<s>
Functional	B-Algorithm
magnetic	I-Algorithm
resonance	I-Algorithm
imaging	I-Algorithm
or	O
functional	B-Algorithm
MRI	I-Algorithm
(	O
fMRI	B-Algorithm
)	O
measures	O
brain	B-Application
activity	I-Application
by	O
detecting	O
changes	O
associated	O
with	O
blood	O
flow	O
.	O
</s>
<s>
The	O
primary	O
form	O
of	O
fMRI	B-Algorithm
uses	O
the	O
blood-oxygen-level	B-Algorithm
dependent	I-Algorithm
(	O
BOLD	O
)	O
contrast	O
,	O
discovered	O
by	O
Seiji	O
Ogawa	O
in	O
1990	O
.	O
</s>
<s>
Since	O
the	O
early	O
1990s	O
,	O
fMRI	B-Algorithm
has	O
come	O
to	O
dominate	O
brain	O
mapping	O
research	O
because	O
it	O
does	O
not	O
involve	O
the	O
use	O
of	O
injections	O
,	O
surgery	O
,	O
the	O
ingestion	O
of	O
substances	O
,	O
or	O
exposure	O
to	O
ionizing	O
radiation	O
.	O
</s>
<s>
Other	O
methods	O
of	O
obtaining	O
contrast	O
are	O
arterial	B-Algorithm
spin	I-Algorithm
labeling	I-Algorithm
and	O
diffusion	B-Algorithm
MRI	I-Algorithm
.	O
</s>
<s>
Diffusion	B-Algorithm
MRI	I-Algorithm
is	O
similar	O
to	O
BOLD	O
fMRI	B-Algorithm
but	O
provides	O
contrast	O
based	O
on	O
the	O
magnitude	O
of	O
diffusion	O
of	O
water	O
molecules	O
in	O
the	O
brain	O
.	O
</s>
<s>
In	O
addition	O
to	O
detecting	O
BOLD	O
responses	O
from	O
activity	O
due	O
to	O
tasks	O
or	O
stimuli	O
,	O
fMRI	B-Algorithm
can	O
measure	O
resting	B-Algorithm
state	I-Algorithm
,	O
or	O
negative-task	O
state	O
,	O
which	O
shows	O
the	O
subjects	O
 '	O
baseline	O
BOLD	O
variance	O
.	O
</s>
<s>
fMRI	B-Algorithm
is	O
used	O
in	O
research	O
,	O
and	O
to	O
a	O
lesser	O
extent	O
,	O
in	O
clinical	O
work	O
.	O
</s>
<s>
It	O
can	O
complement	O
other	O
measures	O
of	O
brain	O
physiology	O
such	O
as	O
electroencephalography	B-Application
(	O
EEG	B-Application
)	O
,	O
and	O
near-infrared	O
spectroscopy	O
(	O
NIRS	O
)	O
.	O
</s>
<s>
Newer	O
methods	O
which	O
improve	O
both	O
spatial	O
and	O
time	O
resolution	O
are	O
being	O
researched	O
,	O
and	O
these	O
largely	O
use	O
biomarkers	O
other	O
than	O
the	O
BOLD	B-Algorithm
signal	I-Algorithm
.	O
</s>
<s>
Some	O
companies	O
have	O
developed	O
commercial	O
products	O
such	O
as	O
lie	O
detectors	O
based	O
on	O
fMRI	B-Algorithm
techniques	O
,	O
but	O
the	O
research	O
is	O
not	O
believed	O
to	O
be	O
developed	O
enough	O
for	O
widespread	O
commercial	O
use	O
.	O
</s>
<s>
The	O
fMRI	B-Algorithm
concept	O
builds	O
on	O
the	O
earlier	O
MRI	B-Algorithm
scanning	O
technology	O
and	O
the	O
discovery	O
of	O
properties	O
of	O
oxygen-rich	O
blood	O
.	O
</s>
<s>
MRI	B-Algorithm
brain	O
scans	O
use	O
a	O
strong	O
,	O
permanent	O
,	O
static	O
magnetic	O
field	O
to	O
align	O
nuclei	O
in	O
the	O
brain	O
region	O
being	O
studied	O
.	O
</s>
<s>
MRI	B-Algorithm
thus	O
provides	O
a	O
static	O
structural	O
view	O
of	O
brain	O
matter	O
.	O
</s>
<s>
The	O
central	O
thrust	O
behind	O
fMRI	B-Algorithm
was	O
to	O
extend	O
MRI	B-Algorithm
to	O
capture	O
functional	O
changes	O
in	O
the	O
brain	O
caused	O
by	O
neuronal	O
activity	O
.	O
</s>
<s>
Seiji	O
Ogawa	O
at	O
AT&T	O
Bell	O
labs	O
recognized	O
that	O
this	O
could	O
be	O
used	O
to	O
augment	O
MRI	B-Algorithm
,	O
which	O
could	O
study	O
just	O
the	O
static	O
structure	O
of	O
the	O
brain	O
,	O
since	O
the	O
differing	O
magnetic	O
properties	O
of	O
dHb	O
and	O
Hb	O
caused	O
by	O
blood	O
flow	O
to	O
activated	O
brain	O
regions	O
would	O
cause	O
measurable	O
changes	O
in	O
the	O
MRI	B-Algorithm
signal	O
.	O
</s>
<s>
BOLD	O
is	O
the	O
MRI	B-Algorithm
contrast	O
of	O
dHb	O
,	O
discovered	O
in	O
1990	O
by	O
Ogawa	O
.	O
</s>
<s>
In	O
a	O
seminal	O
1990	O
study	O
based	O
on	O
earlier	O
work	O
by	O
Thulborn	O
et	O
al.	O
,	O
Ogawa	O
and	O
colleagues	O
scanned	O
rodents	O
in	O
a	O
strong	O
magnetic	O
field	O
(	O
7.0T	O
)	O
MRI	B-Algorithm
.	O
</s>
<s>
As	O
this	O
proportion	O
fell	O
,	O
a	O
map	O
of	O
blood	O
flow	O
in	O
the	O
brain	O
was	O
seen	O
in	O
the	O
MRI	B-Algorithm
.	O
</s>
<s>
To	O
show	O
these	O
blood	O
flow	O
changes	O
were	O
related	O
to	O
functional	O
brain	B-Application
activity	I-Application
,	O
they	O
changed	O
the	O
composition	O
of	O
the	O
air	O
breathed	O
by	O
rats	O
,	O
and	O
scanned	O
them	O
while	O
monitoring	O
brain	B-Application
activity	I-Application
with	O
EEG	B-Application
.	O
</s>
<s>
The	O
first	O
attempt	O
to	O
detect	O
the	O
regional	O
brain	B-Application
activity	I-Application
using	O
MRI	B-Algorithm
was	O
performed	O
by	O
Belliveau	O
and	O
colleagues	O
at	O
Harvard	O
University	O
using	O
the	O
contrast	O
agent	O
Magnevist	O
,	O
a	O
paramagnetic	O
substance	O
remaining	O
in	O
the	O
bloodstream	O
after	O
intravenous	O
injection	O
.	O
</s>
<s>
However	O
,	O
this	O
method	O
is	O
not	O
popular	O
in	O
human	O
fMRI	B-Algorithm
,	O
because	O
of	O
the	O
inconvenience	O
of	O
the	O
contrast	O
agent	O
injection	O
,	O
and	O
because	O
the	O
agent	O
stays	O
in	O
the	O
blood	O
only	O
for	O
a	O
short	O
time	O
.	O
</s>
<s>
Ogawa	O
and	O
Ugurbil	O
