MRI Sequence Physics Explainer
The parameters you turn, the sequence families they build, and how to read a sequence name across scanner vendors.
- Every image is a trade between SNR, spatial resolution, scan time and contrast, so changing one parameter almost always costs one of the others.
- In conventional spin echo, TR drives T1 and TE drives T2: both short gives T1 weighting, both long gives T2, and long TR with short TE gives proton density.
- In inversion recovery, choosing TI at a tissue’s null point removes its signal: a short TI nulls fat (STIR) and a long TI nulls CSF (FLAIR).
- Spin echo forms its echo with a 180° refocusing pulse and gradient echo with a gradient reversal; EPI is a fast blipped readout that can fill k-space after a single excitation.
- Vendors use different trade names for the same physics, such as TSE on Siemens and Philips and FSE on GE; names drift between releases, so check the parameters when unsure.
Explain it simply
- Every scan balances four things: how clean the picture looks, how fine the detail is, how long it takes and how well tissues stand out. Improving one almost always costs another.
- Two timing settings, TR and TE, set a spin echo scan’s contrast: both short gives a T1 image, both long a T2 image, and long TR with short TE a proton density image.
- Some scans add a pulse, then wait before imaging. The right wait blanks out one tissue: a short wait darkens fat (STIR), a long wait darkens the fluid around the brain and spine (FLAIR).
- Spin echo scans bring the signal back with a second radio pulse, gradient echo scans by reversing a magnetic gradient. EPI is a fast reading method that can capture a whole image in one sweep.
- Scanner makers give the same type of scan different names, and the names change between software versions. If unsure what a scan is, look at its settings rather than its name.
A guide only: follow your departmental protocols and regional requirements.
1 · Core parameters
Every image is a compromise between four things: SNR, spatial resolution, scan time, and contrast. Moving one knob almost always spends one of the others.
TR (repetition time)
- Controls
- Time between successive excitations of a slice; governs how much longitudinal (T1) recovery occurs before the next pulse. The dominant lever for T1 weighting.
▲ Long TR → more T1 recovery, more slices, higher SNR, less T1 contrast • ▼ Short TR → strong T1 contrast, lower SNR.
TE (echo time)
- Controls
- Time from excitation to signal sampling; governs how much transverse (T2/T2*) decay is allowed. The dominant lever for T2 weighting.
▲ Long TE → more T2 contrast, lower SNR • ▼ Short TE → higher SNR, minimises T2 effect.
TI (inversion time)
- Controls
- In inversion-recovery, the delay after a 180° inversion pulse before excitation. Choosing TI at a tissue’s null point removes its signal: fat (STIR) or CSF (FLAIR).
Short TI (~150–170 ms @1.5T) nulls fat = STIR • Long TI (~2000–2500 ms) nulls CSF = FLAIR.
Flip angle (α)
- Controls
- How far the RF pulse tips magnetisation. In spin echo usually 90°; in gradient echo the small flip angle is the key contrast/speed lever.
▲ Large α → more T1 weighting; signal peaks at the Ernst angle, then falls • ▼ Small α → less saturation, less T1 weighting. The Ernst angle maximises SNR for a given TR/T1.
Receiver bandwidth (BW)
- Controls
- Range of frequencies sampled per pixel during readout. Sets readout speed and noise admitted.
▼ Low BW → higher SNR, worse chemical-shift/distortion • ▲ High BW → less chemical shift/susceptibility, shorter min TE, lower SNR (SNR ∝ 1/√BW).
ETL / turbo factor
- Controls
- Number of echoes (phase-encode lines) per TR in FSE/TSE and EPI. The main speed multiplier of fast spin echo.
▲ Long ETL → faster scans, more slices • Cost → T2 blurring, higher SAR, bright fat on FSE.
TR / TE map → image weighting
For conventional spin echo, the TR/TE combination alone predicts the dominant contrast. (Rough 1.5 T guide values.)
| Short TE (~10–20 ms) | Long TE (~80–140 ms) | |
|---|---|---|
| Short TR (~400–700 ms) | T1-weighted Anatomy; fat bright, fluid dark; post-Gd. | Mixed / rarely used. |
| Long TR (~2000–4000 ms) | Proton density Cartilage, marrow, menisci. | T2-weighted Fluid/pathology bright; the workhorse. |
Rule of thumb: TR drives T1, TE drives T2. Want T1? Both short. Want T2? Both long. Want PD? Long TR, short TE.
2 · Sequence family tree
Three branches. Two are defined by how the echo is formed: a refocusing 180° pulse (Spin Echo) or a gradient reversal (Gradient Echo). The third, EPI, is a fast blipped readout of gradient echoes that can fill k-space after one excitation, often after a spin-echo preparation.
Spin Echo branch (180° refocusing pulse)
Conventional SE
- What
- 90° excitation then a 180° refocusing pulse forms the echo, undoing static T2* dephasing.
- Strength
- Robust to static B0 inhomogeneity (the 180° pulse refocuses dephasing) → far less susceptibility signal loss than GRE, though distortion near metal still occurs; clean contrast.
- Weakness
- Slow: one line per TR.
FSE / TSE
- What
- Multiple 180° pulses per TR create an echo train (ETL), filling many k-space lines at once.
- Strength
- Fast; retains SE robustness to susceptibility.
- Weakness
- T2 blurring with long ETL; bright fat on T2; higher SAR.
STIR
- What
- IR-prepared FSE with TI set to null fat.
- Strength
- B0-insensitive fat suppression → works off-isocentre and near metal.
