(STIR) Also called Short Tau (t) (inversion time) Inversion Recovery. STIR is a fat suppression technique with an inversion time t = T1 ln2 where the signal of fat is zero (T1 is the spin lattice relaxation time of the component that should be suppressed). To distinguish two tissue components with this technique, the T1 values must be different. Fluid Attenuation Inversion Recovery (FLAIR) is a similar technique to suppress water.
Inversion recovery doubles the distance spins will recover, allowing more time for T1 differences. A 180° preparation pulse inverts the net magnetization to the negative longitudinal magnetization prior to the 90° excitation pulse. This specialized application of the inversion recovery sequence set the inversion time (t) of the sequence at 0.69 times the T1 of fat. The T1 of fat at 1.5 Tesla is approximately 250 with a null point of 170 ms while at 0.5 Tesla its 215 with a 148 ms null point. At the moment of excitation, about 120 to 170 ms after the 180° inversion pulse (depending of the magnetic field) the magnetization of the fat signal has just risen to zero from its original, negative, value and no fat signal is available to be flipped into the transverse plane.
When deciding on the optimal T1 time, factors to be considered include not only the main field strength, but also the tissue to be suppressed and the anatomy. In comparison to a conventional spin echo where tissues with a short T1 are bright due to faster recovery, fat signal is reversed or darkened. Because body fluids have both a long T1 and a long T2, it is evident that STIR offers the possibility of extremely sensitive detection of body fluid. This is of course, only true for stationary fluid such as edema, as the MRI signal of flowing fluids is governed by other factors.

See also Fat Suppression and Inversion Recovery Sequence.

Pulse sequences, designed to be insensitive to flow, e.g. at every even echo, a spin echo sequence is not flow sensitive. Velocity compensation is achieved by using gradients, which are either symmetrical around a 180° pulse and switched on twice as is the case for motion compensated spin echo pulse sequences, or two antisymmetrical gradient lobes without 180° pulse, which is the way to produce a velocity compensated gradient echo pulse sequence.
The signal of the second echo (and all other even echoes) is independent of the velocity of the object. Thus, velocity-based motion effects stemming from the entire voxel or from spins within a voxel (intravoxel incoherent motion) are suppressed with such pulse sequences.
If higher order motion is relevant, as it may be in turbulent jets across valves, acceleration and jerk effects can also be compensated for by the use of appropriate combinations of gradient- and radio frequency pulses.
With the increasingly stronger gradients, echo times in MR systems can be shortened to the point at which effects other than velocity effects hardly ever become relevant.