00:00:00
hello everybody and welcome back so by
00:00:02
now we've spent most of our time looking
00:00:03
at how exactly we generate an MRI signal
00:00:06
we've seen that if we place an element
00:00:08
that has a non-zero spin within an
00:00:10
external magnetic field that element
00:00:12
will process at a frequency known as the
00:00:15
Llama frequency and that processional
00:00:17
frequency is dependent on the
00:00:18
gyromagnetic ratio of that element as
00:00:21
well as the strength of the external
00:00:22
magnetic field
00:00:24
we've also seen that if we apply a radio
00:00:26
frequency pulse perpendicular to that
00:00:28
main magnetic field and that radio
00:00:31
frequency pulse frequency matches the
00:00:33
processional frequency of the spins we
00:00:35
get what's known as nuclear magnetic
00:00:37
resonance where these spins start to
00:00:39
process or resonate in Phase with one
00:00:42
another as well as start to Fan out into
00:00:44
the transverse plane and what we get
00:00:47
then is gaining of transverse
00:00:49
magnetization as well as loss of
00:00:51
longitudinal or z-axis magnetization
00:00:54
and it's that transverse magnetization
00:00:56
that we can actually measure as signal
00:00:58
within our MRI machine
00:01:00
now once we've gained transverse
00:01:02
magnetization we can then stop the radio
00:01:05
frequency pulse once that radio
00:01:07
frequency pulse has been stopped two
00:01:09
independent processes happen T2
00:01:11
relaxation or loss of transverse
00:01:14
magnetization predominantly due to the
00:01:16
dephasing of spins as well as T1
00:01:19
relaxation or gain of longitudinal
00:01:21
magnetization
00:01:23
now different tissues have different
00:01:25
rates of both T2 and T1 relaxation and
00:01:29
it's the differences in those rates that
00:01:30
give us contrast within our image and
00:01:33
we've also seen how we can manipulate
00:01:34
the time to Echo or the te time as well
00:01:37
as the time to repetition or TR time to
00:01:39
weight our images where we can get
00:01:41
images that have predominantly T1
00:01:43
contrast differences shown in the image
00:01:45
or predominantly T2 contrast differences
00:01:48
shown in the image that we are measuring
00:01:50
that signal but we've got no way of
00:01:52
knowing where exactly that signal is
00:01:54
coming from MRI differs from x-ray or
00:01:57
ultrasound or CT in the fact that the
00:01:59
signal is actually being generated by
00:02:01
the patient it's coming from the patient
00:02:03
it's not like x-rays where we cast a
00:02:05
shadow onto a detector or ultrasound
00:02:07
where we wait for these sound waves to
00:02:09
come back and the time it takes for
00:02:11
those waves to come back allows us to
00:02:12
plot the depth of the tissue boundaries
00:02:14
here in MRI the signal is actually
00:02:16
coming from the patient and in order to
00:02:19
know where exactly in space that signal
00:02:21
is coming from we need to try and
00:02:23
separate the various different signals
00:02:25
on the Cartesian plane which we've
00:02:27
looked at before if we have three
00:02:29
different coordinate values a z-axis and
00:02:32
x-axis and a y-axis values if we know
00:02:35
those three coordinates on this
00:02:36
Cartesian plane as a frame of reference
00:02:38
we can say exactly where that signal is
00:02:41
coming from in space within the patient
00:02:43
and that's what's known as spatial
00:02:45
localization within MRI imaging now I'm
00:02:48
going to separate this into three
00:02:50
separate talks the first is what's going
00:02:52
to be known as slice selection here we
00:02:55
are trying to figure out where the
00:02:56
signal is coming from along the z-axis
00:02:59
or along the longitudinal plane of our
00:03:01
patient now if we look at this patient
00:03:03
within the MRI machine when you're
00:03:05
scrolling through an MRI image you're
00:03:07
looking at different slices that have
00:03:09
been stacked upon one another and you
00:03:11
can see that the slice that we select
00:03:12
along the z-axis of the patient has some
00:03:15
width to it so when you're looking at an
00:03:17
MRI image you're looking at a 2D image a
00:03:20
single slice that represents some width
00:03:22
there is some 3D data to those pixels on
00:03:25
your screen and in fact those pixels
00:03:27
represent what's known as a voxel