conducted	O
a	O
similar	O
study	O
using	O
a	O
higher	O
magnetic	O
field	O
(	O
4.0	O
T	O
)	O
in	O
Ugurbil	O
's	O
laboratory	O
at	O
the	O
University	O
of	O
Minnesota	O
,	O
generating	O
higher	O
resolution	O
images	O
that	O
showed	O
activity	O
largely	O
following	O
the	O
gray	O
matter	O
of	O
the	O
brain	O
,	O
as	O
would	O
be	O
expected	O
;	O
in	O
addition	O
,	O
they	O
showed	O
that	O
fMRI	B-Algorithm
signal	O
depended	O
on	O
a	O
decrease	O
in	O
T2*	O
,	O
consistent	O
with	O
the	O
BOLD	O
mechanism	O
.	O
</s>
<s>
The	O
magnetic	O
fields	O
,	O
pulse	O
sequences	O
and	O
procedures	O
and	O
techniques	O
used	O
by	O
these	O
early	O
studies	O
are	O
still	O
used	O
in	O
current-day	O
fMRI	B-Algorithm
studies	O
.	O
</s>
<s>
This	O
changes	O
the	O
magnetic	O
property	O
of	O
the	O
blood	O
,	O
making	O
it	O
interfere	O
less	O
with	O
the	O
magnetization	O
and	O
its	O
eventual	O
decay	O
induced	O
by	O
the	O
MRI	B-Algorithm
process	O
.	O
</s>
<s>
Hence	O
,	O
it	O
distorts	O
the	O
surrounding	O
magnetic	O
field	O
induced	O
by	O
an	O
MRI	B-Algorithm
scanner	O
,	O
causing	O
the	O
nuclei	O
there	O
to	O
lose	O
magnetization	O
faster	O
via	O
the	O
T2*	O
decay	O
.	O
</s>
<s>
The	O
fMRI	B-Algorithm
signal	O
hence	O
needs	O
both	O
a	O
strong	O
magnetic	O
field	O
(	O
1.5T	O
or	O
higher	O
)	O
and	O
a	O
pulse	O
sequence	O
such	O
as	O
EPI	O
,	O
which	O
is	O
sensitive	O
to	O
T2*	O
contrast	O
.	O
</s>
<s>
The	O
physiological	O
blood-flow	O
response	O
largely	O
decides	O
the	O
temporal	O
sensitivity	O
,	O
that	O
is	O
how	O
accurately	O
we	O
can	O
measure	O
when	O
neurons	O
are	O
active	O
,	O
in	O
BOLD	O
fMRI	B-Algorithm
.	O
</s>
<s>
For	O
fMRI	B-Algorithm
specifically	O
,	O
the	O
hemodynamic	O
response	O
lasts	O
over	O
10	O
seconds	O
,	O
rising	O
multiplicatively	O
(	O
that	O
is	O
,	O
as	O
a	O
proportion	O
of	O
current	O
value	O
)	O
,	O
peaking	O
at	O
4	O
to	O
6	O
seconds	O
,	O
and	O
then	O
falling	O
multiplicatively	O
.	O
</s>
<s>
After	O
activity	O
stops	O
,	O
the	O
BOLD	B-Algorithm
signal	I-Algorithm
falls	O
below	O
the	O
original	O
level	O
,	O
the	O
baseline	O
,	O
a	O
phenomenon	O
called	O
the	O
undershoot	O
.	O
</s>
<s>
A	O
single	O
voxel	B-Algorithm
's	O
response	O
signal	O
over	O
time	O
is	O
called	O
its	O
timecourse	O
.	O
</s>
<s>
Typically	O
,	O
the	O
unwanted	O
signal	O
,	O
called	O
the	O
noise	O
,	O
from	O
the	O
scanner	O
,	O
random	O
brain	B-Application
activity	I-Application
and	O
similar	O
elements	O
is	O
as	O
big	O
as	O
the	O
signal	O
itself	O
.	O
</s>
<s>
To	O
eliminate	O
these	O
,	O
fMRI	B-Algorithm
studies	O
repeat	O
a	O
stimulus	O
presentation	O
multiple	O
times	O
.	O
</s>
<s>
Spatial	O
resolution	O
of	O
an	O
fMRI	B-Algorithm
study	O
refers	O
to	O
how	O
well	O
it	O
discriminates	O
between	O
nearby	O
locations	O
.	O
</s>
<s>
It	O
is	O
measured	O
by	O
the	O
size	O
of	O
voxels	B-Algorithm
,	O
as	O
in	O
MRI	B-Algorithm
.	O
</s>
<s>
A	O
voxel	B-Algorithm
is	O
a	O
three-dimensional	O
rectangular	O
cuboid	O
,	O
whose	O
dimensions	O
are	O
set	O
by	O
the	O
slice	O
thickness	O
,	O
the	O
area	O
of	O
a	O
slice	O
,	O
and	O
the	O
grid	O
imposed	O
on	O
the	O
slice	O
by	O
the	O
scanning	O
process	O
.	O
</s>
<s>
Full-brain	O
studies	O
use	O
larger	O
voxels	B-Algorithm
,	O
while	O
those	O
that	O
focus	O
on	O
specific	O
regions	O
of	O
interest	O
typically	O
use	O
smaller	O
sizes	O
.	O
</s>
<s>
Sizes	O
range	O
from	O
4	O
to	O
5mm	O
,	O
or	O
with	O
laminar	O
resolution	O
fMRI	B-Algorithm
(	O
lfMRI	O
)	O
,	O
to	O
submillimeter	O
.	O
</s>
<s>
Smaller	O
voxels	B-Algorithm
contain	O
fewer	O
neurons	O
on	O
average	O
,	O
incorporate	O
less	O
blood	O
flow	O
,	O
and	O
hence	O
have	O
less	O
signal	O
than	O
larger	O
voxels	B-Algorithm
.	O
</s>
<s>
Smaller	O
voxels	B-Algorithm
imply	O
longer	O
scanning	O
times	O
,	O
since	O
scanning	O
time	O
directly	O
rises	O
with	O
the	O
number	O
of	O
voxels	B-Algorithm
per	O
slice	O
and	O
the	O
number	O
of	O
slices	O
.	O
</s>
<s>
A	O
voxel	B-Algorithm
typically	O
contains	O
a	O
few	O
million	O
neurons	O
and	O
tens	O
of	O
billions	O
of	O
synapses	B-Application
,	O
with	O
the	O
actual	O
number	O
depending	O
on	O
voxel	B-Algorithm
size	O
and	O
the	O
area	O
of	O
the	O
brain	O
being	O
imaged	O
.	O
</s>
<s>
The	O
dHb	O
contribution	O
to	O
the	O
fMRI	B-Algorithm
signal	O
is	O
from	O
both	O
the	O
capillaries	O
near	O
the	O
area	O
of	O
activity	O
and	O
larger	O
draining	O
veins	O
that	O
may	O
be	O
farther	O
away	O
.	O
</s>
<s>
With	O
these	O
,	O
fMRI	B-Algorithm
can	O
examine	O
a	O
spatial	O
range	O
from	O
millimeters	O
to	O
centimeters	O
,	O
and	O
can	O
hence	O
identify	O
Brodmann	O
areas	O
(	O
centimers	O
)	O
,	O
subcortical	O
nuclei	O
such	O
as	O
the	O
caudate	O
,	O
putamen	O
and	O
thalamus	O
,	O
and	O
hippocampal	O
subfields	O
such	O
as	O
the	O
combined	O
dentate	O
gyrus/CA3	O
,	O
CA1	O
,	O
and	O
subiculum	O
.	O
</s>
<s>
Temporal	O
resolution	O
is	O
the	O
