Taxonomy of MR Imaging Sequences

Click on one of the Red or Green buttons below to view more information

Spin-echo sequences are shown in red, gradient-echo sequences in green.

The single-echo symbol indicates a sequence in which a single echo is
acquired after a single repetition time, the multiecho symbol indicates a
sequence in which multiple echoes are acquired in a single repetition time
period. Click on a circular sequence node to see an overview of that
sequence and links to the physics and applications associated with that
sequence.

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on the links to read more about that topic.

RARE: Overview
RARE: Rapid Acquisition with Relaxation Enchancement
Synonyms
Turbo spin echo (TSE)
Fast spin echo (FSE)

Siemens Medical Solutions, Philips Medical Systems

Toshiba, GE Medical Systems, Hitachi

Background: In a conventional spin-echo sequence, only one line of k space is acquired

after the single 180 refocusing pulse in each TR period. As a result, conventional spin-echo
imaging is slow, particularly for T2-weighted sequences, in which TR values are long. Even
"short" TRs are relatively long, to give longitundinal magnetization a chance to recover from
the 90 flip. The RARE sequence is based on conventional spin echo but uses multiple
refocusing pulses to acquire several lines of k space in a single TR period.
Features: Allows fast imaging while retaining the advantage of low sensitivity to
susceptibility artifacts and magnet inhomogeneities seen with spin-echo sequences. Speed
allows 3D T2-weighted imaging.
Artifacts/Problems: RARE imaging is prone to blurring, ghosts, and edge enchancement.
Since many 180 pulses are transmitted in a single TR, RARE sequences can have a high
specific absorption rate.
Unusual aspects: Compared to conventional spin echo, the RARE sequence shows brighter
fat at T2-weighted imaging.

GO TO RARE: Physics
GO TO RARE: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on the links to read more about that topic.

RARE: Physics
Timing diagram
A timing diagram for a 2D RARE acquisition is shown in Figure 1. Note the multiple 180
pulses during each TR and the way in which phase-encoding changes from TR to TR encode
each echo to a new line in k space. Every second lobe on the phase-encoding gradient is a
"rewinder": an equal and opposite gradient that undoes the phase encoding of the prior
lobe.

Figure 1: Timing diagram for a RARE sequence. Several refocusing

pulses are transmitted during each TR period. Each pulse refocuses an
echo. The number of echoes in a single TR is called the echo train length
or "turbo factor" (4 in the example shown). As several lines in k space are
now filled during a single TR, imaging time is reduced compared to that
of a conventional spin-echo sequence by the turbo factor.

Effective TE
As in conventional spin-echo imaging, contrast in a RARE sequence is controlled by TE
and TR. It is clear what TR is in a RARE sequence: simply the time between 90 RF pulses.
But what about TE? Every echo in the echo train seems to have its own TE. In the RARE
sequence, the "effective TE" is the TE associated with the echo that is written closest to the
center of k space. Because the center lines of k space have the most effect on contrast, the
TE of this echo is the effective TE (TEeff) for the sequence (Fig 2).

Figure 2: Echo spacing and echo train length or turbo

factor in the RARE sequence.

Artifacts
Note how the echoes in a RARE echo train fall off in amplitude with T2 decay (Fig 3). This
effect is particularly pronounced where tissue T2 is short and the echo train long. This
results in non-uniform weighting of lines stored in k space and image blurring. In a T2weighted RARE sequence, the effective TE must be long. Later echoes fill the inner lines of k
space, and earlier echoes fill the outer lines. The echoes filling the outer lines have decayed
less. This overemphasizes the data stored in the outer lines in k space, which contain detail
on the edges of the image. As result, an artifact called edge enchancement may be seen.

Figure 3: Decay of echoes according to T2 during a RARE

sequence.

Shared-echo (dual-contrast) RARE

In a shared-echo sequence, the earlier echoes in the train are used to fill k space for
a proton-density-weighted image (ie, a short effective TE) and the later echoes to fill
k space for a T2-weighted image (ie, long effective TE). The principle is illustrated in
Figure 4.

Figure 4: Shared-echo RARE sequence. The echoes

written into the center of k space for the proton-density
image have a shorter effective TE than the echoes
written into the center of k space for the T2-weighted
image. Some outer-line echoes (shown here in red) are
shared to save time.

GO TO RARE: Overview
GO TO RARE: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on the links to read more about that topic.

RARE: Applications
Applications of the RARE sequence include T1- or T2-weighted high-resolution neurologic and
orthopedic imaging and T2-weighted breath-hold examinations of the abdomen (Figs 13).

Figure 1: T2-weighted image in the transverse plane at the level of the ventricles demonstrates a
necrotic tumor in the right frontal lobe with marked surrounding vasogenic edema.

Figure 2: T2-weighted RARE image of the knee in the coronal plane.

Figure 3: Shared-echo RARE images of a carpal tunnel. Proton-density (bottom) and T2weighted (top) images are obtained in the same acquisition.

GO TO RARE: Overview
GO TO RARE: Physics
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on the links to read more about that topic

HASTE: Overview
HASTE: Half-Fourier-Acquired Single-Shot Turbo Spin Echo
Synonyms
HASTE

Siemens Medical Solutions

SS-FSE (single-shot fast spin echo)

GE Medical Systems, Philips

Medical Systems

FSE-ADA (fast spin echo, asymmetric data allocation

with half scan)

Hitachi

FASE (fast advanced spin echo)

Toshiba

Background: Filling all of k space in a single shot with a RARE sequence is possible.
However, considerable T2 decay takes place over the relatively long period of acquisition,
and single-shot RAREis effectively restricted to imaging fluid (as in MR
cholangiopancreatography examinations). The HASTE sequence is an adaptation
of RARE that reduces total acquisition time by acquiring only half of k space. Half filling of k
space is sufficient to generate an image.
Features: The HASTE sequence allows acquisitions of 2D slices in less than a second.
Because of its long echo train (all the echoes take place in a single TR), HASTE images
will be T2-weighted.
Artifacts/Problems: Blurring artifact and limited resolution.

GO TO HASTE: Physics
GO TO HASTE: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic

HASTE: Applications
The HASTE sequence provides strong T2 weighting and is used in MR
cholangiopancreatography and MR urography to demonstrate stones and
other filling defects as areas of low signal intensity within bright fluid-filled
structures (Figure). High speed makes the HASTE sequence a suitable choice
for uncooperative patients.

Transverse image through the liver demonstrates stones.

GO TO HASTE: Overview
GO TO HASTE: Physics
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic

HASTE: Physics
Symmetry in k space
Figure 1 shows filled space after an MR image acquisition (the values in k space are plotted
as gray levels).This symmetric property of k space means it is necessary to fill only half of k
space; the rest of the data can be inferred. The HASTE sequence uses this property of k
space to its advantage.

Figure 1: Try to spot the symmetry in k space. If you're having difficulty

(it's not easy!), use the button on the image to reveal k-space symmetry.

Half-Fourier in phase
HASTE imaging uses a technique called half-Fourier in phase. It collects only half the lines of
k space (Fig 2) and infers the rest. (In practice, a little more that half of k space is filled to
correct for some errors in the data).

Figure 2: Because only half of k space is filled, the echo

train length for the sequence need be only half as long.

Single shot
HASTE is a single-shot technique, acquiring sufficient data for an
entire image in only one TR (Fig 3).

Figure 3: Filling of k space in the HASTE sequence. The animation is looped:

Every half filling of k-space represents acquisition of an image.

GO TO HASTE: Overview
GO TO HASTE: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link read more about that topic

Spin Echo: Overview

Spin Echo
Synonyms
Spin echo (SE)

Siemens Medical Solutions, GE Medical

Systems, Philips Medical Systems, Hitachi,
Toshiba

Background: In spin-echo imaging, tissue is subjected to a 90 excitation pulse followed

by a delay of TE/2 and the application of a 180 refocusing pulse. The refocusing pulse
rephases transverse components of magnetization to form an echo at time TE.
Features: Basic robust sequence. T1, T2, and proton-density weighting all possible with
appropriate choice of TE and TR. Spin-echo imaging shows sensitivity to T2 rather than to
T2*, making it less sensitive to magnetic susceptibility artifacts.
Artifacts/Problems: In the conventional form of the sequence, only one line of k space is
collected during every TR period, resulting in long imaging times, particularly for T2weighted imaging, in which TR values need to be long. Has been largely replaced by fastspin-echo (RARE) sequences, especially for T2-weighted imaging.