00:03:29
they've got some volume to them
00:03:32
now when we select the slides we're
00:03:34
selecting along the z-axis of our
00:03:36
Cartesian plane we're selecting along
00:03:38
this blue axis here and the gradient
00:03:41
that we use to select this slice is
00:03:43
what's known as the slice selection
00:03:45
gradient and that's going to be our
00:03:46
focus of today's talk
00:03:47
now if we take this slice out of the
00:03:49
patient and place it on the Cartesian
00:03:51
plane you can see the slice we're
00:03:54
selecting represents a z-axis value here
00:03:57
we can't separate the slice at the
00:04:00
moment into y values and X values now if
00:04:04
we look at this slice head on we can see
00:04:07
that that z-axis is coming in and out of
00:04:09
the screen just like we look at on our
00:04:11
MRI images
00:04:12
we then need a way to separate the
00:04:15
signal coming from this slice into
00:04:17
x-axis values and y-axis values so if we
00:04:20
have a patient here we have an organ and
00:04:22
we have a lesion in that organ we need
00:04:25
to know that that signal coming from the
00:04:26
lesion in the organ comes from a
00:04:28
specific x-coordinate and a specific
00:04:31
y-coordinate in that slice that we have
00:04:33
selected and so the next talk is going
00:04:35
to be looking at the x-axis coordinates
00:04:37
known as the frequency encoding gradient
00:04:39
and after that we're going to see how we
00:04:41
can differentiate the y-axis components
00:04:44
known as the phase encoding gradients
00:04:46
but for now let's focus on how exactly
00:04:48
we can select a specific slice within
00:04:51
the patient now if you take our patient
00:04:53
within the MRI machine here we've
00:04:55
applied a main magnetic field our B
00:04:58
naught magnetic field and the signal
00:05:00
coming from this patient is going to
00:05:01
come from those spins predominantly
00:05:03
dehydrogen protons within water and
00:05:05
within fat so let's substitute our
00:05:07
patient here for processing hydrogen
00:05:10
spins
00:05:11
now this constant B naught means that
00:05:14
these spins are all processing at the
00:05:16
same frequency known as the Llama
00:05:18
frequency the gyromagnetic ratio and the
00:05:20
strength of the magnetic field will
00:05:22
cause these spins to process at a
00:05:24
specific frequency
00:05:25
now in order to select a specific slice
00:05:28
we need these processing frequencies to
00:05:31
be different along the z-axis That's the
00:05:34
basis for slice selection gradient
00:05:36
now in order to change the processional
00:05:39
frequencies of these spins we need to
00:05:41
change the magnetic field strength along
00:05:44
the z-axis and we've seen this before
00:05:46
we've used gradient coils to apply a
00:05:49
gradient magnetic field along the z-axis
00:05:52
or the longitudinal axis of our patient
00:05:54
and that's exactly what we do as a first
00:05:56
step in slice selection we apply a
00:05:59
gradient field across the z-axis of the
00:06:02
patient and that gradient field means
00:06:04
that there's a differential in magnetic
00:06:06
field strength from one end of the
00:06:08
patient to the other end of the patient
00:06:10
now because that magnetic field strength
00:06:13
differs at different locations along the
00:06:15
z-axis we get different processional
00:06:18
frequencies along the z-axis so the
00:06:21
gradient field is causing these
00:06:23
frequencies to differ based on the
00:06:26
strength of the external magnetic field
00:06:28
now remember this line here is not
00:06:31
showing an angle change in the being
00:06:34
naught field it's showing a strength
00:06:36
change we can see that the strength of
00:06:38
the magnetic field changes along that
00:06:40
z-axis the direction of that magnetic
00:06:44
field is still purely along that z-axis
00:06:47
now what we can do is try and select a
00:06:50
specific slice based on these
00:06:52
processional frequencies we've seen that
00:06:54
when we apply a radio frequency pulse
00:06:56
that matches the processional frequency
00:06:59
that's when we get nuclear magnetic
00:07:01
resonance and we get flipping of those
00:07:03
spins into the transverse plane
00:07:06
now we can apply a radio frequency pulse
00:07:09
to the entire length of the patient and
00:07:11
only the spins that match that radio