smallest	O
time	O
period	O
of	O
neural	O
activity	O
reliably	O
separated	O
out	O
by	O
fMRI	B-Algorithm
.	O
</s>
<s>
Most	O
fMRI	B-Algorithm
experiments	O
study	O
brain	O
processes	O
lasting	O
a	O
few	O
seconds	O
,	O
with	O
the	O
study	O
conducted	O
over	O
some	O
tens	O
of	O
minutes	O
.	O
</s>
<s>
This	O
is	O
a	O
fundamental	O
assumption	O
of	O
many	O
fMRI	B-Algorithm
studies	O
that	O
is	O
based	O
on	O
the	O
principle	O
that	O
continuously	O
differentiable	O
systems	O
can	O
be	O
expected	O
to	O
behave	O
linearly	O
when	O
perturbations	O
are	O
small	O
;	O
they	O
are	O
linear	O
to	O
first	O
order	O
.	O
</s>
<s>
To	O
the	O
extent	O
that	O
the	O
behavior	O
is	O
linear	O
,	O
the	O
time	O
course	O
of	O
the	O
BOLD	O
response	O
to	O
an	O
arbitrary	O
stimulus	O
can	O
be	O
modeled	O
by	O
convolution	B-Language
of	O
that	O
stimulus	O
with	O
the	O
impulse	O
BOLD	O
response	O
.	O
</s>
<s>
But	O
they	O
also	O
found	O
deviations	O
from	O
the	O
linear	B-Algorithm
model	I-Algorithm
at	O
time	O
intervals	O
less	O
than	O
2seconds	O
.	O
</s>
<s>
A	O
source	O
of	O
nonlinearity	O
in	O
the	O
fMRI	B-Algorithm
response	O
is	O
from	O
the	O
refractory	O
period	O
,	O
where	O
brain	B-Application
activity	I-Application
from	O
a	O
presented	O
stimulus	O
suppresses	O
further	O
activity	O
on	O
a	O
subsequent	O
,	O
similar	O
,	O
stimulus	O
.	O
</s>
<s>
Researchers	O
have	O
checked	O
the	O
BOLD	B-Algorithm
signal	I-Algorithm
against	O
both	O
signals	O
from	O
implanted	O
electrodes	O
(	O
mostly	O
in	O
monkeys	O
)	O
and	O
signals	O
of	O
field	O
potentials	O
(	O
that	O
is	O
the	O
electric	O
or	O
magnetic	O
field	O
from	O
the	O
brain	O
's	O
activity	O
,	O
measured	O
outside	O
the	O
skull	O
)	O
from	O
EEG	B-Application
and	O
MEG	B-Algorithm
.	O
</s>
<s>
The	O
local	O
field	O
potential	O
,	O
which	O
includes	O
both	O
post-neuron-synaptic	O
activity	O
and	O
internal	O
neuron	O
processing	O
,	O
better	O
predicts	O
the	O
BOLD	B-Algorithm
signal	I-Algorithm
.	O
</s>
<s>
Activation	O
locations	O
detected	O
by	O
BOLD	O
fMRI	B-Algorithm
in	O
cortical	O
areas	O
(	O
brain	O
surface	O
regions	O
)	O
are	O
known	O
to	O
tally	O
with	O
CBF-based	O
functional	O
maps	O
from	O
PET	B-Application
scans	I-Application
.	O
</s>
<s>
Some	O
regions	O
just	O
a	O
few	O
millimeters	O
in	O
size	O
,	O
such	O
as	O
the	O
lateral	O
geniculate	O
nucleus	O
(	O
LGN	O
)	O
of	O
the	O
thalamus	O
,	O
which	O
relays	O
visual	O
inputs	O
from	O
the	O
retina	O
to	O
the	O
visual	O
cortex	O
,	O
have	O
been	O
shown	O
to	O
generate	O
the	O
BOLD	B-Algorithm
signal	I-Algorithm
correctly	O
when	O
presented	O
with	O
visual	O
input	O
.	O
</s>
<s>
In	O
the	O
rat	O
brain	O
,	O
single-whisker	O
touch	O
has	O
been	O
shown	O
to	O
elicit	O
BOLD	B-Algorithm
signals	I-Algorithm
from	O
the	O
somatosensory	O
cortex	O
.	O
</s>
<s>
However	O
,	O
the	O
BOLD	B-Algorithm
signal	I-Algorithm
cannot	O
separate	O
feedback	O
and	O
feedforward	O
active	O
networks	O
in	O
a	O
region	O
;	O
the	O
slowness	O
of	O
the	O
vascular	O
response	O
means	O
the	O
final	O
signal	O
is	O
the	O
summed	O
version	O
of	O
the	O
whole	O
region	O
's	O
network	O
;	O
blood	O
flow	O
is	O
not	O
discontinuous	O
as	O
the	O
processing	O
proceeds	O
.	O
</s>
<s>
Also	O
,	O
both	O
inhibitory	O
and	O
excitatory	O
input	O
to	O
a	O
neuron	O
from	O
other	O
neurons	O
sum	O
and	O
contribute	O
to	O
the	O
BOLD	B-Algorithm
signal	I-Algorithm
.	O
</s>
<s>
The	O
amplitude	O
of	O
the	O
BOLD	B-Algorithm
signal	I-Algorithm
does	O
not	O
necessarily	O
affect	O
its	O
shape	O
.	O
</s>
<s>
More	O
recent	O
characterization	O
of	O
the	O
BOLD	B-Algorithm
signal	I-Algorithm
has	O
used	O
optogenetic	O
techniques	O
in	O
rodents	O
to	O
precisely	O
control	O
neuronal	O
firing	O
while	O
simultaneously	O
monitoring	O
the	O
BOLD	O
response	O
using	O
high	O
field	O
magnets	O
(	O
a	O
technique	O
sometimes	O
referred	O
to	O
as	O
"	O
optofMRI	O
"	O
)	O
.	O
</s>
<s>
These	O
techniques	O
suggest	O
that	O
neuronal	O
firing	O
is	O
well	O
correlated	O
with	O
the	O
measured	O
BOLD	B-Algorithm
signal	I-Algorithm
including	O
approximately	O
linear	O
summation	O
of	O
the	O
BOLD	B-Algorithm
signal	I-Algorithm
over	O
closely	O
spaced	O
bursts	O
of	O
neuronal	O
firing	O
.	O
</s>
<s>
Linear	O
summation	O
is	O
an	O
assumption	O
of	O
commonly	O
used	O
event-related	B-Algorithm
fMRI	I-Algorithm
designs	O
.	O
</s>
<s>
Physicians	O
use	O
fMRI	B-Algorithm
to	O
assess	O
how	O
risky	O
brain	O
surgery	O
or	O
similar	O
invasive	O
treatment	O
is	O
for	O
a	O
patient	O
and	O
to	O
learn	O
how	O
a	O
normal	O
,	O
diseased	O
or	O
injured	O
brain	O
is	O
functioning	O
.	O
</s>
<s>
They	O
map	O
the	O
brain	O
with	O
fMRI	B-Algorithm
to	O
identify	O
regions	O
linked	O
to	O
critical	O
functions	O
such	O
as	O
speaking	O
,	O
moving	O
,	O
sensing	O
,	O
or	O
planning	O
.	O
</s>
<s>
Clinical	O
use	O
of	O
fMRI	B-Algorithm
still	O
lags	O
behind	O
research	O
use	O
.	O
</s>
