Content uploaded by Abolfazl Alipour
Author content
All content in this area was uploaded by Abolfazl Alipour on Feb 01, 2018
Content may be subject to copyright.
UNCORRECTED PROOF
Journal of Integrative Neuroscience 00 (20xx) 1–14 1
DOI 10.3233/JIN-170056
IOS Press
1 1
2 2
3 3
4 4
5 5
6 6
7 7
8 8
9 9
10 10
11 11
12 12
13 13
14 14
15 15
16 16
17 17
18 18
19 19
20 20
21 21
22 22
23 23
24 24
25 25
26 26
27 27
28 28
29 29
30 30
31 31
32 32
33 33
34 34
35 35
36 36
37 37
38 38
39 39
40 40
41 41
42 42
43 43
44 44
45 45
46 46
QEEG-based neural correlates of decision
making in a well-trained eight year-old chess
player
Abolfazl Alipour a,b, Sahar Seifzadeh c, Hadi Aligholi b,dandMohammadNamib,d,e,∗
aDepartment of Psychological and Brain Sciences, Indiana University, Bloomington, Indiana, USA
bNeuroscience Laboratory-NSL (Brain, Cognition and Behavior), Department of Neuroscience, School
of Advanced Medical Sciences and Technologies, Shiraz University of Medical Sciences, Shiraz, Iran
cYoung Researchers and Elite Club, Qazvin Branch, Islamic Azad University, Qazvin, Iran
dDepartment of Neuroscience, School of Advanced Medical Sciences and Technologies, Shiraz
University of Medical Sciences, Shiraz, Iran
eClinical Neurology Research Center, Shiraz University of Medical Sciences, Shiraz, Iran
Received 9 September 2017
Accepted 13 October 2017
Abstract. The neurocognitive substrates of decision making (DM) in the context of chess has appealed to researchers’ interest
for decades. Expert and beginner chess players are hypothesized to employ different brain functional networks when involved in
episodes of critical DM upon chess. Cognitive capacities including, but not restricted to pattern recognition, visuospatial search,
reasoning, planning and DM are perhaps the key determinants of rewarding and judgmental decisions in chess. Meanwhile, the
precise neural correlates of DM in this context has largely remained elusive. The quantitative electroencephalography (QEEG)
is an investigation tool possessing a proper temporal resolution in the study of neural correlates of cognitive tasks at cortical
level. Here, we used a 22-channel EEG setup and digital polygraphy in a well-trained 8 year-old boy while engaged in playing
chess against the computer. Quantitative analyses were done to map and source-localize the EEG signals. Our analyses indicated
a lower power spectral density (PSD) for higher frequency bands in the right hemisphere upon DM-related epochs. Moreover,
the information flow upon DM blocks in this particular case was more of posterior towards anterior brain regions.
Keywords: Chess, decision making, QEEG, power spectra, functional connectivity
1. Introduction
Decision making (DM), characterized as the act of selecting the best among different alternatives,
plays a defining role in various aspects of our life. Complex human cognition, such as DM under uncer-
tainty, is represented through a dynamic spatio-temporal activity in the brain [15]. While one’s wise and
informed decisions may contribute to success and satisfaction, ill-advised decisions often lead to failure.
The multifaceted process of decision making is potentially linked to a wide range of variables including
input, process, output and feedback [10,19].
*Corresponding author: Mohammad Nami, Department of Neuroscience, School of Advanced Medical Sciences and
Technologies, Shiraz University of Medical Sciences, Shiraz 71348-14336, Iran. E-mail: torabinami@sums.ac.ir.
[research-article] p. 1/14
0219-6352/17/$35.00 © 2017 – IOS Press and the authors. All rights reserved
UNCORRECTED PROOF
2A. Alipour et al. / Please provide the short title
1 1
2 2
3 3
4 4
5 5
6 6
7 7
8 8
9 9
10 10
11 11
12 12
13 13
14 14
15 15
16 16
17 17
18 18
19 19
20 20
21 21
22 22
23 23
24 24
25 25
26 26
27 27
28 28
29 29
30 30
31 31
32 32
33 33
34 34
35 35
36 36
37 37
38 38
39 39
40 40
41 41
42 42
43 43
44 44
45 45
46 46
A popular strategy board game like chess brings a compact and easily controllable task environ-
ment for decision making. As such, many researchers have become interested in using chess as the
task paradigm of choice to assess the neural dynamics of various mental and cognitive skills including
DM [7,9,12,24]. In a landmark study on chess players which was done several decades ago by de Groot,
key findings postulated that expertise in chess was more determined by pattern recognition rather than
search. The general image on the cognitive mechanisms involved in critical DM in the context of chess
has been largely transformed since de Groot’s study [7].
Along these lines, results from other investigations [25,26] argued that intelligence plays key part
both in DM and reasoning. In this way, some studies proposed solid correlations between the IQ scores
and the reasoning capacity of individuals [27]. With regard to the processing speed, when players are
forced to play faster, their ability during the play tends to be less predictive. However, expert players
are shown to perform noticeably better than novice chess players in terms of rapid object recognition
abilities [5,29].