GO TO SPIN-ECHO: Physics
GO TO SPIN-ECHO: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

Spin Echo: Physics

The fundamental signature of a conventional spin-echo sequence is a 90-180 RF pulse pair, repeated at intervals
of TR. The spacing between the 90 and 180 pulses is TE/2, and an echo forms at time TE (Fig 1).

Figure 1: Basic elements of a spin-echo sequence

TR and TE
The main user-variable parameters of interest in a spin-echo sequence are TR and TE. Use the interactive image
(Fig 2) to get a sense of how TR and TE affect the timing of the various elements of the sequence.

Figure 2: TR and TE in a spin-echo sequence.

Dephasing and the FID

The 90 pulse flips the longitudinal magnetization into the transverse plane. Immediately, the individual
spins that make up the transverse component begin to dephase (Fig 3).

Figure 3: Dephasing of tranverse components of magnetization

As a result of the dephasing, the signal picked up by the receiver coil decays in the form of a free
induction decay (FID). Dephasing is caused by true T2 dephasing and other dephasing effects caused by
magnet non-uniformities and susceptibility variations. The combination of all these effects causes signal
decay with a time constant T2* (ie, the FID has an exponential envelope with a time constant T2*).

Figure 4: FID curve.

The 180 rephasing flip is only fully effective when (as in Fig 5) the degree of dephasing before and after
the application of the 180 flip is the same. Dephasing caused by suceptibility effects and magnet nonuniformites are quite stable and so can be reversed by the 180 flip. Dephasing caused by true T2 decay
is random in time and is not reversed by the rephasing pulse. As a result, the echo peak lies on a T2
decay curve (Fig 6) .

Figure 5: Rephasing of transverse magnetization by

a 180 pulse.

Figure 6: The echo peak in the spin-echo sequence

lies on the T2 decay curve.

T1 and T2 in spin-echo imaging

Longitudinal magnetization (Mz): At every 90 flip, the longitudinal magnetization is flipped into the
transverse plane. Longitudinal magnetization then starts to recover according to the T1 of the tissue.
Transverse magnetization: The available transverse magnetization (ie, the magnetization than can be
rephased to form an echo) decays according to the T2 of tissue.

Figure 7: Longitundinal and transverse component behavior with

application of a 90 flip. (Changes in sign at the 180 flip are omitted for clarity.)

Echo amplitude
The echo amplitude contributed by a particular tissue voxel determines how bright that voxel will appear in
the final image. In Figure 8, note how TE and TR influence echo amplitude. The echo peak lies on a T2
decay curve. T1 and TR determine how much longitudinal recovery occurs during TR.

Figure 8: Echo amplitude variation with TR and TE.

Contrast in spin-echo imaging

Controlling TE and TR determines the relative echo amplitude of signal from tissues with different T1 and
T2 values. The animation below compares the effect of changing TE and TR on echoes from Tissue
A and two other tissues, one with the same T2 and the other with the same T1.

Figure 9: Image contrast in spin-echo imaging. T1 weighting: Use the sliders to set the shortest possible
TE and TR. Note how the echo amplitude differs mainly between the two tissues with the different T1
values. T2 weighting: Use the sliders to set the longest possible TE and TR values. Note how the echo
amplitude differs mainly between the two tissues with the different T2 values.

Full timing diagram

A full timing diagram for a conventional spin-echo sequence is shown in Figure 10.

Figure 10: Timing diagram for a conventional spin-echo sequence.

Imaging time
Note how only one line in k space is filled per TR. If there are n lines in k space the imaging time is TR xn,
where n is the number of pixels in the phase-encoding direction. Note how the phase-encoding gradient
changes at each TR in order to fill a new line in k space.

GO TO SPIN-ECHO: Overview
GO TO SPIN-ECHO: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

Spin Echo: Applications

T2-weighted conventional spin-echo imaging is time-consuming (long TR for
weighting and only one line of k space filled per TR). In practice, it has been
replaced by RARE imaging. T1-weighted conventional spin-echo is used for
acquiring anatomic images in moderate time periods and at high resolution
(Figure).

T1-weighted conventional spin-echo image of the cervical spine

demonstrates normal marrow.

GO TO SPIN-ECHO: Overview
GO TO SPIN-ECHO: Physics
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

Inversion Recovery: Overview

IR: Inversion Recovery
Synonyms
Siemens Medical
Solutions

IR, TurboIR, True IR, TIRM (turbo IR, magnitude

reconstruction)
IR, IR-TSE (IR Turbo Spin Echo), Real IR

Philips Medical Systems

IR, FSEIR (fast spin echo inversion recovery)

GE Medical Systems

IR, Fast IR
IR, TIR (turbo inversion recovery)

Hitachi
Toshiba

Background: The term inversion recovery refers to a class of sequences that begin each
TR with an inverting 180 RF pulse. After a delay TI (the inversion time), a host sequence is
initiated. The host sequence may be one of several sequence types. In practice, IR is
frequently combined with a fast spin-echo (eg, RARE) host sequence, which gives rise to
names such as turboIR, turbo FSE, and fast IR.
Features: The longitudinal magnetization at the start of the host sequence is determined
by the degree of recovery of the inverted magnetization. Contrast in an IR image is strongly
influenced by TI and tissue T1 (ie, the image is T1-weighted).
Artifacts/Problems: Imaging times can be long, since the TR value must be long enough
to allow recovery between inverting pulses.
Oddities: Two reconstruction methods are possible with IR: standard "magnitude"
reconstruction and "real" reconstruction (also called true IR or real IR). Magnitude IR
sequences show a behavior called contrast reversal, in which the normal appearance of
tissues on a T1-weighted image (bright for short-T1 tissue and dark for long-T1 tissue) is
reversed as the TI is reduced.

GO TO IR: Physics
GO TO IR: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to more about that topic.

Inversion Recovery: Applications

The inversion-recovery sequence is used to introduce T1 weighting. With real
reconstruction, high contrast can be obtained between gray and white matter in the brain
(Figure).

In this real IR reconstruction, the background gray represents the "zero" gray level. Darker
areas on the image represent "negative" signal and brighter areas "positive" signal.

GO TO IR: Overview
GO TO IR: Physics
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

Inversion Recovery: Physics

Inversion-recovery basics
Inversion-recovery (IR) sequences start with an inverting 180 pulse. This is a form of
magnetization preparation. The longitudinal magnetization (Mz) for tissues with different T1 values
will be different at the end of this preparatory phase (Fig 1).

Figure 1: The animation shows the recovery of the

longitudinal magnetization of two different tissues in response to a 180 pulse.
Host sequence and TI
The 180 pulse is a preparatory pulse before the imaging sequence proper starts. The imaging part
of the sequence is called the host sequence. The host sequence could be any one of a number of
sequences, such as RARE or EPI (Fig 2).

Figure 2: The delay between the inversion pulse and the first excitation pulse in the host sequence
is called the inversion time. The longitudinal magnetization available to the host sequence depends
on TI and tissue T1.

Contrast behavior
The TR of an IR sequence will normally be long. This is necessary for longitudinal magnetiztion to
recover between inversion pulses. If the echo time in the host sequence is short, the primary
determinant of contrast behavior is the relative longitudinal magnetisation at time TI. Images will
therefore tend to be T1-weighted with contrast behavior controlled by TI.
Magnitude versus real reconstruction
Two image reconstruction methods are possible in IR imaging: magnitude and real (also called
phase-sensitive reconstruction) (Fig 3).

Figure 3: Pixel intensites in a magnitude image are influenced only by the magnitude of the
longitudinal component of magnetization at time TI, not by its direction. Pixel intensites in a real
image are influenced both by the magnitude and direction of the longitudinal component at time TI.

