Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
and Assessment in
Serious Games
Thomas M. Connolly
University of the West of Scotland, UK
Thomas Hainey
University of the West of Scotland, UK
Elizabeth Boyle
University of the West of Scotland, UK
Gavin Baxter
University of the West of Scotland, UK
Pablo Moreno-Ger
Universidad Complutense de Madrid, Spain
Copyright 2014 by IGI Global. All rights reserved. No part of this publication may be reproduced, stored or distributed in
any form or by any means, electronic or mechanical, including photocopying, without written permission from the publisher.
Product or company names used in this set are for identification purposes only. Inclusion of the names of the products or
companies does not indicate a claim of ownership by IGI Global of the trademark or registered trademark.
This book is published in the IGI Global book series Advances in Game-Based Learning (AGBL) (ISSN: 2327-1825;
eISSN: 2327-1833)
All work contributed to this book is new, previously-unpublished material. The views expressed in this book are those of the
authors, but not necessarily of the publisher.
Chapter 5
Neurofeedback and
Serious Games
Manuel Ninaus Jrgen Kurzmann
University of Graz, Austria University of Graz, Austria
ABSTRACT
Neuroscience as well as computer gaming have rapidly advanced in the last decades. Yet, the combina-
tion of both fields is still in its infancy. One example of an emerging alliance is neurofeedback, where
participants are required to learn controlling their own brain activity. So far, this kind of training is
mostly applied in therapeutic settings, for example improving symptoms in epilepsy, attention-deficit/
hyperactivity disorder, or autism spectrum disorder. However, there are some promising approaches
that used neurofeedback in everyday situations for healthy subjects. This may prove especially valuable
for serious games that aim to improve learning capabilities and cognitive aspects of individual users.
The following chapter introduces the basic concepts and standards of neurofeedback. The different
non-invasive imaging techniques are introduced along with successful applications in neurofeedback.
Finally, benefits and pitfalls for future combinations of neurofeedback and games are discussed: while
the former may profit from realistic and motivating video scenarios, the latter is expected to be a tool
for evaluating and monitoring the direct effects on the users brain.
DOI: 10.4018/978-1-4666-4773-2.ch005
Copyright 2014, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.
Neurofeedback and Serious Games
83
Neurofeedback and Serious Games
NF is often confused with brain computer of slow EEG frequencies (4-7 Hz) and a lack of
interface (BCI). BCI provides an independent faster EEG frequencies (15-20 Hz) compared to
non-muscular channel for communication and normal kids. Therefore, ADHD kids are trained
control by means of translating signals of the to enhance fast EEG frequencies combined with
brain into an output that reflects the users intent inhibiting slow frequencies. This EEG based NF
(Wolpaw, Birbaumer, McFarland, Pfurtscheller, training shows positive effects on behavior, such
& Vaughan, 2002). While in both, NF and BCI as a reduction of negative hyperactive/impulsive
the users brain signals are directly displayed to behaviors with simultaneous improvement of at-
the participant in real-time, these applications tentional capabilities (Gruzelier & Egner, 2005).
differ conceptually. BCI can be seen as a direct Nowadays, NF approach is used to regulate a vari-
communication pathway between a human brain ety of brain dysfunctions and psychiatric disorders,
and external devices such as a wheelchair, a such as schizophrenia, addiction, depression, or
computer, or prosthesis (Kropotov, 2009). In obsessive-compulsive disorder (Kropotov, 2009).
BCI applications, the recorded brain signals are Generally, when participants become successful
translated in real-time into commands that oper- in regulating their own brain activity, e.g. increas-
ate a computer display or other devices (Wolpaw ing voluntary specific EEG frequency bands,
et al., 2002). In contrast, in NF applications, the improvements in cognition and behavior usually
feedback parameters are used for self-regulation follow (Coben & Evans, 2010). However, NF is
of the brain itself (Kropotov, 2009). Participants not only useful for people with deficits in specific
learn to control their own brain activity by means cognitive domains. In specific situations healthy
of contingent feedback of measures of the brain people can benefit from NF as well. Recent studies
activity during NF experiments (Weiskopf, 2012). were able to show that EEG based NF is useful to
In recent studies, both, NF and BCI approaches improve memory performance in healthy young
have been used to control games without muscular students (e.g. Nan et al., 2012). This confirms
activity (e.g. Wang et al., 2011; Yan et al., 2008; findings of prior studies indicating that EEG
Zhao et al., 2009). NF training leads to improvements in cognitive
Different neuroimaging techniques have been performances in healthy adults (Angelakis et al.,
used in NF research and a detailed review of the 2007; Gruzelier, Egner, & Vernon, 2006; Hoedl-
major features will be given later in this chapter. moser et al., 2008; Vernon, 2005; Vernon et al.,
For now, we want to introduce some of the main 2003). The majority of NF studies are based on
applications. Historically, electroencephalography EEG signals, but the derived EEG parameters are
(EEG) based NF was first applied in clinical prac- not the only physiological parameters that reflect
tice for treatment of epileptic patients. Epilepsy functioning of the brain. Recently, the number of
patients who learned to control their electrical NF studies using hemodynamic/metabolic activ-
brain activity using EEG based NF training showed ity of the brain measured by functional magnetic
a reduced seizure rate (Kotchoubey et al., 1999). resonance imaging (fMRI) or near-infrared spec-
Later, NF was mainly applied to reduce attention troscopy (NIRS) as NF signal is rising (Yoo et
deficit-hyperactivity disorder (ADHD) symptoms al., 2008). The fMRI and NIRS signal can be fed
in children, such as inattention and overactivity back to the participants and, consequently, can
(Gruzelier & Egner, 2005). There is ample evi- be used for voluntary control of brain functions
dence that children with ADHD show an excess (Kropotov, 2009).
84
Neurofeedback and Serious Games
85
Neurofeedback and Serious Games
86
Neurofeedback and Serious Games
87
Neurofeedback and Serious Games
thus constantly adjust the level of difficulty to bal- cessful learning. This is often independent from
ance challenging and rewarding aspects for each conscious perception so that instead of verbal
individual user. Another limitation is that a clear reports by users one can rely on more objective
connection between NF training and enhanced patterns of brain activity. Going one step back, a
performance is often hard to validate (Vernon, second way to incorporate knowledge from NF
2005), especially when one aims to report long- is to carry out pre-training screens (Blankertz et
term effects. Here serious games offer additional al., 2010; Maeder, Sannelli, Haufe, & Blankertz,
criteria, for example the amount of training it takes 2012). Short recordings of brain activity during
to progress to the next level of the game. Because rest may help to assess the overall level of relax-
users presumably spend significant time playing ation and arousal. In turn, this could have two
games this will facilitate future exploration of a important consequences: first, it indicates when
potential causal relationship. to avoid lengthy but ineffective training sessions
and second, it indicates and adjusts for the most
Incorporating Brain States into appropriate initial level of difficulty.
Serious Games As we have learned in the previous sections,
the participants internal motivation and the ex-
So far we considered basic principles and issues pected reward have been shown to modulate brain
from the field of neuroscience and cognitive psy- signals. For example, oscillatory 4-8 Hz activity in
chology. But how is that relevant for the gaming fronto-parietal brain areas does not only indicate
industry in any way? The answer for serious games whether an event is subsequently remembered but
is straightforward: if one wants to go beyond the also seems to be related to the anticipated reward
mere fun factor of commercial video games (Fell et al., 2012; Gruber, Watrous, Ekstrom,
and aims at beneficial training effects in terms Ranganath, & Otten, 2013; Musallam, Corneil,
of health, well-being or education, one definitely Greger, Scherberger, & Andersen, 2004). These
needs to acknowledge how the human nervous kinds of signals can help evaluating whether a
system responds to gaming. Human brains do given game design produces the desired effects:
have an immense capacity for plastic changes a highly motivating feedback recruits those brain
throughout the whole lifespan and this is where regions helping to enhance memory. Altogether,
one wants to make a difference. physiological signals and tools from the field of
Recent advancements in neuroscience have neuroscience and psychology can provide ad-
demonstrated that we are able to decode differ- ditional information for the implementation and
ent brain states, for example dissociating periods evaluation of serious games.
of resting from preparatory activity and action-
related brain signals. This is especially important
for neuroprosthetics, i.e. when technical devices DIFFERENT NEUROIMAGING
are used to substitute dysfunctional modalities of TECHNIQUES AND THEIR
the body. Indeed, first successful approaches in APPLICATION IN NEUROFEEDBACK
monkeys have been reported that allow for auto-
matic detection of the transition from one state to A reasonable application of gaming elements
another during arm reaching movements (Kemere in NF supposes that developers have a solid
et al., 2008). This kind of real-time monitoring understanding of the resources and limitations
of fluctuations between different mental states of neuroimaging techniques. That is, all people
is a promising tool for serious games to detect involved from engineer to software programmer
when the user is in an optimal condition for suc- to medical assistant - should know the strengths
88
Neurofeedback and Serious Games
and weaknesses of a given technique and find Silva, 2005). Generally, averaging techniques are
a common language to communicate emerging used to detect such ERP (Pfurtscheller & Lopes
problems from their field to the other disciplines. da Silva, 1999). SCPs reflect changes in cortical
We therefore will introduce the most common polarization, such as negative and positive trends
neurophysiological methods here and give some in the EEG, lasting from about 300ms to several
examples on feedback paradigms that have been seconds after an eliciting event. Several studies
realized so far (see Table 1.) have demonstrated associations between electro-
cortical negativity and behavioral measures such as
short-term memory performance or reaction time
EEG (Birbaumer, Elbert, Canavan, & Rockstroh, 1990).