There is a growing trend especially among young adolescents to join chess clubs, however the question
whether playing chess may improve their global DM skills remains elusive. It is then the matter of debate
whether and how playing chess may empower adolescents’ brain in more productive and error free, fast,
wise and rewarding DM.
In case the quantitative EEG (QEEG) provides us with new insights about the cortical brain regions
involved in critical DM among expert chess players new avenues may open for DM-chess related re-
search. Potentially through reverse engineering, areas involved in a grand average QEEG of elite chess
players may define a pathway toward targeting more or less same networks in the brain of novice players
helping them to gain DM, at least in chess, in a faster and more efficient way.
The emergence of neurotechnological tools such as transcranial magnetic stimulation (TMS), tran-
scranial direct current stimulation (tDCS) and neuro-feedback provides novel approaches for cognitive
empowerment through modulation of the involved brain networks [3].
While our working-team pursues its long way to systematically investigate the above, the present case
study highlights the results of our pilot assessments on the neural correlates of DM at cortical level using
a 22 channel-QEEG recording setup in a well-trained eight year-old male chess player when engaged in
playing chess against the computer.
2. Case study
2.1. Subject
Our case study was done on P.A., an 8-year old right-handed male elite chess player with regional and
national recognition in chess competitions. The subject was instructed to sleep well the night before the
experiment and was not under the effect of any medicine or stimulant food or drink. Both the subject
and his legal guardian read and signed the informed consent to take part in the study.
2.2. Experimental procedure
Subject played the Chess Titans on Windows 10 at a self-claimed challenging difficulty level (level 6)
using the mouse and right hand only. Subject played the game for 15 minutes and eventually lost to the
computer.
[research-article] p. 2/14
UNCORRECTED PROOF
A. Alipour et al. / Please provide the short title 3
1 1
2 2
3 3
4 4
5 5
6 6
7 7
8 8
9 9
10 10
11 11
12 12
13 13
14 14
15 15
16 16
17 17
18 18
19 19
20 20
21 21
22 22
23 23
24 24
25 25
26 26
27 27
28 28
29 29
30 30
31 31
32 32
33 33
34 34
35 35
36 36
37 37
38 38
39 39
40 40
41 41
42 42
43 43
44 44
45 45
46 46
2.3. EEG recording and pre-processing
EEG was recorded at two different conditions including the resting state and critical DM blocks while
playing chess. A 22-channel bipolar EEG montage using a 32-channel amplifier system (3840, NR-
SIGN, BC, Canada) was used for QEEG data acquisition. The channel dipoles, based on the interna-
tional 10-20 system, included FP2-F7, FP2-T5, F8-F3, F8-P3, F4-C3, F4-P1, T4-O1, C4-O1, T6-P1,
FP1-FP2, P1-P2, FP2-C4, FP2-P2, F8-T6, F4-T4, F4-O2, T4-P4, C4-P4, T6-O2, O2-P2, O1-O2 where
FP stands for frontopolar, C for central, F for frontal, P for parietal, T for temporal and O for occipi-
tal.
For the resting state condition, subject was sat in the dim-lit recording room (320 lux) and fixated on
a crosshair for ten minutes. Resting state EEG recording continued for 10 minutes.
EEG was recorded throughout the game period and the subject verbally indicated the start of a critical
DM epoch which eventually ended in moving a piece. The reported critical DM epochs (n=6) were
extracted for further analysis (mean duration 14 ±4 seconds).
Other than the EEG recording, subject was attached to a digital sampling unit for autonomic system
polygraphy. By this, real-time galvanic skin conductance (GSC) and heart rate variability (HRV) were
recorded using the Vilistus DSU, UK [21](Fig.1).
All EEG signals were then imported in EEGLAB version 13.0.0 running on MATLAB R2013a. Pre-
processing of the signals included the rejection of visually detectable artifacts and applying a low-pass
filter of 1 Hz and a high-pass filter of 48 Hz at the first step. Afterwards, an Infomax-based independent
Fig. 1. Data acquisition setup. The study subject engaged in chess against the computer. He reported 6 episodes of critical deci-
sion making challenge during which real-time EEG recording was marked and subsequently grand averaged in brain mapping.
The galvanic skin conductance (GSC) and heart rate variability (HRV) were also being monitored during the whole process
with special marking upon challenging decision making episodes.
