Orientation of Attention in Visual Feedbacks during Neurofeedback
Relaxation
Mehdi Karamnejad, Diane Gromala, Amber Choo, Chris Shaw and Xin Tong
School of Interactive Arts & Technology, Simon Fraser University, Surrey, BC, Canada
Keywords: Brain-Computer Interface, Biofeedback, Neurofeedback, Attention, Ambient Displays, Interface Design,
Human-Computer Interaction.
Abstract: The assumptions underlying differing approaches to interface design result, in part, on how attention is
managed and categorized using theories from media studies. The authors propose the term intraface to refer
to biofeedback or other interfaces that are designed to support users who direct their attention inward to
inner physiological states. In this paper, the role of representing feedback data in abstract forms is compared
in an experiment using Neurosky’s neurofeedback device. Although preliminary, the results suggest that
mapping biofeedback data from a brain-computer interface (BCI) to highly abstract ambient animations is
more effective for relaxation than mapping it to a highly familiar symbolic smiley face icon or to a progress
bar. The authors propose that the relative success of the abstract ambient animation can be explained
because this representation of biofeedback data is the form that requires the least amount of attention, and
that designing biofeedback interfaces that distribute the attention, supports the need of users to the task of
directing most of their attention to their inner physiological states.
1 INTRODUCTION
When designing computational systems for Human-
Computer Interaction (HCI), the commonly-held
assumption is that the interfaces should function as
windows between the user and the information
(Norman, 1988). More recent work in the area of
visualization, such as ambient displays (fig. 1) and
what has been turned “informative art” do not,
however, share this assumption: they are designed to
function in the background of our attention until we
attend to them (Bartram and Woodbury, 2011).
Once we attend to them, we see more abstract
representations of information that provide relatively
imprecise information. There is yet another or third
approach: some interfaces are intentionally designed
to call attention to themselves (fig. 2). The authors
of Windows and Mirrors offer a useful way to
categorize this diversity: the interfaces that are
designed to be invisible function like a “window,”
while those designed to call attention to themselves
function as a sort of “mirror” (Bolter and Gromala,
2005).
In most cases, biofeedback systems need to
function as a sort of mirror. Instead of just reflecting
the user as a mirror would, however, the
biofeedback interface represents the continually
changing states of the user. The goal for feedback is
to not only provide the user with information about
their changing physiological states, but at the same
time, is meant to support the user in trying to affect
those inner states as well. Thus, the information that
is being continuously updated and displayed should
ideally call as little attention to itself as possible.
That is because most of users’ attention must focus
on trying to sense and change their inner
physiological states. Because a user’s focus is
inward, the authors propose the term “intraface” for
these kinds of applications.
2 INTERFACES VS. INTRAFACES
In any biofeedback interface, the user’s continually
changing inner states are displayed and are attended
to by the user, and are designed to help that user
learn to change his/her internal state. In this context,
the interface must accomplish two seemingly
contradictory feats. First, it must provide
information for the user and thus must be in the
user’s attentional zone to some degree. Second,
simultaneously, the user must intentionally focus on
196
Karamnejad M., Gromala D., Choo A., Shaw C. and Tong X..
Orientation of Attention in Visual Feedbacks during Neurofeedback Relaxation.
DOI: 10.5220/0004724801960203
In Proceedings of the International Conference on Physiological Computing Systems (PhyCS-2014), pages 196-203
ISBN: 978-989-758-006-2
Copyright
c
2014 SCITEPRESS (Science and Technology Publications, Lda.)
their inner physiological states. In this case, the
intraface must simultaneously function as both a
window and a mirror. The study described in section
4 focuses on three different ways of representing
biofeedback information, and their effect in
supporting a user’s ability to focus on his or her
inner state. The design challenge for a biofeedback
intraface is to reduce the stimulus-driven attentional
demands of the feedback so the user
can direct as
much attention as possible to gaining an awareness
of their interoceptive processes.
Figure 1: Ambient information display for WestHouse.
