MRI slice selection, encodings, sequences Flashcards
slice selection is along the Z AXIS, it has a width to it therefore the pixals hold a volume known as voxals
what localisations are found on the xyz axis
slice selection - z axis
frequency encoding - x axis
phase encoding - y axis
how does slice selection work
- use gradient coils to apply gradient of magnetic filed along b0
- apply a radio frequency bandwidth according to desired section along z axis
- this will target protons within a certain region/slice to become excited and flip onto transverse plane
what are the 3 ways you can adjust the area of slice selection
- adjust the RF bandwidth
- adjust the baseline of the gradient
- physically move the patient
how does slice thickness, signal and resolution corrolate
increase thickness = increase signal = decreased resolution
what are 2 ways to change the slice thickness
- adjust the rf bandwidth range
- adjust the magnetic field gradient steepness by adjusting the strength of the gradient coils
define the issue known as ‘slice phase’
the protons within the slice selected experience different fields from one end of the slice compared to the other
- despite all being excited by the RF pulse, they may precess slightly out of phase as they experience different magnetic fields on either ends of the slice
how do you fix slice phase
- apply REPHASING GRADIENT
- applies equal and opposite gradient in opp direction along z axis of slice selected
- allows all protons in that slice to now spin in phase
note than ONLY DURING RF PULSE does slice selection and excitation occur, as soon as RF pulse is stopped, slice selection and everything else including gradients pause
explain how frequency encoding works for a selected slice
- each half of one gradient coil has currents going in opposite directions generating a magnetic field in the X AXIS
- this forms a gradient in the x axis thus protons in each column of the selected slice in the x axis has different precessional speeds
- the FE is applied ONLY WHEN READ OUT OCCURS/ONLY WHEN TE IS TAKEN
- multiple samples are taken at the same TE at the same time across the entire x axis of the slice (this can be infinite)
- the samples taken shows the NET magnetisation from the whole slice along the x axis (displayed as grey scale expressing a numerical value from the wave formed) and as it forms a wave of increasing and decreasing amplitudes, you can identify what frequencies are contributing to the signal along the x axis
- only after using the inverter Fourier transformation does it convert into frequencies (based on the comparisons of each sample (the net amplitudes) can they pick out the frequency combination to form that certain net magnetisation and identify the separate frequencies) (converts time based data set into frequency/location based data set)
(the number of times you sample determine the amount of frequencies you can identify/delineate in the same (as it can be infinite))
- TR is then applied (entire sequence with the fe etc) is repeated until the x axis of the whole slice is done eventually forming an entire signal coming from the column along the x-axis of the selected slice ( the same pulse sequence repeated forms the same outcome) (refer to image in camera roll if confused)
when applying a FE, this would affect the phasing during TE. What must be done prior to TE?
- equal and opposite frequency encoding gradient is applied before the actual frequency encoding gradient prior to readout
- by the time read out is happening, the spins will be much more in phase
explain how phase encoding in the y axis works and how it correlates with frequency encoding
- gradient is applied in y axis between 90 and 180 rf pulse, now causing different frequency of precessions of protons in the y axis due to the varying magnetic field (remember that protons at the null region of the gradient spin at the lamour frequency (b0))
- the gradient is the removed, and the protons all return to the same lamour frequency but are all out of phase (as they were previously spinning a different speeds)
- then during readout the frequency encoding gradient is applied, and each column along the x axis spins with the same frequency (due to magnetic gradient in the x-axis)
- overall, there is dephasing of the spins based on their Y AXIS but same frequency based on x axis
the strength of the gradient applied in the y-axis during phase encoding differs each time a sample is taken in order to adjust the amplitude of the wave formed, allowing different data sets in the y axis to be gathered according to the increase/decrease of dephasing caused..
for each column and row of the slice selected, the sequence is repeated hence frequency and phase encoding is also repeated (but phase encoding repeated each time at different ‘magnitudes’)
each time cycle is introduces, small amount of phase in the positive and negative direction is introduced, this is repeated till the resolution required the y axis has been reached.
the grey scale formed from this represents NUMERICAL VALUES overall generating a K-space that can then be applied in Fourier transform to form image
what does each line in k space represent
net magnetisation vector change over a given period
centre of k-space shows the phasing whilst periphery shows dephasing
during spin echo sequence, why is the equal opposite gradient required to counteract the frequency encoding gradient in the x axis, applied prior to the 180 degree rf pulse?