00:07:13
frequency pulse will exhibit nuclear
00:07:16
magnetic resonance the other spins at
00:07:19
differing frequencies won't flip into
00:07:21
the transverse plane because those
00:07:23
processional frequencies don't match
00:07:25
that radio frequency pulse now when we
00:07:28
apply a radiofrequency pulse at a
00:07:30
certain frequency say 60 megahertz we
00:07:33
don't apply that radio frequency pulse
00:07:35
at exactly 60 megahertz there's a slight
00:07:37
range to that radio frequency pulse and
00:07:40
that's what's known as the radio
00:07:41
frequency bandwidth maybe it goes from
00:07:43
55 megahertz to 60 megahertz there's a 5
00:07:46
megahertz band at which that radio
00:07:48
frequency pulse is being applied to the
00:07:50
patient
00:07:51
so when we apply a radio frequency pulse
00:07:53
it's going to match up with certain
00:07:56
processional frequencies within the
00:07:58
patient here
00:07:59
now you can see that that radio
00:08:01
frequency pulse has some width to it
00:08:03
it's got a lower value and a higher
00:08:05
value there's a range known as the
00:08:07
bandwidth of that radio frequency pulse
00:08:10
now this gradient that we've drawn here
00:08:12
you can think of that gradient as
00:08:14
representing the different processional
00:08:16
frequencies here these processional
00:08:18
frequencies are proportional to the
00:08:20
strength of the magnetic field along the
00:08:22
z-axis here and this radio frequency
00:08:25
bandwidth is showing the range of radio
00:08:27
frequencies that we are exposing the
00:08:29
entire patient to within the MRI machine
00:08:33
now because this radio frequency matches
00:08:36
the processional frequency of this
00:08:38
specific slice we'll get nuclear
00:08:40
magnetic resonance within that
00:08:42
particular slice and we'll see that
00:08:44
those spins now gain transverse
00:08:47
magnetization and that gives us a signal
00:08:49
that we can actually measure these other
00:08:52
protons will not exhibit nuclear
00:08:54
magnetic resonance because the radio
00:08:56
frequency pulse frequency does not equal
00:08:59
their processional frequency here we've
00:09:01
selected a specific slice
00:09:03
now this is the basis for slide
00:09:06
selection now today we're going to look
00:09:08
at how we can move that slice along the
00:09:10
patient we want to image multiple points
00:09:12
along the z-axis and we want to see how
00:09:14
we can increase or decrease the
00:09:16
thickness of that slice so let's start
00:09:18
by looking at how we can move the slides
00:09:21
along the Z axis and that's what's known
00:09:22
as slice selection the first thing we
00:09:25
can do is actually change the radio
00:09:27
frequency pulse frequency if we increase
00:09:30
the radio frequency pulse frequency
00:09:32
we're going to match to a higher
00:09:34
processional frequency proton so let's
00:09:36
see what happens as we increase that
00:09:37
radio frequency pulse frequency we shift
00:09:40
our slides along the z-axis because
00:09:43
we're now selecting for a higher
00:09:44
processional frequency and these protons
00:09:47
are processing at a higher frequency
00:09:48
because of that gradient field there's a
00:09:51
higher magnetic field strength further
00:09:53
along the z-axis here now the second
00:09:56
thing that we can do is not change the
00:09:58
radio frequency pulse frequency but
00:10:00
change the gradient field itself if we
00:10:03
increase the magnetic field strength
00:10:05
that these protons experience we will
00:10:07
change the processional frequencies of
00:10:09
these protons so as we increase that
00:10:12
Baseline gradient magnetic field we'll
00:10:15
see that at the same radio frequency
00:10:17
pulse we'll be selecting a different
00:10:19
slice because now these processional
00:10:21
frequencies have changed due to that
00:10:23
increase in external magnetic field not
00:10:26
only can we change the slice that we're
00:10:28
selecting by changing the radio
00:10:30
frequency pulse frequency or changing
00:10:32
the gradient field or external magnetic
00:10:35
field we can in theory also move the
00:10:37
patient along the z-axis and as we move
00:10:41
the patient the slice that we're
00:10:43
selecting will stay the same region
00:10:45
within our z-axis but the part of the
00:10:47
