<s>
A	O
brain	B-Application
–	I-Application
computer	I-Application
interface	I-Application
(	O
BCI	O
)	O
,	O
sometimes	O
called	O
a	O
brain	B-Application
–	I-Application
machine	I-Application
interface	I-Application
(	O
BMI	O
)	O
or	O
smartbrain	O
,	O
is	O
a	O
direct	O
communication	O
pathway	O
between	O
the	O
brain	O
's	O
electrical	O
activity	O
and	O
an	O
external	O
device	O
,	O
most	O
commonly	O
a	O
computer	O
or	O
robotic	O
limb	O
.	O
</s>
<s>
Implementations	O
of	O
BCIs	O
range	O
from	O
non-invasive	O
(	O
EEG	B-Application
,	O
MEG	B-Algorithm
,	O
EOG	O
,	O
MRI	B-Algorithm
)	O
and	O
partially	O
invasive	O
(	O
ECoG	O
and	O
endovascular	O
)	O
to	O
invasive	O
(	O
microelectrode	O
array	O
)	O
,	O
based	O
on	O
how	O
close	O
electrodes	O
get	O
to	O
brain	O
tissue	O
.	O
</s>
<s>
Vidal	O
's	O
1973	O
paper	O
marks	O
the	O
first	O
appearance	O
of	O
the	O
expression	O
brain	B-Application
–	I-Application
computer	I-Application
interface	I-Application
in	O
scientific	O
literature	O
.	O
</s>
<s>
Recently	O
,	O
studies	O
in	O
human-computer	O
interaction	O
via	O
the	O
application	O
of	O
machine	O
learning	O
to	O
statistical	O
temporal	O
features	O
extracted	O
from	O
the	O
frontal	O
lobe	O
(	O
EEG	B-Application
brainwave	I-Application
)	O
data	O
has	O
had	O
high	O
levels	O
of	O
success	O
in	O
classifying	O
mental	O
states	O
(	O
Relaxed	O
,	O
Neutral	O
,	O
Concentrating	O
)	O
,	O
mental	O
emotional	O
states	O
(	O
Negative	O
,	O
Neutral	O
,	O
Positive	O
)	O
,	O
and	O
thalamocortical	O
dysrhythmia	O
.	O
</s>
<s>
The	O
history	O
of	O
brain	B-Application
–	I-Application
computer	I-Application
interfaces	I-Application
(	O
BCIs	O
)	O
starts	O
with	O
Hans	O
Berger	O
's	O
discovery	O
of	O
the	O
electrical	O
activity	O
of	O
the	O
human	O
brain	O
and	O
the	O
development	O
of	O
electroencephalography	B-Application
(	O
EEG	B-Application
)	O
.	O
</s>
<s>
In	O
1924	O
Berger	O
was	O
the	O
first	O
to	O
record	O
human	O
brain	B-Application
activity	I-Application
by	O
means	O
of	O
EEG	B-Application
.	O
</s>
<s>
Berger	O
was	O
able	O
to	O
identify	O
oscillatory	B-Algorithm
activity	I-Algorithm
,	O
such	O
as	O
Berger	O
's	O
wave	O
or	O
the	O
alpha	O
wave	O
(	O
8	O
–	O
13Hz	O
)	O
,	O
by	O
analyzing	O
EEG	B-Application
traces	O
.	O
</s>
<s>
Berger	O
analyzed	O
the	O
interrelation	O
of	O
alternations	O
in	O
his	O
EEG	B-Application
wave	O
diagrams	O
with	O
brain	O
diseases	O
.	O
</s>
<s>
EEGs	B-Application
permitted	O
completely	O
new	O
possibilities	O
for	O
the	O
research	O
of	O
human	O
brain	B-Application
activities	I-Application
.	O
</s>
<s>
Although	O
the	O
term	O
had	O
not	O
yet	O
been	O
coined	O
,	O
one	O
of	O
the	O
earliest	O
examples	O
of	O
a	O
working	O
brain-machine	B-Application
interface	I-Application
was	O
the	O
piece	O
Music	O
for	O
Solo	O
Performer	O
(	O
1965	O
)	O
by	O
the	O
American	O
composer	O
Alvin	O
Lucier	O
.	O
</s>
<s>
The	O
piece	O
makes	O
use	O
of	O
EEG	B-Application
and	O
analog	O
signal	O
processing	O
hardware	O
(	O
filters	O
,	O
amplifiers	O
,	O
and	O
a	O
mixing	O
board	O
)	O
to	O
stimulate	O
acoustic	O
percussion	O
instruments	O
.	O
</s>
<s>
To	O
perform	O
the	O
piece	O
one	O
must	O
produce	O
alpha	B-Application
waves	I-Application
and	O
thereby	O
"	O
play	O
"	O
the	O
various	O
percussion	O
instruments	O
via	O
loudspeakers	O
which	O
are	O
placed	O
near	O
or	O
directly	O
on	O
the	O
instruments	O
themselves	O
.	O
</s>
<s>
A	O
review	O
pointed	O
out	O
that	O
Vidal	O
's	O
1973	O
paper	O
stated	O
the	O
"	O
BCI	O
challenge	O
"	O
of	O
controlling	O
external	O
objects	O
using	O
EEG	B-Application
signals	O
,	O
and	O
especially	O
use	O
of	O
Contingent	B-Algorithm
Negative	I-Algorithm
Variation	I-Algorithm
(	O
CNV	O
)	O
potential	O
as	O
a	O
challenge	O
for	O
BCI	O
control	O
.	O
</s>
<s>
It	O
was	O
a	O
noninvasive	O
EEG	B-Application
(	O
actually	O
Visual	O
Evoked	O
Potentials	O
(	O
VEP	O
)	O
)	O
control	O
of	O
a	O
cursor-like	O
graphical	O
object	O
on	O
a	O
computer	O
screen	O
.	O
</s>
<s>
In	O
1988	O
,	O
a	O
report	O
was	O
given	O
on	O
noninvasive	O
EEG	B-Application
control	O
of	O
a	O
physical	O
object	O
,	O
a	O
robot	O
.	O
</s>
<s>
The	O
experiment	O
described	O
was	O
EEG	B-Application
control	O
of	O
multiple	O
start-stop-restart	O
of	O
the	O
robot	O
movement	O
,	O
along	O
an	O
arbitrary	O
trajectory	O
defined	O
by	O
a	O
line	O
drawn	O
on	O
a	O
floor	O
.	O
</s>
<s>
This	O
1988	O
report	O
written	O
by	O
Stevo	O
Bozinovski	O
,	O
Mihail	O
Sestakov	O
,	O
and	O
Liljana	O
Bozinovska	O
was	O
the	O
first	O
one	O
about	O
a	O
robot	O
control	O
using	O
EEG	B-Application
.	O
</s>
<s>
In	O
1990	O
,	O
a	O
report	O
was	O
given	O
on	O
a	O
closed	O
loop	O
,	O
bidirectional	O
adaptive	O
BCI	O
controlling	O
computer	O
buzzer	O
by	O
an	O
anticipatory	O
brain	O
potential	O
,	O
the	O
Contingent	B-Algorithm
Negative	I-Algorithm
Variation	I-Algorithm
(	O
CNV	O
)	O
potential	O
.	O
</s>
<s>
Monkeys	O
have	O
navigated	O
computer	B-Application
cursors	I-Application
on	O
screen	O
and	O
commanded	O
robotic	O
arms	O
to	O
perform	O
simple	O
tasks	O
simply	O
by	O
thinking	O
about	O
the	O
task	O
and	O
seeing	O
the	O
visual	O
feedback	O
,	O
but	O
without	O
any	O
motor	O
output	O
.	O
</s>
<s>
Phillip	O
Kennedy	O
(	O
who	O
later	O
founded	O
Neural	O
Signals	O
in	O
1987	O
)	O
and	O
colleagues	O
built	O
the	O
first	O
