SCIENCE CHINA
Life Sciences
© The Author(s) 2011. This article is published with open access at Springerlink.com life.scichina.com www.springer.com/scp
*Corresponding author (email: renj@zjnu.cn; luoj@psycy.ac.cn)
• RESEARCH PAPERS • October 2011 Vol.54 No.10: 961–965
doi: 10.1007/s11427-011-4233-3
Meditation promotes insightful problem-solving by keeping people
in a mindful and alert conscious state
REN Jun1*, HUANG ZhiHui1, LUO Jing2*, WEI GaoXia2, YING XiaoPing3, DING ZhiGuang4,
WU YiBin5 & LUO Fei2
1College of Education, Zhejiang Normal University, Jinhua 321004, China;
2Key Laboratory of Mental Health, Chinese Academy of Sciences, Beijing 100101, China;
3Institute of Sociology, Chinese Academy of Social Sciences, Beijing 100026, China;
4Department of Social Sciences, Hebei Medical University, Shijiazhuang 050015, China;
5Beijing Easy Monitor Technology Co., Ltd., Beijing 100101, China
Received December 23, 2010; accepted June 27, 2011
Although previous studies have shown that sleep can inspire insight, it is still unclear whether meditation can promote insight.
Meditation differs from other types of passive rest such as relaxation and sleep because it requires full consciousness and
mindfulness of targets such as one’s breathing. Forty-eight university students without meditation experience were recruited to
learn a simple meditation technique. They were given a list of 10 insight problems to solve (the pre-test session). In this study,
we focused on the unsolved problems and examined if they could be successfully solved after a 20 min rest interval with or
without meditation. Results showed that relative to the control group that listened to Chinese or English words and made a
language judgment, the groups who learned meditation successfully solved significantly more failed problems from the pre-test
session, providing direct evidence for the role of meditation in promoting insight. Further analysis showed that maintaining a
mindful and alert state during meditation (raising a hand to report every 10 deep breaths compared to every 100 deep breaths)
resulted in more insight regarding the failed items from the pre-test session. This implies that it was watchfulness in meditation,
rather than relaxation, that actually contributed to insight. Consistently, in the meditation session or control task, the percentage
of alpha waves—a brain index of mental relaxation—was negatively correlated with insight. These results suggest a medita-
tion-based insight-promoting mechanism different from that involved in passive rest such as relaxation and sleep.
meditation, insight, problem solving
Citation: Ren J, Huang Z H, Luo J, et al. Meditation promotes insightful problem-solving by keeping people in a mindful and alert conscious state. Sci China
Life Sci, 2011, 54: 961–965, doi: 10.1007/s11427-011-4233-3
It has been shown that sleep can significantly increase in-
sightful awareness, for example, discerning a hidden princi-
ple needed to solve the Number Reduction Task (NRT) [1].
However, it is still unclear whether meditation can promote
insight. This question deserves to be explored because med-
itation differs from routine forms of rest such as sleep and
relaxation in that one’s attention is actively focused on the
present state of consciousness [27]. It is widely believed in
Asian cultural traditions such as Buddhism and Taoism that
meditation can lead to insight.
This study recruited undergraduate students without any
previous meditation practice as participants. They were
taught a simple method of meditation and were then tested
to determine if application of meditation could promote
insight. Different from Wagner and colleagues’ experiment
that adopted the NRT [1], we used 10 typical insight prob-
lems [8]. We first instructed participants to try to solve the-
se problems one by one. We selected the unsolved problems
962 Ren J, et al. Sci China Life Sci October (2011) Vol.54 No.10
for each participant for later testing. Then, we required par-
ticipants to engage in either 20 min of meditation or a control
cognitive task. Finally, participants were unexpectedly
re-administered their unsolved problems from the first ses-
sion and asked to re-think and solve these problems. The key
hypothesis of this study was that participants who had learned
meditation would be more capable of insightfully solving the
unsolved problems.
To specify how meditation promotes insight, the effect of
meditation with different levels of alertness and awareness
was examined. These levels were manipulated through
changing the frequency of self-reported (by hand-raising)
number of deep breaths. During meditation, participants
were required to clear their minds, focus on deep breathing,
and count number of breaths. In the M10 condition, partici-
pants were required to raise their hand to report every 10
breaths. In the M100 condition, participants reported every
100 breaths by raising their hand. Both M10 and M100 met
the criterion of meditation, but relative to M100, M10 par-
ticipants stayed in a more alert and awake state because they
were required to frequently report their number of breaths.
