Hi Again,
I forgot to mention in my last email that the same
recommendation that we have for human imaging
applies here. That is, if you are taking *partial brain*
EPI images then you should take a single *whole brain*
EPI image as well, to help the registration. This image
should have the same parameters as your partial brain
images, but just more slices in order to cover the whole
brain. It typically only takes seconds of scanner time
and so is not costly for the acquisition, but makes a
*huge* difference to the robustness and accuracy of
the registration, since dealing with partial brain EPI
images is one of the hardest registration problems.
If you did have such a scan then you should use it.
If you don't, then try and include it in your future scanning
sessions.
All the best,
Mark
On 23 Feb 2011, at 09:54, Mark Jenkinson wrote:
> Hi,
>
> If you have:
> - partial brain coverage
> - substantial, uncorrected, distortions
> - noticeable bias field
> - images far from alignment initially
> then you have given yourself a *very* difficult registration problem.
>
> How hard did you try to get FUGUE to work?
> Normally it is possible to get good results with a bit of effort,
> especially if you take care with the brain extraction.
>
> It is worth trying Wolf's suggestion, but the partial brain coverage
> will make this quite tricky.
>
> If you cannot get this to work then feel free to upload the images to our
> upload site: http://www.fmrib.ox.ac.uk/cgi-bin/upload.cgi
> and send us the reference number. We might be able to find a solution
> to these problems, but it probably depends on what is the "worst"
> thing from the list above in your images.
>
> All the best,
> Mark
>
>
> On 23 Feb 2011, at 09:43, Van Lee wrote:
>
>> Hi,
>>
>> Thanks for your suggestion. I tried the -nosearch option but it gives worse results, maybe because those 2 images are not roughly aligned.
>>
>> I think it might fail because of the distortion in the EPI image. We did acquire the field map before but after FUGUE does not work so well for us, we decided to drop it. And there is a bias field in the EPI image, but I don't know how to correct it.
>>
>> We only acquire the partial volume of the EPI. Does it also cause problem to the registration?
>>
>> I will try zinke's method. Thanks!
>>
>> Hope you can provide me some guidance
>>
>> Thanks a lot
>>
>> Van L
>>
>
|