intracortical	O
brain	B-Application
–	I-Application
computer	I-Application
interface	I-Application
by	O
implanting	O
neurotrophic-cone	O
electrodes	O
into	O
monkeys	O
.	O
</s>
<s>
thumb|Yang	O
Dan	O
and	O
colleagues	O
 '	O
recordings	O
of	O
cat	O
vision	O
using	O
a	O
BCI	O
implanted	O
in	O
the	O
lateral	O
geniculate	O
nucleus	O
(	O
top	O
row	O
:	O
original	O
image	O
;	O
bottom	O
row	O
:	O
recording	O
)	O
In	O
1999	O
,	O
researchers	O
led	O
by	O
Yang	O
Dan	O
at	O
the	O
University	O
of	O
California	O
,	O
Berkeley	O
decoded	O
neuronal	B-Algorithm
firings	I-Algorithm
to	O
reproduce	O
images	O
seen	O
by	O
cats	O
.	O
</s>
<s>
After	O
conducting	O
initial	O
studies	O
in	O
rats	O
during	O
the	O
1990s	O
,	O
Nicolelis	O
and	O
his	O
colleagues	O
developed	O
BCIs	O
that	O
decoded	O
brain	B-Application
activity	I-Application
in	O
owl	O
monkeys	O
and	O
used	O
the	O
devices	O
to	O
reproduce	O
monkey	O
movements	O
in	O
robotic	O
arms	O
.	O
</s>
<s>
By	O
2000	O
,	O
the	O
group	O
succeeded	O
in	O
building	O
a	O
BCI	O
that	O
reproduced	O
owl	O
monkey	O
movements	O
while	O
the	O
monkey	O
operated	O
a	O
joystick	B-Device
or	O
reached	O
for	O
food	O
.	O
</s>
<s>
The	O
BCI	O
operated	O
in	O
real	O
time	O
and	O
could	O
also	O
control	O
a	O
separate	O
robot	O
remotely	O
over	O
Internet	B-Protocol
Protocol	I-Protocol
.	O
</s>
<s>
The	O
monkeys	O
were	O
trained	O
to	O
reach	O
and	O
grasp	O
objects	O
on	O
a	O
computer	O
screen	O
by	O
manipulating	O
a	O
joystick	B-Device
while	O
corresponding	O
movements	O
by	O
a	O
robot	O
arm	O
were	O
hidden	O
.	O
</s>
<s>
John	O
Donoghue	O
's	O
lab	O
at	O
the	O
Carney	O
Institute	O
reported	O
training	O
rhesus	O
monkeys	O
to	O
use	O
a	O
BCI	O
to	O
track	O
visual	O
targets	O
on	O
a	O
computer	O
screen	O
(	O
closed-loop	O
BCI	O
)	O
with	O
or	O
without	O
assistance	O
of	O
a	O
joystick	B-Device
.	O
</s>
<s>
Their	O
BCI	O
used	O
high-density	O
electrocorticography	O
to	O
tap	O
neural	O
activity	O
from	O
a	O
patient	O
's	O
brain	O
and	O
used	O
deep	B-Algorithm
learning	I-Algorithm
methods	O
to	O
synthesize	O
speech	O
.	O
</s>
<s>
BCI2000	O
has	O
been	O
in	O
development	O
since	O
2000	O
in	O
a	O
project	O
led	O
by	O
the	O
Brain	B-Application
–	I-Application
Computer	I-Application
Interface	I-Application
R&D	O
Program	O
at	O
the	O
Wadsworth	O
Center	O
of	O
the	O
New	O
York	O
State	O
Department	O
of	O
Health	O
in	O
Albany	O
,	O
New	O
York	O
,	O
United	O
States	O
.	O
</s>
<s>
The	O
use	O
of	O
BMIs	O
has	O
also	O
led	O
to	O
a	O
deeper	O
understanding	O
of	O
neural	B-Architecture
networks	I-Architecture
and	O
the	O
central	O
nervous	O
system	O
.	O
</s>
<s>
In	O
a	O
secondary	O
,	O
implicit	O
control	O
loop	O
the	O
computer	O
system	O
adapts	O
to	O
its	O
user	O
improving	O
its	O
usability	B-General_Concept
in	O
general	O
.	O
</s>
<s>
The	O
Annual	O
BCI	O
Research	O
Award	O
is	O
awarded	O
in	O
recognition	O
of	O
outstanding	O
and	O
innovative	O
research	O
in	O
the	O
field	O
of	O
Brain-Computer	B-Application
Interfaces	I-Application
.	O
</s>
<s>
One	O
of	O
the	O
first	O
scientists	O
to	O
produce	O
a	O
working	O
brain	B-Application
interface	I-Application
to	O
restore	O
sight	O
was	O
private	O
researcher	O
William	O
Dobelle	O
.	O
</s>
<s>
This	O
also	O
required	O
him	O
to	O
be	O
hooked	O
up	O
to	O
a	O
mainframe	B-Architecture
computer	I-Architecture
,	O
but	O
shrinking	O
electronics	O
and	O
faster	O
computers	O
made	O
his	O
artificial	O
eye	O
more	O
portable	O
and	O
now	O
enable	O
him	O
to	O
perform	O
simple	O
tasks	O
unassisted	O
.	O
</s>
<s>
Ray	O
's	O
implant	O
was	O
installed	O
in	O
1998	O
and	O
he	O
lived	O
long	O
enough	O
to	O
start	O
working	O
with	O
the	O
implant	O
,	O
eventually	O
learning	O
to	O
control	O
a	O
computer	O
cursor	B-Application
;	O
he	O
died	O
in	O
2002	O
of	O
a	O
brain	O
aneurysm	O
.	O
</s>
<s>
Implanted	O
in	O
Nagle	O
's	O
right	O
precentral	O
gyrus	O
(	O
area	O
of	O
the	O
motor	O
cortex	O
for	O
arm	O
movement	O
)	O
,	O
the	O
96-electrode	O
BrainGate	O
implant	O
allowed	O
Nagle	O
to	O
control	O
a	O
robotic	O
arm	O
by	O
thinking	O
about	O
moving	O
his	O
hand	O
as	O
well	O
as	O
a	O
computer	O
cursor	B-Application
,	O
lights	O
and	O
TV	O
.	O
</s>
<s>
One	O
year	O
later	O
,	O
professor	O
Jonathan	O
Wolpaw	O
received	O
the	O
prize	O
of	O
the	O
Altran	O
Foundation	O
for	O
Innovation	O
to	O
develop	O
a	O
Brain	B-Application
Computer	I-Application
Interface	I-Application
with	O
electrodes	O
located	O
on	O
the	O
surface	O
of	O
the	O
skull	O
,	O
instead	O
of	O
directly	O
in	O
the	O
brain	O
.	O
</s>
<s>
The	O
participant	O
imagined	O
moving	O
his	O
hand	O
to	O
write	O
letters	O
,	O
and	O
the	O
system	O
performed	O
handwriting	O
recognition	O
on	O
electrical	O
signals	O
detected	O
in	O
the	O
motor	O
cortex	O
,	O
utilizing	O
hidden	O
Markov	O
models	O
and	O
recurrent	O
neural	B-Architecture
networks	I-Architecture
for	O
decoding	O
.	O
</s>
<s>
There	O
exist	O
a	O
number	O
of	O
technical	O
challenges	O
to	O
recording	O
brain	B-Application
activity	I-Application
with	O
invasive	O
BCIs	O
.	O
</s>
<s>
Advances	O
in	O
CMOS	B-Device
technology	O
are	O
pushing	O
and	O
enabling	O
integrated	O
,	O
invasive	O
BCI	O
designs	O
with	O
smaller	O
size	O
,	O
lower	O
power	O
requirements	O
,	O
and	O
higher	O
signal	O
acquisition	O
capabilities	O
.	O
</s>
<s>
Invasive	O
BCIs	O