During the meditation and control cognitive task period,
electroencephalograms of participants were recorded to
monitor and evaluate mind-brain states. Some EEG indices,
such as percentage of alpha waves, which indicate level of
mental relaxation, as well as the score of i22 related to at-
tention focus [9], were considered in obtaining cross-
domain evidence for the efficiency of our experimental ma-
nipulation and for suggesting possible mental mechanisms
related to how meditation promotes insight.
1 Materials and methods
1.1 Participants
Forty-eight healthy undergraduate students (23 males, mean
age 23.3 years) recruited from a university in Beijing par-
ticipated in the study. They were randomly assigned to three
equal groups with balanced male/female ratios: meditation
group M10 (reported every 10 deep breaths), meditation
group M100 (reported every 100 deep breaths), and the
control group. There were 16 participants in each group.
1.2 Materials
Fifteen insight problems that met the criterion of cognitive
reconstruction were used [8]. Five were used as examples in
the instruction and practice session, and 10 were used for
the formal experiment.
1.3 EEG recording and analysis
EEG data were collected by HXD-I EEG equipment
(HUAXIANG Technology Company, Harbin, Heilongjiang,
China). The collecting electrodes were located in the bilat-
eral prefrontal cortex, reference electrodes were A1 and A2,
and leading electrodes were located in the center of frontal
sites. The frequency was 200 Hz, the time window for
analysis was 1.25, the arithmetic stacking coefficient was 6,
the input impedance was 9 Ω, the resolution of EEG was 3
MV, and the frequency band was 0.5–100 Hz. Wavelet
analysis, spectrum analysis, and pattern recognition were
applied in EEG analysis to extract the wavelet index. The
main indices included alpha waves, i-35 related to relaxa-
tion and i-22 related to alertness. The latter two indices
were developed based on our clinical and experimental tri-
als and can be used as a reference index for common EEG
data [9]. The EEG data recording for the last 15 min was
analyzed.
1.4 Procedure
1.4.1 Meditation and problem-solving training
Before the formal experimental session, participants were
taught Susoku meditation, which requires participants to
mindfully control and focus on their deep abdominal
breathing [1012]. In our pilot study, we found a cartoon
computer animation very helpful for beginners learning the
skill of deep abdominal breathing. Participants were asked
to attentively and slowly take a breath in or out as the “in-
spiration fish” or “expiration fish” gradually and smoothly
went up or down on the computer screen. Each round of
inspiration and expiration lasted 7 s. After participants
grasped the method of deep abdominal breathing, they were
further instructed to attentively take inspiration and expira-
tion breaths without the external aid of the computer anima-
tion (with eyes slightly closed). They were also instructed to
do their best to be mindfully aware and disengage from any
irrelevant thinking, free association, or feelings that might
incidentally occur. During meditation, participants were
asked to count the number of their breaths from 1 to 10 and
then restart (the M10 condition), or from 1 to 100 and then
restart (the M100 condition). They were also asked to raise
their hand to report finishing one round of breaths (i.e., 10
breaths in M10 or 100 breaths in M100). The above-
mentioned training procedures as well as subsequent prob-
lem solving and meditation practice were conducted with
participants individually. Participants in the control group
did not learn meditation. They rested in the waiting room
for the same period of time.
After meditation (the M10 and M100 group) or rest (the
control group), all three groups were given five examples of
insight problems for practice. Every problem was presented
on the screen for a maximum of 3 min. If participants
thought the problem was successfully solved, they were
instructed to press the “2” button as soon as possible and
then write down the solution. If participants could not solve
the problem, they could wait until the problem disappeared
from the screen at the end of the 3 min maximum, or press
the “3” key to stop the presentation and go to the next item.
Ren J, et al. Sci China Life Sci October (2011) Vol.54 No.10 963
It was emphasized that participants should not continue to
think about the unsolved problem once it disappeared from
the screen. Participants did not know they would be given
the unsolved problems later and would have to try to solve
them again.
1.4.2 Formal experimental session
The experimental session included the pre-test, the interval,
and the post-test.
