Content uploaded by Kyongsik Yun
Author content
All content in this area was uploaded by Kyongsik Yun on Feb 02, 2018
Content may be subject to copyright.
1
Neural correlates of flow using auditory evoked potential suppression
Kyongsik Yun1,2,3*, Saeran Doh4*, Elisa Carrus5, Daw-An Wu1,2, Shinsuke Shimojo1,2,6
1Computation and Neural Systems, California Institute of Technology, Pasadena, CA 91125,
USA
2Division of Biology, California Institute of Technology, Pasadena, CA 91125, USA
3BBB Technologies Inc., Seoul, South Korea
4Department of Food Business Management, School of Food, Agricultural and Environmental
Sciences, Miyagi University, Miyagi, Japan
5Division of Psychology, School of Applied Sciences, London South Bank University,
London UK
6Japan Science and Technology Agency, Saitama, Japan
*These authors contributed equally to this work.
Correspondence should be addressed to K.Y. yunks@caltech.edu
2
Abstract
"Flow" is a hyper-engaged state of consciousness most commonly described in athletics,
popularly termed “being in the zone.” Quantitative research into flow has been hampered by
the disruptive nature of gathering subjective reports. Here we show that a passive probe
(suppression of Auditory Evoked Potential in EEG) that allowed our participants to remain
engaged in a first-person shooting game while we continually tracked the depth of their
immersion corresponded with the participants’ subjective experiences, and with their
objective performance levels. Comparing this time-varying record of flow against the overall
EEG record, we identified neural correlates of flow in the anterior cingulate cortex and the
temporal pole. These areas displayed increased beta band activity, mutual connectivity, and
feedback connectivity with primary motor cortex. These results corroborate the notion that
the flow state is an objective and quantifiable state of consciousness, which we identify and
characterize across subjective, behavioral and neural measures.
Keywords: flow; auditory evoked potential; EEG; anterior cingulate cortex; temporal pole
3
Introduction
“Flow” is a mental state of full immersion into an activity in which one is intensively
engaged, accompanied with a feeling of extreme concentration, full control, achievement and
pleasure in the activity 1. In popular culture, this state is often described in professional sports
as “being in the zone.” According to Csikszentmihalyi (1990), flow represents the perfect
experience of control of emotion and cognition during performance and learning. Verbal
protocols he gathered from experts skilled in various activities, mainly via interview, indicate
that the flow state can be experienced during online gaming as well as in various other
activities, such as sports, education, singing, dancing, climbing, and even surgery.
Flow research has received wide interest in the field of marketing, mainly through
the study of Computer Mediated Environments (CME), as a process of optimal experience
during internet use2. It has been found that flow can create compelling consumer online
experiences and can be fun and pleasurable; however, besides marketing research on flow,
very little is known with regards to quantifiable measures of this experience, such as the
neural dynamics underpinning this state.
Investigating the neural correlates of flow requires that the timecourse of neural data
be matched to information about the states of flow during that time period. One challenge to
gathering this information is that flow requires extended, uninterrupted engagement in an
activity. Thus, investigations generally use post-experiment questionnaires to gather data
that accompany flow experiences, this means that flow data lack timing information—they
are usually blanket ratings of the entire performance period.
Several previous studies have gathered neural data from participants experiencing
flow states while performing tasks 3,4, but they all possess this same limitation. Because they
do not track the dynamics of flow in a way that can be aligned to the neural data, they are
4
limited to contrasting data from different blocks of play, in which they have manipulated task
conditions to either increase or decrease the likelihood of entering the flow state. These
studies have provided many candidate regions and neural signatures for the flow state, but the
methodology cannot dissociate between internal flow states and the external conditions that
they set up to induce flow.
A study by Klasen et al. (2011) created a timecourse to align with the neural data by
recording the participants’ gameplay. They analyzed the game videos for events and content
that would support or inhibit flow, and aligned them with the fMRI data. By finding brain
areas that correlated with each content class and then calculating their overlap, they identified
a cerebellar-somatosensory cortex network as a likely substrate for flow. The use of fine-scale
timecourse in this study was a clear advantage, but here too the analysis actually correlated
the brain activity to external game conditions rather than to data about internal states. Thus it
is unclear whether this somato-motor brain network reflects the higher levels of action in the
content-defined gameplay epochs, or the (inferred) higher frequency of flow experience
during those epochs.