<s>
Patients	O
with	O
brain	O
pathologies	O
are	O
more	O
difficult	O
to	O
scan	O
with	O
fMRI	B-Algorithm
than	O
are	O
young	O
healthy	O
volunteers	O
,	O
the	O
typical	O
research-subject	O
population	O
.	O
</s>
<s>
Drugs	O
such	O
as	O
antihistamines	O
and	O
even	O
caffeine	B-Application
can	O
affect	O
HR	O
.	O
</s>
<s>
Despite	O
these	O
difficulties	O
,	O
fMRI	B-Algorithm
has	O
been	O
used	O
clinically	O
to	O
map	O
functional	O
areas	O
,	O
check	O
left-right	O
hemispherical	O
asymmetry	O
in	O
language	O
and	O
memory	O
regions	O
,	O
check	O
the	O
neural	O
correlates	O
of	O
a	O
seizure	O
,	O
study	O
how	O
the	O
brain	O
recovers	O
partially	O
from	O
a	O
stroke	O
,	O
and	O
test	O
how	O
well	O
a	O
drug	O
or	O
behavioral	O
therapy	O
works	O
.	O
</s>
<s>
Recovered	O
depressed	O
patients	O
have	O
shown	O
altered	O
fMRI	B-Algorithm
activity	O
in	O
the	O
cerebellum	O
,	O
and	O
this	O
may	O
indicate	O
a	O
tendency	O
to	O
relapse	O
.	O
</s>
<s>
Pharmacological	O
fMRI	B-Algorithm
,	O
assaying	O
brain	B-Application
activity	I-Application
after	O
drugs	O
are	O
administered	O
,	O
can	O
be	O
used	O
to	O
check	O
how	O
much	O
a	O
drug	O
penetrates	O
the	O
blood	O
–	O
brain	O
barrier	O
and	O
dose	O
vs	O
effect	O
information	O
of	O
the	O
medication	O
.	O
</s>
<s>
These	O
studies	O
can	O
be	O
used	O
both	O
to	O
check	O
or	O
predict	O
human	O
results	O
and	O
to	O
validate	O
the	O
fMRI	B-Algorithm
technique	O
itself	O
.	O
</s>
<s>
The	O
goal	O
of	O
fMRI	B-Algorithm
data	O
analysis	O
is	O
to	O
detect	O
correlations	O
between	O
brain	O
activation	O
and	O
a	O
task	O
the	O
subject	O
performs	O
during	O
the	O
scan	O
.	O
</s>
<s>
Nevertheless	O
,	O
it	O
is	O
possible	O
to	O
predict	O
,	O
for	O
example	O
,	O
the	O
emotions	O
a	O
person	O
is	O
experiencing	O
solely	O
from	O
their	O
fMRI	B-Algorithm
,	O
with	O
a	O
high	O
degree	O
of	O
accuracy	O
.	O
</s>
<s>
The	O
five	O
main	O
sources	O
of	O
noise	O
in	O
fMRI	B-Algorithm
are	O
thermal	O
noise	O
,	O
system	O
noise	O
,	O
physiological	O
noise	O
,	O
random	O
neural	O
activity	O
and	O
differences	O
in	O
both	O
mental	O
strategies	O
and	O
behavior	O
across	O
people	O
and	O
across	O
tasks	O
within	O
a	O
person	O
.	O
</s>
<s>
Heat	O
causes	O
electrons	O
to	O
move	O
around	O
and	O
distort	O
the	O
current	O
in	O
the	O
fMRI	B-Algorithm
detector	O
,	O
producing	O
thermal	O
noise	O
.	O
</s>
<s>
It	O
affects	O
all	O
voxels	B-Algorithm
similarly	O
,	O
independent	O
of	O
anatomy	O
.	O
</s>
<s>
Transforming	O
this	O
into	O
voxels	B-Algorithm
introduces	O
some	O
loss	O
and	O
distortions	O
.	O
</s>
<s>
Head	O
movements	O
cause	O
the	O
voxel-to-neurons	O
mapping	O
to	O
change	O
while	O
scanning	O
is	O
in	O
progress	O
.	O
</s>
<s>
Since	O
fMRI	B-Algorithm
is	O
acquired	O
in	O
slices	O
,	O
after	O
movement	O
,	O
a	O
voxel	B-Algorithm
continues	O
to	O
refer	O
to	O
the	O
same	O
absolute	O
location	O
in	O
space	O
while	O
the	O
neurons	O
underneath	O
it	O
would	O
have	O
changed	O
.	O
</s>
<s>
This	O
consists	O
of	O
an	O
array	O
of	O
voxel	B-Algorithm
intensity	O
values	O
,	O
one	O
value	O
per	O
voxel	B-Algorithm
in	O
the	O
scan	O
.	O
</s>
<s>
The	O
voxels	B-Algorithm
are	O
arranged	O
one	O
after	O
the	O
other	O
,	O
unfolding	O
the	O
three-dimensional	O
structure	O
into	O
a	O
single	O
line	O
.	O
</s>
<s>
The	O
MR	O
scanner	O
acquires	O
different	O
slices	O
within	O
a	O
single	O
brain	O
volume	O
at	O
different	O
times	O
,	O
and	O
hence	O
the	O
slices	O
represent	O
brain	B-Application
activity	I-Application
at	O
different	O
timepoints	O
.	O
</s>
<s>
This	O
is	O
done	O
by	O
assuming	O
the	O
timecourse	O
of	O
a	O
voxel	B-Algorithm
is	O
smooth	O
when	O
plotted	O
as	O
a	O
dotted	O
line	O
.	O
</s>
<s>
Hence	O
the	O
voxel	B-Algorithm
's	O
intensity	O
value	O
at	O
other	O
times	O
not	O
in	O
the	O
sampled	O
frames	O
can	O
be	O
calculated	O
by	O
filling	O
in	O
the	O
dots	O
to	O
create	O
a	O
continuous	O
curve	O
.	O
</s>
<s>
When	O
the	O
head	O
moves	O
,	O
the	O
neurons	O
under	O
a	O
voxel	B-Algorithm
move	O
and	O
hence	O
its	O
timecourse	O
now	O
represents	O
largely	O
that	O
of	O
some	O
other	O
voxel	B-Algorithm
in	O
the	O
past	O
.	O
</s>
<s>
Hence	O
the	O
timecourse	O
curve	O
is	O
effectively	O
cut	O
and	O
pasted	O
from	O
one	O
voxel	B-Algorithm
to	O
another	O
.	O
</s>
<s>
Motion	O
correction	O
tries	O
different	O
ways	O
of	O
undoing	O
this	O
to	O
see	O
which	O
undoing	O
of	O
the	O
cut-and-paste	O
produces	O
the	O
smoothest	O
timecourse	O
for	O
all	O
voxels	B-Algorithm
.	O
</s>
<s>
Bias	O
field	O
estimation	O
is	O
a	O
real	O
preprocessing	O
technique	O
using	O
mathematical	O
models	O
of	O
the	O
noise	O
from	O
distortion	O
,	O
such	O
as	O
Markov	O
random	O
fields	O
and	O
expectation	B-Algorithm
maximization	I-Algorithm
algorithms	I-Algorithm
,	O
to	O
correct	O
for	O
distortion	O
.	O
</s>
<s>
In	O
general	O
,	O
fMRI	B-Algorithm
studies	O
acquire	O
both	O
many	O
