In the spin echo, a 180° pulse rephased the spins and handed us the true T2. What if we leave that pulse out? Then we have a **gradient echo**, and it behaves differently in one important way.

Without the 180°, we rephase the spins by briefly reversing a magnetic field gradient instead. That works — but a gradient reversal only undoes the dephasing the gradient itself caused. It does nothing about the fixed field inhomogeneity. So a gradient-echo signal decays by T2*, not T2, exactly the faster decay from the last post.

That single fact drives most of what you notice clinically:

– **Susceptibility sensitivity.** Anything that distorts the local field —

  hemorrhage/blood products, calcium, metal, air-tissue borders — dephases spins

  and shows up as signal loss that “blooms” larger than the object. This is the basis of sequences like susceptibility-weighted imaging (SWI)¹ and is why gradient echo is your friend for hunting microbleeds.

– Speed. Skipping the 180° and (as we’ll see next) using small flip angles lets gradient echo run with very short TRs, which is why so much fast and 3-D imaging is gradient echo.

– The cost. Near metal or the skull base, that same T2* sensitivity turns into distortion and dropout.

At the PACS station, pull up a GRE or SWI series and look at how vessels and any hemorrhage appear thicker and darker than on the spin-echo images — that is T2* at work.

Gradient echo usually tips the magnetization by less than 90°. Why would you use a small flip angle, and how small? Next post.

1. Invented by E. Mark Haacke: DOI: 10.1002/mrm.20198