Here, we isolate the internal phenomenon of flow by correlating neural data directly
with timecourses of internal states. These timecourses are based on both subjective and
objective measures. Participants provided subjective data via a guided parametric version of
the retropective Think Aloud design 5. While they reviewed videos of their gameplay, they
rated their experience of flow in each time segment. Objective data regarding internal states
was obtained by a novel EEG probing method, which we validate and then apply here.
The EEG-based probe design is based on “telepresence” as a signature feature of the
flow experience during online gaming. This refers to the gameplayer’s tendency to feel like
they “exist” in the game world, which leads to a neglect of sensory stimulation from the real
5
world. The player becomes more sensitive to game-relevant sensory stimuli in the virtual
word, and less sensitive to game-irrelevant, real-world stimuli 6. Accordingly, we hypothesize
that when subjects are in the flow state, they will tend to neglect sounds that aren’t strictly
relevant to the game. Thus, we measure the AEPs (Auditory Evoked Potentials) elicited by
game-irrelevant, random beeping sounds7 inserted throughout gameplay. Attention-based
modulation of AEP amplitude is an extremely robust, and well-established effect in the
evoked-related potential literature8. For example, a previous study showed that the AEPs
elicited by a faint clicking sound were prominent when subjects counted them, but then
became suppressed when subjects read a book intently9. Thus, we expect the AEP to be
suppressed during immersion in the flow state. A recent study applied an auditory oddball as
a secondary task to objectively quantify flow while playing video games 10. Our AEP
suppression has advantages in that it is completely passive and does not explicitly interfere
with the subjects’ cognitive processes during the flow experience.
It must be noted, however, that the AEP suppression alone is not sufficient to define
the flow state during game play, because it directly captures only some of the aspects of flow,
such as intense, focused concentration, and perhaps telepresence indirectly. Subjective
behavioral reporting on the experience of flow is also neither sufficient nor reliable in
defining flow. It was for this reason that we decided to define the flow state by a correlation
of objective and subjective measures, assessing flow state by comparing the AEP amplitude
suppression with the respective post-gameplay ratings of flow for each participant.
The present study aims: 1) to validate AEP suppression as a measure of the flow state,
thereby providing a passive probe that can be used to periodically measure the participant’s
flow state without disrupting their engagement in the task. 2) to use the measured timecourse
of flow to localize its neural correlates, and to characterize functional connectivity between
6
those brain regions. By doing this, we would like to characterize the dynamics of flow in a
quantitative manner, and link it to objective neural substrates.
7
Methods
Participants
Twenty-nine healthy subjects participated in this study (24 males, 5 females, age: 23.5±3.4
years (mean±SD)). We recruited subjects on the basis of previous experience with video
games in general (average 13.6±10.1 hours/week, range 7~49 hours/week) with ads posted on
the main campus of the California Institute of Technology. All participants provided written
informed consent after receiving a detailed explanation of the experimental procedures. The
Institutional Review Board of California Institute of Technology approved all experimental
procedures and this study was carried out in accordance with the approved guidelines.
Participants were excluded if they had a history of neurological disorder such as seizure,
stroke, head injury, or a substance use disorder other than caffeine or nicotine. We obtained a
set of self-reported flow questionnaires 11 at the end of the experiment.
Task
The participants played a FPS video game (Call of Duty : Modern Warfare 2, Activision).
During pre-experiment game training, we adjusted the difficulty of the game depending on
the skill of each participant so that it was right above the player’s skills. During the
experimental session, participants played the game for an hour, consisting of a 30min low
challenge and a 30min high challenge game scenario (Figure 1). During the game play,
randomly distributed beep sounds (inter-beep interval=1~120sec: duration=200ms; 40dB)
were presented via speakers to the participants, in order to measure auditory evoked
potentials.
8
Behavioral analysis
While participants were playing, we recorded the game play using Fraps, a realtime video
capture software (www.fraps.com). This recording was used in the review session following
game play. Here, participants were asked to review their own game play by replaying a video
of their own game session and responding whether they experienced flow or not for each 5
min time period of the video. The 5 min time period was chosen based on the previous
studies of neurofeedback and subjective behavioral ratings 12,13. Player performance (i.e.,
number of kills – deaths) was tracked across the same time bins of game play. Flow
experience and performance distribution histograms were compared using Pearson’s
correlation. For additional skill-based analysis, fifteen highest performing participants were
grouped as “experts” and the others as “beginners.”
Insert figure 1.