because the 180 degree rf pulse flips the proton spins, so to have it in the correct position, you should place it before the 180
- not it doesnt matter if you place the phase encoding before or after the 180
explain what a multiECHO spin echo sequence is
- same 90 RF, 180 RF, PEG, FEG AND READOUT/TE (forms first kspace/image with repeated sequencing)
- then additionally repeat 180 RF and another READOUT (after first readout) then repeat sequence multiple times again but with changing PEG to fill out rest of k space
- you now end up with 2 k-spaces but with different images as they were completed at different TE (so intensity/amplitude of 2nd image is lower than initial image)
- you end up with 2 k spaces that simple represent different weightings for the SAME slice
explain multiSLICE spin echo imaging
- initial 90 RF, 180 RF, PEG, FEG AND READOUT/TE
- simultaneously, the same sequence but for a different slice is going on but both have the same TE AND TR
- you form 2 k-spaces (or how many depending on the number of slices you do simultaneously) with the SAME WEIGHTING but for different regions (due to different slices)
- ensure not to select slices near each other as the RF bandwidth can overlap
- slices are then all combined at the end
number of slices done within one TR corresponds to total reduction of scan time
explain fast spin echo imaging
- same initial spin echo sequence
- prior to another 180 RF, an equal and opposite PEG is applied to negate the prior PEG and bring back the protons in the y-axis in phase
- then do the 180 RF, a different PEG and your readout with FEG and repeat the negating process
- each time you negate and repeat the 180 rf and read out, you fill another line of k-space without having to repeat the entire sequence thus saving a lot of time but each time it is repeated, the TE reduced in signal
- overall, multiple spin echos within one TR
what is ETL and how does this relate to total scan time
Echo train length, the number of pulse sequences between first 90 rf pulse and the TR
- total scan time can be reduced drastically by increasing ETL
- Divides the total scan time by the value of ETL
why would fast spin echo be good for viewing fluid
- water has a long t2 decay
- as multiple 180 RF pulses are applied and readout in one TR, the signal gradually decreases but because water decays slow, there will still be a significant signal still left from
what aspect k-space can be utilised in order to make fast spin echo imaging even faster
- k-space has conjugate symmetry
- therefore if you fill only half of k-space with fast spin echo, you can assume the other half is exactly the same and further reduce scan time by half
explain what gradient echo imaging is
- RF of degree 90 or smaller is applied
- NO additional 180 RF
- equal and opposite FEG is applied in x-axis then intentional/normal FEG is applied during read out (doing this negates the t2* decay (inhomogeneity) and produces a signal of the expected t2 value) (both Frequency gradients must be of equal strength and length of time in order for protons to me completely in-phase during time of readout)
- with no additional 180RF pulse, TE can be applied close to the initial RF forming short TE which is great for t1 weighted images or proton density
main difference b/w spin echo and echo sequence
- spin echo applies 180 RF to negate the inhomogeneity from t2 decay
- gradient echo simple uses additional equal and opposite FEG to negate inhomogeneity from t2 decay
because you cannot apply multiple TE’s during gradient echo, you can decrease TR.
SHORT TR in gradient echo allows rapid imaging which decreases motion artefact and scan time
why is t2 much faster than t1
- t2 decay is due to the spins going out of phase, however the protons are still precessing in the transverse plane
- for t1 recovery, we still need to wait for those protons to return to longitudinal magnetisation from the transverse plane even if the protons there have completely dephased
compare how large and smaller flip angles along with short TR in gradient echo sequences affect the net transverse magnetisation recieved
- larger RF = longer for longitudinal magnetisation recovery, with a short TR, the RF will only excite the small value of regained longitudinal magnetisation back into the transverse plane
- smaller RF = quicker longitudinal magnetisation recovery, with short TR, protons will have already regained max longitudinal magnetisation so the TR will again excite the small RF angle into the transverse plane
- despite smaller RF producing little signal, with the efficiency of it with a short TR compared to larger RF, it may end up producing even larger signals
- gradient echo requires short TR otherwise might as well do spin echo
- gradient echo is fast because you do multiple small RF
- spin echo is fast because you can do multiple sequences in one TR
- know that in gradient echos, each tissue will have a certain flip angle that produces greatest signal compared to others at a certain short TR
What is a steady state flip angle
same repeated signal recieved from the flipping of certain angle that hasn’t been allowed to reach full longitudinal recovery
how can gradient echos still form T2 (fluid bright) images despite having short TR?