This visualization of energy consumption is designed to
function as an ambient display – as a backsplash in a
kitchen, it remains in the background of our attention until
we direct our attention to it. In this case, the greater the
energy use, the more attention it draws (Rodgers, 2010).
Figure 2: (The Wooden Mirror, 1999, Daniel Rozin). An
example of an interface that functions as a “mirror” does
so here in a literal way. The wood of this mirror is novel,
and is intended to call attention to itself. In addition, a
user’s image and self-awareness is meant to be a primary
part of the information or experience.
In the case of biofeedback, although attention is
divided – to the display and to inner states – cannot
be assumed to be equal or invariable. Informal
observations of users during the study suggests that
they do one of two things: they continually shift
their attention back and forth between the
information displayed by the biofeedback device and
their changing internal states, or they manage to
attend to both the display and their inner state,
usually after trial-and-error. During the study, for
example, we observed that most users struggled with
directing their attention between the display and
their inner states; when they attended to the
biofeedback display, their stress levels increased. A
few users, however, appeared to be able to direct
their attention simultaneously to both the display and
their inner state. In the latter case, the user appeared
to not directly look at the display so much as
maintain it in what can be described as a more
distracted or ambient way.
Ideally, in biofeedback contexts, because
attention is necessarily split, the interface should
function more like a partially transparent mirror.
Because the task is to learn how to change one’s
inner state, much of a user’s attention needs to be
directed to this task, especially when it is a task that
the user has not attempted before. At the same time,
however, feedback is intended to support the task.
That feedback should take a secondary role, and
should therefore be designed so that it requires as
little attention as possible. In a pre-test trial, one user,
for example, suggested that her split attention
worked in a figure/ground relationship; that is, s/he
was able to put the displayed information into her
attentional background. It may be possible that users
who are experienced meditators may bring this skill,
and that is a factor we will determine in future
work.
Biofeedback is a task that is unfamiliar for most
users. Indeed, physiological research suggests that
we generally do not pay attention to our inner states.
According to Hermann Helmholtz, we have 100,000
times more information about our inner or
interoceptive states than information about our five
senses or exteroceptive sense (Helmholtz, 1995).
Yet, if we attended to the sheer quantity of this
information, we would have little capacity left to
focus on our senses and on our immediate
environment (Leder, 1990). Thus, information that
our bodies generate about our internal state normally
functions in the background of our conscious
awareness and attention. When threatened, however,
– say, by eating bad oysters – we are able to
attend to at least some of this information. Indeed,
depending on the threat, this information can capture
nearly all of our attention and impel us to take action.
Such action is, in some instances, involuntary – the
information provokes involuntary processes that take
over the control we usually have over, in this case,
our ability to hold food within.
For contexts such as biofeedback or meditation,
OrientationofAttentioninVisualFeedbacksduringNeurofeedbackRelaxation
197
users must overcome this propensity to ignore their
inner states, and must learn to effect change by
processes that seem elusive. While it is easy to exert
conscious control over processes such as breathing,
it is more difficult to exert control and effect change
over other processes such as heart rate, brain waves,
and Galvanic Skin Response (GSR). Nevertheless,
this is an ability that can be learned, and through
practice be made easier to do; in some cases, a long-
term practice can become a habit. Stress reduction
learned via biofeedback is a common example.
To differentiate the ways in which interfaces
serve to direct attention, we propose the term
intraface for technical systems that are designed to
focus a user's attention inward, to their interoceptive
states.
2.1 Biofeedback and Brain-Computer
Interface
Biofeedback technology is typically used to gain an
awareness of physiological processes; the goal is to
learn how to manipulate processes that can be
controlled, such as brainwaves, muscle tone, skin
conductance, and heart rate. Biofeedback sensors
attached to the user’s body capture on-going data on
each process. This data is then sent to a computer or
other device where it is then mapped to visuals or
sounds and displayed (Schwartz, 1987; Montgomery,
2008; Andreassi, 2007). The representation of this
data provides the user with continually updated
information about the activity of these processes,
often in near real-time.