Electrophysiological recording and analysis
We recorded EEG activity from 128 scalp electrodes (EGI System 200; Electrical Geodesics,
Eugene, OR.) during gameplay. Electrode impedance was kept under 40 kΩ for all
recordings14. Electrode nets were covered with a shower cap to prevent electrodes from
drying so that we could maintain low electrode impedance. Vertical and horizontal ocular
movements were also recorded. The EEG was continuously recorded at a 500 Hz sampling
frequency and filtered (high pass 0.1Hz, low pass 200Hz, notch filter 60Hz and 120Hz). The
EEGs were segmented from −500ms to 1000ms relative to the onset of each beep. Ocular
artifact reduction was performed using ICA component rejection in EEGLAB 15.
9
Following pre-processing, we computed the evoked activity aligned to each task-
irrelevant beep sound, and this was done to evaluate epochs with and without AEP
suppression. Consequently, we separated each epoch as flow and non-flow groups based on
the participants’ ratings as well as AEP suppression (Table 1) and computed each epoch's
source localization and partial directed coherence values to define the network of regions
involved and effective connectivity within it. The time window used for source localization
and partial directed coherence was 1000ms following the beep.
Time-frequency analysis
The epoched data were analyzed by means of a event-related spectral perturbation (ERSP) 16
(window length, 250 ms; step, 25 ms; window overlap, 90%). This was done to select the
epochs that showed AEP suppression. We removed the baseline of 500 ms preceding the
beep onset with duration of 500 ms from time frequency ERSP charts. We used a
nonparametric permutation test with 5000 randomizations for comparisons between the
activation and baseline, corrected for multiple comparisons 17.The corrected threshold was set
to p < 0.05, and the time-frequency representation only shows statistically significant results.
sLORETA
Standardized low resolution brain electromagnetic tomography (sLORETA) 18 was used for
source localization. Thousands of synchronized postsynaptic potentials from pyramidal
neurons of the cortex produce scalp EEG activity 19. sLORETA computes the three
dimensional localization of these activities. The subject-specific 3D coordinates of the 128
electrode positions were estimated (Geodesic Photogrammetry System; Electrical Geodesics,
Eugene, OR.) and applied to a digitized MRI version of the Talairach Atlas (McConnell Brain
10
Imaging Centre, Montréal Neurological Institute, McGill University). These Talairach
coordinates were then used to compute the sLORETA transformation matrix for each
participant. Following the transformation to an average reference, the EEG activity of the
flow and non-flow epochs selected from the above time frequency analysis was used to
calculate cross spectra in sLORETA for each participant. We used delta (1~4Hz), theta
(4~8Hz), alpha (8~12Hz), and beta (12~30Hz) frequency bands for the following analyses.
Using the sLORETA transformation matrix, the cross spectra of each participant and
frequency band were then transformed into sLORETA files. These files included the 3D
cortical distribution of the electrical neuronal generators for each participant. The computed
sLORETA image displayed the cortical neuronal oscillators in 6239 voxels, with a spatial
resolution of 5 mm 20. We used a nonparametric permutation test with 5000 randomizations
for comparisons between the flow and non-flow17. The threshold was set to p < 0.01.
Partial directed coherence
Partial directed coherence (PDC) is a form of frequency-domain Granger-causality, which
quantifies the direction of information transfer between brain regions 21,22. The EEG data
from ROIs defined by source localization were windowed in 500-sample long intervals (i.e.,
1000ms in length). The PDC values were evaluated in the delta (1~4Hz), theta (4~8Hz),
alpha (8~12Hz), and beta (12~30Hz) ranges. Model order (i.e., time delay of the
autoregressive parameters) was set to 15, based on previous studies for sufficient frequency
resolution 23,24. We used the nonparametric bootstrap approach for further statistical analyses
25,26. For details of this method, see Snijders and Borgatti 26. The threshold was set to p <
0.001, considering Bonferroni correction for multiple comparisions (3 ROIs, 3X3=9
comparisons). Each arrow pointing from the region i to its target region j represents a
11
significant PDC. SPSS (Windows version 15.0; SPSS, Inc., Chicago, IL) was used for the
statistical analyses.
12
Results
Mean duration of flow was 8.31±3.61 min (13.9% of total time). 53.0% of the flow
experience occurred between the 25 and 45 minute marks of the total one-hour game session.
A significant positive correlation was found between the overall occurrence of the flow
experience and the performance distribution throughout the game play (Pearson’s correlation,
R=0.68, p=0.014) (Figure 2). These results indicate the close relationship between subjective
flow experience and performance, suggesting that the higher the experience of flow, the
better the gaming performance.
Insert Figure 2.