In the following we will discuss some of the
The most frequently applied neurophysiological most interesting and promising NF and BCI-
method for NF and BCI applications is the elec- applications for serious games and virtual environ-
troencephalogram (EEG). This technique uses ments using EEG. In most cases game elements
electrodes to measure the electrical activity of the have been integrated in traditional NF or BCI
brain at the scalp surface, which is reflecting the protocols to increase participants motivation. One
summed potential of ionic currents across mem- reason is that NF is often used for the treatment of
branes of single cells. The EEG contains a wide clinical disorders e.g. ADHD, anxiety-disorders,
frequency spectrum that can be split up in different phobias and social problems where a substantial
frequency sub-bands. From a psychophysiological part of the treatment involves an active contribu-
viewpoint, the most important sub-bands which tion of participants. Thus, a combination of NF
are clinically relevant are: Delta (below 3.5 Hz), and games is highly promising particularly since
Theta (4-7.5 Hz), Alpha (8-13 Hz), Beta (14-30 traditional feedback screens are often monotonous
Hz), and Gamma (above 30 Hz) (Niedermeyer & consisting of two-dimensional moving bars, spots,
Lopes da Silva, 2005). These EEG frequencies are or lines. These modalities of training may quickly
associated with specific brain functions: Delta is become boring, what in turn can lead to decreased
prevalent in the EEG of infants or during sleep, training performance or a stagnated progression
Theta is associated with drowsiness but at the same of the clinical treatment.
time Theta is generally increased during spatial There is empirical evidence that the imple-
navigation, memory and language processes, mentation of simple game elements as feedback
Alpha is prevalent in the EEG during relaxation, modality can improve NF or BCI performance. For
Beta is associated with attentional processes and instance, Ron-Angevin and Daz-Estrella (2009)
concentration, and Gamma is linked to problem compared two different visual feedback protocols:
solving and memory work (Birbaumer & Schmidt, One group of participants received a conventional
2006; Kober & Neuper, 2011). For NF studies, the visual feedback consisting of a horizontal bar
power in different EEG frequency bands can be on a computer screen, which extends in varying
fed back to the user, but there are also EEG based degrees to the left or to the right depending on
NF applications using event-related potentials the classification result of the BCI classifier.
(ERP), such as slow cortical potentials (SCP), as The second group saw a car via a head mounted
feedback signal (Budzynski et al., 2009; Coben & display, which was driving down the middle of
Evans, 2010; Kropotov, 2009). ERP are defined three lanes. Additionally the participants had to
in the time domain as the electrical activity of the avoid obstacles appearing on the left or right lane.
brain that is triggered by the occurrence of par- Participants were able to control the cars position
ticular events or stimuli (Niedermeyer & Lopes da by means of voluntary changes in their EEG signal.
89
Neurofeedback and Serious Games
90
Neurofeedback and Serious Games
In both groups, the BCI training was carried out agery tasks can enhance motivation and mental
discriminating between two mental tasks: mental toughness (Mahmoudi & Erfanian, 2006). Ad-
relaxation and imagining a movement of the right ditionally motor-imagery tasks can also improve
hand. Ron-Angevin and Daz-Estrella (2009) were real motor skills (e.g. Mahmoudi & Erfanian,
able to show that there were significant differ- 2006), which could be a relevant implication for
ences in classification error rates between both rehabilitation purposes. However, conventional
visual feedback protocols. The visual feedback rehabilitation tools or protocols are often mono-
protocol that contained the car produced higher tone and the success of rehabilitation relies on
accuracy than the conventional visual horizontal patients compliance (Cameirao, Bermudez i
bar protocol. These results demonstrate that even Badia, & Verschure, 2008). Adding game ele-
slight changes of a conventional protocol can pro- ments could enhance therapy by increasing the
duce significant effects for BCI. By integrating motivation of the patients to participate in the
further game elements or by implementing virtual rehabilitation process.
reality technology (e.g. head mounted displays) An important issue in paradigms of increased
into traditional protocols it is even possible to complexity is how to design a vivid feedback while
improve therapeutic outcomes and increase the at the same time controlling for possible side ef-
participants perceived motivation (e.g. Leeb et fects. As one step in this direction, the successful
al., 2007). application of virtual reality to NF protocols has
More complex scenarios have already been been demonstrated (e.g. Cho et al., 2002), espe-
implemented and may provide an additional level cially for cognitive training. A NF study by Cho
of control in future serious games. One example and colleagues (2004) examined the effectiveness
is the study by Zhao, Zhang, & Cichocki (2009), of NF, along with virtual reality, for reducing the
where participants were able to control a virtual car level of inattention and impulsiveness in male
in 3D dynamically changing virtual environment. participants with social problems. To this end,
The researchers implemented a BCI system which the researchers created a virtual classroom. One
was able to distinguish between four mental states: group of participants used a head mounted display
motor imagery states of left hand-, right hand- and to visually orient themselves in the virtual world
foot-imagery as well as relaxation. These different (VR group). Another group only used a computer
mental tasks were decoded and used to control the monitor for their NF training. Both groups should
virtual car in the virtual environment: turning left learn to control the beta activity of their brain
or right, speedup and none command, respectively. signal (15-18Hz). Beta activity is closely related
Doud and colleagues (2011) trained participants to attention and impulsiveness. When the beta
to modulate their EEG activity through motor activity of the participants EEG signal was greater
imagery to achieve three-dimensional movement than a baseline threshold, participants earned a
of a virtual helicopter. Participants successfully score for a positive reinforcement outcome in
learned to control the helicopter with the goal of the virtual environment. The environment used
flying through rings in a 3D space. in this NF training was similar to a game. By
The examples mentioned so far can be consid- increasing the beta wave ratio a dinosaurs egg
ered as a proof-of-principle and one can foresee rose from a desk in the virtual classroom, which
different applications in everyday life. With re- then split in two. From the broken egg, each part
spect to serious games, users may experience an of a dinosaur picture puzzle gradually appeared
increased awareness of the learning content and on a whiteboard, if the participants were able to
higher motivation due to the enriched feedback. further increase the beta wave ratio. The game
Furthermore, controlling BCI systems with im- ended when all the puzzle pictures of the dinosaur
91
Neurofeedback and Serious Games
egg were completed on the whiteboard. The VR ability. The authors applied two 2D-(Pipe and
group with the head mounted display achieved Brain Chi) and two 3D-games (Dancing robot
better results than the group without VR in a test and Escape) in combination with a commercial
that was able to measure the ability to respond EEG system that detected the level of concentra-
and pay attention. Furthermore, the VR group tion. All these games used the same general game
paid more attention and made their decisions strategy: points could be earned or lost by the user
more rapidly in a continuous performance task. depending on whether concentration was in- or
The results of this study imply that NF training decreased (game examples are given at http://
within an immersive virtual reality is helpful in www3.ntu.edu.sg/home/EOSourina/projects.
attention enhancement. Similar results, regarding html; Sourina, 2013). Changing the game strategy
a better training outcome with more immersive in this setup, that is either enhancing concentration
NF training, have also been shown by Othmer or relaxation, provided a way of flexible training.
and Kaiser (2000) in a retrospective data analysis. Based on this common design, games may
The researchers examined different 2D and 3D implement different virtual environments and dif-
applications used for different NF trainings. They ferent tasks. The game Brain Chi for example is
discovered that 3D applications outperformed 2D a 2D single-player game, where the player controls
applications. The researchers conclude that some a little boy hero by using his/her brain power
tests of cognitive function show better outcomes (e.g. concentration level). The task was to help
with 3D applications. These findings imply that this little boy hero to fight against evil bats using
participants could benefit from a more complex a protection ball. The size of the protection ball
and information-rich feedback compared to con- is actually controlled by the brain power of the
ventional NF trainings. Virtual reality applications player: by increasing the size of the protection
for NF training seem to improve training outcome ball the player can eliminate all the bats and win
and should therefore be considered as an important the game.
part for future NF protocols. A sophisticated variant of this approach is
Recent studies took advantage of these promis- Escape by Wang, Sourina and Nguyen (2011).