functional	O
images	O
with	O
fMRI	B-Algorithm
and	O
a	O
structural	O
image	O
with	O
MRI	B-Algorithm
.	O
</s>
<s>
Even	O
when	O
whole-brain	O
analysis	O
is	O
done	O
,	O
to	O
interpret	O
the	O
final	O
results	O
,	O
that	O
is	O
to	O
figure	O
out	O
which	O
regions	O
the	O
active	O
voxels	B-Algorithm
fall	O
in	O
,	O
one	O
has	O
to	O
align	O
the	O
functional	O
image	O
to	O
the	O
structural	O
one	O
.	O
</s>
<s>
Typical	O
MRI	B-Algorithm
studies	O
scan	O
a	O
few	O
different	O
subjects	O
.	O
</s>
<s>
A	O
voxel	B-Algorithm
's	O
intensity	O
change	O
over	O
time	O
can	O
be	O
represented	O
as	O
the	O
sum	O
of	O
a	O
number	O
of	O
different	O
repeating	O
waves	O
with	O
differing	O
periods	O
and	O
heights	O
.	O
</s>
<s>
A	O
plot	O
with	O
these	O
periods	O
on	O
the	O
x-axis	O
and	O
the	O
heights	O
on	O
the	O
y-axis	O
is	O
called	O
a	O
power	B-General_Concept
spectrum	O
,	O
and	O
this	O
plot	O
is	O
created	O
with	O
the	O
Fourier	B-Algorithm
transform	I-Algorithm
technique	O
.	O
</s>
<s>
Temporal	O
filtering	O
amounts	O
to	O
removing	O
the	O
periodic	O
waves	O
not	O
of	O
interest	O
to	O
us	O
from	O
the	O
power	B-General_Concept
spectrum	O
,	O
and	O
then	O
summing	O
the	O
waves	O
back	O
again	O
,	O
using	O
the	O
inverse	O
Fourier	B-Algorithm
transform	I-Algorithm
to	O
create	O
a	O
new	O
timecourse	O
for	O
the	O
voxel	B-Algorithm
.	O
</s>
<s>
Smoothing	O
,	O
or	O
spatial	O
filtering	O
,	O
is	O
the	O
idea	O
of	O
averaging	O
the	O
intensities	O
of	O
nearby	O
voxels	B-Algorithm
to	O
produce	O
a	O
smooth	O
spatial	O
map	O
of	O
intensity	O
change	O
across	O
the	O
brain	O
or	O
region	O
of	O
interest	O
.	O
</s>
<s>
The	O
averaging	O
is	O
often	O
done	O
by	O
convolution	B-Language
with	O
a	O
Gaussian	O
filter	O
,	O
which	O
,	O
at	O
every	O
spatial	O
point	O
,	O
weights	O
neighboring	O
voxels	B-Algorithm
by	O
their	O
distance	O
,	O
with	O
the	O
weights	O
falling	O
exponentially	O
following	O
the	O
bell	O
curve	O
.	O
</s>
<s>
If	O
the	O
true	O
spatial	O
extent	O
of	O
activation	O
,	O
that	O
is	O
the	O
spread	O
of	O
the	O
cluster	O
of	O
voxels	B-Algorithm
simultaneously	O
active	O
,	O
matches	O
the	O
width	O
of	O
the	O
filter	O
used	O
,	O
this	O
process	O
improves	O
the	O
signal-to-noise	O
ratio	O
.	O
</s>
<s>
It	O
also	O
makes	O
the	O
total	O
noise	O
for	O
each	O
voxel	B-Algorithm
follow	O
a	O
bell-curve	O
distribution	O
,	O
since	O
adding	O
together	O
a	O
large	O
number	O
of	O
independent	O
,	O
identical	O
distributions	O
of	O
any	O
kind	O
produces	O
the	O
bell	O
curve	O
as	O
the	O
limit	O
case	O
.	O
</s>
<s>
One	O
common	O
approach	O
to	O
analysing	O
fMRI	B-Algorithm
data	O
is	O
to	O
consider	O
each	O
voxel	B-Algorithm
separately	O
within	O
the	O
framework	O
of	O
the	O
general	O
linear	B-Algorithm
model	I-Algorithm
.	O
</s>
<s>
One	O
then	O
assumes	O
a	O
specific	O
shape	O
for	O
the	O
HR	O
,	O
leaving	O
only	O
its	O
amplitude	O
changeable	O
in	O
active	O
voxels	B-Algorithm
.	O
</s>
<s>
The	O
design	O
matrix	O
and	O
this	O
shape	O
are	O
used	O
to	O
generate	O
a	O
prediction	O
of	O
the	O
exact	O
HR	O
of	O
the	O
voxel	B-Algorithm
at	O
every	O
timepoint	O
,	O
using	O
the	O
mathematical	O
procedure	O
of	O
convolution	B-Language
.	O
</s>
<s>
This	O
method	O
is	O
provably	O
optimal	O
if	O
the	O
error	O
were	O
distributed	O
as	O
a	O
bell	O
curve	O
,	O
and	O
if	O
the	O
scaling-and-summing	B-Algorithm
model	I-Algorithm
were	O
accurate	O
.	O
</s>
<s>
For	O
a	O
more	O
mathematical	O
description	O
of	O
the	O
GLM	O
model	O
,	O
see	O
generalized	O
linear	B-Algorithm
models	I-Algorithm
.	O
</s>
<s>
The	O
GLM	O
model	O
does	O
not	O
take	O
into	O
account	O
the	O
contribution	O
of	O
relationships	O
between	O
multiple	O
voxels	B-Algorithm
.	O
</s>
<s>
Whereas	O
GLM	O
analysis	O
methods	O
assess	O
whether	O
a	O
voxel	B-Algorithm
or	O
region	O
's	O
signal	O
amplitude	O
is	O
higher	O
or	O
lower	O
for	O
one	O
condition	O
than	O
another	O
,	O
newer	O
statistical	O
models	O
such	O
as	O
multi-voxel	O
pattern	O
analysis	O
(	O
MVPA	O
)	O
,	O
utilize	O
the	O
unique	O
contributions	O
of	O
multiple	O
voxels	B-Algorithm
within	O
a	O
voxel-population	O
.	O
</s>
<s>
In	O
a	O
typical	O
implementation	O
,	O
a	O
classifier	B-General_Concept
or	O
more	O
basic	O
algorithm	O
is	O
trained	O
to	O
distinguish	O
trials	O
for	O
different	O
conditions	O
within	O
a	O
subset	O
of	O
the	O
data	O
.	O
</s>
<s>
If	O
the	O
classifier	B-General_Concept
is	O
linear	O
,	O
then	O
the	O
training	O
model	O
is	O
a	O
set	O
of	O
weights	O
used	O
to	O
scale	O
the	O
value	O
in	O
each	O
voxel	B-Algorithm
before	O
summing	O
them	O
to	O
generate	O
a	O
single	O
number	O
that	O
determines	O
the	O
condition	O
for	O
each	O
testing	O
set	O
trial	O
.	O
</s>
<s>
More	O
information	O
on	O
training	O
and	O
testing	O
classifiers	B-General_Concept
is	O
at	O
statistical	B-General_Concept
classification	I-General_Concept
.	O
</s>
<s>
Another	O
method	O
used	O