involve	O
electrodes	O
that	O
penetrate	O
brain	O
tissue	O
in	O
an	O
attempt	O
to	O
record	O
action	B-Algorithm
potential	I-Algorithm
signals	O
(	O
also	O
known	O
as	O
spikes	B-Algorithm
)	O
from	O
individual	O
,	O
or	O
small	O
groups	O
of	O
,	O
neurons	O
near	O
the	O
electrode	O
.	O
</s>
<s>
While	O
intracellular	O
recordings	O
of	O
neurons	O
reveal	O
action	B-Algorithm
potential	I-Algorithm
voltages	O
on	O
the	O
scale	O
of	O
hundreds	O
of	O
millivolts	O
,	O
chronic	O
invasive	O
BCIs	O
rely	O
on	O
recording	O
extracellular	O
voltages	O
which	O
typically	O
are	O
three	O
orders	O
of	O
magnitude	O
smaller	O
,	O
existing	O
at	O
hundreds	O
of	O
microvolts	O
.	O
</s>
<s>
Because	O
a	O
typical	O
neuron	O
action	B-Algorithm
potential	I-Algorithm
lasts	O
for	O
one	O
millisecond	O
,	O
BCIs	O
measuring	O
spikes	B-Algorithm
must	O
have	O
sampling	O
rates	O
ranging	O
from	O
300Hz	O
to	O
5kHz	O
.	O
</s>
<s>
They	O
produce	O
better	O
resolution	B-General_Concept
signals	O
than	O
non-invasive	O
BCIs	O
where	O
the	O
bone	O
tissue	O
of	O
the	O
cranium	O
deflects	O
and	O
deforms	O
signals	O
and	O
have	O
a	O
lower	O
risk	O
of	O
forming	O
scar-tissue	O
in	O
the	O
brain	O
than	O
fully	O
invasive	O
BCIs	O
.	O
</s>
<s>
In	O
November	O
2020	O
,	O
two	O
participants	O
with	O
amyotrophic	O
lateral	O
sclerosis	O
were	O
able	O
to	O
wirelessly	O
control	O
an	O
operating	O
system	O
to	O
text	O
,	O
email	O
,	O
shop	O
,	O
and	O
bank	O
using	O
direct	O
thought	O
through	O
the	O
Stentrode	O
brain-computer	B-Application
interface	I-Application
,	O
marking	O
the	O
first	O
time	O
a	O
brain-computer	B-Application
interface	I-Application
was	O
implanted	O
via	O
the	O
patient	O
's	O
blood	O
vessels	O
,	O
eliminating	O
the	O
need	O
for	O
open	O
brain	O
surgery	O
.	O
</s>
<s>
Electrocorticography	O
(	O
ECoG	O
)	O
measures	O
the	O
electrical	O
activity	O
of	O
the	O
brain	O
taken	O
from	O
beneath	O
the	O
skull	O
in	O
a	O
similar	O
way	O
to	O
non-invasive	O
electroencephalography	B-Application
,	O
but	O
the	O
electrodes	O
are	O
embedded	O
in	O
a	O
thin	O
plastic	O
pad	O
that	O
is	O
placed	O
above	O
the	O
cortex	O
,	O
beneath	O
the	O
dura	O
mater	O
.	O
</s>
<s>
In	O
a	O
later	O
trial	O
,	O
the	O
researchers	O
enabled	O
a	O
teenage	O
boy	O
to	O
play	O
Space	B-Application
Invaders	I-Application
using	O
his	O
ECoG	O
implant	O
.	O
</s>
<s>
ECoG	O
is	O
a	O
very	O
promising	O
intermediate	O
BCI	O
modality	O
because	O
it	O
has	O
higher	O
spatial	O
resolution	B-General_Concept
,	O
better	O
signal-to-noise	O
ratio	O
,	O
wider	O
frequency	O
range	O
,	O
and	O
less	O
training	O
requirements	O
than	O
scalp-recorded	O
EEG	B-Application
,	O
and	O
at	O
the	O
same	O
time	O
has	O
lower	O
technical	O
difficulty	O
,	O
lower	O
clinical	O
risk	O
,	O
and	O
may	O
have	O
superior	O
long-term	O
stability	O
than	O
intracortical	O
single-neuron	O
recording	O
.	O
</s>
<s>
Their	O
study	O
achieved	O
word	O
error	O
rates	O
of	O
3%	O
(	O
a	O
marked	O
improvement	O
from	O
prior	O
publications	O
)	O
utilizing	O
an	O
encoder-decoder	O
neural	B-Architecture
network	I-Architecture
,	O
which	O
translated	O
ECoG	O
data	O
into	O
one	O
of	O
fifty	O
sentences	O
composed	O
of	O
250	O
unique	O
words	O
.	O
</s>
<s>
There	O
have	O
also	O
been	O
experiments	O
in	O
humans	O
using	O
non-invasive	O
neuroimaging	B-Algorithm
technologies	O
as	O
interfaces	O
.	O
</s>
<s>
The	O
substantial	O
majority	O
of	O
published	O
BCI	O
work	O
involves	O
noninvasive	O
EEG-based	O
BCIs	O
.	O
</s>
<s>
Noninvasive	O
EEG-based	O
technologies	O
and	O
interfaces	O
have	O
been	O
used	O
for	O
a	O
much	O
broader	O
variety	O
of	O
applications	O
.	O
</s>
<s>
Although	O
EEG-based	O
interfaces	O
are	O
easy	O
to	O
wear	O
and	O
do	O
not	O
require	O
surgery	O
,	O
they	O
have	O
relatively	O
poor	O
spatial	O
resolution	B-General_Concept
and	O
cannot	O
effectively	O
use	O
higher-frequency	O
signals	O
because	O
the	O
skull	O
dampens	O
signals	O
,	O
dispersing	O
and	O
blurring	O
the	O
electromagnetic	O
waves	O
created	O
by	O
the	O
neurons	O
.	O
</s>
<s>
EEG-based	O
interfaces	O
also	O
require	O
some	O
time	O
and	O
effort	O
prior	O
to	O
each	O
usage	O
session	O
,	O
whereas	O
non-EEG-based	O
ones	O
,	O
as	O
well	O
as	O
invasive	O
ones	O
require	O
no	O
prior-usage	O
training	O
.	O
</s>
<s>
A	O
2016	O
article	O
described	O
an	O
entirely	O
new	O
communication	O
device	O
and	O
non-EEG-based	O
human-computer	O
interface	O
,	O
which	O
requires	O
no	O
visual	O
fixation	O
,	O
or	O
ability	O
to	O
move	O
the	O
eyes	O
at	O
all	O
.	O
</s>
<s>
In	O
2014	O
and	O
2017	O
,	O
a	O
BCI	O
using	O
functional	B-Algorithm
near-infrared	I-Algorithm
spectroscopy	I-Algorithm
for	O
"	O
locked-in	O
"	O
patients	O
with	O
amyotrophic	O
lateral	O
sclerosis	O
(	O
ALS	O
)	O
was	O
able	O
to	O
restore	O
some	O
basic	O
ability	O
of	O
the	O
patients	O
to	O
communicate	O
with	O
other	O
people	O
.	O
</s>
<s>
After	O
the	O
BCI	O
challenge	O
was	O
stated	O
by	O
Vidal	O
in	O
1973	O
,	O
the	O
initial	O
reports	O
on	O
non-invasive	O
approach	O
included	O
control	O
of	O
a	O
cursor	B-Application
in	O
2D	O
using	O
VEP	O
(	O
Vidal	O
1977	O
)	O
,	O
control	O
of	O
a	O
buzzer	O
using	O
CNV	O
(	O
Bozinovska	O
et	O
al	O
.	O
</s>
<s>
1988	O
,	O
1990	O
)	O
,	O
control	O
of	O
a	O
physical	O
object	O
,	O
a	O
robot	O
,	O
using	O
a	O
brain	B-Algorithm