ing findings and implemented game elements and This 3D single-player game has an educational
virtual reality technique for their NF trainings (e.g. purpose. The game-story requires the player to
Wang, Sourina, & Nguyen, 2011; Yan et al., 2008). solve different educational puzzles. If the player
While most paradigms mentioned above have been is able to solve the puzzle he or she could get a
driven by questions from a neuroscientific view, password to unlock doors in the virtual environ-
one can also recognize an increasing interest from ment to escape. However, when the player is not
game-inspired research. As mentioned in the study able to solve the puzzle in a conventional way an
by Wang and colleagues, EEG is widely used in alternative possibility is to analyze the concentra-
serious game design and becomes even more tion level of the player via the EEG. The player
prominent since more wireless headsets that meet has to stay concentrated for a specified amount
consumer criteria for affordability, portability and of time to get the password that he/she needs to
ease-of-use are available at the consumer market. pass through the door. In case the player uses his
For game implementation researchers have used brain power to get the password, the overall game
different sorts of game engines, which provide time allocated for the player to escape is reduced.
different tools and utilities designing and creating These examples provide insight on how NF
games. In a study by Wang, Sourina and Nguyen and games may be combined in future training
(2011) 2D and 3D NF games have been imple- approaches. NF might use game design principles
mented to help the user to improve concentration to make such training more fun, immersive and
92
Neurofeedback and Serious Games
93
Neurofeedback and Serious Games
their speed, direction, color, brightness and so on. companies. The graphics were improved as well
This introduces many degrees of freedom of how as the game play, but the essence did not change.
the brain signal could be fed back and was an- The physiological signals are used to drive ele-
other step forward in NF driven computer games ments of the game such as the main objects and
in the commercial sector of NF. the environment and the game present new chal-
Another example of a NF-game is space invad- lenges and rewards in various ways, such as through
ers (Figure 5). In this game, the movement of the new levels. As long as the game sticks to the
laser gun (ship at the bottom of the screen) was principles of operant conditioning and provides
controlled either by the user or therapist (via the meaningful feedback, game technology can enrich
keyboard) or automatically by the computer. The NF training and will continue to do so.
laser gun would shoot whenever all criteria were By highlighting also the perspective of an
met. Example: the laser gun would shoot when industrial company devoted to the manufacturing
SMR activity went over threshold (enough neu- of NF-devices and software we can be confident
rons firing in that frequency) and when Theta and that game elements become more and more
EMG artifact was low enough (below threshold). important for serious applications such as NF,
Besides 2D games, games with a 3D environ- especially for commercially available devices.
ment were developed by Mind Media and other As already mentioned above, the most frequently
Figure 5. Space Invaders: firing the laser gun is driven by brainwave activity
94
Neurofeedback and Serious Games
applied neurophysiological method for NF and of fMRI based feedback studies lies mainly on the
BCI applications is EEG. Due to its affordability, implementation of the continuous monitoring of
portability and ease-of-use EEG is also the most the BOLD response in real time rather than on
popular neurophysiological method for NF in the how these changes in the BOLD response can
industry. Nevertheless EEG is not the only method be fed back to the user optimally. Hence, only a
to fed back neurophysiological signals. few fMRI based feedback studies implemented
gaming elements as feedback modality. Goebel
and colleagues for example created a computer
MRI game based on table tennis that people can play
via self-regulated brain activity (Goebel, Sorger,
There are also attempts to use other neurophysi- Kaiser, Birbaumer, & Weiskopf, 2004).
ological methods than EEG for feedback studies. DeCharms and colleagues (2005) trained par-
Hemodynamic activity of the brain measured by ticipants to gain control over the endogenous pain
functional magnetic resonance imaging (fMRI) or modulatory system to enable voluntarily control
near-infrared spectroscopy (NIRS) is an additional over pain by using real-time fMRI. Therefore,
candidate signal for NF in this sense (S. Yoo, Lee, participants learned to increase or decrease activity
OLeary, Panych, & Jolesz, 2008). in the rostral anterior cingulate cortex (rACC), a
Magnet resonance imaging (MRI) is a non- region putatively involved in pain perception and
invasive brain imaging method that uses strong regulation. Activity in the rACC was fed back to
magnetic fields to create images of biological participants by means of visual feedback. Partici-
tissue. With functional MRI (fMRI) changes in pants received real-time fMRI information from
brain function, typically increases or decreases in the target brain region in the rACC as a scrolling
blood oxygenation, are measured during mental line graph of BOLD signal from the entire brain
activation by assessing changes in magnetization region and a continuous video display depicting
between oxygen-rich and oxygen-poor blood. the same information as a larger or smaller virtual
Hence, fMRI measures brain activity by detecting fire image. Hence, whenever the BOLD signal
associated changes in blood flow in the brain. fMRI in the rACC increased the scrolling line graph
uses the blood-oxygen-level-dependent (BOLD) moved up and the virtual fire got brighter. When
contrast as its basic measure. The BOLD hemo- participants deliberately induced increases or
dynamic response is defined as the relative con- decreases in rACC fMRI activation, there was a
centration change of deoxygenated hemoglobin corresponding change in the perception of pain
(oxygen-poor blood) following neural stimulation. caused by an applied noxious thermal stimulus.
After stimulus onset the concentration of deoxy- Spatial navigation by thoughts is also a hot topic
genated hemoglobin increases rapidly peaking at in the context of real-time fMRI applications. Yoo
about two seconds after stimulus onset, and then et al. (2004) used fMRI based BCI for navigation
declines to a minimum value about six seconds through a simple 2D maze solely through thought
after onset, before returning to baseline about ten processes. For navigation control, four different
seconds after stimulus onset (Huettel, Song, & Mc- mental tasks were used: mental calculation for
Carthy, 2009). One advantage of fMRI over EEG moving up, mental speech generation for moving
is its spatial resolution, which allows researchers down, right hand motor imagery to move to the
to target at the activation in anatomically specific right side, and left hand motor imagery to move
regions of the brain. to the left in the maze. These different mental
Real-time fMRI is a relatively new feedback tasks activate specific regions in the brain. When
technique compared to well-established EEG a participant wanted for instance to turn left in
based feedback applications. Therefore, the focus the virtual maze, he/she imagined a left hand
95
Neurofeedback and Serious Games
movement which led to an increase of activation & Chance, 1997). Activation in a specific brain
in the right (contralateral) motor cortex. This area leads to a localized vascular response that
change in the BOLD signal in the right motor causes an inflow of oxygen-rich blood to the ac-
cortex is detected online by the BCI system and tive brain area and its surrounding tissue. Thus,
consequently the curser moves to the left side in oxy-Hb increases and deoxy-Hb decreases in the
the maze. Hence, real-time fMRI applications can active brain region (Matthews et al., 2008). Such
be used to detect specific brain activation patterns a decrease in deoxy-Hb in active brain areas is the
and to translate these patterns into distinct BCI major source of the BOLD contrast as measured
commands. In the context of spatial navigation with fMRI (Telkemeyer et al., 2011). Hence, NIRS
by thoughts, Mueller and colleagues (2012) de- represents an adequate alternative to measure the
veloped neuroinformatics techniques that enable BOLD effect in cortical areas (Weiskopf, 2012).
real-time fMRI studies in virtual reality (VR) There is only a handful of NIRS based NF
environments. They designed a virtual 3D maze or BCI studies using other feedback modalities
and a virtual 3D city environment which could than moving bars. In one of these rare studies,
be combined with real-time fMRI. a NIRS based BCI was developed, which was
Real-time fMRI is also used to answer multiple- used to generate motion of a humanoid robot.
choice questions. By using different mental tasks Relative concentration changes in oxy-Hb were
(motor imagery, mental calculation, inner speech), assessed over frontal and temporal brain regions
participants voluntarily changed their BOLD while participants performed mental arithmetic
signal in specific brain regions to answer simple tasks compared to a resting period. Whenever the
questions. For instance, they were asked which oxy-Hb concentration level exceeded a specific
color did they like most. They could choose one threshold during the mental task, the brain machine
of four different answers, in this case red, blue, interface sent a control signal to a humanoid robot
green, or black. The participants task was to that was raising its right arm as visual feedback
select one of the four response options displayed (Matsuyama, Asama, & Otake, 2009).
at a screen and encode the corresponding letter Moreover, Coyle and colleagues (2007) used
(A, B, C, or D) by performing a certain mental NIRS based BCI to control a simple game called
task in a specific time window. In 94.4% of all Mindswitch. Mindswitch presents a basic on/
questions the participants answers were encoded off switching option to the user, where selection
correctly (Sorger et al., 2009). of either state takes 1 min by using motor imagery
strategies. NIRS optodes were placed over the
motor cortex to detect changes in oxy-Hb during
NIRS motor imagery.
