Telekinesis, made possible by BCIs [1]
Joseph
Tam Chun Kit (joseph.tam.2013@accountancy.smu.edu.sg), 1st Year
Student, School of Accountancy, Singapore Management University (SMU)
Executive
Summary
The
paper will cover on the development of Brain Computer Interface (BCI)
technology throughout the years, and examine the two major focus of BCI
research today: Invasive BCI, and non-invasive BCI. It will address the applications of BCI in
our current situation, and examine the shortcomings of each of the two major
areas at the moment. Lastly it will analyse the future applications that such a
technology would have, as well as the limitations and issues that it would
face.
1. Introduction
The idea
of being able to control objects with our mind seems like something that is
right out of a science fiction novel. The principle which makes BCIs possible
are the electrical activity that occurs in our brain and the changes in
electrical potential generated by brain activity with regards to stimulation (Lebedev and Nicolelis, 2006), had already been
discovered close to a 100 years ago, but it has only been in the past 20 years
where developments in the technology of BCIs have occurred.
EEG
technology is based on obtaining readings of electrical activity detected on
the scalp, of which the basis that non-invasive BCIs are developed upon.
Invasive BCIs on the other hand require surgery to implant the BCI device into,
or near, the brain. As such non-invasive BCIs offer the advantage of being more
convenient for use as well as socially acceptable and thus offer greater
applications in the future.
As we can see from revolutions such as the i-phone which
offered bought convenience to the users by being able to have many functions
and operate a multitude of applications, and the internet which introduced a
new method of communication compared to the traditional media, innovations in
such areas have more potential to be revolutionary because it is application to
everyone.
This paper will cover the introduction of BCI’s through the
developments in the use of EEG which then led further developments of BCIs
itself. It will then cover the current applications and uses of BCIs in our
industry, which would be mainly the healthcare sector and the entertainment
sector to a lesser extent. It would examine the limitations that BCI technology
faces today in its sectors of influence, and contemplate the factors that are
holding it back. Moving on to the future, the sectors whereby BCI would be able
to influence, as well as the new problems that would inevitably arise. Finally,
the paper would weigh the positive and negative effects that BCI technology
would bring about in the future.
2. Historical
Perspectives
2.1 Electroencephalography
(EEG)
In 1875, Richard Caton reported his findings about the
electrical nature of the brain, whereby the brains of dogs and apes were found
to emit electrical impulses on the surface (Caton, 1875). In 1929, citing
Caton’s research, Hans Berger was able to record brain activity from a closed
skull using EEG and found that brain activity varied according to the
functional stages of the brain (Tudor et al., 2005).
In 1934, Lowenback and Fisher demonstrated the existence of epileptiform spikes (Nielsen, 2009). In 1935 Davis,
Gibbs and Lennox discovered interictal spike waves and the patterns of seizures,
which began the field of clinical electroencephalography (Gibbs et al.,1935).
The usefulness of EEG in understanding epilepsies increased the funding
provided for EEG which bolstered further research in the area.
In the 1950s, William Grey Walter discovered a method, which
is now known as EEG topography. It allowed researches to better map the
electrical activity found on the brain. It gave researchers a more accurate view
of the location of alterations of rhythm and amplitude of the brain waves.
(Collura, 1993)
2.2 Brain Computer Interface
The first BCI was demonstrated by the same William Grey
Walter in 1964. A patient was made to press a button to advance a slide
projector while having electrodes connected to the motor area of his brain at
the same time. It was shown that the slide was able to advance before the patient
pressed the button (Dennett, 1993). This proved the possibility of creating a
desired action in a device by thinking about it without actually having to
carry out the action itself, which is what BCI is about.
Another major breakthrough was in 1969 where monkeys where
shown to be able to successfully operate a biometer arm through their neural
activity through the use of auditory and visual feedback coupled with positive
reinforcement methods such as providing food (Fetz, 1969) This showed that
electrical activity in the brain could be voluntarily modified which is the
basis for the operation of BCIs.
In the 1980s, Apostolos Georgopoulos from Johns Hopkins
University observed that the direction in which the test subject (rhesus
macaque monkeys) moved their arms had a mathematical relationship with
the resulting electrical activity occurring in their motor cortex neurons. He
was also able to identify the groups of neurons in the monkey’s brain that was
responsible for the control of motor commands (Georgopoulos et al, 1989). This gave greater insight to researchers on
the required motor actions that were needed to be thought (such as to imagine
clenching your left fist) in order to create the desired electrical activity.