Often, the physiological changes occur in
conjunction with changes in thoughts or emotional
states. With practice, these changes may be
maintained without the use of the biofeedback
technology itself. One of the major principles of this
approach is that the user gains the skill to control
aspects that otherwise primarily operate
unconsciously (Frank et al., 2010). Biofeedback
regulation techniques have proven to be effective in
treating disorders such as attention deficit
hyperactivity disorder (ADHD), anxiety, chronic
pain, epilepsy and a host of other conditions
(Monastra et al., 2005; Rice et al., 1993; Sterman,
2000).
One of the well-known ways that biofeedback is
used is to enable patients suffering from
neurological disorders to observe and regulate their
neural oscillations towards a healthier direction; this
is termed neurofeedback. In this practice, patients
are usually monitored using a non-invasive brain
activity recording method such as EEG. Patients
obtain an awareness of their brain performance
through forms of feedback, usually while performing
a certain task.
Interacting with software systems using these so-
called brain-triggered commands has led to a
research area known as BCI. Users of these systems
control and drive functions embedded in software by
commands issued by the brain – brain activities are
picked up using a signal acquisition approach and
then translated into actions using signal processing
and machine learning methods. This technology was
initially developed to enable patients with severe
motor injuries to regain mobility (Wolpaw et al.,
2002). For instance, amyotrophic lateral sclerosis
(ALS) patients who gradually lose the ability to
control their muscles can use commands originating
from their brain to control a robotic arm or to
operate a wheelchair via brain activity (Sellers and
Donchin, 2006). Other applications of BCI include
gaming and virtual reality (Bayliss and Ballard,
2000). Creating such systems that can utilize brain
activity as the communication and control medium
enables health practitioners to develop treatments for
patients.
2.2 Examples of Intrafaces
Interfaces designed for biofeedback systems can all
be termed intrafaces because they are intended to
support the user’s task of learning how to change
their inner or interoceptive states. The primary goal
is for the user to direct his or her attention inward.
Because this task is generally novel and relatively
difficult, most of a user’s attention should ideally
remain directed inward. Thus, the design of the
interface to the feedback information should require
as little attention as possible. Early biofeedback
devices displayed a graph, a wave, or a sonic tone.
More contemporary biofeedback devices offer a
number of ways in which the information can be
represented: by numbers, graphs, smiley faces or
more abstract images.
VR researchers have explored displaying
biofeedback in an immersive virtual environment.
Designers of the Virtual Meditative Walk (Shaw et
al., 2007), for example, created a VR system that
incorporates biofeedback in order to for the user to
gain an immersive sense of when and how they may
affect their stress levels. Rather than displaying
biofeedback information in familiar ways, they
mapped the continually changing biofeedback data
to fog that dissipates as users lower their GSR
data. Because the fog surrounds a user and lacks any
specific point to focus on, these researchers suggest
PhyCS2014-InternationalConferenceonPhysiologicalComputingSystems
198
that it is very different from VR systems designed to
distract a user from their pain (Shaw et al., 2007). In
VR pain distraction, the virtual environment (VE) is
designed to capture and hold as much of a user’s
attentional capacity as possible (Hoffman, 2000). In
the Virtual Meditative Walk, in contrast, the fog
demands attention (Shaw et al., 2007) and in this
sense functions more like an ambient display.
Other studies that compare the difference
between a VE that is realistic to a VE that is abstract
have been proposed. The study below is the first
stage of this planned work.
3 THE STUDY: THE ROLE OF
ABSTRACTION IN FOCUSING
ATTENTION INWARD
In order to assess if in fact users are better able to
focus on their interoceptive states if the biofeedback
data is represented as abstractions, we designed the
following study. We used a BCI that monitors
meditation status originating from the brain during
relaxation and maps brain activities to a visual
representation as feedback. 12 participants were
exposed to three types of feedback and 4 people
were monitored without being exposed to any
feedback as control group. They used different
representations of biofeedback data, including an
ambient one to control their relaxation state.
