Hi,
It is possible but certainly not easy and I have no experience
of trying to get this to work - I am only working on an idea.
You will have to write some matlab script (or similar) to
make an appropriate image, and then try to fill it in some
manner. Once you have something that resembles an
image mask then you can try to do the registration.
Hopefully that will work but I cannot guarantee anything.
All the best,
Mark
On 17 Aug 2010, at 10:16, Todor Iordanov wrote:
> Hi Mark,
>
> Thanks for the advices! Yes, I have a head shape file just with
> surface points in three columns (x, y, z) in BTI-coordinates and no
> image file. It's good to know that it is possible to create such an
> image file out of the ascii-file. How can I create a nifti-image
> file from that head shape file?
>
> Thanks again!
>
> Todor
>
> Mark Jenkinson schrieb:
>> Hi,
>>
>> It is common to register individual structural MR images to the MNI
>> template.
>> However, I'm guessing that this is not exactly what you have. If
>> it is then you
>> have a fairly standard problem and the FSL course practicals go
>> through
>> exactly this.
>>
>> If, instead, you do not have an image (your ASCII/digital file
>> format) or you
>> do have an image but it is only of the scalp surface, then this
>> will be much
>> more challenging. If it is the latter then it is probably easier
>> to create a
>> filled in mask of the head from your surface data (by manual
>> painting, or
>> possibly using some of the commands in fslmaths and connectedcomp,
>> depending on what the data looks like). Once you have the overall
>> mask
>> then you need to make a version of the MNI template that is also
>> like this
>> (and a simple thresholding and binarisation with fslmaths will do
>> this) and
>> then register as normal with FLIRT (although non-linear
>> registration with
>> FNIRT is unlikely to work). If you only have coordinates or some
>> other
>> information in your ASCII data then you'll need to figure out a way
>> of
>> turning this into an image like this before proceeding further.
>>
>> I hope this helps.
>> All the best,
>> Mark
>>
>>
>> On 13 Aug 2010, at 09:12, Todor Iordanov wrote:
>>
>>> Hi All!
>>>
>>> Is there any possibility to use FLIRT for registration of MNI-MRI
>>> data set with subject's individual head shape? We make our
>>> functional measurements in MEG and take a digitized head shape (in
>>> ASCII and digital file format) for each subject. The size of the
>>> individual heads is pretty different and it would be very helpful
>>> for me to be able to enlarge or to shrink the MNI-head in order to
>>> match the individual head shape.
>>>
>>> Thanks in advance!
>>>
>>> Kind regards
>>>
>>> Todor Iordanov
>>>
>>>
>>>
>>> --
>>> Dipl.-Math. Todor Iordanov
>>> Department of Psychology
>>> University of Konstanz
>>> Box 25
>>> 78457 Konstanz, GERMANY
>>> Phone: 00 49-75 31-88 30 86
>>> Fax: 00 49-75 31-88 46 01
>>> http://www.clinical-psychology.uni-konstanz.de/member-111.html
>>>
>>
>
>
> --
> Dipl.-Math. Todor Iordanov
> Department of Psychology
> University of Konstanz
> Box 25
> 78457 Konstanz, GERMANY
> Phone: 00 49-75 31-88 30 86
> Fax: 00 49-75 31-88 46 01
> http://www.clinical-psychology.uni-konstanz.de/member-111.html
>
|