Next we tested AEPs as a neural marker for the flow state. Since we did not observe
the conventional clear AEP waveform, due to insufficient number of trials and background
noise from the game play, we analysed the ERSPs of the signal at low frequencies (1-30Hz).
In EEG analysis, the classical ERP model is limited to study complex brain dynamics because
the trial-by-trial frequency components tend to be averaged out by this method. ERSP has
been adapted for this reason27-30. Given the low signal-to-noise ratio, we used ERSPs, which
represent the power of oscillations of the ERP. This was done because evoked potentials
predominantly contain low-frequency information, and by doing this we would still be able to
evaluate the suppression of the auditory evoked potential in the absence of a clear AEP.
Following established protocols, ERSPs were calculated for each tone at the averaged
channels Fz and Pz8,31. We compared the EEG ERSP of the post-stimulus window to the
baseline (500~0ms before the beep onset) and separated epochs with significant post-stimulus
activation from those showing deactivation at averaged. Out of the epochs showing
13
suppressed signal, we categorised them as flow segments only if they were also rated as flow
by participants. The self-reported non-flow trials showed larger evoked potential activation
than the flow trials (t(1861)=3.84, p=0.0001) (Figure 3 and Figure S1). We also found a
significant correlation between suppressed evoked potential and self-reported experience of
flow (Table S1. Chi-square test; Χ2(1)=97.0, p<0.001).
Insert Figure 3.
Following the analysis of the auditory evoked potential, we source localized the
epochs corresponding to the flow state and those corresponding to the non-flow state at time
window of 0~1000ms after the stimulus. This was done to estimate the possible sources of
the activity that gives rise to flow experience, and therefore to find the brain areas associated
with flow. Figure 4 depicts localization results, which showed that the anterior cingulate
cortex (ACC: MNI x=-6, y=25, z=21, Figure 4a); and temporal pole (TP: x=-55, y=10, z=-25,
Figure 4b) were significantly activated in the flow state as compared to the non-flow state
only in the beta frequency range. We also found significant beta frequency oscillations in the
primary motor cortex (precentral gyrus (M1); x=15, y=-20, z=70; Figure 4c) in the beginners
compared to experts.
Insert Figure 4.
To examine the link between these three brain regions and the experience of flow, we
computed correlations between the subjective behavioral flow ratings and beta activity . We
found significant positive correlations between the flow ratings and activation in ACC
14
(Pearson’s correlation; R2=0.22, P=0.017) and TP (R2=0.26, P=0.009) (Figure 5A and B) and
a significant negative correlation with activation in M1 (R2=0.32, P=0.003) (Figure 5C).
These correlation results therefore suggest that ACC and TP are the core areas associated with
the flow state.
Insert Figure 5.
To further understand the network properties of the neural correlates of flow, we analyzed
effective connectivity using PDC across the same three regions of interest. We found that the
PDC connectivity increased during the flow compared to non-flow state. More specifically,
the top-down connectivity increased (ACC->M1, TP->M1) and bottom-up connectivity
decreased (M1->TP) during the flow state (nonparametric jackknife procedure, p < 0.001)
compared to non-flow state (Figure 6).
Insert Figure 6.
15
Discussion
The current study aimed to generate the mental state of “flow” in the laboratory and to
quantify and to characterize it using a novel combination of objective and subjective data,
namely the suppression of AEP, and video-assisted self-reports. By correlating the timecourse
of flow with the EEG data, we localized the potential neural correlates of flow and
characterized effective connectivity between these brain areas. Specifically, the ACC and TP
emerged as important hubs associated with the experience of flow, showing increased beta
frequency power and increased effective connectivity patterns during the flow state compared
to non-flow.
There are several psychological criteria for flow 32, including (1) intense and focused
concentration, (2) a loss of self-consciousness, and (3) a feeling of control and ease of
performance. For example, when a game player experiences flow, his/her attention is
completely focused on the game character and the opponent, and he/she ignores any task-
irrelevant stimuli. The actions he/she performs are perceived as if they were his/her own,
rather than in the virtual world. We therefore exploited this notion by evaluating modulation
of the AEP amplitude in flow and non-flow. We found significantly suppressed neural activity
following the beep when in flow, compared to non-flow in the oscillatory evoked activity,
which represents the power of oscillations of the ERP. This result suggests that when
participants experienced flow, their brains shut off task-irrelevant stimuli in the literal sense
(even at the sensory level). Through the use of the AEP, we were able to quantify the flow
state without disrupting the participant’s engagement with the game.
