e speed, flexibility in contrast, a more SE like T2 contrast (compared to FSE), better slice efficiency (that is, more slices per TR), and can be flexible with respect to resolution by using segmentation. As you would have gathered by constructing the table in question 3, speed is of course the main advantage, and opens up the area of functional rather than anatomical imaging.
All sequences must be fat suppressed due to chemical shift, and the presence of geometric distortions are the two big potential problems. Obviously if you want to image or measure fat, then EPI is not the sequence for you. Also if the patient has braces and you want to image their brain with EPI that is also not going to work - either due to susceptibility distortions or B1 in homogeneities, depending upon what the braces are made of. Also there are some areas where the susceptibility is so great that no degree of segmentation will completely remove the distortion - like the areas at the base of the brain close to the sinuses.
3. Constant phase encoding EPI: to obtain evenly spaced points in ky, data is split into two, 1D FT at each kx, phase shifted to a grid, 2nd FT at each ky, both halves added together applying the Fourier Transform Shift theorem.
4. Spiral scanning methods (square and circular): Points in k space are also not collected uniformly in time (that is, in the line by line method we are familiar with). The square method is, however, evenly spaced in k-space, therefore just needs reordering. Circular spiral scanning points are separated uniformly in RADIAL space, but not in the 2D space we are used to. Either a non-Fourier reconstruction is used (that means you dont need points on a 2D grid) or the data needs to be interpolated to fit a grid.
Badwidth is inversely propotional to the sampling line. The number of Pixels reslting from a shift in phase error is dependent upon the phase per pixel of bandwidth. The change in frequency gives a rise