[research-article] p. 3/14
UNCORRECTED PROOF
4A. Alipour et al. / Please provide the short title
1 1
2 2
3 3
4 4
5 5
6 6
7 7
8 8
9 9
10 10
11 11
12 12
13 13
14 14
15 15
16 16
17 17
18 18
19 19
20 20
21 21
22 22
23 23
24 24
25 25
26 26
27 27
28 28
29 29
30 30
31 31
32 32
33 33
34 34
35 35
36 36
37 37
38 38
39 39
40 40
41 41
42 42
43 43
44 44
45 45
46 46
component analysis (ICA) decomposition algorithm (runica.m in EEGLAB) was used in order to find
possible artifacts. Four of the 22 components were identified as artifactual (3 eye movement artifacts
and 1 muscle artifact component). After removal of these artifactual components, the ICA algorithm
was used once again to achieve 22 artifact-free components. The obtained EEG signals were then used
in two different analysis platforms including EEGLAB and NeuroGuide (v.2.3.8, Applied Neuroscience,
USA).
2.4. Topographic mapping of EEG signals
Corresponding to the 6 critical DM epochs, the resting state EEG was segmented into 6 artifact free
epochs for comparison. The NeuroGuide software was used to analyze and plot the topographic maps
and power spectra of the EEG activity during both resting state and critical DM epochs. The abso-
lute and relative powers of the signals in theta, alpha, and beta bands were calculated and compared
between the resting state and critical DM epochs. Moreover, coherence and power ratio in the same
frequency bands were computed and compared between the two states. The Wilcoxon signed-rank test
was used to compare the difference between relative and absolute power spectra, power ratio, and co-
herence between the two states. Pvalues <0.05 were considered as statistically significant throughout
the analyses.
2.5. EEG source connectivity analysis
The DIPFIT2 plug-in of EEGLAB was used to source localize the EEG signals. Size of the head model
was modified in order to match the size of the 8 year-old subject based on the insights from BenAbdelka-
der et al. 2008 [4,8]. In short, the DIPFIT2 plug-in uses ICA components to estimate sources of EEG
activity in the brain. For the 22 components existing in the signals, 22 sources have been calculated.
The Source Information Flow Toolbox (SIFT) [8,20] was also used to estimate the information flow
between sources. In addition, the direct Directed Transfer Function (dDTF) [16] was used as a measure
of connectivity between the 22 components. The obtained number was averaged for all time-points and
all frequencies between 4 to 46 Hz (with 2 Hz intervals) in both resting state and critical DM epochs
to obtain a 22 ×22 ‘overall connectivity matrix’. Overall connectivity matrix of the decision making
epochs was subtracted from overall connectivity matrix in the resting state in order to obtain a connec-
tivity matrix revealing the differences in information flow between the two states. Consequently, major
sink and source nodes were identified.
3. Results
The grand average QEEG amplitude data for DM epochs versus resting state turned to show sig-
nificantly higher absolute power in the range of theta and alpha frequency bands in the right fronto-
central and temporoparietal brain regions upon task-positive brain cortical activity versus resting state.
Furthermore, except for the F4-C3 dipole, beta amplitude was less in DM states compared to resting
state.
Figures 2–4demonstrate the comparative absolute power values for theta, alpha and beta spectral
bands. As illustrated in Fig. 2, theta amplitude was found to significantly dominate upon DM epochs
grand average versus resting state QEEG in F8-T3 [t(5)=2.1, %95CI =(−1.56)–(−0.03),p=
0.03], F4-C3 [t(5)=2.85, %95CI =(−2.42)–(−0.41),p=0.006], T4-P3 [t(5)=2.02, %95CI =
[research-article] p. 4/14
UNCORRECTED PROOF
A. Alipour et al. / Please provide the short title 5
1 1
2 2
3 3
4 4
5 5
6 6
7 7
8 8
9 9
10 10
11 11
12 12
13 13
14 14
15 15
16 16
17 17
18 18
19 19
20 20
21 21
22 22
23 23
24 24
25 25
26 26
27 27
28 28
29 29
30 30
31 31
32 32
33 33
34 34
35 35
36 36
37 37
38 38
39 39
40 40
41 41
42 42
43 43
44 44
45 45
46 46
Fig. 2. Spiderweb chart plotting the absolute power values for theta frequency band upon decision making (orange) and resting
state (blue) QEEG data. The analysis revealed higher theta power upon decision making task in F8-T3, F4-C3, T4-P3, C4-P3
and T6-O1 dipoles compared to resting state. Values are presented in μV2.#P<0.05, ##P<0.01 and ###P<0.001. FP:
frontopolar, C: central, F: frontal, P: parietal, T: temporal and O: occipital.
(−2.26)–(−0.03),p=0.04], C4-P3 [t(5)=4.48, %95CI =(0.95)–(2.4),p<0.001], and T6-O1
[t(5)=2.76, %95CI =(−1.41)–(−0.005),p=0.01] dipoles.
Figure 3illustrates the difference in absolute power for alpha frequency band in DM grand aver-
age epochs versus resting state in various dipole locations. Alpha amplitude was found to signifi-
cantly dominate upon DM epochs grand average versus resting state QEEG in F4-C3 [t(5)=2.93,
%95CI =(−2.36)–(−0.71),p=0.009]. F8-T3 however, showed a dominant alpha upon resting state
compared to task-positive epochs [t(5)=2.22, %95CI =(−1.57)–(−0.08),p=0.03].