- larger flip angle and applying an addition RF/ apply TR when tissues haven’t reached maximum longitudinal recovery
- this way the signals recieved from transverse plane are due to the T1 differences showing fat as bright
- but if you use small flip angle and apply short TR, the longitudinal magnetisation fo the tissues will have fully recovered so the excitation into the transverse plane shows no T1 differences between the tissues (thus t2 weighting (bright water) can be seen from the slower decay of water)
explain the difference in collection to k space data in fast spin echo and short TR gradient echo
- each TE readout during fast spin echo is a line of k-space and they all vary in contrast as TE gradually decreases as it continues due to it happening within one TR
- each TE readout during short TR gradient echo is a line of k-sapce but they all have a consistent contrast as they all collate the same amplitude of signal as each readout is part of its own sequence/own TR
explain process of echo planar imaging (in gradient echo imaging)
- follows normal gradient echo imaging
- however there is repeated de-phasing and rephrasing frequency encoding gradients applied during TE to gather multiple gradient echos
- right before each de/rephasing frequency gradient, there is an blip of PEG to ordain the readout to a specific line of k-space (starting from the centre with minimal PEG then increasing as FEG/readout are repeated/continued)
- first k-sapce is filled right to left, then left to right etc (to allow rapid filling of k-sapce)
- overall, this allows the filling of large portion of k-space within one free induction decay
- then echo planar process repeats till k-space is completely filled
what is multi-shot EPI (spin echo, echo planar imaging)
- normal spin echo BUT prior to 180 RF there is a large PEG applied to dephase protons and usual FEG applied
- then follow the same format as regular EPI, having multiple PEG blips right before the de/rephasing frequency encoding gradients. Gradually this negates the large dephasing caused by the PEG prior to the 180 RF and nulls the spin to generate large readout before continue to dephase it with more blips and FEG’S.
- this forms K-space (not starting from centre due to initial large PEG applied), going from right to left, left to right etc quickly
compare conventional spin echo and fast(turbo) spin echo
conventional :
- 1 line k-space each TR (256 TRS needed)
fast/turbo:
- multiple lines of k-space per TR
- decreases total acquisition time
- each line has different PEG
define echo train length, effective echo time, echo spacing
echo train length = no of echo / no lines per TR
Effective echo time = TE of echo acquired closest to centre of k-space (e.g in a sequence where there is mutiple echos, the effective echo time is decided by which echo is acquired nearest to centre of k-space (where contrast is located)) (the order of phase encoding is chosen to have effective TE)
echo spacing = time between successive echoes
scan time = TR X TOTAL NO LINES / ETL
what TR and TE is needed for t1/2 weighting
T1 = short TR (ETL 3) (fat bright)
T2 = long TR (2000ms+) (fluid bright)
in general, LONG TE&TR = T2
SHORT TR&TE= T1
How long does a regular spin echo sequence take compared to FSE/turbo
regular = 8 - 12 mins (t2 weighted dual echo)
fast = reduced by 2-3 mins
whats used to convert k-space into image / used to reconstruct image
array processor
remember each line of signal = a line of k-space
low amps of PEG respond to central lines of k-space whilst high amps of PEG respond to outer lines of kspace
note: multi-spiconventional SE
- 1 line k-space each TR
TURBO/FSE
- multiple lines k-space each TR
- reduce total acquisition time by n
- each line has different PEGn echo is the same as multi echo spin echo
Why is it important to apply a negative phase encoding before each 180 RF in a fast spin echo/turbo?
to ensure you return to the centre of k-space befor emaking the next k-space trajectory
does fat appear brighter on FSE or regular SE?
FSE (dont need to know why)
why is there reduces signal from muscle and brain in FSE/turbo?
magnetisation transfer effect
- transfer of nuclear spin polarization and/or spin coherence from one population of nuclei to another population of nuclei,
- the physical process by which macromolecules and their closely associated water molecules (the “bound” pool) cross-relax with protons in the free water pool.
how can you reduce diffusion artefact in FSE
have closely spaced 180 pulses
what can be done to force T2 into longitudinal magnetisation?
use ‘driven equilibrium’ (also known as fast recovery) to tip Mxy into Mz, helps faster recovery esp for tissue with longer recovery like csf
- named as DRIVE / RESOTRE
- as this mixes magnetisation from both planes, contrast can be different from t1/2
how does DRIVE/RESTORE mechanism work
- (-)90 restoring RF pulse applied at end of echo train
- transfers remaining Mxy back along z axis
- allows faster t1 recovery of long t1/2 tissues e.g water
- allows t2 weighted imaging with shorter TR
why is blood black in spin echo sequences aka black blood spin echo
- initial 90 pulse excited both protons in blood and stationary tissue within slice
- excited spins in blood leave slice between RF pulses
- 180 RF pulse, excited spins from prior 90 pulse only in stationary tissue so high signal from tissue but low signal from blood (as fresh blood is affected by 180)
3 causes of magnetic field inhomogeneity
- intrinsic inhomogeneities of b0 (magnet design, shimming)
- magnetic susceptibility (material introduced into field)
- linear magnetic field gradients (deliberate introduction for imaging purposes)
spin echo uses 180 flip to achieve TE according to T2 decay, how does GE do this?
