What you’ll be able to do after this lesson

  • Explain how a 180° pulse recovers signal that a gradient reversal cannot
  • State the single sentence that decides between the two families for a given clinical question
  • Describe what an echo train buys you in fast spin echo, and what it costs
  • Recognise from an image which family produced it

Two ways to make an echo

After a 90° pulse, transverse magnetisation dephases and the signal dies. To measure anything at a useful TE, you have to bring the spins back into phase — make an echo. There are exactly two ways to do it, and the difference between them is the most consequential fork in MRI.

Spin echo: a 180° radiofrequency pulse. At time TE/2 after excitation, transmit a 180° pulse. Every spin’s accumulated phase is flipped. The spins that were running fast — because they sit in a slightly high local field — are now behind, but they are still fast, so they catch up. At time TE, every spin is back in phase and a strong echo forms.

The spin echo

After a 90° pulse the spins precess in step, then fan out because each feels a slightly different field, and the signal fades. A 180° pulse at TE/2 flips the fan over: spins that had moved ahead are now behind, and they come back into step at TE, giving an echo.

Sketch of five runners spread out along a track, all running towards the same line, the fastest furthest from it
The analogy: runners spread out along a track turn round together and run back to the start line. The fastest went furthest and has furthest to come back, so all arrive at once. Its limit: in the scanner the 180° pulse flips the spins’ phase; it does not reverse the direction they precess. Each spin keeps precessing the same way at the same rate, now starting from behind.
Spin echo timing: 90 degree pulse at time zero, 180 degree pulse at TE/2, echo centred at TE. The free induction decay falls with T2 star; the echo peak lies on the dashed T2 decay curve, below the starting signal. RF 90° 180° signal time S0 0 TE/2 TE T2 decay, e−t/T2 FID echo peak on the T2 curve TE (echo time)
The physics, drawn by us: the 90° pulse starts the signal, which fades with T2* (the FID). The 180° pulse at TE/2 brings the spins back into step at TE. The echo peak sits on the dashed T2 curve, below the starting signal S0.

Why the echo follows T2, not T2*: the FID fades with T2*, which adds two kinds of dephasing: random spin-spin interactions (T2) and fixed field inhomogeneities, such as magnet imperfections and susceptibility differences (chemical shift is refocused too). The 180° pulse undoes only the fixed part; the random part (and diffusion) is not refocused. So the echo peak is S0e−TE/T2, and a longer TE gives a smaller echo.
TE = echo time, from the 90° pulse to the centre of the echo; the 180° pulse comes at TE/2.

Illustrative values, not tissue values: T2 = 200 ms, T2* = 20 ms, TE = 80 ms. Curves computed from S(t) = S0e−t/T2* before the 180° pulse and S(t) = S0e−t/T2e−|t−TE|/T2′ after it, with 1/T2* = 1/T2 + 1/T2′ (T2′ ≈ 22 ms, the fixed-inhomogeneity part). This assumes that part decays exponentially, the usual textbook simplification. Because T2 decay continues during the echo, the echo falls a little faster than it rises. The 180° pulse is drawn taller to show the larger flip angle; imaging gradients are left out. The spin echo was first described by Hahn (Physical Review, 1950); the 90°–180° form shown here was introduced by Carr and Purcell (Physical Review, 1954). The runners illustration is only the analogy; the labelled diagram carries the physics.

The elegance is that the 180° pulse does not need to know why a spin was fast. Any constant field offset is reversed: magnet imperfection, imperfect shim, the field distortion around a hip prosthesis, the susceptibility gradient at an air–tissue boundary. All of it cancels.

Gradient echo: a gradient reversal. Instead of an RF pulse, apply a magnetic field gradient that deliberately dephases the spins, then reverse its polarity. Spins that gained phase now lose it at the same rate, and at the reversal-matched time an echo forms.

This also works — but only for the dephasing that the gradient itself caused. Phase accumulated for any other reason keeps accumulating right through the reversal. Field inhomogeneity is untouched.

The one-sentence consequence

Spin echo images T2. Gradient echo images T2*.

Everything else in this lesson follows from that line.

Spin echo, by refocusing static inhomogeneity, sees only the genuinely random spin–spin dephasing. It is therefore robust wherever the field is imperfect: near metal, at the skull base, around the sinuses, at bowel gas, off-isocentre. Its contrast is clean and reproducible.

Gradient echo, having no refocusing pulse, decays at the T2* rate. That is a bug when you want undistorted anatomy near metal — you get a signal void larger than the implant plus geometric warping. It is a feature when the field distortion is the finding: haemosiderin, calcium, iron, deoxygenated blood. T2*-weighted GRE and its phase-enhanced extension SWI exist precisely to exploit blooming.

Two more differences fall out of the same physics:

Speed. A 180° pulse takes time and deposits RF energy. Gradient echo skips it, and typically pairs that with a small flip angle (lesson 07), which means very short TR is possible — milliseconds rather than hundreds of milliseconds. That is why every fast, breath-hold, dynamic and 3D volumetric T1 sequence in the body is a gradient echo.