[research-article] p. 5/14
UNCORRECTED PROOF
6A. Alipour et al. / Please provide the short title
1 1
2 2
3 3
4 4
5 5
6 6
7 7
8 8
9 9
10 10
11 11
12 12
13 13
14 14
15 15
16 16
17 17
18 18
19 19
20 20
21 21
22 22
23 23
24 24
25 25
26 26
27 27
28 28
29 29
30 30
31 31
32 32
33 33
34 34
35 35
36 36
37 37
38 38
39 39
40 40
41 41
42 42
43 43
44 44
45 45
46 46
Fig. 3. Spiderweb chart plotting the absolute power values for alpha frequency band upon decision making (orange) and resting
state (blue) QEEG data. The analysis revealed higher alpha power upon decision making task in F4-C3 dipole compared
to resting state, ##P<0.01. The F8-T3 dipole however showed a dominant alpha power upon resting state compared to
task-positive epochs, *P<0.05. Values are presented in μV2. FP: frontopolar, C: central, F: frontal, P: parietal, T: temporal
and O: occipital.
AsshowninFig.4, beta power was diminished in the anterior brain regions in task-positive states
compared to resting QEEG. In other words, unlike the higher amplitude for beta in many derivations
upon resting state, alpha and theta amplitude took over in right hemisphere dipoles in the task-positive
state. Based on our comparative analyses, beta power was found to be predominantly higher in rest-
ing state than DM epochs in F8-T3 [t(5)=2.72, %95CI =(0.034)–(2.37),p=0.008], F8-F4
[t(5)=2.93, %95CI =(−2.66)–(−0.47),p=0.006], F4-C4 [t(5)=3.11, %95CI =(0.83)–(1.3),
[research-article] p. 6/14
UNCORRECTED PROOF
A. Alipour et al. / Please provide the short title 7
1 1
2 2
3 3
4 4
5 5
6 6
7 7
8 8
9 9
10 10
11 11
12 12
13 13
14 14
15 15
16 16
17 17
18 18
19 19
20 20
21 21
22 22
23 23
24 24
25 25
26 26
27 27
28 28
29 29
30 30
31 31
32 32
33 33
34 34
35 35
36 36
37 37
38 38
39 39
40 40
41 41
42 42
43 43
44 44
45 45
46 46
Fig. 4. Spiderweb chart plotting the absolute power values for beta frequency band upon decision making (orange) and resting
state (blue) QEEG data. The analysis revealed higher beta power in resting state than DM epochs in F8-T3, F8-F4, F4-C4,
C4-P4, and FP2-F3. *P<0.05, **P<0.01, ##P<0.01. Values are presented in μV2. FP: frontopolar, C: central, F: frontal,
P: parietal, T: temporal and O: occipital.
p=0.004], C4-P4 [t(5)=2.22, %95CI =(−1.46)–(−0.07),p=0.03] and FP2-F3 [t(5)=2.16,
%95CI =(−1.54)–(0.04),p=0.04].
The FFT relative power analysis revealed an increased alpha band power in the right centroparietal
region in DM versus resting states. In addition, beta relative power was diminished in FP2 region upon
DM epochs. Such a decrease was evident in FP1 across high beta band frequency.
[research-article] p. 7/14
UNCORRECTED PROOF
8A. Alipour et al. / Please provide the short title
1 1
2 2
3 3
4 4
5 5
6 6
7 7
8 8
9 9
10 10
11 11
12 12
13 13
14 14
15 15
16 16
17 17
18 18
19 19
20 20
21 21
22 22
23 23
24 24
25 25
26 26
27 27
28 28
29 29
30 30
31 31
32 32
33 33
34 34
35 35
36 36
37 37
38 38
39 39
40 40
41 41
42 42
43 43
44 44
45 45
46 46
In terms of FFT power ratio, delta/theta and delta/alpha ratios were increased in FP1 upon DM epochs.
Moreover, centroparietal regions demonstrated a rise in alpha/beta and alpha/high beta power ratio.
Figure 5demonstrates the spatial distribution and connectivity brain maps for resting states (sec-
tion (A)) versus the grand average for the six DM blocks (section (B)) suggesting lower than expected
Fig. 5. The QEEG topographical spectral brain maps. The analysis demonstrates absolute power values across spectra upon
resting state (section (A), upper panel) and decision making QEEG data (panel (B), upper panel). The resting state QEEG’s
centroparietal activity is compatible with default mode network activity. Grand average for the six decision making blocks
(section (B)) suggests lower than expected beta power in the anterior brain regions rather than occipital and centroparietal areas.
The lower panels in sections (A) and (B) in turn represent resting state and decision making beta coherence maps indicating a
frontal beta hypocoherence upon task-positive rather than resting states.
