineering 25, 421–429.

Please cite this article as: Gola, M., et al., EEG beta band activity is related to attention and attentional deficits in the visual performance of elderly
subjects, International Journal of Psychophysiology (2013), http://dx.doi.org/10.1016/j.ijpsycho.2013.05.007
8 M. Gola et al. / International Journal of Psychophysiology xxx (2013) xxx–xxx

Klimesch, W., Sauseng, P., Hanslmayr, S., 2007. EEG alpha oscillations: the inhibition- Oostenveld, R., Praamstra, P., 2001. The five percent electrode system for high-
timing hypothesis. Brain Research Reviews 53, 63–88. resolution EEG and ERP measurements. Clinical Neurophysiology 112, 713–719.
Kühn, A.A., Williams, D., Kupsch, A., Limousin, P., Hariz, M., Schneider, G.H., Yarrow, K., Palva, S., Palva, J.M., 2007. New vistas for alpha-frequency band oscillations. Trends in
Brown, P., 2004. Event-related beta desynchronization in human subthalamic Neurosciences 30, 150–158.
nucleus correlates with motor performance. Brain 127, 735–746. Palva, S., Palva, J.M., 2011. Functional roles of alpha-band phase synchronization in
Lai, Z.C., Moss, M.B., Killiany, R.J., Rosene, D.L., Herndon, J.G., 1995. Executive system local and large-scale cortical networks. Frontiers in Psychology 2, 204.
dysfunction in the aged monkey: spatial and object reversal learning. Neurobiolo- Pfurtscheller, G., Berghold, A., 1989. Patterns of cortical activation during planning of
gy of Aging 16 (6), 947–954. voluntary movement. Electroencephalography and Clinical Neurophysiology 72,
Lakatos, P., Karmos, G., Mehta, A.D., Ulbert, I., Schroeder, C.E., 2008. Entrainment of 250–258.
neuronal oscillation as a mechanism of attentional selection. Science 320, 110–113. Pfurtscheller, G., Lopes da Silva, F.H., 1999. Event-related EEG/MEG synchronization and
Lalo, E., Gilbertson, T., Doyle, L., Di Lazzaro, V., Cioni, B., Brown, P., 2007. Phasic desynchronization: basic principles. Clinical Neurophysiology 110, 1842–1857.
increases in cortical beta activity are associated with alterations in sensory pro- Pogosyan, A., Gaynor, L.D., Eusebio, A., Brown, P., 2009. Boosting cortical activity at
cessing in the human. Experimental brain research. Experimentelle Hirnforschung. beta-band frequencies slows movement in humans. Current Biology 19, 1637–1641.
Expérimentation cérébrale 177, 137–145. Ray, W.J., Cole, H.W., 1985. EEG alpha activity reflects attentional demands, and beta
Leiberg, S., Lutzenberger, W., Kaiser, J., 2006. Effects of memory load on cortical oscillatory activity reflects emotional and cognitive processes. Science 228, 750–752.
activity during auditory pattern working memory. Brain Research 1120, 131–140. Sherman, S.M., 2005. Thalamic relays and cortical functioning. Progress in Brain Re-
Lindström, S., Wróbel, A., 1990. Frequency dependent corticofugal excitation of principal cells search 149, 107–126.
in the cat's dorsal lateral geniculate nucleus. Experimental Brain Research 79, 313–318. Snellen, H., 1862. Test-types for the Determination of the Acuteness of Vision.
MacLean, M.H., Arnell, K.M., Cote, K.A., 2012. Resting EEG in alpha and beta bands Steriade, M., Amzica, F., Contreras, D., 1996. Synchronization of fast (30–40 Hz) spon-
predicts individual differences in attentional blink magnitude. Brain and Cognition taneous cortical rhythms during brain activation. The Journal of Neuroscience 16,
78, 218–229. 392–417.