- applying gradients the dephase and then rephrase Mxy
- because applying a +ve and -ve FEG will only have half an echo readout, you must ensure to have the rephasing gradient for double the time (but the same strength as the dephase) to create a whole echo signal which can form a full line of k-space
what issue comes with have short TR in gradient echo
- With short TR’s around 50ms e.g, this may mean that the TE from before is yet to completely return to max Mz
- therefore when you apply another RF pulse to get another gradient echo, it will further excite residual transverse magnetisation as well as the protons that have returned to Mz
- this creates a stimulated echo (transverse magnetisation left over from the previous echo readouts due to short TR not allowing complete t2 decay) simultaneous to the current echo read out
how can a stimulated echo / coherent gradient echo provide T2 weighted imaging as well as a T1 weighted image?
- residual/stimulated echo is due to tissue that have long t2 as they were unable to completely recovery before the next TR applied
- therefore, if we know all the protons within the stimulated echo are from long T2 tissues, this can give us t2 weighted imaging
- simultaneously, the current echo being read out is following the T2* decay/FID and in real time, this happens extremely early on in FID due to the very short TE so there hasn’t been enough time for T2 differences to occur but we know that fat decays faster (short TE = T1 weighted imaging)
how would you get rid of the stimulated echo in gradient echo sequences in order to get ONLY a t1 weighted image
- apply a spoiler gradient (large enough) after the prior readout/TE in order to completely dephase the transverse magnetisation from before
- this removes that residual echo leaving you with the current echo readout following FID only (which is a t1 reading (due to only factors of having a short TE))
how does adjusting flip angles affect contrast?
the smaller the flip angle, the greater the increase in t2* weighting (as it decays faster)
(short flip angle = increase t2*)
what is SSFP gradient echo
steady state free precession gradient echo
- rewinder gradients used to accelerate the rephasing of the protons from the prior sequences/readout so that the stimulated echo forms before the current readout
- this process separates the stimulated echo from the current readout echo (undergoing FID) so that they can be read seperatly
note that by being able to sample the stimulated/residual echo in gradient echo samples as well as the normal FID means that we can essentially make up for inhomogeneity/t2* decay. Whereas in SE they have the 180 flip to make up for the t2* decay.
state the types of weightings found in coherent gradient echo (stimulated echo + FID), inhoherent/ spoiled gradient echo and SSFP
COHERENT = T1 AND T2 (due to both stimulated and current readout echo together)
INCOHERENT = T1 (due to removed stimulated echo using spoiler gradients)
SSFP = T2 (due to separated stimulated and current readout, allowing full sampling of stimulated echo and essentially making up to decay of T2 (not T2*))
what does flowing blood look like on a gradient echo sequence and why
- GE = bright blood
- RF PULSE, spins in stationary tissue are partially saturated due to repeated excitation
- Blood is continuously replaces so it doesnt become saturated
- steady-state longitudinal magnetisation is greater in blood = higher signal
what is a Hahn echo
When pulses other than 90°-180° are used, the resultant spin echo is sometimes differentiated by calling it a Hahn echo e.g 90 - 90 RF pulse
overlapping signals found in gradient echos can cause what type of issue/artefact and why?
interference patterns, due to presence of magnetic field inhomogeneities
- different signals can go out of phase and cancel
what 2 things make interference patterns better or worse
better = shimming, short TR’S (the longer the TR, the greater the effect of inhomogeneity)
worse = poor homogeneity, long TR
what 2 ways are gradients used in SSFP and how does this effect the interference pattern, signal and flow
- gradients used to refocus ALL SIGNALS for signal sampling (using balanced gradients), interference pattern formed and flow compensated BUT stronger signal (balanced SSFP)
- gradients refocus FID ONLY for signal sampling, removes interference pattern and flow sensitive BUT weak signal (traditional gradient echo)
note a balanced FFE etc is all SSFP, they are just separate names for the same sequence according to the brand
rank balanced GE, unbalanced GE and spoiled GE in terms of signal intensity for fluid with flip angle 90 or less
- balanced GE (SSFP) (T2 weighted image largest signal)
- unbalanced GE (SSFP) ( also T2 weighted so fluid bright but no balancing gradient as found in no 1 so signal isnt as large)
- Spoiled GE (t1 weighted image fluid dark)