Flip angle. Spin echo conventionally uses 90° excitation. Gradient echo uses whatever angle optimises the trade between signal and speed, and that angle becomes the dominant contrast lever in the family.

Fast spin echo: the echo train

Conventional spin echo collects one line of data per TR. With a TR of 4000 ms and 256 phase steps, that is a scan measured in tens of minutes. It is the reason early MRI was slow, and it is why conventional SE has been almost entirely replaced.

Fast (turbo) spin echo applies a train of 180° pulses after each excitation, forming multiple echoes and filling multiple lines of k-space per TR. The number of echoes is the echo train length (Siemens: turbo factor; GE: ETL; Philips: TSE factor). An ETL of 16 makes the scan roughly sixteen times faster.

The costs are real and visible:

  • T2 blurring. Each echo in the train has decayed a little more than the last, so the lines of k-space are filled with different signal amplitudes. That amplitude modulation across k-space acts as a filter and blurs edges — worse with longer trains and shorter tissue T2.
  • RF energy deposition rises with the number of 180° pulses, which is where a sequence typically runs into scanner limits. How that limit is managed is a safety matter and belongs with your local policy and the site’s safety guides.
  • Bright fat on T2. Fat is brighter on FSE/TSE T2-weighted images than on conventional spin echo. The accepted explanation is that the closely spaced refocusing pulses disrupt J-coupling between fat protons, reducing the dephasing that normally darkens fat. It is why fat suppression is routine on FSE T2 and was less essential in the conventional SE era.
  • The effective TE is the TE of the echo that fills the centre of k-space, not the first echo — which matters more than it sounds, and lesson 10 explains why.

Recognising them

  • Clean anatomy, sharp interfaces at the skull base or around a screw, no blooming → spin echo family.
  • Signal void much larger than the metal, geometric warping, dark rims at air–tissue boundaries → gradient echo.
  • Fluid and blood both very bright, thin bright-dark bands across the image → balanced SSFP, a gradient echo variant whose contrast tracks the T2/T1 ratio (Siemens TrueFISP, GE FIESTA, Philips balanced FFE) and whose banding artefacts come from off-resonance.

At the console

An MSK list includes a post-operative shoulder with a metallic anchor. The site protocol offers both a T2* gradient echo and a T2 TSE for the same plane, and the registrar wants the GRE because “it’s quicker”.

The decision is made in one line: is the metal the finding, or the obstacle? Here it is the obstacle — the question is rotator cuff integrity, and the anchor is in the way. Gradient echo will bloom the anchor across the tissue you need to see and will distort the geometry around it. TSE refocuses that inhomogeneity and gives you a smaller, truer void. For the same reason, spectral fat suppression will struggle near the anchor while STIR, which relies on T1 rather than on frequency, will hold up (lesson 06).

Invert the question and the answer inverts. A brain protocol for suspected cerebral amyloid angiopathy wants every microbleed it can find. Now the field distortion is the finding, so you deliberately choose the sequence that exaggerates it: T2*-weighted GRE or SWI, long TE, no refocusing pulse. The blooming that ruined the shoulder is the entire diagnostic yield here.

Same physics; opposite choice. That is what it means to understand a sequence family rather than memorise a protocol.

Check yourself

Explain, without using the words “spin echo” or “gradient echo”, why one sequence shows a 6 mm signal void around a 2 mm surgical clip and another shows a 3 mm void. Then: if fast spin echo is so much faster, why does any department still keep a conventional spin echo sequence available?

This is a private prompt. Nothing you write is stored or assessed.

Key takeaways

  • Spin echo forms its echo with a 180° pulse, which reverses any constant field offset. Gradient echo reverses only the dephasing its own gradient caused.
  • Spin echo therefore images T2 and resists field inhomogeneity; gradient echo images T2* and is sensitive to it.
  • Choose by asking whether the field distortion is the obstacle (spin echo) or the finding (gradient echo).
  • Fast spin echo fills several k-space lines per TR using an echo train. It buys speed and pays in T2 blurring, higher RF deposition, and brighter fat on T2.
  • Gradient echo’s freedom from the 180° pulse is what makes very short TR — and therefore breath-hold, dynamic and 3D volumetric imaging — possible.

References

  1. Hahn EL. Spin echoes. Physical Review. 1950;80(4):580–594 https://doi.org/10.1103/PhysRev.80.580
  2. Hennig J, Nauerth A, Friedburg H. RARE imaging: a fast imaging method for clinical MR. Magnetic Resonance in Medicine. 1986;3(6):823–833 https://doi.org/10.1002/mrm.1910030602
  3. Bernstein MA, King KF, Zhou XJ. Handbook of MRI Pulse Sequences. Elsevier Academic Press, 2004 — ch. 14 (SE), ch. 14.1 (GRE families)
  4. Bitar R, Leung G, Perng R, et al. MR pulse sequences: what every radiologist wants to know but is afraid to ask. RadioGraphics. 2006;26(2):513–537 https://doi.org/10.1148/rg.262055063
  5. Elster AD. Questions and Answers in MRI — ‘Spin echo vs gradient echo’ https://mriquestions.com/spin-echo-vs-gre.html
PreviousNext