Contrast reversal
Magnitude-reconstructed IR images show a type of behavior called contrast reversal (Fig 4). The
real image shown displays typical T1-weighted image behaviour for all values of TI: The shorter-T1
white matter looks brighter than the longer-T1 gray matter.
The magnitude image behaves differently. At a certain value of TI, the gray and white matter
become isointense. At longer values of TI, the image behaves as expected: The white matter
appears brighter than the gray matter. However, at shorter values of TI, contrast reversal is seen,
with white matter appearing darker than gray matter.

Figure 4: Contrast reversal in IR imaging.

GO TO IR: Overview
GO TO IR: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

FLAIR: Overview
FLAIR: Fluid-attenuated Inversion Recovery
Synonyms
FLAIR

Siemens Medical Solutions, Philips Medical

Systems, GE Medical Systems, Toshiba,
Hitachi

Background: The FLAIR sequence is an inversion recovery technique that

uses a long TI for suppressing fluid signal.
Features: In T2-weighted imaging, fluids will be bright because of their long
T2 values. In some situations this can reduce the visibility of other lesions. In
these cases, FLAIR can suppress fluid signal and improve lesion visibility.
Typically FLAIR is used with a RARE host sequence.
Artifacts/Problems: FLAIR requires a long TR, resulting in long imaging
times.

GO TO FLAIR: Physics
GO TO FLAIR: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

FLAIR: Applications
The FLAIR sequence makes periventricular changes easier to visualize by
suppressing bright cerebrospinal fluid (Figure). In addition, cortical changes,
which can also be hard to visualize with T2 sequences, are made more
prominent.

Transverse image at the level of the thalamii demonstrates periventricular high

signal intensity due to myelinolysis secondary to arteriolosclerosis.

GO TO FLAIR: Overview
GO TO FLAIR: Physics
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

FLAIR: Physics
The basic elements of a FLAIR sequence are shown in Figure 1. A 180 pulse is applied, followed
by a TI delay. The host sequence could be a conventional spin-echo sequence. However, as TR
values are long (to let longitundinal magnetization recover), fast host sequences are preferred to
reduce imaging time. A RARE host sequence is typical.

Figure 1: A FLAIR sequence timing diagram.

Continued on next page

Fluid Suppression
After the inverting pulse in FLAIR imaging, longitudinal magnetization begins to recover according
to T1. The magnetization of a given tissue passes through zero after a time equal to approximately
69% of the tissue T1 value. If the TI is selected to equal this time, there is no magnetization
available from that tissue at the start of the host sequence (Fig 2). Signal from the tissue is
therefore suppressed. Unlike in the STIR sequence, most non-nulled tissues are not inverted at
time TI, so contrast reversal is not apparent (see Fig 1, STIR Physics).

Figure 2: In FLAIR imaging, the TI is chosen to match the zero crossing point of fluid (often
cerebrospinal fluid). In a 1.5-T magnet, the T1 of cerebrospinal fluid is approximately 4200
msec, so a TI of approximately 2900 msec will suppress the fluid signal. (In practice, the
exact nulling time depends on the parameters of the host seqeuence; the given figures are
for a theoretically infinite TR.)

GO TO FLAIR: OVERVIEW
GO TO FLAIR: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link read more about that topic.

STIR: Overview
STIR: Short TI Inversion Recovery
Synonyms
Siemens Medical Solutions, Philips
Medical Systems, GE Medical
STIR
Systems, Hitachi, Toshiba

Background: The STIR sequence is an inversion-recovery (IR) technique

that uses a short inversion time (TI) to suppress fat signal.
Features: Useful for suppressing fat, particularly in T1-weighted images and
in RARE T2-weighted images (where fat is abnormally bright). Unlike fatsaturation techniques, STIR is relatively insensitive to poor magnet shim and
susceptibility effects (eg, in presence of a metal implant).
Artifacts/Problems: As an IR technique, STIR requires a relatively long TR.
To improve speed, STIR is frequently used with a RARE host sequence. Any
tissue with a similar T1 to that of fat will also be suppressed with STIR.
Therefore, STIR is not used with contrast agent, where the target lesion may
have a T1 close to that of fat after contrast agent uptake. In a real
reconstruction, STIR shows contrast reversal.

GO TO STIR: Physics
GO TO STIR: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

STIR: Applications
The STIR sequence is useful for suppressing marrow fat to reveal bone
abnormalities (Figure).

T2-weighted STIR view of the ankle joint in the sagittal plane

demonstrates marked marrow edema in the distal tibia due to
osteoid osteoma.

GO TO STIR: Overview
GO TO STIR: Physics
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

STIR: Physics
STIR sequence
The basic elements of a STIR sequence are shown in Figure 1. A 180 pulse is applied, followed by
a TI delay. The host sequence could be a conventional spin-echo sequence. However, since TR
values are long in spin-echo sequences (in order to let longitundinal magnetization recover), fast
host sequences are preferred to reduce imaging time. A RARE host sequence is typical.

Figure 1: Longitudinal magnetization in the STIR sequence.

Continued on next page

Fat signal supression

After the inversion pulse in STIR, longitudinal magnetization begins to recover according to
T1. The magnetization of a given tissue passes through zero after a time equal to
approximately 69% of the tissue T1 value (Fig 2). If the TI is selected to equal this time, no
magnetization is available from that tissue at the start of the host sequence, and signal
from that tissue is suppressed.

Figure 2: In STIR imaging, TI is chosen to match the zero crossing point of fat. In a 1.5-T
imager, the T1 of fat is approximately 230 msec, so a TI of approximately 160 msec will
suppress fat signal. (In practice, the exact nulling time depends on parameters of the host
sequence; these figures are for a theoretically infinite TR). Note from the animation how the
Mz of other tissue (in this case white matter) is also reduced at time TI, and gives reduced
signal in the host sequence.

GO TO STIR: Overview
GO TO STIR: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

Gradient/Spin Echo Hybrid: Overview

Gradient/Spin Echo Hybrid
Synonyms
TurboGSE (turbo gradient spin echo)

Siemens Medical Solutions

GRASE (gradient echo and spin

echo)

GE Medical Systems, Philips Medical

Systems

Hybrid EPI

Toshiba

Background: This is a hybrid sequence that brings together features

of RARE and EPI sequences. As in RARE, multiple 180 pulses are used to generate
multiple spin echoes from an initial excitation pulse. In addition, multiple gradient echoes
are generated by gradient switching after each 180 pulse. This allows several lines of k
space to be filled per refocusing pulse.
Features: Fast sequence mainly used for T2-weighted imaging. Improved resolution
relative to RARE for a given imaging time. Less image distortion than in EPI. Can be used
to give improved T2*-weighted imaging compared to RARE and hence better sensitivity to
hemorrhagic lesions. RARE imaging gives unusually bright fat, whereas fat appearance in
GRASE/TurboGSE imaging is more like that in conventional spin-echo imaging. The short
inter180 pulse spacing in RARE causes the fat-brightening effect; in GRASE/TurboGSE,
this spacing can be longer, as multiple lines of k-space are filling after each 180 pulse. The
longer gap between 180 pulses in GRASE/TurboGSE results in a lower specific
absosrption rate.
Artifacts/Problems: The sequence is subject to a strong ghosting artifact.

GO TO Gradient/Spin Echo Hybrid: Physics

GO TO Gradient/Spin Echo Hybrid: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

Gradient/Spin Echo Hybrid: Physics

Timing diagram
Multiple gradient echoes are generated from each spin echo by gradient fields. In the example, the typical
case of three gradient echoes per spin echo is shown (Figure).
The trajectory through k space is a combination of row-to-row movements between RF pulses and larger
jumps between echo formations. Row-to-row movements are carried out by a conventional phase-encoding
gradient of serially changing strength. The "jumps" are carried out by application of fixed-strength phase
encodings. Siemens Medical Solutions calls the number of spin echoes (ie, no. of 180 pulses) the turbo
factor (as in turbo spin echo). The number of gradient echoes per spin echo is called the EPI factor. At the
extreme of an EPI factor of 1, the sequence becomes a RARE sequence, whereas for a turbo factor of 1, the
sequence becomes a spin-echo EPI sequence.