Current
situation
Research on BCIs in recent years
has increased greatly. An example of the recognition that BCI have started to
garner would be the BCI award (Bci-award.com, 2013). There are now at least 100 BCI research groups
today compared to the past century as well as increased BCI related articles
and talks as well as products, most notably being the Epoc neuroheadset,
designed by Emotiv. This increase in
research arises from several factors. The prices of required computer hardware
needed for the operation of BCIs have dropped drastically over the past decade,
and the ease of access to the software needed have been increased by new
developments in information technology such as cloud computing. The advances in
related medical fields such as neuroscience have also introduced new and
improved methods of recording the electrical activity from the neurons.
The practical
use of BCI in today’s context mainly centres on the use of invasive (implanting
the BCI into the grey matter of the brain) and partially invasive (implanting
the BCI within the skull but not in the brain) techniques. The non-invasive use
of BCI which is made possible through the use of Electrocephalography although
offers greater potential to be commercialised and commonly used, is a
relatively new technology and most of the applications it has today merely
offers potential for the future instead of for practical purposes.
3.1 Invasive and
Partially-Invasive Brain-to-Computer Interface(BCI)
Despite the greater potential that the non-invasive usage of
BCI offers for the future, at the moment the practical uses of BCI such as in
the medical field relies on invasive methods of implanting the BCI devices into
either the grey matter of the brain or partially invasive methods of implanting
the devices within the skull but not in the brain itself. The use of such
methods allows for clearer signals to be received and used, because although the
electrodes in the BCIs measure the same signals used in EEG, the electrodes are
now closer to the brain. In addition to this, signals no longer need to travel
through the dura and the skull which have dampening effects on the signals.
Thus the signals obtained in invasive BCIs have “greater signal amplitude,
wider detectable frequency range, and better topographical resolution” (Mak and
Wolpaw, 2009). More information can thus be transmitted to the BCI which allows
for a greater number of movements and more complex tasks. Currently it has been
used in the vision science area through partially restoring vision to blind
people, and in neuroprosthetics by restoring movement in people such as through
computers or robotic arms (wiseGEEK, 2013). The dangers posed by Invasive BCI
methods such as the increased risk of developing scar tissues as the body’s
immune system reacts to this new device in the brain (of foreign origin) which
will cause the signal to weaken or become completely undetectable. Implantation
surgery, and the presence of a foreign object within a human’s body, can create
scar tissue in the area around the implantation. Scar tissue build-up can even
cause the relatively stronger signal to weaken or be lost altogether (Kennedy
and Bakay, 1998) which would defeat the purpose of using an invasive BCI in the
first place. There is also the risk of the BCI implant malfunctioning or lose
operational capabilities, in which case brain surgery might need to be
performed again (which presents significant risk and cost in the use of invasive
BCI’s) compared to non-invasive BCI’s used by patients which can undergo
consistent maintenance and upgrades. Because of such disadvantages, development
of BCIs have been primary non-invasive in areas such as the EU and Asia.
3.2 Non-Invasive
Brain-to-Computer Interface (BCI)
The non-invasive type of BCI is made possible through the use
of EEG whereby intracranial
electrophysiological signals are recorded using electrodes and then interpreted
by the device in order to decode the intent of subjects and the corresponding
action be used to control external devices (Lebedev and Nicolelis, 2006). The
successful use of this type of BCI has been demonstrated by a group of
researchers in the University of Minnesota where a flying robot has been
successfully controlled. The thoughts of the researcher controlling the robot
were interpreted through the use of an EEG cap and using this, the robot was
capable of being manoeuvred even through hoops (LaFleur et al.,2013). The
ability to execute such relatively complex movements is a significant one because
it was previously thought that only invasive BCI’s could do so.
The ability of non-invasive BCI to demonstrate capabilities
similar to that of invasive BCIs is crucial due to the numerous problems posed
by invasive BCIs. The need of implanting the device into the grey matter of the
brain poses significant risks due to the need for brain operation, and as such
invasive BCIs can only realistically be used as a solution for serious health
issues such as blindness and paralysis.