Power et al. (2012) extended this two-choice
NIRS is a new non-invasive optical neuroimaging BCI system (Coyle et al., 2007), in which only
technique. With this method, relative concentra- two mental states (e.g., a mental task and rest)
tion changes of oxygen-rich and oxygen-poor are discriminated, and used three different mental
blood on the surface of the brain can be measured, states for controlling a game. Participants had to
which are indicative of local changes in brain ac- either perform mental arithmetic tasks, a mental
tivation. More precisely, NIRS measures changes singing task, or they had to allow natural thought
in oxygenated hemoglobin (oxy-Hb) and deoxy- patterns to occur without restriction (no-control
genated hemoglobin (deoxy-Hb) in the cerebral state) and change their oxygenation level in the
vessels based on their different absorption spectra prefrontal cortex voluntarily. With these mental
for light in the near-infrared range (Villringer strategies they could answer multiple choice
96
Neurofeedback and Serious Games
questions. For instance, participants saw three movement imagery were assessed in real-time
different animals at a screen, such as a turtle, over the motor cortex using NIRS. There is ample
a dog, and a fish. The task was to name those evidence that motor imagery activates motor areas
animals capable of walking. The three animals similar to those activated during motor execution
were highlighted one after each other. To choose of the same movement (Neuper, Scherer, Reiner,
the turtle and the dog for the correct answer, & Pfurtscheller, 2005; Wriessnegger, Kurzmann,
participants had to perform the mental strategies & Neuper, 2008). Hence, imagery of a left hand
as long as these animals were highlighted. When movement leads to increased brain activation over
the fish was highlighted, participants had to relax the contralateral (right) motor cortex and vice
themselves to indicate that they did not want to versa. After ten NF training sessions, the majority
select the fish. With this three-choice NIRS BCI of participants successfully learned to control the
system they reached classification accuracies of penguin. Hence, participants learned to play this
62.5% at maximum. Hence, in 62.5% of all pos- simple game, in which no muscular activity is
sible answers the participants could successfully needed for controlling the penguin (unpublished
change their oxygenation level in prefrontal brain data) (see Figure 7.)
regions to give the correct answer.
There are also attempts to use NIRS based Pros and Cons of Different
BCI systems for environmental control. Ayaz Neuroimaging Techniques Used for
et al. (Ayaz, Shewokis, Bunce, & Onaral, 2011) NF
developed and tested a new BCI design that uti-
lizes intention-related cognitive activity within In summary one can see that different neuro-
the dorsolateral prefrontal cortex recorded by physiological methods can be used successfully
functional near infrared spectroscopy (fNIRS). for NF and BCI applications. However, a critical
NIRS signals were used to augment interactive comparison of these brain imaging techniques
behavior within the 3D environment. Navigation commonly used for NF applications reveals that
through the virtual environment was controlled each method is associated with various advantages
by using a keyboard, but interaction with virtual and disadvantages. For instance, a crucial point
objects was controlled via the NIRS based BCI. for NF applications is the temporal resolution of
Participants consistently utilized the NIRS based the feedback signal. EEG has a high temporal
BCI with an overall success rate of 84% and voli- resolution compared to NIRS and fMRI, which
tionally increased their cerebral oxygenation level are based on relatively slow metabolic changes.
to trigger actions within the virtual environment. When using EEG as NF signal, changes in electri-
In one of our NF studies (unpublished data) we cal brain activity can be detected immediately and
developed a NIRS based NF protocol implement- fed back to the user with no time delay. In contrast,
ing a game-like feedback modality to increase NIRS and fMRI based NF systems have a latency
motivation and entertainment during repeated NF of several seconds in responding to a change in
training sessions. The aim of the feedback game users behavioral or mental states. This delayed
was to navigate a penguin either to the left or the feedback response limits the practical use of these
ride side of a feedback screen to catch a fish and systems (Cui, Bray, & Reiss, 2010).
avoid crashing into a barrier (see Figure 6.) Par- A further important issue is the spatial reso-
ticipants could move the penguin to the right side lution of the brain imaging technique. The big
when imaging a right hand movement or to the advantage of fMRI based NF over EEG and NIRS
left side of the screen when imaging a left hand based approaches is the whole-brain coverage,
movement. Changes in blood oxygenation during overcoming limitations of most EEG-based NF
97
Neurofeedback and Serious Games
Figure 6. Feedback screen of a NIRS based NF paradigm. The arrow indicated whether a participant
should imagine a right hand movement to steer the penguin to the right side of the screen or a left hand
movement to steer the penguin to the left. The aim was to catch the fish. The penguin moved in a constant
speed forward, left and right movements of the penguin corresponded to the oxygenation level in the left
and right motor cortex.
Figure 7. Projections of the 24 NIRS channel positions on the cortical surface (over the left and right
motor areas). NIRS positions are overlaid on a MNI-152 compatible canonical brain that is optimized
for NIRS analysis according to a procedure of (Singh, Okamoto, Dan, Jurcak, & Dan, 2005). The left
panel shows relative concentration changes in oxy-Hb over the right motor cortex during a left hand
movement, whereas the right panel pictures increased activation over the left motor cortex during move-
ment of the right hand.
98
Neurofeedback and Serious Games
studies (Weiskopf, 2012). Using fMRI, it is pos- Recently, wireless and portable instruments are
sible to measure activity in deep brain structures available which might be useful when using NIRS
such as the hippocampus or the amygdala. In or EEG in the gaming context (Muehlemann,
the majority of EEG based NF studies only a Haensse, & Wolf, 2008; Sitaram et al., 2009).
rather small number of electrodes were used to While lying in a MR-scanner, some participants
accelerate the measurement preparation process. even experience discomfort because of the noise
Such a small number of electrodes allows only and the narrow scanner. Additionally, fMRI is
for unreliable localization of active brain areas locally bounded to the installation site and re-
and results in limited access to deep sub-cortical stricted in study design due to the limitation of
brain areas. Even with modern multi-channel EEG the participants free moving (Weiskopf, 2012).
systems electric source localization in the brain The last point refers to the costs of the different
is an intrinsically ill-posed problem (Weiskopf, brain imaging methods. While fMRI is the most
2012). Using NIRS, it is only possible to measure expensive technique, both in terms of purchase and
changes in hemodynamic responses at the surface maintenance, EEG is highly cost-effective. Table 2
of the brain. The penetration is a limitation of summarizes the above mentioned advantages and
NIRS, because one can only measure changes in disadvantages of the different neurophysiological
oxy- and deoxy-Hb a few centimeters (0.5 2 cm) methods used for NF and BCI applications. NIRS
from the surface of the head, with a relatively low seems to have great potential for future NF studies.
signal-to-noise ratio (Huppert, Hoge, Diamond,
Franceschini, & Boas, 2006).
The quality of the assessed neurophysiological FUTURE RESEARCH DIRECTIONS
data is also crucial for NF applications. Compared
to fMRI and EEG, NIRS has the big advantage As outlined so far, the different neuroimaging
of a reduced sensitivity to motion-artifacts and techniques have various advantages and disad-
can accommodate a higher degree of movement vantages. However, improvements in technology
(Lloyd-Fox, Blasi, & Elwell, 2010; Nambu et al., will increase the possibilities of using neuroimag-
2009). For instance, NIRS can be adequately used ing techniques during serious gaming especially
to measure brain activation patterns during play- when devices become more user friendly and meet
ing a dance video game, where participants had consumer criteria for affordability, portability and
to move extensively (Tachibana, Noah, Bronner, ease-of-use. In general the combination of NF
Ono, & Onozuka, 2011). While lying in the narrow and serious games can lead to major advance-
space inside the MR-scanner, participants should
not move their head to avoid motion artifacts which
can confound the fMRI results. Table 2. Advantages and disadvantages of EEG,
In neuroscientific studies, factors such as mea- NIRS, and fMRI for NF applications
surement preparation or comfort for participants EEG NIRS fMRI
should be considered as well. The montage of
Spatial resolution - +
the NIRS optodes is very fast and repeated NIRS
Temporal resolution + - -
measurements are also more comfortable for par-
Portability + + -
ticipants compared to EEG measurements since
Costs + -
there is no abrasive gel needed, which could lead
Motion tolerance - + -
to skin irritations. Compared to fMRI, EEG and
NIRS are more flexible and portable (Huppert et Measurement preparation - + +
al., 2006; Sitaram, Caria, & Birbaumer, 2009). Comfort for participants - + -
99
Neurofeedback and Serious Games
ments in both fields. Intuitively, one could come (Salmoni, Schmidt, & Walter, 1984; Schmidt,
to the conclusion that the best way to increase Young, Swinnen, & Shapiro, 1989) suggests that
game efficiency is to provide as much feedback practice with concurrent feedback results in the
information as possible to the user. There is in- fusion of them: the feedback becomes a central
deed evidence that multimodal sensory feedback part of the task as the new skill consolidates.
allows reaching the threshold of a given neural Future work will thus have to assess long-term
activation faster and therefore enhances learning effects of individualized, multimodal NF strate-
(Shams & Seitz, 2008). This finding relates to gies on cognitive improvements in serious games
theories that suggest a distributed cognitive load approaches.
in multimodal tasks (Burke et al., 2006; Sigrist,
Rauter, Riener, & Wolf, 2012).