patient that we're Imaging will change
00:10:49
as that patient moves along the z-axis
00:10:51
so there are three different ways that
00:10:53
we can select the slice that we are
00:10:55
trying to image
00:10:56
now we selected a specific slice say
00:10:59
this slice here and we got nuclear
00:11:01
magnetic resonance only occurring along
00:11:03
this slide now how do we go about
00:11:05
changing the thickness of that slice
00:11:07
remember as we increase the thickness
00:11:09
we'll have more protons that are
00:11:11
experiencing nuclear magnetic resonance
00:11:13
more resonating protons within the
00:11:15
transverse plane and ultimately getting
00:11:17
more signal we will lose some z-axis
00:11:20
resolution but we'll be getting more
00:11:22
signal and there's certain times where
00:11:23
we want our slice to be thicker or our
00:11:25
slice to be thinner now the first way
00:11:27
that we can increase the slice thickness
00:11:29
is by changing the bandwidth of the
00:11:32
radio frequency pulse if we increase the
00:11:35
ranges of the radio frequency pulse
00:11:37
frequencies we are going to be getting
00:11:39
more nuclear magnetic resonance because
00:11:41
we're covering a wider range of
00:11:43
processional frequencies here now you
00:11:46
can see our slices got thicker because
00:11:48
this bandwidth has got thicker we are
00:11:50
covering a wider range of processional
00:11:53
frequencies
00:11:54
now at the same increased bandwidth if
00:11:57
we wanted to decrease the slice
00:11:59
thickness what we could do is actually
00:12:01
change the gradient of the external
00:12:04
magnetic field if we increase that
00:12:07
gradient we make the difference between
00:12:09
the magnetic field at the one end of the
00:12:11
Z axis and the other end of the z-axis
00:12:13
we make that difference bigger we can
00:12:15
see that we ultimately narrow the slice
00:12:18
thickness we've still got that increased
00:12:20
bandwidth but the range of radio
00:12:22
frequency pulses Falls along a smaller
00:12:25
part of this gradient graph here and
00:12:28
again we get a smaller slice thickness
00:12:30
so you can see that changing the radio
00:12:32
frequency pulse bandwidth as well as
00:12:34
changing the steepness of the gradient
00:12:36
magnetic field both play a role in these
00:12:39
slice thickness that we are selecting in
00:12:40
the z-axis now as I mentioned before
00:12:43
this slice has some thickness to it and
00:12:46
we know that the protons within this
00:12:48
slice aren't perfectly on top of one
00:12:50
another in the z-axis there's some width
00:12:52
to that slice now with that width comes
00:12:56
a particular problem and that's what's
00:12:57
known as slice phase you'll see that the
00:13:00
gradient field that is experienced at
00:13:02
this part of the slice will be different
00:13:05
from the gradient field experience at
00:13:07
this part of the slice there's still a
00:13:09
gradient field occurring here that is
00:13:11
being covered by the entire radio
00:13:13
frequency pulse bandwidth now as we
00:13:16
allow those spins to spin we will see
00:13:18
that although they are resonating in
00:13:21
Phase with one another they are
00:13:23
resonating at a frequency that is
00:13:24
dependent on the radio frequency pulse
00:13:26
and the gradient field that these spins
00:13:29
are experiencing and you can see that
00:13:31
these spins are out of phase with one
00:13:33
another because of this differential
00:13:35
ingredients here
00:13:37
now what we can do is apply what's known
00:13:39
as a re-phasing gradient after we've
00:13:42
applied our slice selection gradient
00:13:44
here now the rephasing gradient means
00:13:47
that we will apply an equal and opposite
00:13:49
gradient in the other direction along
00:13:51
the z-axis that equal and opposite
00:13:54
gradient will allow these spins to now
00:13:57
spin in Phase with one another initially
00:14:00
this side of the slice was experiencing
00:14:02
a low magnetic field and once we applied
00:14:05
the rephasing gradient it experienced a
00:14:08
high magnetic field and if we average
00:14:10
our slice selection gradient and the
00:14:12
rephasing gradient out that entire slice
00:14:15
that we've selected will have
00:14:17
experienced these same amount of
00:14:18
external magnetic field allowing these
00:14:21
spins to spin in Phase with one another