(i) Pre-test. In the pre-test phase, participants were in-
structed to solve 10 insight problems appearing randomly
one by one on the screen using the same procedure as in the
practice session.
(ii) Meditation or control cognitive task interval. During
this interval, the M10 and M100 groups were instructed to
count number of breaths for 20 min, while the control group
was instructed to complete an auditory judgment task com-
prising English and Chinese words. The ratio of Chinese
words to English words was 1 to 6. Words appeared ran-
domly with an ITI of 5 s. Participants were instructed to
raise their hand when they heard the Chinese words and not
make any response when they heard the English words.
EEGs were recorded during meditation and the control cog-
nitive task. In view of the possible unstable state of partici-
pants, the EEG recording for the last 15 min was analyzed.
(iii) Post-test. After the meditation or control cognitive
task interval, participants were given the problems they had
failed in the pre-test session and were required to try to
solve these problems again. The procedure was the same as
in the pre-test session.
2 Results
2.1 Behavioral results
2.1.1 Successful problem-solving rate
Problem-solving rate included problem-solving rates in the
pre- and post-tests. It was defined as the number of suc-
cessfully solved items in the pre- or post-test divided by the
total number of items (total number of items in the post-test
was the 10 total items minus successfully solved items in
the pre-test). This step aimed to compare differences in
problem solving ability among these three groups. An
ANOVA showed that there was no significant difference
among the three groups in problem-solving rate for the
pre-test (P>0.05, Table 1), revealing that participants in the
Table 1 Problem-solving rate in the pre-testa)
Group Number of participants M SD SE
M10 16 0.49 0.15 0.04
M100 16 0.54 0.15 0.04
Ctrl 16 0.56 0.14 0.03
Total 48 0.53 0.15 0.02
a) The M10 group reported every 10 breaths, the M100 group reported
every 100 breaths, and Ctrl represents the control group.
three groups did not differ in their problem-solving ability.
Our study was mainly concerned with the problem-
solving rate in the post-test, which would suggest whether
the intervention (meditation) exerted an effect on problem
solving. Nonparametric statistics (Kruskal-Wallis H test)
were used because the data for problem-solving rate in the
post-test were abnormally distributed. The results showed a
significant difference among problem-solving rates in the
post-test (-square=11.414, df=2, P<0.005). Further analy-
sis with the Mann-Whitney U test showed that the prob-
lem-solving rate of the M10 group was higher than that of
the M100 group (P=0.051) and significantly higher than
that of the control group (P<0.005). The problem-solving
rate of the M100 group was higher than that of the control
group (P<0.05, Table 2). These results revealed that (i)
problem-solving ability of the two meditation groups was
better than that of the control group, and (ii) the perfor-
mance of the M10 group was better than that of the M100
group.
2.1.2 Response times (RTs)
First, we calculated the average response time (RT) for
problem solving (including successful and unsuccessful
tries) in the three conditions. Results showed that in the
pre-test, the RTs in the three conditions were not signifi-
cantly different and there was no interaction effect
(P=0.381). Using RTs in the pre-test as covariates in
the analysis, we found no significant difference or interac-
tion effect among the three groups at post-test (P=0.792,
Table 3).
Then, we compared RTs for successfully solved prob-
lems and abandoned (failed) problems among the three
conditions. Results showed that the three groups did not
differ in RTs and showed no interaction effect in the pre-test
for both kinds of problems (P=0.325 for solved problems,
P=0.830 for abandoned problems). With RTs for success-
fully solved problems or abandoned (failed) problems in the
pre-test as covariates, the analysis revealed no significant
difference or interaction effect at post-test (P=0.378 for
successfully solved problems, P=0.945 for abandoned
problems; Tables 4 and 5) among the three conditions.
2.2 EEG results
In this study, we examined the percentage of alpha waves in
Table 2 Problem-solving rate in the post-testa)
Group comparisons
M10-M100 M10-Ctrl
M100-Ctrl
Mann-Whitney U 77.50 50.00 70.50
Wilcoxon W 213.50 186.00 206.50
Z 1.92 3.01 2.23
a) The M10 group reported every 10 breaths, the M100 group reported
every 100 breaths, and Ctrl represents the control group.