GRASE/TurboGSE sequence. During

acquisition, echoes are undergoing T2
decay. There are sharp discontinuities in k
space between rows acquired at different
degrees of T2 decay (ie, the boundaries
between the colored regions in the
animation). This causes a ghosting
artifact.

GO TO Gradient/Spin Echo Hybrid: Overview

GO TO Gradient/Spin Echo Hybrid: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

Gradient/Spin Echo Hybrid: Applications

The turbo gradient echo/spin echo hybrid sequence is used in high-resolution
T2-weighted brain (Figure) and orthopedic imaging.

T2-weighted transverse image of a normal brain at the level of the thalamii.

GO TO Gradient/Spin Echo Hybrid: Overview

GO TO Gradient/Spin Echo Hybrid: Physics
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on the red text to read more about that topic.

Gradient Echo: Overview

Gradient Echo
Synonyms
Gradient-recalled echo
(GRE)

Siemens Medical Solutions

Gradient echo

Hitachi

Field echo

GE Medical Systems, Philips Medical Systems,

Toshiba

Background: The term gradient echo refers to an entire class of sequences. These
sequences use a gradient to generate an RF echo, rather than a 180 refocusing pulse, as
is the case in spin-echo sequences. Gradient-echo sequences generally use an excitation
flip angle of less than 90.
Features: Gradient-echo sequences allow faster imaging than spin-echo sequences and
are frequently used in fast 3D imaging. The echo peak of a gradient echo lies on a T2*
curve, rather than on a T2 curve as in spin-echo imaging. Since the T2* value of tissue is
sensitive to variations in magnetic field strength, gradient-echo images are sensitive to
factors perturbing the static field. A benefit of this is good sensitivity to hemorraghic lesions,
since magnetic suceptibility is disturbed by hemoglobin degradation products.
Artifacts/Problems: Being sensitive to T2*, gradient-echo sequences are more prone to
susceptibility artifact. At a tissue/air boundary or in the presence of a metal implant, this
may manifest as a signal void on the image.

GO TO Gradient Echo: Physics

GO TO Echo: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on the link to read more about that topic.

Gradient Echo: Physics

Excitation pulse
Spin-echo sequences use a 90 excitation pulse. Gradient-echo sequences generally use a smaller
flip angle as the excitation pulse. The 90 flip used in spin-echo imaging turns the longitundinal
component of magnetization completely into the transverse plane. This is great for signal
acquisition, since a strong transverse component is needed for signal, but not so good for imaging
time. Because the longitudinal component has to recover from zero, TR must be relatively long. The
smaller flip angle used in gradient-echo imaging doesn't "wipe out" the longitudinal component, so a
shorter TR can be used (Fig 1). With a moderate flip angle, the transverse component is still quite
strong immediately after the flip.

Figure 1: Set a flip angle of 90 in the

animation and observe the magnitude
of the transverse component (yellow)
immediately after the flip, and the time
it takes for the longitudinal component
(red) to recover. Then do the same for
a flip angle of ~45.

T2* decay
Immediately after the application of an RF flip in the gradient-echo sequence, a transverse
component appears. This transverse component dephases and decays according to T2*,
generating an FID (Fig 2). T2* is a combination of true tissue T2 and decay caused by magnetspecific non-uniformities and local variations in magnetic suceptibility (eg, at an air/tissue interface).

Figure 2: Generation of an FID. The

sinusoidal form of the FID is a result of
the transverse component spinning at
the Larmor frequency. The decaying
envelope of the signal is a result of
dephasing of the transverse component
of individual spins.

Dephasing and rephasing with gradients

The gradient-echo sequence drives the FID dephasing even faster by applying a gradient. The
dephasing introduced by the gradient is then reversed and an echo generated. Use the animation in
Figure 3 to study how a gradient mimics the normal T2* dephasing effect. Note how reversing the
gradient reverses the gradient-induced dephasing.

Figure 3: The animation shows the

transverse component for five spins
precessing at the Larmor
frequency.The location of the spins
in a gradient field is indicated at the
top left. Apply a positive gradient
and note how the resultant change
in frequency causes dephasing.
Then apply a negative gradient and
see how the spins are rephased.

Echo formation
In a gradient-echo sequence, the frequency-encoding gradient generates the echo. The first lobe of
the gradient dephases the FID rapidly; the second lobe rephases the FID. Note now the echo peak
lies on the T2* decay curve (ie, the envelope of the original FID) (Fig 4). In spin-echo imaging, the
echo peak lies on the T2 decay curve. A basic gradient-echo sequence is sensitive to anything that
disturbs T2* (eg, poor magnetic uniformity, susceptibility effects due to metal prostheses, presence
of hemoglobin degradation products in hemorrhage).

Figure 4: The gradient echo lies

under the T2* curve.

Basic gradient-echo sequence

A generic 2D gradient-echo sequence timing diagram is shown in Figure 5. In clinical practice,
gradient-echo sequences build on this basic format and fall into a number of subcategories.

Figure 5: Timing diagram for a basic gradient-echo sequence.

GO TO Gradient Echo: Overview

GO TO Gradient Echo: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

Gradient Echo: Applications

The gradient-echo sequences used in practice split into many varieties that differ in terms of
contrast behavior, speed, and susceptibility to artifact. The sensitivity of gradient echoes to T2* can
be both beneficial and detrimental (Figs 1, 2). Note that not all sequences classed as "gradient
echo" show this classic sensitivity to T2*. Some, like balanced SSFP can demonstrate T2 rather
than T2* weighting.

Figure 1: T2* sensitivity as a source of artifact in gradient-echo images. This fatsuppressed T2-weighted gradient-echo image shows susceptibility artifact from a
left hip prosthesis.

Figure 2: T2* sensitivity as a diagnostic aid in gradient-echo imaging. This coronal

T2-weighted gradient-echo image in a patient with hemochromatosis demonstrates
that iron deposition causes marked signal loss due to suceptibility effects..

The speed and variety of contrast behavior in gradient-echo imaging allow a range of
applications not possible with spin-echo imaging (Figs 3, 4).

Figure 3: MIP from a time of Flight (TOF) MR angiographic study of the circle of Willis. TOF
sequences use gradient echoes to generate high signal intensity from fresh blood flowing into a
slice or slab, while background stationary tissue becomes relatively saturated.

Figure 4: Balanced SSFP gradient-echo image demonstrates that high

speed and good blood-tissue contrast allow cine acquistions of the cardiac cycle.

GO TO Gradient Echo: Overview

GO TO Gradient Echo: Physics
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

Spoiled Gradient Echo: Overview

Incoherent or Spoiled Gradient Echo
Synonyms
FLASH (fast low-angle shot)
T1-FFE (T1-weighted fast field echo)
SPGR (spoiled gradient echo)
Gradient echo, normal, with low flip angle
T1-FFE (T1-weighted fast field echo)

Siemens Medical
Solutions
Philips Medical Systems
GE Medical Systems
Hitachi
Toshiba

Background: Spoiled gradient-echo sequences destroy or "spoil" any residual transverse

magnetization remaining at the end of each TR. In that way, every RF flip applied in a
sequence acts solely on longitudinal magnetization.
Features: Used in fast T1-weighted imaging. The short TRs allow 3D and breath-hold 2D
acquisitions.

GO TO Gradient Echo: Physics

GO TO Gradient Echo: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

Spoiled Gradient Echo: Physics

Residual transverse magnetization
The RF flip applied at the start of each TR in a gradient-echo sequence creates a
transverse component of magnetization. This decays according to T2*. Since TRs are
relatively short in gradient-echo imaging, there is normally some residual transverse
magnetization left at the end of the TR period (Fig 1). In "spoiled" or incoherent gradientecho imaging this residual transverse magnetization is destroyed. This prevents the
transverse magnetization from building up and feeding back to the longitudinal
magnetization after each flip.

Figure 1: The animation shows the residual transverse magnetization remaining after
application of a single RF flip (flip angle is 35 in this case).