Non-invasive BCIs have been
successfully used in
patients both partially or severely paralyzed, allowing them to have basic
forms of communication and control in their interaction with the external
world. The EPOC headset today is an example of the application of non-invasive
BCI in today’s world, whereby it can be used as a form of art, aid medical
patients, used as a form of market research and most importantly in the area
with the biggest market, gaming.
3.3 Brain-to-Brain
interface (BBI)
The field of BBI is still in its early stages whereby
different methods of transmitting the data are being used in different studies.
A technique known as focus ultrasound (FUS) has been found that a low-energy
blast can be able to trigger activity in the tissues of the brain without any
harmful effects. Using this technique, rats were able to be controlled with the
human mind by connecting the rat to a FUS-based computer-to-brain interface
(CBI) and the human controller is linked to an EEG-based BCI. The human is then
given a specific visual cue which generates electrical brain activity of
corresponding frequency to the aforementioned visual cue. The brain activity
triggers the BCI to send a unique command to the CBI that deploys the
ultrasound blast to the motor region of the rat’s brain (the area responsible
in controlling the tail), which leads to the mouse moving its tail under the
control of the researcher. (Starr, 2013)
Another
more well-known experiment involved a researcher being able to transmit brain
signals to his colleague via the internet causing his hand to move. The
researcher attempting to “control” the subject with his brain signals was connected
to a non-invasive BCI device and the signal was sent over through the internet
to another computer which controlled a transcranial magnetic stimulation
device. and the signal was sent to the receiver
via the Internet to a computer connected to a transcranial magnetic stimulation
device (which, if used on the motor cortex of the brain, can create muscle
activity) placed on the precise area of
the brain responsible for movement for the right hand. (Rao and Stocco, 2013)
3.4
Limitations that BCI technology is facing at the moment
A significant limitation of using non-invasive BCI & BBI
by the use of electroencephalography (EEG) is the corruption of EEG data by
electrical signals that are picked up by the sensors that are actually not
coming from the cortical neurons, such
as through eye movement, blinking, and the use of surrounding muscles and
tissue as well as from insufficient contact between the electrodes and the
scalp (Rowan andTolunsky,2003).
This makes it impractical to use EEG based remote control in practical
situations because rarely will there be such controlled environments to prevent
the corruption of the electrical signals. The distortion of electrical signals
by even the use of muscles makes it hard to be used while on-the-go, which
defeats the purpose of convenience intended for the user. However there are
ways to overcome this, by reducing such distractions and “noise” through the
use of reference electrodes placed in locations where there is little cortical
activity so as to filter out correlated patterns (Ludwig et al., 2009)
Another
problem of EEG in the use of BCI is that
much of the signals are lost due to overlapping electrical activity generated
by different cortical areas as well as the passive conductance of these signals
through braise tissue, bone and skin( Lebedev
and Nicolelis,2006). As a result, the signal is insufficient for the
control of highly complex equipment. For example in the case of a prosthetic
arm, the signals necessary to control the arm , including the closed loop
controls between the brain, arm and
fingers, would be too blurred for the necessary arm movement (Dietrich et al., 2010)
Lastly,
the use of EEG based BCI requires the individuals to be trained for a
significant period of time. In research conducted by Bin he, the individuals were
trained extensively over a couple of months to be familiar with the mechanics
of flying the remote controlled droid as well as practicing the thoughts needed
to create the brainwave required to control the droid. Similarly, in the BBI
experiment by Rao and Stocco (2013), subjects had to spend time training their
minds so as to emit the correct brainwave responsible for movement of the
right-hand such that it could be recognised by the computer. This seems to be a
rather pressing issue because people are not receptive to devices which require
lots of effort to master when the current form of remote control is a more
convenient alternative.
The testing and research that is needed to be done in the
field of invasive BCIs require brain surgery and as such is limited to patients
undergoing treatment to engage in studies on a voluntary basis, as well as lab
animals such as monkeys and rats. The patients who would volunteer would
normally be fixated positive aspects of the BCI technology and would not be
fully aware regarding the side effects and complications that could arise. An
example would be in the case of Johnny Ray who was a patient who “locked-in” after
suffering a brain-stem
stroke. After receiving the implant he was able to establish a
certain level of communication through the use of computer devices such as
being able to control a computer cursor. However, he died of brain aneurysm 4
years later (Kennedy & Bakay,1998)
Because of the complexity of the human brain compared to
other animals, the use of less controversial animals (such as rats, mice, dogs
and rabbits) would be less desirable compared to using Non-human primates
because their brains share similar structural and functional features to our
human brain. However, such experimentation is controversial because the
similarity raises the possibility of the animals being able to have emotions
and feelings comparable to what we feel (Nuffieldbioethics.org, 2013). As a
result research conducted on Non-human primates has been the subject of many
animal rights attacks and limits the progress in the field of invasive BCI.