However, one should be careful not to gen- CONCLUSION
eralize this concept. Firstly, the amount and
complexity of feedback seems to have different In the present review, we focused on the combina-
effects depending on the task complexity (for a tion of NF and serious games. We outlined the
review see Sigrist et al., 2012). For instance, in benefits of this combination such as improved
simple visuomotor tasks a vivid, real-time feed- learning outcomes and increased motivation and
back approach can actually interfere negatively attention of users. But we also tried to point out
with learning processes. Secondly, the impact of its limitations, for instance highly complex games
artifacts has to be considered with great care. Eye can interfere with the learning outcome. Overall,
movements and muscle activity, for example, can we foresee that neuroscience and gaming will
produce large amplitude signals superimposed to become more tightly coupled but future studies
brain activity reflecting the genuine activation of are mandatory to address remaining issues. We
cognitive functions. These kinds of artifacts, often particularly identified shortcomings in a common
not consciously perceived by users, can easily language and a coherent study design between
provide an effective way of control over feedback both disciplines. Overcoming this limitation will
and mimic the recruitment of cognitive functions. likely be a key point for success.
The failure to detect and isolate the influence of After an extensive literature search it turned
this kind of artifact on signal may completely out that feedback studies using games as feedback
prevent effective use of NF to optimize learning. modality are rare. The majority of studies used
Therefore, more than simply trying to combine simple two-dimensional moving objects as feed-
feedback on EEG signal with games, one should back modality. For instance, voluntary changes
check for plausibility and aim for an artifact-free in EEG power spectra were visually fed back to
control signal that presumably will result in more the participant as a bar moving up and down on
targeted impact on brain functions. a computer screen. Nevertheless, there are some
Ultimately, one may be able to take advantage promising approaches that started to implement
of the optimal sources, i.e. one can combine game-like aspects. So far, those examples have
sensory modalities in a task-specific manner. One often focused on spatial aspects, for example
can take into consideration the fine spatial details producing a control signal in virtual reality envi-
our visual system is able to resolve and add to that ronments. The next step is to develop true serious
the high temporal resolution of the hearing system. games that incorporate knowledge about attention
The challenge is to create individual solutions: and memory capacities coming from the cognitive
which feedback is optimal for different tasks and neurosciences. This novel aspect should take us
different users? And will this change during the one step closer to the ultimate goal of boosting
learning process? Evidence from motor learning human brain performance.
100
Neurofeedback and Serious Games
101
Neurofeedback and Serious Games
Budzynski, T., Budzynski Kogan, H., Evans, J., & Coyle, S. M., Ward, T. E., & Markham, C. M.
Abarbanel, A. (2009). Introduction to quantitative (2007). Brain-computer interface using a sim-
EEG and neurofeedback. The Journal of Head plified functional near-infrared spectroscopy
Trauma Rehabilitation. system. Journal of Neural Engineering, 4(3),
219226. doi:10.1088/1741-2560/4/3/007
Burde, W., & Blankertz, B. (2006). Is the locus
PMID:17873424.
of control of reinforcement a predictor of brain-
computer interface performance. In Proceedings Cui, X., Bray, S., & Reiss, A. L. (2010). Speeded
of the 3rd International Braincomputer Inferface near infrared spectroscopy (NIRS) response de-
Workshop and Training Course 2006 (pp. 7677). tection. PLoS ONE, 5(11), e15474. doi:10.1371/
IEEE. journal.pone.0015474 PMID:21085607.
Burke, J., Prewett, M., Gray, A., Yang, L., Stil- Daum, I., Rockstroh, B., Birbaumer, N., Elbert,
son, F., Coovert, M., et al. (2006). Comparing T., Canavan, A., & Lutzenberger, W. (1993). Be-
the effects of visual-auditory and visual-tactile havioural treatment of slow cortical potentials in
feedback on user performance: A meta-analysis. intractable epilepsy: Neuropsychological predic-
In Proceedings of the 8th International Conference tors of outcome. Journal of Neurology, Neurosur-
on Multimodal Interfaces (pp. 108117). IEEE. gery, and Psychiatry, 56(1), 9497. doi:10.1136/
jnnp.56.1.94 PMID:8429329.
Buzski, G. (2006). Rhythms of the brain. New
York: Oxford University Press. doi:10.1093/acp deCharms, R. C., Maeda, F., Glover, G. H., Lud-
rof:oso/9780195301069.001.0001. low, D., Pauly, J. M., & Soneji, D. etal. (2005).
Control over brain activation and pain learned by
Cameirao, M. S., Bermudez i Badia, S., & Ver-
using real-time functional MRI. Proceedings of the
schure, P. F. M. J. (2008). Virtual reality based
National Academy of Sciences of the United States
upper extremity rehabilitation following stroke:
of America, 102(51), 1862618631. doi:10.1073/
A review. Journal of CyberTherapy & Rehabilita-
pnas.0505210102 PMID:16352728.
tion, 1(1), 6373.
Doud, A. J., Lucas, J. P., Pisansky, M. T., & He,
Cho, B.-H., Kim, S., Shin, D. I., Lee, J. H., Lee,
B. (2011). Continuous three-dimensional control
S. M., Kim, I. Y., & Kim, S. I. (2004). Neurofeed-
of a virtual helicopter using a motor imagery
back training with virtual reality for inattention
based brain-computer interface. PLoS ONE,
and impulsiveness. Cyberpsychology & Behavior:
6(10), e26322. doi:10.1371/journal.pone.0026322
The Impact of the Internet. Multimedia and Virtual
PMID:22046274.
Reality on Behavior and Society, 7(5), 519527.
Egner, T., & Gruzelier, J. H. (2001). Learned
Cho, B.-H., Ku, J., Jang, D. P., Kim, S., Lee, Y.
self-regulation of EEG frequency components
H., & Kim, I. Y. etal. (2002). The effect of virtual
affects attention and event-related brain poten-
reality cognitive training for attention enhance-
tials in humans. Neuroreport, 12(18), 4155
ment. Cyberpsychology & Behavior: The Impact
4159. doi:10.1097/00001756-200112210-00058
of the Internet. Multimedia and Virtual Reality on
PMID:11742256.
Behavior and Society, 5(2), 129137.
Coben, R., & Evans, J. (2010). Neurofeedback and
neuromodulation techniques and applications.
London: Elsevier Academic Press.
102
Neurofeedback and Serious Games
Egner, T., & Gruzelier, J. H. (2004). EEG Green, C. S., & Bavelier, D. (2003). Action video
biofeedback of low beta band components: game modifies visual selective attention. Nature,
Frequency-specific effects on variables of atten- 423(6939), 534537. doi:10.1038/nature01647
tion and event-related brain potentials. Clinical PMID:12774121.
Neurophysiology, 115(1), 131139. doi:10.1016/
Green, C. S., Pouget, A., & Bavelier, D. (2010).
S1388-2457(03)00353-5 PMID:14706480.
Improved probabilistic inference as a general
Fell, J., Axmacher, N., & Haupt, S. (2010). From learning mechanism with action video games. Cur-
alpha to gamma: Electrophysiological correlates of rent Biology, 20(17), 15731579. doi:10.1016/j.
meditation-related states of consciousness. Medi- cub.2010.07.040 PMID:20833324.
cal Hypotheses, 75(2), 218224. doi:10.1016/j.
Green, R., & Turner, G. (2010). Growing evidence
mehy.2010.02.025 PMID:20227193.
for the influence of meditation on brain and behav-
Fell, J., Staresina, B., Do Lam, A., Widman, G., iour. Neuropsychological Rehabilitation, 20(2),
Helmstaedter, C., Elger, C., & Axmacher, N. 306311. doi:10.1080/09602010903172239
(2012). Memory modulation by weak synchro- PMID:20204915.
nous deep brain stimulation: A pilot study. Brain
Gruber, M. J., Watrous, A. J., Ekstrom, A. D., Ran-
Stimulation. PMID:22939277.
ganath, C., & Otten, L. J. (2013). Expected reward
Feng, J., Spence, I., & Pratt, J. (2007). Playing an modulates encoding-related theta activity before
action video game reduces gender differences in an event. NeuroImage, 64, 6874. doi:10.1016/j.
spatial cognition. Psychological Science, 18(10), neuroimage.2012.07.064 PMID:22917987.
850855. doi:10.1111/j.1467-9280.2007.01990.x
Gruzelier, J., & Egner, T. (2005). Critical
PMID:17894600.
validation studies of neurofeedback. Child and
Garris, R., & Driskell, J. E. (2002). Games, mo- Adolescent Psychiatric Clinics of North America,
tivation, and learning: A research and practice 14(1), 83104. doi:10.1016/j.chc.2004.07.002
model. Practice, 117. PMID:15564053.
Gevensleben, H., Holl, B., Albrecht, B., Schlamp, Gruzelier, J., Egner, T., & Vernon, D. (2006). Vali-
D., Kratz, O., & Studer, P. et al. (2010). Neu- dating the efficacy of neurofeedback for optimising
rofeedback training in children with ADHD: performance. Progress in Brain Research, 159,
6-month follow-up of a randomised controlled 421431. doi:10.1016/S0079-6123(06)59027-2
trial. European Child & Adolescent Psychiatry, PMID:17071246.