Our source localization results are consistent with the flow characteristics mentioned
above. Activation of the ACC known to be involved in the processing attention and focus 33,34,
which is in line with the first criterion of flow. The TP, an empathy-related brain region, was
16
also activated and we interpret this in light of the second characteristic of flow, a loss of self-
consciousness and telepresence. The TP has been known to be crucial in the process of
distinguishing self vs. other as well as in perspective-taking in social affective processes35,36.
Lastly, our current results show a pattern of negative correlation between motor activity and
the flow experience. We also found that the motor activity decreased in experts compared to
beginners. These results are consistent with previous studies, which have found that motor
activity decreases when motor behavior is processed more efficiently37,38, and that experts
develop a focused and efficient organization of neural networks39. This accords with the third
characteristic of flow, which is a feeling of control and ease of performance.
Our effective connectivity results showed a top-down connectivity from the ACC and
the TP to M1 in the flow state. Enhanced top-down processing has been known to be
consistent with heightened performance and concentration40,41. However, our top-down
connectivity results are different from the typical top-down executive control represented by
explicit (conscious) cognitive process of the prefrontal cortex. Rather, ACC and TP
performed as implicit (unconscious) cognitive and social processes, including extreme focus,
empathy, and telepresence. In summary, we argue that the ACC and the TP are the two main
regions that may provide the neural basis of the flow state, or “being in the zone”.
Our study took several measures to maximize the likelihood of inducing a deep flow
state. First, we used a first-person-shooter video game known to be very addictive42, and our
participants were all experienced players, many of whom routinely devoted large amounts of
time to gameplay. This allowed us to recreate a state of flow in the laboratory in players who
had spontaneously experienced flow in the past. The laboratory setting was comfortable, and
provided a typical computer game playing environment. This contrasts with an MRI
environment, where players are prone, immobilized and in a confined space. Flow is less
17
likely to occur if the environment is unfamiliar and/or uncomfortable43. Most importantly, we
allowed the participants to play the game continuously for an hour. According to our results,
87.5% of the participants require at least 25min to get into the flow state (Figure 2). This
suggests that the relative strength of flow experiences in previous neuroscientific studies may
be limited, as individual blocks of their experiments range from 3-12 minutes.
Lastly, we implemented an objective electrophysiological evaluation (AEP
suppression) to test whether participants were in flow and cross-validated these results
through subjective behavioral flow ratings. This technique allowed us to track the strength of
flow and build a timeline of the flow experience. Previous studies have been limited in their
characterization of flow: they tended to quantify objective external features such as game
performance (killing vs. being killed, active fighting vs. boring scenes in the case of shooting
games; easy vs. hard levels in general). Although there is a positive correlation between such
external criteria and the flow experience, using them as substitutes for direct measures of
flow causes clear confounds in the analysis of neural correlates. We argue that our cross-
validation method (across neural, behavioral and subjective measures) captures the actual
emergence of flow in the closest possible way to the original definition and the
phenomenology of the state.
One limitation of our study is related to our use of EEG, in that it cannot capture the
activity of deep brain structures, such as amygdala and midbrain reward circuits. These
regions have been known to be closely related with empathy, attention, motivation, and
reward, which are highly probable to be correlated with flow experience44. Future studies
using different techniques are necessary to pursue such targets. For example a combination of
EEG and fMRI could be used to simultaneously measure the AEPs and image deep brain
areas.
18
We should also mention the limitation that we did not resolve a clear AEP waveform,
probably because our experimental setting was ill-suited to extracting classical ERPs through
trial averaging. Rather than having hundreds of regularly repeated trials, we had a relatively
small number of beeps randomly inserted into the dynamic and irregular auditory game
environment. However, anlaysis in the time-frequency domain revealed clear patterns of
evoked oscillatory activity, showing large positive components of the AEP phase-locked to
the beep onset and suppression of those components during the flow state.
Techniques of detection and quantification of the flow state can contribute not only
to neuroscience of the altered state of consciousness, but also to game content design. The
behavioral characteristics of the flow experience, including its mean duration, temporal
location, and the relationship with gaming performance, would be very useful for effective,
immersive game design as well as marketing research on internet shopping, etc. Degrees of
the feeling of being in the zone have been known to be highly associated with amount of
information search in game players and purchasing intention among internet shoppers45. Thus,
whether our technique can be generalized to other situations of flow, particularly in the
domain of marketing, would be an intriguing question to address in future. Thus all together,
the flow is a psychological and neurophysiological phenomenon which can be uniquely
defined, and feasibly assessed by a combination of subjective and objective techniques as
demonstrated in the current study. Whether or not the neural correlates and the neural
measure are applicable to the flow state in other kinds of context, such as sports, gambling,
singing, dancing, internet surfing/shopping, etc. would be an obvious next step of future
research.