4 Future considerations
The use of BCIs offer great potential, especially its
non-invasive form due to the lesser constraints it faces in testing compared to
invasive BCIs which would require operating on animals or get medical patients
who are willing to volunteer. The ease of using noninvasive BCIs opens up many
areas that it could influence, however we would have to have to be aware of the
possible problems that could arise as a result of its possible widespread use
in the future.
4.1
Possible Applications in the future
The use of EEG based remote control provides the opportunity
for patients suffering from paralysis to operate equipment around them and
interact with their surroundings through the use of telepresence robotics, enabling
them a certain level of independence that they would otherwise not achieve. Combined with the ability to express their
thoughts in a simple form (Morrison, 2013),
this greatly improves the quality of life that these patients will enjoy.
In a study by Cruse et al (2011), a small percentage of
patients known to be in a persistent vegetative state (PVS) were shown to be
able to generate appropriate EEG responses in response to various commands. This
presence of brain activity could offer the possibility of using the EEG based
remote control which will allow these patients to lead a more normal life.
As
mentioned before, the use the EEG based remote control reduces the time lag
between thought and action. An example of where this would be useful would be
for medical robots controlled by doctors to perform operations. Robots are
being increasingly used for medical operations because they allow for smaller
and more specific movements to be carried out and access to the areas of the
body which are harder to reach (Urmc.rochester.edu,
2013).In this context, it is important that the actions carried out by
the robot are as synchronised as possible to the surgeon’s intentions because a
very low margin of error is allowed. This form of remote control makes humans
to be more connected to machines, allowing for better reaction times reduce
accidents from occurring. The potential to generate faster responses would also
be an extremely useful asset in the field of competitive gaming.
The ability to control equipment
without physically doing so offers convenience in scenarios by offering users
another method of control. It also allows for greater multitasking, an example
would be in warfare scenario whereby soldiers could be carrying weapons on both
hands and the background noise makes audio communication impossible, the
ability to communicate through the use of BCIs would be very useful. In
scenarios where the user suddenly loses his physical capabilities such as in
the case of traffic accidents or sudden health conditions like heart attacks,
EEG based remote control provides the means for the victims to seek for methods
of escape by still being able or communicate to others for help.
BBI,
which makes brain-to-brain communication possible, offers a host of
possibilities in the future. Animals, some who possess a small size capable of
entering small nooks and crannies, and others with acute sensory mechanisms,
could be used in search and rescue operations as well as environmental
surveillance. As shown in the experiment Rao and Stocco (2013), BBI can be used
to induce actions in the receiver. The ability to control movements of other
people would be a useful educational tool especially for the teaching of
complicated movements such as for skills like dancing, self-defence and sports
manoeuvres. The role of internet in BBI also opens up the potential for
physical interactions to occur through virtual networks which would be a new
element of communication
BCI,
combined with the capabilities of BBI through the internet, offers great
potential in the gaming industry because it brings in an additional physical
aspect in the world of virtual reality. Its relevance and potential for the
gaming industry is significant because of the large amount of people it could
reach out to, compared to the medical field, and thus the large market of the
gaming industry will attract resources for further development of relevant
technologies and developments in its medical applications would benefit as part
of its spill over effects.
4.2 Ethical
issues
With
further developments in BCIs and BBIs, there will inevitably be concerns
regarding ethics such as the invasion of privacy such as in cases where the
patient does not have means of expressing consent due to various forms of
communicative disorder. As a result, it is unclear if it would be ethical to
either implant BCI’s or to subject them to non-invasive BCIs which would be
tapping into their minds without prior consent, amounting to a breach of
privacy.
The
availability of medical BCI usages also creates new problems regarding the
distribution of such medical devices. The technology of BCIs is relatively
complex and requires certain qualifications and hence people in rural areas
would be left out of the benefits. With healthcare being something that is
supposed to be a public right in some countries, but yet with a limited number
of BCIs, ethical issues may arise on the priority given to different types of
patients; The patients requiring BCIs are mainly the disabled with varying
degrees of paralysis, both which could be in-born or as a result of accidents.