4 METHODOLOGY
4.1 Apparatus
We used MindWave Mobile from NeuroSky Inc. to
acquire neural brain activity associated with
relaxation state. The device is designed for practical
applications of BCI; it consists of a dry electrode
that picks up EEG signals and transmits data to the
receiving end via Bluetooth technology (fig. 3).
4.2 Participants
Participants were a convenience sample comprised
of 16 male and female university students between
the ages of 21 and 26 (M = 23, SD = 3.49) with no
previous meditation or biofeedback experience.
They were recruited using Doodle event manager.
Participants were briefed about the experiment via
email before arrival, and signed a consent form
before the experiment.
Figure 3: MindWave Mobile (side and front view).
4.3 Experimental Design
To assess different types of visual feedback in a
brain-triggered system, we designed a between-
subject experiment. The independent variable was
the type of feedback or control group that the
participants were randomly exposed to and the
dependent variable was the level of relaxation in a
percentage reported by NeuroSky’s Mindwave
Mobile. We developed three visual types of
feedbacks using processing programming language
consisted of the following: a progress bar, an
animation of cartoon face, and a slow moving
abstract animation.
4.3.1 Control Group
In this group, participants were instructed to relax by
deep breathing and to try to keep their mind free of
distractions. They were not exposed to any visual
feedback during the relaxation process.
4.3.2 Progress Bar
Participants of this group were instructed to look at a
progress bar which shows how relaxed a participant
was during relaxation (fig. 4). The more relaxed the
participant was, the more filled the progress bar was;
a fully filled progress bar indicated the most relaxed
state (out of one hundred, from 0 to 100).
Figure 4: The amount the bar is filled directly correlates
with the participant’s relaxation state; an empty bar
represents a state of agitation, and a fully filled bar
represents the most relaxed state.
OrientationofAttentioninVisualFeedbacksduringNeurofeedbackRelaxation
199
4.3.3 Smiley Face Animation
In this group, the percentage of relaxation affected
the animation of a cartoon-like smiley face. When
the participant was less relaxed, the face looked sad,
darker, and gray. As the participant reached a more
relaxed state, the face became happy, lighter and
more colorful (fig. 5).
Figure 5: The image gradually transitions between these
three images.
4.3.4 Ambient Feedback
The relaxation level in this group was mapped to a
video that changed color with respect to participant's
relaxation status. The original video constituted
abstract animations and was meant to be the main
focus of participant’s attention (fig. 6). As the
participants moved from a non-relaxed state to a
relaxed state, the colors of the video changed from
red to orange to green to blue, respectively. The goal
in this mode was to display the feedback in a
passive, implicit yet informative manner.
4.4 Procedure
We randomly assigned each participant to a
feedback type and recorded relaxation levels in
percentage for 5 minutes. The sampling rate
provided by NeuroSky MindWave is one data point
per second; hence there were 300 data points for
each participant at the end of the trial. This was
saved as plain text file for further analysis.
NeuroSky MindWave conveys the patient’s relaxed
mental state with percentages utilizing a proprietary
algorithm called eSense. According to NeuroSky,
eSense constitutes artifact rejection and machine
learning methods that can distinguish among higher
cognitive mental states such as meditation and
attention. The manufacturer of MindWave
conducted an extensive study to distinguish when
participants are in a calm and relaxing state by
providing relaxation levels with percentages
(“NeuroSky’s eSense Meters and Detection of
Mental State,” 2013). This output as opposed to
conventional raw EEG that requires preprocessing
along with machine learning techniques to classify
mental states is used for practical application-
oriented studies.
4.5 Statistical Analysis
We initially averaged 300 data points for each
participant into a single value representing the
relaxation level. To test for potential significant
difference among means of relaxation levels, we ran
a one-way analysis of variance (ANOVA) test. To
further investigate the effects of each feedback, we
ran a pair-wise Tukey HSD test in order to compare
all pair of feedback types.
Figure 6: The top left image (red) represents agitation. As
the user relaxes, the image transitions to the bottom right
image (blue), which represents the most relaxed state. This
slow-moving animation does not offer any one area that
demands more attention than any other area.