19
References
1 Csikszentmihalyi, M. The domain of creativity. Theories of creativity 4, 61-91 (1990).
2 Hoffman, D. L. & Novak, T. P. Marketing in hypermedia computer-mediated
environments: conceptual foundations. The Journal of Marketing, 50-68 (1996).
3 Plotnikov, A. et al. in Advanced Learning Technologies (ICALT), 2012 IEEE 12th
International Conference on. 688-689 (IEEE).
4 Berta, R., Bellotti, F., De Gloria, A., Pranantha, D. & Schatten, C.
Electroencephalogram and physiological signal analysis for assessing flow in games.
Computational Intelligence and AI in Games, IEEE Transactions on 5, 164-175
(2013).
5 Van Someren, M. W., Barnard, Y. F. & Sandberg, J. A. The think aloud method: A
practical guide to modelling cognitive processes. Vol. 2 (Academic Press London,
1994).
6 Cowley, B., Charles, D., Black, M. & Hickey, R. Toward an understanding of flow in
video games. Computers in Entertainment (CIE) 6, 20 (2008).
7 Allison, B. Z. & Polich, J. Workload assessment of computer gaming using a single-
stimulus event-related potential paradigm. Biol. Psychol. 77, 277-283 (2008).
8 Picton, T. & Hillyard, S. Human auditory evoked potentials. II: Effects of attention.
Electroencephalogr. Clin. Neurophysiol. 36, 191-200 (1974).
9 Picton, T., Hillyard, S., Krausz, H. & Galambos, R. Human auditory evoked potentials:
I. Evaluation of components. Electroencephalography & Clinical Neurophysiology;
Electroencephalography & Clinical Neurophysiology (1974).
10 Nuñez Castellar, E. P., Antons, J.-N., Marinazzo, D. & van Looy, J. Being in the zone:
20
Using behavioral and EEG recordings for the indirect assessment of flow. PeerJ
Preprints 4, e2482v2481 (2016).
11 Csikszentmihalyi, M. Toward a psychology of optimal experience. Review of
personality and social psychology 2, 13-36 (1982).
12 Prinsloo, G. E., Derman, W. E., Lambert, M. I. & Rauch, H. L. The effect of a single
session of short duration biofeedback-induced deep breathing on measures of heart
rate variability during laboratory-induced cognitive stress: A pilot study. Appl.
Psychophysiol. Biofeedback 38, 81-90 (2013).
13 Potteiger, J. A., Schroeder, J. M. & Goff, K. L. Influence of music on ratings of
perceived exertion during 20 minutes of moderate intensity exercise. Percept. Mot.
Skills 91, 848-854 (2000).
14 Ferree, T. C., Luu, P., Russell, G. S. & Tucker, D. M. Scalp electrode impedance,
infection risk, and EEG data quality. Clin. Neurophysiol. 112, 536-544 (2001).
15 Delorme, A. & Makeig, S. EEGLAB: an open source toolbox for analysis of single-
trial EEG dynamics including independent component analysis. J. Neurosci. Methods
134, 9-21 (2004).
16 Pfurtscheller, G. & Lopes da Silva, F. H. Event-related EEG/MEG synchronization
and desynchronization: basic principles. Clin. Neurophysiol. 110, 1842-1857 (1999).
17 Nichols, T. E. & Holmes, A. P. Nonparametric permutation tests for functional
neuroimaging: a primer with examples. Hum. Brain Mapp. 15, 1-25 (2002).
18 Pascual-Marqui, R. D., Esslen, M., Kochi, K. & Lehmann, D. Functional imaging
with low-resolution brain electromagnetic tomography (LORETA): a review. Methods
Find. Exp. Clin. Pharmacol. 24, 91-95 (2002).
19 Martin, J. H. The collective electrical behavior of cortical neurons: the
21
electroencephalogram and the mechanisms of epilepsy. Principles of Neural Science.
4th ed. New York: McGraw-Hill, Health Professions Division, 777-791 (2000).
20 Pascual-Marqui, R. D. Standardized low resolution brain electromagnetic tomography
(sLORETA): technical details. Methods Find. Exp. Clin. Pharmacol. 24, 5-12 (2002).