[research-article] p. 8/14
UNCORRECTED PROOF
A. Alipour et al. / Please provide the short title 9
1 1
2 2
3 3
4 4
5 5
6 6
7 7
8 8
9 9
10 10
11 11
12 12
13 13
14 14
15 15
16 16
17 17
18 18
19 19
20 20
21 21
22 22
23 23
24 24
25 25
26 26
27 27
28 28
29 29
30 30
31 31
32 32
33 33
34 34
35 35
36 36
37 37
38 38
39 39
40 40
41 41
42 42
43 43
44 44
45 45
46 46
Fig. 5. (Continued.)
beta power in the anterior brain regions rather than occipital and centroparietal areas. This perhaps indi-
cates that DM process in our subject was more of an automatic and less of cognitive nature regulated by
subcortical (corticostrial) networks rather than cortico-cortical pathways. Interestingly, beta coherence
in anterior brain regions diminished upon DM epochs (grand average for 6 blocks, section (B), lower
panel) compared to resting state (Section (A), lower panel) (Fig. 5).
[research-article] p. 9/14
UNCORRECTED PROOF
10 A. Alipour et al. / Please provide the short title
1 1
2 2
3 3
4 4
5 5
6 6
7 7
8 8
9 9
10 10
11 11
12 12
13 13
14 14
15 15
16 16
17 17
18 18
19 19
20 20
21 21
22 22
23 23
24 24
25 25
26 26
27 27
28 28
29 29
30 30
31 31
32 32
33 33
34 34
35 35
36 36
37 37
38 38
39 39
40 40
41 41
42 42
43 43
44 44
45 45
46 46
Fig. 6. Autonomic response upon decision making task. Panels (A)-(C) represent heart rate variability (HRV), pNN50 (number
of pairs of successive NN intervals that differ by more than 50 ms) and GSR (in μSiemens). The yellow shades represent
decision making (DM) blocks of which the EEG data were extracted and analyzed. Despite no significant difference in GSR
in DM and non-DM blocks there was an apparent increase in HRV and PNN50 in DM blocks. This suggests an autonomic
component involved when the subject is engaged in DM task.
With regard to the autonomic response data upon DM blocks, Fig. 6shows an increased HRV and
interbeat group change but not GSR in DM blocks as highlighted in yellow. This suggests an autonomic
component involved when the subject is engaged (even subcortically) in DM task (Fig. 6).
Figure 7represents the EEG source connectivity analysis. As outlined in the methods section, an
Infomax-based ICA decomposition algorithm using the EEGLAB in MATLAB yielded 22 artifact-free
components. Based on our analysis, the source of information flow upon DM block EEG analyses
were localized at components 3, 4, 13, 14, 15 and 16 which were anatomically linked to a left ven-
tromedial prefrontal cortex (vmPFC) as well as right occipital and right medial temporal cortices. On
the other hand, the information flow sink were localized at components 9, 11, 19 and 22 which were
linked to the left orbitofrontal cortex (OFC), left PPC and right intraparietal sulcus (IPS). As such, it
turned out that the information flow upon DM blocks were more of posterior towards anterior brain
regions.
[research-article] p. 10/14
UNCORRECTED PROOF
A. Alipour et al. / Please provide the short title 11
1 1
2 2
3 3
4 4
5 5
6 6
7 7
8 8
9 9
10 10
11 11
12 12
13 13
14 14
15 15
16 16
17 17
18 18
19 19
20 20
21 21
22 22
23 23
24 24
25 25
26 26
27 27
28 28
29 29
30 30
31 31
32 32
33 33
34 34
35 35
36 36
37 37
38 38
39 39
40 40
41 41
42 42
43 43
44 44
45 45
46 46
Fig. 7. Information flow between major sources and sinks upon critical decision making. (A) (left to right): position of informa-
tion sources in sagittal, coronal, and transverse sections. (B): the color matrix for information flow between all signal sources.
Each square represents the information flow difference between Resting and DM states for a specific pair of signal sources. For
instance, the red color in element (4, 3) indicates an increased flow of information and the causal effect from signal source 3 on
signal source 4 during DM epochs. (C): Position of information sources in 3D space. (D) (left to right): position of “information
sinks” in sagittal, coronal, and transverse sections. (E): the reciprocal color matrix (versus (B)) for information flow between
all signal sources. (F): Position of information sinks in 3D space. Sources and sinks were identified by thresholding the amount
of their causal effect. While the sources of information flow upon DM blocks were localized at components 3, 4, 13, 14, 15 and
16 (anatomically linked to the left vmPFC, right occipital and right medial temporal cortices); the information flow sinks were
localized at components 9, 11, 19 and 22 ( anatomically linked to the left OFC, left PPC and right IPS).
4. Discussion
The present case study was an attempt to investigate the plausible cortical networks potentially in-
volved in DM task while a well-trained 8 year-old boy was engaged in chess play against the computer.
Results from the QEEG analysis indicated a lower power spectral density (PSD) for higher frequency
bands power in the right hemisphere upon DM epochs. Further analyses suggested that the information
flow upon DM blocks in this particular case were more of posterior towards anterior brain regions.