Spoiling
In RF spoiling, the phase of each RF pulse is varied from TR to TR. Use the animation in
Figure 2 to examine what is meant by the phase of the RF pulse. At each TR, the "new"
transverse component will be out of phase with any residual component from the last TR.
This prevents the RF flip from feeding the transverse component, and so a transverse
steady-state magnetization cannot build up. Gradient spoiling is an alternative method of
spoiling the transverse conponent. By applying a gradient, the transverse component can
be dephased. In practice, RF spoiling produces more uniform and effective spoiling.

Figure 2: The animation shows

how the phase of the RF pulse
determines the phase of the
transverse component at each
flip. In a spoiled gradient-echo
sequence, the RF phase is
changed in a systematic manner
at each TR, preventing a build
up of transverse magnetization.

Steady state
A normal gradient-echo sequence consists of repeated RF flips at intervals TR. In a spoiled
gradient-echo sequence, we need only consider the longitudinal component after each
applied flip, since the transverse component is deliberately destroyed. After a sufficient
number of RF pulses, the longitudinal magnetization settles into a steady state (Fig 3).

Figure 3: In the animation, the yellow graph shows the magnitude of the longitundinal
magnetization (Mz) in response to a series of RF pulses. Each sharp drop is due to an RF
pulse. The interpulse increases in magnitude are due to T1 recovery. Mz eventually settles
into a steady state, where recovery just offsets the effect of the applied flip at each TR. Try
varying the flip angle. Note how this affects the amplitude of the steady state.

Ernst angle
For any given ratio of TR/T1, there is an optimum flip angle that will give the strongest
transverse component at each applied RF pulse and therefore the strongest signal. This is
called the Ernst angle. The concept of the Ernst angle can be explored by using the
animation in Figure 4.

Figure 4: The yellow graph shows the longitudinal magnetization at the start of a spoiled
gradient-echo sequence. The orange spikes show the transverse magnetization
immediately after each flip. Try varying the flip angle. Note, for example, when the flip angle
is 90, all the longitudinal magnetization appears in the transverse plane after the first flip.
Very little longitundinal magnetization will have recovered for the next flip. The Ernst angle
in this case (TR/T1 = 0.1) is approximately 25. Set the flip angle to about 25 and note
how this flip angle gives the highest transverse signal once the system reaches steady
state.

Timing diagram
Figure 5 shows the basic elements of a 2D spoiled gradient-echo sequence.

Figure 5: Timing diagram of a FLASH sequence. Note that the phase-encoding gradient is
"rewound" at the end of each TR by an equal and opposite gradient. Without a rewinder,
the effectiveness of RF spoiling would vary across the image.

GO TO Spoiled Gradient Echo: Overview

GO TO Spoiled Gradient Echo: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

Spoiled Gradient Echo: Applications

The fast low-angle shot (FLASH) sequence finds application in contrast-enhanced MR
angiography, where T1 weighting makes the T1-shortened blood appear bright (Fig 1)

Figure 1: Source image (left) and MIP (right) in contrast-enhanced MR angiography of the
renal arteries. Short imaging times combined with centric k-space filling allow acquisition of
pure arterial images.

The FLASH sequence is sensitive to the susceptibiliy effect of hemoglobin

breakdown products. As such it can be used to detect hemosiderin deposition from
remote trauma (Fig 2).

Figure 2: Transverse FLASH image of the brain above the level of the ventricles
demonstrates an old area of hemorrhage in the right frontal lobe.

Transverse vectors of fat and water fall in and out of phase periodically because of "chemical
shift." Chemical shift is the small difference in precession frequency between fat and water
(~210 Hz at 1.5 T). This may be applied usefully in in-phase and out-of-phase images
generated by running two FLASH sequences at slightly different TE values (2.4 and 4.8 msec
at 1.5 T). For voxels containing fat/water mixtures, the signal is reduced in the out-of-phase
image relative to the in-phase image. In spin-echo imaging, rephasing by the 180 pulse does
not allow this phenomenon to be demonstrated. The technique is useful for detecting fatcontaining structures such as the adrenal adenoma shown in Figure 3.

Figure 3: In-phase (left) and out-of-phase (right) FLASH images show adreanl adenoma.

GO TO SPIN-ECHO: Overview
GO TO Gradient Echo: Physics
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

Ultrafast Gradient Echo: Overview

Ulltrafast Gradient Echo
Synonyms
TurboFLASH
TFE (turbo field echo)
Fast SPGR
RSSG (RF spoiled SARGE [steady
state acquisition with rewound
gradient echo])
T1-FFE (fast field echo)

Siemens Medical Solutions,

GE Medical Systems,
Philips Medical Systems,
Hitachi,
Toshiba

Background: Gradient-echo sequences that acquire an image in less than a second are
considered ultrafast sequences. Achieving adequate signal and contrast in a simple very
short TE, very short TR gradient-echo sequence is a problem; for good T1 contrast, a
larger flip angle is needed, but the optimal flip angle for a good signal-to-noise ratio (SNR)
(ie, the Ernst angle) at short TRs is small. Ultrafast sequences overcome this problem by
incorporating a magnetization preparation phase prior to the gradient-echo host. Typically,
a 180 RF pulse is applied in the preparatory phase. The 180 flip introduces increased T1
weighting. All lines in k space are acquired after this single preparatory pulse.
Features: Used for high-speed T1-weighted imaging

GO TO Ultrafast Gradient Echo: Physics

GO TO Ultrafast Gradient Echo: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

Ultrafast Gradient Echo: Applications

The ultrafast gradient-echo sequence can acquire T1-weighted images in less than a second. This
forms the basis of "fluoroscopic" type methods of detecting the arrival of contrast agent in a region
of interest. The live image allows the operator to trigger imaging once contrast agent is seen
(Figure).

Figure 1: Contrast agent arrival imaged with a turbo FLASH technique.

GO TO Ultrafast Gradient Echo: Overview

GO TO Ultrafast Gradient Echo: Physics
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

Ultrafast Gradient Echo: Physics

Turbo FLASH timing diagram
In the ultrafast gradient-echo sequence, a preparatory phase is applied, during which the
magnetization is flipped with a 180 RF pulse (Fig 1). This is followed by repeated RF alpha
pulses, with a k-space line acquisition at each echo.

Figure 1: Timing diagram for an ultrafast gradient-echo sequence (eg, turbo FLASH). The
basic structure of the sequence is a preparatory phase (shaded purple) followed by
repeated gradient-echo acquisitions (shaded green). In, for example, an image with 256
pixels in the phase-encoding direction, these acquisitions will be repeated 256 times to
complete k-space filling. The entire length of image acquisition will be on the order of a
second or less. In the animation, only the first three lines of k space are shown being filled.

TI in turbo FLASH
TI is the time to the midpoint of the acquisition phase of the sequence, where the middle
lines of k-space are being filled (Fig 2).

Figure 2: Note how the inversion pulse introduces T1 weighting. TI corresponds to the time
of filling of the middle lines in k space. (The plot of Mz is somewhat simplified: The effect of
the repeated RF flips on recovery during the acquisition period has been omitted).

GO TO Ultrafast Gradient Echo: Overview

GO TO Ultrafast Gradient Echo: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

3D Ultrafast Gradient Echo: Overview

3D Ulltrafast Gradient Echo
Synonyms
MP-RAGE (magnetization-prepared
rapid acquisition gradient echo)
3D TFE (turbo field echo)
3D fast SPGR (spoiled gradientrecalled acquisition in the steady
state)

Siemens Medical Solutions,

GE Medical Systems,
Philips Medical Systems,
Hitachi,
Toshiba

3D RSSG (radiofrequency-spoiled
steady state acquisition with
rewound gradient echo)
3D FFE (fast field echo)

Background: This is the 3D variant of the ultrafast gradient-echo sequence (eg, turbo
FLASH [fast low-angle shot], TFE, FFE).
As in two dimensions, a preparatory pulse introduces T1 weighting. Usually, the imaging
time for a 3D gradient-echo sequence would be too long for a single preparatory pulse at
the start of the sequence to maintain its effect. Instead, several preparatory pulses are
applied over the course of the acquisition, with new lines in k space acquired after each
pulse.
Features: Used for high-speed 3D T1-weighted imaging.