The
ability to control the minds of animals and use them as tools for human
activities may also outrage animal rights activists and face public
condemnation, despite the many benefits it may confer. Brain-to-brain communication
might also give rise to the possibility of mind control without consent, such
as in the case where researchers were able to control the actions of the rat.
4.3 Social Issues
The high
cost of EEG related technology means it would not be accessible to the poor.
With regards to the medical benefits that could be derived, funds channelled
into its research could arguably do more good in being used to address the
basic needs and concerns of the common man instead of providing specialised
healthcare that only the rich could afford.
Should
non-invasive forms of EEG become popular in the future, its popularity might
compel the government to make it compulsory for everyone for administrative
purposes due to the convenience it brings for everyone to be under the same
health plan, even though some people may actually be opposed to it.
The
usage of non-invasive BCIs in the gaming industry, although poses many benefits
such as offering gamers a whole new experience, can pose significant threats.
The proliferation of BCIs may gradually lead to getting information from the
internet itself (instead of only sending information) which would definitely
lead to new kinds of internet viruses that may have disastrous consequences
given that they can “hack” into a person’s brain, not just his computer. The
potential for such activity to occur is shown through studies which show that
the brain activity can be influenced by external sources, such as the internet.
In other words transference of information can be two-way as displayed by the
experiment of Rao & Stocco (2013). This
could mean that should the use of BCIs be mainstream, large amounts of people
could be “brainwashed” by a virus transferred by BCI which will lead the to
birth of new elements of cybercrime and terrorism.
4.4 Issues
relating to privacy
Present
studies have also discovered the possibility of obtaining the private
information of its users through the use of a cheap BCI gaming device (Martinovic et al.,2012). Although such extraction of
information could be potentially useful in webpages and search devices to
greater understand the user’s needs, this would amount to an invasion of
privacy.
The issue of privacy has been raised in recent years in
response to counter terrorism measures such as x-ray scanning in the airports.
Despite complaints that their privacy was breached, full-body scanners are used
in the airports of many countries (Crislip, 2013). Similar disputes are likely
to arise when BCIs could introduce an additional layer of detection but again
at the expense of revealing personal information of the travellers. Privacy
would again most likely be compromised for the purpose of security. This could
raise the question of the legal liabilities of an intention to commit a crime,
even though the action has not been committed.
4.5 Legal Issues
Despite the convenience that non-invasive BCIs offer,
invasive BCIs have an advantage of being used in secret. BCIs have been shown
to be capable of enhancing motor capabilities and brain performance in healthy
individuals (Warwick, (2011). Controversy would arise in the area of
competitive sports because this would constitute to artificial enhancement.
Conclusion
There has been much recent developments and
interest in the field of BCIs which would continue into the future. The realm
of non-invasive BCIs offers significant advantages as its ease of use offers
the potential for it to be applied not just in the medical field, but in areas
for the common man such as communication, gaming, as well as possible social
applications. Hence in my opinion non-invasive BCIs will become a revolution in
the future. However, it also presents to us many new problems should it become
a reality. Issues of mind control become an actual reality which we could face
decades later, and we may need to ask the question whether BCI, with the risk
that it offers, should be commercialised and be used for recreation or for the
sake of convenience instead of the main medical usages currently.
References
Bci-award.com (2013). bci-award.com.
[online] Retrieved from: http://www.bci-award.com/ [Accessed: 24 Oct 2013].
Caton, R. (1875). Electrical currents of the brain. The Journal of Nervous and Mental
Disease, 2(4), 610.
Collura, T. F. (1993). History and evolution of
electroencephalographic instruments and techniques. Journal of Clinical Neurophysiology, 10(4), 476-504.
Crislip, K. (2013). Which
Airports Have Full Body Scanners?. [online] Retrieved from:
http://studenttravel.about.com/od/planes/f/Which-Airports-Have-Full-Body-Scanners.htm
[Accessed: 25 Oct 2013].
Cruse, D., Chennu, S., Chatelle, C., Bekinschtein, T. A.,
Fernández-Espejo, D., Pickard, J. D., ... & Owen, A. M. (2012). Bedside
detection of awareness in the vegetative state: a cohort study. The Lancet, 378(9809), 2088-2094.