5 RESULTS
Primary observations of data are demonstrated in
Figure 7. The results suggest that the participants
who are exposed to ambient feedback have the
highest level of relaxation compared to the other
visual feedback scenarios and the control group
(Mean= 77.61%). The Face Smiley mapping
(Mean=68.89), Progress Bar (Mean =58.68), and the
control group (Mean=46.78) are rated afterwards,
respectively.
The results of one-way ANOVA suggests that
there was a significant effect of independent variable
feedback type on dependent variable relaxation level
for 4 conditions F(3,12)=23.74, p<0.0001 at the p
value <0.05. Figure 8 demonstrates the results of
analysis for all the conditions.
The results of pair wise Tukey HSD test suggest
a significant difference between ambient feedbacks
when compared to other feedback types. Table 1
provides result of Tukey test along with statistical
PhyCS2014-InternationalConferenceonPhysiologicalComputingSystems
200
Table 1: Comparing feedback types pair-wise using Tukey
HSD.
Feedback type 1 Feedback type 2 p-Value
Ambient Control group <.0001
Ambient Progress bar 0.0013
Ambient Face 0.0139
Face Control group 0.0029
Face Progress bar 0.5205
Progress bar Control group 0.0329
significant at p value <0.05. It also supports the fact
that the difference between progress bar mapping
and the face mapping is not statically significant.
Figure 7: average relaxation level for each participant
along with its standard deviation. The average relaxation
level for each feedback condition is also represented.
6 DISCUSSION
The results clearly show that subjects achieve a
higher level of relaxation when they use the ambient
video display compared to the other conditions. The
cartoon face and the progress bar are equivalent
(p=0.52 no significant difference). The control group
(relaxing without feedback) resulted in the lowest
Figure 8: Visualization of data points for different
participants in all conditions.
relaxation readings.
Our suggestion as to why this is the case has to
do with the role that visual attention plays in each of
the treatments. In the face condition, the visual
stimulus can be strongly attended to, namely the
curve of the mouth and the color of the face.
Moreover, the smiley face is a schematic face,
and is thus a highly familiar symbolic cue. Cues like
these are often referred to as a gaze cue: they appear
to be processed faster and more accurately, and are
thought to be so very well-learned – or even
overlearned – that responses to them may seem
reflexive, but are automatic (Vecera & Rizzo, 2006).
If this is true, and if gaze cues are more related to
goal-directed attention than stimulus-directed
attention, why then, does the ambient video function
better in supporting users’ attention to their
interoceptive processes? First, unlike direct cues,
gaze cues persist longer, and do not produce do not
produce inhibition of return (Friesen &
Kingstone ,1998). Second, the human face is the
most important social stimulus we process (Itier et
al., 2007), and is fundamental to social cognition.
Thus, schematic faces actually function to direct
attention, or by “popping out,” call attention to
themselves. Although schematic faces are thought to
support task-oriented attention (Wright & Ward,
2008), the biofeedback task may be qualitatively so
different from those used in attention studies that
they function as a unique case. Put another way,
when the task is to try to focus one’s attention
inward, toward one’s interoceptive senses,
recognition of a face, no matter how schematized,
implies that that we are impelled to look back at its
highly salient visual features. In this case, the
OrientationofAttentioninVisualFeedbacksduringNeurofeedbackRelaxation
201
schematic face functions more like a stimulus that
cannot be easily ignored. In addition, that it is a face
means that it bears visual features perceived as a
facial feature, one that users can attend to in a
focused manner.
Similarly, the progress bar has a single visual
feature – the bar location. In addition, some users in
terms of a performance can interpret the progress bar
as a goal or challenge. That is, the half empty
progress bar can be perceived as a challenge to fill,
and that challenge seems to continually draw
attention outward, to the bar.
By contrast, the ambient video distributes visual
attention across a wider range of the visual field. In
addition, the video has no specific element to focus
on. Also, the video changes slowly and continuously,
frame-to-frame. Slow changes are difficult to attend
to, particularly when the changes are taking place in
visual stimuli that do not contain a clearly
identifiable central object of interest (Auvray et al.,
2003).