The inferior frontal gyrus has also been proposed to be centrally involved in DM processes [28]. In the
present investigation however, we observed predominant theta frequency upon DM epochs in the right
[research-article] p. 11/14
UNCORRECTED PROOF
12 A. Alipour et al. / Please provide the short title
1 1
2 2
3 3
4 4
5 5
6 6
7 7
8 8
9 9
10 10
11 11
12 12
13 13
14 14
15 15
16 16
17 17
18 18
19 19
20 20
21 21
22 22
23 23
24 24
25 25
26 26
27 27
28 28
29 29
30 30
31 31
32 32
33 33
34 34
35 35
36 36
37 37
38 38
39 39
40 40
41 41
42 42
43 43
44 44
45 45
46 46
inferior frontal included dipole. The hypothetical explanation for this observation may revolve around
the role of dorsolateral prefrontal cortex and inferior frontal gyrus in the process of DM when cognitive
control plays a central role [6]. This was apparently found to be less employed in this particular case
study when the subject was engaged in critical DM process. The neural correlates of decision making
is known to involve at least the anterior cingulate cortex, middle frontal gyrus, and inferior frontal
gyrus/insula, with recent insights suggesting that decisions may emerge from distributed processes [28,
31].
Some earlier reports have indicated the role of left parietal theta power as a correlate of memory
retrieval and DM [14,31]. Our finding was in agreement with the above results since a major sink of
information flow in our QEEG analysis was found to be the left posterior parietal cortex (PPC). Indeed,
unraveling the neural mechanism of our observation might help to explain some key electrophysiological
determinants of DM. As outlined in the results section, the electrophysiological difference associated
with decision-making epochs mainly corresponded to the distribution and power of theta frequency band
in fronto-central and posterior parietal cortices. Based on our QEEG findings and source connectivity
analysis, we propose a neural model more of an automatic and less of cognitive nature regulating the
process of DM in our tested elite chess player.
Studies have considered chess as a task paradigm to evaluate the players’ brain activity under am-
biguous circumstances and time pressure. For instance, a number of neuroimaging investigations using
the functional magnetic resonance imaging (fMRI) strived to localize the neural activity upon percep-
tual DM. According to such evidence, cortical and subcortical brain regions and structures including
the frontal and parietal cortices, thalamus and striatum were found to be largely involved in modulating
decisions’ accuracy and uncertainty [11,13,22,23]. In other well-designed studies, expert chess play-
ers’ memory recall was compared with that of beginners or less skilled players with the whole-brain
analysis conducted especially on regions of interest such as anterior cingulate cortex (ACC), bilateral
intraparietal sulci (IPS), bilateral ventromedial and dorsolateral prefrontal cortices (vmPFC and dlPFC)
and prefrontal cortices (PFC). Results from such studies on chess player brains corroborated that com-
ponents of the frontoparietal network (FPN) are not only linked to consciousness and attention but also
working memory [9,11,24]. Despite the above, in our case study, the FPN was shown to be less involved
once this well-trained chess player was making critical decisions. Instead centroparietal areas were found
to shower higher amplitude for beta frequency band in QEEG may and the sinks for information flow
(estimated through ICA) turned to be the left PPC and right IPS.
Assessment tools such as EEG are not only less costly and more convenient to administer but also
capable of rendering a proper temporal resolution, hence may be considered appropriate to address the
temporal sequencing of DM signals [1]. For instance, computational model-based approach to EEG data
acquired during simple binary choice task have shown to yield dependable data on the temporal sequence
of information flow in the brain [17,18].
In studies which examined the pattern recognition of four simple conditions in chess, evoked coher-
ences of EEG signals were found to be sensitive to sensory as well as mental activity (theta and beta
coherence, respectively). Meanwhile, beta coherence considerably depended on the task type [30].
When it comes to perceptual DM, the process is more error-prone, especially near the threshold.
Despite the proposed internal noise in neural systems which seems to be responsible for such errors,
it appears that a mixture of bottom-up and top-down sources drive this potential complexity [2]. Such
a complexity in interpreting data should be considered when EEG is used as the method of choice in
neurocognitive studies.
[research-article] p. 12/14
UNCORRECTED PROOF
A. Alipour et al. / Please provide the short title 13
1 1
2 2
3 3
4 4
5 5
6 6
7 7
8 8
9 9
10 10
11 11
12 12
13 13
14 14
15 15
16 16
17 17
18 18
19 19
20 20
21 21
22 22
23 23
24 24
25 25
26 26
27 27
28 28
29 29
30 30
31 31
32 32
33 33
34 34
35 35
36 36
37 37
38 38
39 39
40 40
41 41
42 42
43 43
44 44
45 45
46 46
5. Conclusion
Our present QEEG findings propose a less cognitive load upon DM task in this particular subject
based on the dominance of right and posterior alpha/theta versus beta frequency bands. As such, the
FPN was shown to be less involved once our subject was involved in critical DM. This data may suggest
that perhaps an early system responds preferentially to outcomes only in order to initiate a fast automatic
alertness response when someone is an expert chess player.