rhythm	I-Algorithm
(	O
alpha	O
)	O
(	O
Bozinovski	O
et	O
al	O
.	O
</s>
<s>
1988	O
)	O
,	O
control	O
of	O
a	O
text	O
written	O
on	O
a	O
screen	O
using	O
P300	B-Algorithm
(	O
Farwell	O
and	O
Donchin	O
,	O
1988	O
)	O
.	O
</s>
<s>
In	O
the	O
early	O
days	O
of	O
BCI	O
research	O
,	O
another	O
substantial	O
barrier	O
to	O
using	O
electroencephalography	B-Application
(	O
EEG	B-Application
)	O
as	O
a	O
brain	B-Application
–	I-Application
computer	I-Application
interface	I-Application
was	O
the	O
extensive	O
training	O
required	O
before	O
users	O
can	O
work	O
the	O
technology	O
.	O
</s>
<s>
For	O
example	O
,	O
in	O
experiments	O
beginning	O
in	O
the	O
mid-1990s	O
,	O
Niels	O
Birbaumer	O
at	O
the	O
University	O
of	O
Tübingen	O
in	O
Germany	O
trained	O
severely	O
paralysed	O
people	O
to	O
self-regulate	O
the	O
slow	O
cortical	O
potentials	O
in	O
their	O
EEG	B-Application
to	O
such	O
an	O
extent	O
that	O
these	O
signals	O
could	O
be	O
used	O
as	O
a	O
binary	O
signal	O
to	O
control	O
a	O
computer	O
cursor	B-Application
.	O
</s>
<s>
The	O
experiment	O
saw	O
ten	O
patients	O
trained	O
to	O
move	O
a	O
computer	O
cursor	B-Application
by	O
controlling	O
their	O
brainwaves	B-Algorithm
.	O
</s>
<s>
The	O
process	O
was	O
slow	O
,	O
requiring	O
more	O
than	O
an	O
hour	O
for	O
patients	O
to	O
write	O
100	O
characters	O
with	O
the	O
cursor	B-Application
,	O
while	O
training	O
often	O
took	O
many	O
months	O
.	O
</s>
<s>
Another	O
research	O
parameter	O
is	O
the	O
type	O
of	O
oscillatory	B-Algorithm
activity	I-Algorithm
that	O
is	O
measured	O
.	O
</s>
<s>
Together	O
with	O
Birbaumer	O
and	O
Jonathan	O
Wolpaw	O
at	O
New	O
York	O
State	O
University	O
they	O
focused	O
on	O
developing	O
technology	O
that	O
would	O
allow	O
users	O
to	O
choose	O
the	O
brain	O
signals	O
they	O
found	O
easiest	O
to	O
operate	O
a	O
BCI	O
,	O
including	O
mu	B-Algorithm
and	O
beta	B-Algorithm
rhythms	I-Algorithm
.	O
</s>
<s>
A	O
further	O
parameter	O
is	O
the	O
method	O
of	O
feedback	O
used	O
and	O
this	O
is	O
shown	O
in	O
studies	O
of	O
P300	B-Algorithm
signals	O
.	O
</s>
<s>
Patterns	O
of	O
P300	B-Algorithm
waves	O
are	O
generated	O
involuntarily	O
(	O
stimulus-feedback	B-Algorithm
)	O
when	O
people	O
see	O
something	O
they	O
recognize	O
and	O
may	O
allow	O
BCIs	O
to	O
decode	O
categories	O
of	O
thoughts	O
without	O
training	O
patients	O
first	O
.	O
</s>
<s>
By	O
contrast	O
,	O
the	O
biofeedback	O
methods	O
described	O
above	O
require	O
learning	O
to	O
control	O
brainwaves	B-Algorithm
so	O
the	O
resulting	O
brain	B-Application
activity	I-Application
can	O
be	O
detected	O
.	O
</s>
<s>
In	O
2005	O
it	O
was	O
reported	O
research	O
on	O
EEG	B-Application
emulation	O
of	O
digital	O
control	O
circuits	O
for	O
BCI	O
,	O
with	O
example	O
of	O
a	O
CNV	O
flip-flop	O
.	O
</s>
<s>
In	O
2009	O
it	O
was	O
reported	O
noninvasive	O
EEG	B-Application
control	O
of	O
a	O
robotic	O
arm	O
using	O
a	O
CNV	O
flip-flop	O
.	O
</s>
<s>
In	O
2015	O
it	O
was	O
described	O
EEG-emulation	O
of	O
a	O
Schmitt	O
trigger	O
,	O
flip-flop	O
,	O
demultiplexer	O
,	O
and	O
modem	O
.	O
</s>
<s>
While	O
an	O
EEG	B-Application
based	O
brain-computer	B-Application
interface	I-Application
has	O
been	O
pursued	O
extensively	O
by	O
a	O
number	O
of	O
research	O
labs	O
,	O
recent	O
advancements	O
made	O
by	O
Bin	O
He	O
and	O
his	O
team	O
at	O
the	O
University	O
of	O
Minnesota	O
suggest	O
the	O
potential	O
of	O
an	O
EEG	B-Application
based	O
brain-computer	B-Application
interface	I-Application
to	O
accomplish	O
tasks	O
close	O
to	O
invasive	O
brain-computer	B-Application
interface	I-Application
.	O
</s>
<s>
Using	O
advanced	O
functional	O
neuroimaging	B-Algorithm
including	O
BOLD	O
functional	B-Algorithm
MRI	I-Algorithm
and	O
EEG	B-Application
source	O
imaging	O
,	O
Bin	O
He	O
and	O
co-workers	O
identified	O
the	O
co-variation	O
and	O
co-localization	O
of	O
electrophysiological	O
and	O
hemodynamic	O
signals	O
induced	O
by	O
motor	O
imagination	O
.	O
</s>
<s>
Refined	O
by	O
a	O
neuroimaging	B-Algorithm
approach	O
and	O
by	O
a	O
training	O
protocol	O
,	O
Bin	O
He	O
and	O
co-workers	O
demonstrated	O
the	O
ability	O
of	O
a	O
non-invasive	O
EEG	B-Application
based	O
brain-computer	B-Application
interface	I-Application
to	O
control	O
the	O
flight	O
of	O
a	O
virtual	O
helicopter	O
in	O
3-dimensional	O
space	O
,	O
based	O
upon	O
motor	O
imagination	O
.	O
</s>
<s>
In	O
addition	O
to	O
a	O
brain-computer	B-Application
interface	I-Application
based	O
on	O
brain	B-Algorithm
waves	I-Algorithm
,	O
as	O
recorded	O
from	O
scalp	O
EEG	B-Application
electrodes	O
,	O
Bin	O
He	O
and	O
co-workers	O
explored	O
a	O
virtual	O
EEG	B-Application
signal-based	O
brain-computer	B-Application
interface	I-Application
by	O
first	O
solving	O
the	O
EEG	B-Application
inverse	O
problem	O
and	O
then	O
used	O
the	O
resulting	O
virtual	O
EEG	B-Application
for	O
brain-computer	B-Application
interface	I-Application
tasks	O
.	O
</s>
<s>
Well-controlled	O
studies	O
suggested	O
the	O
merits	O
of	O
such	O
a	O
source	O
analysis	O
based	O
brain-computer	B-Application
interface	I-Application
.	O
</s>
<s>
A	O
2014	O
study	O
found	O
that	O
severely	O
motor-impaired	O
patients	O
could	O
communicate	O
faster	O
and	O
more	O
reliably	O
with	O
non-invasive	O
EEG	B-Application
BCI	O
,	O
than	O
with	O
any	O
muscle-based	O