the	O
same	O
fMRI	B-Algorithm
dataset	O
for	O
visual	O
object	O
recognition	O
in	O
the	O
human	O
brain	O
is	O
depending	O
on	O
multi-voxel	O
pattern	O
analysis	O
(	O
fMRI	B-Algorithm
voxels	B-Algorithm
)	O
and	O
multi-view	O
learning	O
which	O
is	O
described	O
in	O
,	O
this	O
method	O
used	O
meta-heuristic	O
search	O
and	O
mutual	O
information	O
to	O
eliminate	O
noisy	O
voxels	B-Algorithm
and	O
select	O
the	O
significant	O
BOLD	B-Algorithm
signals	I-Algorithm
.	O
</s>
<s>
It	O
is	O
common	O
to	O
combine	O
fMRI	B-Algorithm
signal	O
acquisition	O
with	O
tracking	O
of	O
participants	O
 '	O
responses	O
and	O
reaction	O
times	O
.	O
</s>
<s>
Physiological	O
measures	O
such	O
as	O
heart	O
rate	O
,	O
breathing	O
,	O
skin	O
conductance	O
(	O
rate	O
of	O
sweating	O
)	O
,	O
and	O
eye	O
movements	O
are	O
sometimes	O
captured	O
simultaneously	O
with	O
fMRI	B-Algorithm
.	O
</s>
<s>
The	O
method	O
can	O
also	O
be	O
combined	O
with	O
other	O
brain-imaging	O
techniques	O
such	O
as	O
transcranial	O
stimulation	O
,	O
direct	O
cortical	O
stimulation	O
and	O
,	O
especially	O
,	O
EEG	B-Application
.	O
</s>
<s>
The	O
fMRI	B-Algorithm
procedure	O
can	O
also	O
be	O
combined	O
with	O
near-infrared	O
spectroscopy	O
(	O
NIRS	O
)	O
to	O
have	O
supplementary	O
information	O
about	O
both	O
oxyhemoglobin	O
and	O
deoxyhemoglobin	O
.	O
</s>
<s>
The	O
fMRI	B-Algorithm
technique	O
can	O
complement	O
or	O
supplement	O
other	O
techniques	O
because	O
of	O
its	O
unique	O
strengths	O
and	O
gaps	O
.	O
</s>
<s>
It	O
can	O
noninvasively	O
record	O
brain	O
signals	O
without	O
risks	O
of	O
ionising	O
radiation	O
inherent	O
in	O
other	O
scanning	O
methods	O
,	O
such	O
as	O
CT	O
or	O
PET	B-Application
scans	I-Application
.	O
</s>
<s>
It	O
can	O
also	O
record	O
signal	O
from	O
all	O
regions	O
of	O
the	O
brain	O
,	O
unlike	O
EEG/MEG	O
,	O
which	O
are	O
biased	O
toward	O
the	O
cortical	O
surface	O
.	O
</s>
<s>
But	O
fMRI	B-Algorithm
temporal	O
resolution	O
is	O
poorer	O
than	O
that	O
of	O
EEG	B-Application
since	O
the	O
HR	O
takes	O
tens	O
of	O
seconds	O
to	O
climb	O
to	O
its	O
peak	O
.	O
</s>
<s>
Combining	O
EEG	B-Application
with	O
fMRI	B-Algorithm
is	O
hence	O
potentially	O
powerful	O
because	O
the	O
two	O
have	O
complementary	O
strengths	O
—	O
EEG	B-Application
has	O
high	O
temporal	O
resolution	O
,	O
and	O
fMRI	B-Algorithm
high	O
spatial	O
resolution	O
.	O
</s>
<s>
But	O
simultaneous	O
acquisition	O
needs	O
to	O
account	O
for	O
the	O
EEG	B-Application
signal	O
from	O
varying	O
blood	O
flow	O
triggered	O
by	O
the	O
fMRI	B-Algorithm
gradient	O
field	O
,	O
and	O
the	O
EEG	B-Application
signal	O
from	O
the	O
static	O
field	O
.	O
</s>
<s>
For	O
details	O
,	O
see	O
EEG	B-Application
vs	O
fMRI	B-Algorithm
.	O
</s>
<s>
While	O
fMRI	B-Algorithm
stands	O
out	O
due	O
to	O
its	O
potential	O
to	O
capture	O
neural	O
processes	O
associated	O
with	O
health	O
and	O
disease	O
,	O
brain	O
stimulation	O
techniques	O
such	O
as	O
transcranial	O
magnetic	O
stimulation	O
(	O
TMS	O
)	O
have	O
the	O
power	B-General_Concept
to	O
alter	O
these	O
neural	O
processes	O
.	O
</s>
<s>
The	O
current	O
state-of-the-art	O
setup	O
for	O
these	O
concurrent	O
TMS/fMRI	O
experiments	O
comprises	O
a	O
large-volume	O
head	O
coil	O
,	O
usually	O
a	O
birdcage	O
coil	O
,	O
with	O
the	O
MR-compatible	O
TMS	O
coil	O
being	O
mounted	O
inside	O
that	O
birdcage	O
coil	O
.	O
</s>
<s>
For	O
these	O
reasons	O
new	O
MR	O
coil	O
arrays	O
were	O
currently	O
developed	O
dedicated	O
to	O
concurrent	O
TMS/fMRI	O
experiments	O
.	O
</s>
<s>
Another	O
limitation	O
on	O
experimental	O
design	O
is	O
head	O
motion	O
,	O
which	O
can	O
lead	O
to	O
artificial	O
intensity	O
changes	O
of	O
the	O
fMRI	B-Algorithm
signal	O
.	O
</s>
<s>
Each	O
block	O
will	O
have	O
a	O
duration	O
of	O
a	O
certain	O
number	O
of	O
fMRI	B-Algorithm
scans	I-Algorithm
and	O
within	O
each	O
block	O
only	O
one	O
condition	O
is	O
presented	O
.	O
</s>
<s>
By	O
making	O
the	O
conditions	O
differ	O
in	O
only	O
the	O
cognitive	O
process	O
of	O
interest	O
,	O
the	O
fMRI	B-Algorithm
signal	O
that	O
differentiates	O
the	O
conditions	O
should	O
represent	O
this	O
cognitive	O
process	O
of	O
interest	O
.	O
</s>
<s>
The	O
increase	O
in	O
fMRI	B-Algorithm
signal	O
in	O
response	O
to	O
a	O
stimulus	O
is	O
additive	O
.	O
</s>
<s>
As	O
such	O
,	O
we	O
conclude	O
that	O
block	O
designs	O
offer	O
considerable	O
statistical	B-General_Concept
power	I-General_Concept
.	O
</s>
<s>
Event-related	B-Algorithm
designs	O
allow	O
more	O
real	O
world	O
testing	O
,	O
however	O
,	O
the	O
statistical	B-General_Concept
power	I-General_Concept
of	O
event	O
related	O
designs	O
is	O
inherently	O
low	O
,	O
because	O
the	O
signal	O
change	O
in	O
the	O
BOLD	O
fMRI	B-Algorithm
signal	O
following	O
a	O
single	O
stimulus	O
presentation	O
is	O
small	O
.	O
</s>
<s>
Both	O
block	O
and	O
event-related	B-Algorithm
designs	O
are	O