19(9), 715724. doi:10.1007/s00787-010-0109-5
Gruzelier, J., Inoue, A., Smart, R., Steed, A.,
PMID:20499120.
& Steffert, T. (2010). Acting performance and
Goebel, R., Sorger, B., Kaiser, J., Birbaumer, flow state enhanced with sensory-motor rhythm
N., & Weiskopf, N. (2004). BOLD brain pong: neurofeedback comparing ecologically valid im-
Self regulation of local brain activity during mersive VR and training screen scenarios. Neuro-
synchronously scanned, interacting subjects. In science Letters, 480(2), 112116. doi:10.1016/j.
Proceedings of 34th Annual Meeting of the Society neulet.2010.06.019 PMID:20542087.
for Neuroscience. IEEE.
103
Neurofeedback and Serious Games
Guger, C., Edlinger, G., Harkam, W., Niedermeyer, Kober, S. E., Kurzmann, J., & Neuper, C. (2012).
I., & Pfurtscheller, G. (2003). How many people Cortical correlate of spatial presence in 2D and
are able to operate an EEG-based brain-computer 3D interactive virtual reality: An EEG study.
interface (BCI)? IEEE Transactions on Neural International Journal of Psychophysiology: Of-
Systems and Rehabilitation Engineering, 11(2), ficial Journal of the International Organization of
145147. doi:10.1109/TNSRE.2003.814481 Psychophysiology, 83(3), 365374. doi:10.1016/j.
PMID:12899258. ijpsycho.2011.12.003 PMID:22206906.
Hammer, E. M., Halder, S., Blankertz, B., Sannelli, Kober, S. E., & Neuper, C. (2011). Sex differences
C., Dickhaus, T., & Kleih, S. etal. (2012). Psycho- in human EEG theta oscillations during spatial
logical predictors of SMR-BCI performance. Bio- navigation in virtual reality. International Journal
logical Psychology, 89(1), 8086. doi:10.1016/j. of Psychophysiology: Official Journal of the Inter-
biopsycho.2011.09.006 PMID:21964375. national Organization of Psychophysiology, 79(3),
347355. doi:10.1016/j.ijpsycho.2010.12.002
Hoedlmoser, K., Pecherstorfer, T., Gruber, G.,
PMID:21146566.
Anderer, P., Doppelmayr, M., Klimesch, W., &
Schabus, M. (2008). Instrumental conditioning of Kotchoubey, B., Strehl, U., Holzapfel, S., Blanken-
human sensorimotor rhythm (12-15 Hz) and its horn, V., Frscher, W., & Birbaumer, N. (1999).
impact on sleep as well as declarative learning. Negative potential shifts and the prediction of the
Sleep, 31(10), 14011408. PMID:18853937. outcome of neurofeedback therapy in epilepsy.
Clinical Neurophysiology: Official Journal of
Holzapfel, S., Strehl, U., Kotchoubey, B., &
the International Federation of Clinical Neuro-
Birbaumer, N. (1998). Behavioral psychophysi-
physiology, 110(4), 683686. doi:10.1016/S1388-
ological intervention in a mentally retarded
2457(99)00005-X PMID:10378738.
epileptic patient with brain lesion. Applied Psy-
chophysiology and Biofeedback, 23(2), 189202. Kropotov, J. D. (2009). Quantitative EEG, event-
doi:10.1023/A:1022299422116 PMID:10384250. related potentials and neurotherapy. San Diego,
CA: Elsevier Academic Press.
Huettel, S. A., Song, A. W., & McCarthy, G.
(2009). Functional magnetic resonance imaging Leeb, R., Scherer, R., Friedman, D., Lee, F., Kein-
(2nd ed.). Boston: Sinauer Associates, Inc. rath, C., Bischog, H., et al. (2007). Combining BCI
and virtual reality: Scouting virtual worlds. In G.
Huppert, T. J., Hoge, R. D., Diamond, S. G., Fran-
Dornhege, J. del R. Milln, T. Hinterberger, D. J.
ceschini, M. A., & Boas, D. A. (2006). A temporal
McFarland, & K.-R. Mller (Eds.), Toward brain
comparison of BOLD, ASL, and NIRS hemody-
computer interfacing (pp. 393408). Cambridge,
namic responses to motor stimuli in adult humans.
MA: MIT Press.
NeuroImage, 29(2), 368382. doi:10.1016/j.
neuroimage.2005.08.065 PMID:16303317. Lloyd-Fox, S., Blasi, a, & Elwell, C. E. (2010). Il-
luminating the developing brain: The past, present
Kemere, C., Santhanam, G., Yu, B. M., Afshar, A.,
and future of functional near infrared spectroscopy.
Ryu, S. I., Meng, T. H., & Shenoy, K. V. (2008).
Neuroscience and Biobehavioral Reviews, 34(3),
Detecting neural-state transitions using hidden
269284. doi:10.1016/j.neubiorev.2009.07.008
Markov models for motor cortical prostheses.
PMID:19632270.
Journal of Neurophysiology, 100(4), 24412452.
doi:10.1152/jn.00924.2007 PMID:18614757.
104
Neurofeedback and Serious Games
Lofthouse, N., Arnold, L., & Hurt, E. (2012). Mueller, C., Luehrs, M., Baecke, S., Adolf, D.,
Current status of neurofeedback for attention- Luetzkendorf, R., Luchtmann, M., & Bernarding,
deficit/hyperactivity disorder. Current Psychiatry J. (2012). Building virtual reality fMRI para-
Reports, 14(5), 536542. doi:10.1007/s11920- digms: A framework for presenting immersive
012-0301-z PMID:22890816. virtual environments. Journal of Neuroscience
Methods, 209(2), 290298. doi:10.1016/j.jneu-
Maeder, C. L., Sannelli, C., Haufe, S., & Blankertz,
meth.2012.06.025 PMID:22759716.
B. (2012). Pre-stimulus sensorimotor rhythms
influence brain-computer interface classifica- Musallam, S., Corneil, B. D., Greger, B.,
tion performance. IEEE Transactions on Neural Scherberger, H., & Andersen, R. A. (2004).
Systems and Rehabilitation Engineering: A Pub- Cognitive control signals for neural prosthetics.
lication of the IEEE Engineering in Medicine Science, 305(5681), 258262. doi:10.1126/sci-
and Biology Society, 20(5), 65362. doi:10.1109/ ence.1097938 PMID:15247483.
TNSRE.2012.2205707
Nambu, I., Osu, R., Sato, M., Ando, S., Kawato,
Mahmoudi, B., & Erfanian, A. (2006). Electro- M., & Naito, E. (2009). Single-trial reconstruc-
encephalogram based brain-computer interface: tion of finger-pinch forces from human motor-
Improved performance by mental practice and con- cortical activation measured by near-infrared
centration skills. Medical & Biological Engineer- spectroscopy (NIRS). NeuroImage, 47(2), 628
ing & Computing, 44(11), 959969. doi:10.1007/ 637. doi:10.1016/j.neuroimage.2009.04.050
s11517-006-0111-8 PMID:17028907. PMID:19393320.
Matsuyama, H., Asama, H., & Otake, M. (2009). Nan, W., Rodrigues, J. P., Ma, J., Qu, X., Wan, F.,
Design of differential near-infrared spectroscopy & Mak, P.-I. etal. (2012). Individual alpha neu-
based brain machine interface. In Proceedings of rofeedback training effect on short term memory.
18th IEEE International Symposium on Robot and International Journal of Psychophysiology: Of-
Human Interactive Communication, (pp. 775 ficial Journal of the International Organization
780). IEEE. doi:10.1109/ROMAN.2009.5326215 of Psychophysiology, 86(1), 8387. doi:10.1016/j.
ijpsycho.2012.07.182 PMID:22864258.
Matthews, R., Turner, P. J., McDonald, N. J.,
Ermolaev, K., Manus, T. M., Shelby, R. A., & Neuper, C., Scherer, R., Reiner, M., & Pfurtschell-
Steindorf, M. (2008). Real time workload classi- er, G. (2005). Imagery of motor actions: Dif-
fication from an ambulatory wireless EEG system ferential effects of kinesthetic and visual-motor
using hybrid EEG electrodes. In Proceedings: mode of imagery in single-trial EEG. Brain
Annual International Conference of the IEEE Research. Cognitive Brain Research, 25(3),
Engineering in Medicine and Biology Society. 668677. doi:10.1016/j.cogbrainres.2005.08.014
IEEE. doi:10.1109/IEMBS.2008.4650550 PMID:16236487.
Muehlemann, T., Haensse, D., & Wolf, M. (2008). Niedermeyer, E., & Lopes da Silva, F. H. (2005).
Wireless miniaturized in-vivo near infrared im- Electroencephalography: Basic principles, clini-
aging. Optics Express, 16(14), 1032310330. cal applications, and related fields. Academic
doi:10.1364/OE.16.010323 PMID:18607442. Press.