communication	O
channel	O
.	O
</s>
<s>
A	O
2016	O
study	O
found	O
that	O
the	O
Emotiv	B-Algorithm
EPOC	O
device	O
may	O
be	O
more	O
suitable	O
for	O
control	O
tasks	O
using	O
the	O
attention/meditation	O
level	O
or	O
eye	O
blinking	O
than	O
the	O
Neurosky	B-Device
MindWave	O
device	O
.	O
</s>
<s>
A	O
2019	O
study	O
found	O
that	O
the	O
application	O
of	O
evolutionary	O
algorithms	O
could	O
improve	O
EEG	B-Application
mental	O
state	O
classification	O
with	O
a	O
non-invasive	O
Muse	B-Device
device	O
,	O
enabling	O
high	B-General_Concept
quality	I-General_Concept
classification	O
of	O
data	O
acquired	O
by	O
a	O
cheap	O
consumer-grade	O
EEG	B-Application
sensing	O
device	O
.	O
</s>
<s>
In	O
a	O
2021	O
systematic	O
review	O
of	O
randomized	O
controlled	O
trials	O
using	O
BCI	O
for	O
upper-limb	O
rehabilitation	O
after	O
stroke	O
,	O
EEG-based	O
BCI	O
was	O
found	O
to	O
have	O
significant	O
efficacy	O
in	O
improving	O
upper-limb	O
motor	O
function	O
compared	O
to	O
control	O
therapies	O
.	O
</s>
<s>
Another	O
2021	O
systematic	O
review	O
focused	O
on	O
robotic-assisted	O
EEG-based	O
BCI	O
for	O
hand	O
rehabilitation	O
after	O
stroke	O
.	O
</s>
<s>
The	O
single	O
channel	O
dry	O
EEG	B-Application
electrode	O
construction	O
and	O
results	O
were	O
published	O
in	O
1994	O
.	O
</s>
<s>
The	O
advantages	O
of	O
such	O
electrodes	O
are	O
:	O
(	O
1	O
)	O
no	O
electrolyte	O
used	O
,	O
(	O
2	O
)	O
no	O
skin	O
preparation	O
,	O
(	O
3	O
)	O
significantly	O
reduced	O
sensor	O
size	O
,	O
and	O
(	O
4	O
)	O
compatibility	O
with	O
EEG	B-Application
monitoring	O
systems	O
.	O
</s>
<s>
The	O
electrode	O
was	O
tested	O
on	O
an	O
electrical	O
test	O
bench	O
and	O
on	O
human	O
subjects	O
in	O
four	O
modalities	O
of	O
EEG	B-Application
activity	O
,	O
namely	O
:	O
(	O
1	O
)	O
spontaneous	O
EEG	B-Application
,	O
(	O
2	O
)	O
sensory	O
event-related	B-Algorithm
potentials	I-Algorithm
,	O
(	O
3	O
)	O
brain	O
stem	O
potentials	O
,	O
and	O
(	O
4	O
)	O
cognitive	O
event-related	B-Algorithm
potentials	I-Algorithm
.	O
</s>
<s>
In	O
1999	O
researchers	O
at	O
Case	O
Western	O
Reserve	O
University	O
,	O
in	O
Cleveland	O
,	O
Ohio	O
,	O
led	O
by	O
Hunter	O
Peckham	O
,	O
used	O
64-electrode	O
EEG	B-Application
skullcap	O
to	O
return	O
limited	O
hand	O
movements	O
to	O
quadriplegic	O
Jim	O
Jatich	O
.	O
</s>
<s>
As	O
Jatich	O
concentrated	O
on	O
simple	O
but	O
opposite	O
concepts	O
like	O
up	O
and	O
down	O
,	O
his	O
beta-rhythm	O
EEG	B-Application
output	O
was	O
analysed	O
using	O
software	O
to	O
identify	O
patterns	O
in	O
the	O
noise	O
.	O
</s>
<s>
As	O
well	O
as	O
enabling	O
Jatich	O
to	O
control	O
a	O
computer	O
cursor	B-Application
the	O
signals	O
were	O
also	O
used	O
to	O
drive	O
the	O
nerve	O
controllers	O
embedded	O
in	O
his	O
hands	O
,	O
restoring	O
some	O
movement	O
.	O
</s>
<s>
The	O
researchers	O
who	O
developed	O
this	O
BCI-headband	O
also	O
engineered	O
silicon-based	O
microelectro-mechanical	O
system	O
(	O
MEMS	B-Architecture
)	O
dry	O
electrodes	O
designed	O
for	O
application	O
in	O
non-hairy	O
sites	O
of	O
the	O
body	O
.	O
</s>
<s>
These	O
electrodes	O
were	O
secured	O
to	O
the	O
DAQ	B-Algorithm
board	O
in	O
the	O
headband	O
with	O
snap-on	O
electrode	O
holders	O
.	O
</s>
<s>
In	O
2011	O
,	O
researchers	O
reported	O
a	O
cellular	O
based	O
BCI	O
with	O
the	O
capability	O
of	O
taking	O
EEG	B-Application
data	O
and	O
converting	O
it	O
into	O
a	O
command	O
to	O
cause	O
the	O
phone	O
to	O
ring	O
.	O
</s>
<s>
The	O
electrodes	O
were	O
placed	O
so	O
that	O
they	O
pick	O
up	O
steady	O
state	O
visual	O
evoked	O
potentials	O
(	O
SSVEPs	B-Algorithm
)	O
.	O
</s>
<s>
SSVEPs	B-Algorithm
are	O
electrical	O
responses	O
to	O
flickering	O
visual	O
stimuli	O
with	O
repetition	O
rates	O
over	O
6Hz	O
that	O
are	O
best	O
found	O
in	O
the	O
parietal	O
and	O
occipital	O
scalp	O
regions	O
of	O
the	O
visual	O
cortex	O
.	O
</s>
<s>
While	O
the	O
cellular	O
based	O
BCI	O
technology	O
was	O
developed	O
to	O
initiate	O
a	O
phone	O
call	O
from	O
SSVEPs	B-Algorithm
,	O
the	O
researchers	O
said	O
that	O
it	O
can	O
be	O
translated	O
for	O
other	O
applications	O
,	O
such	O
as	O
picking	O
up	O
sensorimotor	O
mu/beta	O
rhythms	O
to	O
function	O
as	O
a	O
motor-imagery	O
based	O
BCI	O
.	O
</s>
<s>
In	O
2013	O
,	O
comparative	O
tests	O
were	O
performed	O
on	O
android	B-Application
cell	I-Application
phone	I-Application
,	O
tablet	O
,	O
and	O
computer	O
based	O
BCIs	O
,	O
analyzing	O
the	O
power	O
spectrum	B-Algorithm
density	I-Algorithm
of	O
resultant	O
EEG	B-Application
SSVEPs	B-Algorithm
.	O
</s>
<s>
The	O
stated	O
goals	O
of	O
this	O
study	O
,	O
which	O
involved	O
scientists	O
supported	O
in	O
part	O
by	O
the	O
U.S.	O
Army	O
Research	O
Laboratory	O
,	O
were	O
to	O
"	O
increase	O
the	O
practicability	O
,	O
portability	O
,	O
and	O
ubiquity	O
of	O
an	O
SSVEP-based	O
BCI	O
,	O
for	O
daily	O
use	O
"	O
.	O
</s>
<s>
The	O
amplitudes	O
of	O
the	O
SSVEPs	B-Algorithm
for	O
the	O
laptop	O
and	O
tablet	O
were	O
also	O
reported	O
to	O
be	O
larger	O
than	O
those	O
of	O
the	O
cell	O
phone	O
.	O
</s>
<s>
In	O
2011	O
,	O
researchers	O
stated	O
that	O
continued	O
work	O
should	O
address	O
ease	B-General_Concept
of	I-General_Concept
use	I-General_Concept
,	O
performance	O
robustness	O
,	O
reducing	O
hardware	O
and	O
software	O
costs	O
.	O
</s>