964 Ren J, et al. Sci China Life Sci October (2011) Vol.54 No.10
Table 3 RTs in the pre-test and post-test
Group Number of participants M (ms) SD
Pre-test M10 16 71185.43 16873.03
M100 16 76823.79 17143.16
Ctrl 16 81040.91 24739.61
Total 48 76350.04 19915.76
Post-test M10 16 99166.95 34191.55
M100 16 101111.43 33481.85
Ctrl 16 96973.40 40199.79
Total 48 99083.93 35349.00
Table 4 RTs of successfully solved problems in the pre-test and post-testa)
Group Number of participants M (ms) SD
Pre-test M10 16 56369.91 18393.58
M100 16 61182.66 15213.48
Ctrl 16 67644.36 31260.17
Total 48 61732.31 22704.73
Post-test M10 15 78005.15 36924.43
M100 15 67062.57 41478.93
Ctrl 6 59237.92 38748.66
Total 36 70317.87 38716.58
a) Participants who did not solve any problems correctly were elimi-
nated from the post-test and were regarded as missing values.
Table 5 RTs of the abandoned (failed) problemsa)
Group Number of participants M (ms) SD
Pre-test M10 16 90526.28 24606.78
M100 13 94610.68 25597.84
Ctrl 13 96884.51 34869.44
Total 42 93758.52 27867.09
Post-test M10 9 102690.22 41354.98
M100 10 104566.24 44002.92
Ctrl 13 109053.26 43440.90
Total 32 105861.46 41733.89
a) Participants who solved all problems correctly were eliminated from
the pre- and post-tests and were regarded as missing values.
the three conditions. A higher percentage of alpha waves
have been associated with a more relaxed mental state. A
one-way within-groups ANOVA showed a significant dif-
ference in the percentage of alpha waves among the three
groups (F(2, 45)=4.144, P<0.05). Post-hoc testing showed
that the percentage of alpha waves in the M10 group was
significantly lower than in the control group (P<0.05) and
showed a decrease in tendency relative to that of the M100
group (P=0.098). There was no difference in the percentage
of alpha waves between the M10 group and control group
(Table 6). This indicated that the M10 group stayed more
alert during the meditation interval session. Product-
moment correlation analysis across all participants in the
three groups revealed that percentage of alpha waves was
negatively correlated with problem-solving rate in the
post-test (r=0.471, P=0.001).
We also analyzed i-35 and i-22, which are shown to be
related to states of attention and general relaxation, respec-
tively. Results showed there was a significant difference in
i-35 among the three groups (F(2, 45)=6.516, P<0.005).
Post-hoc testing revealed that the M10 group had higher
i-35 than the M100 group (P<0.05) and control group
(P=0.073), but no significant difference existed between the
M100 group and the control group. This again indicated that
the latter two groups were more relaxed than the M10 group.
As for i-22, the results showed differences among the three
groups (F(2, 45)=3.377, P<0.05). Post-hoc testing showed
that the M10 group showed a lower i-22 tendency than the
M100 group (P=0.084) and control group (P=0.093). How-
ever, there was no difference between the M100 group and
control group. This implied that participants in the M10
group were more mentally focused than the M100 and con-
trol group. The correlation analysis showed that i-35 was
not correlated with problem-solving rate in the post-test,
and i-22 was negatively correlated with problem-solving
rate in the post-test (r=.403, P<0.005). These results in-
dicated that a more alert and focused mind in the meditation
or control cognitive task interval may have been associated
with better problem-solving performance in the post-test.
3 Discussion
Our results showed that after 20 min of meditation or a con-
trol cognitive task, participants were able to solve some of
the problems they had failed in the pre-test. This is general-
ly in line with previous research on the role of incubation.
However, participants who engaged in meditation solved
more previously unsolved problems compared to partici-
pants in the control condition, thereby providing direct evi-
dence for the role of meditation in promoting insight.