Thus, in the ambient video condition, attentional
resources that could be devoted outward to highly
salient visual features can be instead directed inward
towards managing the internal sense of relaxation,
with occasional reference to the color and general
appearance of the ambient video.
The NeuroSky EEG sensing device generates the
numerical measures of relaxation. The relaxation
measure was generated at NeuroSky’s labs by
having a number of subjects enter a relaxed state and
recording a sequence of raw sensor readings during
this state. A neural net recognizer was trained on this
set of subject data, and the output of this trained
neural net is the relaxation reading.
The advantage of this is that we did not have to
train our own machine learning system to recognize
raw readings as “relaxed” or otherwise. However,
the disadvantage of using NeuroSky’s recognizer is
that we have to trust that the NeuroSky device
measures what it claims to measure. Thus, we plan
in future work to compare NeuroSky’s measures
with data from other biofeedback devices.
We observed that subjects almost always
attended to the progress bar and smiley faces, but
not the ambient animation. For reasons that are
unclear, subjects tended to look at the animation,
close their eyes, and look back at the animation from
time to time, ostensibly for updates from the
feedback. It is unclear whether this resulted from the
kind of representation: the progress bar and smiley
face may have appeared to users to be more “formal,”
while the animation may have been perceived to be
more informal, especially because no one-to-one
data mapping appeared evident.
7 CONCLUSIONS
Differing approaches to interface design operate on
underlying assumptions of how attention is directed
and managed. We draw upon a theory from media
studies that characterize these approaches as
transparent windows onto the information displayed,
or mirrors that demand that users pay attention to the
interface itself, and in some cases, include
reflections of users themselves in the interface.
Interfaces for contexts that support users is usually
an unfamiliar task of directing their attention inward,
to their physiological processes, like biofeedback
should call as little attention to themselves as
possible. This enables users to maintain focus on
their inner states, and on their ability to learn how to
change them. Biofeedback presents a challenge in
that the feedback tends to split attention to the
representation of the feedback and to users’ inner
states. The authors propose the term intraface to
refer to biofeedback or other interfaces that are
designed to support users who direct their attention
to inner physiological states. The role of
representing feedback data in abstract forms is
compared in an experiment using Neurosky’s
neurofeedback device. Preliminary results suggest
that mapping biofeedback data from a brain-
computer interface (BCI) to highly abstract ambient
animations is more effective than mapping it to a
highly familiar symbolic smiley face icon or to a
progress bar. The relative success of the abstract
ambient animation over the schematic face or
progress bar may be because ambient representation
of biofeedback data requires the least amount of
attention. Therefore, designing biofeedback
interfaces should be designed so that they require the
least amount of attention, supporting the need of
users to the task of directing most of their attention
to their inner physiological states.
ACKNOWLEDGEMENTS
This research is partially supported by GRAND
Network of Centres of Excellence. We also thank
Jeremy Mamisao for helping with graphic design.
PhyCS2014-InternationalConferenceonPhysiologicalComputingSystems
202
REFERENCES
Andreassi, J. L., 2007. Psychophysiology: Human
behavior and physiological response. Psychology
Press, 5th edition.
Auvray, M., & O'Regan, J.K., 2003. Influence of semantic
factors on blindness to progressive changes in visual
scenes / L'influence des facteurs sémantiques sur la
cécité aux changements progressifs dans les scènes
visuelles. L'année Psychologique. 103(1), 9-32.
Bayliss, J. D., & Ballard, D. H., 2000. A virtual reality
testbed for brain-computer interface research. IEEE
transactions on rehabilitation engineering: a
publication of the IEEE Engineering in Medicine and
Biology Society, 8(2), 188–190.
Bolter, J. D. and Gromala, D. J., 2005. Windows and
Mirrors: Interaction Design, Digital Art, and the Myth
of Transparency. Cambridge: the MIT Press.