00:14:23
and that's what's known as the rephasing
00:14:24
gradient
00:14:25
so let's now recap the entire process
00:14:28
that has allowed us to select a specific
00:14:30
slice along the z-axis of our patient
00:14:34
now the MRI machine constantly has a b
00:14:37
naught an external magnetic field along
00:14:40
the z-axis that never gets Switched Off
00:14:43
no matter which type of pulse sequence
00:14:45
you're doing there will always be a b
00:14:47
naught along the z-axis of the patient
00:14:50
and that's why when we represent what's
00:14:52
happening along our specific pulse
00:14:54
sequence we don't actually have a line
00:14:56
here for B naught it's always there
00:14:59
whether we're taking an image or not
00:15:00
that MRI machine is on and that b naught
00:15:03
is on
00:15:04
now we've seen that when we want to
00:15:06
generate signal within our MRI image we
00:15:09
need to apply a radio frequency pulse
00:15:12
and here we've applied a 90 degree radio
00:15:14
frequency pulse these graphs if you're
00:15:16
unfamiliar with them represent time
00:15:18
along this axis here now we apply a
00:15:22
radial frequency pulse for a specific
00:15:24
period of time that causes the spins to
00:15:27
flip to 90 degrees in the transverse
00:15:29
plane they've got maximum transverse
00:15:31
magnetization
00:15:32
at that same time we need to be applying
00:15:36
this slide selection gradient the
00:15:38
gradient that we've been looking at
00:15:39
throughout this talk that gradient along
00:15:41
the z-axis of the patient now because we
00:15:44
are applying the slice selection
00:15:45
gradient this 90 degree RF pulse that is
00:15:49
released at a specific frequency or at
00:15:51
least a specific frequency bandwidth
00:15:53
will only cause certain spins to exhibit
00:15:56
nuclear magnetic resonance and flip into
00:15:58
the transverse plane and it's only those
00:16:01
spins with a processional frequency that
00:16:04
matches the frequency of the radio
00:16:05
frequency pulse that will flip and
00:16:07
that's the basis for selecting the
00:16:09
specific slice
00:16:10
we then looked at why we need to apply a
00:16:14
re-phasing gradient after the slice
00:16:16
selection gradient to account for the
00:16:19
differences in the spins along that
00:16:21
thickness of the slice and this re-phase
00:16:24
ingredient will mean that all the spins
00:16:26
within that slice along the entire Z
00:16:28
axis of that slice will be resonating in
00:16:31
Phase with one another
00:16:32
now as you'll remember from the t2
00:16:35
relaxation talk we then apply a 180
00:16:38
degree radio frequency pulse and that
00:16:41
180 degree radio frequency pulse will
00:16:44
allow those spins to start to re-phase
00:16:46
with one another and account for the t2
00:16:48
star differences within the tissues now
00:16:51
if that sounds like Greek to you go back
00:16:53
to that T2 relaxation talk and see why
00:16:56
we apply this 180 degree radio frequency
00:16:59
pulse now at the same time difference
00:17:02
between the 90 and 180 degree radio
00:17:04
frequency pulse we then sample the
00:17:07
signal within our tissue at this point
00:17:09
all the signal that we measure in the
00:17:12
MRI machine is coming from this specific
00:17:15
slice because only this specific slice
00:17:18
has been tipped into the transverse
00:17:21
plane
00:17:22
now a lot of people get confused this
00:17:24
slice selection gradient the gradient
00:17:27
that we're applying along the z-axis is
00:17:30
only on while the radio frequency pulse
00:17:33
is on
00:17:34
the times between the radio frequency
00:17:37
poles or between the slice selection
00:17:39
gradient the only magnetic field at this
00:17:42
stage that the patient is experiencing
00:17:44
is the main magnetic field we flip those
00:17:48
spins into the transverse plane we then
00:17:50
switch off the radio frequency pulse and
00:17:53
we switch off the slice selection
00:17:55
gradient we get that T2 and T1
00:17:58
relaxation of those spins and at this
00:18:00
time to Echo when we sample we are only
00:18:03
measuring the differences in the t2
00:18:06
relaxation at this time to Echo
00:18:09
now some people get confused thinking
00:18:11
but if we've switched the slice
00:18:13
selection gradient off how do we know
00:18:15
that that signal is coming from exactly