<s>
One	O
of	O
the	O
difficulties	O
with	O
EEG	B-Application
readings	O
is	O
the	O
large	O
susceptibility	O
to	O
motion	O
artifacts	O
.	O
</s>
<s>
In	O
2013	O
,	O
researchers	O
tested	O
mobile	O
EEG-based	O
BCI	O
technology	O
,	O
measuring	O
SSVEPs	B-Algorithm
from	O
participants	O
as	O
they	O
walked	O
on	O
a	O
treadmill	O
at	O
varying	O
speeds	O
.	O
</s>
<s>
Stated	O
results	O
were	O
that	O
as	O
speed	O
increased	O
the	O
SSVEP	B-Algorithm
detectability	O
using	O
CCA	O
decreased	O
.	O
</s>
<s>
As	O
independent	B-Algorithm
component	I-Algorithm
analysis	I-Algorithm
(	O
ICA	B-Algorithm
)	O
had	O
been	O
shown	O
to	O
be	O
efficient	O
in	O
separating	O
EEG	B-Application
signals	O
from	O
noise	O
,	O
the	O
scientists	O
applied	O
ICA	B-Algorithm
to	O
CCA	O
extracted	O
EEG	B-Application
data	O
.	O
</s>
<s>
They	O
stated	O
that	O
the	O
CCA	O
data	O
with	O
and	O
without	O
ICA	B-Algorithm
processing	O
were	O
similar	O
.	O
</s>
<s>
One	O
of	O
the	O
major	O
problems	O
in	O
EEG-based	O
BCI	O
applications	O
is	O
the	O
low	O
spatial	O
resolution	B-General_Concept
.	O
</s>
<s>
Several	O
solutions	O
have	O
been	O
suggested	O
to	O
address	O
this	O
issue	O
since	O
2019	O
,	O
which	O
include	O
:	O
EEG	B-Application
source	O
connectivity	O
based	O
on	O
graph	O
theory	O
,	O
EEG	B-Application
pattern	O
recognition	O
based	O
on	O
Topomap	O
,	O
EEG-fMRI	O
fusion	O
,	O
and	O
so	O
on	O
.	O
</s>
<s>
In	O
2009	O
Alex	O
Blainey	O
,	O
an	O
independent	O
researcher	O
based	O
in	O
the	O
UK	O
,	O
successfully	O
used	O
the	O
Emotiv	B-Algorithm
EPOC	O
to	O
control	O
a	O
5	O
axis	O
robot	O
arm	O
.	O
</s>
<s>
The	O
OpenEEG	O
Project	O
marked	O
a	O
significant	O
moment	O
in	O
the	O
emergence	O
of	O
DIY	O
brain-computer	B-Application
interfacing	I-Application
.	O
</s>
<s>
In	O
2010	O
,	O
the	O
Frontier	O
Nerds	O
of	O
NYU	O
's	O
ITP	O
program	O
published	O
a	O
thorough	O
tutorial	O
titled	O
How	O
To	O
Hack	O
Toy	O
EEGs	B-Application
.	O
</s>
<s>
The	O
tutorial	O
,	O
which	O
stirred	O
the	O
minds	O
of	O
many	O
budding	O
DIY	O
BCI	O
enthusiasts	O
,	O
demonstrated	O
how	O
to	O
create	O
a	O
single	O
channel	O
at-home	O
EEG	B-Application
with	O
an	O
Arduino	O
and	O
a	O
Mattel	O
Mindflex	O
at	O
a	O
very	O
reasonable	O
price	O
.	O
</s>
<s>
Magnetoencephalography	B-Algorithm
(	O
MEG	B-Algorithm
)	O
and	O
functional	B-Algorithm
magnetic	I-Algorithm
resonance	I-Algorithm
imaging	I-Algorithm
(	O
fMRI	B-Algorithm
)	O
have	O
both	O
been	O
used	O
successfully	O
as	O
non-invasive	O
BCIs	O
.	O
</s>
<s>
In	O
a	O
widely	O
reported	O
experiment	O
,	O
fMRI	B-Algorithm
allowed	O
two	O
users	O
being	O
scanned	O
to	O
play	O
Pong	B-Application
in	O
real-time	O
by	O
altering	O
their	O
haemodynamic	O
response	O
or	O
brain	O
blood	O
flow	O
through	O
biofeedback	O
techniques	O
.	O
</s>
<s>
fMRI	B-Algorithm
measurements	O
of	O
haemodynamic	O
responses	O
in	O
real	O
time	O
have	O
also	O
been	O
used	O
to	O
control	O
robot	O
arms	O
with	O
a	O
seven-second	O
delay	O
between	O
thought	O
and	O
movement	O
.	O
</s>
<s>
In	O
2008	O
research	O
developed	O
in	O
the	O
Advanced	O
Telecommunications	O
Research	O
(	O
ATR	O
)	O
Computational	O
Neuroscience	O
Laboratories	O
in	O
Kyoto	O
,	O
Japan	O
,	O
allowed	O
the	O
scientists	O
to	O
reconstruct	O
images	O
directly	O
from	O
the	O
brain	O
and	O
display	O
them	O
on	O
a	O
computer	O
in	O
black	O
and	O
white	O
at	O
a	O
resolution	B-General_Concept
of	O
10x10	O
pixels	B-Algorithm
.	O
</s>
<s>
In	O
2011	O
researchers	O
from	O
UC	O
Berkeley	O
published	O
a	O
study	O
reporting	O
second-by-second	O
reconstruction	O
of	O
videos	O
watched	O
by	O
the	O
study	O
's	O
subjects	O
,	O
from	O
fMRI	B-Algorithm
data	O
.	O
</s>
<s>
This	O
was	O
achieved	O
by	O
creating	O
a	O
statistical	O
model	O
relating	O
visual	O
patterns	O
in	O
videos	O
shown	O
to	O
the	O
subjects	O
,	O
to	O
the	O
brain	B-Application
activity	I-Application
caused	O
by	O
watching	O
the	O
videos	O
.	O
</s>
<s>
This	O
model	O
was	O
then	O
used	O
to	O
look	O
up	O
the	O
100	O
one-second	O
video	O
segments	O
,	O
in	O
a	O
database	O
of	O
18	O
million	O
seconds	O
of	O
random	O
YouTube	B-General_Concept
videos	I-General_Concept
,	O
whose	O
visual	O
patterns	O
most	O
closely	O
matched	O
the	O
brain	B-Application
activity	I-Application
recorded	O
when	O
subjects	O
watched	O
a	O
new	O
video	O
.	O
</s>
<s>
Motor	O
imagery	O
involves	O
the	O
imagination	O
of	O
the	O
movement	O
of	O
various	O
body	O
parts	O
resulting	O
in	O
sensorimotor	O
cortex	O
activation	O
,	O
which	O
modulates	O
sensorimotor	O
oscillations	O
in	O
the	O
EEG	B-Application
.	O
</s>
<s>
In	O
some	O
cases	O
,	O
biofeedback	O
does	O
not	O
monitor	O
electroencephalography	B-Application
(	O
EEG	B-Application
)	O
,	O
but	O
instead	O
bodily	O
parameters	O
such	O
as	O
electromyography	O
(	O
EMG	O
)	O
,	O
galvanic	O
skin	O
resistance	O
(	O
GSR	O
)	O
,	O
and	O
heart	O
rate	O
variability	O
(	O
HRV	O
)	O
.	O
</s>
<s>
EEG	B-Application
biofeedback	O
systems	O
typically	O
monitor	O
four	O
different	O
bands	O
(	O
theta	O
:	O
4	O
–	O
7Hz	O
,	O
alpha:8	O
–	O
12Hz	O
,	O
SMR	O
:	O
12	O
–	O
15Hz	O
,	O
beta	B-Algorithm
:	O
15	O
–	O
18Hz	O
)	O
and	O
challenge	O
the	O
subject	O
to	O
control	O
them	O
.	O
</s>
<s>
Steady-state	O
visually	O
evoked	O
potentials	O
(	O
SSVEPs	B-Algorithm
)	O
use	O
potentials	O
generated	O