Though our research opens up new avenues for the investigation of the neural system underlying nor-
mal DM, future studies should demonstrate the level of FPN and subcortical nuclei involvement when
elite versus novice chess players engage in critical DM tasks. The combined use of advanced neurotech-
nological tools such as fMRI, functional near infra-red spectroscopy (fNIRS) and magnetoencephalog-
raphy (MEG) would offer novel opportunities for more in-depth investigation of the above.
6. Conflict of interest
None.
Acknowledgement
Authors would like to thank Ms. S. Safaie for her assistance with this report.
References
[1] A. Alipour, A. Mojdehfarahbakhsh, A. Tavakolian, T. Morshedzadeh, M. Asadi, A. Mehdizadeh and M. Nami, Neural
communication through theta-gamma cross-frequency coupling in a bistable motion perception task, Journal of integrative
neuroscience 04 (2016), 539–551. doi:10.1142/S0219635216500291.
[2] S. Amitay, J. Guiraud, E. Sohoglu, O. Zobay, B.A. Edmonds, Y.X. Zhang and D.R. Moore, Human decision making based
on variations in internal noise: An EEG study, PloS one 7(2013), e68928. doi:10.1371/journal.pone.0068928.
[3] N. Ashjazadeh, R. Boostani, H. Ekhtiari, M. Emamghoreishi, M. Farrokhi, A. Ghanizadeh, G. Hatam, H. Hadianfard,
M. Lotfi and S.M.J. Mortazavi, Operationalizing Cognitive Science and Technologies’ Research and Development; The
“Brain and Cognition Study Group (BCSG)” Initiative from Shiraz, Iran, Basic and Clinical Neuroscience 2, 2014, p. 104.
[4] C. BenAbdelkader and Y. Yacoob, Statistical estimation of human anthropometry from a single uncalibrated image, Com-
putational Forensics (2008), 200–220.
[5] B.D. Burns, The effects of speed on skilled chess performance, Psychological Science 7(2004), 442–447. doi:10.1111/j.
0956-7976.2004.00699.x.
[6] E.C. Cieslik, K. Zilles, S. Caspers, C. Roski, T.S. Kellermann, O. Jakobs, R. Langner, A.R. Laird, P.T. Fox and S.B. Eick-
hoff, Is there “one” DLPFC in cognitive action control? Evidence for heterogeneity from co-activation-based parcellation,
Cerebral cortex 11 (2013), 2677–2689. doi:10.1093/cercor/bhs256.
[7] M.H. Connors, B.D. Burns and G. Campitelli, Expertise in complex decision making: The role of search in chess 70 years
after de Groot, Cognitive science 8(2011), 1567–1579. doi:10.1111/j.1551-6709.2011.01196.x.
[8] A. Delorme, T. Mullen, C. Kothe, Z.A. Acar, N. Bigdely-Shamlo, A. Vankov and S. Makeig, EEGLAB, SIFT, NFT,
BCILAB, and ERICA: New tools for advanced EEG processing, Computational intelligence and neuroscience 2011
(2011), 130714.
[9] P. Eisele, Judgment and decision-making: Experts’ and novices’ evaluation of chess positions, Perceptual and motor skills
1(2004), 237–248. doi:10.2466/pms.98.1.237-248.
[10] D.P. Forbes and F.J. Milliken, Cognition and corporate governance: Understanding boards of directors as strategic
decision-making groups, Academy of management review 3(1999), 489–505.
[11] J. Grinband, J. Hirsch and V.P. Ferrera, A neural representation of categorization uncertainty in the human brain, Neuron
5(2006), 757–763. doi:10.1016/j.neuron.2006.01.032.
[research-article] p. 13/14
UNCORRECTED PROOF
14 A. Alipour et al. / Please provide the short title
1 1
2 2
3 3
4 4
5 5
6 6
7 7
8 8
9 9
10 10
11 11
12 12
13 13
14 14
15 15
16 16
17 17
18 18
19 19
20 20
21 21
22 22
23 23
24 24
25 25
26 26
27 27
28 28
29 29
30 30
31 31
32 32
33 33
34 34
35 35
36 36
37 37
38 38
39 39
40 40
41 41
42 42
43 43
44 44
45 45
46 46
[12] D.H. Gustafson, K. Bosworth, R.P. Hawkins, E.W. Boberg and E. Bricker, CHESS: A computer-based system for provid-
ing information, referrals, decision support and social support to people facing medical and other health-related crises, in:
Proceedings Symposium on Computer Applications in Medical Care, 1992, pp. 161–165.
[13] H.R. Heekeren, S. Marrett, P.A. Bandettini and L.G. Ungerleider, A general mechanism for perceptual decision-making
in the human brain, Nature 7010 (2004), 859–862. doi:10.1038/nature02966.
