Dear Ruth,
> Thank you Jesper,
> So I see that, if I do my own motion correction I will want to give EDDY the matching fieldmap. But what if I provide dwi data that has been corrected for susceptibility but NOT motion. Does EDDY still need the fieldmap?
If you provide eddy with data that has been corrected for susceptibility but NOT motion there is a big risk that it has not been corrected properly for susceptibility. Imagine a fieldmap acquired with the subject in location A and further imagine that your DWI data was acquired shortly after with the subject still in location A at the start of acquisition. Also imagine that at some point during the DWI acquisition the subject moves to location B, for example by a 10mm translation in the y-direction.
When you now attempt to correct your DWI data for susceptibility using you fieldmap you will be fine up to the point of that movement. After that movement the areas of large distortion will no longer line up with the areas of large off-resonance in the fieldmap, and your correction will not work.
This all, of course, depends on how much your subjects move but in general I think that all of these corrections need to be performed simultaneously as they all affect each other. That is what you get if you supply a fieldmap to eddy and let eddy do everything.
For this to work it is crucial that the fieldmap is well aligned with the first volume of your DWI data set (and with DWI I mean the mixture of b0:s and dwi:s that constitute your diffusion data, so it is conceivable that the first volume is actually a b0).
Jesper
>
> --Ruth
|