21 Baccal, L. A. & Sameshima, K. Partial directed coherence: a new concept in neural
structure determination. Biol. Cybern. 84, 463-474 (2001).
22 Baccal, L. A. & Sameshima, K. Comments on 'Is partial coherence a viable technique
for identifying generators of neural oscillations?'. Biol. Cybern. 95, 135-141 (2006).
23 Brovelli, A. et al. Beta oscillations in a large-scale sensorimotor cortical network:
Directional influences revealed by Granger causality. Proceedings of the National
Academy of Sciences 101, 9849-9854 (2004).
24 Supp, G. G., Schlogl, A., Trujillo-Barreto, N., Muller, M. M. & Gruber, T. Directed
cortical information flow during human object recognition: analyzing induced EEG
gamma-band responses in brain's source space. PLoS One 2 (2007).
25 Efron, B. Nonparametric estimates of standard error: the jackknife, the bootstrap and
other methods. Biometrika 68, 589-599 (1981).
26 Snijders, T. A. B. & Borgatti, S. P. Non-parametric standard errors and tests for
network statistics. Connections 22, 61-70 (1999).
27 Makeig, S. Auditory event-related dynamics of the EEG spectrum and effects of
exposure to tones. Electroencephalogr. Clin. Neurophysiol. 86, 283-293 (1993).
28 Grandchamp, R. & Delorme, A. Single-trial normalization for event-related spectral
decomposition reduces sensitivity to noisy trials. Front Psychol 2, 236 (2011).
29 Rossi, A., Parada, F. J., Kolchinsky, A. & Puce, A. Neural correlates of apparent
motion perception of impoverished facial stimuli: a comparison of ERP and ERSP
22
activity. Neuroimage 98, 442-459 (2014).
30 Engell, A. D., Huettel, S. & McCarthy, G. The fMRI BOLD signal tracks
electrophysiological spectral perturbations, not event-related potentials. Neuroimage
59, 2600-2606 (2012).
31 Hegerl, U. & Frodl-Bauch, T. Dipole source analysis of P300 component of the
auditory evoked potential: a methodological advance? Psychiatry Research:
Neuroimaging 74, 109-118 (1997).
32 Nakamura, J. & Csikszentmihalyi, M. Flow theory and research. Oxford handbook of
positive psychology, 195-206 (2009).
33 Cohen, R., Kaplan, R., Moser, D., Jenkins, M. & Wilkinson, H. Impairments of
attention after cingulotomy. Neurology 53, 819-819 (1999).
34 Lane, R. D. et al. Neural correlates of levels of emotional awareness: evidence of an
interaction between emotion and attention in the anterior cingulate cortex. J. Cogn.
Neurosci. 10, 525-535 (1998).
35 Ruby, P. & Decety, J. How would you feel versus how do you think she would feel? A
neuroimaging study of perspective-taking with social emotions. J. Cogn. Neurosci. 16,
988-999 (2004).
36 Yun, K., Watanabe, K. & Shimojo, S. Interpersonal body and neural synchronization
as a marker of implicit social interaction. Scientific Reports 2, 959,
doi:doi:10.1038/srep00959 (2012).
37 Mazaheri, A., Nieuwenhuis, I. L. C., van Dijk, H. & Jensen, O. Prestimulus alpha and
mu activity predicts failure to inhibit motor responses. Hum. Brain Mapp. 30, 1791-
1800 (2009).
38 Henson, R. & Rugg, M. Neural response suppression, haemodynamic repetition
23
effects, and behavioural priming. Neuropsychologia 41, 263-270 (2003).
39 Milton, J., Solodkin, A., Hluštík, P. & Small, S. L. The mind of expert motor
performance is cool and focused. Neuroimage 35, 804-813 (2007).
40 Engel, A. K., Fries, P. & Singer, W. Dynamic predictions: oscillations and synchrony
in top-down processing. Nature Reviews Neuroscience 2, 704-716 (2001).
41 Pessoa, L., Kastner, S. & Ungerleider, L. G. Neuroimaging studies of attention: from
modulation of sensory processing to top-down control. The Journal of neuroscience
23, 3990-3998 (2003).
42 Montag, C. et al. Internet addiction and personality in first-person-shooter video
gamers. Journal of Media Psychology: Theories, Methods, and Applications 23, 163
(2011).
43 Csikszentmihalyi, M. Finding flow: The psychology of engagement with everyday life.
(Basic Books, 1998).