[14] J. Jacobs, G. Hwang, T. Curran and M.J. Kahana, EEG oscillations and recognition memory: Theta correlates of memory
retrieval and decision making, Neuroimage 2(2006), 978–987. doi:10.1016/j.neuroimage.2006.02.018.
[15] L. Jollans, R. Whelan, L. Venables, O.H. Turnbull, M. Cella and S. Dymond, Computational EEG modelling of decision
making under ambiguity reveals spatio-temporal dynamics of outcome evaluation, Behavioural brain research 321 (2017),
28–35. doi:10.1016/j.bbr.2016.12.033.
[16] A. Korzeniewska, M. Ma´
nczak, M. Kami´
nski, K.J. Blinowska and S. Kasicki, Determination of information flow direction
among brain structures by a modified directed transfer function (dDTF) method, Journal of neuroscience methods 1
(2003), 195–207. doi:10.1016/S0165-0270(03)00052-9.
[17] S.P. Kyathanahally, A. Franco-Watkins, X. Zhang, V.D. Calhoun and G. Deshpande, A realistic framework for investigat-
ing decision making in the brain with high spatiotemporal resolution using simultaneous EEG/fMRI and joint ICA, IEEE
Journal of Biomedical and Health Informatics 3(2017), 814–825. doi:10.1109/JBHI.2016.2590434.
[18] T. Larsen and J.P. O’Doherty, Uncovering the spatio-temporal dynamics of value-based decision-making in the human
brain: A combined fMRI–EEG study, Phil Trans R Soc B 1655 (2014), 20130473. doi:10.1098/rstb.2013.0473.
[19] M. Mahmoud, Y. Liu, H. Hartmann, S. Stewart, T. Wagener, D. Semmens, R. Stewart, H. Gupta, D. Dominguez and
F. Dominguez, A formal framework for scenario development in support of environmental decision-making, Environmen-
tal Modelling & Software 7(2009), 798–808. doi:10.1016/j.envsoft.2008.11.010.
[20] T.R. Mullen, The Dynamic Brain: Modeling Neural Dynamics and Interactions from Human Electrophysiological Record-
ings, University of California, San Diego, 2014.
[21] F. Nazaraghaie, F. Torkamani, B. Kiani and M. Torabi-Nami, EEG-guided meditative training through geometrical ap-
proach: An interim analysis, Avicenna Journal of Phytomedicine (2015).
[22] M.P. Paulus, N. Hozack, L. Frank and G.G. Brown, Error rate and outcome predictability affect neural activation in
prefrontal cortex and anterior cingulate during decision-making, Neuroimage 4(2002), 836–846. doi:10.1006/nimg.2001.
1031.
[23] M.N. Shadlen and W.T. Newsome, Neural basis of a perceptual decision in the parietal cortex (area LIP) of the rhesus
monkey, Journal of neurophysiology 4(2001), 1916–1936.
[24] M. Sigman, P. Etchemendy, D.F. Slezak and G.A. Cecchi, Response time distributions in rapid chess: A large-scale
decision making experiment, Frontiers in neuroscience 4(2010), 60.
[25] K. Stanovich, R. West and R. Hertwig, Individual differences in reasoning: Implications for the rationality debate? – Open
Peer Commentary – The questionable utility of cognitive ability in explaining cognitive illusions, 2000.
[26] K.E. Stanovich and R.F. West, Individual differences in reasoning: Implications for the rationality debate?, Behavioral
and brain sciences 5(2000), 645–665. doi:10.1017/S0140525X00003435.
[27] R.J. Sternberg, The ability is not general, and neither are the conclusions, Behavioral and Brain Sciences 5(2000), 697–
698. doi:10.1017/S0140525X00573439.
[28] A. Thielscher and L. Pessoa, Neural correlates of perceptual choice and decision making during fear-disgust discrimi-
nation, The Journal of neuroscience: The official journal of the Society for Neuroscience 11 (2007), 2908–2917. doi:10.
1523/JNEUROSCI.3024-06.2007.
[29] H.L. Van Der Maas and E.-J. Wagenmakers, A psychometric analysis of chess expertise, The American journal of psy-
chology 118 (2005), 29–60.
[30] H.-J. Volke, P. Dettmar, P. Richter, M. Rudolf and U. Buhss, On-coupling and off-coupling of neocortical areas in chess
experts and novices as revealed by evoked EEG coherence measures and factor-based topological analysis – a pilot study,
Journal of Psychophysiology 16 (2002), 23–36. doi:10.1027//0269-8803.16.1.23.
[31] D. Zhang, R. Gu, L.S. Broster, Y. Jiang, W. Luo, J. Zhang and Y.J. Luo, Linking brain electrical signals elicited by current
outcomes with future risk decision-making, Frontiers in behavioral neuroscience 8(2014), 84.
[research-article] p. 14/14