based	O
on	O
the	O
subtraction	O
paradigm	O
,	O
which	O
assumes	O
that	O
specific	O
cognitive	O
processes	O
can	O
be	O
added	O
selectively	O
in	O
different	O
conditions	O
.	O
</s>
<s>
Any	O
difference	O
in	O
blood	O
flow	O
(	O
the	O
BOLD	B-Algorithm
signal	I-Algorithm
)	O
between	O
these	O
two	O
conditions	O
is	O
then	O
assumed	O
to	O
reflect	O
the	O
differing	O
cognitive	O
process	O
.	O
</s>
<s>
In	O
fact	O
,	O
in	O
Stark	O
and	O
Squire	O
's	O
,	O
2001	O
study	O
When	O
zero	O
is	O
not	O
zero	O
:	O
The	O
problem	O
of	O
ambiguous	O
baseline	O
conditions	O
in	O
fMRI	B-Algorithm
,	O
activity	O
in	O
the	O
medial	O
temporal	O
lobe	O
(	O
as	O
well	O
as	O
in	O
other	O
brain	O
regions	O
)	O
was	O
substantially	O
higher	O
during	O
rest	O
than	O
during	O
several	O
alternative	O
baseline	O
conditions	O
.	O
</s>
<s>
Neuroimaging	O
methods	O
such	O
as	O
fMRI	B-Algorithm
and	O
MRI	B-Algorithm
offer	O
a	O
measure	O
of	O
the	O
activation	O
of	O
certain	O
brain	O
areas	O
in	O
response	O
to	O
cognitive	O
tasks	O
engaged	O
in	O
during	O
the	O
scanning	O
process	O
.	O
</s>
<s>
Perusal	O
of	O
the	O
discussion	O
sections	O
of	O
a	O
few	O
fMRI	B-Algorithm
articles	O
will	O
quickly	O
reveal	O
,	O
however	O
,	O
an	O
epidemic	O
of	O
reasoning	O
taking	O
the	O
following	O
form	O
:	O
</s>
<s>
The	O
most	O
common	O
risk	O
to	O
participants	O
in	O
an	O
fMRI	B-Algorithm
study	O
is	O
claustrophobia	O
and	O
there	O
are	O
reported	O
risks	O
for	O
pregnant	O
women	O
to	O
go	O
through	O
the	O
scanning	O
process	O
.	O
</s>
<s>
Implanted	O
medical	O
devices	O
such	O
as	O
pacemakers	B-Algorithm
could	O
malfunction	O
because	O
of	O
these	O
currents	O
.	O
</s>
<s>
However	O
,	O
genotoxic	O
(	O
i.e.	O
,	O
potentially	O
carcinogenic	O
)	O
effects	O
of	O
MRI	B-Algorithm
scanning	O
have	O
been	O
demonstrated	O
in	O
vivo	O
and	O
in	O
vitro	O
,	O
leading	O
a	O
recent	O
review	O
to	O
recommend	O
"	O
a	O
need	O
for	O
further	O
studies	O
and	O
prudent	O
use	O
in	O
order	O
to	O
avoid	O
unnecessary	O
examinations	O
,	O
according	O
to	O
the	O
precautionary	O
principle	O
"	O
.	O
</s>
<s>
In	O
a	O
comparison	O
of	O
genotoxic	O
effects	O
of	O
MRI	B-Algorithm
compared	O
with	O
those	O
of	O
CT	O
scans	O
,	O
Knuuti	O
et	O
al	O
.	O
</s>
<s>
reported	O
that	O
even	O
though	O
the	O
DNA	O
damage	O
detected	O
after	O
MRI	B-Algorithm
was	O
at	O
a	O
level	O
comparable	O
to	O
that	O
produced	O
by	O
scans	O
using	O
ionizing	O
radiation	O
(	O
low-dose	O
coronary	O
CT	O
angiography	O
,	O
nuclear	O
imaging	O
,	O
and	O
X-ray	O
angiography	O
)	O
,	O
differences	O
in	O
the	O
mechanism	O
by	O
which	O
this	O
damage	O
takes	O
place	O
suggests	O
that	O
the	O
cancer	O
risk	O
of	O
MRI	B-Algorithm
,	O
if	O
any	O
,	O
is	O
unknown	O
.	O
</s>
<s>
The	O
first	O
fMRI	B-Algorithm
studies	O
validated	O
the	O
technique	O
against	O
brain	B-Application
activity	I-Application
known	O
,	O
from	O
other	O
techniques	O
,	O
to	O
be	O
correlated	O
to	O
tasks	O
.	O
</s>
<s>
By	O
the	O
early	O
2000s	O
,	O
fMRI	B-Algorithm
studies	O
began	O
to	O
discover	O
novel	O
correlations	O
.	O
</s>
<s>
Still	O
their	O
technical	O
disadvantages	O
have	O
spurred	O
researchers	O
to	O
try	O
more	O
advanced	O
ways	O
to	O
increase	O
the	O
power	B-General_Concept
of	O
both	O
clinical	O
and	O
research	O
studies	O
.	O
</s>
<s>
MRI	B-Algorithm
,	O
in	O
general	O
,	O
has	O
better	O
spatial	O
resolution	O
than	O
EEG	B-Application
and	O
MEG	B-Algorithm
,	O
but	O
not	O
as	O
good	O
a	O
resolution	O
as	O
invasive	O
procedures	O
such	O
as	O
single-unit	O
electrodes	O
.	O
</s>
<s>
While	O
typical	O
resolutions	O
are	O
in	O
the	O
millimeter	O
range	O
,	O
ultra-high-resolution	O
MRI	B-Algorithm
or	O
MR	O
spectroscopy	O
works	O
at	O
a	O
resolution	O
of	O
tens	O
of	O
micrometers	O
.	O
</s>
<s>
Temporal	O
resolution	O
of	O
fMRI	B-Algorithm
is	O
limited	O
by	O
:	O
(	O
1	O
)	O
the	O
feedback	O
mechanism	O
that	O
raises	O
the	O
blood	O
flow	O
operating	O
slowly	O
;	O
(	O
2	O
)	O
having	O
to	O
wait	O
until	O
net	O
magnetization	O
recovers	O
before	O
sampling	O
a	O
slice	O
again	O
;	O
and	O
(	O
3	O
)	O
having	O
to	O
acquire	O
multiple	O
slices	O
to	O
cover	O
the	O
whole	O
brain	O
or	O
region	O
of	O
interest	O
.	O
</s>
<s>
This	O
spiral	O
imaging	O
sequence	O
acquires	O
images	O
faster	O
than	O
gradient-echo	O
sequences	O
,	O
but	O
needs	O
more	O
math	O
transformations	O
(	O
and	O
consequent	O
assumptions	O
)	O
since	O
converting	O
back	O
to	O
voxel	B-Algorithm
space	O
requires	O
the	O
data	O
be	O
in	O
grid	O
form	O
(	O
a	O
set	O
of	O
equally	O
spaced	O
points	O
in	O
both	O
horizontal	O
and	O
vertical	O
directions	O
)	O
.	O
</s>
<s>
Calcium-sensitive	O
agents	O
make	O
MRI	B-Algorithm
more	O
sensitive	O
to	O
calcium	O
concentrations	O
,	O
with	O
calcium	O
ions	O
often	O
being	O
the	O
messengers	O
for	O
cellular	O
signalling	O
pathways	O
in	O
active	O
neurons	O
.	O
</s>
<s>
Samuel	O
M	O
