Hi,
At the moment the only option for B0 unwarping through the GUI is to use the BBR cost function.
If you are not getting good results with that then the first thing to figure out would be why this is the case. Do you have unusual images in any way? Is the segmentation of your structural image sufficiently good? Or is there very little gray-white contrast in your EPI images?
It would generally be better to get the BBR working, as we find that it gives much better results, but if you cannot for some reason then it would be possible to script everything using flirt and fugue, but it is not that simple. Let us know if this is something that you are likely to need to do.
Don't worry about seeing the skull in the webpage images if it is in the right place. You can always check the individual nifti images in the *.feat/reg and *.feat/reg/unwarp directories if you want to be sure that the brain extracted versions have the skull removed correctly.
All the best,
Mark
> On 14 Feb 2017, at 21:09, Gabriella <[log in to unmask]> wrote:
>
> Hello,
>
> I'm in the stage of preprocessing some functional data I've acquired using the FEAT GUI but noticed that when the B0 unwarping button is clicked/selected, I am unable to change the type of registration for my anatomical input (for the main structural image). The default and only option is BBR, however, I am running into issues with registration and would like to try using 12 DOF instead.
>
> Is there a way around this (possibly by running from command line)?
>
> Also, is it normal to see skull in the unwarping report log (as the fieldmap magnitude image in the first row)? I have ensured that the fieldmap mag image inputted in the GUI is the one that had BET run on it (and made sure the original file could be found). However, I'm under the impression that skull should not be visible in the fieldmap image based on the FSL course examples...
>
> Any help would be greatly appreciated!
>
> Many thanks,
> Gabriella
|