Dear FSL Experts and Community,
I have a question about susceptibility distortion correction in fMRI
using opposite phase encoded EPIs (not the "conventional" double echo
method). I am basing myself largely on this description:
https://lcni.uoregon.edu/kb-articles/kb-0003
I apply TOPUP to the opposite encoded images: inputAP (which has the
same phase encoding as our fMRI sequence of interest) and inputPA (which
has opposite phase encoding). I need to provide a data file which tells
TOPUP for each scan what the phase encoding direction is and the total
readout time. For example, the file could have these contents:
0 -1 0 .0351
0 1 0 .0351
(1) Am I right to assume that I should give the phase encode directions
in "EPI space" and not in "Scanner space"? What I mean is that the EPI
images themselves are acquired in an oblique orientation (tilted 22.6
degrees from transverse towards coronal), so I think the direction of
the phase encoding gradients in scanner space are some oblique vectors
like [0 -cos(22.6) -sin(22.6)] for the AP scan and [0 cos(22.6)
sin(22.6)] for the PA scan. However, in the coordinate system of the
EPI images, then I think the phase encoding gradients would be: [0 -1 0]
for the AP scan and [0 1 0] for the PA scan.
(2) TOPUP then outputs a fieldmap as well as "corrected" versions of the
inputAP and inputPA images. However, to unwarp my actual fMRI sequence
of interest, I need to specify the phase unwarp direction (e.g. y, -y or
x etc.). Given that my inputAP image has [0 -1 0] phase encoding
direction does this mean that the unwarp direction should be "y"? Or is
it not that straight-forward?
Also on a more conceptual level I am wondering why this additional piece
of information is required because it was able to "correct" the inputAP
image in the TOPUP step without needing this?
Thanks for any answers/suggestions/help.
Best wishes!
Floris
--
Floris van Vugt, Ph.D.
Motor Control Lab, Psychology Dept.
McGill University
1205 Dr. Penfield Ave
Stewart Biology Building
Montreal QC Canada H3A 1B1
www.florisvanvugt.com
|