by	O
exciting	O
the	O
retina	O
,	O
using	O
visual	O
stimuli	O
modulated	O
at	O
certain	O
frequencies	O
.	O
</s>
<s>
SSVEP	B-Algorithm
's	O
stimuli	O
are	O
often	O
formed	O
from	O
alternating	O
checkerboard	O
patterns	O
and	O
at	O
times	O
simply	O
use	O
flashing	O
images	O
.	O
</s>
<s>
The	O
frequency	O
of	O
the	O
phase	O
reversal	O
of	O
the	O
stimulus	O
used	O
can	O
be	O
clearly	O
distinguished	O
in	O
the	O
spectrum	O
of	O
an	O
EEG	B-Application
;	O
this	O
makes	O
detection	O
of	O
SSVEP	B-Algorithm
stimuli	O
relatively	O
easy	O
.	O
</s>
<s>
SSVEP	B-Algorithm
has	O
proved	O
to	O
be	O
successful	O
within	O
many	O
BCI	O
systems	O
.	O
</s>
<s>
In	O
addition	O
,	O
the	O
SSVEP	B-Algorithm
signal	O
is	O
exceptionally	O
robust	O
;	O
the	O
topographic	O
organization	O
of	O
the	O
primary	O
visual	O
cortex	O
is	O
such	O
that	O
a	O
broader	O
area	O
obtains	O
afferents	O
from	O
the	O
central	O
or	O
fovial	O
region	O
of	O
the	O
visual	O
field	O
.	O
</s>
<s>
SSVEP	B-Algorithm
does	O
have	O
several	O
problems	O
however	O
.	O
</s>
<s>
As	O
SSVEPs	B-Algorithm
use	O
flashing	O
stimuli	O
to	O
infer	O
a	O
user	O
's	O
intent	O
,	O
the	O
user	O
must	O
gaze	O
at	O
one	O
of	O
the	O
flashing	O
or	O
iterating	O
symbols	O
in	O
order	O
to	O
interact	O
with	O
the	O
system	O
.	O
</s>
<s>
Another	O
type	O
of	O
VEP	O
used	O
with	O
applications	O
is	O
the	O
P300	B-Algorithm
potential	I-Algorithm
.	O
</s>
<s>
The	O
P300	B-Algorithm
event-related	B-Algorithm
potential	I-Algorithm
is	O
a	O
positive	O
peak	O
in	O
the	O
EEG	B-Application
that	O
occurs	O
at	O
roughly	O
300	O
ms	O
after	O
the	O
appearance	O
of	O
a	O
target	O
stimulus	O
(	O
a	O
stimulus	O
for	O
which	O
the	O
user	O
is	O
waiting	O
or	O
seeking	O
)	O
or	O
oddball	B-Algorithm
stimuli	I-Algorithm
.	O
</s>
<s>
The	O
P300	B-Algorithm
amplitude	O
decreases	O
as	O
the	O
target	O
stimuli	O
and	O
the	O
ignored	O
stimuli	O
grow	O
more	O
similar.The	O
P300	B-Algorithm
is	O
thought	O
to	O
be	O
related	O
to	O
a	O
higher	O
level	O
attention	O
process	O
or	O
an	O
orienting	O
response	O
using	O
P300	B-Algorithm
as	O
a	O
control	O
scheme	O
has	O
the	O
advantage	O
of	O
the	O
participant	O
only	O
having	O
to	O
attend	O
limited	O
training	O
sessions	O
.	O
</s>
<s>
The	O
first	O
application	O
to	O
use	O
the	O
P300	B-Algorithm
model	O
was	O
the	O
P300	B-Algorithm
matrix	O
.	O
</s>
<s>
The	O
rows	O
and	O
columns	O
of	O
the	O
grid	O
flashed	O
sequentially	O
and	O
every	O
time	O
the	O
selected	O
"	O
choice	O
letter	O
"	O
was	O
illuminated	O
the	O
user	O
's	O
P300	B-Algorithm
was	O
(	O
potentially	O
)	O
elicited	O
.	O
</s>
<s>
The	O
P300	B-Algorithm
is	O
a	O
BCI	O
that	O
offers	O
a	O
discrete	O
selection	O
rather	O
than	O
a	O
continuous	O
control	O
mechanism	O
.	O
</s>
<s>
The	O
advantage	O
of	O
P300	B-Algorithm
use	O
within	O
games	O
is	O
that	O
the	O
player	O
does	O
not	O
have	O
to	O
teach	O
himself/herself	O
how	O
to	O
use	O
a	O
completely	O
new	O
control	O
system	O
and	O
so	O
only	O
has	O
to	O
undertake	O
short	O
training	O
instances	O
,	O
to	O
learn	O
the	O
gameplay	O
mechanics	O
and	O
basic	O
use	O
of	O
the	O
BCI	O
paradigm	O
.	O
</s>
<s>
In	O
the	O
1960s	O
a	O
researcher	O
was	O
successful	O
after	O
some	O
training	O
in	O
using	O
EEG	B-Application
to	O
create	O
Morse	O
code	O
using	O
their	O
brain	O
alpha	B-Application
waves	I-Application
.	O
</s>
<s>
Researchers	O
have	O
built	O
devices	O
to	O
interface	O
with	O
neural	O
cells	O
and	O
entire	O
neural	B-Architecture
networks	I-Architecture
in	O
cultures	O
outside	O
animals	O
.	O
</s>
<s>
In	O
2004	O
Thomas	O
DeMarse	O
at	O
the	O
University	O
of	O
Florida	O
used	O
a	O
culture	O
of	O
25,000	O
neurons	O
taken	O
from	O
a	O
rat	O
's	O
brain	O
to	O
fly	O
a	O
F-22	B-Application
fighter	O
jet	O
aircraft	B-Application
simulator	I-Application
.	O
</s>
<s>
After	O
collection	O
,	O
the	O
cortical	O
neurons	O
were	O
cultured	O
in	O
a	O
petri	O
dish	O
and	O
rapidly	O
began	O
to	O
reconnect	O
themselves	O
to	O
form	O
a	O
living	O
neural	B-Architecture
network	I-Architecture
.	O
</s>
<s>
In	O
May	O
2011	O
,	O
Yijun	O
Wang	O
and	O
Tzyy-Ping	O
Jung	O
published	O
,	O
"	O
A	O
Collaborative	O
Brain-Computer	B-Application
Interface	I-Application
for	O
Improving	O
Human	O
Performance	O
"	O
,	O
and	O
in	O
January	O
2012	O
Miguel	O
Eckstein	O
published	O
,	O
"	O
Neural	O
decoding	O
of	O
collective	O
wisdom	O
with	O
multi-brain	O
computing	O
"	O
.	O
</s>
<s>
Recently	O
a	O
number	O
of	O
companies	O
have	O
scaled	O
back	O
medical	O
grade	O
EEG	B-Application
technology	O
to	O
create	O
inexpensive	O
BCIs	O
for	O
research	O
as	O
well	O
as	O
entertainment	O
purposes	O
.	O
</s>
<s>
For	O
example	O
,	O
toys	O
such	O
as	O
the	O
NeuroSky	B-Device
and	O
Mattel	O
MindFlex	O
have	O
seen	O
some	O
commercial	O
success	O
.	O
</s>
<s>
In	O
2006	O
Sony	O
patented	O
a	O
neural	B-Application
interface	I-Application
system	O
allowing	O
radio	O
waves	O
to	O
affect	O
signals	O
in	O
the	O
neural	O
cortex	O
.	O
</s>
<s>
In	O
2007	O
NeuroSky	B-Device
released	O
the	O
first	O
affordable	O
consumer	O
based	O
EEG	B-Application
along	O
with	O
the	O
game	O
NeuroBoy	O
.	O
</s>
<s>
This	O
was	O
also	O
the	O
first	O
large	O
scale	O
EEG	B-Application
device	O
to	O
use	O
dry	O
sensor	O
technology	O
.	O
</s>
<s>
In	O
2008	O
Final	B-Application
Fantasy	I-Application
