To add to Matt's answer, the magnitude of the B0 geometric distortions is proportional to the total length of the EPI readout. So, for a given echo spacing, the fewer the number of phase-encoding (PE) lines, the smaller the distortions. Given the fact that the head is narrower in the RL direction, what we recommend our users is to set the PE direction to RL, and then decrease the % of the field of view in the PE direction. That way you can get the same brain coverage with less distortions (although with a little lower SNR). Of course, the images don't look symmetric, so some users don't like it...
For the same reason, we also recommend to choose the bandwidth that minimizes the echo spacing.
Best,
-Pablo
|