GO TO SPIN-ECHO: Physics
GO TO SPIN-ECHO: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on the red text to read more about that topic.

3D Ultrafast Gradient Echo: Applications

Three-dimensional ultrafast gradient-echo sequences are used to produce high-resolution T1weighted 3D images of the brain. Imaging times are several minutes.

Sagittal T1-weighted MP-RAGE (magnetization-prepared rapid acquisition gradient-echo) image

of the brain

GO TO 3D Ultrafast Gradient Echo: Overview

GO TO 3D Ultrafast Gradient Echo: Physics
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

3D Ultrafast Gradient Echo: Physics

Timing diagram
In the 3D ultrafast gradient-echo sequence, there is a preparatory phase at each repetition
time (TR), before the collection of several lines in 3D k space (Fig 1). Typically, the k-space
lines for an entire partition of k space may be collected in a single TR period (a partition is a
slice though 3D k space corresponding to all the slice-encoding steps for a given phaseencoding step).

Figure 1. MP-RAGE timing diagram. In the preparatory phase (shaded purple) an inversion
pulse is applied. Repeated gradient-echo acquisitions follow (shaded green). In each TR, the
slice (depth) encoding is varied to acquire all lines in the slice-encoding direction. The phaseencoding gradient is varied from TR to TR. For illustration, there are only four slices in the case
shown.

3D k-space filling in ultrafast imaging

The animation in Figure 2 illustrates 3D k-space filling for an ultrafast 3D imaging sequence.

Figure 2. The upper graph shows the effect of the inversion pulse on longitudinal
magnetization (Mz) over the acquisition period. The inversion time (TI) is the time from the
inversion pulse to the midpoint of the acquisition, where the mean or middle partition is filled.
Note how the inversion pulse introduces T1 weighting. After data acquisition, there is a
recovery period for the longitudinal magnetization prior to application of the next preparatory
pulse.

GO TO 3D Ultrafast Gradient Echo: Overview

GO TO 3D Ultrafast Gradient Echo: Physics
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

EPI: Overview
EPI: Echo Planar Imaging
Synonyms
EPI

Siemens Medical Solutions, GE

Medical Systems, Philips Medical
Systems, Toshiba, Hitachi

Background: Fast gradient-echo sequences (eg, TurboFLASH; see Ultrafast Gradient

Echo) generate a single echo with every RF excitation flip applied. In the EPI sequence,
multiple echoes and k-space lines are acquired with a single RF pulse. In single-shot EPI,
the entire k-space matrix is filled within a single TR.
Features: Extremely fast sequence, often used to generate 2D images in a single
subsecond shot. Has T2* weighting. Adaptations of the basic EPI sequence are used in
diffusion imaging, perfusion imaging, and functional MR imaging.
Artifacts/Problems: Artifact prone. Sensitive to susceptibility effects and field nonuniformities that show signal voids and image distortions. Demonstrates a ghosting artifact
in the phase-encoding direction. Resolution is limited.

GO TO EPI: Physics
GO TO EPI: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

EPI: Applications
EPI sequences and sequence variants form the basis of practical diffusion-weighted
imaging. (Figs 1, 2), perfusion imaging, and functional MR imaging.

Figure 1: Trace diffusion-weighted (b = 1000) axial

image shows an infarct in the region of the left
thalamic nucleus. The infarct appears as an area of
restricted diffusion with high signal intensity.

Figure 2: Apparent diffusion coefficient (ADC) image in

same subject as in Figure 1. The image is calculated
from the diffusion-weighted image and a second T2weighted image. The ADC image shows the infarction as
a dark area of restricted diffusion. Regions of elevated
signal intensity in the diffusion-weighted image may be
due to elevated T2 rather than restricted diffusion ("T2
shine through"). Calculation of the ADC image removes
this ambiguity.

GO TO EPI: Overview
GO TO EPI: Physics
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

EPI: Physics
Timing diagram
A timing diagram for an EPI sequence is shown in Figure 1. Note how only one RF flip is applied
(90), and all k-space lines can be acquired in a single shot. In the sequence shown, the alternating
frequency-encoding gradient sweeps the location in k-space from side to side. A weak constant
phase-encoding gradient simultaneously sweeps the k-space location from bottom to top. The weak
pixel-to-pixel frequency difference in the phase-encoding direction emphasizes chemical shift
artifact. To avoid this, EPI sequences are normally used with fat supression.

Figure 1: Timing diagram for EPI sequence.

Image distortion
The shallow phase encoding means there is a very narrow bandwidth across the tissue in the
phase-encoding direction. In other words, there is only a small pixel-to-pixel precession frequency
difference implementing phase encoding. In EPI, any effect that disturbs the local magnetic field by
even a small amount will have a proportionally large disruptive effect on these small frequency
differences. For example, at air/tissue interfaces, transistions in magnetic susceptibility will cause
geometric distortion of the image. Imperfect magnet shimming will also cause geometric distortion.
Ghosting
EPI is subject to a ghost arifact in which a ghost of the image appears shifted in the phaseencoding direction by a distance FOV (field of view)/2. This is called an N/2 or Nyquist ghost.
T2* decay
EPI generates gradient echoes under the FID curve created by the RF flip (Fig 2). As the curve
decays according to T2*, gradient EPI sequences willl be T2* weighted. Sensitivity to T2*, of
course, introduces sensitivity to artifacts caused by changes in magnetic susceptibility (eg,
air/tissue interfaces) and imperfect magnet shim.
In a single-shot EPI acquisition, all echoes for the image are acquired under a single T2* decay
curve. Echo spacings must be extremely short so that signal levels at the later echoes are not
completely degraded. Because all k-space lines are acquired at different stages of T2* decay, EPI
images are subject to a blurring effect in the phase direction.

Figure 2: Echoes in an EPI sequence form under the T2* decay curve.

Gradient hardware requirements

The extremely short echo spacing in EPI places demands on gradient system performance (Fig 3).
In order to sweep across k space during each echo acquisition period, each frequency-encoding
lobe must have a defined area determined by the size of the FOV. From an imaging time
perspective, the ideal is to achieve the required area under the gradient lobe in the shortest
possible time.

Figure 3: The two frequency-encoding gradient lobes shown have the same area A under the curve
and so can perform the same function in the EPI sequence timing diagram. However, the upper
gradient is applied for a shorter time than the lower gradient and so can achieve shorter echo
spacing. In practice, this leads to a requirement for high-performance gradient hardware in EPI.

Blipped EPI
There are many different schemes for driving the phase- and frequency-encoding gradients during
EPI. The timing diagram for "blipped" EPI is shown in Figure 4. Here, a brief application of the
phase-encoding gradient between echoes nudges the trajectory in k space to a new row.

Figure 4: Blipped EPI. The trajectory corresponds directly to the rectilinear spacing of k-space
locations. Unlike the "zigzag'' trajectory, corrective measures such as reformatting of k space are
not necessary.

Spiral EPI
The timing diagram for spiral EPI is shown in Figure 5. The phase- and frequency-encoding
gradient waveforms are out of phase sinusoids.

Figure 5: Spiral EPI timing diagram. Sinusoidally driven gradients place less demands on the
gradient hardware. The spiral acquisition requires special k-space reformatting and recontruction
methods.

GO TO EPI: Overview
GO TO EPI: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

SSFP-FID: Overview
SSFP-FID: Steady State Free Precession FID
Synonyms
FISP (fast imaging with steady-state
precession)
GRASS (gradient-recalled acquisition in the
steady state)

Siemens Medical Solutions

GE Medical Systems
Philips Medical Systems
Hitachi
Toshiba

GE Medical Systems
FFE (fast field echo)
SARGE (steady-state acquisition with
rewound gradient echo)

Background: SSFP sequences are a type of gradient-echo sequence in which the

transverse component is not spoiled. In an SSFP-FID sequence an echo of the FID signal
component is sampled.
Features: Used for its T1/T2* contrast at short TR values. A long TR produces similar
contrast to that of a spoiled gradient-echo sequence. Signal-to-noise ratio is better than that
of a spoiled gradient-echo (FLASH) sequence, since the SSFP-FID sequence "recycles"
the transverse component that is destroyed in a spoiled gradient-echo sequence. The short
TRs means SSFP-FID sequences are suitable for 3D and sequential 2D imaging, but not
for interleaved 2D imaging.
Artifacts/Problems: The SSFP-FID sequence is sensitive to dephasing caused by flow.
Because of this, it is of limited use in imaging the brain or cervical spine.