Makeig, S., Onton, J., 2011. ERP features and EEG dynamics: an ICA perspective. In: Luck, Townsend, R.E., Johnson, L.C., 1979. Relation of frequency-analyzed EEG to monitoring
S., Kappenman, E. (Eds.), Oxford Handbook of Event-Related Potential Components. behavior. Electroencephalography and Clinical Neurophysiology 47, 272–279.
Makeig, S., Bell, A.J., Jung, T.-P., Sejnowski, T.J., 1996. Independent component analysis VanRullen, R., Busch, N.A., Drewes, J., Dubois, J., 2011. Ongoing EEG phase as a trial-by-trial
of electroencephalographic data. In: Touretzky, D., Mozer, M., Hasselmo, M. (Eds.), predictor of perceptual and attentional variability. Frontiers in Psychology 2, 60.
Advances in Neural Information Processing Systems, 8, pp. 145–151. Wang, X.-J., 2010. Neurophysiological and computational principles of cortical rhythms
Mathewson, K.E., Gratton, G., Fabiani, M., Beck, D.M., Ro, T., 2009. To see or not to see: in cognition. Physiological Reviews 90, 1195–1268.
prestimulus alpha phase predicts visual awareness. The Journal of Neuroscience Wróbel, A., 2000. Beta activity: a carrier for visual attention. Acta Neurobiologiae
29, 2725–2732. Experimentalis 60, 247–260.
Mathewson, K.E., Fabiani, M., Gratton, G., Beck, D.M., Lleras, A., 2010. Rescuing stimuli Wróbel, A., 2014. Attentional activation in cortico-thalamic loops of the visual system.
from invisibility: inducing a momentary release from visual masking with pre-target In: Werner, J.S., Chalupa, L.M. (Eds.), New Visual Neurosciences. MIT Press, pp.
entrainment. Cognition 115, 186–191. 339–350.
Matlab software by Mathworks; www.mathworks.com. Wróbel, A., Bekisz, M., Waleszczyk, W., 1994. 20 Hz bursts of activity in the cortico-
Moss, M.B., Rosene, D.L., Peters, A., 1988. Effects of aging on visual recognition memory thalamic pathway during attentive perception. In: Pantev, C., Elbert, T., Lutkenhoner, B.
in the rhesus monkey. Neurobiology of Aging 9, 495–502. (Eds.), Oscillatory Event Related Brain Dynamics. Plenum Press, London, pp. 311–324.
Mundy-Castle, A.C., 1951. Theta and beta rhythm in the electroencephalograms of Wróbel, A., Bekisz, M., Kublik, E., Waleszczyk, W., 1994. 20 Hz bursting beta activity in
normal adults. Electroencephalography and Clinical Neurophysiology 3, 477–486. cortico-thalamic system of visually attending cats. Acta Neurobiologiae Experimentalis
Onton, J., Westerfield, M., Townsend, J., Makeig, S., 2006. Interpreting EEG brain 54, 95–107.
dynamics using independent component analysis. Development Through the Life Wróbel, A., Ghazaryan, A., Bekisz, M., Bogdan, W., Kamiński, J., 2007. Two streams of atten-
Cycle, Neuroscience and Behavioral Reviews. tion dependent beta activity in the striate recipient zone of cat's lateral posterior-
Oostendorp, T.F., van Oosterom, A., 1989. Source parameter estimation in inhomoge- pulvinar complex. The Journal of Neuroscience 27, 2230–2240.
neous volume conductors of arbitrary shape. IEEE Transactions on Biomedical Zhang, Y., Chen, Y., Bressler, S.L., Ding, M., 2008. Response preparation and inhibition:
Engineering 36, 382–391. the role of the cortical sensorimotor beta rhythm. Neuroscience 156, 238–246.

Please cite this article as: Gola, M., et al., EEG beta band activity is related to attention and attentional deficits in the visual performance of elderly
subjects, International Journal of Psychophysiology (2013), http://dx.doi.org/10.1016/j.ijpsycho.2013.05.007