developer	O
Square	O
Enix	O
announced	O
that	O
it	O
was	O
partnering	O
with	O
NeuroSky	B-Device
to	O
create	O
a	O
game	O
,	O
Judecca	O
.	O
</s>
<s>
In	O
2009	O
Mattel	O
partnered	O
with	O
NeuroSky	B-Device
to	O
release	O
the	O
Mindflex	O
,	O
a	O
game	O
that	O
used	O
an	O
EEG	B-Application
to	O
steer	O
a	O
ball	O
through	O
an	O
obstacle	O
course	O
.	O
</s>
<s>
It	O
is	O
by	O
far	O
the	O
best	O
selling	O
consumer	O
based	O
EEG	B-Application
to	O
date	O
.	O
</s>
<s>
In	O
2009	O
Uncle	O
Milton	O
Industries	O
partnered	O
with	O
NeuroSky	B-Device
to	O
release	O
the	O
Star	B-Application
Wars	I-Application
Force	O
Trainer	O
,	O
a	O
game	O
designed	O
to	O
create	O
the	O
illusion	O
of	O
possessing	O
the	O
Force	O
.	O
</s>
<s>
In	O
2009	O
Emotiv	B-Algorithm
released	O
the	O
EPOC	O
,	O
a	O
14	O
channel	O
EEG	B-Application
device	O
that	O
can	O
read	O
4	O
mental	O
states	O
,	O
13	O
conscious	O
states	O
,	O
facial	O
expressions	O
,	O
and	O
head	O
movements	O
.	O
</s>
<s>
The	O
company	O
announced	O
that	O
it	O
expected	O
to	O
launch	O
a	O
consumer	O
version	O
of	O
the	O
garment	O
,	O
consisting	O
of	O
catlike	O
ears	O
controlled	O
by	O
a	O
brain-wave	O
reader	O
produced	O
by	O
NeuroSky	B-Device
,	O
in	O
spring	O
2012	O
.	O
</s>
<s>
Basic	O
diagnostic	O
software	O
is	O
available	O
for	O
Android	B-Application
devices	O
,	O
as	O
well	O
as	O
a	O
text	O
entry	O
app	O
for	O
Unity	B-Application
.	O
</s>
<s>
In	O
2020	O
,	O
NextMind	O
released	O
a	O
dev	O
kit	O
including	O
an	O
EEG	B-Application
headset	O
with	O
dry	O
electrodes	O
at	O
$399	O
.	O
</s>
<s>
A	O
2015	O
publication	O
led	O
by	O
Dr.	O
Clemens	O
Brunner	O
describes	O
some	O
of	O
the	O
analyses	O
and	O
achievements	O
of	O
this	O
project	O
,	O
as	O
well	O
as	O
the	O
emerging	O
Brain-Computer	B-Application
Interface	I-Application
Society	O
.	O
</s>
<s>
This	O
focused	O
attention	O
produces	O
reliable	O
changes	O
in	O
EEG	B-Application
patterns	I-Application
that	O
can	O
help	O
determine	O
that	O
the	O
patient	O
is	O
able	O
to	O
communicate	O
.	O
</s>
<s>
Research	O
in	O
recent	O
years	O
has	O
demonstrated	O
the	O
utility	O
of	O
EEG-based	O
BCI	O
systems	O
in	O
aiding	O
motor	O
recovery	O
and	O
neurorehabilitation	O
in	O
patients	O
who	O
have	O
had	O
a	O
stroke	O
.	O
</s>
<s>
So	O
far	O
,	O
BCIs	O
for	O
motor	O
recovery	O
have	O
relied	O
on	O
the	O
EEG	B-Application
to	O
measure	O
the	O
patient	O
's	O
motor	O
imagery	O
.	O
</s>
<s>
However	O
,	O
studies	O
have	O
also	O
used	O
fMRI	B-Algorithm
to	O
study	O
different	O
changes	O
in	O
the	O
brain	O
as	O
persons	O
undergo	O
BCI-based	O
stroke	O
rehab	O
training	O
.	O
</s>
<s>
Imaging	O
studies	O
combined	O
with	O
EEG-based	O
BCI	O
systems	O
hold	O
promise	O
for	O
investigating	O
neuroplasticity	O
during	O
motor	O
recovery	O
post-stroke	O
.	O
</s>
<s>
Future	O
systems	O
might	O
include	O
the	O
fMRI	B-Algorithm
and	O
other	O
measures	O
for	O
real-time	O
control	O
,	O
such	O
as	O
functional	O
near-infrared	O
,	O
probably	O
in	O
tandem	O
with	O
EEGs	B-Application
.	O
</s>
<s>
In	O
2016	O
,	O
scientists	O
out	O
of	O
the	O
University	O
of	O
Melbourne	O
published	O
preclinical	O
proof-of-concept	O
data	O
related	O
to	O
a	O
potential	O
brain-computer	B-Application
interface	I-Application
technology	O
platform	O
being	O
developed	O
for	O
patients	O
with	O
paralysis	O
to	O
facilitate	O
control	O
of	O
external	O
devices	O
such	O
as	O
robotic	O
limbs	O
,	O
computers	O
and	O
exoskeletons	O
by	O
translating	O
brain	B-Application
activity	I-Application
.	O
</s>
<s>
Flexible	O
electronics	O
are	O
polymers	B-Language
or	O
other	O
flexible	O
materials	O
(	O
e.g.	O
</s>
<s>
silk	O
,	O
pentacene	O
,	O
PDMS	O
,	O
Parylene	O
,	O
polyimide	O
)	O
that	O
are	O
printed	O
with	O
circuitry	O
;	O
the	O
flexible	O
nature	O
of	O
the	O
organic	O
background	O
materials	O
allowing	O
the	O
electronics	O
created	O
to	O
bend	O
,	O
and	O
the	O
fabrication	B-Architecture
techniques	I-Architecture
used	O
to	O
create	O
these	O
devices	O
resembles	O
those	O
used	O
to	O
create	O
integrated	O
circuits	O
and	O
microelectromechanical	B-Architecture
systems	I-Architecture
(	O
MEMS	B-Architecture
)	O
.	O
</s>
<s>
Flexible	O
neural	B-Application
interfaces	I-Application
have	O
been	O
extensively	O
tested	O
in	O
recent	O
years	O
in	O
an	O
effort	O
to	O
minimize	O
brain	O
tissue	O
trauma	O
related	O
to	O
mechanical	O
mismatch	O
between	O
electrode	O
and	O
tissue	O
.	O
</s>
<s>
Neural	O
dust	O
is	O
a	O
term	O
used	O
to	O
refer	O
to	O
millimeter-sized	O
devices	O
operated	O
as	O
wirelessly	B-Application
powered	I-Application
nerve	O
sensors	O
that	O
were	O
proposed	O
in	O
a	O
2011	O
paper	O
from	O
the	O
University	O
of	O
California	O
,	O
Berkeley	O
Wireless	O
Research	O
Center	O
,	O
which	O
described	O
both	O
the	O
challenges	O
and	O
outstanding	O
benefits	O
of	O
creating	O
a	O
long	O
lasting	O
wireless	O
BCI	O
.	O
</s>
<s>
In	O
one	O
proposed	O
model	O
of	O
the	O
neural	O
dust	O
sensor	O
,	O
the	O
transistor	O
model	O
allowed	O
for	O
a	O
method	O
of	O
separating	O
between	O
local	B-Algorithm
field	I-Algorithm
potentials	I-Algorithm
and	O
action	B-Algorithm
potential	I-Algorithm
"	O
spikes	B-Algorithm
"	O
,	O
which	O
would	O
allow	O
for	O
a	O
greatly	O
diversified	O
wealth	O
of	O
data	O
acquirable	O
from	O
the	O
recordings	O
.	O
</s>