105
Neurofeedback and Serious Games
Nijboer, F., Furdea, A., Gunst, I., Mellinger, J., Roberts, L., Birbaumer, N., Rockstroh, B., Lut-
McFarland, D. J., Birbaumer, N., & Kbler, A. zenberger, W., & Elbert, T. (1989). Self-report
(2008). An auditory brain-computer interface during feedback regulation of slow cortical
(BCI). Journal of Neuroscience Methods, 167(1), potentials. Psychophysiology, 26(4), 392
4350. doi:10.1016/j.jneumeth.2007.02.009 403. doi:10.1111/j.1469-8986.1989.tb01941.x
PMID:17399797. PMID:2798689.
Oei, A. C., & Patterson, M. D. (2013). Enhancing Ron-Angevin, R., & Daz-Estrella, A. (2009).
cognition with video games: A multiple game train- Brain-computer interface: Changes in perfor-
ing study. PLoS ONE, 8(3), e58546. doi:10.1371/ mance using virtual reality techniques. Neuro-
journal.pone.0058546 PMID:23516504. science Letters, 449(2), 123127. doi:10.1016/j.
neulet.2008.10.099 PMID:19000739.
Othmer, S., & Kaiser, D. (2000). Implementation
of virtual reality in EEG. Biofeedback, 3(3). Rotter, J. (1966). Generalized expectancies for
internal versus external control of reinforce-
Ott, M., & Tavella, M. (2009). A contribution to
ment. Psychological Monographs, 88(609).
the understanding of what makes young students
PMID:5340840.
genuinely engaged in computer-based learning
tasks. Procedia - Social and Behavioral Sciences, Rowe, J. P., Shores, L. R., Mott, B. W., & Lester,
1(1), 184188. doi:10.1016/j.sbspro.2009.01.034 J. C. (2010). Integrating learning and engagement
in narrative-centered learning environments. Aca-
Pfurtscheller, G. (1992). Event-related synchro-
demic Press. doi:10.1007/978-3-642-13437-1_17.
nization (ERS): An electrophysiological cor-
relate of cortical areas at rest. Electroencepha- Salmoni, A., Schmidt, R., & Walter, C. (1984).
lography and Clinical Neurophysiology, 83(1), Knowledge of results and motor learning: A review
6269. doi:10.1016/0013-4694(92)90133-3 and critical reappraisal. Psychological Bulletin,
PMID:1376667. 95(3), 355386. doi:10.1037/0033-2909.95.3.355
PMID:6399752.
Pfurtscheller, G., & Lopes da Silva, F. H. (1999).
Event-related EEG/MEG synchronization and Schmidt, R., Young, D., Swinnen, S., & Shap-
desynchronization: Basic principles. Clinical Neu- iro, D. (1989). Summary knowledge of results
rophysiology: Official Journal of the International for skill acquisition: Support for the guidance
Federation of Clinical Neurophysiology, 110(11), hypothesis. Journal of Experimental Psychol-
18421857. doi:10.1016/S1388-2457(99)00141- ogy. Learning, Memory, and Cognition, 15(2),
8 PMID:10576479. 352359. doi:10.1037/0278-7393.15.2.352
PMID:2522520.
Posner, M., & Raichle, M. (1994). Images of mind.
New York: Scientific American Library. Serruya, M. D., & Kahana, M. J. (2008).
Techniques and devices to restore cognition.
Power, S. D., Kushki, A., & Chau, T. (2012).
Behavioural Brain Research, 192(2), 149165.
Automatic single-trial discrimination of mental
doi:10.1016/j.bbr.2008.04.007 PMID:18539345.
arithmetic, mental singing and the no-control
state from prefrontal activity: Toward a three- Shams, L., & Seitz, A. R. (2008). Benefits of mul-
state NIRS-BCI. BMC Research Notes, 5(1), 141. tisensory learning. Trends in Cognitive Sciences,
doi:10.1186/1756-0500-5-141 PMID:22414111. 12(11), 411417. doi:10.1016/j.tics.2008.07.006
PMID:18805039.
106
Neurofeedback and Serious Games
Sigrist, R., Rauter, G., Riener, R., & Wolf, P. Vernon, D. (2005). Can neurofeedback train-
(2012). Augmented visual, auditory, haptic, and ing enhance performance? An evaluation of the
multimodal feedback in motor learning: A review. evidence with implications for future research.
Psychonomic Bulletin & Review. doi: doi:10.3758/ Applied Psychophysiology and Biofeedback,
s13423-012-0333-8 PMID:23132605. 30(4), 347364. doi:10.1007/s10484-005-8421-4
PMID:16385423.
Singh, A. K., Okamoto, M., Dan, H., Jurcak, V., &
Dan, I. (2005). Spatial registration of multichannel Vernon, D., Egner, T., Cooper, N., Compton, T.,
multi-subject fNIRS data to MNI space without Neilands, C., Sheri, A., & Gruzelier, J. (2003).
MRI. NeuroImage, 27(4), 842851. doi:10.1016/j. The effect of training distinct neurofeedback
neuroimage.2005.05.019 PMID:15979346. protocols on aspects of cognitive performance.
International Journal of Psychophysiology: Of-
Sitaram, R., Caria, A., & Birbaumer, N. (2009).
ficial Journal of the International Organization
Hemodynamic brain-computer interfaces for com-
of Psychophysiology, 47(1), 7585. doi:10.1016/
munication and rehabilitation. Neural Networks:
S0167-8760(02)00091-0 PMID:12543448.
The Official Journal of the International Neural
Network Society, 22(9), 13201328. doi:10.1016/j. Villringer, A., & Chance, B. (1997). Non-invasive
neunet.2009.05.009 PMID:19524399. optical spectroscopy and imaging of human
brain function. Trends in Neurosciences, 20(10),
Sorger, B., Dahmen, B., Reithler, J., Gosseries,
435442. doi:10.1016/S0166-2236(97)01132-6
O., Maudoux, A., Laureys, S., & Goebel, R.
PMID:9347608.
(2009). Another kind of bold response: Answer-
ing multiple-choice questions via online decoded Wang, Q., Sourina, O., & Nguyen, M. K. (2010).
single-trial brain signals. London: Elsevier. EEG-based serious games design for medical ap-
doi:10.1016/S0079-6123(09)17719-1. plications. In Proceedings of 2010 International
Conference on Cyberworlds (pp. 270276). Sin-
Sourina, O. (2013). Emotion-based personalised
gapore: IEEE. doi:10.1109/CW.2010.56
digital media experience in co-spaces (EmoDEx
Project). Retrieved January 5, 2013, from http:// Wang, Q., Sourina, O., & Nguyen, M. K. (2011).
www3.ntu.edu.sg/home/EOSourina/projects.html Fractal dimension based neurofeedback in serious
games. The Visual Computer, 27(4), 299309.
Tachibana, A., Noah, J. A., Bronner, S., Ono, Y.,
doi:10.1007/s00371-011-0551-5.
& Onozuka, M. (2011). Parietal and temporal
activity during a multimodal dance video game: Weiskopf, N. (2012). Real-time fMRI and its ap-
an fNIRS study. Neuroscience Letters, 503(2), plication to neurofeedback. NeuroImage, 62(2),
125130. doi:10.1016/j.neulet.2011.08.023 682692. doi:10.1016/j.neuroimage.2011.10.009
PMID:21875646. PMID:22019880.
Telkemeyer, S., Rossi, S., Nierhaus, T., Steinbrink, Wolpaw, J. R., Birbaumer, N., McFarland, D.
J., Obrig, H., & Wartenburger, I. (2011). Acoustic J., Pfurtscheller, G., & Vaughan, T. M. (2002).
processing of temporally modulated sounds in Brain-computer interfaces for communication
infants: Evidence from a combined near-infrared and control. Clinical Neurophysiology: Official
spectroscopy and EEG study. Frontiers in Psy- Journal of the International Federation of Clinical
chology, 1, 62. doi:10.3389/fpsyg.2011.00062 Neurophysiology, 113(6), 767791. doi:10.1016/
PMID:21716574. S1388-2457(02)00057-3 PMID:12048038.
107
Neurofeedback and Serious Games
Wriessnegger, S. C., Kurzmann, J., & Neuper, Birbaumer, N., Elbert, T., Canavan, A., & Rock-
C. (2008). Spatio-temporal differences in brain stroh, B. (1990). Slow potentials of the cerebral
oxygenation between movement execution and im- cortex and behavior. Physiological Reviews, 70(1),
agery: A multichannel near-infrared spectroscopy 128. PMID:2404287.
study. International Journal of Psychophysiology:
Birbaumer, N., Ghanayim, N., Hinterberger,
Official Journal of the International Organization
T., Iversen, I., & Kotchoubey, B., Kbler, a,
of Psychophysiology, 67(1), 5463. doi:10.1016/j.
Perelmouter, J., et al. (1999). A spelling device
ijpsycho.2007.10.004 PMID:18006099.
for the paralysed. Nature, 398(6725), 297298.
Yan, N., Wang, J., Liu, M., Zong, L., Jiao, Y., & doi:10.1038/18581 PMID:10192330.