00:18:17
this slice while all the other spins
00:18:20
here throughout this pulse sequence have
00:18:22
just been processing with the main
00:18:24
magnetic field and when we apply that
00:18:27
slice selection gradient those
00:18:28
processional frequencies aren't matching
00:18:30
up with this radio frequency pulse so
00:18:32
they never flip into the transverse
00:18:34
plane so even though the slice selection
00:18:37
gradients and the radio frequency pulse
00:18:39
has been switched off prior to the time
00:18:41
to Echo it's only those that were
00:18:44
selected during this period of time that
00:18:47
will exhibit transverse magnetization
00:18:49
and remember we can only measure
00:18:51
transverse magnetization
00:18:53
so now we've got to the stage where
00:18:55
we've selected a specific slice within
00:18:58
our patient and that slice covers one
00:19:01
transverse plane across the patient now
00:19:04
the signal that we're measuring within
00:19:06
the coils of our MRI machine is coming
00:19:08
from that entire slice there's no way at
00:19:11
the moment for that MRI machine to
00:19:13
figure out where along that slice this
00:19:15
signal is coming from all those net
00:19:18
transverse magnetization vectors are
00:19:20
being added to and taken away from one
00:19:22
another and we're just getting one long
00:19:24
Trace coming into our coil as those
00:19:27
spins resonate within the slice we need
00:19:30
to now figure out a way in which we can
00:19:31
take that signal that is being generated
00:19:33
from this slice and tease out where
00:19:36
exactly that signal is coming from along
00:19:38
the x-axis as well as along the y-axis
00:19:41
of this particular slice at the moment
00:19:44
with the pulse sequence that we've done
00:19:46
here we've got all of these spins
00:19:49
resonating at the same resonance
00:19:52
frequency in Phase with one another now
00:19:55
they're resonating at the same frequency
00:19:56
because they are experiencing these same
00:19:59
external magnetic field they're
00:20:01
experiencing that being naught magnetic
00:20:03
field and at this time to Echo they have
00:20:06
lost a certain amount of transverse
00:20:07
magnetization and gained a certain
00:20:09
amount of longitudinal magnetization now
00:20:12
these spins are all spinning in Phase
00:20:14
coming from this entire slice and we've
00:20:17
got no way as it currently stands of
00:20:19
differentiating where that signal is
00:20:22
coming from along the x-axis as well as
00:20:24
the y-axis now in the next talk we're
00:20:27
going to add another line to our pole
00:20:30
sequence here that will allow us to
00:20:32
differentiate where the signals are
00:20:33
coming from along the x-axis of that
00:20:36
slice once we figured out where the
00:20:38
signal is coming from along the x-axis
00:20:40
of the slice we need to then add another
00:20:42
line to our pulse sequence in order then
00:20:45
to differentiate where the signal is
00:20:47
coming from along the y-axis of our
00:20:49
slice and those two happen in completely
00:20:51
separate processes so we're going to
00:20:53
talk about them in two separate talks
00:20:55
and hopefully by the end of these three
00:20:57
talks we'll be able to understand
00:20:59
exactly how an MRI machine knows exactly
00:21:02
where the specific signal is coming from
00:21:05
within the patient now these three
00:21:07
sections are very confusing and you
00:21:09
might need to go back to them multiple
00:21:11
times and as you've seen you might need
00:21:13
to actually revisit the talks that we've
00:21:15
done prior to these I'd encourage you to
00:21:17
spend a lot of time making sure you
00:21:19
understand the concepts before moving on
00:21:22
to the next talk and again if you're
00:21:24
studying for a specific Physics Exam
00:21:26
I've Linked In The Top Line in the
00:21:28
description below question banks that
00:21:29
I've curated from past papers that allow
00:21:32
you to test yourself and see where your
00:21:34
knowledge gaps are and where you need to
00:21:36
focus on more before for your exam so go
00:21:38
and check those out if you are studying
00:21:40
for a specific exam otherwise I'll see
00:21:42
you all in the next talk where we're
00:21:43
going to look at x-axis frequency
00:21:46
encoding gradients so until then goodbye
00:21:48
everybody