- Weakness
- Low SNR; don’t combine with gadolinium (may null enhancing tissue).
FLAIR
- What
- Long-TI IR-FSE nulling CSF while keeping T2 weighting.
- Strength
- Unmasks periventricular pathology hidden by bright CSF (MS, cortical lesions, SAH).
- Weakness
- CSF-flow & incomplete-nulling artifacts; long scan.
SSFSE / HASTE
- What
- Half-Fourier single-shot FSE: one long echo train captures a whole slice.
- Strength
- Very fast; motion-insensitive. MRCP, fetal, uncooperative patients.
- Weakness
- T2 blurring; lower SNR/resolution.
Gradient Echo branch (gradient reversal, no 180°)
Spoiled GRE (FLASH / SPGR / T1-FFE)
- What
- Small flip angle + gradient echo; residual transverse magnetisation spoiled each TR.
- Strength
- Very fast T1: breath-hold abdomen, dynamic post-contrast, 3D T1, in/opposed-phase.
- Weakness
- Susceptibility/T2* sensitive; lower SNR per shot.
Balanced SSFP (TrueFISP / FIESTA / bFFE)
- What
- Fully-balanced gradients maintain steady state; contrast ≈ T2/T1 → fluid very bright.
- Strength
- High SNR & speed; bright fluid/blood without gadolinium (cardiac cine, fetal).
- Weakness
- Dark banding from off-resonance/poor shim, worse at 3 T.
T2*-weighted GRE / SWI
- What
- Long-TE gradient echo sensitive to T2* dephasing; SWI adds phase.
- Strength
- Exquisitely sensitive to microhaemorrhage, iron, calcium, cavernomas.
- Weakness
- Distorts near air/metal; not for clean anatomy.
EPI branch (echo-planar readout)
Diffusion (DWI / DTI)
- What
- Diffusion-sensitising gradients on an EPI readout; DTI adds directions for tractography.
- Strength
- Unique contrast; very fast. Stroke, abscess vs tumour, cellularity.
- Weakness
- Distortion & N/2 ghosts near air/bone; T2 shine-through; always confirm on ADC.
fMRI (BOLD)
- What
- T2*-weighted single-shot EPI repeated rapidly to detect blood-oxygen-level-dependent change.
- Strength
- Whole-brain every ~1–3 s; pre-surgical eloquent-cortex mapping.
- Weakness
- Low CNR (needs statistics); dropout at sinuses; motion-sensitive.
Why EPI at all?
- Trade
- Single-shot EPI reads all of k-space after one excitation, which makes diffusion and BOLD practical.
- The cost
- Long readout → geometric distortion + Nyquist (N/2) ghost. Mitigate with good shim, higher BW, reduced-FOV / readout-segmented EPI, field-map/reversed-PE correction.
3 · Read a sequence name across vendors
Same physics, different trade names. Match the generic concept to the console label.
| Generic concept | Siemens | GE | Philips | Canon |
|---|---|---|---|---|
| Fast/turbo SE (T2/PD) | TSE | FSE | TSE | FSE / FASE |
| Single-shot FSE | HASTE | SSFSE | SShTSE | FASE |
| STIR (fat-null IR) | STIR / TIRM | STIR | STIR | STIR |
| FLAIR (CSF-null IR) | FLAIR | FLAIR | FLAIR | FLAIR |
| Spoiled (T1) GRE | FLASH | SPGR / FSPGR | T1-FFE | Fast FE |
| 3D volumetric T1 (IR-prep GRE) | MPRAGE | BRAVO / IR-FSPGR | 3D T1 TFE | 3D Fast FE |
| Balanced SSFP | TrueFISP | FIESTA | bFFE (bTFE) | True SSFP |
| GRE T2* / SWI | GRE · SWI | MERGE · SWAN | T2-FFE · SWIp | Field Echo · SWI |
| Diffusion EPI | DWI / RESOLVE | DWI / eDWI | DWI | DWI |
Names drift between software releases; treat this as a decoder, not gospel. When unsure, check the sequence’s parameters.
Quick decision guide
Clean T1/T2 anatomy or near metal? Spin echo (SE/FSE/TSE).
Robust fat or CSF suppression? STIR (fat) or FLAIR (CSF); STIR near metal.
Speed / breath-hold T1 / dynamic / 3D? Spoiled GRE; MPRAGE/BRAVO/TFE for 3D T1.
Bright fluid without gadolinium? Balanced SSFP (watch banding).
Microbleeds/iron/calcium? T2*-GRE / SWI.
Stroke/function? EPI (DWI/ADC, BOLD).
The four-way trade-off, restated
You can’t maximise SNR, resolution, scan time and contrast at once; pick three, pay in the fourth. Bigger voxels → SNR at cost of resolution; more averages → SNR at cost of time; lower BW → SNR at cost of chemical shift, distortion and longer minimum TE. Smaller voxels → resolution at cost of SNR/time. Long ETL/parallel imaging → speed at cost of blurring/g-factor noise. TR/TE/TI/flip set contrast, each with its own SNR & time cost.
Representative timing values are 1.5 T approximations and vary with field strength, vendor and protocol.
Sources
Key references: Westbrook & Talbot, MRI in Practice, 5th ed (Wiley, 2018). Elster, Questions and Answers in MRI (mriquestions.com). Bitar R et al., “MR Pulse Sequences: What Every Radiologist Wants to Know but Is Afraid to Ask,” RadioGraphics 2006;26:513–537. Cross-vendor terminology per Siemens MRI Acronyms card and MRImaster.