Yue, J. etal. (2008). Designing a brain-computer
Birbaumer, N., & Schmidt, R. (2006). Biologische
interface device for neurofeedback using virtual
Psychologie (6th ed.). Heidelberg: Springer-
environments. Journal of Medical and Biological
Verlag.
Engineering, 28(3), 167172.
Blankertz, B., Sannelli, C., Halder, S., Hammer,
Yoo, S., Lee, J., OLeary, H., Panych, L. P.,
E. M., Kbler, A., & Mller, K.-R. etal. (2010).
& Jolesz, F. A. (2008). Neurofeedback fMRI-
Neurophysiological predictor of SMR-based
mediated learning and consolidation of regional
BCI performance. NeuroImage, 51(4), 1303
brain activation during motor imagery. Inter-
1309. doi:10.1016/j.neuroimage.2010.03.022
national Journal of Imaging Systems and Tech-
PMID:20303409.
nology, 18(1), 6978. doi:10.1002/ima.20139
PMID:19526048. Buzski, G. (2006). Rhythms of the Brain
(1st ed.). New York, USA: Oxford Uni-
Yoo, S.-S., Fairneny, T., Chen, N.-K., Choo, S.-E.,
versity Press. doi:10.1093/acprof:o
Panych, L. P., & Park, H. etal. (2004). Braincom-
so/9780195301069.001.0001.
puter interface using fMRI: Spatial navigation
by thoughts. Neuroreport, 15(10), 15911595. Cho, B.-H., Kim, S., Shin, D. I., Lee, J. H., Lee,
doi:10.1097/01.wnr.0000133296.39160.fe S. M., Kim, I. Y., & Kim, S. I. (2004). Neurofeed-
PMID:15232289. back training with virtual reality for inattention
and impulsiveness. Cyberpsychology & behavior:
Zhao, Q., Zhang, L., & Cichocki, A. (2009).
the impact of the Internet, multimedia and virtual
EEG-based asynchronous BCI control of a car in
reality on behavior and society, 7(5), 519527.
3D virtual reality environments. Chinese Science
Retrieved from http://online.liebertpub.com/doi/
Bulletin, 54(1), 7887. doi:10.1007/s11434-008-
abs/10.1089/cpb.2004.7.519
0547-3.
Coben, R., & Evans, J. (2010). Neurofeedback
and neuromodulation techniques and applications
(R. Coben, & J. Evans, Eds.). 1st ed.). London:
ADDITIONAL READING
Elsevier Academic Press.
Angelakis, E., Stathopoulou, S., Frymiare, Green, C. S., & Bavelier, D. (2003). Action video
J. L., Green, D. L., Lubar, J. F., & Kounios, game modifies visual selective attention. Nature,
J. (2007). EEG neurofeedback: a brief over- 423(6939), 534537. doi:10.1038/nature01647
view and an example of peak alpha frequency PMID:12774121.
training for cognitive enhancement in the el-
derly. The Clinical Neuropsychologist, 21(1),
110129. doi:10.1080/13854040600744839
PMID:17366280.
108
Neurofeedback and Serious Games
Green, C. S., Pouget, A., & Bavelier, D. (2010). Oei, A. C., & Patterson, M. D. (2013). Enhancing
Improved probabilistic inference as a general Cognition with Video Games: A Multiple Game
learning mechanism with action video games. Cur- Training Study. (J. J. Geng, Ed.)PLoS ONE, 8(3),
rent biology: CB, 20(17), 15739. doi:10.1016/j. e58546. doi: doi:10.1371/journal.pone.0058546.
cub.2010.07.040
Othmer, S., & Kaiser, D. (2000). Implementation
Gruzelier, J., Egner, T., & Vernon, D. (2006). Vali- of Virtual Reality in EEG Biofeedback, 3(3).
dating the efficacy of neurofeedback for optimising
Pfurtscheller, G. (1992). Event-related synchro-
performance. Progress in Brain Research, 159,
nization (ERS), An electrophysiological cor-
421431. doi:10.1016/S0079-6123(06)59027-2
relate of cortical areas at rest. Electroencepha-
PMID:17071246.
lography and Clinical Neurophysiology, 83(1),
Huettel, S. A., Song, A. W., & McCarthy, G. 6269. doi:10.1016/0013-4694(92)90133-3
(2009). Functional Magnetic Resonance Imaging PMID:1376667.
(2nd ed., pp. 147154). Massachusetts: Sinauer
Pfurtscheller, G., & Lopes da Silva, F. H. (1999).
Associates, Inc.
Event-related EEG/MEG synchronization and
Kober, S. E., Kurzmann, J., & Neuper, C. (2012). desynchronization: basic principles. Clinical
Cortical correlate of spatial presence in 2D and neurophysiology: official journal of the Interna-
3D interactive virtual reality: an EEG study. tional Federation of Clinical Neurophysiology,
International journal of psychophysiology: offi- 110(11), 184257.
cial journal of the International Organization of
Ron-Angevin, R., & Daz-Estrella, A. (2009).
Psychophysiology, 83(3), 36574. doi:10.1016/j.
Brain-computer interface: changes in performance
ijpsycho.2011.12.003
using virtual reality techniques. Neuroscience
Kober, S. E., & Neuper, C. (2011). Sex differences Letters, 449(2), 123127. doi:10.1016/j.neu-
in human EEG theta oscillations during spatial let.2008.10.099 PMID:19000739.
navigation in virtual reality. International journal
Sigrist, R., Rauter, G., Riener, R., & Wolf, P.
of psychophysiology: official journal of the Inter-
(2012). Augmented visual, auditory, haptic, and
national Organization of Psychophysiology, 79(3),
multimodal feedback in motor learning: A review.
34755. doi:10.1016/j.ijpsycho.2010.12.002
Psychonomic Bulletin & Review. doi: doi:10.3758/
Kropotov, J. D. (2009). Quantitative EEG, event- s13423-012-0333-8 PMID:23132605.
related potentials and neurotherapy (1st ed.). San
Vernon, D. (2005). Can neurofeedback train-
Diego: Elsevier Academic Press.
ing enhance performance? An evaluation of the
Leeb, R., Scherer, R., Friedman, D., Lee, F., Kein- evidence with implications for future research.
rath, C., Bischog, H., et al. (2007). Combining BCI Applied Psychophysiology and Biofeedback,
and Virtual Reality: Scouting Virtual Worlds. In 30(4), 347364. doi:10.1007/s10484-005-8421-4
G. Dornhege, J. del R. Milln, T. Hinterberger, PMID:16385423.
D. J. McFarland, & K.-R. Mller (Eds.), Toward
Vernon, D., Egner, T., Cooper, N., Compton, T.,
brain computer interfacing (pp. 393408). Mas-
Neilands, C., Sheri, A., & Gruzelier, J. (2003).
sachusetts: MIT Press.
The effect of training distinct neurofeedback
Niedermeyer, E., & Lopes da Silva, F. H. (2005). protocols on aspects of cognitive performance.
Electroencephalography: Basic Principles. Clini- International journal of psychophysiology: of-
cal Applications, and Related Fields. ficial journal of the International Organization
of Psychophysiology, 47(1), 7585.
109
Neurofeedback and Serious Games
Wang, Q., Sourina, O., & Nguyen, M. K. (2010). functions and processes such as brain activity,
EEG-Based Serious Games Design for Medical muscle tone, skin conductance, heart rate, etc. and
Applications. 2010 International Conference on consequently can learn to control them voluntarily.
Cyberworlds (pp. 270276). Singapore: Ieee. Brain-Computer Interface: A direct com-
doi:10.1109/CW.2010.56 munication pathway between a human brain and
external devices such as a wheelchair, a computer,
Wang, Q., Sourina, O., & Nguyen, M. K. (2011).
or prosthesis.
Fractal dimension based neurofeedback in serious
Neurofeedback: A type of biofeedback. The
games. The Visual Computer, 27(4), 299309.
users brain activation is depicted in real-time with
doi:10.1007/s00371-011-0551-5.
the goal of helping the user to gain control over
Weiskopf, N. (2012). Real-time fMRI and its ap- specific aspects of the activity in his/her central
plication to neurofeedback. NeuroImage, 62(2), nervous system.
682692. doi:10.1016/j.neuroimage.2011.10.009 Neuroscience: An interdisciplinary science
PMID:22019880. that studies the nervous system.
Non-Invasive Neuroimaging: Various tech-
Yan, N., Wang, J., Liu, M., Zong, L., Jiao, Y., &
niques to directly or indirectly image the function
Yue, J. etal. (2008). Designing a Brain-computer
or structure of the brain without penetrating the
Interface Device for Neurofeedback Using Virtual
body of the participant.
Environments. Journal of Medical and Biological
Serious Games: Games that have educational
Engineering, 28(3), 167172.
and/or health-related aims besides entertainment.
Virtual Reality: A computer generated envi-
ronment that simulates physical presence in real
KEY TERMS AND DEFINITIONS: or artificial worlds.
110