GO TO SSFP-FID: Physics
GO TO SSFP-FID: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

SSFP-FID: Physics
Timing diagram
The timing diagram in Figure 1 is for an SSFP-FID sequence. At first glance the diagram
does not appear to differ significantly from a spoiled gradient-echo (eg, FLASH)
sequence. However, there is a key difference: an absence of the spoiling mechansims (RF
or gradient) used in FLASH imaging.

Figure 1: SSFP-FID timing diagram. If TR is sufficiently long in an SSFP-FID sequence,

the transverse component is effectively spoiled by normal T2 decay, and the sequence
behavior approaches that of a spoiled gradient-echo (FLASH) sequence.

Steady-state free precession

The concept of a steady state in the longitudinal direction is familiar from spoiled gradientecho sequences, where the longitudinal component settles down to a steady value after the
application of a number of RF pulses. The steady state in spoiled gradient-echo imaging is
a subset of a generally more complex steady state behavior called steady-state free
precession. In unspoiled gradient-echo sequences, the transverse component is allowed to
contribute to the steady state, creating a complex mixing over and back between the
longitundinal and transverse components at every flip. To get a sense of this complexity,
use the animation in Figure 2 to examine transverse components after the first TR of an
unpoiled gradient-echo sequence.

Figure 2: The animation shows transverse components of spins immediately after an alpha
flip and the effect of the gradient on the spins (as used, say, in FISP or FLASH imaging).
Note how at the end of a TR period, a dephased set of spins is left. (Dephasing by T2, T2*,
etc, is ignored here: Only the effects of the gradients are considered.)

After many RF flips, the result is a complex interchange of the "histories" of dephasing,
rephasing, decay, and recovery between the transverse and longitundinal directions: The
remaining transverse magnetization feeds into the longitundinal magnetization at a
subsequent RF flip (Fig 3). "New" transverse magnetization from the longitudinal direction
mixes into the transverse plane at each flip. Any remaining transverse component can be
partially rephased by a subsequent RF flip. The conditions for the transverse component to
contribute to the steady state (ie, unspoiled gradient echo) are
1. A short TR
2. Constant dephasing during each TR
3. Same net application of gradients during each TR (to meet condition 2)
4. Constant flip angle

Figure 3: Partially dephased transverse components remaining at the end of a TR period

contribute to the longitudinal component at the next application of an RF pulse. This leads
to complex mixing of transverse and longitundinal components.

Signal pathways
A simplified view of the mixing of signal in an unspoiled SSFP sequence can be considered
by referring to Figure 4, which shows the response of tissue to a train of RF pulses during
an SSFP sequence. The action of the gradients has been ignored.

The signal received can be divided into two components. Not unexpectedly, there is an FID
component after each RF flip; but there is also a spin-echo component. This is because
each alpha flip has the potential to partially refocus magnetization from a prior FID. This is
analogous to the more complete refocusing of an FID seen with a 180 flip in spin-echo
imaging.

Figure 4: Response of tissue to a a train of equally spaced RF pulses. In an actual imaging

sequence, gradients are used to form gradient echoes. By appropriate use of the gradients,
a particular "pathway" can be favored: The gradient echo can be of the steady-state FID
component or of the steady-state echo component.
Contrast in SSFP-FID imaging
In the SSFP-FID sequence, gradients act to favor the echo from an FID component
generated by a train of RF pulses. The complex mixing of longitudinal and transverse
components in SSFP-FID imaging leads to T1/T2* contrast. With long TR values, the
transverse component is effectively spoiled and the contrast behavior becomes that of a
spoiled gradient-echo sequence (eg, FLASH). For small flips angles, the contrast becomes
proton-density weighted.

GO TO SSFP-FID: Overview
GO TO SSFP-FID: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on the red text to read more about that topic.

SSFP-FID: Applications
The SSFP-FID sequence is applied in 3D imaging during orthopedic
examinations (Figure).

FISP image of the ankle joint.

GO TO SSFP-FID: Overview
GO TO SSFP-FID: Physics
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on the red text to read more about that topic.

SSFP-Echo: Overview
SSFP-Echo: Steady-State Free-Precession-Echo
Synonyms
PSIF (mirrored FISP)
CE-FFE-T2 (contrast-enhanced T2weighted FFE)
CE-GRASS (contrast-enhanced
gradient-recalled-echo in the steady
state)

Siemens Medical Solutions

GE Medical Systems
Philips Medical Systems
Hitachi
Toshiba

TRSG (time-reverse SARGE)

Background: SSFP sequences are a type of gradient-echo sequence in which the

transverse component is not spoiled. In an SSFP-echo sequence an echo of the spin-echo
signal component is sampled. The timing diagram of an SSFP-echo sequence is the same
as that of a SSFP-FID sequence reversed in time.
Features: Used for its heavily weighted T2 contrast (ie, bright fluid). The strong fluid signal
makes SSFP-echo sequences useful in imaging the cerebrospinal fluid space.
Artifacts/Problems: Signal-to-noise ratio can be poor. Being supplanted by newer
balanced SSFP techniques.

GO TO SSFP-Echo: Physics
GO TO SSFP-Echo: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

SSFP-Echo: Physics
Timing diagram
In an SSFP-echo type sequence (eg, PSIF), the gradients act to rephase the echo part the
signal rather than the FID part as in SSFP-FID (eg, FISP). The timing diagram for an
SSFP-echo sequence is shown in the Figure.

SSFP-echo timing diagram. The echo produced is a spin echo and shows T2 rather than
T2* weighting. The echo is in effect a rephasing of transverse components generated by
the RF flip in the previous TR. This gives, effectively, a long TE (approximately twice the
TR) and heavy T2 weighting.

GO TO SSFP-Echo: Overview
GO TO SSFP-Echo: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

SSFP-Echo: Applications
The SSFP-echo sequence provides bright fluid, suitable for imaging
cerebrospinal fluid (Figure). Balanced SSFP sequences are now providing
similar bright fluid, but at an improved signal-to-noise ratio and with less
artifact sensitivity than SSFP-echo imaging.

Heavily T2-weighted PSIF sequence demonstrates the anatomy of the inner

ear.

GO TO SSFP-Echo: Overview
GO TO SSFP-Echo: Physics
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

DESS: Overview
DESS: Dual-Echo Steady State
Synonyms
Spin echo (SE)

Siemens Medical Solutions

Background: The DESS sequence generates both an SSFP-FID (eg, FISP) signal and an
SSFP-echo (eg, PSIF [mirrored FISP]) signal in a single TR period. The two generated
echoes are then combined to form a single image.
Artifacts/Problems: In the DESS sequence, the SSFP-echo component provides
additional T2 weighting to long-T2 tissues, showing bright fluid. The SSFP-FID component
shows other tissues with weighting similar to that obtained with a FISP sequence. The
sequence is used for high-resolution bright-fluid 3D examinations, particularly in
orthopedics.

GO TO DESS: Physics
GO TO DESS: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on the link to read more about that topic.

DESS: Applications
The DESS sequence is used in orthopedic imaging, yielding high-resolution
images with bright fluid and good contrast between synovial fluid and cartilage
(Figure).

T1-weighted conventional spin-echo image of the cervical spine demonstrates

normal marrow.

GO TO DESS: Overview
GO TO DESS: Physics
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

DESS: Physics
Timing diagram
In the DESS sequence, the gradients act to separate out the FISP element and the
PSIF echoes (see Fig 4, SSFP-FID Physics). The images can be combined to
form a single image. A DESS timing diagram is shown in the Figure.