Dennett, D. C. (1993). Consciousness
explained. Penguin UK.Dietrich, Dietmar, et al. "Limitations,
possibilities and implications of Brain-Computer Interfaces." Human System Interactions (HSI),
2010 3rd Conference on. IEEE, 2010.
Fetz,
E. E. (1969). "Operant Conditioning of Cortical Unit Activity". Science 163 (3870):
955–8. doi:10.1126/science.163.3870.955
Georgopoulos, A.; Lurito, J.; Petrides, M; Schwartz, A.;
Massey, J. (1989). "Mental rotation of the neuronal population
vector". Science 243 (4888): 234–6. doi:10.1126/science.2911737
Gibbs, F. A., Davis, H., & Lennox, W. G. (1935). The
electro-encephalogram in epilepsy and in conditions of impaired consciousness. Archives of neurology and
psychiatry, 34(6),
1133.
Kennedy, P. R., & Bakay, R. A. (1998). Restoration of
neural output from a paralyzed patient by a direct brain connection. Neuroreport, 9(8), 1707-1711.
LaFleur, K., Cassady, K., Doud, A., Shades, K., Rogin, E.,
& He, B. (2013). Quadcopter control in three-dimensional space using a
noninvasive motor imagery-based brain–computer interface. Journal of neural engineering, 10(4), 046003.Lebedev, M. A.,
& Nicolelis, M. A. (2006). Brain–machine interfaces: past, present and
future. TRENDS in
Neurosciences, 29(9),
536-546.
Ludwig, Kip A., et al. "Using a common average reference
to improve cortical neuron recordings from microelectrode arrays." Journal of neurophysiology 101.3 (2009): 1679-1689. Biomedical Engineering, IEEE
Reviews in, 2,
187-199.
Martinovic, I., Davies, D., Frank, M., Perito, D., Ros, T.,
& Song, D. (2012). On the feasibility of side-channel attacks with brain-computer
interfaces. In 21st USENIX
Security Symp.
Morrison, D. (2013). Reading
minds: University research may help those who cannot move : UMNews : University
of Minnesota. [online] Retrieved from: http://www1.umn.edu/news/features/2004/UR_28115_REGION1.html [Accessed: 13
Oct 2013]
Nielsen, B. (2009). Towards
24-7 Brain Mapping Technology (Doctoral
dissertation, Technical University of Denmark, DTU, DK-2800 Kgs. Lyngby,
Denmark).
Nuffieldbioethics.org (2013). Animal research | Nuffield Council
on Bioethics. [online] Retrieved from:
http://www.nuffieldbioethics.org/animal-research [Accessed: 24 Oct 2013].
Rao, R. & Stocco, A. (2013). Brain2Brain Study: Experiment.
[online] Retrieved from: http://homes.cs.washington.edu/~rao/brain2brain/experiment.html [Accessed: 13 Oct 2013]
Rowan, A. J., & Tolunsky, E. (2003). Primer of EEG: With a Mini-atlas.
Butterworth-Heinemann.
Starr, M. (2013). Control
a rat with your mind. [online] Retrieved from:
http://www.cnet.com.au/control-a-rat-with-your-mind-339345047.htm [Accessed: 12
Oct 2013]
Tudor, M., Tudor, L., & Tudor, K. I. (2005). Hans Berger
(1873-1941)--the history of electroencephalography]. Acta medica Croatica: c̆asopis
Hravatske akademije medicinskih znanosti, 59(4),
307.
Urmc.rochester.edu (2013). Robots:
Now Appearing in an Operating Room Near You - Online Medical Encyclopedia -
University of Rochester Medical Center. [online] Retrieved from:
http://www.urmc.rochester.edu/encyclopedia/content.aspx?ContentTypeID=1&ContentID=24247
[Accessed: 13 Oct 2013].
Warwick, K. (2011). Possibilities for Human Enhancement
through BCI. InUKIERI Workshop (p.
7).
wiseGEEK (2013). What
are Some Brain-Computer Interfaces (BCI)?. [online] Retrieved from:
http://www.wisegeek.com/what-are-some-brain-computer-interfaces-bci.htm
[Accessed: 13 Oct 2013].
No comments:
Post a Comment