Spin Echo (SE) Sequences

Figure: Diagrams illustrating sequential (a) and segmental (b) filling of k-space. In sequential filling (such as for the spin echo sequence shown in fig. 2.4), one line of k-space, shown in bold, is acquired per $T_{R }$period. In segmental filling (b) several lines of k-space (in bold) are acquired in one $T_{R }$period by the use of multiple spin echos.

[Sequential] [Segmental]

As described earlier, spin echo sequences utilise a 180$^{\circ }$ RF pulse to rephase the dephasing spins and produce a detectable RF signal. There are three main types of spin echo sequence; single echo, multi-echo and echo-train spin echo. An example of a single echo sequence is shown in figure 2.4. RF excitation and slice selection are followed by phase encoding of the signal. The 180$^{\circ }$ rephasing pulse and slice selection gradient are applied and the echo is read-out at time $T_{E }$. The sequence is repeated using a different phase-encoding magnitude each repetition until k-space is filled, as shown in fig. 2.5(a). Multiple slices can be acquired during one $T_{R }$period using a multislice loop structure. Single echo sequences are typically used to produce $T_{1 }$-weighted images, using a short $T_{R }$($<$700ms) and short $T_{E }$($<$30ms) [5].

Multi-echo sequences use multiple 180$^{\circ }$ pulses to produce signals at different $T_{E }$
values, where the only differences between the signals at the two echo times will be due to the relaxation rates of the tissues. Multi-echo sequences are typically used to produce two simultaneous images: a proton-density weighted image where $T_{E }$$<$30ms and a $T_{2 }$-weighted image with $T_{E }$$>$80ms [5]. $T_{R }$is sufficiently great to allow (almost) complete $T_{1 }$relaxation.

Echo-train spin echo sequences are similar to the multi-echo sequences, except that each spin echo is acquired with a different phase encoding as well as its own $T_{E }$. The echo-train length (or turbo factor) is the number of echos acquired in each $T_{R }$period. Here, k-space is filled segmentally (see figure 2.5(a)), with one echo from each echo-train filling each segment of k-space, and as such is an efficient sequence. These sequences are typically used to create $T_{2 }$-weighted images, with a typical echo-train of 9 echos used in brain imaging.