GO TO DESS: Overview
GO TO Dess: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

Balanced SSFP: Overview

Balanced SSFP : Balanced Steady State Free Precession
Synonyms
TrueFISP (fast imaging with steady-state precession)

Siemens Medical Solutions

GE Medical Systems

FIESTA (fast imaging employing steady-state aquisition)

Balanced FFE (fast field echo)

Philips Medical Systems

Hitachi

BASG (balanced SARGE)

True-SSFP (steady-state free precession)

Toshiba

Background: Balanced SSFP sequences are sequences in which the net phase change
caused by all three gradients (frequency, phase, and slice) is zero.
Features: Delivers contrast related to the ratio T2/T1. Tissues with a T2 value approaching
their T1 value appear brightest. Gives good blood/muscle contrast. Very fast sequence with
excellent signal-to-noise ratio.
Artifacts/Problems: The T2/T1 weighting has limited application. Prone to a banding
artifact caused by imperfect shimming or suceptibility effects; problem is minimized with
use of a short TR.
Oddities: While a balanced SSFP sequences settles into its steady state ("transient"
phase), its contrast will be a mix of proton-density and T2/T1 weighting. So a 2D image
may show different contrast than a 3D image (since a higher proportion of data acquisition
in the 2D case takes place during the transient phase). Balanced SSFP does not show high
sensitivity to T2* effects, as might a gradient-echo sequence.

GO TO SSFP: Physics
GO TO SSFP: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

Balanced SSFP: Physics

Balanced gradients
Each of the three gradients (frequency, phase, and slice) in this type of sequence is
"balanced." In the timing diagram (Fig 1), note how the total negative area on each gradient
is counterbalanced by the same total positive area during each TR. Implementing the
theory of a completely balanced sequence required a wait for the development of highperformance gradients: All that switching has to happen within a short TR (otherwise the
steady state would be lost). Hence, the name TrueFISP used by Siemens Medical Systems
to distinguish it from the earlier, only partially balanced FISP sequence.

Figure 1: Balanced SSFP timing diagram.

The effects of balanced gradients can be studied by using the animation in Figure 2. Note
how net dephasing is zero at the end of TR. In a balanced SSFP sequence, all three
gradients are balanced in this way.

Figure 2: Effect of a balanced gradient. The colored vectors represent transverse

components of spin at different spatial positions in a gradient.

Steady state in a balanced sequence

Given that the dephasing caused by the gradients is zero in a balanced sequence, the
transverse component at each TR is fully rephased, with phase zero. The resulting steady
state is relatively uncomplicated (Fig 3). In practice, "alternating sign" RF pulses are used
in a balanced SSFP sequence. This means that at every second TR, the applied flip is
+alpha or alpha. Alernating the sign gives a stronger signal than not doing so.

This figure is not available

Figure 3: The animation shows the resultant vector in a steady-state balanced sequence
from TR to TR. The flip angle is alpha, with an alternating sign. The vector "settles down" to
a position alpha/2 around the longitudinal axis.

Banding artifact
The balanced steady state shown in Figure 4 is for conditions of perfect magnet uniformity.
In reality, small local differences in precession frequency cause small deviations from the
ideal case of zero net phase at the end of each TR. When the phase angle approaches
180, there is a rapid loss of the steady state and a decrease in signal. Areas of relatively
poor magnet uniformity or changes in susceptibility introduce sufficient dephasing to cause
a dark "banding artifact" on the image. The requirement to keep dephasing in any voxel to
less than 180 to avoid banding is stringent, and balanced SSFP requires good magnet
uniformity. If TR is kept short, less time is given for phase differences to develop and the
artifact will be reduced.

Figure 4: In practice, the ideal steady state (Fig 3) is disturbed by small local variations in
precession frequency. Where the phase change introuced by these varaiations is severe, a
rapid decrease in signal occurs and results in the appearance of a banding artifact.

GO TO SSFP: Overview
GO TO SSFP: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

Balanced SSFP: Applications

Balanced SSFP sequences provide high-speed T2*/T1 contrast images. Fluid appears
bright (Fig 1). Without contrast material administration, soft-tissue contrast is poor.
Although this sequence is frequently used as a localizing sequence in the abdomen, the
high speed allows rapid assessment of acute abdominal conditions in patients who are
unstable or unable to stay in the magnet a long time (Fig 2).

Figure 1: Balanced SSFP image in a coronal plane through the liver. High signal intensity
is noted in the gall bladder and portal vein and within bowel loops. The sequence can be
used to demonstrate venous anatomy without contrast agent.

Figure 2: Balanced SSFP image in a coronal plane through the abdomen demonstrates
fetus

Good blood/myocardium contrast and high speed make balanced SSFP sequences
suitable for cardiac imaging applications (Fig 3).

Figure 3: Cardiac imaging with balanced SSFP sequence.

GO TO SSFP: Overview
GO TO SSFP: Physics
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

CISS: Overview
CISS: Constructive Interference in the Steady State
Synonyms
CISS
FIESTA-C (fast imaging with steady-state
acquisition and phase cycling)

Siemens Medical Solutions

GE Medical Systems

Background: Balanced SSFP sequences are subject to a "banding artifact." In the CISS
sequence, two balanced SSFP images in each acquisition are obtained. One image is
acquired with an "alternating phase" RF pulse as in a standard balanced SSFP sequence,
and the other is acquired with a nonalternating phase RF pulse. The banding artifact in
each image will be displaced relative to the banding artifact in the other image. By
combining the two images, the banding artifact is eliminated.

Features: The CISS sequence provides T2-weighted 3D images with submillimeter

resolution and is used for imaging the cochlea, labyrinth, and cerebellum. Slower than
balanced SSFP, it uses a longer TR.

GO TO CISS: Physics
GO TO CISS: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

CISS: Physics
CISS and alternating phase
The CISS sequence is simply a combination of two balanced SSFP acquisitions.
The images are combined to remove banding artifact. The acquisitions differ in that
one is "phase alternating," and the other is not. This refers to the sign of the applied
RF pulse in each sequence (Fig 1): In one acquisition, the angle alternates at each
TR (eg, 70, -70, 70 etc), and in the other it does not. Phase alternation (or
cycling) is the "normal" case for a stand-alone balanced SSFP acquisition.

Figure 1: RF flip angle. In a phase-alternating acquisition, the sign of the flip angle
alternates.

Off-resonance effect and banding artifact in balanced SSFP (true FISP)

The individual balanced SSFP images acquired in the CISS sequence show
banding artifact due to off-resonance effects. An off-resonance voxel is a voxel in
which the precession frequency differs slightly from the ideal reference frequency
(ie, approx 63.8 MHz in a 1.5-T imager). Off-resonance effects occur when there is
an imperfect magnet shim or at a boundary between two materials with different

magnetic susceptibilities. The animation in Figure 2 simulates this effect for

transverse vectors near such a tissue boundary.
In an ideal balanced SSFP sequence, the accumulated phase in the transverse
plane is zero at the end of a TR period. Deviations from zero phase cause signal
loss. In a phase-alternated balanced SSFP sequence, this loss is most severe for a
phase of 180.

Figure 2: Balanced SSFP banding artifact. Small local variations in precession

frequency result in phase error accumulating over a TR period and signal loss.

CISS sequence and banding artifact removal

The second image acquired in the CISS sequence is also a balanced SSFP image,
but without RF phase alternation. It too shows banding artifact, but the bands are
displaced relative to the bands in the phase-alternated sequence (Fig 3). While
signal level for an RF phase-alternated sequence is lowest when the accumulated
phase over a TR is about 180 and highest for an accumulated phase of zero, the
reverse is true for the nonphase alternating sequence. When the two images
acquired in CISS are recombined, a single image without banding artifact results.

Figure 3: Banding artifact removal with the CISS sequence.

GO TO CISS: Overview
GO TO CISS: Applications
BACK TO Sequence Map

An Interactive Taxonomy of MR Imaging Sequences

This is an interactive PDF. Click on a link to read more about that topic.

CISS: Applications
CISS provides heavily T2-weighted images at high resolution. It is most
commonly used to demonstrate the seventh and eighth cranial nerves and the
cererbropontine angle (Figure).

CISS acquisition demonstrates the seventh and eighth cranial nerves in a

maximum-intensity projection (MIP) image.

GO TO CISS: Overview
GO TO CISS: Physics
BACK TO Sequence Map