44 Klasen, M., Weber, R., Kircher, T. T. J., Mathiak, K. A. & Mathiak, K. Neural
contributions to flow experience during video game playing. Social cognitive and
affective neuroscience 7, 485-495 (2012).
45 Doh, S. Flow Construct: Its Mediating Roles in an On-line Search Model. 11th
International Marketing Trends Conference proceedings (2012).
24
Acknowledgments
This work was supported by JST-ERATO, JST-CREST, Tamagawa-Caltech gCOE programs,
Grant-in-Aid for Scientific Research (Kakenhi) in Japan, and Basic Science Research
Program through the National Research Foundation of Korea (NRF) funded by the Ministry
of Education, Science and Technology (2013R1A6A3A03020772).
Author Contributions
K.Y., S.D., E.C., and S.S. designed and performed the experiments. K.Y. analyzed the data.
K.Y., S.D., E.C., and S.S. wrote the manuscript.
Additional information
Competing financial interests: The authors declare no competing financial interests.
25
Figure Legends
Figure 1. Experimental procedure.
Participants played a first-person shooter video game for an hour, consisting of 30min low
challenge and 30min high challenge. Randomly distributed beep sounds (inter-beep
interval=1~120sec) were presented to participants via speakers during the game.
Figure 2. Histograms of (A) the flow distribution and (B) the performance distribution across
time of game play. Performance was defined as the mean number of “kills” minus “deaths”
across all participants (R=0.68, p=0.014).
Figure 3. Time-frequency analyses of auditory evoked activity during (A) flow and (B) non-
flow states. The non-flow epochs showed significant evoked potential activation and the flow
trials showed significant evoked potential suppression compared to the baseline (500~0ms
before the beep onset) (average of channels Fz and Pz; nonparametric permutation test,
p<0.05). The vertical line represents the onset of the beep.
Figure 4. (A) Source localization showing the contrast of flow state minus non-flow state for
(A) the anterior cingulate cortex (ACC) (x=-6, y=25, z=21), and (B) temporal pole (TP) (x=-
55, y=10, z=-25). (C) the flow minus non-flow contrast for beginners minus experts in the
precentral gyrus (M1) (x=15, y=-20, z=70) (all: non-parametric permutation test, red: p<0.05,
yellow: p<0.01).
Figure 5. Correlation between the behavioral flow ratings and source localized EEG activity.
For the (A) temporal pole (TP) (R2=0.26, P=0.009), (B) anterior cingulate cortex (ACC)
26
(R2=0.22, P=0.017), and (C) primary motor cortex (M1) (R2=0.32, P=0.003).
Figure 6. Effective connectivity using PDC between regions of interest, including anterior
cingulate cortex (ACC), temporal pole (TP), and primary motor (M1) cortex. These regions
were selected based on a priori source localized activity in Figure 4. Red arrows indicate the
contrast for PDC connectivity of flow with non-flow, and the blue arrows indicate the
contrast of non-flow with flow.
27
Figure 1
28
Figure 2
29
Figure 3
30
Figure 4
31
Figure 5
32
Figure 6
Supplementary Information
Neural correlates of flow using auditory evoked potential suppression
Kyongsik Yun1,2,3*, Saeran Doh4*, Elisa Carrus5, Daw-An Wu1,2, Shinsuke Shimojo1,2,6
1Computation and Neural Systems, California Institute of Technology, Pasadena, CA 91125, USA
2Division of Biology, California Institute of Technology, Pasadena, CA 91125, USA
3BBB Technologies Inc., Seoul, South Korea
4Department of Food Business Management, School of Food, Agricultural and Environmental Sciences, Miyagi
University, Miyagi, Japan
5Division of Psychology, School of Applied Sciences, London South Bank University, London UK
6Japan Science and Technology Agency, Saitama, Japan
*These authors contributed equally to this work.
Correspondence should be addressed to K.Y. yunks@caltech.edu
S1 Table. Percentage of epochs with and without suppressed AEP rated as flow and non-flow
AEP suppression
No AEP suppression
Flow rating
26.1%
18.7%
Non-flow rating
14.6%
40.6%
Epochs rated as flow showed significantly larger number of suppressed AEP compared with
the AEP of epochs rated as non-flow (chi-square test; Χ2(1)=97.0, p<0.001).
Table S1. Venn diagram of the percentage distribution of all the trials
S1 Figure. Peak auditory evoked potential (0~500ms) in non-flow and flow state. The non-
flow trials showed larger evoked potential activation than the flow trials (t(1861)=3.84,
p=0.0001).