The following considerations excluded, or partially ex-
cluded, the possibility that the observed differences between
M10, M100, and Ctrl conditions were caused by other un-
related factors. First, in the initial problem solving session
(pre-test), there were no significant differences in solution
rate and RTs among the three groups, thus excluding the
possibility that the observed key differences were caused by
differences in problem solving ability in the groups. Second,
in the final problem solving session (post-test) after medita-
tion or the control cognitive task, there was no detectable
difference in RTs for problem solving among the three
groups, thus excluding the possibility that the observed dif-
ferences were caused by participants in the M10 and M100
groups having learned to be more patient and insistent on
solving the problems after meditation. Further, we also cal-
culated the RTs for the successfully solved items and failed
items for the three groups. We found that in neither the ini-
tial session nor the final session was there any detectable dif-
ference among the groups. This excluded the possibility that
the better performance of the M10 and M100 groups was
related to the increase of participants’ meta-cognitive ability
in identifying the solvable and unsolvable items, thus ena-
Ren J, et al. Sci China Life Sci October (2011) Vol.54 No.10 965
bling them to wisely spend more time on the solvable items.
Rather, it is reasonable to propose that the increased in-
sight ability brought about by meditation may be related to
the fact that the participants had maintained a more alert and
mindful state of consciousness. First of all, participants in
M10, relative to M100, showed a stronger tendency for in-
sight. Although the difference was not significant, it ap-
proached significance (P=0.051). Secondly, there was a
significant negative correlation between insightful problem
solving in the final session and the percentage of alpha
waves known to be closely related to mental relaxation.
This implies that the more relaxed the participants were in
the meditation or control cognitive task interval, the less
insight they achieved later. Thirdly, the percentage of alpha
waves in M10 was not only significantly lower than in the
Ctrl condition, but also lower than that in M100 (but did not
reach significance, P=0.098). Taken together, these findings
imply that meditation promoted insight by keeping people
in an alert and mindful state. This appears to be different
from the possible mechanism underlying promotion of in-
sight during a passive type of relaxation, like sleep, that
does not require any mental effort.
This work was supported by the Education Program of the National Social
Science Foundation of China (Grant No. BBA100016), the National Basic
Research Program of China (Grant No. 2010CB833904), the National
Natural Science Foundation of China (Grant Nos. 30970890 and
30770708), and the National High Technology Research and Development
Program of China (Grant No. 2008AA022604).
1 Wagner U, Gais S, Haider H, et al. Sleep inspires insight. Nature,
2004, 427: 352–355
2 Tang Y Y, Lu Q, Geng X, et al. Short-term meditation induces white
matter changes in the anterior cingulate. Proc Natl Acad Sci USA,
2010, 107: 15649–15652
3 Tang Y Y, Ma Y, Fan Y, et al. Central and autonomic nervous sys-
tem interaction is altered by short-term meditation. Proc Natl Acad
Sci USA, 2009, 106: 8865–8870
4 Tang Y Y, Ma Y, Wang J, et al. Short-term meditation training im-
proves attention and self-regulation. Proc Natl Acad Sci USA, 2007,
104: 17152–17156
5 Slagter H A, Lutz A, Greischar L L, et al. Mental training affects dis-
tribution of limited brain resources. PLoS Biol, 2007, 5: e138
6 Lutz A, Slagter H A, Rawlings N B, et al. Mental training enhances
attentional stability: neural and behavioral evidence. J Neurosci, 2009,
29: 13418–13427
7 Lutz A, Slagter H A, Dunne J D, et al. Attention regulation and mon-
itoring in meditation. Trends Cogn Sci, 2008, 12: 163–169
8 Weisberg R W. Prolegomena to theories of insight in problem solv-
ing: A taxonomy of problems. In: Sternberg R J, Davidson J E, eds.
The Psychology of Problem Solving. Cambridge: Cambridge Univer-
sity Press, 1995. 157–196
9 Luo F, Luo J, Wu Y B, et al. Psychosomatic recovery by relaxing
and rebuilding the brain: Possible application in post traumatic psy-
chological recovery (in Chinese). Adv Psychol Sci, 2009, 17: 594–
601
10 Takahashi T, Murata T, Hamada T, et al. Changes in EEG and auto-
nomic nervous activity during meditation and their association with
personality traits. Int J Psychophysiol, 2005, 55: 199–207
11 Murata T, Takahashi T, Hamada T, et al. Individual trait anxiety lev-
els characterizing the properties of Zen meditation. Neuropsychobi-
ology, 2004, 50: 189–194
12 Kubota Y, Sato W, Toichi M, et al. Frontal midline theta rhythm is
correlated with cardiac autonomic activities during the performance
of an attention demanding meditation procedure. Brain Res Cogn
Brain Res, 2001, 11: 281–287
Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction
in any medium, provided the original author(s) and source are credited.