.	O
McClure	O
used	O
fMRI	B-Algorithm
to	O
show	O
the	O
dorsolateral	O
prefrontal	O
cortex	O
,	O
hippocampus	O
and	O
midbrain	O
were	O
more	O
active	O
when	O
people	O
knowingly	O
drank	O
Coca-Cola	O
as	O
opposed	O
to	O
when	O
they	O
drank	O
unlabeled	O
Coke	O
.	O
</s>
<s>
Other	O
studies	O
have	O
shown	O
the	O
brain	B-Application
activity	I-Application
that	O
characterizes	O
men	O
's	O
preference	O
for	O
sports	O
cars	O
,	O
and	O
even	O
differences	O
between	O
Democrats	O
and	O
Republicans	O
in	O
their	O
reaction	O
to	O
campaign	O
commercials	O
with	O
images	O
of	O
the	O
9/11	O
attacks	O
.	O
</s>
<s>
Another	O
is	O
Oxford	O
,	O
UK-based	O
Neurosense	O
,	O
which	O
advises	O
clients	O
how	O
they	O
could	O
potentially	O
use	O
fMRI	B-Algorithm
as	O
part	O
of	O
their	O
marketing	O
business	O
activity	O
.	O
</s>
<s>
At	O
least	O
two	O
companies	O
have	O
been	O
set	O
up	O
to	O
use	O
fMRI	B-Algorithm
in	O
lie	O
detection	O
:	O
No	O
Lie	O
MRI	B-Algorithm
and	O
the	O
Cephos	O
Corporation	O
.	O
</s>
<s>
No	O
Lie	O
MRI	B-Algorithm
charges	O
close	O
to	O
$5000	O
for	O
its	O
services	O
.	O
</s>
<s>
A	O
federal	O
magistrate	O
judge	O
in	O
Tennessee	O
prohibited	O
fMRI	B-Algorithm
evidence	O
to	O
back	O
up	O
a	O
defendant	O
's	O
claim	O
of	O
telling	O
the	O
truth	O
,	O
on	O
the	O
grounds	O
that	O
such	O
scans	O
do	O
not	O
measure	O
up	O
to	O
the	O
legal	O
standard	O
of	O
scientific	O
evidence	O
..	O
</s>
<s>
Most	O
researchers	O
agree	O
that	O
the	O
ability	O
of	O
fMRI	B-Algorithm
to	O
detect	O
deception	O
in	O
a	O
real	O
life	O
setting	O
has	O
not	O
been	O
established	O
.	O
</s>
<s>
Use	O
of	O
the	O
fMRI	B-Algorithm
has	O
been	O
left	O
out	O
of	O
legal	O
debates	O
throughout	O
its	O
history	O
.	O
</s>
<s>
Use	O
of	O
this	O
technology	O
has	O
not	O
been	O
allowed	O
due	O
to	O
holes	O
in	O
the	O
evidence	O
supporting	O
fMRI	B-Algorithm
.	O
</s>
<s>
First	O
,	O
most	O
evidence	O
supporting	O
fMRIs	B-Algorithm
accuracy	O
was	O
done	O
in	O
a	O
lab	O
under	O
controlled	O
circumstances	O
with	O
solid	O
facts	O
.	O
</s>
<s>
Lastly	O
,	O
there	O
is	O
an	O
ethical	O
question	O
relating	O
to	O
fMRI	B-Algorithm
scanning	O
.	O
</s>
<s>
This	O
testing	O
of	O
BOLD	O
has	O
led	O
to	O
controversy	O
over	O
if	O
fMRIs	B-Algorithm
are	O
an	O
invasion	O
of	O
privacy	O
.	O
</s>
<s>
Because	O
of	O
these	O
factors	O
and	O
more	O
,	O
fMRI	B-Algorithm
evidence	O
has	O
been	O
excluded	O
from	O
any	O
form	O
of	O
legal	O
system	O
.	O
</s>
<s>
Therefore	O
,	O
it	O
has	O
been	O
stated	O
that	O
fMRI	B-Algorithm
has	O
much	O
more	O
testing	O
to	O
do	O
before	O
it	O
can	O
be	O
considered	O
viable	O
in	O
the	O
eyes	O
of	O
the	O
legal	O
system	O
.	O
</s>
<s>
Some	O
scholars	O
have	O
criticized	O
fMRI	B-Algorithm
studies	O
for	O
problematic	O
statistical	O
analyses	O
,	O
often	O
based	O
on	O
low-power	O
,	O
small-sample	O
studies	O
.	O
</s>
<s>
Other	O
fMRI	B-Algorithm
researchers	O
have	O
defended	O
their	O
work	O
as	O
valid	O
.	O
</s>
<s>
In	O
2018	O
,	O
Turner	O
and	O
colleagues	O
have	O
suggested	O
that	O
the	O
small	O
sizes	O
affect	O
the	O
replicability	O
of	O
task-based	O
fMRI	B-Algorithm
studies	O
and	O
claimed	O
that	O
even	O
datasets	O
with	O
at	O
least	O
100	O
participants	O
the	O
results	O
may	O
not	O
be	O
well	O
replicated	O
,	O
although	O
there	O
are	O
debates	O
on	O
it	O
.	O
</s>
<s>
In	O
one	O
real	O
but	O
satirical	O
fMRI	B-Algorithm
study	O
,	O
a	O
dead	O
salmon	O
was	O
shown	O
pictures	O
of	O
humans	O
in	O
different	O
emotional	O
states	O
.	O
</s>
<s>
The	O
study	O
was	O
used	O
to	O
highlight	O
the	O
need	O
for	O
more	O
careful	O
statistical	O
analyses	O
in	O
fMRI	B-Algorithm
research	O
,	O
given	O
the	O
large	O
number	O
of	O
voxels	B-Algorithm
in	O
a	O
typical	O
fMRI	B-Algorithm
scan	O
and	O
the	O
multiple	B-General_Concept
comparisons	I-General_Concept
problem	I-General_Concept
.	O
</s>
<s>
Before	O
the	O
controversies	O
were	O
publicized	O
in	O
2010	O
,	O
between	O
25	O
and	O
40%	O
of	O
studies	O
on	O
fMRI	B-Algorithm
being	O
published	O
were	O
not	O
using	O
the	O
corrected	O
comparisons	O
.	O
</s>
<s>
In	O
2015	O
,	O
it	O
was	O
discovered	O
that	O
a	O
statistical	O
bug	O
was	O
found	O
in	O
the	O
fMRI	B-Algorithm
computations	O
which	O
likely	O
invalidated	O
at	O
least	O
40,000	O
fMRI	B-Algorithm
studies	O
preceding	O
2015	O
,	O
and	O
researchers	O
suggest	O
that	O
results	O
prior	O
to	O
the	O
bug	O
fix	O
cannot	O
be	O
relied	O
upon	O
.	O
</s>
<s>
In	O
2020	O
professor	O
Ahmad	O
Hariri	O
,	O
(	O
Duke	O
University	O
)	O
one	O
of	O
the	O
first	O
researchers	O
to	O
use	O
fMRI	B-Algorithm
,	O
performed	O
a	O
largescale	O
experiment	O
that	O
sought	O
to	O
test	O
the	O
reliability	O
of	O
fMRI	B-Algorithm
on	O
individual	O
people	O
.	O
</s>
<s>
In	O
the	O
study	O
,	O
he	O
copied	O
protocols	O
from	O
56	O
published	O
papers	O
in	O
psychology	O
that	O
used	O
fMRI	B-Algorithm
.	O
</s>