Bartram, Lyn, and Rob Woodbury. Smart Homes or Smart
Occupants? Reframing, 2011. Artificial Intelligence
and Sustainable Design, AAAI Spring Symposium.
Rodgers, Johnny. 2010. Westhouse. Accessed 13 Sept.
2013, http://johnny.hcssl.iat.sfu.ca/tag/ambient-infovis.
Frank, D. L., Khorshid, L., Kiffer, J. F., Moravec, C. S., &
McKee, M. G., 2010. Biofeedback in medicine: who,
when, why and how? Mental Health in Family
Medicine, 7(2), 85–91.
Friesen C. K., Kingstone A., 1998. The eyes have it!
Reflexive orienting is triggered by nonpredictive gaze.
Psychonomic Bulletin & Review, 5:490–495.
Helmholtz, H., 1995. Science and Culture: Popular and
Philosophical Essays. Chicago: University of Chicago
Press, 1
st
ed.
Hoffman, H., 2000. “Virtual reality as an adjunctive pain
control during burn wound care in adolescent patients,”
Pain, 85, no. 1 (2000), 305–309.
Itier R. J., Alain C, Sedore K, McIntosh AR, 2007. Early
face processing specificity: it’s in the eyes!. J Cogn
Neurosci 19:1815–1826.
Leder, D., 1990. The Absent Body. Chicago: the
University of Chicago Press.
Monastra, V. J., Lynn, S., Linden, M., Lubar, J. F.,
Gruzelier, J., & LaVaque, T. J. , 2005.
Electroencephalographic biofeedback in the treatment
of attention-deficit/hyperactivity disorder. Applied
psychophysiology and biofeedback, 30(2), 95–114.
NeuroSky’s eSense Meters and Detection of Mental State,
2013. Neurosky. Retrieved from http://
company.neurosky.com/files/neurosky_esense_whitep
aper.pdf.
Norman, D., 1988. The Psychology of Everyday Things.
New York: Basic Books.
Norman, D., 1999. The Invisible Computer: Why Good
Products Can Fail, the Personal Computer Is So
Complex, and Information Appliances Are the Solution.
Cambridge: the MIT Press.
Rice, K. M., Blanchard, E. B., & Purcell, M., 1993.
Biofeedback treatments of generalized anxiety
disorder: preliminary results. Biofeedback and self-
regulation, 18(2), 93–105.
Schwartz, M., 1987. Biofeedback: A Practitioner's Guide.
New York: The Guilford Press.
Sellers, E. W., & Donchin, E., 2006. A P300-based brain–
computer interface: Initial tests by ALS patients.
Clinical Neurophysiology, 117(3), 538–548.
doi:10.1016/j.clinph.2005.06.027.
Shaw, C., Gromala, D , and Seay, F., 2007. “The
Meditation Chamber” Enacting Autonomic Senses.
Proceedings of ENACTIVE/07, 4th International
Conference on Enactive Interfaces.
Sterman, M. B., 2000. Basic concepts and clinical findings
in the treatment of seizure disorders with EEG operant
conditioning. Clinical EEG (electroencephalography),
31(1), 45–55.
Vecera, S. P., & Rizzo, M. , 2006. Eye gaze does not
produce reflexive shifts of visual attention: Evidence
from frontal-lobe damage. Neuropsychologia, 44, 150-
159.
Wolpaw, J. R., Birbaumer, N., McFarland, D. J.,
Pfurtscheller, G., & Vaughan, T. M., 2002. Brain-
computer interfaces for communication and control.
Clinical neurophysiology: official journal of the
International Federation of Clinical Neurophysiology,
113(6), 767–791.
Wright, R. D. and Ward, L. M., 2008. Orienting of
Attention. Oxford: Oxford University Press.
Yucha, C; Montgomery D, 2008. Evidence-based practice
in biofeedback and neurofeedback (PDF). Wheat
Ridge, CO: AAPB. Archived from the original on
2010-10-09.
OrientationofAttentioninVisualFeedbacksduringNeurofeedbackRelaxation
203
